The Effect of Bean Consumption on Urinary and Serum Metabolites in Patients with Peripheral
Artery Disease (PAD): A Non-Targeted Metabolomics Approach
by
Le Wang
A Thesis submitted to the Faculty of Graduate Studies of
The University of Manitoba
in partial fulfilment of the requirements of the degree of
MASTERS OF SCIENCE
Department of Human Nutritional Science
University of Manitoba
Winnipeg
Copyright © 2016 by Le Wang
ABSTRACT
Non-targeted liquid chromatography-mass spectrometry (LC-MS) based metabolomics methods
were used to determine whether bean consumption affects the profile of metabolites in serum and
urine of individuals with peripheral artery disease. Urine and serum collected at baseline or after
8 weeks of consuming 1.5 or 3 cups per week of cooked mixed beans (pinto, red kidney, black
and navy) or pulse-free foods, was extracted with acetonitrile and analyzed by non-targeted LC-
QTOF-MS methods.
Several compounds found in beans were present in the biological fluids indicating that these
compounds may be useful markers of bean consumption. Bean consumption significantly
(P<0.05) altered several endogenous metabolites including amino acids, glutathione, bile salts,
phospholipids and products of arachidonic acid metabolism by cyclooxygenase. The decrease in
the metabolites of an anti-hypertensive drug in urine after 8 weeks of bean consumption
suggested the existence of potential drug-diet interactions that could affect the required dosage of certain anti-hypertensive medications.
I
ACKNOWLEDGMENTS
I sincerely thank my supervisor, Dr. Michel Aliani, and my co-advisor, Dr. Carla Taylor. They
gave me valuable suggestions, taught me how to analyze, organize and write, as well as offered
me warmest help. I appreciate their patience and contributions in the past three years. I realize
that I could not continue to produce this thesis without them. Thank you!
In the meantime, to my committee members, Dr. Peter Zahradka and Dr. Chris Siow, I thank you very much for your encouragement, critical comments and positive feedback throughout our meetings.
Thank you to Drs. Zahradka and Taylor for providing me with the samples (urine, serum, and
cooked beans) from their clinical trial (ClinicalTrials.gov Identifier: NCT01382056) that were
analyzed for my thesis project.
I really enjoy working in Dr. Aliani’s lab because of the nicest colleagues. Especially, Mrs. Shiva
Shariati Ievari, who taught me how to perform lab work and has become my best friend.
I also thank and appreciate receiving the Graduate Students` Association Conference Grant and
the Faculty of Graduate Studies Travel Award.
Last, I sincerely thank my parents for your support as I could not successfully complete my
Master`s program without you. I always love you so much.
II
TABLE OF CONTENTS
ABSTRACT...... I
ACKNOWLEDGMENTS...... II
TABLE OF CONTENTS...... III
LIST OF FIGURES...... VI
LIST OF TABLES...... VII
LIST OF ABBREVIATIONS...... VIII
Chapter 1. Introduction ...... 1
Chapter 2. Literature Review ...... 5
2.1. Peripheral Artery Disease...... 5
2.1.1. Causes and risk factors...... 6
2.1.2. Medical management...... 8
2.1.3. Nutritional interventions...... 11
2.1.4. Food-drug interactions...... 13
2.2. Beans...... 16
2.2.1. Compositions of beans ...... 16
2.2.2. Nutrients and health ...... 19
2.2.3. Beneficial effects of beans on diseases conditions...... 27
2.3. Metabolomics...... 29
2.3.1. Metabolomics in disease-related studies...... 30
2.3.2. Metabolomics in drug studies...... 32
2.3.3. Metabolomics in food and nutrition studies...... 33
Chapter 3. Hypotheses and Objectives…………………………………………………..……36
III
3.1. Hypotheses...... 36
3.2. Objectives...... 36
Chapter 4. Materials and Methods...... 38
4.1. Chemicals and standards…………………………………………………………….……. 38
4.2. Bio-fluids and diet materials………………………………………………………….…... 38
4.3. Sample preparation for metabolomics studies………………………………….………….41
4.3.1. Urine……………………………………………………………………………….…...41
4.3.2. Serum……………………………………………………………………………….…..41
4.3.3. Beans…………………………………………………………………………………... 42
4.4. HPLC-QTOF-MS…………………………………………………………………...……. 42
4.4.1. LC conditions for metabolites extraction……………………………………………… 43
4.4.2. QTOF-MS conditions for metabolites extraction………………………………...…… 43
4.5. Data processing…………………………………………………………………………… 44
4.6. Statistical analyses…………………………………………………………………...…… 45
4.7. Identification of metabolites…………………………………………………………...…. 46
4.8. Pathway analyses……………………………………………………………………...….. 46
Chapter 5. Results ……………….……………………………………………………………. 47
5.1. General results……………………………………………………………………..……... 47
5.1.1. Urinary metabolites……………………………………………………………………. 47
5.1.2. Serum metabolites……………………………………………………………………... 50
5.2. Significant results ………………………………………………………..…………..…… 53
5.2.1. Urinary metabolites with statistical significance…………………………………...…. 53
5.2.2. Serum metabolites with statistical significance…………………………………...…... 58
IV
5.3. Administrated drugs and their metabolites………………………………………..……… 62
5.3.1. Detected metabolites of drugs…………………………………………………..…….. 62
5.3.2. Metoprolol………………………………………………………………………..…… 68
5.4. Beans and rice extracts …………………………………………………………..………. 69
Chapter 6. Discussion……………………………………………………………………..……78
6.1. Endogenous metabolites……………………………………………………………..…….78
6.2. Exogenous metabolites (administered drugs) ……………………………………….…….83
6.3. Compounds in beans ………………………………………………………………………85
Chapter 7. General Discussion……………………………………………………………..…. 88
Chapter 8. Conclusions………………………………………………………………….…….. 94
References ………………………………………………………………………………..……. 98
Appendix I...... 124
Appendix II. ………………………………………………………………………………...... 129
Appendix III. ………………………………………………………………………………… 133
V
LIST OF FIGURES
Figure 1. Metabolomics workflow. …………………………………………………………...... 44
Figure 2. Partial least square discrimination (PLSD) of urine samples of PAD patients. …...... 48
Figure 3. Lorenz Curves generated for different treatments used in the PLSD model for urine.. 49
Figure 4. Partial least square discrimination (PLSD) of serum samples of PAD patients...…..... 51
Figure 5. Lorenz Curves generated for different treatments used in the PLSD model for serum.52
Figure 6. Heatmap of 9 selected significant (P<0.05) urinary metabolites affected by bean
consumption. …………………………………………………………………………...…. 55
Figure 7. Heatmap of 15 selected significant (P<0.05) serum metabolites affected by bean
consumption. …………………………………………………………………………..…...59
Figure 8. α-Hydroxymetoprolol in urine with 0.6 cups of beans consumption. ………………...68
Figure 9. PLSD of beans and rice extracts. …………………………………………………...... 70
Figure 10. Lorenz Curves generated for different treatments (bean and rice extracts) used
in the PLSD model.…….…………………………………………………………………... 71
Figure 11. Venn diagram for the 4 types of beans..……….………….…………………………. 72
Figure 12. Heatmap of 125 compounds with significant differences among 4 types of beans
and rice extracts.……….…..……………………………………….……………………… 73
Figure 13. Structures of PGF2α, PGE2 and their derivatives. ………..………………………... 81
Figure 14. Structures of metoprolol and its 3 metabolites: α-hydroxymetoprolol,
o-demethylmetoprolol and metroprolol acid…...………………………………………….. 84
VI
LIST OF TABLES
Table 1. Selected nutrient composition of four bean types….. ………………………...…….… 17
Table 2. Flavonoids in selected common beans. …………….…………………….……….…... 18
Table 3. Phytosterols, saponins and phytates in selected types of beans. ….…….………….…. 18
Table 4. Selected dietary phytochemicals and their effects. ………………..…………….……. 24
Table 5. Metabolomics in disease-related studies…………………………..……………………32
Table 6. Metabolomics in drug-related studies……………………………..……………………33
Table 7. Recruitment criteria for PAD participants. ………………………..………………..…. 39
Table 8. Sample preparation and weights of beans and cooking water. …….………………..... 40
Table 9. Workflow of metabolomics and statistical analyses of urine of PAD patients….……...47
Table 10. Workflow of metabolomics and statistical analyses of serum of PAD patients…….…50
Table 11. Urinary metabolites significantly modified by bean consumption and their associated
pathways…………………………………………………………………………………….56
Table 12. Serum metabolites significantly modified by bean consumption and their associated
pathways…………………………………………………………………………………….60
Table 13. The list of medications administered to participants in the “Bean-PAD study”. ….… 62
Table 14. Administered drugs and their corresponding metabolites detected in urine and
serum of 62 PAD participants. …………..………………………………..….……….…… 64
Table 15. Workflow of metabolomics and statistical analyses of bean extracts………....………69
Table 16. Compounds found in black, kidney, navy and pinto beans. …...………..………….... 72
Table 17. The distribution of detected entities in beans, rice and biological extracts of PAD
patients. ………………………………………………………………………………….… 75
Table 18. Pathways of 11 significant compounds in beans and bio-fluids of PAD patients…...... 77
VII
LIST OF ABBREVIATIONS
25-(OH)-D 25-hydroxyvitamin D
ABI Ankle-brachial index
ACEI Angiotensin converting enzyme inhibitor
ALA Alpha-linolenic acid
ARB Alpha-receptor blocker
CAMs Cell adhesion molecules
DB DrugBank database
DBP Diastolic blood pressure
HbA1c Hemoglobin A1c
HCTZ Hydrochlorothiazide
HMDB Human metabolome database
HMG-CoA 3-hydroxy-3-methyl-glutary CoA
HPLC High-performance liquid chromatography-mass spectrometry
IC Intermittent claudication
IOP Intraocular pressure
KEGG Kyoto encyclopedia of genes and genomes
LC-MS Liquid chromatography–mass spectrometry
LC-QTOF-MS Liquid chromatography-quadrupole time-of-flight-mass spectrometry
LDL Low-density lipoprotein
LysoPE Lysophosphatidylethanolamine
MHQ Mass hunter qualitative
MI Myocardial infarction
VIII
MPP Mass profiler professional
MS Mass spectrometry
MUM Main urinary metabolite
NMR Nuclear magnetic resonance
NO Nitric oxide
PAD Peripheral artery disease
PG Prostaglandin
PLSD Partial least square discrimination
ROS Reactive oxygen species
SBP Systolic blood pressure
SHR Hypertensive rats
VLDL Very low-density lipoprotein
IX
Chapter 1. Introduction
Peripheral artery disease (PAD) is systemic atherosclerosis present in the legs (Lau et al.,
2011). PAD is diagnosed when the ankle-brachial pressure (ABI) is ≤0.9. Individuals with PAD are at increased risk for chronic cardiovascular diseases such as myocardial infarction (MI) and stroke and they may experience intermittent claudication (IC) in the legs.
Every decade there is an approximately 25% increase in the worldwide prevalence of
PAD (Fowkes et al., 2013). Furthermore, approximately 20% of PAD patients suffer from MI and stroke, and 15-30% of PAD patients die from cardiovascular diseases every half decade (Lau et al., 2011).
Age is considered as a risk factor in PAD patients, with higher morbidity and mortality incidences found in PAD patients aged 40 or higher, and especially after 70 years old
(Fowkes et al., 2013). Other risk factors for developing PAD include diabetes mellitus, smoking, hypertension, hyperlipidemia and hyperhomocysteinemia (Lau et al., 2011;
Hiatt, 2001).
Four main categories of medications are generally administered to PAD individuals.
These include cholesterol-lowering drugs (statins), anti-hypertensive agents (beta- blockers, angiotensin converting enzyme inhibitors (ACEI), α-receptor blockers (ARB)), anti-platelet agents (acetylsalicyclic acid) and anti-hyperglycemic medications
(metformin) (Hirsch et al., 2001). In addition to these medications, catheter-based endovascular interventions such as angioplasty and surgery are also considered for PAD patients with severe conditions.
- 1 -
Exercise and nutrition are of interest as they have potential for lowering the risk factors associated with PAD. For example, flaxseed is a good source of fibers, lignans and alpha- linolenic acid (ALA) and it has been shown in a recent study (Rodriguez-Leyva et al.,
2013) to help PAD patients with high blood pressure. Common beans are also good sources of fiber, protein and necessary minerals as well as phytochemicals such as flavonoids with beneficial effects on human health (Kadouh and Zhou, 2012). Bean consumption has been shown to reduce blood lipids levels in diabetic rats
(Venkateswaran et al., 2002), to control blood glucose concentrations in individuals with type 2 diabetes (Rizkalla et al., 2002) and to decrease the proliferation of tumor cells in cancer patients (Wong et al., 2010). Furthermore, a mixed pulse diet containing beans has been shown to improve blood flow to the legs of individuals with PAD (Zahradka et al.,
2013).
However, in most nutritional studies where participants are patients with severe health conditions, a variety of drugs is also administered; thus, the potential benefits may be due to the action of either the bioactive compounds present in the selected foods and/or an interaction between bioactive compounds and the administered drugs. Food-drug interactions can involve in nutrients and non-nutrient components including bioactive compounds that influence drug absorption, metabolism and excretion (Pronsky and
Crowe, 2012). Investigation of all administered drugs, their metabolites and all potential bioactive compounds present in foods requires robust and reliable analytical techniques.
Metabolomics can provide a unique solution to study hundreds of compounds in a single run and as such can be used as an analytical platform to investigate potential interactions between drugs and food components in biological fluids. Metabolomics is a rapid,
- 2 -
accurate and sensitive approach to investigate both metabolic changes and the interactions between food and drugs. Comprehensive and systematic analysis via metabolomics aims at identifying and quantifying small molecule metabolites in biological samples at a given time point (Bujak et al., 2015). Profiling techniques including nuclear magnetic resonance (NMR) spectroscopy and mass spectrometry (MS)- based methods can be applied for the separation and identification of endogenous and exogenous metabolites.
Metabolomics is a useful method with applications in foods (foodomics), drugs (pharma- metabolomics) and disease studies. For examples, metabolomics was successfully applied to food interventions with fruits, vegetables, seafood, beans and rice (Andersen et al.,
2014; Mannina et al., 2012; Mensack et al., 2012; Kusano et al., 2015); pharmaceutical studies with anti-hypertensive drugs (beta-blockers, ACEI), cholesterol-lowering drugs
(statins, fibrates) and other agents (Altmaier et al., 2014); as well as diseases states including diabetes, hypertension, coronary artery disease and PAD (Dudzik et al., 2014; van Deventer et al., 2015; Turer et al., 2009; Huang et al., 2013).
The main objective of this work was to use a liquid chromatography–mass spectrometry
(LC-MS) based metabolomics approach to investigate the effects of 8 weeks of bean consumption on urinary and serum endogenous and exogenous compounds in individuals with PAD. Urine and serum samples were obtained at baseline and 8 weeks from PAD patients (n=62) randomly assigned to 1 of 3 groups (n=22, 19, 21; intake of one study food per day, 5 days/week and for 8 weeks): i. Pulse-free foods (rice instead of beans = control); ii. Study foods containing 0.3 cup and iii. 0.6 cup of mixed cooked beans (navy, pinto, red kidney, black). The PAD patients were under the care of their physician and
- 3 -
were taking multiple drugs such as cholesterol-lowering and anti-hypertensive agents.
Thus, in this thesis, a metabolomics study was performed using the baseline and 8 week urine and serum samples to investigate the changes in metabolites (endogenous and exogenous compounds), and to determine potential interactions among bean consumption and medications in individuals with PAD. The use of a non-targeted metabolomics approach as an invaluable screening tool to monitor both endogenous and exogenous compounds in urine and serum is discussed in the present thesis. This approach provided insight into biochemical pathways affected by diet, and helped to generate new hypotheses for mechanism of action.
- 4 -
Chapter 2. Literature Review
2.1. Peripheral Artery Disease (PAD)
PAD is a typical symptom of systemic atherosclerosis and characterized by atherosclerotic occlusion of arteries, especially in the lower extremities including iliac, femoral, popliteal and tibial arteries (Lau et al., 2011; Hiatt, 2001). IC, a major manifestation of PAD, is cramping in the muscles of the legs. This produces pain in the lower legs when walking and is relieved by rest (Hiatt, 2001; Beebe, 2001). With the development of PAD, most IC patients will undergo worsening of claudication that may proceed to serious leg ischemia, gangrene and amputation (Hiatt, 2001). Apart from IC, other symptoms include numbness, weakness, coldness, sores, and a weak pulse (Mayo
Clinic, 2015).
However, over half of PAD patients are asymptomatic. The gold standard for diagnosis of PAD is the ankle-brachial index (ABI), the ratio of the systolic blood pressure (SBP) in legs (the dorsalis pedis or posterior tibial arteries) to the SBP in arms (the brachial artery) (Lau et al., 2011). ABI values less than 0.9 (Hiatt, 2001) indicate diagnosis of
PAD. In addition, diseased arterial segments localized by segmental limb pressures and pulse-volume recordings and challenged with exercise treadmill testing may have a 20 mmHg drop in SBP in individuals with PAD. Imaging methods, including duplex ultrasonography, computed tomographic and magnetic resonance angiographies, illustrating >50% stenosis, have been used to identify the severity of PAD (Lau et al.,
2011).
The prevalence of PAD is higher in age groups >40 years old (Fowkes et al., 2013).
Especially after 70 years of age, the morbidity (Lau et al., 2011) and mortality (Sampson
- 5 -
et al., 2014) of PAD significantly increase. Furthermore, there is a trend for a higher
incidence of PAD in younger groups (<40 years old) but the evidence is scarce (Fowkes
et al., 2013).
2.1.1. Causes and risk factors
Atherosclerosis, a main cause of PAD (Mayo Clinic, 2015), is a chronic inflammation
located in the arterial wall triggered by low-density lipoprotein (LDL) cholesterol
(Raymond and Couch, 2012). The changes in function and morphology of endothelial
cells, platelets, and leukocytes are crucial to the progression and development of
atherosclerosis; furthermore, reactive oxygen species (ROS) induce damage in
endothelial cells by inhibiting the release of nitric oxide (NO), and platelets promote the
formation of atherosclerosis by stimulating the release of cellular adhesion molecules
(Carnevale et al., 2014).
Fibrous and advanced plaques develop from fatty streaks and contribute to the formation
of thrombi, thereby leading to blockages and restricted blood flow. This is the underlying
cause of many forms of cardiovascular disease including coronary arterial diseases
(angina, myocardial infarction), cerebral arterial diseases (strokes, transient ischemic
attacks), and peripheral arterial diseases (IC, limb ischemia, gangrene) (Raymond and
Couch, 2012).
Various inflammatory markers are associated with vascular disease. Biomarkers of
inflammation that have been investigated include C-reactive protein (Scirica and
Morrow, 2006), cytokines (interleukin-6, interleukin-8) (Kaptoge et al., 2014), fibrinogen
(Bartolucci and Howard, 2006), homocysteine (Towfighi et al., 2010) and lipoprotein- associated phospholipase A2 (Berger et al., 2011). In addition, Mlekusch et al. (2004)
- 6 -
reported that serum creatinine, which is increased in end-stage renal disease, can predict the death rate in patients with PAD independent of hypertension and diabetes.
In general, the risk factors of PAD include advanced age (over 40 years), diabetes mellitus, smoking, hypertension, hyperlipidemia and hyperhomocysteinemia (Lau et al.,
2011; Wood and Hiatt, 2001). Among those risk factors, smoking is the most influential element for promoting PAD and has an effect on other factors which are associated with
PAD (Beebe, 2001). In recent studies, considerable attention was given to diabetes and hypertension, two important risk factors in development of PAD.
PAD is commonly accompanied with coronary artery disease, MI, and ischemic stroke; these usually lead to limited organ function and poor quality life (Lau et al., 2011; Ix et al., 2011; Hirsch et al., 2001; Hiatt, 2001) causing a high rate of morbidity and mortality in PAD patients.
Between 2000 and 2010, the prevalence of PAD increased by 28.7% in low or middle income countries and 13.1% in high income countries due to advanced age, especially in women and those who smoke (Fowkes et al., 2013). In the meantime, mortality and disability of PAD patients grew dramatically and was associated with older age, female gender, and developing regions of the world (Sampson et al., 2014).
Annually, more than 15% of Americans suffer from PAD and the number increases dramatically with age (Ix et al., 2011). In Canada, the increase in PAD patients with age was shown to be as follows: approximately 4% after 40 years old, 10-20% after 55 years old and 20% after 75 years old (Gakhar, 2013). Patients with PAD had an equal chance of death from cardiovascular disease regardless of whether they had a history of coronary artery disease or not (Lau et al., 2011; Hiatt, 2001). Furthermore, among PAD patients
- 7 -
including the asymptomatic patients, both women and men had approximately an equal
chance of death (Hiatt, 2001). Every 5 years, approximately 20% of PAD patients had a
MI or stroke, and 15-30% of PAD patients died from cardiovascular diseases (Lau et al.,
2011). According to Statistics Canada (2015), in 2012, there were 66,952 individuals who
died of circulatory diseases, a quarter of the all-cause deaths.
To date, several therapeutic strategies have been applied to reduce the morbidity and mortality of patients with PAD and they are briefly discussed in the next section.
2.1.2. Medical management
Cholesterol-lowering and anti-hypertensive medications are prescribed to manage blood lipids and blood pressure, respectively.
Statins, niacin, fibrates and bile acid binding resin agents are cholesterol-lowering drugs widely used to reduce the blood total cholesterol concentration to less than 240 mg/dL
(6.2 mmol/L), and decrease the LDL-cholesterol concentrations to less than 160 mg/dL
(4.1 mmol/L) in PAD patients (Hirsch et al., 2001). According to Weinberg et al. (2011),
LDL-cholesterol should be less than 70 mg/dL (1.81 mmol/L) in PAD patients with severe ischemia. In a drug-claudication interaction study where 354 PAD patients with classic symptoms of claudication were randomly assigned to 3 groups receiving either a placebo or atorvastatin (10 mg or 80 mg per day for 1 year) (Mohler et al., 2003), both total cholesterol and LDL-cholesterol were significantly decreased to less than 160 mg/dL (4.14 mmol/L) in atorvastatin groups with no significance changes in ABI and maximal walking time indicating that cholesterol-lowering may not be linked to improvements in vascular function. However, concentrations of total cholesterol and
- 8 -
LDL-cholesterol were much lower in the 80 mg dose group and they had a significantly
increased pain-free walking time compared to the 10 mg dose group.
Different varieties of antihypertensive medications such as ACEI (e.g. ramipril), diuretic
(e.g. thiazides), beta-blocker (e.g. metoprolol), and ARB (e.g. irbesartan) are commonly used to decrease systolic blood pressure to less than 140 mmHg and diastolic blood pressure to less than 90 mmHg in individuals with PAD (Lane and Lip, 2013; Weinberg et al., 2011). A study with 212 claudication patients with PAD who were randomly and equally allocated to two groups to either receive placebo or 10 mg of ramipril per day for
6 months (Ahimastos et al., 2013) showed a significant decline (systolic: -3.1 ±0.6 mmHg; diastolic: -4.3 ±0.9 mmHg) in the blood pressure to average values of 135.9/77.7 mmHg in the ramipril group; furthermore, both pain-free walking time and maximal walking time significantly increased by 75 and 255 seconds, and the ABI significantly increased by 0.1. A review by The Cochrane Collaboration summarized the effects of 10 anti-hypertensive drugs on 3610 PAD participants in 8 clinical trials (Lane and Lip,
2013). Three medications including perindopril, verapamil and telmisartan significantly increased the walking distance of PAD patients with no significant effect on ABI values.
Apart from the aforementioned two major types of medications administered to PAD individuals, others such as anti-platelet agents including aspirin and clopidogrel, and hypoglycemic agents including insulin and metformin, are also employed to treat atherosclerosis and diabetes, respectively (Hirsch et al., 2001). A subgroup consisting of
11,592 PAD patients from a randomized and blinded trial of clopidogrel versus aspirin in patients at risk of ischemic events (CAPRIE) were analyzed after a 3-year intake of 325 mg aspirin (n=5797) or 75 mg clopidogrel (n=5795) per day. Compared with aspirin
- 9 -
group, clopidogrel showed a reduction of ischemic risks including stroke, MI and
vascular death by 23.8% (CAPRIE, 1996). However, these studies did not report data for
functional indicators (i.e. ABI, claudication, or walking time) which are important
endpoints affecting PAD patients.
Furthermore, a class of phosphodiesterase inhibitors has been applied to treat IC in PAD
patients. They were believed to promote vasodilation and decrease blood viscosity and
platelet activity, thereby improving the blood flow through narrowed vessels (De Backer
et al., 2009; Thompson et al., 2002; Hood et al., 1996). Three meta-analyses have found
associations between utilization of naftidrofuryl (De Backer et al., 2009), cilostazol
(Thompson et al., 2002) or pentoxifylline (Hood et al., 1996), and improvement of
claudication in PAD. The clinical trials in those meta-analyses have shown that both naftidrofuryl and cilostazol, but not pentoxifylline, significantly increased walking distance in PAD patients. Also, there was evidence that propionyl-L-carnitine improved claudication (Hiatt et al., 2001). However, there is no evidence available to indicate that this class of medications result in decreased mortality.
In addition to numerous medications prescribed to PAD patients, exercise training has
also been shown to effectively improve conditions of PAD. For example, after 6-month
walking exercise intervention, the average for the 6-minute walking distance and the
maximal walking time increased by 53.5 meters and 1.01 minutes, respectively, in 97
individuals with PAD compared to a control group of another 97 PAD patients
(McDermott et al., 2013).
In addition to the medications, catheter-based endovascular therapies such as angioplasty and surgery are also commonly used to treat PAD patients with severe limb ischemia. A
- 10 -
review of endovascular treatment on PAD patients stated that percutaneous revascularization was only applied to those who had persistently serious claudication after exercise and medication therapies, and to those who had critical limb ischemia
(Weinberg et al., 2011). Although revascularization was the first-line approach administrated to isolated iliac artery stenosis due to the initially >90% success rate, it was limited by costs and techniques.
In addition to the therapeutic strategies mentioned above, nutrition might also be useful as a feasible therapy for PAD patients and it has been considered in several recent clinical trials.
2.1.3. Nutritional interventions
One nutritional study correlated dietary macronutrients and exercise performance of 46
PAD patients (Gardner et al., 2011). This study showed that higher intakes of monounsaturated fats were associated with significantly reduced walking time and less ischemia, while greater dietary fiber intake was associated with a significantly decreased onset time for claudication. In addition, this study revealed that PAD patients with diets high in saturated fats, sodium and cholesterol, and low in fiber, vitamin E and folate were highly associated with increased cardiovascular disease risk as well as ambulatory and vascular function impairment.
On the other hand, diets rich in vitamins, fibers and polyunsaturated fatty acids, and low in sodium and saturated fats were associated with a lower prevalence of PAD. As a result, fish oil, multivitamins and essential minerals have been recommended to PAD patients.
- 11 -
Several epidemiological studies have shown the importance of diet in PAD patients. For
example, Reis et al. (2008) compared the concentrations of 25-hydroxyvitamin D (25-
(OH)-D) in 866 African-American individuals with PAD and 2987 Caucasian individuals
with PAD. They reported that the concentrations of 25-(OH)-D in African-American
patients were significantly lower than Caucasian patients. In addition, they suggested a
7% reduction in the morbidity of PAD patients for every 5 nmol/L increase in 25-(OH)-D
concentration. Antonelli-Incalzi et al. (2006) reported that daily intakes of ≥ 34.4 g of
vegetable lipids and ≥ 7.7 mg vitamin E were significantly correlated with a reduced risk
of having an ABI less than 0.9 in 1251 PAD patients. Furthermore, an association
between daily nut intake and a 21% reduction in prevalence of PAD in 219,527 patients
has also been reported (Heffron et al., 2015).
Leyva et al. (2011) hypothesized that daily intake of 30 g of milled flaxseed, a rich source
of alpha-linolenic acid (ALA, an omega-3 fatty acid), might be effective to control blood
lipids and lower blood pressure in 110 participants with PAD who were also treated with
cholesterol-lowering (74%) and anti-hypertensive (79%) medications, and confirmed it in
2013 (Rodriguez-Leyva et al., 2013). Based on this hypothesis, Caligiuri et al. (2014) designed a six-month trial with flaxseed (HYPERFlax trial) to study the association between daily intake of 30 g milled flax and blood pressure lowering in PAD patients not taking anti-hypertensive drugs. They found a significant mean reduction of 7.97 mmHg in systolic blood pressure due to the effect of ALA from flaxseed. Also, plasma ALA concentration was positively correlated with flax consumption and negatively associated with the levels of six plasma soluble epoxide hydrolase-derived oxylipins in PAD
patients (Caligiuri et al., 2014). Edel et al. (2015) investigated both independent and
- 12 -
combined cholesterol-lowering effects of flaxseed and statins in PAD patients. This study
reported a significant reduction (11-15%) in total cholesterol and LDL-cholesterol that was maintained by daily intake of 30 g of milled flaxseed for 6 months. In addition, a subgroup analysis of their data found a significant decrease (8.5%) of LDL-cholesterol in the flaxseed + statin group compared to control + statin group after the 12-month daily consumption of 30 g milled flaxseed (Edel et al., 2015).
As shown from the results of the flax-PAD study, a combination of an appropriate diet and drugs may risk factors for PAD, however, there was no significant effects on ABI values and claudication. The interaction between the diet and medications might be an important factor.
2.1.4. Food-drug interactions
The presence of food and nutrients in the human digestive system might influence the absorption, metabolism and excretion of administrated medications, and likewise, drugs might affect the actions of food and nutrients (Pronsky and Crowe, 2012).
Consumption of oral agents has direct positive and negative associations with diet intake.
For example, in the alimentary tract, consumption of misoprostol with food reduced the intestinal absorption of misoprostol, a synthetic prostaglandin E1 analogue (Karim et al.,
1989), but intake of troglitazone (Young et al., 1998) with meals enhanced the bioavailability of this drug used in diabetes management. Meals could also affect oral agents which target blood formation, cardiovascular and respiratory health, or have anti- neoplastic and anti-infective effects (Schmidt and Dalhoff, 2002).
- 13 -
Cardiovascular agents such as cholesterol-lowering drugs, anti-hypertensive medications, diuretics and ACE inhibitors might be affected by regular diets. For instance, several studies showed that regular meal intake decreased the bioavailability of hydralazine
(Shepherd et al., 1984), captopril (Singhvi et al., 1982) and furosemide (McCrindle et al.,
1996), but increased the bioavailability of hydrochlorothiazide (Melander, 1977) and furosemide (Kim et al., 1993) through reduced gastrointestinal and first pass hepatic metabolism.
In addition, diets high in protein enhanced the bioavailability of propranolol (Semple and
Xia, 1995) and metoprolol (Wang and Semple, 1997) without relevant clinical effects, however, consumption of grapefruit juice (Bailey et al., 1998) increases the clinical effects and toxicity of anti-hypertensive drugs through enhanced levels of cytochrome
P450 (CYP3A4), an enzyme involved in drug metabolism. With respect to lovastatin,
regular meals increased its bioavailability, but fruit and fiber decreased the effect of
lovastatin (Schmidt and Dalhoff, 2002).
Furthermore, interactions between food and cardiovascular agents such as ACE
inhibitors, ARBs, and calcium-channel blockers have been previously reported (Jáuregui-
Garrido and Jáuregui-Lobera, 2012).
Administered drugs affect the absorption of several types of nutrients. For example, these
reactions include the chelation of dairy calcium by tetracycline; binding of fat-soluble
vitamins by cholestyramine; a change in the GI environment via inhibition of gastric acid
secretion by omeprazole followed by an impairment of vitamin B12 absorption; damage
of the intestinal mucosa by nonsteroidal anti-inflammatory drugs or colitis leads to iron
- 14 -
deficiency anemia; and folate transport is inhibited by pyrimethamine (Pronsky and
Crowe, 2012).
Drug-drug interactions are also common. Cholesterol lowering drugs such as atorvastatin inhibit the activity of 3-hydroxy-3-methyl-glutary CoA (HMG-CoA) reductase (EC
1.1.1.88 and EC 1.1.1.34) in the mevalonate pathway and thus decrease cholesterol synthesis, but it may also decrease the formation of coenzyme Q10 (Ghirlanda, 1993).
Furosemide increases the urinary excretion of potassium, magnesium, sodium, and calcium; hydrochlorothiazide (HCTZ) increases urinary excretion of potassium and magnesium, and decreases renal reabsorption of calcium; ACE inhibitors decrease urinary excretion of potassium; and chlorpromazine increases urinary excretion of riboflavin (Pronsky and Crowe, 2012).
With respect to food-drug interactions, the nutrients affected by drugs are present in regular daily diets, and most of the medications can be influenced by the food consumed.
According to Health Canada (2015), the Canadian daily diet should include 4 categories of food: vegetables and fruits, grain products, dairy products, and meat /meat alternatives.
Among those, 175 mL (3/4 cup) cooked beans per serving has been highly recommended especially for vegetarians and those wishing to decrease their meat consumption. Beans contain various nutrients that are similar to flaxseed, and specifically, high levels of dietary fiber. Therefore, beans might provide a variety of benefits to help PAD patients either in combination with administered drugs or independently.
- 15 -
2.2. Beans
Dried beans are considered a healthy food due to their low saturated fat content, high
amounts of essential amino acids, protein, fiber and useful phytochemicals. Beans can be
used as a substitute for animal products or animal-based protein in meals (Messina,
2014). According to Statistics Canada (2014), the production of pulses in Canada increased dramatically from 1981 to 2011, with 2.2 million hectares and 12,110 farms currently dedicated to pulse production, which makes Canada one of the top producers and exporters of pulses in the world (Bekkering, 2014). Although Canada is a top producer of beans, the average Canadian consumption is only 294±40 g/d (Mudryj et al.,
2012). However, vegetarians are more willing than non-vegetarians to consume higher amount of beans. In the United States, the Dietary Guidelines Advisory Committee decreased recommendations of bean intake from 3 cups per week in 2005 to 1.5 cups per week in 2010 (Messina, 2014).
2.2.1. Compositions of beans
Beans (Phaseolus vulgaris L.) are a good source of protein and fiber (Table 1) and contain a variety of phytochemicals. The protein content of beans is commonly between
20% and 30% of calories and a half cup serving of cooked beans (about 90 g) provides 7-
8 g protein, especially lysine, an indispensable amino acid (Messina, 2014). High fiber and low fat content makes beans unique among high protein and high carbohydrate containing foods. Beans are rich in resistant starch and fibers especially soluble fibers.
One half cup serving of cooked beans (about 90 g) provides 5-8 g fibers including approximately 1-3 g soluble fiber (Messina, 2014).
- 16 -
Table 1. Selected nutrient composition of 4 beans (Adopted from Messina, 2014). Bean (g per Total Total Iron Zinc Calcium Potassium Magnesium 0.5 cup of cooked beans) protein (g) fiber (g) (mg) (mg) (mg) (mg) (mg) Black beans (86 g) 7.6 7.5 1.8 0.96 23 305 60 Kidney beans (88.5 g) 7.7 5.7 2.0 0.9 31 358 37 Navy beans (85.5 g) 7.5 5.2 2.2 0.9 63 354 48 Pinto beans (85.5 g) 7.7 7.7 1.8 0.8 40 373 43
The common beans, including pinto, kidney, black and navy beans, are also rich in
flavonols such as kaempferol (Hu et al., 2006), anthocyanins and tannins (Díaz et al.,
2010), and quercetin and rutin (Mishra et al., 2012) as well as isoflavonoids such as
daidzein and genistein (De Lima et al., 2014). These compounds in beans are generally
expressed as free forms and their corresponding 3-O-glycosides forms. For example,
black beans mainly contain 3-O-glycosides of anthocyanins (delphinidin, petunidin and
malvidin), pinto beans primarily contain free and the 3-O-glycoside of kaempferol, and kidney beans majorly contain quercetin and quercetin-3-O-glycoside (Lin et al., 2008).
However, the Lin et al. (2008) study did not detect or report flavonoids. Díaz et al. (2010) showed that anthocyanins are present in common beans mainly as delphinidin-3- glucoside, while tannins are present primarily as catechin (60.3%), gallocatechin (25%) and afzelechin (14.7%). The presence of other flavonoids in beans needs to be confirmed in future studies. The concentrations of the various flavonoids are distinctive in diverse genotypes of common beans and vary from study to study. Two examples of phenolic distribution are shown in Table 2.
- 17 -
Table 2. Flavonoids in selected common beans.
De Lima et al. (2014) Díaz et al. (2010) Flavonoids (average Category Flavonoids (average value, μg/g) Genotype value, %) (bean) kaempferol quercetin daidzein genistein myricetin coumestrol anthocyanins tannins Raz 55 Black 5.40 8.34 8.05 5.98 26.15 17.67 Brancão Navy ------Argentino Flor de 0.08 20.04 Kidney 6.47 - - 1.93 - - Mayo IAC Pinto 60.07 9.88 - 8.93 - 16.85 Formoso
Most of the bioactive compounds present in common beans are found in the seed coat
(Kadouh and Zhou, 2012). In addition, the seed coat of beans also contains vitamins,
minerals and several other phytochemicals such as phytosterols (Chávez-Santoscoy et al.,
2014), saponins (Chavez-Santoscoy et al., 2014) and phytates (Petry et al., 2013) as
shown in Table 3.
Table 3. Phytosterols, saponins and phytates in selected types of beans. Category (bean) Phytochemicals Black Navy (white) Kidney/Pinto (brown) Phytosterols Stigmasterol 3.21 - - (mean value, Campesterol 66.71 - -
mg/100 g seed β-Sitosterol 70.65 - - coats)1 Δ5-Avenasterol 5.50 - - Phaseoside l 0.17 - - Soyasaponin Af 1.42 - - Deacetyl Saponins 0.26 - - soyasaponin Af (mean value, Soyasaponin Ba 0.34 - - mg/g seed coat)2 Soyasaponin αg 1.26 - - Soyasaponin βg 0.50 - - Soyasaponin γg 0.20 - - Phytates (mean value, Phytic acid - 1360 1030 mg/100 g seed)3 1 Chávez-Santoscoy et al., 2014 b; 2Chavez-Santoscoy et al., 2014 a; 3Petry et al., 2013
- 18 -
2.2.2. Nutrients and health
Both soluble fibers and insoluble fibers are abundant in the common beans, and they are
especially beneficial for cardiovascular and digestive health, and contribute to lowering
blood cholesterol concentrations, as well as controlling blood pressure and blood glucose
(Kadouh and Zhou, 2012). One meta-analysis of 67 controlled clinical trials reported that
daily intake of 2-10 g soluble fibers contributed to a small but significant decrease of
total cholesterol (-0.045 mmol/L/g soluble fiber) and LDL-cholesterol (-0.057 mmol/L)
concentrations in both healthy and ill individuals (Brown et al., 1999). Another meta-
analysis of 25 clinical studies reported a positive association between intake of dietary
fibers (7.2-18.9 g/d) and control of blood pressure (-3.40 mmHg in systolic blood pressure (SBP); -1.97 mmHg in diastolic blood pressure (DBP)) in humans (Whelton et al., 2005). Another meta-analysis of 11 qualified clinical trials showed that dietary (42.5 g per day) or supplemental (15 g per day) fibers were associated with a significant reduction of glycated hemoglobin (HbA1c; -0.55%) and fasting plasma glucose (-9.97
mg/dL or -0.55 mmol/L) (Silva et al., 2013).
Apart from fibers, beans are also good sources of resistant starch. Resistant starch (1.7-
4.2 g/100 g cooked beans) has been shown to contribute to the low glycemic index of
beans (Messina, 2014). Jenkins et al. (2012) reported a greater decrease in HbA1c values
(-0.2%) with bean fibers compared to wheat fiber. Syed and Khan (2011) concluded that
a 1% decrease in HbA1c values results in a 15–18% reduced risk of ischemic heart
disease. Furthermore, Merchant et al. (2003) reported cereal fibers reduce the risk of
PAD in men. Thus, the high fiber content of beans might be a factor capable of
improving the health of PAD patients.
- 19 -
Phytochemicals such as phenolic acids including flavonoids are antioxidants that might
be useful therapies in the management of oxidative-stress-induced diseases (Tsao et al.,
2012). de Bock et al. (2012) reported associations between polyphenols, glucose
homeostasis, and risk reduction for diabetes and cardiovascular disease. Two review
articles have reported anti-hypertensive, cholesterol-lowering and anti-atherosclerotic effects of 5 categories of flavonoids containing flavones (luteolin and chrysin), flavonols
(quercetin, kaempferol and myricetin), flavanones (naringenin and hesperidin), flavan-3- ol derivatives (catechins), and anthocyanins both in animal and human studies (Clark et al., 2015; Salvamani et al., 2014).
For instance, quercetin and rutin showed a significant dose-dependent effect on lowering blood pressure, blood glucose and LDL-cholesterol concentrations in rats fed a high sodium diet; in addition, quercetin and rutin produced greater reductions than nifedipine, an anti-hypertensive drug (Olaleye et al., 2014). In the meantime, a clinical trial investigating the effect of daily 500 mg quercetin intake on blood pressure in 62 diabetic women demonstrated a mean 8.8 mmHg reduction in SBP (Zahedi et al., 2013). Potenza et al. (2007) studied the effect of catechins (epigallocatechin gallate) (200 mg/kg/d) compared with enalapril (30 mg/kg/d), an anti-hypertensive drug, in spontaneously hypertensive rats (SHR); both catechins and enalapril significantly decreased systolic blood pressure and increased insulin sensitivity in SHR, while catechins stimulated production of NO and improved cardiovascular dysfunction. In a two-year human study,
Matsuyama et al. (2008) demonstrated that daily intake of 576 mg catechins from tea significantly reduced SBP (-10.2 mmHg) and LDL-cholesterol (-0.68 mmol/L) in 21 obese children compared to 19 controls not consuming the catechins in tea. In addition,
- 20 -
Yamamoto et al. (2008) and Morand et al. (2011) reported the anti-hypertensive effect of
hesperidin in SHR and overweight men, respectively.
With respect to other animal studies, Kong et al. (2013) reported that kaempferol had a
stronger effect than fenofibrate (a cholesterol-lowering drug) for reducing total
cholesterol and LDL-cholesterol concentrations of rabbits fed a high cholesterol diet;
therefore, a high dose of kaempferol was correlated to the inhibition of atherosclerosis at
an early stage through attenuating the expression of intercellular adhesion molecules. El-
Bassossy et al. (2013) showed that daily administration of 25 mg/kg chrysin and 100
mg/kg luteolin was associated with a significant attenuation of diastolic blood pressure,
and total cholesterol and LDL- cholesterol levels in diabetic rats. In this study, chrysin
reduced advanced glycation end products and luteolin ameliorate NO production in
diabetic aorta. Furthermore, the dose-dependent associations between myricetin (Godse et al., 2010) and naringin (Alam et al., 2013) and anti-hypertensive, hypoglycemic and cholesterol-lowering effects in fructose-fed rats were confirmed.
However, kaempferol, chrysin, luteolin, myricetin and naringin have been rarely investigated in human studies. In addition to the in vivo studies described above, the anti- atherosclerotic potential of fisetin, morin, myricetin (Lian et al., 2008) and gossypetin
(Chen et al., 2013) for inhibiting oxidation of LDL-cholesterol have been demonstrated in vitro. The anti-inflammatory potential of kaempferol (Kowalski et al., 2005) and quercetin (Kleemann et al., 2011) has also been reported in in vitro studies.
A study by Carnevale et al. (2014) reported that PAD patients had higher platelet activation and lower flow-mediated dilation than healthy individuals, however, catechins
(mixture of epicatechin (0.1–10 M) and catechin (0.1–10 M), for 30 mins )
휇 - 21 - 휇
significantly ameliorated endothelial activation via a decrease in cell adhesion molecules
(CAMs) and an increase in NO products in vitro.
Daily intake of anthocyanins (320 mg/d for 24 weeks) was shown to significantly reduce
β-thromboglobulin and soluble platelet selectin in 73 hyperlipidemic participants (Song et al., 2014). The inhibition of platelet activation by anthocyanins was confirmed in vitro.
Teede et al. (2003) found that isoflavones, containing abundant formononetin, had estrogen-like activity and decreased arterial stiffness and total vascular resistance in humans. Ohkawara et al. (2005) illustrated that pterocarpan but not formononetin inhibited proliferation and DNA synthesis of vascular smooth muscle cells. The results from those research studies were associated with the onset and progression of atherosclerosis.
Other compounds, such as phytosterols, saponins and phytates, present in pulses might be beneficial to cardiovascular patients. For example, a cohort study found that the plasma concentrations of plant sterols in 125 participants with coronary heart disease were significantly lower compared to individuals without vascular disease, and that a 2-fold increase of sitosterol levels in plasma was correlated with a significant 22% decrease of coronary heart disease risk (Fassbender et al., 2008). Cha´vez-Santoscoy et al. (2014a) reported that phytosterols from black beans significantly decreased cholesterol synthesis and lipogenesis in rat hepatocytes. Another study from Cha´vez-Santoscoy et al. (2014b) showed the saponins from black beans decreased lipogenesis and increased bile acid synthesis both in vivo and in vitro.
- 22 -
The relevant studies and effects of the aforementioned phytochemicals are shown in
Table 4. These phytochemicals are likely present in beans, but this needed to be confirmed in our study.
- 23 -
Table 4. Selected dietary phytochemicals and their effects. Phytocemicals Formula Structure Mass Animals studies Human studies In vitro studies Kleemann et al., Olaleye et al., 2013 2011 Zahedi et al., 2013 Anti-hypertensive Anti-inflammatory Anti-hypertensive Quercetin C H O 302.236 hypoglycemic Anti-proliferation 15 10 7 (serum of type 2 Cholesterol-lowering Anti-atherosclerosis diabetics) (serum of rats) (cultured human endothelial cells) Potenza et al., 2007 Carnevale et al., Matsuyama et al., Anti-hypertensive 2014 2008 Anti-platelet Anti-platelet Anti-hypertensive Insulin sensitivity- (platelet rich plasma Catechins C H O 290.079 Cholesterol-lowering 15 14 6 increased of PAD patients, (serum and urine of (blood, mesenteric human umbilical obese children) vascular beds, heart of vein endothelial
SHR) cells) -
24 24 Yamamoto et al., 2008 Morand et al., 2011 Anti-hypertensive Anti-hypertensive -
Hesperidin C H O 610.1898 - 28 34 15 (thoracic aorta rings of (plasma and urine of SHR) heathy volunteers)
Flavonoids Kong et al., 2013 Kowalski et al., Cholesterol-lowering 2005 Kaempferol C15H10O6 286.0477 Anti-atherosclerosis - Anti-inflammatory (serum, intercostals (cultured mouse artery of rabbits) macrophage cell) El-Bassossy et al., 2013 Anti-hypertensive Cholesterol-lowering Chrysin C H O 254.0579 - - 15 10 4 Hypoglycemic (serum, thoracic aorta rings of rats) El-Bassossy et al., 2013 Anti-hypertensive Cholesterol-lowering Luteolin C H O 286.0477 - - 15 10 6 Anti-platelet (serum, thoracic aorta rings of Wistar rats)
Phytocemicals Formula Structure Mass Animals studies Human studies In vitro studies Olaleye et al., 2013 Anti-hypertensive Rutin C27H30O16 610.1534 hypoglycemic - - Cholesterol-lowering (serum of rats)
Godse et al., 2010 Lian et al., 2008 Anti-hypertensive Anti-atherosclerosis Myricetin C H O 318.0376 Cholesterol-lowering - 15 10 8 (U937-derived Hypoglycemic macrophages) (serum of rats) Alam et al., 2013 Anti-hypertensive Cholesterol-lowering Naringin C H O 580.1792 - - 27 32 14 Hypoglycemic -
(plasma, thoracic 25 25 aortic ring of rats) - Chen et al., 2013 Anti-atherosclerosis (blood of healthy Gossypetin C H O 318.0376 - - 15 10 8 individuals, cultured murine macrophage cell)
Lian et al., 2008 Anti-atherosclerosis Fisetin C H O 286.2363 - - 15 10 6 (U937-derived macrophages)
Song et al., 2014 Song et al., 2014 Anti-atherosclerosis Anti-platelet Anthocyanins C H O 287.246 - (plasma of 15 11 6 (gel-filtered hypercholesterolaemic platelets) patients)
Phytocemicals Formula Structure Mass Animals studies Human studies In vitro studies Teede et al., 2003 Anti-arterial stiffness Formononetin C H O 268.0736 - 16 12 4 (plasma and urine of healthy participants)
Ohkawara et al., 2005 Anti-proliferation Astrapterocarpan C H O 212.1593 - - 17 16 5 (cultured Al0 cells derived from the rat thoracic aorta) Cha´vez-Santoscoy et al., 2014b Cholesterol- β-Sitosterol C29H50O 414.707 - - lowering Anti-lipogensis -
26 26 Phytosterols (primary rat
- hepatocytes)
Cha´vez-Santoscoy Cha´vez-Santoscoy et et al., 2014b al., 2014a Anti-lipognesis soyasaponin Af C H O - Cholesterol-lowering - 27 42 3 Pro-bile synthesis (serum, liver and ileum Saponins (cultured primary rat of C57BL/6 mice) hepatocytes)
All phytochemicals involved in the above studies were administered to subjects as pure chemicals and not dietary extracts.
2.2.3. Beneficial effects of beans on diseases conditions
An epidemiological study has shown that pulse consumption more than 4 times per week
compared to less than once per week reduces the risk of chronic heart disease by 22% and
chronic vascular disease by 11% (Bazzano et al., 2001). Common beans might benefit
cardiovascular health mainly through lowering serum cholesterol levels. Consumption of
common beans for 45 days significantly decreased total cholesterol, LDL-cholesterol,
very low-density lipoprotein (VLDL)-cholesterol, and serum triglycerides in diabetic rats
(Venkateswaran et al., 2002). A Guatemalan epidemiological study reported that black beans were associated with a lower glycemic response and decreased serum cholesterol levels (Serrano and Goñi, 2004). Winham et al (2007) showed daily intake of 0.5 cups pinto beans for 8 weeks compared to black-eyed peas or carrots (placebo) significantly lowered serum total cholesterol and LDL-cholesterol thereby reducing risk for coronary heart disease in individuals with diabetes.
Hanson et al. (2014) studied the effects of different pulse varieties in SHR, a model of hypertension and arterial stiffness, which are components of chronic arterial diseases.
This study reported that various pulse types including beans (mixture of black, navy, red kidney, and pinto) reduced the LDL-cholesterol levels without any difference among the pulse types studied (beans, peas, lentils, chickpeas). Although all pulses reduced circulating TC and LDL-cholesterol in the SHR, only lentils significantly attenuated the rise in blood pressure and had positive effects on large-artery remodeling in the SHR, without affecting pulse wave velocity. Interestingly, the benefits obtained with lentils were not explained by cholesterol lowering.
- 27 -
Apart from the cholesterol-lowering effect, an in vitro study recently reported that pinto
bean peptides have ACE-inhibitory activity (Tagliazucchi et al., 2015); however, in vivo
studies are scarce and the anti-hypertensive effects of beans need to be confirmed.
Hanson et al. (2014) reported that lentils but not beans (mixture of black, white navy, red
kidney, and pinto) attenuated the rise in blood pressure of SHR compared to the pulse-
free control SHR.
In addition, some studies have explored the hypoglycemic effect of beans. In a review,
Rizkalla et al. (2002) summarized that pulses, as foods with a low glycemic index and high fiber content, helped to control the blood glucose and blood lipids in patients with type 2 diabetes and prevent chronic heart disease and diabetes. Venkateswaran et al.
(2002) demonstrated that water-soluble extracts from beans significantly reduced the
blood glucose concentration in diabetic rats. Jenkins et al. (2012) reported that a diet
containing legumes reduced HbA1c, a risk factor for cardiovascular disease, by 5% in
type 2 diabetes patients.
Apart from cholesterol-lowering, hypotensive and hypoglycemic activities, Potischman et
al. (1999) showed an association between increased bean intake and breast cancer risk
reduction. Wong et al. (2010) reported a strong anti-proliferative effect of red beans,
which was related to anti-carcinogenic effects in human hepatoma cells. The common
beans have also been associated with weight control. Celleno et al. (2007) reported
significant weight loss in 60 overweight participants after consuming an extract (which
inhibits α-amylase) from white kidney beans for 30 days.
Many of the beneficial effects of beans outlined in aforementioned studies might help
with symptoms associated with PAD. However, nutritional studies investigating the
- 28 -
effect of pulses in PAD patients are scarce and the mechanisms by which pulses might provide beneficial results to atherosclerotic patients need to be elucidated.
A recent nutritional intervention showed that the daily consumption of mixed pulses (0.5 cup of mixed pulses including beans, peas, lentils and chickpeas) decreased the total cholesterol and LDL-cholesterol levels of PAD patients by 5.0% and 8.7%, respectively, and increased the ABI of PAD patients by 5.5% (Zahradka et al., 2013). However, the independent effects of each pulse on PAD are not known.
One approach for investigating the effects of a dietary component (e.g. pulses) in a chronic disease state (e.g. PAD) is to apply a non-targeted metabolomics approach to learn more about endogenous and exogenous metabolites.
2.3. Metabolomics
Metabolomics is a relatively new discipline and is generally defined as “the quantitative measurement of the dynamic multi-parametric metabolic response of living systems to pathophysiological stimuli or genetic modification” (Nicholson et al., 1999). Others have defined it as “comprehensive and quantitative analysis of all metabolites in a system”
(Fiehn, 2001), or as “a comprehensive and systematic identification and quantification of small molecule metabolites (<1500 Da) in biological samples at a given point of time”
(Bujak et al., 2015).
As a novel, integrated and holistic approach, metabolomics is commonly applied to investigate the quantitative changes of endogenous and exogenous metabolites in biochemical and clinical studies. Comprehensive analyses of low molecular weight
- 29 -
metabolites in biological samples such as urine and blood frequently rely on profiling
techniques such as NMR spectroscopy and LC-MS-based methods, combined with diverse statistical analyses (Duarte et al., 2014).
Along with metabolomics, genomics, transcriptomics, and proteomics are categorized as
“omic” sciences, which aim at examining the metabolome, genome, transcriptome, and proteome, respectively (Bujak et al., 2015). In combination, they reflect how information goes from genes to transcripts through proteins and finally to metabolites present in
biological systems. Therefore, in specific biological systems, the dynamic perturbations
of genome, transcriptome and proteome are revealed by the changes of corresponding
metabolome. This is a reflection of the specific molecular phenotype and is frequently
implemented to investigate complex and dynamic pathophysiological processes in
disease progression as well as to seek new diagnostic or prognostic biomarkers related to
various disorders (Bujak et al., 2015).
2.3.1. Metabolomics in disease-related studies
As a feasible method in clinical research, metabolomics is frequently applied to study a
variety of systemic disorders such as diabetes, cancer and cardiovascular disease. For
example, metabolomics has identified lysoglycerophospholipids as potential prognostic
markers of gestational diabetes mellitus and they are closely associated with the
development of severe glucose intolerance in pregnant women (Dudzik et al., 2014).
Metabolite profiling has shown that metabolites from lipid metabolism are correlated
with mammary carcinoma in women with breast cancer (Mensack et al., 2012).
Moreover, one study on the metabolome demonstrated that the metabolic perturbation
caused by alcohol abuse was linked with development of hypertension (van Deventer et
- 30 -
al., 2015). Using a metabolomic approach, Turer et al. (2009) showed that acylcarnitine,
a myocardial substrate released after surgical ischemia or reperfusion in patients with
coronary artery disease or left ventricular dysfunction, was a useful targeted metabolic
biomarker to monitor the prognosis of these patients.
With respect to PAD, plasma metabolomic profiles might be a useful predictor for
identifying increased risk of all cause near-term mortality in patients with PAD (Huang et al., 2013). In addition, a study on the metabolome has identified serum tyrosine and oxidized LDL as metabolic biomarkers of artery stiffness in patients with PAD (Zagura et al., 2014).
Metabolomics has been recently implemented in research on Alzheimer's disease
(Gonzalez-Domínguez et al., 2015), glaucoma (McNally and O’Brien, 2014), ulcerative colitis and Crohn’s disease (Bjerrum et al., 2015), Niemann-Pick disease (Maekawa et al., 2015), psoriatic disease (Armstrong et al., 2014) as well as inherited metabolic disorders (Janeckova et al., 2015). Duarte et al. (2014) concluded that metabolomics has been applied to study 10 types of diseases related to 10 different systems including nervous, digestive, respiratory, and circulatory systems. The details are shown in Table 4.
- 31 -
Table 5. Metabolomics in disease-related studies. Reference Disease Outcomes Combination of GC-MS and LC-MS Gonzalez-Domínguez provided urine and serum biomarkers in Alzheimer's disease et al., 2015 accordance with pathological alterations in human Alzheimer's disease LC-MS/MS provided urine and serum McNally and O’Brien, Glaucoma biomarkers to monitor progression of 2014 exfoliation glaucoma 1H NMR spectroscopy of fecal extracts as a potential non-invasive diagnostic tool to Ulcerative colitis and Crohn’s Bjerrum et al., 2015 characterize inflammation driven changes disease in metabolic profiles related to malabsorption and dysbiosis LC-MS/MS lead to the discovery of new Maekawa et al., 2015 Niemann-Pick disease urinary biomarkers for disturbances in cholesterol catabolism and transport GC-TOF-MS revealed altered serum metabolites to help elucidate the Armstrong et al., 2014 Psoriatic disease pathogenesis of psoriasis and psoriatic arthritis; this may provide insights for therapeutic development LC-QTOF-MS can be applied for diagnosis of mild inherited metabolic Janeckova et al., 2015 Inherited metabolic disorders disorders which have less clear biochemical profiles Neoplasms; Digestive, endocrine, genitourinary, circulatory, repiratory and NMR metabolomics provides insight into nervous system; Nutritional and underlying disease pathogenesis and can Duarte et al., 2014 metabolic disturbances; Mental unveil new metabolic markers for disease state and behavioural changes; diagnosis and follow up Pregnancy, childbirth and the puerperium; Infectious and parasitic diseases
Apart from involvement in diseases studies, metabolomics is used in pharmaceutical studies as well as nutrition research.
2.3.2. Metabolomics in drug studies
Altmaier et al. (2014) has used metabolic variations to reveal the effects of anti- hypertensives and cholesterol-lowering drugs in the general population. For example, the metabolites of beta-blockers and ACE inhibitors were used to investigate the side effects and action of these drugs, respectively; the metabolites of statins and fibrates were used
- 32 -
to monitor reduced cholesterol synthesis as well as increased cholesterol degradation
(Altmaier et al., 2014).
Metabolomic approaches offer a specific insight into various classes of medications and
their corresponding metabolites, and can reveal interactions among multiple variables.
Apart from anti-hypertensive and cholesterol-lowering drugs, the metabolomics method has been implemented in trials investigating acetaminophen (Kim et al., 2013), indomethacin (Um et al., 2012), nonsteroidal drugs (Um et al., 2009), and antimicrobial drugs (Derewacz et al., 2013) as well as anti-psychotic medications (Chan et al., 2011).
The applications of metabolomics on those drugs are shown in Table 5.
Table 6. Metabolomics in drug-related studies. References Drugs Outcomes 1H NMR-based metabolomics was used to predict or Kim et al., 2013 Acetaminophen screen potential hepatotoxicity caused by acetaminophen 1H-NMR urine spectra suggested that screening for gastric Um et al., 2012 Indomethacin damage induced by indomethacin is possible in the preclinical stage of drug development 1H NMR-based metabolomics indicated putative Nonsteroidal Um et al., 2009 biomarkers that might be useful for predicting adverse drugs effects caused by NSAIDs LC-MS revealed that derepression of biosynthetic Antimicrobial Derewacz et al., 2013 pathways is a general consequence of some antibiotic drugs resistance mutations LC-MS and 1H NMR-based metabolomics revealed novel protein and metabolite changes in low-cumulative- Anti-psychotic Chan et al., 2011 medication subjects associated with synaptogenesis, medications neuritic dynamics, presynaptic vesicle cycling, amino acid and glutamine metabolism, and energy buffering systems
2.3.3. Metabolomics in food and nutrition studies
As a powerful tool applied to nutrition research, metabolomics offers a novel opportunity
to identify new biomarkers of dietary intake, to study diet-related diseases, and to
elucidate molecular mechanisms of dietary interventions (Gibbons et al., 2015). Markers
- 33 -
of potential food exposure to several plant foods including berries, fruits, vegetables and
herbs, oils, nuts and chocolate have been discovered and validated through metabolomics
(Andersen et al., 2014). Metabolomics has been applied to the study of rhubarb aglycone
(Guan et al., 2014). Mannina et al. (2012) concluded the applications of NMR-based metabolomics in various food trials were potentially versatile to explain the composition, processing, and storage procedures of truffle, kiwi fruit, lettuce, and sea bass over time.
Metabolomics offered a clear view to understand the influence and regulation of nutrition on diet-related disorders such as diabetes, cancers and cardiovascular diseases (Gibbons et al., 2015).
At the same time, metabolomics has been extensively applied to discriminate the species and geographical origin of green coffee beans (Wei et al., 2012), and to identify the general discrimination factors in immature (Setoyama et al., 2013) and roasted (Wei et al., 2014) coffee beans. Furthermore, the effect of dry common beans for down- regulating lipid metabolism in mammary carcinomas has been demonstrated via metabolomics analysis (Mensack et al., 2012). Metabolomics combined with proteomics has been successfully applied to validate the genetic distance between two common beans, the navy bean and the white kidney bean (Mensack et al., 2010). Apart from these, metabolites in rice plants revealed the chemical diversity of rice and it was useful to identify crucial potential metabolite biomarkers and their corresponding genes for rice growth using a metabolomics approach (Kusano et al., 2015). Other lipid metabolites such as oxylipins (Willenberg et al., 2015) and beverages such as wine (Arbulu et al.,
2015) have been classified using metabolic profiling. Tagliazucchi et al. (2015) has reported the ACE-inhibitory activity and related hypotensive properties of pinto bean
- 34 -
peptides in vitro through a high-performance liquid chromatography-mass spectrometry
(HPLC-MS) method. However, the ACE-inhibitory activity still needs to be confirmed in vivo.
Given the capabilities of metabolomics and its application to various fields discussed above, the objectives of this thesis were to apply an LC-MS based metabolomics approach to biological fluids (urine and serum) collected from PAD patients who participated in a nutritional intervention with beans and to investigate the effect of bean consumption on endogenous and exogenous detected metabolites. Since the PAD patients were taking various medications, it was also important to investigate any potential drug- diet interactions.
- 35 -
Chapter 3. Hypotheses and Objectives
3.1. Hypotheses
Metabolomics is a powerful tool and its application to nutrition (nutritional
metabolomics) and/or a population with cardiovascular disease seems appropriate,
therefore, we hypothesized that:
1. The majority of endogenous and exogenous metabolites would be successfully
detected and identified using a liquid chromatography coupled to quadrupole
time-of-flight mass spectrometry (LC-QTOF-MS)-based non-targeted
metabolomics approach;
2. The daily consumption of beans or rice (0.3 or 0.6 cups cooked mixed beans/day
compared to 0.6 cups cooked rice/day; 5 days/week; for 8 weeks) would
significantly affect the metabolic profile of administrated drugs and endogenous
metabolites present in biological fluids (urine and serum) of PAD patients.
3.2. Objectives
The main objectives of this study were:
1. To develop LC-MS based metabolomics methods capable of monitoring the
majority of known endogenous and exogenous (administered drugs) compounds
in biological fluids (urine and serum) collected from PAD patients who
participated in a nutritional intervention with beans;
- 36 -
2. To investigate the effect of bean consumption (0.3 or 0.6 cups mixed beans/day
compared to 0.6 cups rice/day; 5 days/week; for 8 weeks) on administered drugs
and detected endogenous metabolites;
3. To investigate the chemical composition of the four beans types (black, navy, red
kidney, pinto) and rice used in the nutritional intervention.
- 37 -
Chapter 4. Materials and Methods
4.1. Chemicals and standards
Acetonitrile, methanol, formic acid (HPLC grade) and D-norvaline were purchased from
Sigma-Aldrich (ON, Canada). Reference solutions, TFANH4 (Trifluoroacetic acid
ammonium salt), Hp (Hexakis (1H, 1H, 3H-tetrafluoropropoxy) phosphazine) and purine were purchased from Agilent Technologies (CA, USA). Deionized water was prepared using the Milli-Q system from Millipore (MA, USA).
4.2. Bio-fluids and diet materials
The biological fluids (urine and serum) and cooked beans used for this Thesis were from a clinical trial (ClinicalTrials.gov Identifier: NCT01382056) conducted by Drs. Peter
Zahradka and Carla Taylor at the Asper Clinical Research Institute, St Boniface Hospital,
Winnipeg, Manitoba, Canada. These samples (urine and serum) were obtained during the
period of March 15, 2012 to September 17, 2013 (i.e. baseline visit of first participant
and 8 week visit of the last participant); the samples were processed immediately and
stored at -80°C prior to their delivery (May 30, 2014) to our laboratory for the
metabolomics studies described here. The bean and rice samples were cooked and stored
at -20°C for approximately one month before metabolomics analysis.
The urine and serum samples were from an 8-week food intervention study in which 62
participants with PAD were randomly assigned to one of 3 groups: rice (0.6 cups rice per
food item; pulse-free control; n=22), or 0.3 cups (n=19) or 0.6 cups (n=21) of mixed
cooked beans (navy, pinto, red kidney, black) per food item. Participants were provided
- 38 -
with a selection of frozen food items (soups, side dishes, entrees) and they were instructed to consume one study food item per day for 5 days per week for 8 weeks as part of their usual dietary intake. The overall goal of the clinical trial was to determine the effects of mixed bean consumption on vascular function in individuals with PAD.
Assessments included anthropometric, biochemical, clinical and dietary parameters.
Relevant to this thesis, a detailed medical history, including all medications, was obtained from the participants. Urine and serum samples were collected at baseline and after 8 weeks.
The inclusion and exclusion criteria used to recruit the PAD participants are summarized in Table 5. These patients were recruited from the vascular clinic at St. Boniface
Hospital.
Table 7. Recruitment criteria for PAD participants (ClinicalTrials.gov Identifier: NCT01382056, conducted by Drs. Peter Zahradka and Carla Taylor). Inclusion Criteria: Exclusion Criteria: 1. Smoking for last 6 months; 2. Type 1 diabetes; 3. Renal failure on dialysis; 4. Acute cardiovascular or medical illness 1. Male or female more than 40 years old; within the past 3 months; 2. Having documented diagnosis of PAD: 5. Hormone (estrogen) replacement therapy; Intermittent claudication defined by an ABI of 6. Amputation of upper or lower extremity on ≤ 0.90; or both sides; Asymptomatic carotid stenosis > 50% or 7. Allergies or severe gastrointestinal reactions confirmed by imaging; or to pulses or ingredients used to prepare the Having a previous intervention for PAD or pulse-containing and pulse-free foods; carotid disease; 8. Currently participating in or having 3. Stable medication for the past 3 months; participated in a food study within the last 4. Willing to comply with the protocol 3 months except involvement in a control requirements; group and consumed no study foods; 5. Willing to provide informed consent; 9. Inability to adhere to a regular diet. 10. High pulse consumption of more than 2 servings per week, or additional intake of pulses apart from the study requirements.
- 39 -
After excluding participants who did not complete the study and missing samples, there
were 108 urine and 106 serum samples used for this metabolomics study.
The beans were high quality seed sourced by Pulse Canada from local pulse growers. The
cooking conditions for the four types of beans (pinto, kidney, black, and navy) used in the
food intervention study are shown in Table 6. The beans were soaked overnight at 4°C,
drained, and fresh water added for cooking. The beans were brought to a boil and
simmered for the indicated times.
The clinical trial investigated two doses of beans (0.3 and 0.6 cups per day for 5 days per
week, or 1.5 and 3 cups per week, respectively) and this is within the range of various
dietary recommendations. For example, the Dietary Guidelines for Americans
recommended 3 cups per week in 2005 but this has since be reduced to 1.5 cups per week
in 2010 (Messina, 2014). Furthermore, a previous study by Drs. Zahradka and Taylor had
found that daily consumption of ½ cup mixed pulses (beans, peas, lentils and chickpeas)
for 8 weeks improved blood flow to the legs of participants with PAD (Zahradka et al.,
2013), and they wanted to see if the effects on vascular function could be replicated with
beans.
Table 8. Sample preparation and weights of beans and cooking water (unpublished data provided by Drs. Peter Zahradka and Carla Taylor). Soaking Dry Soaked water Cooked Drained Cook Soaking Cooking Beans beans beans after beans cooking time water (g) water (g) (g) (g) soaking (g) water (g) (min) (g) Black 400 1200 789 807 2367 888 1805 50 Navy 500 1500 950 1032 2850 1017 2175 64 Pinto 400 1200 766 801 2100 899 1727 58 Kidney 433 1299 887 835 1661 960 2073 56
- 40 -
4.3. Sample preparation for metabolomics studies
4.3.1. Urine
Urine samples (kept at -80°C) were thawed at room temperature for approximately 20
minutes and vortexed for 10 seconds. A sub-sample (250 µL) of the urine sample was
placed in a 2.0 mL Eppendorf tube, to which 10 µL of norvaline solution (0.03 mg/mL)
was added as the internal standard. The samples were then centrifuged at 10,000g for 10
minutes at 4°C. The supernatant was placed in a new 2.0 mL Eppendorf tube, to which
500 µL of acetonitrile was added and vortexed for 20 seconds. The tube was kept at
‒20°C for 30 minutes and then the mixture was centrifuged at 10,000g for 20 minutes at
4°C. The supernatant was transferred into a new Eppendorf tube and dried under vacuum
(Settings V-HV) at room temperature for 3 hours. The dried samples were kept at -20°C prior to reconstitution for LC-MS analysis. The samples were reconstituted in 200 µL of acetonitrile:ddH2O (1:4) using glass insert and brown LC vials for analysis. In total, 54
pairs of samples (baseline and week 8) were prepared.
4.3.2. Serum
Frozen serum samples (kept at -80°C) were thawed on ice for approximately 20 minutes.
A sub-sample (100 µL) was placed in a 2.0 mL Eppendorf tube, to which 200 µL acetonitrile was added. Samples were vortexed vigorously for 30 seconds, and centrifuged at 10,000g for 10 minutes at 4°C. The supernatant (200 µL) was transferred into a new Eppendorf tube and completely dried under vacuum (Settings V-HV) at room temperature for 1.5 hours. The dried samples were stored at -20°C. Prior to HPLC-
QTOF-MS, the dried samples were reconstituted with 100 µL of 80% acetonitrile
- 41 -
prepared in deionized water using glass insert and brown LC vials. In total, 53 pairs of
samples (baseline and week 8) were prepared.
4.3.3. Beans
Samples of cooked pinto, black, kidney, and navy beans and cooked rice were crushed
into a mash. One hundred fifty mg crushed bean or rice sample was placed in an 8 mL
tube, to which 2 mL of methanol:water (3:2) was added. Samples were vortexed for 1
minute, and sonicated in a Branson 1200 sonicator bath containing deionized water for
one hour (the water was changed every 15 minutes). The samples were then vortexed for
10s and centrifuged at 14,000g for 10 minutes at 4°C. The supernatant was transferred
into a clean 2 mL Eppendorf tube, and dried under a gentle stream of nitrogen gas over
three hours. The dried samples were stored at -20°C prior to HPLC-QTOF-MS analysis.
The samples were reconstituted in 1 mL of 1:4 acetonitrile:ddH2O, vortexed, and centrifuged at 14,000g for 10 minutes 4°C. The clear supernatant was transferred into a clean LC vial for LC-MS analysis. In total, 25 samples (5 samples for each of the four bean types and for rice) were prepared.
4.4. HPLC-QTOF-MS
Metabolomics analysis was performed using a 1260 Infinity HPLC coupled to a 6538
UHD Accurate Q-TOF MS system (Agilent Technologies, CA, USA) equipped with a electrospray ionization (ESI) source.
- 42 -
4.4.1. LC conditions for metabolites extraction
Urine and bean metabolites were separated on a 2.1 mm × 100 mm, 1.8 μm Zorbax SB-
Aq column (Agilent Technologies) with the column temperature maintained at 45°C.
Serum metabolites were separated on a 2.1 mm × 50 mm, 2.7 µm Poroshell column
(Agilent Technologies) with the column temperature maintained at 60°C. The mobile phase A was deionized water with 0.1% formic acid and the mobile phase B was acetonitrile with 0.1% formic acid. A 2 µL sample was injected. The HPLC flow rate was maintained at 0.4 mL/min for urine and bean metabolites and at 0.7 mL/min for serum metabolites. The gradient program was 0 min and 16 min with 2% and 98% of solvent B, respectively, for urine and bean metabolites while the gradient program was 0, 0.5 and 16 min with 30%, 30% and 100% of solvent B, respectively, for serum metabolites. There was a post-run time of 2 min with buffer before the next sample injection. The auto- sampler was maintained at a temperature of 5°C.
4.4.2. QTOF-MS conditions for metabolites extraction
The gas temperature for MS was maintained at 300°C while the drying N2 gas flow rate was kept at 11 L/min. The nebulizer pressure was maintained at 45 psig. Capillary voltage and fragmentor voltage were 4000V and 175V, respectively. The mass detection was operated by using electrospray with reference ions of m/z 121.050873 and
922.009798 for the positive mode. Targeted MS/MS mode was used to identify the significant metabolites. A full range mass scan from 50 to 1700 m/z was used. The data acquisition rate was maintained at 2 spectra/s.
- 43 -
4.5. Data processing
The workflow (Figure 1) applied for data processing comprised several algorithms used
by Agilent Mass Hunter Qualitative (MHQ, B.05) and by Mass Profiler Professional
(MPP, 12.6).
Figure 1. Metabolomics workflow. Features were extracted from samples by LC-MS and selected by MHQ. Raw data (“*.d”) was exported to compound exchange format files (“*.cef”) for statistical analyses in MPP.
When detected and separated in LC-MS, the raw data files were acquired and stored as
“*.d” files using Agilent Mass Hunter Acquisition software (B.05). Then “*.d” files were processed in MHQ where Molecular Feature Extraction, Generate Formulas and Export to CEF procedures were applied.
The molecular feature extraction was the first algorithm applied to raw ion chromatogram files and this is considered a premier extraction procedure. The parameters were set to allow the detection and extraction of features satisfying absolute abundances of more than 3,000 counts and to provide information regarding [M+H] +, isomers and their
+ + + corresponding NH4 , Na , and K adducts. Potential formulas were generated for the
extracted compounds by the Generate Formulas algorithm using collected information
- 44 -
such as retention time (RT), exact masses and abundances. After converting into
compound exchange format files (“*.cef ”) by the Export to CEF algorithm, extracted
features were ready for further comparative and statistical analyses by Mass Profiler
Professional (MPP, 12.6).
Individual “*.cef” files were created from the original individual “*.d” files and exported
into MPP software for statistical and differential analysis. Filter by Frequency and Filter
by Flags algorithms were used to implement acceptable frequency filtrations. The filters
were set to accept entities detected in at least 50% of one of the two conditions (baseline and week 8) or to accept entities present in at least one out of 5 samples for each bean or rice sample. This filtration step was implemented to ensure an elimination of potential feature extraction artifacts.
4.6. Statistical analyses
The prediction models were generated by a partial least square discrimination (PLSD) algorithm (MPP 12.6.1), and viewed as 3 dimensional graphs and Lorenz Curves.
One way ANOVA (P<0.05) followed by a post-hoc test (Tukey HSD) and an asymptotic
P value computation and a multiple-testing correction (Benjamini-Hochberg) were used to compare urine and serum metabolites in different groups (baseline with n=54 and n=22, 17, 15 for the three diet groups (urine); baseline with n=53 and n=20, 18, 15 for the three diet groups (serum)) and to compare all 4 bean and rice extracts.
Fold change analysis using paired T-Tests (P<0.05; ≥ 2 fold changes) was applied to examine statistical significance of 1) α-hydroxymetoprolol in urine samples of
- 45 -
participants with 0.6 cups of beans and 2) compounds present in beans and rice extracts
also detected in urine and serum samples.
An asymptotic P value computation and a multiple-testing correction (Benjamini-
Hochberg) were used to perform all paired T-Tests.
4.7. Identification of metabolites
Using an untargeted approach, a list of metabolites with at least two fold significant changes (P < 0.05) was generated and monitored against Metlin (>79,000 metabolites,
39,000 lipids and 168,000 peptides) databases. Identification of those compounds for which pure standards were available in databases was confirmed by mass spectra and retention time. Confirmation and identification of the select metabolites for which pure standards were not available in databases was done by comparison with MS/MS spectra from previous studies and by confidence scores.
4.8. Pathway analyses
Biochemical pathways of selected metabolites were analyzed using the KEGG (Kyoto
Encyclopedia of Genes and Genomes) pathway database in which standard pathways were available to understand the mechanism of action for targeted metabolites.
- 46 -
Chapter 5. Results
5.1. General results
5.1.1. Urinary metabolites
The detection and separation of metabolites in urine samples used an LC-QTOF-MS system, one of the several existing methods implemented in determination of total metabolites in biological fluids. The workflow for the metabolomics analysis is shown in
Step 1 of Table 9. The method employed here was successful to detect a total of 10,487 entities in urine samples from all PAD patients combined.
Table 9. Workflow of metabolomics and statistical analyses of urine of PAD patients. Step 1: Metabolomics of Urine in PAD Patients (LC-QTOF-MS and MHQ 7.01) Non-targeted analysis of all urine samples by LC-QTOF-MS in ESI+ mode Molecular Feature Extraction algorithm was used to extract all detectable compounds Generate Formulas algorithm was used to generate potential formulas for extracted compounds Export to CEF algorithm was used to convert “*.d” files into “*.cef ” files for further analyses
Step 2: Statistical Analysis (MPP 12.6.1) Partial least square discrimination algorithm on the entity list containing 10,487 entities in ESI+ mode (Figure 2), and generation of Lorenz Curves (Figure 3) Filter by Frequency algorithm was used to accept entities detected in at least 50% of one of the two conditions (baseline and week 8) One way ANOVA (P<0.05) followed by a post-hoc test (Tukey HSD) and an asymptotic P value computation and a multiple-testing correction (Benjamini-Hochberg) were applied to the filtered list (299 entities) to compare urine metabolites in different groups (baseline with n=54 and n=22, 17, 15 for the three diet groups) and this generated 9 significant compounds (Figure 6)
Step 3: Pathway Analysis Biochemical pathways for the 9 metabolites significantly changed in urine of PAD patients were identified using the KEGG database (Table 7)
- 47 -
For the first stage of the data analysis (Step 2, Table 9), a partial least square discrimination
(PLSD) algorithm was applied to the list containing all detected entities in urine to build a
prediction model. This prediction model can help to classify the detected entities based on the 3
treatments (rice (n=22), 0.3 (n=17) and 0.6 (n=15) cups mixed beans) vs. baseline values (n=54).
The generated PLSD graph showed a real diet effect (all 3 diets) after 8 weeks with no obvious
separation among beans and rice groups (Figure 2).
Figure 2. Partial least square discrimination (PLSD) of urine samples of PAD patients (n=54). Note: 1. Each circle on the graph represents one individual PAD patient; 2. Different colors represent the baseline and the 3 treatments.
As opposed to principal component analysis (PCA), PLSD is a supervised clustering method since the groups and individuals belonging to each group are known to the method and the algorithm is forced to find entities that can separate the clusters. It is therefore important to examine the quality and accuracy of the predictions when PLSD is used. In order to measure the quality of the predictive classifications, Lorenz Curves generated for each class (treatment) were examined (Figure 3a, b, c and d). As shown in these figures, a straight line for each treatment was obtained between 0 and 1 on Y axis, indicating true positive fractions, whereas all other
- 48 -
samples belonging to the other 3 treatments were portioned out (flattened on the horizontal line).
Thus, the Lorenz Curves showed that each of the 4 treatments may be partitioned out.
Figure 3. Lorenz Curves generated for different treatments (baseline (a), rice (b), 0.3 cups of cooked beans (c) and 0.6 cups of cooked beans (d)) used in the PLSD model for urine. Note: 1. The red points represent the individual PAD patients within each treatment; 2. Straight red lines between 0 and 1 on Y axis show the true positive fraction corresponding to each treatment; 3. The horizontal line shows all other PAD patients who do not belong to the selected treatment.
The PLSD graph showed a real effect of the 8 week diet intervention for all 3 treatments
compared to baseline. It is also important to note that the clusters related to the 3 diets were
spread over the Y axis while the baseline cluster seemed to be more compact thus showing a
lower degree of variation in urine samples in all 54 PAD patients combined at baseline.
- 49 -
5.1.2. Serum metabolites
Similar to the urine, serum entities were detected and determined by LC-QTOF-MS. A total of
20,805 entities were found in the serum samples of PAD patients using the method which was applied for detection and separation. Table 10 summarizes the workflow for the metabolomics and statistical analyses of the serum samples.
Table 10. Workflow of metabolomics and statistical analyses of serum of PAD patients. Step 1: Metabolomics of Serum in PAD Patients (LC-QTOF-MS and MHQ 7.01) Non-targeted analysis of all serum samples by LC-QTOF-MS in ESI+ mode Molecular Feature Extraction algorithm was used to extract all detectable compounds Generate Formulas algorithm was used to generate potential formulas for extracted compounds Export to CEF algorithm was used to convert “*.d” files into “*.cef ” files for further analyses
Step 2: Statistical Analysis (MPP 12.6.1) Partial least square discrimination algorithm on the entity list containing 20,805 entities in ESI+ mode (Figure 4), and generation of Lorenz Curves (Figure 5) Filter by Frequency algorithm was used to accept entities detected in at least 50% of one of the two conditions (baseline and week 8) One way ANOVA (P<0.05) followed by a post-hoc test (Tukey HSD) and an asymptotic P value computation and a multiple-testing correction (Benjamini-Hochberg) were applied to the filtered list (444 entities) to compare serum metabolites in different groups (baseline with n=54 and n=22, 17, 15 for the three diet groups) and this generated 15 significant compounds (Figure 7)
Step 3: Pathway Analysis Biochemical pathways for the 15 metabolites which were significantly changed in serum of PAD patients were identified using the KEGG database (Table 7)
A prediction model was established by the application of a PLSD algorithm to the list including all detected entities in serum. The classification produced by this prediction model for detected serum entities was in accordance with three treatments (rice (n=20), 0.3 (n=18) and 0.6 (n=15) cups mixed beans) vs. baseline (n=53). The PLSD (Figure 4) provides a three dimensional view of the real effect of all three diets after 8 weeks with no obvious separation among beans and rice groups.
- 50 -
Figure 4. Partial least square discrimination (PLSD) of serum samples of PAD patients (n=53). Note: 1. Each circle on the graph represents one individual PAD patient, 2. Different colors represent 4 treatments.
The classification of the serum metabolites seemed to suggest a different pattern for rice compared to bean groups as all grey circles seems to be lower compared to other groups along the Y axis. To measure the quality and accuracy of this predictive classification, the examination was done by Lorenz Curves generated for each treatment (Fig 5a, b, c and d). As shown in these figures, straight lines were obtained between 0 and 1 on the Y axis as true positive fractions for the classes of baseline and rice whereas all other samples belonging to the other classes were flattened on the horizontal line. However, the classes for the two beans groups were not straight along the slope; this showed that these two treatments may be not partitioned out perfectly.
- 51 -
Figure 5. Lorenz Curves generated for different treatments (baseline (a), rice (b), 0.3 cups of cooked beans (c) and 0.6 cups of cooked beans (d)) used in the PLSD model for serum. Note: 1. The red points represent the individual PAD patients within each treatment; 2. Straight red lines between 0 and 1 on Y axis show the true positive fraction corresponding to each treatment; 3. The horizontal line shows all other PAD patients who do not belong to the selected treatment.
The PLSD model showed a real effect of the 8-week diet intervention for all three treatments compared to baseline. Neither three treatments (0.3 or 0.6 beans or rice) nor baseline showed compact clusters indicating that there was a high degree of variation in serum samples from all the PAD patients. The class fractions obtained for bean groups were not satisfactory as steps were observed on the slope for both 0.3 and 0.6 cups of beans.
- 52 -
5.2. Significant results
5.2.1. Urinary metabolites with statistical significance
The workflow for the statistical analyses of urine results is shown in Step 2 of Table 9. The Filter
by Frequency algorithm was used to implement acceptable frequency filtrations. The filter was
set to accept entities detected in at least 50% of one of the two conditions (baseline and week 8).
A filtered list containing 299 entities was used for one way ANOVA analysis.
One way ANOVA and Tukey HSD tests were used to determine metabolites with significant
changes (P<0.05) in urine of PAD patients. These results are presented as Log 2 values obtained
for average abundances in each group and a normalization procedure using D-norvaline, the
internal standard. The urine analysis is based on a quantified volume (250 µL) of a urine sample obtained from participants on the mornings of their study visits; this was not a 24 h urine collection and the results have not been corrected for creatinine or osmolality. A heatmap was
generated to provide a visual graph depicting the 9 metabolites with statistically significant
changes (Figure 6). The statistical comparisons among the dietary groups and the associated
pathways for these compounds are shown in Table 11.
Based on the heatmap for the 9 detected significant metabolites (Figure 6), α-hydroxymetoprolol
was decreased in urine of PAD patients after 8 weeks of diet intervention compared to baseline,
with no obvious separation among beans and rice groups. Similar to α-hydroxymetoprolol, 5-L-
Glutamyl-L-alanine, Nα-methyl-L-histidine, L-glutamic acid n-butyl ester and 3-
methylthiopropanamine were decreased in the 3 diet groups compared to baseline.
N(alpha)-t-Butoxycarbonyl-L-leucine, a derivative of L-leucine, declined less in those
consuming 0.3 cups of mixed beans for 8 weeks compared to the other two groups, while 4-
- 53 -
Amino-6,7-dimethoxy-2-(1-piperazinyl)quinazoline and Leu Arg Ala decreased more in the groups consuming rice than beans.
Based on the statistical comparisons shown in Table 11, there were some differences between the bean and rice groups at 8 weeks. α-hydroxymetoprolol and 5-L-Glutamyl-L-alanine were reduced in the two bean groups compared to rice at 8 weeks. There were no statistical differences among the groups for creatinine. The rice and 0.6 cups bean groups had less
N(alpha)-t-Butoxycarbonyl-L-leucine and L-glutamic acid n-butyl ester compared to group consuming 0.3 cups beans. The rice group had elevated Nα-methyl-L-histidine compared to the other groups. 3-methylthiopropanamine was present at baseline but much lower in all groups at
8 weeks. 4-Amino-6,7-dimethoxy-2-(1-piperazinyl)quinazoline was reduced in all 3 groups at 8 weeks compared to baseline. Leu Arg Ala was elevated in the rice and 0.3 cups bean groups compared to 0.6 cups and baseline.
- 54 -
Figure 6. Heatmap of 9 selected significant (P<0.05) urinary metabolites affected by bean consumption. Note: 1. Each row represents one metabolite; 2. Each column represents one treatment and n is the sum of participants; 3. Red means higher concentration while blue means lower concentration.
- 55 -
Table 11. Urinary metabolites significantly modified by bean consumption and their associated pathways. 8 weeks Baseline 0.3 Cups 0.6 Cups Compound P (Corr.) Rice Associated Pathways References (n=54) of Beans of Beans (n=22) (n=17) (n=15) Metoprolol: Adrenergic signaling in cardiomyocytes; -3 a b c c KEGG DRUG: α-Hydroxymetoprolol 1.53 10 -2.24 -9.34 -6.44 -7.75 Neuroactive ligand-receptor D02358 interaction; Salivary secretion; Calcium signaling pathway KEGG COMPOUND: -8 a b c c C03740 5-L-Glutamyl-L-alanine 9.30 10 -0.49 -8.15 -5.48 -5.76 Glutathione metabolism KEGG REACTION:
- R01262
56 56 KEGG
- Arginine and proline metabolism; Creatinine - 0.82 0.64 0.77 1.81 COMPOUND: Metabolic pathways C00791 N(alpha)-t- KEGG -3 a c b c L-Leucine: Biosynthesis of amino Butoxycarbonyl-L- 1.21 10 8.52 2.76 4.88 2.85 COMPOUND: acids leucine C00123 KEGG N(alpha)- -2 b a b b Histidine and methionine 3.41 10 0.00 1.38 0.00 0.00 REACTION: methylhistidine metabolism R10523, R10524 L-Glutamic acid n-butyl -2 a c b d Tomita et al., 7.43 10 3.17 0.69 1.52 0.00 Sediment formation ester 1999 KEGG 3- -2 a b b b 5.30 10 0.25 0.00 0.00 0.00 Methionine metabolism REACTION: Methylthiopropanamine R00656 Doxazosin: 4-Amino-6,7- Vascular smooth muscle -2 a b b b KEGG DRUG: dimethoxy-2-(1- 6.49 10 2.72 1.37 1.49 1.99 contraction; D07874 piperazinyl)quinazoline Neuroactive ligand-receptor interaction;
Salivary secretion; Calcium signaling pathway -2 b a a b Amino acid, peptide and protein Leu Arg Ala 5.82 10 3.63 7.91 6.66 3.90 Metlin database metabolism Notes: 1. p (Corr) the corrected P values after Bonferroni Multiple Correction test (P<0.05) 2. Mean values (Log2 Normalized values) followed by the same letter (a, b, c, d) in the same row are not significantly different when a (probability level of <0.05 and Bonferroni multiple correction is applied -
57 57 -
5.2.2. Serum metabolites with statistical significance
The workflow of statistical analyses of serum is shown in Step 2 of Table 10. The Filter by Frequency algorithm was used to accept entities detected in at least 50% of one of the two conditions (baseline and week 8). A filtered list containing 444 entities was used for one way ANOVA analysis.
One-way ANOVA and Tukey HSD tests detected significant changes in 15 metabolites in serum as shown in the heatmap (Figure 7). Log 2 values obtained for average abundance for each compound in each group were used to generate the heatmap. Several of the
compounds in this heatmap had similar abundances in all 3 diets and therefore cannot be
directly related to bean consumption. However, several compounds had a lower
abundance in the rice group (N-Oleoyl-L-Serine, Glycocholic Acid, Syringomethyl
Reserpate, Methyl 10,13-dihydroxy-9-oxo-11-octadecenoate, LysoPE (0:0/20:0), and
GalNAcbeta1-4Galbeta1-4Glcbeta-Cer (d18:1/26:1(17Z)) compared to the bean groups and they were also lower in the baseline group compared to the rice groups, indicating that these compounds may be related to bean consumption.
Table 12 provides the statistical analyses among groups for the 15 metabolites modified by the dietary intervention and also the associated pathways for these compounds.
Several compounds were significantly increased in both bean groups compared to rice:
25-Hydroxyvitamin D2-25-glucuronide, Glycocholic Acid, Syringomethyl Reserpate,
Prostaglandin E2 p-acetamidophenyl ester, Nadolol, PGF2α diethyl amide, PS(P-
16:0/20:0) and GalNAcbeta1-4Galbeta1-4Glcbeta-Cer (d18:1/26:1(17Z). Both bean
groups had less N-Oleoyl-L-Serine compared to the rice group. Methyl 10,13-dihydroxy-
- 58 -
9-oxo-11-octadecenoate was present at baseline but not in the 8 week groups. Lys Ala
His and Lactone of PGF-MUM was present in the rice group but not the other groups.
The rice group had more Ubiquinone 8 than the other groups. The 0.3 cup beans group had more LysoPE (0:0/20:0) than the other groups while the 0.6 beans group had more
Isomethiozin than the other groups.
Figure 7. Heatmap of 15 selected significant (P<0.05) serum metabolites affected by bean consumption. Note: 1. Each row represents one metabolite; 2. Each column represents one treatment and n is the sum of participants; 3. Red means higher concentration while blue means lower concentration.
- 59 -
Table 12. Serum metabolites significantly modified by bean consumption and their associated pathways.
8 weeks Baseline 0.3 Cups 0.6 Cups Compound P (Corr.)† Rice Pathways References (n=53) of Beans of Beans (n=20) (n=18) (n=15) 25-Hydroxyvitamin D -25- -2 c b a a Shimada et al., 2 1.32 10 1.20 3.21 5.54 6.03 Vitamin D metabolism glucuronide 2 1997 Stimulate bone formation, Cayman N-Oleoyl-L-Serine 1.52 10-2 -2.24a -9.34b -6.44c -7.75c Inhibit bone resorption Chemical Item Number 13058 Primary bile acid biosynthesis; KEGG -2 c c a b Secondary bile acid Glycocholic Acid 6.28 10 2.90 2.59 4.78 3.11 COMPOUND: biosynthesis; C01921 Metabolic pathways; Bile secretion
- Winer and
60 60 -2 b b a a Syrosingopine and reserpine Syringomethyl Reserpate 0.75 10 -2.30 -2.12 0.41 0.35 Sahay, 1960;
- metabolism Suzuki Y., 1981 Methyl 10,13-dihydroxy-9- -2 a b b b 1.08 10 1.93 0.00 0.00 0.00 Fatty acid oxygenation Sessa et al., 1977 oxo-11-octadecenoate -2 b a b b Amino acid, peptide and Lys Ala His 3.41 10 0.00 1.38 0.00 0.00 Metlin database protein metabolism Ubiquinone and other terpenoid-quinone KEGG Ubiquinone 8 2.39 10-5 1.48b 3.76a 2.95b 1.37b biosynthesis; COMPOUND: Biosynthesis of secondary C17569 metabolites Cayman Prostaglandin E p- -5 c b a a 2 7.20 10 2.24 5.09 8.37 9.00 Reduce intraocular pressure Chemical Item acetamidophenyl ester Number 14053 HMDB11481 LysoPE (0:0/20:0) 1.35 10-3 3.40c 6.36b 10.88a 6.77b Lipid metabolism LIPID MAPS LMGP02050045 -6 c b b a KEGG Isomethiozin 3.87 10 2.95 8.08 9.74 13.07 Pesticides COMPOUND:
8 weeks Baseline 0.3 Cups 0.6 Cups Compound P (Corr.)† Rice Pathways References (n=53) of Beans of Beans (n=20) (n=18) (n=15) C19107 Calcium signaling pathway; Neuroactive ligand-receptor -10 c b a a interaction; KEGG DRUG: Nadolol 2.28 10 2.83 9.91 12.13 13.27 Adrenergic signaling in D00432 cardiomyocytes; Salivary secretion -3 b a b b Kitamura et Lactone of PGF-MUM 3.07 10 0.00 0.76 0.00 0.00 PGF metabolism 2α al.,1977 Cayman -5 c b a a PGF2α diethyl amide 1.37 10 0.25 3.86 4.62 6.00 Reduce intraocular pressure Chemical Item Number 16023 -4 d c b a LIPID MAPS PS(P-16:0/20:0) 1.03 10 3.51 5.78 7.65 10.03 Lipid metabolism LMGP03030018 -
61 61 GalNAcbeta1-4Galbeta1- -2 c c a b LIPID MAPS
- 4Glcbeta-Cer 1.10 10 0.84 1.33 2.56 1.75 Lipid metabolism LMSP0502AB08 (d18:1/26:1(17Z)) Notes: 1. p (Corr) the corrected P values after Bonferroni Multiple Correction test (P<0.05) 2. Mean values followed by the same letter in the same row are not significantly different when a probability level of <0.05 and Bonferroni multiple correction is applied
5.3. Administrated drugs and their metabolites
5.3.1. Detected metabolites of drugs
The second and the main objective of this Thesis was to assess the suitability of the LC-
QTOF-MS methods developed for biological fluids in our laboratory for rapid screening and detection of commonly administered drugs to PAD patients. Table 13 provides a list of administered drugs to PAD patients who took part in this bean trial. The administered drugs belonged to four main following categories: 1) cholesterol-lowering, 2) anti- hypertensive, 3) anti-platelet and 4) anti-hyperglycemic drugs.
Table 13. The list of medications administered to participants in the “Bean-PAD study”. Drug Categories Administered Drugs Metoprolol, HCTZ, Ramipril, Nifedipine, Furosemide, Telmisartan, Amlodipine, Losartan, Atenolol, Irbesartan, Cilazapril, Quinapril, Anti-hypertensive Triamterene, Diltiazem, Nadolol, Fosinopril, Perindopril, Enalapril, Trandolapril, Indapamide, Valsartan Simvastatin, Atorvastatin, Rosuvastatin, Pravastatin, Niacin Cholesterol lowering (Nicotinic acid), Fenofibrate Anti-hyperglycemic Metformin, Gliclazide, Glyburide, Sitagliptin, Repaglinide Anti-platelet Aspirin, Clopidogrel Symbicort (Budesonide/Formoterol), Flomax (Tamsulosin), Gabapentin, Diclofenac, Lorazepam, Dilantin (Phenytoin), Others Pentoxifylline, Prednisone, Omeprazole/Esomeprazole, Primidone, Finasteride, Ranitidine, Venlafaxine, COSOPT (Dorzolamide/Timolol eye drops), Citalopram, Cortisone.
The list of 40 detected drugs and their corresponding metabolites found in biological fluids of PAD participants using our developed methods is provided in Table 14. It is
- 62 -
important to study the effect of diet on these detected drugs and/or drug metabolites as it provides critical information on potential diet and/or nutrient-drug interactions in these patients. The current study provides preliminary data that can be used to evaluate the feasibility of this approach. A proper statistical analysis would require a sufficiently large sample size for each drug group and with sufficient numbers of PAD patients taking the same drug and allocated to the same diet group. However, this was not the case for the present study, and it must be noted that the allocation of PAD patients to the 3 different diet groups was performed in a random manner and not based on their administered drugs.
Among all drugs listed in Table 14, metoprolol (n = 23) and statins (n = 46 for all forms of statins combined) were the ones most administered. This Table also shows the breakdown of participants within each diet group for a given drug.
- 63 -
Table 14. Administered drugs and their corresponding metabolites detected in urine and serum of 62 PAD participants. Detected in Detected in Participants (n) Dosage Drugs Metabolites Formula m/z urine serum in each group (per day) RT‡ RT‡ Metoprolol C H NO 268.1907 3.91 0.26 Rice: 6 15 25 3 Metoprolol α-Hydroxymetoprolol C H NO 284.1856 3.85 0.27 0.3 cups bean: 9 25-200mg 15 25 4 (n=23) o-Demethylmetoprolol C H NO 254.1751 3.18 ND 0.6 cups bean: 8 14 23 3 Metoprolol acid C14H21NO4 268.1543 3.85 0.24 Rice: 1 Atenolol (n=2) 50-100mg Atenolol C H N O 267.1708 2.89 ND 0.6 cups bean: 1 14 22 2 3 Nadolol (n=1) Rice: 1 20mg Nadolol C17H27NO4 310.2018 4.41 0.41 Atorvastatin C33H35FN2O5 559.2603 6.61 3.55 p/o-OH-atorvastatin C H FN O 575.2552 5.85 3.22 Rice: 10 33 35 2 6 Atorvastatin o-OH-atorvastatin 0.3 cups bean: 5 5-80mg C H FN O 751.2893 5.26 10.57 (n=23) (glucuronide_1/2) 39 43 2 12 0.6 cups bean: 8 Atorvastatin Metabolite C H FN O 543.2642 ND 6.41 (ortho/para/meta) 33 35 2 4 Rice: 3 Rosuvastatin C H FN O S 482.1764 5.92 1.73 Rosuvastatin 22 28 3 6 0.3 cups bean: 6 5-40mg (n=16) Desmethyl rosuvastatin C H FN O S 468.1608 5.09 ND 0.6 cups bean: 7 21 26 3 6 -
64 64 Simvastatin Rice: 2 Simvastatin-6'-carboxylic acid C H O 449.2524 4.73 ND 10-40mg 25 36 7 (n=7) 0.3 cups bean: 5 6'-Exomethylenesimvastatin C H O 417.2633 ND 3.25
- 25 36 5 Pravastatin / 6'-Epipravastatin / C23H36O7 425.2537 ND 2.17 Pravastatin Rice: 1 3'α-Isopravastatin 10-20mg (n=2) 0.3 cups bean: 1 3''-Hydroxy Pravastatin C23H36O8 441.2488 ND 6.14 pravastatin dihydrodiol C23H38O9 459.2574 ND 6.14 Rice: 5 HCTZ (n=15) 0.3 cups bean: 7 12.5-125mg Hydrochlorothiazide C7H8ClN3O4S2 297.9721 3.54 ND 0.6 cups bean: 3 Ramipril C23H32N2O5 417.2388 5.01 0.59 Ramiprilat C H N O 389.2075 4.43 4.25 Rice: 5 21 28 2 5 Ramipril Ramipril glucuronide C H N O 593.2712 4.83 0.43 0.3 cups bean: 5 2.5-40mg 29 40 2 11 (n=12) Ramiprilat glucuronide C H N O 565.2413 4.17 7.60 0.6 cups bean: 2 27 36 2 11 Ramipril diketopiperazine C23H30N2O4 399.2282 6.65 2.95 Ramiprilat diketopiperazine C21H26N2O4 371.1961 6.45 1.21 Quinapril C25H30N2O5 439.2231 5.89 0.87 Quinaprilat C23H26N2O5 411.1913 5.42 0.39 Quinapril Diketopiperazine derivative of 0.6 cups bean: 3 20-40mg C H N O 421.2106 7.11 3.22 (n=3) Quinapril 25 28 2 4 Diketopiperazine derivative of C H N O 393.1815 6.65 1.54 Quinaprilat 23 24 2 4 Cilazapril (n=2) 0.3 cups bean: 2 5mg Cilazapril C22H31N3O5 418.2323 5.41 0.47
Detected in Detected in Participants (n) Dosage Drugs Metabolites Formula m/z urine serum in each group (per day) RT‡ RT‡ Fosinopril C30H46NO7P 564.3012 4.83 3.97 Fosinopril p-hydroxyfosinopril C H NO P 580.3062 ND 6.96 0.6 cups bean: 1 40mg 30 46 8 (n=1) Fosinoprilat C27H42NO6P 508.2808 4.17 12.92 p-hydroxyfosinoprilat C27H42NO7P 524.2791 5.65 6.12 Perindopril C H N O 369.2315 8.03 0.74 Perindopril 19 32 2 5 Rice: 1 4mg Perindopril glucuronide C H N O 545.2698 ND 0.53 (n=1) 25 40 2 11 Perindoprilat C17H28N2O5 341.2069 ND 11.06 Rice: 4 Irbesartan (n=8) 0.3 cups bean: 2 150-600mg Irbesartan C25H28N6O 429.2403 5.56 1.29 0.6 cups bean: 2 Telmisartan (n=2) Rice: 2 80mg Telmisartan C33H30N4O2 515.2453 5.18 0.94 Rice: 1 Candesartan (n=2) 16-32mg Candesartan C H N O 441.1668 5.95 1.40 0.6 cups bean: 1 24 20 6 3 Losartan C22H23ClN6O 423.1698 6.33 1.36 Losartan Metabolite (E3174) C22H21ClN6O2 437.1484 6.84 2.04 Losartan Metabolite (1H- Imidazole-2-propanol, 4-
- chloro-5-(hydroxymethyl)-a-
65 65 methyl-1-[[2'-(1H-tetrazo / C22H23ClN6O2 439.1643 6.11 0.73 - Losartan 0.3 cups bean: 1 Losartan Metabolite (1H- 50-100mg (n=2) 0.6 cups bean: 1 Imidazole-2,5-dimethanol,4- chloro-a2-propyl-1-[[2'-(1H- tetrazol-5-yl)[1,1'- Losartan Metabolite (b-D- Glucopyranuronicacid, 1-[5-[4'- C H ClN O 599.2016 5.04 1.14 [[2-butyl-4-chloro-5- 28 31 6 7 (hydroxymethyl)-1H-imi Valsartan (n=1) Rice: 1 120mg Valsartan C24H29N5O3 436.228 7.13 2.66 Rice: 1 Metformin (n=11) 0.3 cups bean: 4 250-2550mg Metformin C4H11N5 130.1088 1.46 0.22 0.6 cups bean: 6 Rice: 16 Aspirin (n=47) 0.3 cups bean: 16 81-325mg Aspirin C9H8O4 181.0504 2.80 1.20 0.6 cups bean: 15
Rice: 4 2-Oxoclopidogrel C16H16ClNO3S 338.0624 5.80 ND Clopidogrel (n=11) 0.3 cups bean: 4 75-150mg Clopidogrel Carboxylic Acid C H ClNO S 308.0508 5.39 0.32 0.6 cups bean: 3 15 14 2 Rice: 2 Amlodipine C H ClN O 409.1515 4.88 ND Amlodipine (n=7) 2.5-10mg 20 25 2 5 0.3 cups bean: 4 Dehydroamlodipine C20H23ClN2O5 407.1363 4.65 0.35
Detected in Detected in Participants (n) Dosage Drugs Metabolites Formula m/z urine serum in each group (per day) RT‡ RT‡ 0.6 cups bean: 1 Desmethyldehydroamlodipine C19H21ClN2O5 393.1172 ND 0.38 Nifedipine C H N O 347.1244 3.64 ND Nifedipine Rice: 2 17 18 2 6 30-60mg Dehydronifedipine C H N O 345.1099 ND 0.40 (n=3) 0.3 cups bean: 1 17 16 2 6 Desmethyldehydronifedipine C16H14N2O6 331.0928 5.13 0.37 Rice: 3 Furosemide (n=5) 0.3 cups bean: 1 20-40mg Furosemide C12H11ClN2O5S 331.0154 4.54 ND 0.6 cups bean: 1 Diltiazem C22H26N2O4S 415.1692 5.36 0.41 N-Monodesmethyldiltiazem C21H24N2O4S 401.1532 5.43 0.38 Diltiazem Rice: 1 Deacetyldiltiazem C20H24N2O3S 373.1589 5.12 0.29 180-300mg (n=2) 0.6 cups bean: 1 Deacetyl-N- monodemethyldiltiazem/Deacet C19H22N2O3S 359.1415 4.34 10.87 yl-O-demethyldiltiazem Triamterene C H N 254.1161 4.19 6.66 Triamterene (n=1) 0.3 cups bean: 1 50mg 12 11 7 p-Hydroxytriamterene C12H11N7O 270.1109 3.15 ND 0.3 cups bean: 3 Gliclazide C H N O S 324.1305 6.29 2.29 Gliclazide (n=4) 30-160mg 15 21 3 3 0.6 cups bean: 1 N-oxide gliclazide C15H21N3O4S 340.1329 4.95 0.71 -
Pentoxifylline Pentoxifylline C13H18N4O3 279.1448 3.58 0.33 66 66 Rice: 1 1200mg (n=1) Pentoxifylline alcohol C13H20N4O3 281.1527 4.57 0.31 - Rice: 1 Budesonide (n=3) 1 puff Budesonide C H O 431.2367 5.76 7.84 0.3 cups bean: 3 25 34 6 Rice: 1 Tamsulosin (n=5) 0.4mg Tamsulosin C H N O S 409.1790 4.34 0.28 0.3 cups bean: 4 20 28 2 5 Gabapentin Rice: 1 Gabapentin C H NO 172.1335 3.59 1.16 100-3600mg 9 17 2 (n=2) 0.3 cups bean: 1 N-methyl-Gabapentin C10H19NO2 186.1492 3.23 ND Rice: 1 Cis-3 / trans-4- Glyburide (n=4) 0.3 cups bean: 1 2.5-20mg C H ClN O S 510.1456 5.60 ND Hydroxyglyburide 23 28 3 6 0.6 cups bean: 2 Lorazepam Rice: 1 Lorazepam C H Cl N O 321.0191 4.78 ND 1-2mg 15 10 2 2 2 (n=3) 0.6 cups bean: 2 Lorazepam glucuronide C21H18Cl2N2O8 497.0516 4.77 0.47 Diclofenac Diclofenac C H Cl NO 296.0234 6.29 1.60 Rice: 1 75mg 14 11 2 2 (n=1) 5- / 4'-Hydroxydiclofenac C14H11Cl2NO3 312.0188 5.33 ND Phenytoin C15H12N2O2 253.0974 4.97 0.76 Phenytoin p-Hydroxyphenytoin 0.3 cups bean: 1 300mg C21H20N2O9 445.1248 4.25 8.91 (n=1) glucuronide Phenytoin-N-glucuronide C21H20N2O8 429.1313 4.95 5.91 Rice: 1 Prednisone / 20α- Prednisone (n=3) 0.3 cups bean: 1 4-5mg Dihydroprednisone / 20β- C21H28O5 361.2009 5.448 2.52 0.6 cups bean: 1 Dihydroprednisone
Detected in Detected in Participants (n) Dosage Drugs Metabolites Formula m/z urine serum in each group (per day) RT‡ RT‡ Omeprazole sulfone / 5- Omeprazole/ C H N O S 362.1169 4.774 0.53 0.3 cups bean: 2 Hydroxyomeprazole 17 19 3 4 Esomeprazole 20-40mg 0.6 cups bean: 7 (n=9) (S)-Esomeprazole / C H N O S 346.1223 4.72 0.34 Omeprazole 17 19 3 3
Primidone (n=1) 0.3 cups bean: 1 125mg p-Hydroxyprimidone C12H14N2O3 235.1091 3.11 0.21
Finasteride (n=2) 0.3 cups bean: 2 5mg o-Monocarboxylfinasteride C23H34N2O4 403.2593 6.11 0.58 Ranitidine C H N O S 315.1489 2.96 1.86 Ranitidine (n=1) 0.3 cups bean: 1 300mg 13 22 4 3 Desmethylranitidine C12H20N4O3S 301.1336 2.89 ND Note: 1. n is the number of participants taking the drug; ‡ 2. RT is the retention time of each entity monitored; 3. ND is not detected in bio-fluids. -
67 67 -
5.3.2. Metoprolol
In our study, metoprolol was administrated to 23 PAD patients including 9 individuals
participating in the group who received 0.3 cups of mixed beans/day five days per week,
8 participants in the group who received 0.6 cups of mixed beans/day five days per week
as well as 6 patients in the control group (rice). A paired T-test comparing urine samples
in PAD patients at 0 and 8 weeks for the 0.6 cups bean group (n= 8; P<0.05 followed by a
Bonfernoni multiple correction) revealed significant decreases in α-hydroxymetoprolol, a metabolite of metoprolol, after 8 weeks as presented in Figure 8. A reduction in exogenous metabolites such as α-hydroxymetoprolol suggests a diet-drug interaction that
could be affecting the metabolism of metoprolol.
Figure 8. α-hydroxymetoprolol in urine before and after 0.6 cups of bean consumption for 8 weeks (n=8 PAD Patients, Paired T-Test (P<0.05)).
- 68 -
5.4. Beans and rice extracts
Another objective of this Thesis was to screen beans and rice extracts used in this study, to explore the potential differences in their composition and to determine whether compounds in the beans and rice extracts were present in serum and urine. The workflow for the metabolomics and statistical analyses of the bean and rice extracts is summarized in Table 15.
Table 15. Workflow of metabolomics and statistical analyses in bean extracts.
Step 1: Metabolomics analysis of Bean and Rice Extracts (LC-QTOF-MS and MHQ 7.01) Non-targeted analysis of all bean and rice samples by LC-QTOF-MS in ESI Positive mode Molecular Feature Extraction algorithm was used to extract all detectable compounds Generate Formulas algorithm was used to generate potential formulas for extracted compounds Export to CEF algorithm was used to convert “*.d” files into “*.cef ” files for further analyses
Step 2: Statistical Analysis (MPP 12.6.1) Partial least square discrimination algorithm on the entity list containing 1,781 entities in ESI+ mode (Figure 9), and generation of Lorenz Curves (Figure 10) Filter by Flag algorithm was used to accept entities present in at least one out of 5 samples for each bean or rice sample and generated lists contained 127, 105, 183, 230 and 139 out of 1,781 entities for navy, kidney, black and pinto beans, and rice, respectively. A Venn diagram was used to select and to identify the compounds that were common or different among the different bean types (Figure 11) and this generated 11 compounds common to all 4 bean types (Table 16) One way ANOVA (P<0.05) followed by a post-hoc test (Tukey HSD) and an asymptotic P value computation and a multiple-testing correction (Benjamini-Hochberg) were applied to entire list (1,781 entities) to compare compounds in different bean varieties and rice and this generated 125 significant compounds (Figure 12) Paired T-Tests (P<0.05) was applied to examine statistical significance of compounds present in beans and rice extracts also detected in urine and serum of PAD patients, 11 compounds out of 125 sig. ones were also significantly changed in biological samples of PAD patients
Step 3: Pathway Analysis Biochemical pathways for the 11 bean compounds significantly changed in bio-fluids of PAD patients were identified using KEGG and HMDB database (Table 13)
A total of 1,781 entities were detected in all beans and rice extracts using the LC-QTOF-
MS method employed. A ‘filtration by flag’ algorithm was applied to select entities with
- 69 -
high concentrations in each type of bean. The filter was set to accept entities present in at least one out of 5 samples for each bean or rice sample. The resulting entity lists contained 127, 105, 183, 230 and 139 out of 1,781 entities for navy, kidney, black and pinto beans, and rice, respectively.
To classify the five different diet components (pinto, black, kidney, navy beans and rice)
based on the average values of all detected entities, a predictive model was generated
using the PLSD algorithm. This algorithm was applied to all detected entities in all 4
beans and rice extracts. The PLSD plot (Figure 9) showed a clear separation among the
extracts for the different beans and the rice. Based on this graph, pinto and navy beans
appeared to be clustered differently compared to rice and other two bean types.
Figure 9. PLSD of bean and rice extracts. Each circle on the graph represents one replicate for individual bean or rice extracts, with 5 replicates for each bean type and for rice.
In addition, Lorenz Curves (Fig 10a, b, c, d and e) were generated for each class (4 beans
and rice) and examined for the predictive classification. The Lorenz curves provided
additional support for the cluster separations observed in the PLSD graph; straight lines
- 70 -
for each bean type and rice were obtained between 0 and 1 on the Y axis indicating true positive fractions whereas all other samples belonging to other classes were flattened on the horizontal line.
Figure 10. Lorenz curves for different treatments (rice (a), pinto bean (b), kidney bean (c) navy bean (d) and black bean (e)) used in the PLSD model.
Note: 1. The red points represent the individual bean or rice samples; 2. Straight red lines between 0 and 1 on Y axis showed the true positive fraction corresponding to each treatment; 3. The horizontal lines show all other classes that do not belong to the selected treatment.
A Venn diagram (Figure 11) was used to select and to identify the compounds that were common or different among the different bean types. In the Venn diagram, 11 entities seemed to be commonly present in all 4 bean types and these are listed in Table 16. Two of the 11 compounds listed in Table 16, 1-O-Feruloyl-β-D-glucose and (1S,4S)-4-
- 71 -
Hydroxy-3-oxocyclohexane-1-carboxylate were found in high concentrations in navy and pinto beans compared to black and kidney beans.
Figure 11. Venn diagram for the 4 types of beans.
Table 16. Compounds found in black, kidney, navy and pinto beans. Bean Extracts Compound ‡ m/z Black Kidney Navy Pinto † 1-O-Feruloyl-β-D-glucose 357.1107 -0.27 -4.27 1.20 1.10 2-(4-Methyl-5-thiazolyl)ethyl octanoate 270.1449 -2.24 -2.52 -6.63 0.01 Isopropyl tetradecanoate 271.4507 -0.28 -0.74 0.84 0.88 Dikegulac 275.2671 -0.33 -0.70 0.38 0.21 Furmecyclox 252.3214 -6.45 -3.02 -9.52 -6.32 6-Hydroxy-4-nonadecanone 299.2872 -9.52 -3.46 0.82 0.82 4-Hydroxy-6-methylpyran-2-one 127.0317 0.04 -6.04 -0.11 0.46 Dioctyl phthalate 391.5561 -6.59 -2.90 -2.54 1.80 Esmolol 296.1784 -3.10 -3.14 -9.43 -6.21 Sorbitan laurate 247.4601 -11.35 -7.87 1.07 0.97 (1S,4S)-4-Hydroxy-3-oxocyclohexane-1- 159.1519 -4.52 -1.13 1.23 1.24 carboxylate Note: ‡ The Identification is performed by MS Spectra comparisons using the Metlin database † All values are the Log2 normalized values for average abundances
- 72 -
While the Venn diagram used above was effective to help summarize the main common compounds among all 4 beans extracts it is critical to determine the major statistical differences among compounds in the bean and the rice extracts. To this end, one way
ANOVA (P<0.05) followed by Tukey HSD tests was used to generate a list of 125 entities with significant different among the extracts for the 4 types of beans and rice. Those differences are presented in the heatmap in Figure 12.
Figure 12. Heatmap of 125 compounds with significant differences among extracts for 4 types of beans and rice Note: 1. Each row represents one entity; 2. Each column shows one food extract; 3. The concentration for each entity (Log2 Normalized values) is color coded with a range from -14.1 to 14.1. Red means a higher concentration and blue means a lower concentration.
As shown in this heat map, many compounds were found at high concentrations in pinto
and navy beans. This confirms the PLSD cluster separations corresponding to pinto and
navy beans shown in Figure 11.
- 73 -
Table 17 summarizes several compounds that were found not only in bean and/or rice extracts but also in either or both biological fluids of participants who consumed the corresponding beans or rice.
Among the 30 compounds listed, 9 were also present in rice and will not be considered as specific compounds related to beans or discussed in this Thesis.
Most of the 125 compounds were detected in urine and/or serum of the PAD participants, but only 11 compounds showed significant changes in the biological samples. Notably, 2- Dodecylbenzenesulfonic acid, Absinthin, Absinthin and 3α,7α-Dihydroxy-5β- cholestanate were increased in the serum after bean consumption (all compounds for 0.3 cups beans versus baseline, and only 3α,7α-Dihydroxy-5β-cholestanate for 0.6 cups beans versus baseline). In the urine, there was an increase in 9-(beta-D- Ribofuranosyl)zeatin and Dimethyl trisulfide in those consuming 0.3 cups beans versus baseline. Other compounds decreased in the urine and some of these were not specific to bean consumption.
Table 18 provides the Pathways for the 11 compounds with significant changes in the biological samples.
- 74 -
Table 17. The distribution of detected entities in rice, bean, and biological extracts of PAD patients.
Log2 Normalized Mean Values Bio-fluids of PAD Patients compounds m/z (Week 8 vs. Baseline) Rice Black Kidney Navy Pinto Urine Serum 1,2-Dimethoxy-4-[2-(2-propenyloxy)ethenyl]-benzene 221.2643 ND ND ND ND 8.95 ↓α √ 16-Oxo-palmitate 271.2195 ND ND ND 12.59 12.34 ↓α, β, γ √ 2-Dodecylbenzenesulfonic acid 327.1915 ND ND ND ND 14.94 √ ↑α 3-(8,11,14-Pentadecatrienyl)phenol 299.2297 ND ND ND ND 12.14 ↑α, β, γ ND 4Z,7Z,10Z-octadecatrienenitrile 260.4296 ND ND ND ND 9.34 ↑α, β, γ √ 9-(beta-D-Ribofuranosyl)zeatin 352.1542 ND ND ND 16.66 16.63 ↑ α √ Absinthin 497.6350 ND ND ND ND 9.51 ↓α, β, γ ↑α Ammothamnine 265.3633 10.25 ND ND 17.34 6.94 √ ↑α
- cis-1,2-Diphenylcyclobutane 209.1252 ND ND 3.29 ND 12.47 ↓α, β, γ √
75 75 Dimethyl trisulfide 127.2640 ND ND ND 16.85 6.47 ↑α ND -
Hexadecyl Acetyl Glycerol 359.5558 ND ND ND 16.11 16.47 ↓α, β ↑α, β 3α,7α-Dihydroxy-5β-cholestanate 435.6517 ND ND ND ND 8.55 √ √ 5-Pentadecylresorcinol 321.5093 ND ND 3.28 9.35 15.10 √ √ 6''-O-(3-Hydroxy-3-methylglutaroyl)astragalin 593.5022 10.77 ND ND 11.09 18.17 √ √ 7-beta-D-Glucopyranosyloxybutylidenephthalide 367.3625 ND ND 3.08 3.32 15.87 √ √ Armillaramide 556.9160 ND ND ND ND 9.17 √ √ Avermectin A2a monosaccharide 761.9504 ND ND ND ND 8.98 ND √ Butamben 194.1102 -2.86 -14.51 -11.55 0.54 0.54 √ √ Palmitic acid 257.4241 ND ND ND 9.32 11.93 √ √ Cerulenin 224.2695 -4.85 -1.86 -4.60 8.38 8.62 √ √ Erinacine D 479.6182 ND ND ND ND 8.69 √ √ Etamiphylline 280.3381 ND ND ND 9.99 13.36 √ ND Garcinia lactone dibutyl ester 303.3203 ND ND ND 3.22 12.58 √ √
Log2 Normalized Mean Values Bio-fluids of PAD Patients compounds m/z (Week 8 vs. Baseline) Rice Black Kidney Navy Pinto Urine Serum Goyaglycoside c 663.8935 -2.09 -8.54 -5.71 6.17 9.78 √ √ Isopropyl tetradecanoate 271.4507 -0.19 -0.28 -0.74 0.84 0.88 √ √ Magnoshinin 415.4914 -5.81 -8.89 -15.20 2.97 2.86 √ √ Netilmicin 476.5795 0.85 0.66 -2.33 14.07 14.06 √ √ O-Acetylcyclocalopin A 339.3524 -6.14 -9.29 -3.17 8.08 7.84 √ √ Palmitic amide 256.4393 ND ND ND ND 8.11 √ √ Stearic acid 285.4772 ND ND ND 6.19 11.96 √ √
Notes: √: Detected in the sample; ND: Not Detected in the sample; -
76 76 ↑: Significantly increased; ↓: Significantly decreased; paired T-test, p<0.05; α: Significantly changed in 0.3 cups of beans group versus baseline, paired T-test (n=17 in urine; n=18 in serum); p<0.05; - β: Significantly changed in 0.6 cups of beans group versus baseline, paired T-test (n=15 in urine; n=15 in serum), p<0.05; γ: Significantly changed in control (rice) group versus baseline, paired T-test (n=22 in urine; n=20 in serum), p<0.05.
Table 18. Pathways of 11 significant compounds in beans and bio-fluids of PAD patients. Compounds Pathways References KEGG COMPOUND: 1,2-Dimethoxy-4-[2-(2-propenyloxy)ethenyl]-benzene - C12263 KEGG COMPOUND: 16-Oxo-palmitate Cutin, suberine and wax biosynthesis; metabolic pathways C19614 Benzenesulfonic Acid Derivatives, a detergent in food 2-Dodecylbenzenesulfonic acid HMDB31031 processing 3-(8,11,14-Pentadecatrienyl)phenol A constituent of lipids HMDB33862 KEGG COMPOUND: 4Z,7Z,10Z-octadecatrienenitrile Fatty nitriles C13832 9-(beta-D-Ribofuranosyl)zeatin Cytokinins HMDB30388 Absinthin Anti-inflammatory agent, plant metabolite CHEBI:2366 Ammothamnine Alkaloid metabolite CHEBI:2672 cis-1,2-Diphenylcyclobutane Styrene dimer HMDB31821 Dimethyl trisulfide Volatile compound HMDB13780 Hexadecyl Acetyl Glycerol Inhibitory role in protein kinase C signaling CAS 77133-35-8 -
77 77 -
Chapter 6. Discussion
6.1. Endogenous metabolites
In urine, there were 9 metabolites significantly affected by consumption of study food by
the 62 PAD patients. Some of these compounds are involved in metabolism of amino
acids thereby could be affecting progression of PAD.
Both bean groups had a greater abundance of 5-L-Glutamyl-L-alanine in urine compared to the rice group (Table 11). The suggested formation route for 5-L-Glutamyl-L-alanine is a reaction catalyzed by ‘γ-glutamylcyclotransferase’ where glutathione loses a glutamic acid moiety to generate a dipeptide and Cys-Gly as shown below:
Glutathione + L-Amino acid Cys-Gly + (5-L-Glutamyl)-L-amino acid
The mammalian enzyme is part of the cell antioxidant defense mechanism. It initiates
extracellular glutathione (GSH) breakdown, provides cells with a local cysteine supply
and contributes to maintain intracellular GSH levels (Okada et al., 2006; Boanca et al.,
2007). In addition, this cell antioxidant defense system contributes to regulation role of
glutathione in NO circle. The up-regulation of NO leads to improvement of
atherosclerosis and hypertension (Lapenna et al., 1998; Robaczewska et al., 2016). The
bean groups had more 5-L-Glutamyl-L-alanine in urine than the rice group and this might
indicate increased formation of glutathione in the body thereby improving vascular
function and reducing hypertension of PAD patients.
- 78 -
Other urinary metabolites, such as Nα-methylhistidine, which was present at higher
abundance in the rice group compared to the bean groups (Table 11) is generated via a methylation step by S-Adenosyl-L-methionine which can go through a second methylation to produce N,N-Dimethylhistidine:
S-Adenosyl-L-methionine + L-Histidine S-Adenosyl-L-homocysteine
+ Nα-Methylhistidine
S-Adenosyl-L-methionine + Nα-Methylhistidine S-Adenosyl-L-homocysteine
+ N,N-Dimethylhistidine
These reactions are catalyzed by L-histidine Nα-methyltransferase (EC 2.1.1.44).
In addition, 3-Methylthiopropanamine, which was present at higher abundance at
baseline (Table 11) participates in the following reaction which might contribute to the
histidine and methionine metabolism.
L-Methionine 3-Methylthiopropanamine + CO2
The reactions related to methionine metabolism are also associated with homocysteine
metabolism. As a risk factor of PAD, hyperhomocysteinemia promotes atherosclerosis by
- 79 -
damaging endothelial cells, increasing vascular smooth muscle cell growth and platelet
adhesiveness, enhancing oxidation and deposition of LDL-cholesterol in the arterial wall,
as well as activating the coagulation cascade (Fonseca et al., 1999). Therefore,
consumption of study food (both rice and beans) might affect the metabolism of
methionine and homocysteine thereby influencing the progression of PAD.
There were no significant differences among the groups for urinary creatinine based on
metabolomics analysis (Table 11). This indicates that reporting the samples on a per
volume basis was not a confounding factor. However, it should be noted that participants
provided a spot urine sample when they attended their morning study visits and this may
or may not have been their first voided urine sample for the morning.
In serum of 62 PAD individuals, several classes of compounds such as prostaglandins,
bile acid, coenzyme and lipids were also significantly affected by bean consumption
(Table 12).
For instance, PGF2α diethyl amide is an analog of PGF2α and PGE2 p-acetamidophenyl
ester is a derivative of PGE2, and both of these compounds were elevated in the serum of the bean groups versus the rice group (Table 12). PGF2α and PGE2 are primary COX
products (prostaglandin) of arachidonic acid which have ocular hypotensive potency
(Bito, 1984). Bito (1984) compared the effect of 50 µl application of various PGs (PGE2,
PGF1α, PGD2, PGI2 and PGF2α), and their isomers or analogues on intraocular pressure
(IOP) of cats. PGE2 had the highest ocular hypotensive potency followed by PGF2α.
However, the E type of PGs has been reported to have serious side effects on IOP such as
transient ocular hypertension due to local vasodilatation which leads to sudden shifts in
intraocular volume (Waitzman and King, 1967) as well as increased permeability of the
- 80 -
blood-aqueous barrier (Beitch and Eakins, 1969). Therefore, the F type of PGs has been
suggested to be administrated as a long-term ocular hypotensive agent. Also, PGF2α
esters (methyl-, ethyl- and isopropyl-esters) are reported to have at least 10-fold higher
potency in reduction of ocular hypertension than PGF2α itself (Bito, 1984). In addition,
Woodward et al. (2001) reported that a dose of 0.03% bimatoprost (PGF2α diethyl
amide), which is administrated as an anti-glaucoma drug, had high ocular hypotensive
efficiency and long-term action in both animal and human studies with reduction of IOP
by approximately 35%-50%. Bimatoprost reduces IOP as a prostamide by increasing
aqueous humor outflow through trabecular and uveoscleral pathways (Woodward et al.,
2001).
The chemical structures of PGF2α, PGE2 and their derivatives detected in serum of PAD patients are shown in Figure 13.
- 81 -
Figure 13. Structures of PGF2α, PGE2 and their derivatives.
Instead of the carboxyl group of PGF2α and PGE2, a diethyl amide group is added to
PGF2α to produce PGF2α diethyl amide and an ester group of acetamidophenyl is added to PGE2 to produce PGE2 p acetamidophenyl ester. Those two derivatives were
significantly increased in serum of PAD patients after diet intervention with beans (Table
12). Thus, bean consumption may benefit PAD patients who have hypertension especially
ocular hypertension.
Glycocholic acid, a secondary bile acid which is involved in biosynthesis and secretion of
bile salts, was increased in the serum of both bean groups compared to rice. Bean
consumption might stimulate production of bile acids thereby enhancing excretion of
cholesterol from the body of PAD patients.
Ubiquinone 8 was elevated in the rice group compared to both bean groups and to
baseline (Table 12). Ubiquinone 8 is Coenzyme Q8 which is involved in pathways of
ubiquinone and the biosynthesis of other terpenoid-quinones and secondary metabolites
as well as the following reaction:
2-Octaprenyl-3-methyl-5-hydroxy-6-methoxy-1,4-benzoquinone + S-Adenosyl-L-
methionine Ubiquinone-8 + S-Adenosyl-L-homocysteine
Thus, it has been considered as a biomarker of a biological state. Our results suggested that consumption of rice might be more effective than beans for elevating Ubiquinone 8 in serum and thus promoting the conversion from methionine to homocysteine and inhibiting the progression of atherosclerosis in PAD.
- 82 -
PAD patients consuming beans had less N-oleoyl-L-serine in serum compared to the rice group, but more than at baseline (Table 12). N-oleoyl-L-serine is an endogenous lipid endocannabinoid present in trabecular bone, brain and blood of mammals; in addition, it has been reported to regulate bone remodelling and bone mass through stimulating bone formation (osteoblasts) and inhibiting bone resorption (osteoclasts) via CB2 receptor
signalling (Bab et al., 2008; Smoum et al., 2010). Therefore, N-oleoyl-L-serine and its
derivatives might be developed as a novel anti-osteoporotic drug. Higher concentrations
of N-oleoyl-L-serine in serum of PAD patients consuming beans may thereby be
benefiting the bone health of these patients.
Syringomethyl reserpate was elevated in the serum of both bean groups compared to the rice group and baseline (Table 12). Syringomethyl reserpate is a common metabolite of syrosingopine and reserpine in blood of animals and humans; furthermore, the hypotensive activity of syrosingopine and reserpine as well as syringomethyl reserpate has been confirmed in animal trials but there is a lack of human studies (Winer and
Sahay, 1960; Suzuki Y., 1981). The consumption of beans seemed to increase its concentration in serum of PAD patients.
6.2. Exogenous metabolites (administered drug)
Metoprolol is given more attention in this discussion simply because of its high frequency of use and its random distribution among diet groups provided an opportunity to test for statistical changes. Metoprolol and its 3 known metabolites, α- hydroxymetoprolol, O-demethylmetoprolol and metroprolol acid were successfully detected in both urine and serum of PAD patients (Table 14). Compared to metoprolol, α-
- 83 -
hydroxymetoprolol contains a hydroxyl group while in o-demethylmetoprolol one hydrogen atom is replaced by a methyl group. The hydroxyl group in O- demethylmetoprolol is replaced by a carboxylic acid group in metroprolol acid (Figure
14). In fact, the fat soluble metoprolol can be converted into more water-soluble forms for their ultimate urinary excretion.
Figure 14. Structures of metoprolol and its 3 metabolites: α-hydroxymetoprolol, o-demethylmetoprolol and metroprolol acid.
In humans, metoprolol is primarily metabolized by hepatic cytochrome P450 2D6
(CYP2D6) to two main metabolites, α-hydroxymetoprolol and O-demethylmetoprolol, which contribute 5% of the β1-blocking activity of metoprolol (Prakash and Markham,
2000). The main metabolic pathways of metoprolol are O-demethylation, oxidation and
α-hydroxylation; in addition, the O-demethylmetoprolol undergoes further oxidation to produce metroprolol acid which accounts for the elimination of approximately 65% of the dose in humans (Cerqueira et al., 2003). Thus, metoprolol has a bioavailability of approximately 50% due to hepatic first pass metabolism (Prakash and Markham, 2000).
- 84 -
α-Hydroxymetoprolol is a metabolite of metoprolol, an anti-hypertensive drug which acts
as β1 receptor blocker and is commonly administrated to PAD patients. The main known
pathways targeted by this drug include: the adrenergic signaling pathway in
cardiomyocytes, neuroactive ligand-receptor interaction, salivary secretion and calcium signaling pathways. The concentration of α-hydroxymetoprolol was decreased in the urine of PAD patients consuming of 0.6 cups of beans for 8 weeks compared to baseline
(Figure 8) indicating a diet (bean) – drug interaction. This result may predict the increased formation of metoprolol acid, or may suggest the enhanced bioavailability of metoprolol.
6.3. Compounds in beans
The PAD patients consumed the 4 kinds of beans throughout the intervention period, thus, the common compounds which are present in all 4 beans were studied first.
Among the 11 compounds common to all 4 beans (Table 16), one of the most interesting compounds found in these bean extracts was ‘esmolol’, a known β1-Adrenergic blocker with similar actions as for metoprolol. Esmolol is rapidly metabolized by hydrolysis of the ester linkage, chiefly by the esterases in the cytosol of red blood cells and not by plasma cholinesterases or red cell membrane acetylcholinesterase (DB00187). Although not present in high concentrations, this compound may explain the benefits of mixed beans on PAD.
One of the highly concentrated compounds in navy and pinto beans was 1-O-Feruloyl-β-
D-glucose (Table 16), which is associated with several known reactions including:
- 85 -
1) Amaranthin + 1-O-Feruloyl-beta-D-glucose <=> Celosianin II + D-Glucose
catalyzed by amaranthin feruloyltransferase
2) Betanin + 1-O-Feruloyl-beta-D-glucose <=> Lampranthin II + D-Glucose
catalyzed by betanin feruloyltransferase
3) UDP-glucose + Ferulate <=> 1-O-Feruloyl-beta-D-glucose + UDP catalyzed
by ferulate D-glucosyltransferase.
If the identity of this compound is further confirmed using different techniques, it might be an important biomarker in beans for glucose metabolism.
The second highly concentrated compound in Table 16 was (1S, 4S)-4-Hydroxy-3- oxocyclohexane-1-carboxylate, a substrate for the following reaction catalyzed by hydroxycyclohexanecarboxylate dehydrogenase:
(1S, 3R, 4S)-3, 4-Dihydroxycyclohexane-1-carboxylate + NAD+ <=> (1S, 4S)-4-
Hydroxy-3-oxocyclohexane-1-carboxylate + NADH + H+
These two compounds were high in pinto and navy beans and low in black and pinto
beans, but they were not the best markers to monitor bean consumption due to their
absence in bio-fluids of PAD patients.
One of the interesting aspects of the non-targeted approach is that any given compound detected in a food item may be tracked in biological samples. This might be a challenging task as most compounds present in foods might be metabolized and/or chemically modified due to their interactions with other endogenous and exogenous compounds and thus will not be present in biological samples without modifications. However, the non- targeted approach will help to identify potential compounds as a key signature of a given
- 86 -
food that can be used as a compliancy test for specific nutritional interventions and/or to understand how various nutrients or other compounds are metabolized. Different algorithms used in MPP and MHQ software allow one to build a database to track similar compounds and/or their metabolites in different samples.
Among 125 significantly different compounds present in the 4 beans (Figure 12), 11 were significantly changed in biological samples of PAD patients (Table 17). Although discussing the biological significance of every single compound listed in Table 17 is beyond the scope of this Thesis, a few key compounds will be covered in this section.
One interesting compound, 16-Oxo-palmitate, was present in high concentrations in navy and pinto beans and significantly decreased in urine of all PAD groups after 8 weeks of bean or rice consumption.
This compound has been associated with the following two reactions:
1) 16-Hydroxypalmitate <=> 16-Oxo-palmitate
2) 16-Oxo-palmitate <=> Hexadecanedioate
Interestingly, hexadecanedioate has been reported to have anti-tumor activity in vitro
(You et al., 2004).
Another compound detected in navy and pinto bean extracts was 9-(beta-D-
Ribofuranosyl) zeatin, and it was increased in urine of the 0.3 cups bean group after 8 weeks. This compound is not unique to beans; it has been previously reported in kiwifruit
(Moncalean et al., 2001). Zeatin is a known plant hormone (cytokinin) derived from the purine adenine and has been reported to have in vitro anti-aging effects on human skin fibroblasts (Rattan and Sodagam, 2005).
- 87 -
Absinthin is a naturally produced sesquiterpene lactone which has been reported to have
anti-inflammatory activity in lung cells (Guo et al., 2015). This compound was found in high concentrations in pinto bean extract, but it was significantly decreased in urine of all participants and significantly increased in serum of the 0.3 cups of beans group after 8
weeks (Table 17).
Hexadecyl acetyl glycerol was found in navy and pinto beans with significant decreases
and increases in urine and serum samples, respectively. It has been reported to inhibit the
growth of the human promyelocytic leukemia cell line (HL-60 cells) and induce
differentiation in cells resembling mononuclear phagocytes (McNamara et al., 1984) via
its inhibitory roles in protein kinase C signaling (Daniel et al., 1988).
Absinthin and hexadecyl acetyl glycerol were significantly decreased in urine and
increased in serum of PAD patients and highly found in pinto bean, therefore, they may
be considered as biomarkers of pinto bean consumption. In addition, 9-(beta-D-
Ribofuranosyl) zeatin and hexadecyl acetyl glycerol were highly detected in pinto and
navy beans and may be investigated as candidates of pinto and navy bean consumption in
the future studies.
Although there were no significant changes in urine or serum of PAD participants 6''-O-
(3-Hydroxy-3-methylglutaroyl) astragalin might be of interest (Table 17). It belongs to the class of organic compounds known as flavonoid-3-o-glycosides, and its other names are kaempferol 3-[6''-(3-hydroxy-3-methylglutaryl) glucoside] or kaempferol 3-[6-O-(3- hydroxy-3-methylglutaroyl) glucoside]. This 3-o-glycoside of kaempferol was mostly found in pinto bean, but also found in navy beans and rice; furthermore, it was successfully detected in both urine and serum (Table 17). The cholesterol-lowering effect
- 88 -
of kaempferol and its 3-o-glycoside form have been reviewed in Chapter 2. The results
from this study confirmed that pinto beans have the highest concentrations of kaempferol, which agrees with previous reports (De Lima et al. 2014) (Table 2).
In general, the non-targeted metabolomics approach applied in this study revealed that daily consumption of beans might affect the metabolism of several classes of compounds in bio-fluids of PAD patients, however, in some cases, similar responses were observed with the rice group (Figure 2 & 4, Tables 10 & 12). Compounds of interest that were specific to the bean groups and suggested for further study are urinary 5-L-Glutamyl-L- alanine (both doses of beans), N(alpha)-t-Butoxycarbonyl-L-leucine and L-Glutamic acid n-butyl ester (0.3 cups beans)(Table 11), and serum 25-Hydroxyvitamin D2-25-
glucuronide, Glycocholic Acid, Syringomethyl Reserpate, Prostaglandin E2 p-
acetamidophenyl ester, Nadolol, PGF2α diethyl amide, PS(P-16:0/20:0), and
GalNAcbeta1-4Galbeta1-4Glcbeta-Cer (d18:1/26:1(17Z))(Table 12).
- 89 -
Chapter 7. General Discussion
A non-targeted metabolomics approach was used to investigate potential significant changes in biological fluids (urine and serum) in PAD patients before and after 8 weeks of bean consumption. Any given food or food item is expected to contain several
hundreds and/or thousands of chemical compounds. Upon ingestion, hundreds of
compounds are digested, absorbed and metabolized. Biological fluids such as serum and
urine are rich sources of information to detect and develop profiles of these compounds.
Although specific analytical methods may be developed and optimized for each given
compound using sophisticated targeted analytical approaches, a non-targeted
metabolomics approach provides a rapid tool to detect a large number of these
compounds in a single run.
In this Thesis, biological fluids collected at baseline and after 8 weeks from 62 patients
with PAD who were randomly assigned to receive beans-free (beans were replaced by
rice), or 0.3 or 0.6 cups of mixed beans per day for 5 days per week were analyzed using
a non-targeted metabolomics approach.
In total, thousands of entities were successfully detected in both biological fluids. Of
those entities, some has been identified in the literature as 'metabolites' or 'compounds',
while the rest are breakdown products of identified compounds and many are unknown.
The Human Metabolome Database (HMDB) (Bouatra et al., 2013; Psychogios et al.,
2011) reports a total number of 345 urinary and 281 serum metabolites that are routinely
identified in general populations by using various metabolomics approaches such as GC-
MS, NMR, ICP-MS and DFI/LC-MS/MS. Our methods were successful to detect 273 of
- 90 -
the 345 reported urinary compounds (Appendix I) and 194 of the 281 reported serum
metabolites (Appendix II) among the entities detected in biological samples from PAD
patients in this study.
In addition to all endogenous metabolites detected in these samples, exogenous
compounds, including all 80 medications administrated to these participants, were
included in the database searches. Among these, all 23 anti-hypertensive and 4
cholesterol-lowering drugs (statins) were successfully detected in bio-fluid samples using our non-targeted LC-MS-based methods. Of the remaining 53 medications, 48 were also detected and identified (Appendix III). As a result, approximately 80% of the literature identified urinary compounds, 70% of the literature determined serum metabolites and
90% of the administrated agents were successfully identified by the method developed using LC-QTOF-MS. Perhaps the disease status of our population affected the percentage
of endogenous metabolites detected compared to the literature based on general
populations. However, it is evident that detection of all human metabolites with one
single method might be a very difficult task to achieve. The non-targeted approach
applied failed to detect 6% of the administrated agents as presented in Appendix III. The
medication profile was self-reported and we did not verify whether patients had taken all
the medications recorded in their medical history within a specified timeframe of their
study visit and providing serum and urine samples. Or, some drugs might not be detected
due to their rapid metabolism in the body and/or the extraction and ionization methods
used. Some medications were detected successfully but infrequently. For instance, 47
PAD individuals were administrated aspirin but only 4 frequencies were detected.
These results suggest that our developed methods were capable of monitoring both
- 91 -
endogenous and exogenous (administrated drugs) metabolites in urine and serum of
patients with PAD.
Statistical analyses applying one-way ANOVA (p<0.05) and Tukey HSD tests generated
several significantly changed entities in biological fluids of PAD patients. Nine and 15
selected and identified urinary and serum metabolites, respectively, were shown and
discussed in Chapter 5.2; these compounds illustrated the real effects of incorporating
beans into the diet. In urine, α-hydroxy-metoprolol was significantly attenuated after an
8-week diet intervention; paired T-tests (p<0.05) revealed a significant reduction of both metoprolol and α-hydroxy-metoprolol in the urine of the 0.6 cups beans group. This might suggest a drug-nutrient interaction with a reduction in hepatic metabolism of metoprolol. Nutrients may enhance action of drugs, thus requiring a lower dose, with potentially less adverse effects during absorption and metabolism and a careful examination of these types of interactions may be of interest to both medical groups and nutritionists. The results reported here for metoprolol will have to be confirmed in a clinical trial with appropriate experimental design.
Although there was no obvious separation among the 3 groups in PLSD graphs presented in this Thesis (Figures 2 & 4; Chapter 5.1), several classes of metabolites such as 5-L-
glutamyl-L-alanine (glutathione metabolism), nalpha-methylhistidine (histidine metabolism), 3-methylthiopropanamine (methionine metabolism), L-glutamic acid n- butyl ester (sediment formation) and creatinine (renal function) in urine as well as 25- hydroxyvitamin D2-25-glucuronide (vitamin D2 metabolism), glycocholic acid (bile acid synthesis), PGF2α diethyl amide and PGE2 p-acetamidophenyl ester (prostaglandin, hypotensive potency), lactone of PGF-MUM (PGF metabolite), ubiquinone 8
- 92 -
(methionine metabolism), N-oleoyl-L-serine (anti-osteoporotic agent) and syringomethyl
reserpate (hypotensive potency) in serum of PAD individuals were significantly affected
by the diets (Figures 6 & 7; Chapter 5.2). Therefore, it is clear that bean consumption significantly affects endogenous metabolites in urine and serum and administrated drugs in urine, and in many instances, with no obvious differences between the groups consuming 0.6 or 0.3 cups of mixed beans. These metabolites will need to be monitored in any future nutritional interventions using beans.
In addition, the metabolomics studies of the 4 types of beans used in this Thesis (pinto, navy, black and kidney) provided several compounds with significant concentration differences across bean types and versus rice. Among these, 11 were shown to be significantly changed in biological fluids of PAD patients (Table 17; Chapter 5.4). 9-
(beta-D-ribofuranosyl) zeatin, absinthin and hexadecyl acetyl glycerol were highly present in pinto and navy beans and significantly changed in urine and serum of PAD individuals. Hence, they may be considered as markers of bean consumption. Compounds such as kaempferol 3-O-glucoside were present in both beans and bio-fluids of PAD patients but with no significant changes.
It seemed that the consumption of beans might affect the concentration of their corresponding metabolites in biological samples of PAD patients; this may contribute to the changes of metoprolol metabolites in urine, but the correlations between bean compounds and bioavailability of metoprolol still need to be elucidated.
It is worth to noting that most of the 125 entities with significant difference among beans and rice were highly concentrated in pinto and navy beans. It is not clear why these two beans have high concentration of these compounds but it is possible that genetic and
- 93 -
environmental factors may be affecting the levels of some of these key compounds.
Chapter 8. Conclusions
1) The non-targeted metabolomics approach implemented in this study was capable
of monitoring endogenous metabolites and exogenously administered drugs in
urine and serum of patients with PAD.
2) Several classes of metabolites were significantly affected by the consumption of
mixed beans which might suggest changes in the metabolism of several classes of
compounds including amino acids (His, Arg, Ser, Pro, Leu, Glu), peptides (Lys
Ala His), glutathione metabolism (5-L-glutamyl-L-alanine), bile secretion
(Glycocholic acid), lipids (PE, PS, PI and LysoPE) and products of arachidonic
acid metabolism by COX. These pathways may be monitored using specific and
optimized analytical methods.
3) Several bean-related compounds were detected in biological samples and may be
used to measure compliance in bean studies. These include 9-(beta-D-
ribofuranosyl) zeatin, absinthin and hexadecyl acetyl glycerol. It is possible that
these compounds are in part responsible for the metabolic effects of the diets.
Metabolomics tools are a novel way of providing insight into biochemical
pathways affected by diet, and can help generate hypotheses for mechanism of
action.
4) Our results support the idea that potential drug-diet interactions might affect the
metabolism of administered drugs as demonstrated by the reduced urinary levels
of the anti-hypertensive medication metoprolol in PAD patients consuming beans.
- 94 -
Metabolomics is an effective method for concurrently monitoring endogenous and
exogenous metabolites.
Strengths
1) It was the first time to apply metabolomics to food items (foodomics) and
biological fluids from a nutritional intervention study. This provided a feasible
way to monitor the compliance of participants in a clinical trial and might direct a
construction of database for foodomics.
2) The HPLC-QTOF-MS method implemented in this study is capable of monitoring
anti-hypertensive agents such as metoprolol; as well, all 4 metoprolol metabolites
were successfully and frequently detected in both urine and serum. Approximately
70-80% of defined endogenous compounds, as well as bean metabolites, were
successfully detected in this study using the method developed. This method
should apply generally to non-PAD persons as well.
Limitations
1) We fully recognize that the main weakness of this study is that a single method
may never be capable of monitoring all drugs and metabolites present in a sample.
2) Two dosages of mixed beans were distributed to two different groups of
participants in PAD-Bean study, and the real dose-based effect of beans might be
affected by individual differences.
3) A mixture of 4 kinds of beans added to diet had a significant effect on
- 95 -
metabolites; however, whether a single kind of bean can influence the metabolites
was not addressed by present study.
4) As a clinical study, multiple medications as well as diets were administrated to
PAD patients at the same time, and thus drug-drug or drug-food interactions might
interfere with each other. For example, 18 out of 23 PAD patients who took
metoprolol were also administrated with statins. Although consumption of beans
significantly reduced the excretion of metoprolol, the real effect of diet on this
drug should be investigated in a separate and independent study.
5) This was an older population with a chronic disease and the results may or may
not reflect bean metabolites (and/or drug metabolites) in a younger healthier
population.
Recommendations
1) Due to multiple sources of variations in a clinical trial, a clinical study using a
cross-over design with all PAD patients receiving all 3 diets with adequate wash-
out periods is highly recommended for any further studies. A cross-over design
will provide the needed statistical power for metabolomics studies of human
samples.
2) This study used a mixture of 4 beans at two different dosages. A similar study
using individual beans will provide additional information with respect to each
bean used while the results obtained in this study reflect a combination of 4 beans.
3) Randomization of groups may need to be performed based on the types of
administered drugs to PAD patients to provide a more statistically robust design to
- 96 -
investigate the potential drug-food interactions.
4) Other methods such as Nuclear Magnetic Resonance (NMR) Spectroscopy would
provide unequivocal identification of selected key compounds in rice, beans and
biological samples.
5) Comparison between PAD patients and healthy individuals may elucidate
differences in metabolites between these two groups.
- 97 -
References
Ahimastos, A. A., Walker, P. J., Askew, C., Leicht, A., Pappas, E., Blombery, P., et al. (2013).
Effect of ramipril on walking times and quality of life among patients with peripheral artery
disease and intermittent claudication: A randomized controlled trial. JAMA, 309(5), 453-
460. doi:10.1001/jama.2012.216237
Alam, M. A., Kauter, K., & Brown, L. (2013). Naringin improves diet-induced cardiovascular
dysfunction and obesity in high carbohydrate, high fat diet-fed rats. Nutrients, 5(3), 637-
650. doi:10.3390/nu5030637
Altmaier, E., Fobo, G., Heier, M., Thorand, B., Meisinger, C., Römisch-Margl, W., et al. (2014).
Metabolomics approach reveals effects of antihypertensives and lipid- lowering drugs on the
human metabolism. Eur. J. Epidemiol., 29(5), 325-336. doi:10.1007/s10654-014-9910-7
Andersen, M., Kristensen, M., Manach, C., Pujos-Guillot, E., Poulsen, S., Larsen, T., et al.
(2014). Discovery and validation of urinary exposure markers for different plant foods by
untargeted metabolomics. Anal. Bioanal. Chem., 406(7), 1829-1844. doi:10.1007/s00216-
013-7498-5
Antonelli-Incalzi, R., Pedone, C., Mcdermott, M. M., Bandinelli, S., Miniati, B., Lova, R. M., et
al. (2006). Association between nutrient intake and peripheral artery disease: Results from
the InCHIANTI study. Atherosclerosis, 186(1), 200-206.
doi:10.1016/j.atherosclerosis.2005.07.016
Arbulu, M., Sampedro, M. C., Gómez-Caballero, A., Goicolea, M. A., & Barrio, R. J. (2015).
Untargeted metabolomic analysis using liquid chromatography quadrupole time-of- flight
98
mass spectrometry for non- volatile profiling of wines. Anal. Chim. Acta, 858, 32-41.
doi:10.1016/j.aca.2014.12.028
Armstrong, A. W., Wu, J., Mary, A. J., Grapov, D., Azizi, B., Dhillon, J., et al. (2014).
Metabolomics in psoriatic disease: Pilot study reveals metabolite differences in psoriasis
and psoriatic arthritis [version 1; referees: 2 approved]. F1000research, 3, 248-261.
doi:10.12688/f1000research.4709.1
Bab, I., Ofek, O., Tam, J., Rehnelt, J., & Zimmer, A. (2008). Endocannabinoids and the
regulation of bone metabolism. J. Neuroendocrinol., 20(s1), 69-74. doi:10.1111/j.1365-
2826.2008.01675.x
Bailey, D. G., Malcolm, J., Arnold, O., & David Spence, J. (1998). Grapefruit juice–drug
interactions. Br. J. Clin. Pharmacol., 46(2), 101-110. doi:10.1046/j.1365-
2125.1998.00764.x
Bartolucci, A. A., & Howard, G. (2006). Meta-analysis of data from the six primary prevention
trials of cardiovascular events using aspirin. Am. J. Cardiol., 98(6), 746–750.
doi:10.1016/j.amjcard.2006.04.012
Bazzano, L. A., He, J., Ogden, L. G., Loria, C., Vupputuri, S., Myers, L., et al. (2001). Legume
consumption and risk of coronary heart disease in US men and women: NHANES I
epidemiologic follow-up study. Arch. Intern. Med., 161(21):2573-2578.
doi:10.1001/archinte.161.21.2573
Beebe, H. G. (2001). Intermittent claudication: Effective medical management of a common
circulatory problem. Am. J. Cardiol., 87(12), 14-18. doi:10.1016/S0002-9149(01)01672-1
99
Beitch, B. R., & Eakins, K. E. (1969). The effects of prostaglandins on the intraocular pressure
of the rabbit. Br. J. Pharmacol., 37(1), 158-167. doi:10.1111/j.1476-5381.1969.tb09533.x
Bekkering, E. (2014). Pulses in Canada. Retrieved June 1, 2015, from
http://www.statcan.gc.ca/pub/96-325-x/2014001/article/14041-eng.htm
Berger, J. S., Ballantyne, C. M., Davidson, M. H., Johnson, J. L., Tarka, E. A., Lawrence, D., et
al. (2011). Peripheral artery disease, biomarkers, and darapladib. Am. Heart J., 161(5), 972-
978. doi:10.1016/j.ahj.2011.01.017
Bito, L. Z. (1984). Comparison of the ocular hypotensive efficacy of eicosanoids and related
compounds. Exp. Eye Res., 38(2), 181-194. doi:10.1016/0014-4835(84)90102-7
Bjerrum, J., Wang, Y., Hao, F., Coskun, M., Ludwig, C., Günther, U., et al. (2015).
Metabonomics of human fecal extracts characterize ulcerative colitis, Crohn’s disease and
healthy individuals. Metabolomics, 11(1), 122-133. doi:10.1007/s11306-014-0677-3
Boanca, G., Sand, A., Okada, T., Suzuki, H., Kumagai, H., Fukuyama, K., et al. (2007).
Autoprocessing of helicobacter pylori gamma - glutamyltranspeptidase leads to the
formation of a threonine- threonine catalytic dyad. J. Biol. Chem., 282(1), 534-541.
doi:10.1074/jbc.M607694200
Bouatra, S., Aziat, F., Mandal, R., Guo, A. C., Wilson, M. R., Knox, C., et al. (2013). The human
urine metabolome. PLoS One, 8(9), 1-29. doi:10.1371/journal.pone.0073076
Brown, L., Rosner, B., Willett, W., & Sacks, F. (1999). Cholesterol-lowering effects of dietary
fiber: A meta-analysis. Am. J. Clin. Nutr., 69(1), 30-42.
100
Bujak, R., Struck-Lewicka, W., Markuszewski, M. J., & Kaliszan, R. (2015). Metabolomics for
laboratory diagnostics. J. Pharm. Biomed. Anal., 113, 108-120.
doi:10.1016/j.jpba.2014.12.017
Caligiuri, S. P. B., Penner, B., & Pierce, G. N. (2014). The HYPERFlax trial for determining the
anti- HYPERtensive effects of dietary flaxseed in newly diagnosed stage 1 hypertensive
patients: Study protocol for a randomized, double- blinded, controlled clinical trial. Trials,
15, 232-238. doi:10.1186/1745-6215-15-232
CAPRIE Steering Committee (1996). A randomised, blinded, trial of clopidogrel versus aspirin
in patients at risk of ischaemic events (CAPRIE). Lancet, 348(9038), 1329-1339.
doi:10.1016/S0140-6736(96)09457-3
Carnevale, R., Loffredo, L., Nocella, C., Bartimoccia, S., Bucci, T., De Falco, E., et al. (2014).
Epicatechin and catechin modulate endothelial activation induced by platelets of patients
with peripheral artery disease. Oxid. Med. Cell. Longev., 2014, 1-9.
doi:10.1155/2014/691015
Celleno, L., Tolaini, M. V., D'Amore, A., Perricone, N. V., & Preuss, H. G. (2007). A dietary
supplement containing standardized phaseolus vulgaris extract influences body composition
of overweight men and women. Int. J. Med. Sci., 4(1), 45-52. doi:10.7150/ijms.4.45
Cerqueira, P. M., Boralli, V. B., Coelho, E. B., Lopes, N. P., Guimarães, L.,Fernando Lopes,
Bonato, P. S., et al. (2003). Enantioselective determination of metoprolol acidic metabolite
in plasma and urine using liquid chromatography chiral columns: Applications to
pharmacokinetics. J. Chromatogr. B, 783(2), 433-441. doi:10.1016/S1570-0232(02)00705-5
101
Chan, M. K., Tsang, T. M., Harris, L. W., Guest, P. C., Holmes, E., & Bahn, S. (2011). Evidence
for disease and antipsychotic medication effects in post-mortem brain from schizophrenia
patients. Mol. Psychiatry, 16(12), 1189-1202. doi:10.1038/mp.2010.100
Chavez-Santoscoy, R., Gutierrez-Uribe, J., Granados, O., Torre-Villalvazo, I., Serna-Saldivar, S.,
Torres, N., et al. (2014a). Flavonoids and saponins extracted from black bean (Phaseolus
vulgaris L.) seed coats modulate lipid metabolism and biliary cholesterol secretion in
C57BL/6 mice. Br. J. Nutr., 112(6), 886-899. doi:10.1017/S0007114514001536
Chávez-Santoscoy, R., Tovar, A., Serna-Saldivar, S., Torres, N., & Gutiérrez-Uribe, J. (2014b).
Conjugated and free sterols from black bean (Phaseolus vulgaris L.) seed coats as
cholesterol micelle disruptors and their effect on lipid metabolism and cholesterol transport
in rat primary hepatocytes. Genes Nutr., 9(1), 1-9. doi:10.1007/s12263-013-0367-1
Chen, J., Tsai, C., Wang, C., & Lin, H. (2013). Anti- atherosclerotic potential of gossypetin via
inhibiting LDL oxidation and foam cell formation. Toxicol. Appl. Pharmacol., 272(2), 313-
324. doi:10.1016/j.taap.2013.06.027
Clark, J. L., Zahradka, P., & Taylor, C. G. (2015). Efficacy of flavonoids in the management of
high blood pressure. Nutr. Rev., 73(12), 799-822. doi:10.1093/nutrit/nuv048
Coleman, J. J., & Butt, T. F. (2009). Antihypertensive drugs. Side Effects of Drugs Annual, 31,
349-370. doi:10.1016/S0378-6080(09)03120-1
Daniel, L. W., Small, G. W., Schmitt, J. D., Marasco, C. J., Ishaq, K., & Piantadosi, C. (1988).
Alkyl-linked diglycerides inhibit protein kinase C activation by diacylglycerols. Biochem.
Biophys. Res. Commun., 151(1), 291-297. doi:10.1016/0006-291X(88)90592-X
102
De Backer, T., Vander Stichele, R., Lehert, P., & Van Bortel, L. (2009). Naftidrofuryl for
intermittent claudication: Meta- analysis based on individual patient data. BMJ, 338, 1-8.
doi:10.1136/bmj.b603 de Bock, M., Derraik, J. G. B., & Cutfield, W. S. (2012). Polyphenols and glucose homeostasis
in humans. J. Acad. Nutr. Diet., 112(6), 808-815. doi:10.1016/j.jand.2012.01.018 de Lima, P. F., Colombo, C. A., Chiorato, A. F., Yamaguchi, L. F., Kato, M. J., & Carbonell, S.
(2014). Occurrence of isoflavonoids in brazilian common bean germplasm (Phaseolus
vulgaris L.). J. Agric. Food Chem., 62(40), 9699–9704. doi:10.1021/jf5033312
Derewacz, D. K., Goodwin, C. R., Mcnees, C. R., Mclean, J. A., & Bachmann, B. O. (2013).
Antimicrobial drug resistance affects broad changes in metabolomic phenotype in addition
to secondary metabolism. Proc. Natl. Acad. Sci. U. S. A., 110(6): 2336–2341.
doi:10.1073/pnas.1218524110
Díaz, A. M., Caldas, G. V., & Blair, M. W. (2010). Concentrations of condensed tannins and
anthocyanins in common bean seed coats. Food Res. Int., 43(2), 595-601.
doi:10.1016/j.foodres.2009.07.014
Duarte, I. F., Diaz, S., & Gil, A. M. (2014). NMR metabolomics of human blood and urine in
disease research. J. Pharm. Biomed. Anal., 93, 17-26. doi:10.1016/j.jpba.2013.09.025
Dudzik, D., Zorawski, M., Skotnicki, M., Zarzycki, W., Kozlowska, G., Bibik-Malinowska, K.,
et al. (2014). Metabolic fingerprint of gestational diabetes mellitus. J. Proteomics, 103, 57-
71. doi:10.1016/j.jprot.2014.03.025
Edel, A. L., Rodriguez-Leyva, D., Maddaford, T. G., Caligiuri, S. P., Austria, J. A., Weighell,
W., et al. (2015). Dietary flaxseed independently lowers circulating cholesterol and lowers it
103
beyond the effects of cholesterol- lowering medications alone in patients with peripheral
artery disease. J. Nutr., 145(4), 749-757. doi:10.3945/jn.114.204594
El-Bassossy, H., Abo-Warda, S., & Fahmy, A. (2013). Chrysin and luteolin attenuate diabetes-
induced impairment in endothelial-dependent relaxation: Effect on lipid profile, AGEs and
NO generation. Phytother. Res., 27(11), 1678-1684. doi:10.1002/ptr.4917
Fassbender, K., Lütjohann, D., Dik, M. G., Bremmer, M., König, J., Walter, S., et al. (2008).
Moderately elevated plant sterol levels are associated with reduced cardiovascular risk—
The LASA study. Atherosclerosis, 196(1), 283-288.
doi:10.1016/j.atherosclerosis.2006.10.032
Fiehn, O. (2001). Combining genomics, metabolome analysis, and biochemical modelling to
understand metabolic networks. Comp. Funct. Genom., 2(3), 155–168. doi:10.1002 / cfg.82
Fonseca, V., Guba, S. C., & Fink, L. M. (1999). Hyperhomocysteinemia and the endocrine
system: implications for atherosclerosis and thrombosis. Endocrine reviews, 20(5), 738-759.
Fowkes, F. G., Rudan, D., Rudan, I., Aboyans, V., Denenberg, J. O., Mcdermott, M. M., et al.
(2013). Comparison of global estimates of prevalence and risk factors for peripheral artery
disease in 2000 and 2010: A systematic review and analysis. The Lancet, 382(9901), 1329-
1340. doi: 10.1016/ S0140-6736(13)61249-0
Gakhar, N. (2013). Dietary intake of peripheral artery disease patients (Thesis). Winnipeg,
Manitoba: University of Manitoba. Retrieved May 15, 2015,
from http://hdl.handle.net/1993/19448
104
Gardner, A. W., Bright, B. C., Ort, K. A., & Montgomery, P. S. (2011). Dietary intake of
participants with peripheral artery disease and claudication. Angiology, 62(3), 270-275.
doi:10.1177/0003319710384395
Ghirlanda, G., Oradei, A., Manto, A., Lippa, S., Uccioli, L., Caputo, S., et al. (1993). Evidence
of plasma CoQ10‐ lowering effect by HMG‐ CoA reductase inhibitors: A Double‐Blind,
Placebo‐Controlled study. J. Clin. Pharmacol., 33(3), 226-229. doi:10.1002/j.1552-
4604.1993.tb03948.x
Gibbons, H., O'Gorman, A., & Brennan, L. (2015). Metabolomics as a tool in nutritional
research. Curr. Opin. Lipidol., 26(1), 30-34. doi:10.1097/MOL.0000000000000140
Godse, S., Mohan, M., Kasture, V., & Kasture, S. (2010). Effect of myricetin on blood pressure
and metabolic alterations in fructose hypertensive rats. Pharm. Biol., 48(5), 494-498.
doi:10.3109/13880200903188526
González-Domínguez, R., García-Barrera, T., Vitorica, J., & Gómez-Ariza, J. L. (2015).
Deciphering metabolic abnormalities associated with alzheimer's disease in the APP/ PS1
mouse model using integrated metabolomic approaches. Biochimie, 110, 119-128.
doi:10.1016/j.biochi.2015.01.005
Guan, Q., Liang, S., Wang, Z., Yang, Y., & Wang, S. M. (2014). H-1 NMR-based metabonomic
analysis of the effect of optimized rhubarb aglycone on the plasma and urine metabolic
fingerprints of focal cerebral ischemia-reperfusion rats. J. Ethnopharmacol., 154(1), 65-75.
doi:10.1016/j.jep.2014.03.002
105
Guo, N., Xu, Y., & Cao, Z. (2015). Absinthin attenuates LPS- induced ALI through MIP-1
alpha- mediated inflammatory cell infiltration. Exp.Lung Res., 41(9), 514-524.
doi:10.3109/01902148.2015.1093566
Hanson, M. G., Zahradka, P., & Taylor, C. G. (2014). Lentil-based diets attenuate hypertension
and large-artery remodelling in spontaneously hypertensive rats. Br. J. Nutr., 111(4), 690-
698. doi:10.1017/S0007114513002997
Health Canada. (2015). Examples of one food guide serving. Retrieved June 15, 2015,
from http://www.hc-sc.gc.ca/fn-an/food-guide-aliment/basics-base/serving-portion-eng.php
Heffron, S. P., Rockman, C. B., Gianos, E., Guo, Y., & Berger, J. S. (2015). Greater frequency
of nut consumption is associated with lower prevalence of peripheral arterial disease. Prev.
Med., 72, 15-18. doi:10.1016/j.ypmed.2014.12.014
Hiatt, W. R. (2001). Medical treatment of peripheral arterial disease and claudication. N. Engl. J.
Med., 344(21), 1608-1621. doi:10.1056/NEJM200105243442108
Hiatt, W. R., Regensteiner, J. G., Creager, M. A., Hirsch, A. T., Cooke, J. P., Olin, J. W., et al.
(2001). Propionyl- L- carnitine improves exercise performance and functional status in
patients with claudication. Am. J. Med., 110(8), 616-622. doi:10.1016/S0002-
9343(01)00704-5
Hirsch, A., Criqui, M., Treat-Jacobson, D., & Regensteiner, J. (2001). Peripheral arterial disease
detection, awareness, and treatment in primary care. JAMA, 286(11), 1317-1324.
doi:10.1001/jama.286.11.1317
Hood, S., Moher, D., & Barber, G. (1996). Management of intermittent claudication with
pentoxifylline: Meta- analysis of randomized controlled trials. CMAJ, 155(8), 1053-1059.
106
Hu, Y., Cheng, Z., Heller, L. I., Krasnoff, S. B., Glahn, R. P., & Welch, R. M. (2006).
Kaempferol in red and pinto bean seed (Phaseolus vulgaris L.) coats inhibits iron
bioavailability using an in vitro digestion/human Caco-2 cell model. J. Agric. Food Chem.,
54(24), 9254-9261. doi:10.1021/jf0612981
Huang, C., Mcdermott, M. M., Liu, K., Kuo, C., Wang, S., Tao, H., et al. (2013). Plasma
metabolomic profiles predict near-term death among individuals with lower extremity
peripheral arterial disease. J. Vasc. Surg., 58(4), 989-996. doi:10.1016/j.jvs.2013.04.022
Hutchins, A. M., Winham, D. M., & Thompson, S. V. (2012). Phaseolus beans: Impact on
glycaemic response and chronic disease risk in human subjects. Br. J. Nutr., 108, S52-S65.
doi:10.1017/S0007114512000761
Ix, J. H., Biggs, M. L., Kizer, J. R., Mukamal, K. J., Djousse, L., Zieman, S. J., et al. (2011).
Association of body mass index with peripheral arterial disease in older adults: The
cardiovascular health study. Am. J. Epidemiol., 174(9), 1036-1043. doi:10.1093/aje/kwr228
Janeckova, H., Kalivodova, A., Najdekr, L., Friedecky, D., Hron, K., Bruheim, P., et al. (2015).
Untargeted metabolomic analysis of urine samples in the diagnosis of some inherited
metabolic disorders. Biomed. Pap. Med. Fac. Univ. Palacky Olomouc Czech. Repub.,
159(4), 582-585. doi:10.5507/bp.2014.048
Jáuregui-Garrido, B., & Jáuregui-Lobera, I. (2012). Interactions between antihypertensive drugs
and food. Nutr. Hosp., 27(5), 1866-1875. doi:10.3305/nh.2012.27.6.6127
Jenkins, D. J. A., Kendall, C. W. C., Augustin, L. S. A., Sandra Mitchell, , Sahye-Pudaruth, S.,
Mejia, S. B., et al. (2012). Effect of legumes as part of a low glycemic index diet on
glycemic control and cardiovascular risk factors in type 2 diabetes mellitus: A randomized
107
controlled trial. Arch. Intern. Med., 172(21), 1653-1660.
doi:10.1001/2013.jamainternmed.70
Kadouh, H., & Zhou, K. (2012). Chapter 13. Nutraceutical and health properties of common
beans (Phaseolus vulgaris). In L. L. Yu, R. Tsao & F. Shahidi (Eds.), Cereals and pulses
nutraceutical properties and health benefits (1st ed., pp. 187-196). Oxford, UK: Wiley-
Blackwell.
Kaptoge, S., Seshasai, S. R. K., Gao, P., Freitag, D. F., Butterworth, A. S., Borglykke, A., et al.
(2014). Inflammatory cytokines and risk of coronary heart disease: New prospective study
and updated meta-analysis. Eur. Heart J., 35(9), 578-589. doi:10.1093/eurheartj/eht367
Karim, A., Rozek, L. F., Smith, M. E., & Kowalski, K. G. (1989). Effects of food and antacid on
oral absorption of misoprostol, a synthetic prostaglandin E1 analog. J. Clin. Pharmacol.,
29(5), 439-443. doi:10.1002/j.1552-4604.1989.tb03358.x
Kim, J., Ryu, S., Kim, S., Lee, H., Lim, M., Seong, S., et al. (2013). Pattern recognition analysis
for hepatotoxicity induced by acetaminophen using plasma and urinary 1H NMR- based
metabolomics in humans. Anal. Chem., 85(23), 11326-11334. doi:10.1021/ac402390q
Kim, S., Choi, Y., & Lee, M. (1993). Pharmacokinetics and pharmacodynamics of furosemide in
protein-calorie malnutrition. J. Pharmacokinet. and Biopharm., 21(1), 1-17.
doi:10.1007/BF01061772
Kitamura, S., Ishihara, Y., & Kosaka, K. (1977). Radioimmunoassay of main urinary metabolite
of prostaglandin F2α in normal subjects. Prostaglandins, 14(5), 961-965. doi:10.1016/0090-
6980(77)90310-0
108
Kleemann, R., Verschuren, L., Morrison, M., Zadelaar, S., van Erk, M. J., Wielinga, P. Y., et al.
(2011). Anti-inflammatory, anti-proliferative and anti-atherosclerotic effects of quercetin in
human in vitro and in vivo models. Atherosclerosis, 218(1), 44-52.
doi:10.1016/j.atherosclerosis.2011.04.023
Kong, L., Luo, C., Li, X., Zhou, Y., & He, H. (2013). The anti-inflammatory effect of
kaempferol on early atherosclerosis in high cholesterol fed rabbits. Lipids Health Dis., 12,
115-126. doi:10.1186/1476-511X-12-115
Kowalski, J., Samojedny, A., Paul, M., Pietsz, G., & Wilczok, T. (2005). Effect of kaempferol on
the production and gene expression of monocyte chemoattractant protein-1 in J774.2
macrophages. Pharmacol. Rep., 57(1), 107-112.
Kusano, M., Yang, Z., Okazaki, Y., Nakabayashi, R., Fukushima, A., & Saito, K. (2015). Using
metabolomic approaches to explore chemical diversity in rice. Mol. Plant, 8(1), 58-67.
doi:10.1016/j.molp.2014.11.010
Lane, D. A., & Lip, G. Y. H. (2013). Treatment of hypertension in peripheral arterial disease.
Cochrane Database Syst. Rev., 12, 1-55. doi:10.1002/14651858.CD003075.pub3
Lapenna, D., de Gioia, S., Ciofani, G., Mezzetti, A., Ucchino, S., Calafiore, A. M., et al. (1998).
Glutathione-related antioxidant defenses in human atherosclerotic plaques. Circulation,
97(19), 1930-1934. doi: 10.1161/01.CIR.97.19.1930
Lau, J. F., Weinberg, M. D., & Olin, J. W. (2011). Peripheral artery disease. Part 1: Clinical
evaluation and noninvasive diagnosis. Nat. Rev. Cardiol., 8(7), 405-418.
doi:10.1038/nrcardio.2011.66
109
Leyva, D. R., Zahradka, P., Ramjiawan, B., Guzman, R., Aliani, M., & Pierce, G. N. (2011). The
effect of dietary flaxseed on improving symptoms of cardiovascular disease in patients with
peripheral artery disease: Rationale and design of the FLAX-PAD randomized controlled
trial. Contemp. Clin. Trials, 32(5), 724-730. doi:10.1016/j.cct.2011.05.005
Lian, T., Wang, L., Lo, Y., Huang, I., & Wu, M. (2008). Fisetin, morin and myricetin attenuate
CD36 expression and oxLDL uptake in U937-derived macrophages. Biochim. Biophys.
Acta, 1781(10), 601-609. doi:10.1016/j.bbalip.2008.06.009
Lin, L., Harnly, J. M., Pastor-Corrales, M., & Luthria, D. L. (2008). The polyphenolic profiles of
common bean (Phaseolus vulgaris L.). Food Chem., 107(1), 399-410.
doi:10.1016/j.foodchem.2007.08.038
Maekawa, M., Shimada, M., Ohno, K., Togawa, M., Nittono, H., Iida, T., et al. (2015). Focused
metabolomics using liquid chromatography/electrospray ionization tandem mass
spectrometry for analysis of urinary conjugated cholesterol metabolites from patients with
Niemann-Pick disease type C and 3-hydroxysteroid dehydrogenase deficiency. Ann. Clin.
Biochem., 52(5), 576-587. doi:10.1177/0004563214568871
Mannina, L., Sobolev, A. P., & Capitani, D. (2012). Applications of NMR metabolomics to the
study of foodstuffs: Truffle, kiwifruit, lettuce, and sea bass. Electrophoresis, 33(15), 2290-
2313. doi:10.1002/elps.201100668
Matsuyama, T., Tanaka, Y., Kamimaki, I., Nagao, T., & Tokimitsu, I. (2008). Catechin safely
improved higher levels of fatness, blood pressure, and cholesterol in children. Obesity
(Silver Spring), 16(6), 1338-1348. doi:10.1038/oby.2008.60
110
Mayo Clinic. (2015). Peripheral artery disease (PAD): Symptoms and causes. Retrieved June 12,
2015, from http://www.mayoclinic.org/diseases-conditions/peripheral-artery-
disease/symptoms-causes/dxc-20167421
Mccrindle, J. L., Li, K. W., Barron, W., & Prescott, L. F. (1996). Effect of food on the
absorption of frusemide and bumetanide in man. Br. J. Clin. Pharmacol., 42(6), 743-746.
doi:10.1046/j.1365-2125.1996.00494.x
McDermott, M. M., Liu, K., Guralnik, J. M., Criqui, M. H., Spring, B., Tian, L., et al. (2013).
Home-based walking exercise intervention in peripheral artery disease: A randomized
clinical trial. JAMA, 310(1), 57-65. doi:10.1001/jama.2013.7231
McNally, S., & O’Brien, C. J. (2014). Metabolomics/proteomics strategies used to identify
biomarkers for exfoliation glaucoma. J. Glaucoma, 23(8), S51-S54.
doi:10.1097/IJG.0000000000000117
McNamara, M. J., Schmitt, J. D., Wykle, R. L., & Daniel, L. W. (1984). 1-0-hexadecyl-2-acetyl-
sn-glycerol stimulates differentiation of HL-60 human promyelocytic leukemia cells to
macrophage-like cells. Biochem. Biophys. Res. Commun., 122(2), 824-830.
doi:10.1016/S0006-291X(84)80108-4
Melander, A., Danielson, K., Hanson, A., Rudell, B., Scherstén, B., Thulin, T., et al. (1977).
Enhancement of hydralazine bioavailability by food. Clin. Pharmacol. Ther., 22(1), 104-
107.
Mensack, M. M., Fitzgerald, V. K., Ryan, E. P., Lewis, M. R., Thompson, H. J., & Brick, M. A.
(2010). Evaluation of diversity among common beans (Phaseolus vulgaris L.) from two
111
centers of domestication using 'omics' technologies. BMC Genomics, 11, 686-696.
doi:10.1186/1471-2164-11-686
Mensack, M. M., McGinley, J. N., & Thompson, H. J. (2012). Metabolomic analysis of the
effects of edible dry beans (Phaseolus vulgaris L.) on tissue lipid metabolism and
carcinogenesis in rats. Br. J. Nutr., 108, S155-S165. doi:10.1017/S0007114512000827
Merchant, A. T., Hu, F. B., Spiegelman, D., Willett, W. C., Rimm, E. B., & Ascherio, A. (2003).
Dietary fiber reduces peripheral arterial disease risk in men. J. Nutr., 133(11), 3658-3663.
Messina, V. (2014). Nutritional and health benefits of dried beans. Am. J. Clin. Nutr., 100, 437S-
442S. doi:10.3945/ajcn.113.071472
Mishra, J., Aradhana, Rattan, D. S., Gausain, M. P., & Vikas, S. J. (2012). HPTLC profile of
quercetin content of common bean (Uttarakhand) landraces growing in Uttarakhand. Am. J.
Food Technol., 7(2), 96-100. doi:10.3923/ajft.2012.96.100
Mlekusch, W., Exner, M., Sabeti, S., Amighi, J., Schlager, O., Wagner, O., et al. (2004). Serum
creatinine predicts mortality in patients with peripheral artery disease: Influence of diabetes
and hypertension. Atherosclerosis, 175(2), 361-367.
doi:10.1016/j.atherosclerosis.2004.04.008
Mohler, E. R., Hiatt, W. R., & Creager, M. A. (2003). Cholesterol reduction with atorvastatin
improves walking distance in patients with peripheral arterial disease. Circulation, 108(12),
1481-1486. doi:10.1161/01.CIR.0000090686.57897.F5
Moncaleán, P., López-Iglesias, C., Fernández, B., & Rodríguez, A. (2001). Immunocytochemical
location of endogenous cytokinins in buds of kiwifruit (Actinidia deliciosa) during the first
hours of in vitro culture. Histochem. J., 33(7), 403-411.
112
Morand, C., Dubray, C., Milenkovic, D., Lioger, D., Martin, J. F., Scalbert, A., et al. (2011).
Hesperidin contributes to the vascular protective effects of orange juice: A randomized
crossover study in healthy volunteers. Am. J. Clin. Nutr., 93(1), 73-80.
doi:10.3945/ajcn.110.004945
Mudryj, A. N., Yu, N., Hartman, T. J., Mitchell, D. C., Lawrence, F. R., & Aukema, H. M.
(2012). Pulse consumption in Canadian adults influences nutrient intakes. Br. J. Nutr., 108,
S27-S36. doi:10.1017/S0007114512000724
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.
doi:10.1080/004982599238047
Ohkawara, S., Okuma, Y., Uehara, T., Yamagishi, T., & Nomura, Y. (2005). Astrapterocarpan
isolated from Astragalus membranaceus inhibits proliferation of vascular smooth muscle
cells. Eur. J. Pharmacol, 525(1), 41-47. doi:10.1016/j.ejphar.2005.08.063
Okada, T., Suzuki, H., Wada, K., Kumagai, H., & Fukuyama, K. (2006). Crystal structures of
gamma-glutamyltranspeptidase from Escherichia coli, a key enzyme in glutathione
metabolism, and its reaction intermediate. Proc. Natl. Acad. Sci. U. S. A., 103(17), 6471-
6476. doi:10.1073/pnas.0511020103
Olaleye, M. T., Crown, O. O., Akinmoladun, A. C., & Akindahunsi, A. A. (2014). Rutin and
quercetin show greater efficacy than nifedipin in ameliorating hemodynamic, redox, and
metabolite imbalances in sodium chloride-induced hypertensive rats. Hum. Exp. Toxicol.,
33(6), 602-608. doi:10.1177/0960327113504790
113
Petry, N., Egli, I., Campion, B., Nielsen, E., & Hurrell, R. (2013). Genetic reduction of phytate
in common bean (Phaseolus vulgaris L.) seeds increases iron absorption in young women. J.
Nutr., 143(8), 1219-1224. doi:10.3945/jn.113.175067
Potenza, M. A., Marasciulo, F. L., Tarquinio, M., Tiravanti, E., Colantuono, G., Federici, A., et
al. (2007). EGCG, a green tea polyphenol, improves endothelial function and insulin
sensitivity, reduces blood pressure, and protects against myocardial I/R injury in SHR. Am.
J. Physiol. Endocrinol. Metab., 292(5), E1378-E1387. doi:10.1152/ajpendo.00698.2006
Potischman, N., Swanson, C. A., Coates, R. J., Gammon, M. D., Brogan, D. R., Curtin, J., et al.
(1999). Intake of food groups and associated micronutrients in relation to risk of early-stage
breast cancer. Int. J. Cancer, 82(3), 315-321.
Prakash, A., & Markham, A. (2000). Metoprolol: A review of its use in chronic heart failure.
Drugs, 60(3), 647-678.
Pronsky, Z., M., Crowe, Sr. J. P. (2012). Chapter 9. Clinical: Food-drug interactions. In L. K.
Mahan, S. Escott-Stump, & J. L. Raymond (Eds.), Krause's food and the nutrition care
process (13th ed., pp. 209-228). St. Louis, MO: Elsevier/Saunders.
Psychogios, N., Hau, D. D., Peng, J., Guo, A. C., Mandal, R., Bouatra, S., et al. (2011). The
human serum metabolome. PLoS One, 6(2), 1-23. doi:10.1371/journal.pone.0016957
Rattan, S. I. S., & Sodagam, L. (2005). Gerontomodulatory and youth-preserving effects of
zeatin on human skin fibroblasts undergoing aging in vitro. Rejuvenation Res., 8(1), 46-57.
doi:10.1089/rej.2005.8.46
114
Raymond, J. L., & Couch, S. C. (2012). Chapter 34. medical nutrition therapy for cardiovascular
disease. In L. K. Mahan, S. Escott-Stump, & J. L. Raymond (Eds.), Krause's food and the
nutrition care process (13th ed., pp. 742-781). St. Louis, MO: Elsevier/Saunders.
Reis, J. P., Michos, E. D., von Muhlen, D., & Miller, E. R. 3rd. (2008). Differences in vitamin D
status as a possible contributor to the racial disparity in peripheral arterial disease. Am. J.
Clin. Nutr., 88(6), 1469-1477. doi:10.3945/ajcn.2008.26447
Rizkalla, S. W., Bellisle, F., & Slama, G. (2002). Health benefits of low glycaemic index foods,
such as pulses, in diabetic patients and healthy individuals. Br. J. Nutr., 88, S255-S262.
Robaczewska, J., Kedziora-Kornatowska, K., Kozakiewicz, M., Zary-Sikorska, E., Pawluk, H.,
Pawliszak, W., et al. (2016). Role of glutathione metabolism and glutathione-related
antioxidant defense systems in hypertension. J Physiol Pharmacol, 67(3), 331-337.
Rodriguez-Leyva, D., Weighell, W., Edel, A. L., Lavallee, R., Dibrov, E., Pinneker, R., et al.
(2013). Potent antihypertensive action of dietary flaxseed in hypertensive patients.
Hypertension, 62(6), 1081-1089. doi:10.1161/HYPERTENSIONAHA.113.02094
Salvamani, S., Gunasekaran, B., Shaharuddin, N. A., Ahmad, S. A., & Shukor, M. Y. (2014).
Antiartherosclerotic effects of plant flavonoids. Biomed. Res. Int., 2014, 1-11.
doi:10.1155/2014/480258
Sampson, U., Fowkes, F., Mcdermott, M., Criqui, M., Aboyans, V., Norman, P., et al. (2014).
Global and regional burden of death and disability from peripheral artery disease: 21 world
regions, 1990 to 2010. Glob. Heart., 9(1), 145-158. doi:10.1016/j.gheart.2013.12.008
Schmidt, L. E., & Dalhoff, K. (2002). Food-drug interactions. Drugs, 62(10), 1481-1502.
doi:10.2165/00003495-200262100-00005
115
Scirica, B. M., & Morrow, D. A. (2006). Is C-reactive protein an innocent bystander or
proatherogenic culprit? the verdict is still out. Circulation, 113(17), 2128-2134.
doi:10.1161/CIRCULATIONAHA.105.611350
Semple, H. A., & Xia, F. (1995). Interaction between propranolol and amino acids in the single-
pass isolated, perfused rat liver. Drug Metab. Dispos, 23(8), 794-798.
Serrano, J., & Goñi, I. (2004). Role of black bean phaseolus vulgaris on the nutritional status of
guatemalan population. Arch. Latinoam. Nutr., 54(1), 36-44.
Sessa D.J., Gardner H.W., Kleiman R., & Weisleder D. (1977). Oxygenated fatty acid
constituents of soybean phosphatidylcholines. Lipids, 12(7), 613-619.
doi:10.1007/BF02533391
Setoyama, D., Iwasa, K., Seta, H., Shimizu, H., Fujimura, Y., Miura, D., et al. (2013). High-
throughput metabolic profiling of diverse green coffea arabica beans identified tryptophan
as a universal discrimination factor for immature beans. PLoS One, 8(8), 1-7.
doi:10.1371/journal.pone.0070098
Shepherd, A. M. M., Irvine, N. A., & Ludden, T. M. (1984). Effect of food on blood hydralazine
levels and response in hypertension. Clin. Pharmacol. Ther., 36(1), 14-18.
doi:10.1038/clpt.1984.131
Shimada, K., Mitamura, K., & Nakatani, I. (1997). Characterization of monoglucuronides of
vitamin D-2 and 25-hydroxyvitamin D-2 in rat bile using high-performance liquid
chromatography-atmospheric pressure chemical ionization mass spectrometry. J.
Chromatogr. B, 690(1-2), 348-354.
116
Silva, F. M., Kramer, C. K., Almeida, J. C., Steemburgo, T., Gross, J. L., & Azevedo, M. J.
(2013). Fiber intake and glycemic control in patients with type 2 diabetes mellitus: A
systematic review with meta‐analysis of randomized controlled trials. Nutr. Rev., 71(12),
790-801. doi:10.1111/nure.12076
Singhvi, S. M., Mckinstry, D. N., Shaw, J. M., Willard, D. A., & Migdalof, B. H. (1982). Effect
of food on the bioavailability of captopril in healthy subjects. J. Clin. Pharmacol., 22(2),
135-140. doi:10.1002/j.1552-4604.1982.tb02661.x
Smoum, R., Bar, A., Tan, B., Milman, G., Attar-Namdar, M., Ofek, O., et al. (2010). Oleoyl
serine, an endogenous N-acyl amide, modulates bone remodeling and mass. Proc. Natl.
Acad. Sci. U. S. A., 107(41), 17710-17715. doi:10.1073/pnas.0912479107
Song, F., Zhu, Y., Shi, Z., Tian, J., Deng, X., Ren, J., et al. (2014). Plant food anthocyanins
inhibit platelet granule secretion in hypercholesterolaemia: Involving the signalling pathway
of PI3K-Akt. Thromb. Haemost., 112(5), 981-991. doi:10.1160/TH13-12-1002
Statistics Canada (2015). Deaths, by cause, chapter IX: Diseases of the circulatory system (I00
to I99), age group and sex, Canada annual (number). Retrieved April 23, 2015,
from http://www5.statcan.gc.ca/cansim/pick-choisir?lang=eng&p2=33&id=1020529
Strocchi, A., & Lercker, G. (1979). Cardanol in germ and seed oils extracted from cashew nuts
obtained by the oltremare process. J. Am. Oil Chem. Soc., 56(6), 616-619.
doi:10.1007/BF02679334
Suzuki Y. (1981). Metabolic and pharmacological features of syrosingopine reserpine and their
active metabolite syringo methyl reserpate (abstract). J. Med. Soc. Toho Univer., 260-274.
117
Retrieved May 12, 2016, from http://eurekamag.com/research/005/880/005880109.php -
close
Syed, I. A., & Khan, W. A. (2011). Glycated haemoglobin - a marker and predictor of
cardiovascular disease. J. Pak. Med. Assoc., 61(7), 690-695.
Tagliazucchi, D., Martini, S., Bellesia, A., & Conte, A. (2015). Identification of ACE-inhibitory
peptides from phaseolus vulgaris after in vitro gastrointestinal digestion. Int. J. Food Sci.
Nutr., 66(7), 774-782. doi:10.3109/09637486.2015.1088940
Teede, H. J., Mcgrath, B. P., Desilva, L., Cehun, M., Fassoulakis, A., & Nestel, P. J. (2003).
Isoflavones reduce arterial stiffness: A placebo-controlled study in men and postmenopausal
women. Arterioscler. Thromb, Vasc. Biol., 23(6), 1066-1071.
doi:10.1161/01.ATV.0000072967.97296.4A
Thompson, P. D., Zimet, R., Forbes, W. P., & Zhang, P. (2002). Meta- analysis of results from
eight randomized, placebo- controlled trials on the effect of cilostazol on patients with
intermittent claudication. Am. J. Cardiol., 90(12), 1314-1319. doi:10.1016/S0002-
9149(02)02869-2
Tomita, M., Motomura, Y., Kitahara, H., Yoshiki, Y., Okubo, K. (1999). Isomers and
homologues of L-glutamic acid 5-n-butyl ester as promoters of sediment formation in raw
soy sauce. J. Biosci. Bioeng., 87(3), 357-360. doi:10.1016/S1389-1723(99)80045-0
Towfighi, A., Markovic, D., & Ovbiagele, B. (2010). Pronounced association of elevated serum
homocysteine with stroke in subgroups of individuals: A nationwide study. J. Neurol. Sci.,
298(1-2), 153-157. doi:10.1016/j.jns.2010.07.013
118
Tsao, R., Yu, L. L., & Shahidi, F. (2012). Cereals and pulses – an overview. In L. L. Yu, R.
Tsao, F. Shahidi (Eds.), Cereals and pulses: Nutraceutical properties and health benefits
(first ed., pp.1-5). Oxford, UK: Wiley‐Blackwell.
Turer, A. T., Stevens, R. D., Bain, J. R., Muehlbauer, M. J., van, d. W., Mathew, J. P., et al.
(2009). Metabolomic profiling reveals distinct patterns of myocardial substrate use in
humans with coronary artery disease or left ventricular dysfunction during surgical
ischemia/reperfusion. Circulation, 119(13), 1736-1746.
doi:10.1161/CIRCULATIONAHA.108.816116
Um, S. Y., Park, J. H., Chung, M. W., Kim, K., Kim, S. H., Choi, K. H., et al. (2012). Nuclear
magnetic resonance-based metabolomics for prediction of gastric damage induced by
indomethacin in rats. Anal. Chim. Acta, 722, 87-94. doi:10.1016/j.aca.2012.01.062
Um, S., Chung, M., Kim, K., Kim, S., Oh, J., Oh, H., et al. (2009). Pattern recognition analysis
for the prediction of adverse effects by nonsteroidal anti-inflammatory drugs using 1H
NMR-based metabolomics in rats. Anal. Chem., 81(12), 4734-4741. doi:10.1021/ac9000282
van Deventer, C. A., Lindeque, J. Z., van Rensburg, Peet J. Jansen, Malan, L., van, D. W., &
Louw, R. (2015). Use of metabolomics to elucidate the metabolic perturbation
associated with hypertension in a black south african male cohort: The SABPA study. J.
Am. Soc. Hypertens., 9(2), 104-114. doi:10.1016/j.jash.2014.11.007
Venkateswaran, S., & Pari, L. (2002). Antioxidant effect of phaseolus vulgaris in streptozotocin‐
induced diabetic rats. Asia Pac. J. Clin. Nutr., 11(3), 206-209.
Waitzman, M. B., & King, C. D. (1967). Prostaglandin influences on intraocular pressure and
pupil size. Am. J. Physiol., 212(2), 329-334.
119
Wang, B. O., & Semple, H. A. (1997). Inhibition of metoprolol metabolism by amino acids in
perfused rat livers. Insights into the food effect? Drug Metabo. Dispos., 25(3), 287-295.
Wei, F., Furihata, K., Koda, M., Hu, F., Kato, R., Miyakawa, T., et al. (2012). (13)C NMR-based
metabolomics for the classification of green coffee beans according to variety and origin. J.
Agric. Food Chem., 60(40), 10118-10125. doi:10.1021/jf3033057
Wei, F., Furihata, K., Miyakawa, T., & Tanokura, M. (2014). A pilot study of NMR- based
sensory prediction of roasted coffee bean extracts. Food Chem., 152, 363-369.
doi:10.1016/j.foodchem.2013.11.161
Weinberg, M. D., Lau, J. F., Rosenfield, K., & Olin, J. W. (2011). Peripheral artery disease. Part
2: Medical and endovascular treatment. Nat. Rev. Cardiol., 8(8), 429-441.
doi:10.1038/nrcardio.2011.81
Whelton, S. P., Hyre, A. D., Pedersen, B., Yi, Y., Whelton, P. K., & He, J. (2005). Effect of
dietary fiber intake on blood pressure: A meta- analysis of randomized, controlled clinical
trials. J. Hypertens., 23(3), 475-481.
Willenberg, I., Ostermann, A. I., & Schebb, N. H. (2015). Targeted metabolomics of the
arachidonic acid cascade: Current state and challenges of LC–MS analysis of oxylipins.
Anal. and Bioanal. Chem., 407(10), 2675-2683. doi:10.1007/s00216-014-8369-4
Winer, B. M., & Sahay, B. M. (1960). Comparative observations of reserpine and syrosingopine.
N. Engl. J. Med., 263, 1165-1169. doi:10.1056/NEJM196012082632303
Winham, D. M., Hutchins, A. M., & Johnston, C. S. (2007). Pinto bean consumption reduces
biomarkers for heart disease risk. J. Am. Coll. Nutr., 26(3), 243-249.
doi:10.1080/07315724.2007.10719607
120
Wong, J. H., Wan, C. T., & Ng, T. B. (2010). Characterisation of a haemagglutinin from
Hokkaido red bean (Phaseolus vulgaris cv. Hokkaido red bean). J. Sci. Food Agric., 90(1),
70-77. doi:10.1002/jsfa.3782
Woodward, D. F., Krauss, A. H.-P., Chen, J., Lai, R. K., Spada, C. S., Burk, R. M., et al. (2001).
The pharmacology of bimatoprost (Lumigan™). Surv. Ophthalmol., 45(4), S337-S345.
doi:10.1016/S0039-6257(01)00224-7
Yamamoto, M., Suzuki, A., & Hase, T. (2008). Short-term effects of glucosyl hesperidin and
hesperetin on blood pressure and vascular endothelial function in spontaneously
hypertensive rats. J. Nutr. Sci. Vitaminol. (Tokyo), 54(1), 95-98. doi:10.3177/jnsv.54.95
You, Y., Kim, Y., Nam, N., Bang, S., & Ahn, B. (2004). Alkyl and carboxylalkyl esters of 4′-
demethyl-4-deoxypodophyllotoxin: Synthesis, cytotoxic, and antitumor activity. Eur. J.
Med. Chem., 39(2), 189-193. doi:10.1016/j.ejmech.2003.10.002
Young, M. A., Lettis, S., & Eastmond, R. (1998). Improvement in the gastrointestinal absorption
of troglitazone when taken with, or shortly after, food. Br. J. Clin. Pharmacol., 45(1), 31-35.
doi:10.1046/j.1365-2125.1998.00653.x
Zagura, M., Kals, J., Paapstel, K., Kilk, K., Serg, M., Kampus, P., et al. (2014). Metabolomic
signature of arterial stiffness in patients with peripheral arterial disease. Hypertens. Res.,
38(12), 840-846. doi:10.1038/hr.2015.71
Zahedi, M., Ghiasvand, R., Feizi, A., Asgari, G., & Darvish, L. (2013). Does quercetin improve
cardiovascular risk factors and inflammatory biomarkers in women with type 2 diabetes: A
double-blind randomized controlled clinical trial. Int. J. Prev. Med., 4(7), 777-785.
121
Zahradka, P., Wright, B., Weighell, W., Blewett, H., Baldwin, A., O, K., et al. (2013). Daily non-
soy legume consumption reverses vascular impairment due to peripheral artery disease.
Atherosclerosis, 230(2), 310-314. doi:10.1016/j.atherosclerosis.2013.07.048
122
APPENDICES
123
Appendix I. Identified metabolites (273 out of 345) detected in urine of PAD patients compared to the Human Urine Metabolome.1 Baseline (n=54) Week 8 (n=54) Compounds HMDB ID Formula (occurrence, %) (occurrence, %) 1,3-Dimethyluric acid HMDB01857 C7H8N4O3 93 83 1-Methyladenosine HMDB03331 C11 H15N5O4 96 98 1-Methylhistidine HMDB00001 C7H11 N3O2 96 98 1-Methylnicotinamide HMDB00699 C7H9N2O 98 96 2-Hydroxy-3-methylpentanoic acid HMDB00317 C6H12O3 100 91 2-Furoylglycine HMDB00439 C7H7NO4 98 100 2-Hydroxybutyric acid HMDB00008 C4H8O3 56 81 2-Hydroxyisobutyric acid HMDB00729 C4H8O3 56 81 L-2-Hydroxyglutaric acid HMDB00694 C5H8O5 80 89 2-Ketobutyric acid HMDB00005 C4H6O3 83 52 2-Methyl-3-hydroxybutyric acid HMDB00354 C5H10O3 100 100 2-Methyl-3-ketovaleric acid HMDB00408 C6H10O3 96 96 2-Methylbutyrylglycine HMDB00339 C7H13NO3 96 98 2-Methylerythritol HMDB11659 C5H12O4 20 11 2-Methylglutaric acid HMDB00422 C6H10O4 41 46 3-Aminoisobutanoic acid HMDB03911 C4H9NO2 9 35 3-Hydroxybutyric acid HMDB00357 C4H8O3 56 81 3-Hydroxyhippuric acid HMDB06116 C9H9NO4 31 35 3-Hydroxyisovaleric acid HMDB00754 C5H10O3 100 100 3-Hydroxymethylglutaric acid HMDB00355 C6H10O5 31 67 3-Hydroxyphenylacetic acid HMDB00440 C8H8O3 48 52 3-Methyl-2-oxovaleric acid HMDB00491 C6H10O3 96 96 3-Methyladipic acid HMDB00555 C7H12O4 41 50 3-Methylglutaconic acid HMDB00522 C6H8O4 50 59 3-Methylhistidine HMDB00479 C7H11 N3O2 96 98 3-Methylxanthine HMDB01886 C6H6N4O2 98 96 γ-Aminobutyric acid HMDB00112 C4H9NO2 2 2 4-Aminohippuric acid HMDB01867 C9H10N2O3 17 15 4-Ethylphenol HMDB29306 C8H10O 33 26 4-Hydroxybenzoic acid HMDB00500 C7H6O3 98 100 4-Hydroxybutyric acid HMDB00710 C4H8O3 56 81 4-Hydroxyhippuric acid HMDB13678 C9H9NO4 31 35 Hydroxyphenyllactic acid HMDB00755 C9H10O4 81 89 4-Pyridoxic acid HMDB00017 C8H9NO4 81 76 Dihydrouracil HMDB00076 C4H6N2O2 100 100 5-Aminolevulinic acid HMDB01149 C5H9NO3 54 37 5-Aminopentanoic acid HMDB03355 C5H11 NO2 100 100 7-Methylxanthine HMDB01991 C6H6N4O2 98 96 Acetaminophen HMDB01859 C8H9NO2 87 89 Acetaminophen glucuronide HMDB10316 C14H17NO8 54 50 Paracetamol sulfate HMDB59911 C8H9NO5S 31 24 Acetoacetic acid HMDB00060 C4H6O3 83 52 Acetone HMDB01659 C3H6O ND 2 L-Acetylcarnitine HMDB00201 C9H17NO4 91 94 Adenine HMDB00034 C5H5N5 20 41 Adenosine HMDB00050 C10H13N5O4 20 13 Adipic acid HMDB00448 C6H10O4 31 17 Aminoadipic acid HMDB00510 C6H11 NO4 41 35 Anserine HMDB00194 C10H16N4O3 7 7 Ascorbic acid HMDB00044 C6H8O6 54 52 Asymmetric dimethylarginine HMDB01539 C8H18N4O2 98 100 Azelaic acid HMDB00784 C9H16O4 2 4 Benzoic acid HMDB01870 C7H6O2 30 43
124
Baseline (n=54) Week 8 (n=54) Compounds HMDB ID Formula (occurrence, %) (occurrence, %) Betaine HMDB00043 C5H11 NO2 100 100 Butyric acid HMDB00039 C4H8O2 6 6 Caffeine HMDB01847 C8H10N4O2 28 28 Carnosine HMDB00033 C9H14N4O3 22 26 Choline HMDB00097 C5H14NO 6 ND Cinnamic acid HMDB00567 C9H8O2 46 41 Citraconic acid HMDB00634 C5H6O4 100 100 Citramalic acid HMDB00426 C5H8O5 19 6 Citric acid HMDB00094 C6H8O7 57 20 Creatine HMDB00064 C4H9N3O2 69 72 Creatinine HMDB00562 C4H7N3O 100 100 Cytosine HMDB00630 C4H5N3O 81 50 D-Galactose HMDB00143 C6H12O6 15 11 D-Glucose HMDB00122 C6H12O6 15 11 D-Fructose HMDB00660 C6H12O6 15 11 Dihydrothymine HMDB00079 C5H8N2O2 72 57 Dimethylglycine HMDB00092 C4H9NO2 6 ND D-Threitol HMDB04136 C4H10O4 2 50 D-Xylitol HMDB02917 C5H12O5 7 7 D-Xylose HMDB00098 C5H10O5 19 19 Erythritol HMDB02994 C4H10O4 2 50 Ethanolamine HMDB00149 C2H7NO 2 ND Ethyl glucuronide HMDB10325 C8H14O7 22 24 Ethylmalonic acid HMDB00622 C5H8O4 19 28 Fumaric acid HMDB00134 C4H4O4 4 ND Glucaric acid HMDB00663 C6H10O8 2 2 Gluconic acid HMDB00625 C6H12O7 13 6 D-Glucuronic acid HMDB00127 C6H10O7 13 7 Glutaconic acid HMDB00620 C5H6O4 28 24 Glutaric acid HMDB00661 C5H8O4 19 28 Glyceric acid HMDB00139 C3H6O4 6 ND Glycerol HMDB00131 C3H8O3 2 ND Guanidoacetic acid HMDB00128 C3H7N3O2 52 50 Hippuric acid HMDB00714 C9H9NO3 89 94 Homocitrulline HMDB00679 C7H15N3O3 35 20 Homogentisic acid HMDB00130 C8H8O4 22 17 Homovanillic acid HMDB00118 C9H10O4 63 67 3-(3-Hydroxyphenyl)-3-hydroxypropanoic acid HMDB02643 C9H10O4 63 67 4-Hydroxyproline HMDB00725 C5H9NO3 48 30 Hypoxanthine HMDB00157 C5H4N4O 67 69 Indoleacetic acid HMDB00197 C10H9NO2 89 85 Inosine HMDB00195 C10H12N4O5 2 ND Isobutyric acid HMDB01873 C4H8O2 13 9 Isocitric acid HMDB00193 C6H8O7 57 20 Isovaleric acid HMDB00718 C5H10O2 7 ND Isovalerylglycine HMDB00678 C7H13NO3 80 80 Kynurenic acid HMDB00715 C10H7NO3 87 87 Kynurenine HMDB00684 C10H12N2O3 30 13 Lactose HMDB00186 C12H22O11 100 100 L-Arabinose HMDB00646 C5H10O5 19 19 L-Arginine HMDB00517 C6H14N4O2 15 11 L-Aspartic acid HMDB00191 C4H7NO4 2 2 L-Carnitine HMDB00062 C7H15NO3 100 100 L-Cystathionine HMDB00099 C7H14N2O4S 9 4 L-Cysteine HMDB00574 C3H7NO2S 24 13 C H N O S L-Cystine HMDB00192 6 12 2 4 20 9 2
125
Baseline (n=54) Week 8 (n=54) Compounds HMDB ID Formula (occurrence, %) (occurrence, %) Levoglucosan HMDB00640 C6H10O5 31 67 Levulinic acid HMDB00720 C5H8O3 30 35 L-Fucose HMDB00174 C6H12O5 4 ND L-Glutamic acid HMDB00148 C5H9NO4 4 ND L-Glutamine HMDB00641 C5H10N2O3 98 100 L-Histidine HMDB00177 C6H9N3O2 43 43 L-Isoleucine HMDB00172 C6H13NO2 89 83 L-Leucine HMDB00687 C6H13NO2 89 83 L-Lysine HMDB00182 C6H14N2O2 28 9 L-Methionine HMDB00696 C5H11 NO2S 17 4 L-Phenylalanine HMDB00159 C9H11 NO2 9 4 L-Threonine HMDB00167 C4H9NO3 15 2 L-Tryptophan HMDB00929 C11 H12N2O2 94 91 L-Tyrosine HMDB00158 C9H11 NO3 81 65 L-Valine HMDB00883 C5H11 NO2 100 100 Maleic acid HMDB00176 C4H4O4 4 ND D-Maltose HMDB00163 C12H22O11 100 100 Mandelic acid HMDB00703 C8H8O3 31 6 Mannitol HMDB00765 C6H14O6 65 44 Methanol HMDB01875 CH4O 2 ND Methylglutaric acid HMDB00752 C6H10O4 41 46 Methylmalonic acid HMDB00202 C4H6O4 15 22 Methylsuccinic acid HMDB01844 C5H8O4 89 94 Monomethyl glutaric acid HMDB00858 C6H10O4 41 46 Thymol HMDB01878 C10H14O 39 35 Myoinositol HMDB00211 C6H12O6 57 48 N-Acetyl-L-aspartic acid HMDB00812 C6H9NO5 26 7 N-Acetylneuraminic acid HMDB00230 C11 H19NO9 69 57 N-Acetylputrescine HMDB02064 C6H14N2O 6 6 N-Methylhydantoin HMDB03646 C4H6N2O2 69 76 Ortho-Hydroxyphenylacetic acid HMDB00669 C8H8O3 48 52 Oxoglutaric acid HMDB00208 C5H6O5 70 52 Oxypurinol HMDB00786 C5H4N4O2 94 91 Pantothenic acid HMDB00210 C9H17NO5 22 33 p-Cresol sulfate HMDB11635 C7H8O4S 6 ND Phenol HMDB00228 C6H6O 2 2 Phenylacetic acid HMDB00209 C8H8O2 59 44 α-N-Phenylacetyl-L-glutamine HMDB06344 C13H16N2O4 98 100 Phenylglyoxylic acid HMDB01587 C8H6O3 35 33 Phosphorylcholine HMDB01565 C5H15NO4P 39 31 p-Hydroxyphenylacetic acid HMDB00020 C8H8O3 48 52 Picolinic acid HMDB02243 C6H5NO2 96 98 Pimelic acid HMDB00857 C7H12O4 41 50 Proline betaine HMDB04827 C7H13NO2 100 100 Pseudouridine HMDB00767 C9H12N2O6 91 81 Pyrocatechol HMDB00957 C6H6O2 44 50 Pyroglutamic acid HMDB00267 C5H7NO3 100 100 Quinolinic acid HMDB00232 C7H5NO4 20 24 Scyllitol HMDB06088 C6H12O6 57 48 Sebacic acid HMDB00792 C10H18O4 4 ND Sorbitol HMDB00247 C6H14O6 65 44 Suberic acid HMDB00893 C8H14O4 6 6 Succinic acid HMDB00254 C4H6O4 15 22 Succinylacetone HMDB00635 C7H10O4 7 4 Sucrose HMDB00258 C12H22O11 100 100 5-(hydroxymethyl)furan-2-carboxylic acid HMDB02432 C6H6O4 67 44 Symmetric dimethylarginine HMDB03334 C8H18N4O2 98 100
126
Baseline (n=54) Week 8 (n=54) Compounds HMDB ID Formula (occurrence, %) (occurrence, %) Syringic acid HMDB02085 C9H10O5 28 30 Taurine HMDB00251 C2H7NO3S 17 13 Threonic acid HMDB00943 C4H8O5 11 15 Thymidine HMDB00273 C10H14N2O5 30 15 Aconitic acid HMDB00958 C6H6O6 28 17 Ferulic acid HMDB00954 C10H10O4 59 48 Trigonelline HMDB00875 C7H7NO2 98 100 Trimethylamine N-oxide HMDB00925 C3H9NO 2 ND Uracil HMDB00300 C4H4N2O2 28 26 Vanillic acid HMDB00484 C8H8O4 22 17 Vanillylmandelic acid HMDB00291 C9H10O5 28 20 Urea HMDB00294 CH4N2O 6 4 Allantoin HMDB00462 C4H6N4O3 87 93 L-Ornithine HMDB00214 C5H12N2O2 6 4 L-Asparagine HMDB00168 C4H8N2O3 46 48 L-Serine HMDB00187 C3H7NO3 2 ND Cholenic acid HMDB00308 C24H38O3 65 37 Deoxycholic acid HMDB00626 C24H40O4 19 6 Ursodeoxycholic acid HMDB00946 C24H40O4 19 6 Glycerol 3-phosphate HMDB00126 C3H9O6P 2 ND 5-Hydroxyindole HMDB59805 C8H7NO 4 2 2-Ethylhydracrylic acid HMDB00396 C5H10O3 100 100 2-Hydroxy-2-methylbutyric acid HMDB01987 C5H10O3 100 100 2-Hydroxyvaleric acid HMDB01863 C5H10O3 100 100 Dihydroxyphenylacetic acid HMDB01336 C8H8O4 24 39 3,4-Dihydroxyphenylpropionate HMDB00423 C9H10O4 81 89 Aconitic acid HMDB00072 C6H6O6 28 17 Erythronic acid HMDB00613 C4H8O5 11 15 β-Lactic acid HMDB00700 C3H6O3 4 ND L-Lactic acid HMDB00190 C3H6O3 4 ND Indolelactic acid HMDB00671 C11 H11 NO3 98 100 Isobutyrylglycine HMDB00730 C6H11 NO3 59 54 Malic acid HMDB00744 C4H6O5 2 2 Malonic acid HMDB00691 C3H4O4 2 ND m-Chlorobenzoic acid HMDB01544 C7H5ClO2 6 4 m-Coumaric acid HMDB01713 C9H8O3 89 87 Salicyluric acid HMDB00840 C9H9NO4 19 30 Palmitic acid HMDB00220 C16H32O2 54 54 Pyruvic acid HMDB00243 C3H4O3 2 ND Stearic acid HMDB00827 C18H36O2 94 72 3-Hydroxyisobutyric acid HMDB00023 C4H8O3 56 81 3-Hydroxysebacic acid HMDB00350 C10H18O5 43 37 5-Hydroxyhexanoic acid HMDB00525 C6H12O3 100 91 5-Hydroxyindoleacetic acid HMDB00763 C10H9NO3 59 43 Aminomalonic acid HMDB01147 C3H5NO4 2 ND Uric acid HMDB00289 C5H4N4O3 98 100 Vanillactic acid HMDB00913 C10H12O5 17 20 D-Xylulose HMDB01644 C5H10O5 19 19 2-Pentanone HMDB34235 C5H10O 2 ND 3-Hexanone HMDB00753 C6H12O 100 100 2-Hexanone HMDB05842 C6H12O 100 100 4-Heptanone HMDB04814 C7H14O 9 4 Diallyl sulfide HMDB36491 C6H10S 7 6 P-Menthane HMDB31455 C10H20 4 2 P-Cresol HMDB01858 C7H8O 98 96 Neoisomenthol HMDB35764 C10H20O 9 2 Menthol HMDB03352 C10H20O 9 2
127
Baseline (n=54) Week 8 (n=54) Compounds HMDB ID Formula (occurrence, %) (occurrence, %) Methyl salicylate HMDB34172 C8H8O3 48 52 Carvone HMDB35824 C10H14O 33 44 O-Thymol HMDB35770 C10H14O 33 44 Piperitone HMDB34975 C10H16O 46 31 β-Cyclocitral HMDB41011 C10H16O 46 31 Cuminaldehyde HMDB02214 C10H12O 22 13 Oxidized latia luciferin HMDB32913 C13H22O 13 4 α-Ionene HMDB59826 C13H20O 17 2 Butyrylcarnitine HMDB02013 C11 H21NO4 98 100 Butenylcarnitine HMDB13126 C11 H19NO4 100 100 Linoleyl carnitine HMDB06469 C25H45NO4 81 59 Decanoylcarnitine HMDB00651 C17H33NO4 98 91 9-Decenoylcarnitine HMDB13205 C17H31NO4 98 100 Dodecanoylcarnitine HMDB02250 C19H37NO4 94 85 Trans-2-Dodecenoylcarnitine HMDB13326 C19H35NO4 81 63 Glutarylcarnitine HMDB13130 C12H21NO6 98 100 L-Hexanoylcarnitine HMDB00756 C13H25NO4 91 61 2-Hexenoylcarnitine HMDB13161 C13H23NO4 87 80 Hydroxytetradecenoyl-L-carnitine HMDB13330 C21H39NO5 69 48 Malonylcarnitine HMDB02095 C10H17NO6 85 94 Methylglutaryl-L-carnitine HMDB00552 C13H23NO6 98 100 Methylmalonyl-L-carnitine HMDB13133 C11 H19NO6 98 100 Nonanoylcarnitine HMDB13288 C16H31NO4 98 100 L-Octanoylcarnitine HMDB00791 C15H29NO4 74 80 Octenoyl-L-carnitine HMDB13324 C15H27NO4 100 100 Pimelylcarnitine HMDB13328 C14H25NO6 83 93 Propenoylcarnitine HMDB13124 C10H17NO4 54 52 Propionylcarnitine HMDB00824 C10H19NO4 96 100 Tetradecanoylcarnitine HMDB05066 C21H41NO4 11 11 Cis-5-Tetradecenoylcarnitine HMDB02014 C21H39NO4 33 11 Tiglylcarnitine HMDB02366 C12H21NO4 85 85 Valerylcarnitine HMDB13128 C12H23NO4 76 81 Citrulline HMDB00904 C6H13N3O3 44 39 Acetyl-ornithine HMDB03357 C7H14N2O3 96 96 Dopamine HMDB00073 C8H11 NO2 94 100 Histamine HMDB00870 C5H9N3 2 6 Hydroxykynurenine HMDB00732 C10H12N2O4 2 2 L-Dopa HMDB00181 C9H11 NO4 96 98 Phenylethylamine HMDB12275 C8H11 N 87 85 Sarcosine HMDB00271 C3H7NO2 ND 2 Serotonin HMDB00259 C10H12N2O 83 56 Biochanin A HMDB02338 C16H12O5 2 2 Coumesterol HMDB02326 C15H8O5 ND 2 Daidzein HMDB03312 C15H10O4 2 ND Equol HMDB02209 C15H14O3 2 ND Formononetin HMDB05808 C16H12O4 2 ND Genistein HMDB03217 C15H10O5 2 6 Cysteinylglycine HMDB00078 C5H10N2O3S 7 2 C H N O Glutathione HMDB00125 10 17 3 6 2 2 S Homocysteine HMDB00742 C4H9NO2S 6 2 ND: not detected; HMBD ID: Human Metabolome Database identification number 1Percentage of identified metabolites based on the Human Urine Metabolome (Bouatra et al., 2013 and http://www.urinemetabolome.ca/): 273/345=79%
128
Appendix II. Identified metabolites (194 out of 281) detected in serum of PAD patients compared to the Human Serum Metabolome.1 Baseline (n=53) Week 8 (n=53) Compound HMDB ID Formula (occurrence, %) (occurrence, %) 2-Hydroxybutyric acid HMDB00008 C4H8O3 13 38 α-ketoisovaleric acid HMDB00019 C5H8O3 19 34 3-Hydroxybutyric acid HMDB00357 C4H8O3 13 38 Acetaminophen HMDB01859 C8H9NO2 34 49 Acetoacetic acid HMDB00060 C4H6O3 2 13 L-Alanine HMDB00161 C3H7NO2 6 2 L-Arginine HMDB00517 C6H14N4O2 13 36 L-Asparagine HMDB00168 C4H8N2O3 17 15 L-Aspartic acid HMDB00191 C4H7NO4 19 13 Betaine HMDB00043 C5H11 NO2 4 4 L-Carnitine HMDB00062 C7H15NO3 36 87 Choline HMDB00097 C5H14NO 25 75 Creatine HMDB00064 C4H9N3O2 34 74 Creatinine HMDB00562 C4H7N3O 34 75 L-Cysteine HMDB00574 C3H7NO2S 8 4 C H N O S L-Cystine HMDB00192 6 12 2 4 ND 8 2 D-Glucose HMDB00122 C6H12O6 42 87 L-Glutamic acid HMDB00148 C5H9NO4 19 30 L-Glutamine HMDB00641 C5H10N2O3 4 8 L-Histidine HMDB00177 C6H9N3O2 13 4 Hypoxanthine HMDB00157 C5H4N4O 11 8 Isobutyric acid HMDB01873 C4H8O2 36 83 L-Isoleucine HMDB00172 C6H13NO2 26 70 L-Leucine HMDB00687 C6H13NO2 26 70 L-Lactic acid HMDB00190 C3H6O3 ND 2 L-Lysine HMDB00182 C6H14N2O2 2 ND L-Methionine HMDB00696 C5H11 NO2S 4 6 Methylmalonic acid HMDB00202 C4H6O4 4 ND L-Ornithine HMDB00214 C5H12N2O2 ND 2 L-Phenylalanine HMDB00159 C9H11 NO2 13 25 L-Proline HMDB00162 C5H9NO2 19 34 Pyruvic acid HMDB00243 C3H4O3 2 ND L-Threonine HMDB00167 C4H9NO3 6 4 L-Tryptophan HMDB00929 C11 H12N2O2 25 19 L-Tyrosine HMDB00158 C9H11 NO3 6 4 L-Valine HMDB00883 C5H11 NO2 4 4 α-Hydroxyisobutyric acid HMDB00729 C4H8O3 13 38 Xanthine HMDB00292 C5H4N4O2 6 4 4-Hydroxybutyric acid HMDB00710 C4H8O3 13 38 Aminomalonic acid HMDB01147 C3H5NO4 8 ND Benzoic acid HMDB01870 C7H6O2 6 ND Citric acid HMDB00094 C6H8O7 4 2 Erythronic acid HMDB00613 C4H8O5 8 ND Fumaric acid HMDB00134 C4H4O4 4 2 Glyceric acid HMDB00139 C3H6O4 8 2 Tartaric acid HMDB00956 C4H6O6 8 ND Citraconic acid HMDB00634 C5H6O4 15 4 Nicotinic acid HMDB01488 C6H5NO2 2 2 Pyroglutamic acid HMDB00267 C5H7NO3 17 15 Succinic acid HMDB00254 C4H6O4 4 ND
129
Baseline (n=53) Week 8 (n=53) Compound HMDB ID Formula (occurrence, %) (occurrence, %) Uric acid HMDB00289 C5H4N4O3 9 21 Arachidonic acid HMDB01043 C20H32O2 34 51 Cholesterol HMDB00067 C27H46O 38 72 Capric acid HMDB00511 C10H20O2 30 40 Dodecanoic acid HMDB00638 C12H24O2 ND 9 Arachidic acid HMDB02212 C20H40O2 21 38 Heptadecanoic acid HMDB02259 C17H34O2 26 58 Linoleic acid HMDB00673 C18H32O2 38 87 Oleic acid HMDB00207 C18H34O2 32 85 Vaccenic acid HMDB03231 C18H34O2 32 85 Palmitelaidic acid HMDB12328 C16H30O2 34 77 Palmitoleic acid HMDB03229 C16H30O2 34 77 Palmitic acid HMDB00220 C16H32O2 34 83 Stearic acid HMDB00827 C18H36O2 34 83 Myristic acid HMDB00806 C14H28O2 34 81 D-Fructose HMDB00660 C6H12O6 42 87 D-Galactose HMDB00143 C6H12O6 42 87 D-Glucose HMDB00122 C6H12O6 42 87 Myoinositol HMDB00211 C6H12O6 42 87 L-Iditol HMDB11632 C6H14O6 4 2 Hydroxyproline HMDB00725 C5H9NO3 2 2 D-Maltose HMDB00163 C12H22O11 2 15 Acetylglycine HMDB00532 C4H7NO3 4 4 N-Alpha-acetyllysine HMDB00446 C8H16N2O3 6 6 Ribitol HMDB00508 C5H12O5 2 ND D-Xylitol HMDB02917 C5H12O5 2 ND Salicylic acid HMDB01895 C7H6O3 28 55 Pentadecanoic acid HMDB00826 C15H30O2 26 77 Nonadeca-10(Z)-enoic acid HMDB13622 C19H36O2 ND 2 Eicosenoic acid HMDB02231 C20H38O2 25 40 Gamma-Linolenic acid HMDB03073 C18H30O2 34 40 Bovinic acid HMDB03797 C18H32O2 38 87 Alpha-Linolenic acid HMDB01388 C18H30O2 34 40 5,8,11-Eicosatrienoic acid HMDB10378 C20H34O2 11 4 8,11,14-Eicosatrienoic acid HMDB02925 C20H34O2 11 4 Adrenic acid HMDB02226 C22H36O2 2 9 Stearidonic acid HMDB06547 C18H28O2 4 2 Eicosapentaenoic acid HMDB01999 C20H30O2 9 11 Docosapentaenoic acid HMDB06528 C22H34O2 15 6 Docosahexaenoic acid HMDB02183 C22H32O2 30 23 Acetylglycine HMDB00532 C4H7NO3 4 4 Erythronic acid HMDB00613 C4H8O5 8 ND Palmitoylethanolamide (C16:0) HMDB02100 C18H37NO2 34 85 Stearoylethanolamide (C18:0) HMDB13078 C20H41NO2 21 34 Oleoylethanolamine (C18:1n9) HMDB02088 C20H39NO2 8 6 Linoleoyl ethanolamide (C18:2n6) HMDB12252 C20H37NO2 ND 8 α-Linolenoyl ethanolamide (C18:3n3) HMDB13624 C20H35NO2 2 ND Dihomo-γ-Linolenoyl ethanolamide (C20:3n6) HMDB13625 C22H39NO2 15 9 Arachidonylethanolamide (C20:4n6) HMDB04080 C22H37NO2 2 2 Docosatetraenoylethanolamine (C22:4n6) HMDB13626 C24H41NO2 6 4 Cervonoyl ethanolamide (C22:6n3) HMDB13627 C24H36O3 26 57 PGF2a ethanolamide HMDB13628 C22H39NO5 4 4 20-HETE ethanolamide HMDB13630 C22H37NO3 ND 2 MG(18:1n9/0:0) HMDB11567 C21H40O4 40 89 MG(18:2n6/0:0) HMDB11568 C21H38O4 36 79 MG(20:4n6/0:0) HMDB11578 C23H38O4 30 72 MG(0:0/18:1n9) HMDB11537 C21H40O4 40 89
130
Baseline (n=53) Week 8 (n=53) Compound HMDB ID Formula (occurrence, %) (occurrence, %) MG(0:0/18:2n6) HMDB11538 C21H38O4 36 79 MG(0:0/20:4n6) HMDB04666 C23H38O4 30 72 N-Oleoyl glycine (C18:1n9) HMDB13631 C20H37NO3 ND 9 N-Arachidonoyl glycine (C20:4n9) HMDB05096 C22H35NO3 30 70 15-HETE HMDB03876 C20H32O3 15 11 11-HETE HMDB04682 C20H32O3 15 11 5-HETE HMDB11134 C20H32O3 15 11 15-HPETE HMDB04244 C20H32O4 ND 2 12-HPETE HMDB04243 C20H32O4 ND 2 Lipoxin A4 HMDB04385 C20H32O5 2 2 20-Carboxy-leukotriene B4 HMDB06059 C20H30O6 23 42 14,15-DiHETrE HMDB02265 C20H34O4 6 6 11,12-DiHETrE HMDB02314 C20H34O4 6 6 8,9-DiHETrE HMDB02311 C20H34O4 6 6 5,6-DiHETrE HMDB02343 C20H34O4 6 6 14(15)-EpETrE HMDB02283 C20H32O3 15 11 Leukotriene E4 HMDB02200 C23H37NO5S 6 2 Thromboxane B2 HMDB03252 C20H34O6 ND 2 6-Keto-prostaglandin F1a HMDB02886 C20H34O6 ND 2 Prostaglandin F2a HMDB01139 C20H34O5 4 4 Prostaglandin D2 HMDB01403 C20H32O5 2 ND 15(S)-Hydroxyeicosatrienoic acid (15-HETrE) HMDB05045 C20H34O3 38 49 Prostaglandin E3 HMDB02664 C20H30O5 2 ND Resolvin E1 HMDB10410 C20H30O5 4 2 Resolvin D1 HMDB03733 C22H32O5 8 6 12,13-DiHOME HMDB04705 C18H34O4 2 6 9,10-DiHOME HMDB04704 C18H34O4 2 6 12(13)-EpOME HMDB04702 C18H32O3 25 23 9(10)-EpOME HMDB04701 C18H32O3 25 23 9-HODE HMDB10223 C18H32O3 23 9 12(13)Ep-9-KODE HMDB13623 C18H30O4 2 4 13-KODE HMDB04668 C18H30O3 2 ND 9,12,13-TriHOME HMDB04708 C18H34O5 4 17 9,10,13-TriHOME HMDB04710 C18H34O5 4 17 15(16)-EpODE HMDB10206 C18H30O3 2 ND 12(13)-EpODE HMDB10200 C18H30O3 2 ND Decanoylcarnitine HMDB00651 C17H33NO4 36 85 9-Decenoylcarnitine HMDB13205 C17H31NO4 34 85 Linoleyl carnitine HMDB06469 C25H45NO4 17 13 Dodecanoylcarnitine HMDB02250 C19H37NO4 36 83 Tetradecanoylcarnitine HMDB05066 C21H41NO4 32 70 cis-5-Tetradecenoylcarnitine HMDB02014 C21H39NO4 49 85 Hexadecenoylcarnitine HMDB13207 C23H43NO4 15 25 Stearoylcarnitine HMDB00848 C25H49NO4 30 81 Octadecenoylcarnitine HMDB13338 C25H47NO4 38 85 Propionylcarnitine HMDB00824 C10H19NO4 2 11 L-Acetylcarnitine HMDB00201 C9H17NO4 36 81 Butyrylcarnitine HMDB02013 C11 H21NO4 4 9 Valerylcarnitine HMDB13128 C12H23NO4 8 23 L-Octanoylcarnitine HMDB00791 C15H29NO4 2 4 Nonanoylcarnitine HMDB13288 C16H31NO4 30 72 Palmitoleic acid (C16:1) HMDB03229 C16H30O2 8 23 CE(16:1n7/0:0) HMDB00658 C43H74O2 4 4 CE(14:1n5/0:0) HMDB10367 C41H70O2 ND 2 CE(14:0) HMDB06725 C41H72O2 ND 2 CE(24:0) HMDB10376 C51H92O2 2 4 CE(18:1n7/0:0) HMDB05189 C45H78O2 9 30
131
Baseline (n=53) Week 8 (n=53) Compound HMDB ID Formula (occurrence, %) (occurrence, %) CE(18:1n9/0:0) HMDB00918 C45H78O2 9 30 CE(20:1n9/0:0) HMDB05193 C47H82O2 ND 2 CE(20:3n9/0:0) HMDB10373 C47H78O2 ND 2 CE(20:3n6/0:0) HMDB06736 C47H78O2 ND 2 CE(22:1n9/0:0) HMDB10372 C49H86O2 ND 6 CE(18:3n3/0:0) HMDB10370 C45H74O2 ND 2 CE(18:3n6/0:0) HMDB10369 C45H74O2 ND 2 CE(20:2n6/0:0) HMDB06734 C47H80O2 ND 4 CE(22:2n6/0:0) HMDB06737 C49H84O2 ND 2 CE(22:4n6/0:0) HMDB06729 C49H80O2 26 66 CE(22:5n3/0:0) HMDB10375 C49H78O2 30 74 CE(22:5n6/0:0) HMDB10374 C49H78O2 30 74 CE(20:5n3/0:0) HMDB06731 C47H74O2 32 76 CE(22:6n3/0:0) HMDB06733 C49H76O2 25 72 LysoPC(15:0) HMDB10381 C23H48NO7P 11 6 LysoPC(14:0) HMDB10379 C22H46NO7P 2 ND LysoPC(24:0) HMDB10405 C32H66NO7P 2 ND LysoPC(16:0) HMDB10382 C24H50NO7P 2 ND LysoPC(24:1n9/0:0) HMDB10406 C32H64NO7P 4 2 LysoPC(18:3n3/0:0) HMDB10388 C26H48NO7P 8 15 LysoPC(18:3n6/0:0) HMDB10387 C26H48NO7P 8 15 LysoPC(20:3n6/0:0) HMDB10394 C28H52NO7P ND 2 LysoPC(20:3n9/0:0) HMDB10393 C28H52NO7P ND 2 LysoPC(20:4n3/0:0) HMDB10396 C28H50NO7P 2 ND LysoPC(20:4n6/0:0) HMDB10395 C28H50NO7P 2 ND LysoPC(22:5n6/0:0) HMDB10402 C30H52NO7P 6 8 LysoPC(22:5n3/0:0) HMDB10403 C30H52NO7P 6 8 LysoPC(18:4n3/0:0) HMDB10389 C26H46NO7P ND 4 LysoPC(20:5n3/0:0) HMDB10397 C28H48NO7P 4 23 LysoPC(22:6n3/0:0) HMDB10404 C30H50NO7P 8 15 ND: not detected; HMBD ID: Human Metabolome Database identification number 1Percentage of identified metabolites based on the Human Serum Metabolome (Psychogios et al., 2011; http://www.serummetabolome.ca/): 194/281=69%
132
Appendix III. Medications and vitamin/mineral supplements consumed by study participants. Drug or Vitamin/Mineral Drug or Vitamin/Mineral Participants Detected Participants Detected Supplement Supplement Multi-vitamin supplement n=6 √ Ezetrol (Ezetimibe) n=3 √ Advair (Fluticasone/Salmeterol) n=3 √ Famotidine n=1 √ Albuterol (Salbutamol-4'-O- n=2 √ Fenofibrate n=1 × sulfate) Altace (Ramipril) n=12 √ Finasteride n=2 √ Amlodipine n=9 √ Flowmax (Tamsulosin (rINN)) n=5 √ Aspirin n=47 √ Folic acid n=4 √ Atacand (Candesartan) n=3 √ Fosamax (Alendronic acid) n=3 √ Atenolol n=2 √ Fosinopril n=1 √ Atorvastatin n=23 √ Furosemide n=5 √ Calcium n=5 √ Gabapentin n=2 √ Cilazapril n=2 √ Gliclazide n=4 √ Citalopram n=1 √ Glucosamine n=3 √ Clopidogrel n=10 √ Glyburide n=4 √ Cortisone n=1 √ HCTZ (Hydrochlorothiazide) n=15 √ COSOPT (Dorzolamide/Timolol n=1 √ Indapamide n=1 √ eye drops) Diclofenac n=1 √ Indomethacin n=2 × Dilantin (Phenytoin) n=1 √ Irbesartan n=8 √ Diltiazem n=2 √ Iron (Ferrous Sulphate) n=2 √ Docosahexaenoic n=4 √ Levothyroxine n=7 √ Acid/Eicosapentaenoic acid Enalapril n=3 √ Lisinopril n=1 √ Lorazepam n=3 √ Repaglinide n=1 √ Losartan n=2 √ Robaxacet (Methocarbamol) n=1 √ Magnesium n=1 √ Rosuvastatin n=16 √ Metformin n=11 √ Sertraline n=1 × Metoprolol n=23 √ Simvastatin n=7 √ Mirapex (Pramipexole) n=1 × Sitagliptin n=1 √ Nadolol n=1 √ Spiriva (Tiotropium bromide) n=3 √ Symbicort Naproxen n=1 √ n=3 √ (Budesonide/Formoterol) Niacin (Nicotinic acid) n=1 √ Telmisartan n=2 √ Nifedipine n=3 √ Trandolapril n=1 √ Omeprazole/Esomeprazole n=9 √ Triamterene n=1 √ Pentoxifylline n=1 √ Tylenol n=4 √ Perindopril n=1 √ Valsartan n=1 √ Pravastatin n=2 √ Venlafaxine n=1 √ Prednisone n=3 √ Vitamin B12 n=10 √ Primidone n=1 √ Vitamin B6 n=1 √ Quinapril n=6 √ Vitamin C n=8 √ Quinine n=1 √ Vitamin D n=15 √ Rabeprazole n=4 √ Vitamin E n=4 √ Ranitidine n=1 √ Zopiclone n=1 × Note: √: The drug (or vitamin/mineral) and its metabolites can be detected in biological fluids of PAD patients; ×: The drug (or vitamin/mineral) and its metabolites cannot be detected in biological fluids of PAD patients; n: The sum of PAD patients who took the drug (or vitamin/mineral).
133
Percentage of medications and/or vitamin/mineral supplements detected by the method applied: 75/80=94%
134