University of South FloridaScholar Commons

Graduate Theses and Dissertations Graduate School

March 2018

Impact of Obesity and Expression of Obesity-Related Genes in the Progression of ProstateCancer in African American MenMmadili Nancy IlozumbaUniversity of South Florida, ilozumbammadili@gmail.com

Follow this and additional works at: http://scholarcommons.usf.edu/etd

Part of the Epidemiology Commons

This Thesis is brought to you for free and open access by the Graduate School at Scholar Commons. It has been accepted for inclusion in GraduateTheses and Dissertations by an authorized administrator of Scholar Commons. For more information, please contact scholarcommons@usf.edu.

Scholar Commons CitationIlozumba, Mmadili Nancy, "Impact of Obesity and Expression of Obesity-Related Genes in the Progression of Prostate Cancer inAfrican American Men" (2018). Graduate Theses and Dissertations.http://scholarcommons.usf.edu/etd/7171

Impact of Obesity and Expression of Obesity-

Related Genes in the Progression of Prostate Cancer

in African American Men

by

Mmadili Nancy Ilozumba

A thesis submitted in partial fulfillment

of the requirements of the degree of

Master of Science in Public Health

Department of Epidemiology and Biostatistics

with a concentration in Epidemiology

College of Public Health

University of South Florida

Major Professor: Janice Zgibor, M.P.H., RPh, Ph.D.

Committee: Jong Park, M.S., M.P.H., Ph.D.

Committee: Skai Schwartz, B.A., M.A., Ph.D.

Date of Approval:

March 13, 2018

Keywords: biochemical recurrence of prostate cancer, gene expression, obesity, health disparity

Copyright (c) 2018, Mmadili Nancy Ilozumba

DEDICATION

This thesis is dedicated to my loving family. Thank you daddy for sponsoring my

Master’s program in the US, despite the economic recession in our country, Nigeria. You

provided all my needs and I really do not know how to thank you enough. Thank you for

encouraging me to do my best and prove myself. Thank you for being proud of me. Thank you

mummy for being my rock, for your ceaseless prayers and for your encouragement. Indeed, your

answered prayers are evident in my life. Thank you for your love, Mum. Thank you mum and

dad for being my role models and for raising me to be the kind of woman I am today. I love you

so much, mum and dad. To my lovely sisters and brother: Chinasa, Usochi, Odichi, Eziolu and

Ebube, thanks for the moral support, encouragement and care. I would not have come this far

without your presence in my life. I love you, guys! God bless everyone of you!

ACKNOWLEDGEMENTS

I immensely thank the Almighty God who made this thesis a dream come true. I would

not have been able to get this work done without His infinite mercy and grace. I am also thankful

to God for an opportunity to further my education far away from my country, Nigeria. The

struggles and challenges were so real but God kept me strong even when it seemed like I could

not go on. Thank You, Jesus!

My profound gratitude goes to Dr. Jong Park who has been an incredible and a kind

mentor. Your humility and great sense of humor are commendable. I remember how you swiftly

responded to my email when I asked for an opportunity in prostate cancer research. You

immediately added me to your team and working with you has been swell. You always spared

some time for me whenever I had questions regarding my thesis work despite how busy you

were. You always welcomed me in your office and patiently answered all my questions. I would

not have been able to start and finish my thesis without you providing me with data and

mentorship. Thank you so much!

My profound gratitude also goes to Dr. Janice Zgibor, my academic advisor and faculty

mentor. Thank you so much for being so kind, welcoming and for helping me navigate my

Masters’ program. You selflessly decided to be a major professor in my thesis so that you could

assist me in the best way possible. Thank you for your advices, corrections, writing aid,

supervision and mentorship.

My profound appreciation goes to Dr. Skai Schwartz and Dr Hung Luu. Thank you for

the contributions you made in my thesis. Thank you for the insightful suggestions and

corrections. I really appreciate your mentorship.

I want to thank Ganna Chornokur, who contributed to this project during her post-

doctoral training at Moffitt Cancer Center and Research Institute. I also want to thank my

colleague Bashir Dabo for his contribution to this project.

i

TABLE OF CONTENTS

LIST OF TABLES ii

LIST OF FIGURES iii

ABSTRACT iv

INTRODUCTION 1

Health Disparity of Prostate Cancer 1

Obesity and Prostate Cancer Risk 3

Overview of Genetic Variation and Prostate Cancer Risk 4

Gene Mutation 4

Genetic Polymorphism 5

Gene Expression 6

Significance of the Study 8

METHODOLOGY 10

Tissue Processing and Sample Selection 10

Expression Assay 10

Gene Selection 11

Statistical Analysis 11

RESULTS 14

DISCUSSION 32

Limitation 34

CONCLUSION 35

REFERENCES 37

APPENDICES 43

Appendix A 44

ii

LIST OF TABLES

Table 1: Descriptive and Clinical Characteristics for the 48 African American Prostate

Cancer Patients 18

Table 2: Dichotomous Clinical and Demographic Characteristics for the 48 African

American Prostate Cancer Patients 19

Table 3: Comparing the Means for Gene Expression among Tumor Tissue Prostatic

Intraepithelial Neoplasia (PIN) and Normal Tissue 20

Table 4: Comparing the Means for Gene Expression among the Tissue Types Stratified

by Obesity (Obese and Non-obese) 21

Table 5: Testing the Association between Obesity and the Recurrence of Prostate

Cancer 22

Table 6: Simple Logistic Regression Testing the Association between Gene Expression

and Biochemical Recurrence of Prostate cancer 23

Table 7: Multiple Logistic Regression Testing the Association between Gene Expression

and Biochemical recurrence of Prostate cancer 24

Table 8: An overlap of Genes 25

Table 9: P-value for a Kaplan Meier Curve Depicting the Role of Obesity in Prostate

Cancer progression 25

iii

LIST OF FIGURES

Figure 1: Kaplan Meier Curve depicting the role of obesity in prostate cancer Progression 26

Figure 2: Trend of Gene Expression in Normal, PIN and Tumor Tissue Sample 27

iv

ABSTRACT

In the US, the incidence and mortality rates of prostate cancer (PCa) are higher among

African American men compared to European American men. Obesity is an important risk factor

of PCa. Obesity is known to alter the gene expression profiles in prostate tumors. This study

evaluates the impact of obesity and the expression of obesity-related genes on the progression of

PCa in African American men.

The primary outcome of interest is biochemical recurrence (BCR) of PCa. There were

48 African American prostate cancer patients in the study. The tissue samples included 42

normal tissues, 40 Prostate Intraepithelial Neoplasia (PIN) and 45 tumor tissues (127 tissue

samples in total). We assembled 99 obesity-related genes and determined the levels of their

expression in the three types of tissue samples using Nanostring Technologies. An ANOVA test

was used to compare the means for gene expression among normal, PIN and tumor tissue

samples. Unconditional logistic regression models were used to calculate odds ratios (ORs) and

their respective 95% confidence intervals (95% CIs) to determine the association between

obesity and BCR as well as gene expression and BCR. Results were regarded as statistically

significant if p-values were less than 0.05. A Kaplan Meier Curve was constructed to depict the

survival time and time to event (BCR) among obese and non-obese African American prostate

cancer patients. Patients were followed up from the date of first surgery to the date of

biochemical recurrence or date of last follow-up. Statistical analysis was done with SAS 9.4

software.

v

Forty-three obesity-related genes were statistically significantly associated with

biochemical recurrence. There was no association between obesity and biochemical recurrence

(BCR) in obese African American men compared to non-obese African American men (OR=

2.03, 95% CI = 0.22 - 18.77, p-value= 0.53). Twenty genes showed an upward trend in gene

expression among normal, PIN and tumor tissue samples including ADIPOR1, AKRIC4,

ALOX12, ALOX15, CRYBB2, EIF5A, ERG, GNPDA2, HNF1B, HSD3B1, KLK4, LEP, MC4R,

MTCH2, PCSK1, PIK3CB, SLC2A2, STAT1, SULT1A1, YY1. The probability of survival (not

having BCR) is lower in obese African American men compared to non-obese African American

men as indicted in the Kaplan Meier curve. In other words, the probability of developing BCR is

higher in obese African American men compared to non-obese African American men.

We did not find a significant association between obesity and biochemical recurrence.

However, we elucidated some obesity-related genes that could explain PCa carcinogenesis.

Further studies are needed to determine functional significance of these selected obesity-related

genes and the role they play in encouraging PCa progression in African American men.

1

INTRODUCTION

Health Disparities in Prostate Cancer

There is an apparent disparity in prostate cancer (PCa) among different races in the

United States (US). African American men (AAM) are disproportionately affected by the burden

of prostate cancer. Amongst European American men, the incidence rate of prostate cancer is

approximately 113 per 100,000 persons while the prostate cancer incidence rate is approximately

189 per 100,000 persons in African American males (National Cancer Institute, 2017b). In terms

of prostate cancer death rates, approximately 19 per 100,000 White men die from prostate cancer

while 42 per 100, 000 AAM die from prostate cancer (National Cancer Institute, 2017b). Hence,

African American males have both the highest incidence rate and mortality rate of prostate

cancer compared to other races in the US. There are risk factors that are likely responsible for

these disparities. The socio-economic status (SES) of an individual is a potential risk factor for

prostate cancer disparity. A study by (Byers et al., 2008) showed that low SES significantly

elevated the risk of dying from prostate cancer among men. The risk was elevated because of

later diagnosis of prostate cancer and less aggressive treatment. Being a minority population in

the US, AAM are more likely to have a low SES. Low SES compromises the health of

individuals. These individuals tend to lack access to health care and health insurance coverage

(Becker & Newsom, 2003, American Cancer Society, 2016, Fiscella & Williams, 2004). In terms

of median value of assets, Whites have a net worth of $103976 while Blacks have a net worth of

$9211 (US Census Bureau, 2017). Low SES encourages behavioral risk factors that increase the

risk of prostate cancer development.

2

Behavioral risk factors for prostate cancer disparities including lack of access to healthy

foods, obesity, physical inactivity, cigarette smoking, alcohol consumption and other health

behaviors are more common among people with low SES and minority population (Pampel,

Krueger, & Denney, 2010). In terms of diet and prostate cancer risk, consumption of red meat,

high fat diets and low carotenoid intake increases the risk of prostate cancer (Jianjun, Dhakal,

Greene, Lang, & Kadlubar, 2011). A study by Taksler, Keating, & Cutler (2012) that was aimed

at elucidating the differences in prostate cancer attributed to race revealed that the estimated

disparity in prostate cancer mortality was 1320 more cases per 100000 males amongst Black

males compared with White males. This disparity was due to an increased number of new cases

of prostate cancer among Black males (76%) as well as an increased stage-specific death rate at

diagnosis (24%). Twenty-nine percent of the difference in the number of new cases of metastatic

prostate cancer between Black men compared to White men was attributed to the variation in

income, pre-existing health conditions and prostate specific antigen (PSA) testing while the

characteristics of tumor at diagnosis explained fifty percent in racial differences of prostate

cancer (Taksler et al., 2012).

3

Obesity and Prostate Cancer Risk

Obesity, which is a chronic health condition is a risk factor for other chronic diseases

including high blood pressure, stroke and different types of cancer like breast and prostate cancer

(American Cancer Society, 2016). Epidemiological evidence revealed that obesity is associated

with poorer treatment outcomes and higher risk of mortality due to prostate cancer

(Buschemeyer & Freedland, 2007). Obese men are more likely to experience prostate cancer

recurrence after treatment than non-obese men. A study by (Freedland et al., 2004) revealed that

obesity rate is 31% in African American men compared to 21% in White men. They also showed

that obesity is linked with higher grade tumors. In the US, the prevalence of obesity is greater in

Black men than in Whites men. The obesity prevalence is 38% in Black men compared to 35%

in White men (American Cancer Society, 2016). Physical activity reduces the risk of cancer

remission but the prevalence of no leisure physical activity is 34% in Black men compared to

25% in White men (American Cancer Society, 2016). Among men treated with radical

prostatectomy, obese patients were more likely to have more pronounced biochemical failure

rates (Freedland et al., 2004). Obesity also modifies the levels of certain hormones that are

related to prostate cancer risk such as testosterone, leptin, insulin-like-growth-factor-1 and

estrogen (Buschemeyer & Freedland, 2007, Freedland & Aronson, 2004).

The association between obesity and aggressive prostate cancer as a health outcome can

be explained through different mechanisms of action. These mechanisms may be more

pronounced in an obese patient, who has general hormone changes that leads to increased

estradiol, lower sex hormone binding protein and low free testosterone, which results in an

aggressive form of prostate cancer (Freedland & Aronson, 2004). Another mechanism in obese

males is insulin resistance leading to high insulin levels, which results in an aggressive form of

4

prostate cancer (Buschemeyer & Freedland, 2007). Adiponectin, a protein hormone found in

adipose tissues is responsible for the regulation of fatty acids and glucose in the body and plays a

crucial role in obesity-related cancers including cancer of the prostate (Kelesidis, Kelesidis, &

Mantzoros, 2006, Rider et al., 2015, RJ & English, 2006). A low level of adiponectin in the body

suggests risk for the progression and aggressiveness of prostate cancer (Obeid & Hebbard,

2012).

Overview of Genetic Variation and Prostate Cancer Risk

Variation occurs whenever the order of the bases in a DNA sequence changes. Variations

can involve only one base or many bases (National Institute of Health, 2017). If the two strands

of a chromosome are thought of as nucleotides threaded on a string, then, for example, a string

can break and the order of the bases can vary. One or more nucleotides may be changed, added,

or removed. In chromosomes, these changes are called Single Nucleotide Polymorphisms

(SNPS) (National Institute of Health, 2017), insertions, and deletions (Mullaney, Mills, Stephen

Pittard, & Devine, 2010).

In addition to these changes, some DNA sequences called “repeats” like to insert extra

copies of themselves several times. Chromosomes can also undergo more dramatic changes

called translocations (Nambiar & Raghavan, 2011). These occur when an entire section of DNA

on one chromosome switches places with a section on another. The risk for prostate cancer can

be exacerbated in the presence of a variation in the gene of an individual. This variation could be

brought about by some mechanisms such as gene mutations, miRNA alterations, genetic

polymorphisms and epigenetic modification (Arun et al., 2017).

5

Gene Mutation

An inheritance of a mutated gene poses a risk for prostate cancer as well as the

aggressiveness of the disease. Family history of prostate cancer is important in determining the

risk of prostate cancer (American Cancer Society, 2017). Mutation in genes such as BRCA1,

BRCA2, HOXB13 (National Cancer Institute, 2017c) are good predictors of prostate cancer risk.

A study conducted by (Pritchard et al., 2016) found that men with metastatic prostate cancer

were more likely to have an inherited mutation in the gene responsible for DNA-repair compared

with men diagnosed with a localized prostate cancer. They identified mutations in 16 genes

including BRCA2, ATM, CHEK2, BRCA1, PALB2 and RAD51D. The number of new cases of

mutations in DNA repair genes was approximately 12% in men with metastatic prostate cancer,

which exceeded the number of DNA repair gene mutation in men with localized prostate cancer

(Pritchard et al., 2016). DNA-repair genes are responsible for the maintenance of genes,

prevention and correction of errors during DNA replication (Cancer Information and Support

Network, 2017).

Genetic Polymorphisms

Single Nucleotide Polymorphisms (SNP) is a type of genetic variation in which there is a

change in the arrangement of nucleotides in a DNA (National Institute of Health, 2017). A study

carried out by Lipson et al., (2014) was aimed at determining whether the GWAS replication

SNPs and candidate SNPs that showed evidence for prostate cancer susceptibility can also

indicate prostate cancer aggressiveness. They found that among the 34 GWAS replicated SNPs,

4 were associated with prostate cancer aggressiveness in African American males. These SNPs

were rs2660753, rs13254738, rs10090154, rs2735839. Two of the SNPs were found on 8q24

while the other two SNPs were found on 3p12 and 19q13. Jianjun et al., (2011) evaluated

6

whether there was an association between the polymorphism in hOGG1 and XRCC1 genes and

risk of prostate cancer. They found a positive association and the effect of the genes were

modified by plasma antioxidants. These genes are involved in the repair of oxidative DNA

damage. Thus, an existence of variation in these genes increases the risk of prostate cancer.

Genetic variation in the HNF1B is associated with prostate cancer risk (Chornokur et al., 2013),

although the underlying mechanistic pathway(s) remain elusive. Obese men with rs7501939 C-

allele were found to have a higher risk of prostate cancer. rs7501939 is located in the HNF1B

gene that regulates proliferation of prostate cancer (Ross-Adams et al., 2015). Furthermore, 4 of

the 12 single nucleotide polymorphisms (rs10993994, rs5945619, rs4430796, and rs9364554)

studied by Grisanzio et al., (2012) showed a strong association with five genes: MSMB, NCOA4,

NUDT11, HNF1B, and SLC22A3. These genes were involved in prostate cancer pathogenesis.

Gene Expression and Prostate Cancer Risk

Ideally, DNA are the building blocks of genes. Genes produce functional end-products

such as proteins through the process of gene-expression (National Cancer Institute, 2018). For

every gene, individuals have different levels of proteins. Enzymes are proteins and different

people have variation of the same enzyme. The process of gene expression starts with a DNA,

which undergoes transcription to form an RNA. The RNA is then translated into a protein

(National Cancer Institute, 2018). In other words, the presence or absence of a functional enzyme

depends on whether an individual expresses the gene responsible for the enzyme as well as the

level at which the gene is expressed. Consequently, expression levels of certain enzymes likely

have an association with prostate cancer risk (Donkena & Young, 2011, Nghiem et al., 2016).

Proteins may also come in the form of hormones. Androgen metabolism genetics and prostate

cancer risk has been well established. Higher levels of androgen increases the risk of prostate

7

cancer (Michaud, Billups, & Partin, 2015). African American males usually have higher levels of

androgen compared to other races (C. M. Zeigler-Johnson et al., 2008). Although impact of

testosterones on the risk of prostate cancer is still inconclusive, testosterone metabolites, such as

dihydotestosterone (DHL) increase the risk of prostate cancer (Gann, Hennekens, Ma,

Longcope, & Stampfer, 1996). Variations in certain enzymes such as CYP3A4, CYP3A43 and

CYP3A5 which are responsible for the activation of testosterone hormone, are associated with the

aggressiveness of prostate cancer both in African American males and European American males

(C. Zeigler-Johnson et al., 2004). Park et al., (2006) found that there was an association between

the deletion polymorphism of an enzyme called UDP-Glucuronosyltransferase 2B17 and an

increased risk of prostate cancer among AAM and Caucasian men. The enzyme is responsible

for the glucuronidation of many chemicals as well as androgen. In other words, UDP-

Glucuronosyltransferase 2B17 facilitates the excretion of these chemicals and hormones.

Preliminary studies have evaluated the association of gene expression and the risk of

prostate cancer progression. Nghiem et al., (2016) analyzed the association between the

expression of mismatch repair enzymes and prostate cancer risk. They found no association

between the mismatch repair enzymes and biochemical recurrence. Powell et al.,( 2013) showed

that genes associated with prostate cancer are expressed differently between European American

males and African American males. Singh et al., (2002) demonstrated that the expression of

genes can predict the behavioral outcome of prostate cancer in patients. Chandran et al., (2007)

delineated the potential genes expressed in prostate cancer patients that influence the progression

of the disease. In a study conducted by Yu et al., (2004), there was evidence of a difference in

the genes expressed in the tumor and non-tumor samples. They also found that the expression of

these genes might serve as a prognosticative factor in determining the aggressiveness of prostate

8

cancer. Mortensen et al., (2015) discovered the expression of twelve genes that were predictors

of biochemical recurrence of prostate cancer. They also found that patients with recurrence had

an up-regulated expression of SFRP gene.

Significance of the Study

Obesity is known to alter the gene expression profiles in prostate tumors (Ornish et al.,

2008, Ribeiro et al., 2012). Hence, it is imperative to understand the role of obesity in prostate

cancer risk in order to inform adequate control of prostate cancer. There is evidence that tumors

derived from African American men are more aggressive in nature compared to tumors from

European American men (Castro et al., 2009, Lipson et al., 2014, Timofeeva et al., 2009,

Zeigler-Johnson et al., 2004). It is possible that obesity would influence prostate carcinogenesis

in African- American males (AAM), although the studies aimed to verify this hypothesis are

lacking.

Our study hypothesis is that obesity may be associated with an increased risk for prostate

cancer progression in obese African American men compared to non-obese African American

men. We selected a novel candidate-driven selective approach over the traditional non-candidate

gene expression profiling. We tested this hypothesis by:

a. Comparing the means for gene expression among tumor tissue, Prostatic Intraepithelial

Neoplasia (PIN) and normal tissue.

b. Testing the trend of gene expression among tumor, PIN and normal tissues.

c. Comparing the means for gene expression among tumor tissue, Prostatic Intraepithelial

Neoplasia (PIN) and normal tissue stratified by obese and non-obese patients.

d. Testing the association between obesity and the biochemical recurrence of prostate

cancer after initial surgery with the use of a multiple logistic regression

9

e. Testing the association between gene expression and biochemical recurrence of prostate

cancer using logistic regression models at a 95% confidence interval.

f. Constructing a Kaplan Meier Curve depicting the time to event (biochemical recurrence)

among obese and non-obese African American prostate cancer patients.

10

METHODOLOGY

Tissue Processing and Sample Selection

We obtained tissue samples from 48 African American prostate cancer patients Total

Cancer Care bank (a biological specimen repository at Moffitt). The tissue samples included 42

normal tissues, 40 Prostate intraepithelial Neoplasia (PIN) and 45 Tumor tissues. There was a

total of 127 tissue samples in our study. The samples are formalin fixed paraffin embedded

(FFPE). These samples are linked to the epidemiological, clinical and follow up data. Our

primary outcome is the biochemical recurrence of prostate cancer (BCR). Patients’ demographic

and clinical data were obtained, including, age, height, weight, BMI, Gleason score (including

primary and secondary Gleason scores), Prostate Specific Antigen (PSA), Cancer of the Prostate

Risk Assessment (CAPRA) Score, seminal vesicle invasion, surgical margins, lymph node

invasion and extra capsular extension. Patients were followed up from the date of first surgery to

the date of biochemical recurrence or date of last follow-up. A Moffitt pathologist identified and

marked tumor, PIN and normal regions. RNA samples were extracted from marked tissues for

gene expression profiling.

Expression Assay

We used the nCounter gene expression assays by NanoString technologies (Nanostring

Technologies, 2017). These innovative assays can analyze the formalin fixed, paraffin embedded

(FFPE) blocks, as well as fresh-frozen (FF) tumor tissues with comparable accuracy. The nCounter

gene expression assays can analyze up to 800 genes in a single reaction using 100ng or less of total

RNA or FFPE extract. These assays were provided at the Moffitt Center genomics core.

11

Gene Selection

We assembled a list of “obesity-related” genes. A total of 99 obesity-related genes were

included in the gene expression profiling. The databases including “Gene” by NCBI (National

Center for Biotechnology Information, 2017), GeneCards (GeneCards, 2017) and Cancer

Genome Anatomy Project (National Cancer Institute, 2017a) were used to ensure comprehensive

search.

Statistical Analysis

We recoded and categorized the BMI scores in order to identify patients who were

overweight and obese in the sample. The obese and overweight group were those with BMI ≥ 25

while the non-obese and non-overweight group had BMI < 25. We decided to use a cut point of

25 for BMI because overweight and obesity are likely related to an increased risk of prostate cancer

recurrence after treatment (American Cancer Society, 2018). We also recoded and categorized the

Gleason scores in order to determine the number of patients and their respective tissue samples

with an aggressive Gleason score. Gleason scores determines the likelihood of prostate cancer

spreading to other normal tissues (Stark et al., 2009). Gleason score ≥ 7 were recoded as aggressive

while Gleason scores less than 7 were coded as non-aggressive Gleason scores.

Descriptive statistics were used to summarize the participants’ demographic and clinical

characteristics. Mean, standard deviation, minimum, maximum and range were calculated for

continuous variables whereas frequency and percentage were generated for categorical variables.

Gene expressions was normalized using internal housekeeping reference genes, and summarized

using descriptive statistics. We also summarized the descriptive statistics of the expression of 99

genes for the three types of tissue sample using the mean, standard deviation and range. To

12

compare the means for gene expression among tumor tissue, prostatic intraepithelial neoplasia

(PIN) and normal tissue, we used F test. Genes less than the alpha level of 0.05 were regarded as

being differently expressed among the three types of tissue samples and statistically significant.

We constructed a logistic regression model to test the association between obesity and the

biochemical recurrence of prostate cancer after initial surgery. In order to control for

confounding, we carried out a stepwise selection procedure for all covariates including age,

height, weight, BMI, Gleason score, Prostate Specific Antigen (PSA), Cancer of the Prostate

Risk Assessment (CAPRA) Score, seminal vesicle invasion, surgical margins, lymph node

invasion and extra capsular extension. The result of the stepwise selection determined the

significant covariates that would be adjusted for in the logistic regression. No covariate met the

significance level of entry (0.15) and significant level of stay (0.15) in the stepwise selection

model. Hence, we did not adjust for any covariate in the final logistic model.

To determine the association between biochemical recurrence and gene expression, we

first carried out a simple logistic regression for each of the 99 genes present in the study. We

went further to generate a multiple logistic regression in order to control the presence of

confounder. A stepwise selection procedure was done for the 99 genes individually. Variables

such as age, height, weight, BMI, Gleason score, Prostate Specific Antigen (PSA), Cancer of the

Prostate Risk Assessment (CAPRA) Score, seminal vesicle invasion, surgical margins, lymph

node invasion and extra capsular extension were included in the stepwise selection procedure.

The result of each stepwise selection procedure determined the significant covariates adjusted for

each gene in the logistic regression. P-values ≤ 0.05 were regarded as statistically significant.

The odds-ratio and 95% confidence interval for the association between gene expression and

biochemical recurrence were recorded for the results with p-values ≤ 0.05.

13

We constructed a Kaplan Meier’s Curve to determine the time to event, in this case,

prostate cancer recurrence among obese and non-obese African American prostate cancer

patients. Data analyses were performed using SAS 9.4 software (SAS Institute Inc., Cary, NC,

USA).

14

RESULTS

Table 1 shows a summary of the demographic and clinical characteristics of the prostate

cancer patients in our sample. The statistics were stratified into the three types of tissues that we

obtained, namely, normal tissue, prostatic intraepithelial neoplasia (PIN) and tumor tissue.

Patients with normal tissue and PIN had a mean age of 54 years while patients with tumor tissue

had a mean age of 55 years. The minimum and maximum heights of the patients were 161.50cm

and 188cm respectively, while they had a minimum and maximum weights of 62.46kg and

119.67kg respectively. The maximum BMI in our sample was 39.42 kg/m2.The average BMI for

normal and tumor tissues was 29.08 kg/m2 while it was 28.90 kg/m2 in prostatic intraepithelial

neoplasia. The prostate specific antigen (PSA) had a mean above 7ng/ml and a maximum value

of 33.10ng/ml in the sample. This is as expected since all the patients in our sample already had

prostate cancer. PSA levels ≤ 4ng/ml is regarded as normal, levels above 4ng/ml but less than

10ng/ml could pose a threat while levels >10ng/ml significantly increase the risk of prostate

cancer (American Cancer Society, 2017). On a scale of 0-12, the Capra score had a maximum

value of 11 for both normal and tumor tissues but a score of 9 for PIN. Capra score predicts the

likelihood of prostate cancer progression after initial prostatectomy (Punnen et al., 2014). The

average Gleason score for normal tissue, PIN and tumor tissue were 6.64, 6.58 and 6.62

respectively. Normal and tumor tissues had a maximum Gleason score of 9 while PIN had a

maximum value of 9.

Table 2 shows the dichotomous clinical and demographic characteristics of the patients.

Ten patients had a biochemical recurrence while thirty-eight patients did not have a biochemical

15

recurrence. Eighty-three percent (83%) of the patients were obese while 16.67% were non-obese.

Aggressive Gleason score was evident in twenty-nine patients. In terms of metastasis of prostate

cancer, two patients had a seminal vesicle invasion, twelve patients had surgical margins, one

patient had a lymph node invasion while 14.58% of the patients had an extracellular extension.

Table 3 shows the result of the F test used to compare the means of gene expression

among the three types of tissue samples. Only expressed genes that were statistically

significantly different among the three tissue samples are presented. A total of 29 genes were

significantly expressed differently in the normal tissue, PIN and tumor tissue.

Table 4 shows the result of the F test used to compare the means of gene expression

among normal, PIN and tumor tissues stratified by obese and non-obese patients. Genes that

were statistically significantly different among obese and non-obese are marked with asterisk. A

total of 6 genes out of the 29 obesity-related genes that showed a difference in the means for

gene expression among tumor tissue, Prostatic Intraepithelial Neoplasia (PIN) and normal tissue,

were significantly expressed differently between the obese and non-obese African American

prostate cancer patients.

Table 5 shows the result of the test of association between obesity and recurrence of

prostate cancer after initial surgery. We found no association between obesity and biochemical

recurrence of prostate cancer in African American obese patients compared to African American

non-obese patients (OR= 2.032, 95% CI = 0.220- 18.765, p-value= 0.5318). There is no

association between obesity and biochemical recurrence because the p-value is above the alpha

level of 0.05.

Table 6 illustrates the result of the test of association between gene expression and

biochemical recurrence of prostate cancer with no adjustment of covariates. A total of 35 genes

16

showed a statistically significant association with biochemical recurrence of prostate cancer with

a p-value less than 0.05. Their respective odds ratios and 95% confidence intervals are also

represented.

Table 7 illustrates the result of the test of association between gene expression and

biochemical recurrence of prostate cancer with the adjustment of potential covariates. The

covariates we adjusted for include age, height, weight, BMI, Gleason score, Prostate Specific

Antigen (PSA), Cancer of the Prostate Risk Assessment (CAPRA) Score, seminal vesicle

invasion, surgical margins, lymph node invasion and extra capsular extension. A total of 43

genes out of the 99 obesity-related genes showed a statistically significant association with

biochemical recurrence of prostate cancer with a p-value less than 0.05. Their respective odds

ratios and 95% confidence intervals are also represented.

In the Kaplan Meier curves illustrated in Figures 1, the probability of having no

biochemical recurrence of prostate cancer at day zero is 100% or 1 for both the obese and non-

obese African American (AA) patients. However, over time the probability reduces such that at

time 1000days, approximately 15% of obese patients have a biochemical recurrence (BCR) of

prostate cancer while 0% of non-obese patients have BCR. In other words, the probability that an

obese patient will not have a biochemical recurrence of prostate cancer (BCR) beyond the 1000th

day of follow up is approximately 85% while the probability that a non-obese patient will not

have BCR beyond the 1000th day is 100%. At 2000th day of follow up, the probability that an

obese patient will not have BCR is 70% while the probability of not experiencing BCR is

approximately 80% in non-obese AA prostate cancer patients. The probability that an obese

patient will not have BCR beyond 3000days of follow-up is 62% in the obese AA patients

compared to approximately 80% in the non-obese AA patients. The comparison between the two

17

survival curves for obese and no-obese was not statistically significant as indicated by the (p-

value = 0.3118).

Figure 2 shows the trend analysis of gene expression in normal, prostatic intraepithelial

neoplasia and tumor tissue samples. Twenty genes showed an upward trend in gene expression

including ADIPOR1, AKRIC4, ALOX12, ALOX15, CRYBB2, EIF5A, ERG, GNPDA2, HNF1B,

HSD3B1, KLK4, LEP, MC4R, MTCH2, PCSK1, PIK3CB, SLC2A2, STAT1, SULT1A1, YY1.

Table 8 shows the list of genes that showed an overlap between the result from the trend

test for gene expression among the three tissue samples and the result from the means

comparison of gene expression that were significantly different between obese and non-obese

patients among normal, PIN and tumor tissues.

18

Table 1: Descriptive and Clinical Characteristics for the 48 African American Prostate

Cancer Patients

Normal

Variable Mean Standard

Deviation

Minimum Maximum Range

Age 54.29 7.20 40 72 32

Height 177.17 6.31 161.50 188.00 26.50

Weight 91.66 14.29 62.46 119.67 57.21

PSA 7.50 5.77 1.30 33.10 31.80

Capra

Score

2.38 2.46 0 11 11

BMI 29.08 4.80 19.97 39.42 19.45

Gleason

Score

6.64 0.62 6 9 3

PIN

Variable Mean Standard

Deviation

Minimum Maximum Range

Age 54.03 6.90 40 66 26

Height 177.41 6.13 161.50 188.00 26.50

Weight 90.78 14.81 62.46 118.01 55.55

PSA 7.26 5.72 1.40 33.10 31.70

Capra

Score

2.15 2.17 0 9.0 9.0

BMI 28.90 4.93 19.97 39.42 19.45

Gleason

Score

6.58 0.50 6 7 1

Tumor

Variable Mean Standard

Deviation

Minimum Maximum Range

Age 54.51 6.66 40 70 32

Height 176.81 6.10 161.50 188.00 26.50

Weight 90.75 14.41 62.46 119.67 57.21

PSA 7.13 5.63 1.30 33.10 31.80

Capra

Score

2.04 2.40 0 11.00 11.00

BMI 29.08 4.80 19.97 39.42 19.45

Gleason

Score

6.62 0.61 6 9 3

19

Table 2: Dichotomous Clinical and Demographic Characteristics for the 48 African

American Prostate Cancer Patients

Variable Frequency (%)

Biochemical Recurrence

Yes 10 (20.83)

No 38 (79.17)

Obesity

High BMI 40 (83.33)

Low BMI 8 (16.67)

Aggressive Gleason Score

Yes 29 (60.42)

No 19 (39.58)

Seminal Vesicle Invasion

Yes 2 (4.17)

No 46 (95.83)

Surgical Margins

Yes 12 (25)

No 36 (75)

Lymph Node Invasion

Yes 1 (2.08)

No 47 (97.92

Extracellular Extension

Yes 7 (14.58)

No 41 (85.42)

20

Table 3: Comparing the means for gene expression among Tumor Tissue, Prostatic

Intraepithelial Neoplasia (PIN) and Normal Tissue.

Gene Normal

Mean ± SD

PIN

Mean ± SD

Tumor

Mean ± SD

Pr > F

ADIPOR1 8.67 ± 0.33 8.84±0.31 8.85±0.35 0.00238

AKRIC4 0.96 ± 0.84 1.18±1.01 1.78±1.34 0.0021

ALOX12 1.91±1.12 2.27±1.06 2.51±1.16 0.0475

ALOX15 2.01±0.85 2.84±0.95 3.24±1.15 <.0001

ANGPT1 7.95±0.60 7.59±0.61 7.36±0.78 0.0004

CRYBB2 2.56±0.91 2.75±0.87 3.11±1.09 0.0282

EIF5A 10.67±0.20 10.81±0.31 10.90±0.30 0.0012

ERG 4.62±0.69 4.90±1.20 5.29±1.53 0.0366

FADS1 7.08±0.71 7.01±0.75 6.69±0.83 0.0488

GNPDA2 6.75±0.34 6.82±0.32 6.98±0.50 0.0255

HNF1B 5.27±0.66 5.64±0.59 5.74±0.78 0.0053

HSD3B1 10.70±0.56 1.46±1.21 2.15±1.57 0.0127

HSPB8 10.70±0.56 5.64±0.59 10.24±0.89 0.0115

IL6 7.46±1.73 6.99±1.60 6.57±1.56 0.0441

INSR 8.97±0.26 9.12±0.27 9.03±0.24 0.0464

KLK4 11.26±0.80 11.55±0.65 11.77±0.83 0.0101

LEP 0.96±0.80 1.48±0.93 1.86±1.09 0.0002

MC4R 5.05±1.48 1.39±1.24 1.75±1.29 0.0427

MTCH2 8.66±0.24 8.76±0.34 8.83±0.30 0.0206

PCSK1 1.53±0.97 1.89±0.99 2.30±1.29 0.0063

PIK3CB 6.57±0.39 6.72±0.32 6.79±0.47 0.0355

PSCA 7.19±1.73 7.69±2.34 6.54±1.69 0.0305

SLC2A2 0.85±0.93 1.04±1.04 1.64±1.40 0.0048

SRD5A2 8.27±0.51 7.99±0.61 7.92±0.71 0.0248

STAT1 9.57±0.33 9.77±0.41 9.85±0.47 0.0066

SULT1A1 6.55±0.62 6.70±0.61 6.95±0.70 0.0168

TFF1 4.48±1.04 4.92±1.35 4.15±1.01 0.0127

TIMP3 11.41±0.48 10.97±0.48 10.83±0.55 <.0001

YY1 8.72±0.22 8.85±0.28 8.86±0.23 0.0108

21

Table 4: Comparing the Means for Gene Expression among the Tissue types stratified by

Obesity (Obese and Non-obese)

Gene Normal

PIN Tumor

Obese by Non-obese Obese by Non-obese Obese by Non-obese

Pr > F Pr > F Pr > F

ADIPOR1 0.7443 0.2154 0.5617

AKRIC4 0.4752 0.2755 0.7141

ALOX12 0.1407 0.3353 0.9110

ALOX15 0.5497 0.6929 0.3508

ANGPT1 0.6079 0.8102 0.3083

CRYBB2 0.4120 0.9135 0.2009

EIF5A 0.1239 0.6994 0.0329*

ERG 0.4665 0.7902 0.6115

FADS1 0.8594 0.4347 0.7493

GNPDA2 0.9875 0.3207 0.1825

HNF1B 0.7697 0.8566 0.07710

HSD3B1 0.9710 0.5307 0.9136

HSPB8 0.7937 0.7953 0.6937

IL6 0.5406 0.5112 0.0242*

INSR 0.4926 0.5637 0.8049

KLK4 0.9703 0.9860 0.0527

LEP 0.7613 0.5702 0.3050

MC4R 0.5935 0.0362* 0.9865

MTCH2 0.2883 0.9962 0.0479*

PCSK1 0.7244 0.5192 0.8320

PIK3CB 0.4733 0.1993 0.3659

PSCA 0.5922 0.3288 0.4886

SLC2A2 0.9801 0.7198 0.6501

SRD5A2 0.5081 0.6497 0.8810

STAT1 0.3280 0.4771 0.0894

SULT1A1 0.0894 0.8403 0.0227*

TFF1 0.8745 0.2061 0.1437

TIMP3 0.8546 0.3463 0.0057*

YY1 0.3644 0.4651 0.6183

22

Table 5: Testing the association between obesity and the recurrence of prostate cancer

after initial surgery with the use of a multiple logistic regression

Analysis of Maximum Likelihood Estimates

Parameter DF Estimates Standard

Error

Wald

Chi-Square

Pr>ChiSq

Intercept 1 -1.9459 1.0690 3.3132 0.0687

BMI-Group 1 0.7091 1.1341 0.3910 0.5318

Odds Ratio Estimates

Effect Point Estimate 95% Wald

Confidence Limits

BMI-Group 2.032 0.220 18.765

23

Table 6: Simple Logistic Regression Testing the association between gene expression and

biochemical recurrence of prostate cancer

Gene Pr > ChiSq OR 95% Confidence Interval

ADIPOR1 0.0253 0.233 0.065 0.835

AKT1 <0.0001 12.633 3.669 43.498

ALOX12 0.0055 1.754 1.180 2.609

ALOXI5B 0.0227 1.504 1.059 2.137

APOLD1 <.0001 11.561 4.290 31.157

AR 0.0053 0.214 0.073 0.632

BMP2 0.0053 2.372 1.293 4.351

EIF2A 0.0001 0.138 0.049 0.383

EIF5A 0.0153 0.122 0.022 0.667

FADS1 <.0001 4.354 2.126 8.919

FOXO3 0.0188 5.806 1.339 25.178

GPER1 0.0046 2.438 1.316 4.518

HSPB8 0.0016 3.072 1.529 6.175

INS 0.0175 1.617 1.088 2.404

INSR 0.0109 9.271 1.670 51.460

ITGBS 0.0071 0.134 0.031 0.579

LEP 0.0434 1.532 1.013 2.317

LEGALS8 <.0001 0.032 0.007 0.152

MAPK1 <.0001 0.029 0.006 0.136

NFKB1 0.0010 0.214 0.086 0.534

NPC1 0.0372 0.145 0.024 0.892

NPYIR 0.0272 1.922 1.076 3.434

OR2LI3 0.0007 2.189 1.389 3.451

PAH 0.0008 1.809 1.278 2.559

RAB6A 0.0170 0.093 0.013 0.654

RHOA 0.0059 6.710 1.730 26.027

SHBG 0.0387 1.612 1.025 2.534

SIRT1 0.0034 0.101 0.022 0.467

SLC39A2 0.0244 1.380 1.043 1.827

STAT1 0.0183 0.258 0.084 0.795

STAT3 0.0002 15.674 3.758 65.365

TFF1 0.0141 1.605 1.100 2.342

TGFB1 0.0238 2.988 1.157 7.718

TP53 <.0001 0.077 0.023 0.252

YY1 0.0011 0.043 0.007 0.287

24

Table 7: Multiple Logistic Regression Testing the association between gene expression and

biochemical recurrence of prostate cancer

Gene Pr > ChiSq OR 95% Confidence Interval

ADIPOR1 0.0094 0.134 0.030 0.611

ADIPOR2 0.0450 0.294 0.089 0.973

AKT1 <.0001 30.802 6.500 145.963

ALOX12 0.0089 1.815 1.161 2.838

APOLD1 <.0001 12.807 4.597 35.685

AR 0.0136 0.260 0.089 0.758

BMP2 0.0035 2.908 1.422 5.947

CXCR4 0.0320 1.814 1.053 3.126

CYLD 0.0082 1.081 1.001 1.168

EIF2A 0.0001 0.132 0.046 0.375

EIF5A 0.0100 0.061 0.007 0.513

ESR2 0.0456 1.742 1.011 3.000

FADS1 0.0001 6.206 2.466 15.615

FOXO3 0.0167 7.479 1.439 38.881

GPER1 0.0047 2.918 1.388 6.132

HSPB8 0.0009 3.857 1.739 8.553

IGFBP3 0.0088 2.207 1.221 3.988

INS 0.0215 1.725 1.084 2.747

INSR 0.0119 14.381 1.800 114.915

ITGB5 0.0046 0.121 0.028 0.522

LGALS8 <.0001 0.036 0.008 0.172

MAPK1 <.0001 0.009 0.001 0.068

MMP9 0.0275 0.667 0.465 0.956

NFKB1 0.0009 0.185 0.068 0.504

NPC1 0.0116 0.062 0.007 0.538

NPYIR 0.0024 3.127 1.496 6.534

OR2LI3 0.0007 2.339 1.427 3.832

PAH 0.0007 1.858 1.300 2.654

PIK3C3 0.0303 5.950 1.185 29.864

PIK3CB 0.0256 0.264 0.082 0.850

PIK3RI 0.0469 3.340 1.017 10.973

PLA2G2A 0.0388 0.705 0.506 0.982

RAB6A 0.0320 0.116 0.016 0.830

RHOA 0.0015 17.569 2.995 103.049

SIRT1 0.0016 0.068 0.013 0.363

SLC39A2 0.0048 1.585 1.151 2.183

STAT1 0.0068 0.161 0.043 0.604

STAT3 <.0001 126.617 14.071 >999.999

TCF7L2 0.0312 9.453 1.225 72.931

25

Table 7: Multiple Logistic Regression Testing the association between gene expression and

biochemical recurrence of prostate cancer (continued)

TFF1 0.0170 1.685 1.098 2.587

TGFB1 0.0126 4.007 1.347 11.914

TP53 <.0001 0.061 0.017 0.220

YY1 0.0006 0.030 0.004 0.222

Table 8: An overlap of Genes

Table 9: P-value for a Kaplan Meier Curve depicting the role of obesity in prostate cancer

progression.

Test of Equality over Strata

Test Chi-Square DF Pr > Chi-Square

Log-Rank 1.0230 1 0.3118

Wilcoxon 0.7888 1 0.3745

-2Log(LR) 1.5797 1 0.2088

Gene

EIF5A

MC4R

MTCH2

SULT1A1

26

Figure 1: Kaplan Meier Curve depicting the role of obesity in prostate cancer progression.

27

Figure 2: Trend of Gene Expression in Normal, PIN and Tumor Tissue Sample

28

Figure 2: Trend of Gene Expression in Normal, PIN and Tumor Tissue Sample (continued)

29

Figure 2: Trend of Gene Expression in Normal, PIN and Tumor Tissue Sample (continued)

30

Figure 2: Trend of Gene Expression in Normal, PIN and Tumor Tissue Sample (continued)

31

Figure 2: Trend of Gene Expression in Normal, PIN and Tumor Tissue Sample (continued)

32

DISCUSSION

Our study examined the association between obesity and biochemical recurrence of

prostate cancer as well as the association between the RNA expression of obesity-related genes

and biochemical recurrence of prostate cancer among African American men. The analysis

between obesity and biochemical recurrence (Table 5) showed no association between obesity

and biochemical recurrence of prostate cancer in obese African American men (AAM) compared

to non-obese African American men. Reasons for this lack of association could be due to low

statistical power and small sample size in our study. The Kaplan Meier Curve (Figure 1) depicts

the survival time and time to event (biochemical recurrence) among obese and non-obese

African American prostate cancer patients. The probability of survival (not having BCR) is lower

in obese African American men compared to non-obese African American men. In other words,

the probability of having biochemical recurrence is higher in obese African American men

compared to non-obese African American men.

In terms of the association between gene expression and biochemical recurrence of

prostate cancer, we found forty-three obesity-related genes that were associated with biochemical

recurrence after adjusting for potential confounders (Table 7). The confounders adjusted for in

the multivariate models varied for each individual gene. The typical adjusted confounders were

age at diagnosis, height, prostate specific antigen (PSA), Gleason score, surgical margins,

CAPRAS score and extracapsular extension. The results (Table 7) suggests the importance of

these genes in the progression of prostate cancer. Our findings is consistent with the study by

(Wallace et al., 2008) which found that genes CXCR4 and MMP9 were highly expressed

33

metastasis-associated genes in African American men. Taichman et al., (2002) showed the

mechanism of prostate cancer invasion of bones through the SDF-1/CXCR4 pathway. Literature

supports the role of BMP2 in prostate cancer metastasis (Jin, Dayyani, & Gallick, 2011, Powell

et al., 2013). TP53 was responsible for prostate cancer recurrence in a study by (Ecke et al.,

2010), which is similar to our result.

The analysis comparing the means of gene expression among normal, prostate

intraepithelial neoplasia (PIN) and tumor tissues generated twenty-nine obesity-related genes that

were differentially expressed out of the ninety-nine obesity-related genes (Table 3). We further

stratified each tissue sample by obesity status. Out of the twenty-nine obesity-related genes,

EIF5A, IL6, MC4R, MTCH2, SULT1A1 and TIMP3 were expressed differently among normal,

PIN and tumor tissue samples (Table 4). IL6 which is called interleukin-6 causes inflammation

and prostate cancer metastasis as discussed by (Nguyen, Li, & Tewari, 2014).

A trend test comparing the means of gene expression among the three tissue samples

(Figure 2) generated twenty obesity-related genes that showed a positive (upward) trend of the

expression of genes and progression of the disease from normal tissue to prostate intraepithelial

neoplasia to tumor tissue. This shows a potential biological plausibility of the expression of the

genes in relation to the progression of the prostate cancer. Flajollet et al., (2011) showed how

ERG increases the expression of Osteopontin, a protein that is highly implicated in prostate

cancer metastasis. Our finding showed an expression of ERG in a manner that explains prostate

cancer progression.

We determined the presence of an overlap between EIF5A, IL6, MC4R, MTCH2,

SULT1A1, TIMP3 and the twenty obesity-related genes that showed a trend in gene expression.

We found that four genes: EIF5A, MC4R, MTCH2, SULT1A1 showed an overlap (Table 8).

34

EIF5A has been isolated as a cancer biomarker because of its overexpression and cancer

proliferation function (Acta, 2016). Studies have shown the association between SULT1A1 and

prostate cancer risk (Arslan, Silig, & Pinarbasi, 2011, Nowell et al., 2004).

Limitations

Sample size is a major limitation in our study. We had only 48 African American prostate

cancer patients. Hence, we were unable to determine the true association between obesity and

biochemical recurrence of prostate cancer due to a low statistical power. Larger sample size in

future analysis is required in order to give a better estimate of this association. Additionally, we

had nine obese patients with biochemical recurrence and only one non-obese patient with

biochemical failure rate in our study. This suggests evidence of biased results since the size of

the high risk subgroup (biochemical recurrence) is much larger for the exposed group (obese)

than for the unexposed group (non-obese). Sample size is also the main reason why we had a cut

point of 25 for BMI although overweight and obese individuals are likely to experience prostate

cancer remission after treatment (American Cancer Society, 2018)

In determining the presence of within subject correlation in our sample, we analyzed the

43 obesity-related genes that were significant in the multiple logistic model testing the

association between gene expression and biochemical recurrence. A spearman rho correlation

test was conducted on the 43 obesity-related genes with BMI as a continuous variable. Only gene

TCF7L2 showed a moderate negative correlation with BMI (-0.336, p-value = 0.0196). Hence, a

linear relationship exists between gene TCF7L2 and BMI.

35

CONCLUSION

In summary, we found no association between obesity and biochemical recurrence among

obese African American men compared to non-obese African American men. We also elucidated

obesity-related genes, which may play a role in biochemical recurrence of prostate cancer.

Further analysis is vital to determine the functional significance of the selected genes in prostate

carcinogenesis.

Although we found no association between obesity and biochemical recurrence in our

study, a study investigating the effect of obesity on biochemical control showed that obese

patients were more likely to have a more pronounced biochemical failure rates (Freedland et al.,

2004). This simply suggests that more research on this subject is needed to determine the true

impact of obesity on biochemical recurrence. Obesity is a chronic disease and it is a risk factor

for other chronic diseases including heart disease and cancer. Hence, it was imperative to

understand the role of obesity in prostate cancer risk in order to inform adequate control of

prostate cancer. According to the American Cancer Society, (2016), physical activity reduces

cancer recurrence and mortality in general. Obesity can impact the gene expression profiles in

African American prostate cancer patients and influences their prostate cancer biochemical

recurrence as shown in our study. Gene expression predicts the behavioral outcome of prostate

cancer in patients (Singh et al., 2002) . Studies have shown that gene expression explains the

progression and aggressiveness of prostate cancer (Chandran et al., 2007, Mortensen et al., 2015,

Yu et al., 2004) . With the use of the selected genes from our study, genotyping and other

molecular research methods can further explain the functional role of these obesity-related genes

36

in prostate cancer proliferation. Our research findings can be used to develop hypothesis for

individualized prostate cancer risk estimation in African American men. Our research findings

can also be used in the future to select the most appropriate treatment modalities and to estimate

the risk of prostate cancer progression.

37

REFERENCES

Acta, B. B. (2016). HHS Public Access, 1849(7), 836–844.

https://doi.org/10.1016/j.bbagrm.2015.05.002.The

American Cancer Society. (2016). Cancer Facts & Figures for African Americans 2016-2018,

12–14. Retrieved from https://www.cancer.org/content/dam/cancer-org/research/cancer-

facts-and-statistics/cancer-facts-and-figures-for-african-americans/cancer-facts-and-figures-

for-african-americans-2016-

2018.pdf%5Cnhttp://www.cancer.org/acs/groups/content/@editorial/documents

American Cancer Society. (2017). Prostate Cancer Risk Factors. Retrieved December 7, 2017,

from https://www.cancer.org/cancer/prostate-cancer/causes-risks-prevention/risk-

factors.html

American Cancer Society. (2018). What does the American Cancer Society recommend about

body weight? Retrieved January 18, 2018, from https://www.cancer.org/cancer/cancer-

causes/diet-physical-activity/body-weight-and-cancer-risk/acs-recommendations.html

Arslan, S., Silig, Y., & Pinarbasi, H. (2011). Sulfotransferase 1A1 Arg213His polymorphism and

prostate cancer risk. Experimental and Therapeutic Medicine, 2(6), 1159–1162.

https://doi.org/10.3892/etm.2011.334

Arun, B., Sanjeev K, S., Mohammad Aslam, K., Prajapati1, V. K., Seema, S., James E., C., &

Ajay P., S. (2017). HHS Public Access, 10(4), 4173–4183.

https://doi.org/10.1021/acsnano.5b07425.Molecular

Becker, G., & Newsom, E. (2003). Socioeconomic Status and Dissatisfaction with Health Care

among Chronically Ill African Americans. American Journal of Public Health, 93(5), 742–

748. https://doi.org/10.2105/AJPH.93.5.742

Bureau, U. C. (2017). Wealth, Asset Ownership, &amp; Debt of Households Detailed Tables:

2013. Retrieved from https://www.census.gov/data/tables/2013/demo/wealth/wealth-asset-

ownership.html

Buschemeyer, W. C., & Freedland, S. J. (2007). Obesity and prostate cancer: epidemiology and

clinical implications. European Urology, 52(2), 331–343.

https://doi.org/10.1016/j.eururo.2007.04.069

Byers, T. E., Wolf, H. J., Bauer, K. R., Bolick-Aldrich, S., Chen, V. W., Finch, J. L., … Yin, X.

(2008). The impact of socioeconomic status on survival after cancer in the United States:

Findings from the National Program of Cancer Registries patterns of care study. Cancer,

113(3), 582–591. https://doi.org/10.1002/cncr.23567

38

Cancer Information and Support Network. (2017). CISN - Oncogenes, Tumor-Suppressor Genes

and DNA Repair Genes. Retrieved December 7, 2017, from

http://cisncancer.org/research/what_we_know/advances/oncogenes.html

Castro, P., Creighton, C. J., Ozen, M., Berel, D., Mims, M. P., & Ittmann, M. (2009). Genomic

Profiling of Prostate Cancers from African American Men. Neoplasia, 11(3), 305–312.

https://doi.org/10.1593/neo.81530

Chandran, U. R., Ma, C., Dhir, R., Bisceglia, M., Lyons-Weiler, M., Liang, W., … Monzon, F.

A. (2007). Gene expression profiles of prostate cancer reveal involvement of multiple

molecular pathways in the metastatic process. BMC Cancer, 7, 1–21.

https://doi.org/10.1186/1471-2407-7-64

Chornokur, G., Amankwah, E. K., Davis, S. N., Phelan, C. M., Park, J. Y., Pow-Sang, J., &

Kumar, N. B. (2013). Variation in HNF1B and Obesity May Influence Prostate Cancer Risk

in African American Men: A Pilot Study. Prostate Cancer, 2013, 384594.

https://doi.org/10.1155/2013/384594

Donkena, K. V, & Young, C. Y. (2011). Vitamin d, sunlight and prostate cancer risk.

Adv.Prev.Med, 2011, 281863-. https://doi.org/10.4061/2011/281863

Ecke, T. H., Schlechte, H. H., Schiemenz, K., Sachs, M. D., Lenk, S. V., Rudolph, B. D., &

Loening, S. A. (2010). TP53 gene mutations in prostate cancer progression. Anticancer

Research, 30(5), 1579–1586.

Fiscella, K., & Williams, D. R. (2004). Health Disparities Based on SE Inequalities. Academic

Medicine.

Flajollet, S., Tian, T. V., Flourens, A., Tomavo, N., Villers, A., Bonnelye, E., … Duterque-

Coquillaud, M. (2011). Abnormal Expression of the ERG Transcription Factor in Prostate

Cancer Cells Activates Osteopontin. Molecular Cancer Research, 9(7), 914–924.

https://doi.org/10.1158/1541-7786.MCR-10-0537

Freedland, S. J., & Aronson, W. J. (2004). Examining the Relationship Between Obesity and

Prostate Cancer. Reviews in Urology, 6(2), 73–81. Retrieved from

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1550782/%5Cnhttps://www.ncbi.nlm.nih.go

v/pmc/articles/PMC1550782/pdf/RIU006002_0073.pdf

Freedland, S. J., Aronson, W. J., Kane, C. J., Presti, J. C., Amling, C. L., Elashoff, D., & Terris,

M. K. (2004). Impact of obesity on biochemical control after radical prostatectomy for

clinically localized prostate cancer: A report by the shared equal access regional cancer

hospital database study group. Journal of Clinical Oncology, 22(3), 446–453.

https://doi.org/10.1200/JCO.2004.04.181

Gann, P. H., Hennekens, C. H., Ma, J., Longcope, C., & Stampfer, M. J. (1996). Prospective

Study of Sex Hormone Levels and Risk of Prostate Cancer. JNCI Journal of the National

Cancer Institute, 88(16), 1118–1126. https://doi.org/10.1093/jnci/88.16.1118

GeneCards. (2017). GeneCards - Human Genes | Gene Database | Gene Search. Retrieved

November 29, 2017, from http://www.genecards.org/

39

Grisanzio, C., Werner, L., Takeda, D., Awoyemi, B. C., Pomerantz, M. M., Yamada, H., …

Freedman, M. L. (2012). Genetic and functional analyses implicate the NUDT11, HNF1B,

and SLC22A3 genes in prostate cancer pathogenesis. Proceedings of the National Academy

of Sciences, 109(28), 11252–11257. https://doi.org/10.1073/pnas.1200853109

Jianjun, Z., Dhakal, I. B., Greene, G., Lang, N. P., & Kadlubar, F. F. (2011). Polymorphisms in

hOGG1 and XRCC1 and Risk of Prostate Cancer: Effects Modified by Plasma

Antioxidants. Urology, 75(4), 779–785.

https://doi.org/10.1016/j.urology.2009.08.063.Polymorphisms

Jin, J. K., Dayyani, F., & Gallick, G. E. (2011). Steps in prostate cancer progression that lead to

bone metastasis. International Journal of Cancer, 128(11), 2545–2561.

https://doi.org/10.1002/ijc.26024

Kelesidis, I., Kelesidis, T., & Mantzoros, C. S. (2006). Adiponectin and cancer: a systematic

review. British Journal of Cancer, 94(9), 1221–1225.

https://doi.org/10.1038/sj.bjc.6603051

Lipson, E. J., Vincent, J. G., Loyo, M., Kagohara, L. T., Brandon, S., Wang, H., … Taube, J. M.

(2014). NIH Public Access, 73(1), 1–20. https://doi.org/10.1158/2326-6066.CIR-13-

0034.PD-L1

Michaud, J. E., Billups, K. L., & Partin, A. W. (2015). Testosterone and prostate cancer: An

evidence-based review of pathogenesis and oncologic risk. Therapeutic Advances in

Urology, 7(6), 378–387. https://doi.org/10.1177/1756287215597633

Mortensen, M. M., Høyer, S., Lynnerup, A. S., Ørntoft, T. F., Sørensen, K. D., Borre, M., &

Dyrskjøt, L. (2015). Expression profiling of prostate cancer tissue delineates genes

associated with recurrence after prostatectomy. Scientific Reports, 5(June), 1–11.

https://doi.org/10.1038/srep16018

Mullaney, J. M., Mills, R. E., Stephen Pittard, W., & Devine, S. E. (2010). Small insertions and

deletions (INDELs) in human genomes. Human Molecular Genetics, 19(R2), 131–136.

https://doi.org/10.1093/hmg/ddq400

Nambiar, M., & Raghavan, S. C. (2011). How does DNA break during chromosomal

translocations? Nucleic Acids Research, 39(14), 5813–5825.

https://doi.org/10.1093/nar/gkr223

Nanostring Technologies. (2017). Gene Expression Panels. Retrieved November 29, 2017, from

https://www.nanostring.com/products/gene-expression-panels

National Cancer Institute. (2017a). Cancer Genome Anatomy Project. Retrieved November 29,

2017, from https://cgap.nci.nih.gov/Pathways/BioCarta_Pathways

National Cancer Institute. (2017b). Cancer of the Prostate - Cancer Stat Facts. Retrieved

December 7, 2017, from https://seer.cancer.gov/statfacts/html/prost.html

National Cancer Institute. (2017c). Genetics of Prostate Cancer. Retrieved December 1, 2017,

from http://www.ncbi.nlm.nih.gov/pubmed/26389227

40

National Cancer Institute. (2018). Definition of gene expression - NCI Dictionary of Cancer

Terms. Retrieved January 7, 2018, from

https://www.cancer.gov/publications/dictionaries/cancer-terms?CdrID=537335

National Center for Biotechnology Information. (2017). Genes - NCBI. Retrieved November 29,

2017, from https://www.ncbi.nlm.nih.gov/home/genes/

National Institute of Health. (2017). What are single nucleotide polymorphisms (SNPs)? -

Genetics Home Reference. Retrieved December 7, 2017, from

https://ghr.nlm.nih.gov/primer/genomicresearch/snp

Nghiem, B., Zhang, X., Lam, H.-M., True, L. D., Coleman, I., Higano, C. S., … Morrissey, C.

(2016). Mismatch repair enzyme expression in primary and castrate resistant prostate

cancer. Asian Journal of Urology, 3(4), 223–228. https://doi.org/10.1016/j.ajur.2016.09.002

Nguyen, D. P., Li, J., & Tewari, A. K. (2014). Inflammation and prostate cancer: The role of

interleukin 6 (IL-6). BJU International, 113(6), 986–992. https://doi.org/10.1111/bju.12452

Nowell, S., Ratnasinghe, D. L., Ambrosone, C. B., Williams, S., Teague-ross, T., Trimble, L., …

Lang, N. P. (2004). Association of SULT1A1 Phenotype and Genotype with Prostate

Cancer Risk in African-Americans and Caucasians Association of SULT1A1 Phenotype

and Genotype with Prostate Cancer Risk in African-Americans and Caucasians,

13(February), 270–276. https://doi.org/10.1158/1055-9965.EPI-03-0047

Obeid, S., & Hebbard, L. (2012). Role of adiponectin and its receptors in cancer. Cancer Biology

& Medicine, 9(4), 213–20. https://doi.org/10.7497/j.issn.2095-3941.2012.04.001

Ornish, D., Magbanua, M. J. M., Weidner, G., Weinberg, V., Kemp, C., Green, C., … Carroll, P.

R. (2008). Changes in prostate gene expression in men undergoing an intensive nutrition

and lifestyle intervention. Proceedings of the National Academy of Sciences, 105(24),

8369–8374. https://doi.org/10.1073/pnas.0803080105

Pampel, F. C., Krueger, P., & Denney, J. (2010). Socioeconomic disparities in health behaviors.

Annual Review of Sociology, 36, 349–370.

https://doi.org/10.1146/annurev.soc.012809.102529.Socioeconomic

Park, J., Chen, L., Ratnashinge, L., Sellers, T. A., Tanner, J. P., Lee, J. H., … Pow-Sang, J.

(2006). Deletion polymorphism of UDP-glucuronosyltransferase 2B17 and risk of prostate

cancer in African American and Caucasian men. Cancer Epidemiology Biomarkers and

Prevention, 15(8), 1473–1478. https://doi.org/10.1158/1055-9965.EPI-06-0141

Powell, I. J., Dyson, G., Land, S., Ruterbusch, J., Bock, C. H., Lenk, S., … Bollig-Fischer, A.

(2013). Genes Associated with Prostate Cancer Are Differentially Expressed in African

American and European American Men, 5(22), 891–897. https://doi.org/10.1158/1055-

9965.EPI-12-1238

Pritchard, C. C., Mateo, J., Walsh, M. F., De Sarkar, N., Abida, W., Beltran, H., … Nelson, P. S.

(2016). Inherited DNA-Repair Gene Mutations in Men with Metastatic Prostate Cancer.

New England Journal of Medicine, 375(5), 443–453.

https://doi.org/10.1056/NEJMoa1603144

41

Punnen, S., Freedland, S. J., Presti, J. C., Aronson, W. J., Terris, M. K., Kane, C. J., …

Cooperberg, M. R. (2014). Multi-institutional validation of the CAPRA-S score to predict

disease recurrence and mortality after radical prostatectomy. European Urology, 65(6),

1171–1177. https://doi.org/10.1016/j.eururo.2013.03.058

Ribeiro, R., Monteiro, C., Catalan, V., Hu, P., Cunha, V., Rodriguez, A., … Fruhbeck, G. (2012).

Obesity and prostate cancer: gene expression signature of human periprostatic adipose

tissue. BMC.Med., 10(1741–7015 (Electronic)), 108. https://doi.org/10.1186/1741-7015-10-

108

Rider, J. R., Fiorentino, M., Kelly, R., Gerke, T., Jordahl, K., Sinnott, J. A., … Finn, S. (2015).

Tumor expression of adiponectin receptor 2 and lethal prostate cancer. Carcinogenesis,

36(6), 639–647. https://doi.org/10.1093/carcin/bgv048

RJ, M., & English, D. (2006). Body size and composition and prostate cancer risk: systematic

review and meta-regression analysis. Cancer Causes & Control, 17(December), 989–1003.

https://doi.org/10.1007/s10552-006-0049-z

Ross-Adams, H., Ball, S., Lawrenson, K., Halim, S., Russell, R., Wells, C., … Mills, I. G.

(2015). &lt;i&gt;HNF1B&lt;/i&gt; variants associate with promoter methylation and

regulate gene networks activated in prostate and ovarian cancer. Oncotarget, 7(46).

https://doi.org/10.18632/oncotarget.12543

Singh, D., Febbo, P. G., Ross, K., Jackson, D. G., Manola, J., Ladd, C., … Sellers, W. R. (2002).

Gene expression correlates of clinical prostate cancer behavior. Cancer Cell, 1(2), 203–209.

https://doi.org/10.1016/S1535-6108(02)00030-2

Stark, J. R., Perner, S., Stampfer, M. J., Sinnott, J. A., Finn, S., Eisenstein, A. S., … Mucci, L. A.

(2009). Gleason score and lethal prostate cancer: Does 3 + 4 = 4 + 3? Journal of Clinical

Oncology, 27(21), 3459–3464. https://doi.org/10.1200/JCO.2008.20.4669

Taichman, R. S., Cooper, C., Keller, E. T., Pienta, K. J., Taichman, N. S., & Mccauley, L. K.

(2002). Use of the Stromal Cell-derived Factor-1 / CXCR4 Pathway in Prostate Cancer

Metastasis to Bone Use of the Stromal Cell-derived Factor-1 / CXCR4 Pathway in Prostate

Cancer Metastasis to Bone 1, 1832–1837.

Taksler, G. B., Keating, N. L., & Cutler, D. M. (2012). Explaining racial differences in prostate

cancer mortality. Cancer, 118(17), 4280–4289. https://doi.org/10.1002/cncr.27379

Timofeeva, O. A., Zhang, X., Ressom, H. W., Varghese, R. S., Kallakury, B. V. S., Wang, K., …

Dritschilo, A. (2009). Enhanced expression of SOS1 is detected in prostate cancer epithelial

cells from African-American men. International Journal of Oncology, 35(4), 751–760.

https://doi.org/10.3892/ijo-00000388

Wallace, T. A., Prueitt, R. L., Yi, M., Howe, T. M., Gillespie, J. W., Yfantis, H. G., … Ambs, S.

(2008). Tumor immunobiological differences in prostate cancer between African-American

and European-American men. Cancer Research, 68(3), 927–936.

https://doi.org/10.1158/0008-5472.CAN-07-2608

42

Yu, Y. P., Landsittel, D., Jing, L., Nelson, J., Ren, B., Liu, L., … Luo, J. H. (2004). Gene

expression alterations in prostate cancer predicting tumor aggression and preceding

development of malignancy. Journal of Clinical Oncology, 22(14), 2790–2799.

https://doi.org/10.1200/JCO.2004.05.158

Zeigler-Johnson, C., Friebel, T., Walker, A. H., Wang, Y., Spangler, E., Panossian, S., …

Rebbeck, T. R. (2004). CYP3A4, CYP3A5, and CYP3A43 genotypes and haplotypes in the

etiology and severity of prostate cancer. Cancer Research, 64(22), 8461–7.

https://doi.org/10.1158/0008-5472.CAN-04-1651

Zeigler-Johnson, C. M., Spangler, E., Jalloh, M., Gueye, S. M., Rennert, H., & Rebbeck, T. R.

(2008). Genetic susceptibility to prostate cancer in men of African descent: implications for

global disparities in incidence and outcomes. The Canadian Journal of Urology, 15(1),

3872–3882. https://doi.org/10.1016/j.bbi.2008.05.010

43

APPENDICES

44

Appendix A: Descriptive Statistics for the Gene Expression of 99 Obesity-Related Genes

among the Three Tissue Samples

Type Variable Label Mean Std Dev Range Normal ACTN2 ACTN2 4.63 2.82 11.12 ADIPOQ ADIPOQ 1.36 1.05 4.13 ADIPOR1 ADIPOR1 8.67 0.33 1.37 ADIPOR2 ADIPOR2 6.51 0.49 2.89 AKR1C4 AKR1C4 0.96 0.84 2.87 ALOX12 ALOX12 1.91 1.12 4.76 ALOX15 ALOX15 2.01 0.85 3.24 ALOX15B ALOX15B 8.80 1.48 6.47 ANGPT1 ANGPT1 7.95 0.60 2.05 APOLD1 APOLD1 7.49 0.70 3.08 AR AR 8.67 0.40 1.87 BMP2 BMP2 5.09 0.74 3.20 CGA CGA 2.01 0.94 3.48 CRYBB2 CRYBB2 2.56 0.91 4.02 CSNK2A1 CSNK2A1 7.99 0.27 1.35 CXCR4 CXCR4 8.97 0.97 4.50 CYLD CYLD 7.06 0.33 1.48 CYP17A1 CYP17A1 2.36 1.09 4.25 CYP19A1 CYP19A1 2.01 0.98 3.76 CYP3A4 CYP3A4 1.85 0.99 3.64 EIF2A EIF2A 7.34 0.44 1.82 EIF5A EIF5A 10.67 0.20 0.86 EPHB2 EPHB2 5.14 0.64 2.50 ERG ERG 4.62 0.68 3.73 ESR1 ESR1 6.92 0.80 3.53 ESR2 ESR2 4.62 0.81 2.71 FADS1 FADS1 7.07 0.70 2.78 FASN FASN 10.13 0.79 3.26 FOXO3 FOXO3 8.98 0.27 1.21 FTO FTO 8.41 0.29 1.17 GNPDA2 GNPDA2 6.74 0.33 1.64 GOLIM4 GOLIM4 8.32 0.30 1.68 GPER1 GPER1 5.56 0.72 3.03 HLA_DRB3 HLA-DRB3 10.30 0.83 3.79 HNF1A HNF1A 1.96 1.01 4.03 HNF1B HNF1B 5.26 0.65 3.63 HSD3B1 HSD3B1 1.37 1.03 3.76 HSPB8 HSPB8 10.69 0.55 2.57 IGF1 IGF1 9.35 0.74 2.72 IGFBP3 IGFBP3 8.92 0.87 3.79 IL1B IL1B 5.27 1.175 5.01 IL6 IL6 7.45 1.72 7.79 INS INS 1.14 0.85 3.35 INSIG2 INSIG2 5.85 0.35 1.92 INSR INSR 8.97 0.25 1.08 ITGB5 ITGB5 8.47 0.34 1.28 KLK4 KLK4 11.26 0.79 3.46 LEP LEP 0.96 0.83 2.86 LEPR LEPR 7.02 0.43 1.85 LGALS8 LGALS8 7.61 0.31 1.38 MAPK1 MAPK1 8.49 0.40 1.71 MAPKAPK2 MAPKAPK2 7.08 0.47 2.93 MC4R MC4R 1.11 0.97 3.35 MMP9 MMP9 5.04 1.47 5.65 MTCH2 MTCH2 8.65 0.23 1.10 NFKB1 NFKB1 6.21 0.43 2.55 NPC1 NPC1 6.59 0.17 0.74

45

Appendix A: Descriptive Statistics for the Gene Expression of 99 Obesity-Related Genes

among the Three Tissue Samples (Continued)

NPY1R NPY1R 5.50 0.77 3.89 NPY5R NPY5R 2.50 0.85 3.97 OR2L13 OR2L13 2.18 1.06 4.18 PAH PAH 3.29 1.35 5.97 PCGEM1 PCGEM1 3.72 1.79 7.00 PCSK1 PCSK1 1.53 0.96 3.31 PI3 PI3 3.58 1.28 5.81 PIK3CB PIK3CB 6.56 0.39 1.67 PIK3CA PIK3CA 7.06 0.27 1.34 PIK3R1 PIK3R1 9.02 0.40 1.63 PLA2G2A PLA2G2A 10.28 1.33 6.19 POMC POMC 2.47 1.07 4.42 PRKAA2 PRKAA2 6.84 0.39 1.96 PSCA PSCA 7.19 1.73 6.87 PSPH PSPH 6.79 0.55 3.03 PTEN PTEN 8.95 0.35 1.50 RAB6A RAB6A 9.50 0.19 0.86 RHOA RHOA 9.59 0.37 1.52 SCD SCD 8.86 0.99 3.68 SHBG SHBG 2.97 0.75 2.69 SIRT1 SIRT1 7.08 0.29 1.36 SLC2A2 SLC2A2 0.84 0.92 2.66 SLC2A3 SLC2A3 2.54 0.71 3.41 SLC2A4 SLC2A4 7.01 0.72 2.98 SLC39A1 SLC39A1 8.38 0.27 1.15 SLC39A2 SLC39A2 3.06 1.29 5.44 SRD5A2 SRD5A2 8.26 0.50 2.62 STAT1 STAT1 9.57 0.32 1.28 STAT3 STAT3 10.87 0.29 1.33 SULT1A1 SULT1A1 6.55 0.61 3.19 TCF7L2 TCF7L2 8.59 0.27 1.31 TFF1 TFF1 4.48 1.03 5.16 TGFB1 TGFB1 8.60 0.48 2.20 TIMP3 TIMP3 11.41 0.48 2.08 TNF TNF 3.75 0.93 4.10 TP53 TP53 7.05 0.45 2.25 VDR VDR 3.91 0.67 3.19 YY1 YY1 8.71 0.22 1.01 Type Variable Label Mean Std Dev Range Pin ACTN2 ACTN2 3.57 2.08 8.00 ADIPOQ ADIPOQ 1.51 0.90 3.70 ADIPOR1 ADIPOR1 8.84 0.31 1.30 ADIPOR2 ADIPOR2 6.54 0.34 1.49 AKR1C4 AKR1C4 1.17 1.01 3.42 ALOX12 ALOX12 2.26 1.06 4.32 ALOX15 ALOX15 2.84 0.95 4.13 ALOX15B ALOX15B 8.81 1.25 5.01 ANGPT1 ANGPT1 7.58 0.61 2.40 APOLD1 APOLD1 7.34 0.77 3.07 AR AR 8.67 0.50 2.21 BMP2 BMP2 4.98 0.55 2.83 CGA CGA 2.40 1.14 4.59 CRYBB2 CRYBB2 2.75 0.87 3.24 CSNK2A1 CSNK2A1 8.12 0.29 1.32 CXCR4 CXCR4 8.74 0.80 3.62

46

Appendix A: Descriptive Statistics for the Gene Expression of 99 Obesity-Related Genes

among the Three Tissue Samples (Continued)

CYLD CYLD 7.02 0.43 2.16 CYP17A1 CYP17A1 2.51 1.34 5.22 CYP19A1 CYP19A1 2.19 0.97 4.64 CYP3A4 CYP3A4 2.09 1.27 4.18 EIF2A EIF2A 7.40 0.53 2.26 EIF5A EIF5A 10.80 0.30 1.09 EPHB2 EPHB2 5.12 0.50 2.23 ERG ERG 4.90 1.19 5.06 ESR1 ESR1 7.11 0.57 2.51 ESR2 ESR2 4.59 0.95 4.16 FADS1 FADS1 7.00 0.74 3.18 FASN FASN 10.45 0.80 3.29 FOXO3 FOXO3 8.95 0.30 1.26 FTO FTO 8.42 0.23 0.92 GNPDA2 GNPDA2 6.82 0.32 1.36 GOLIM4 GOLIM4 8.23 0.28 1.39 GPER1 GPER1 5.62 0.74 3.07 HLA_DRB3 HLA-DRB3 10.11 0.69 2.40 HNF1A HNF1A 2.45 1.03 4.15 HNF1B HNF1B 5.63 0.58 2.17 HSD3B1 HSD3B1 1.46 1.21 4.60 HSPB8 HSPB8 10.40 0.59 2.32 IGF1 IGF1 9.10 0.72 3.02 IGFBP3 IGFBP3 8.88 0.80 3.13 IL1B IL1B 5.26 1.36 6.21 IL6 IL6 6.98 1.60 5.86 INS INS 1.34 0.98 3.60 INSIG2 INSIG2 5.89 0.48 2.20 INSR INSR 9.12 0.26 1.19 ITGB5 ITGB5 8.54 0.29 1.28 KLK4 KLK4 11.55 0.65 3.01 LEP LEP 1.47 0.92 3.50 LEPR LEPR 6.99 0.41 1.65 LGALS8 LGALS8 7.64 0.34 1.50 MAPK1 MAPK1 8.52 0.43 1.79 MAPKAPK2 MAPKAPK2 7.04 0.41 1.89 MC4R MC4R 1.38 1.24 4.00 MMP9 MMP9 5.36 1.47 5.59 MTCH2 MTCH2 8.76 0.34 1.65 NFKB1 NFKB1 6.09 0.51 2.04 NPC1 NPC1 6.55 0.21 0.96 NPY1R NPY1R 5.54 0.71 2.96 NPY5R NPY5R 2.77 0.77 3.58 OR2L13 OR2L13 2.45 0.98 3.85 PAH PAH 3.83 1.42 5.99 PCGEM1 PCGEM1 3.94 1.84 6.75 PCSK1 PCSK1 1.89 0.99 4.13 PI3 PI3 3.70 1.32 5.45 PIK3CB PIK3CB 6.71 0.31 1.34 PIK3CA PIK3CA 7.01 0.26 0.91 PIK3R1 PIK3R1 8.88 0.36 1.42 PLA2G2A PLA2G2A 11.05 1.31 5.38 POMC POMC 2.50 0.93 3.95 PRKAA2 PRKAA2 6.93 0.35 1.32 PSCA PSCA 7.68 2.33 9.62 PSPH PSPH 6.84 0.53 2.59 PTEN PTEN 8.95 0.33 1.79 RAB6A RAB6A 9.55 0.26 1.17 RHOA RHOA 9.49 0.32 1.39 SCD SCD 9.13 0.92 3.65 SHBG SHBG 2.79 0.80 2.92

47

Appendix A: Descriptive Statistics for the Gene Expression of 99 Obesity-Related Genes

among the Three Tissue Samples (Continued)

SIRT1 SIRT1 7.09 0.31 1.37 SLC2A2 SLC2A2 1.04 1.03 3.60 SLC2A3 SLC2A3 2.60 0.72 3.24 SLC2A4 SLC2A4 6.86 0.62 2.94 SLC39A1 SLC39A1 8.43 0.24 1.36 SLC39A2 SLC39A2 3.51 1.86 6.44 SRD5A2 SRD5A2 7.99 0.61 2.75 STAT1 STAT1 9.76 0.40 1.50 STAT3 STAT3 10.93 0.34 1.46 SULT1A1 SULT1A1 6.69 0.61 2.76 TCF7L2 TCF7L2 8.67 0.22 1.09 TFF1 TFF1 4.92 1.35 5.23 TGFB1 TGFB1 8.47 0.40 1.61 TIMP3 TIMP3 10.97 0.47 2.01 TNF TNF 3.54 0.88 4.13 TP53 TP53 7.17 0.56 2.38 VDR VDR 3.90 0.71 2.86 YY1 YY1 8.85 0.27 1.18 Type Variable Label Mean Std Dev Range Tumor ACTN2 ACTN2 4.45 2.66 10.65 ADIPOQ ADIPOQ 1.82 1.43 4.83 ADIPOR1 ADIPOR1 8.85 0.35 1.49 ADIPOR2 ADIPOR2 6.54 0.37 2.01 AKR1C4 AKR1C4 1.77 1.34 4.40 ALOX12 ALOX12 2.51 1.15 5.25 ALOX15 ALOX15 3.23 1.14 4.78 ALOX15B ALOX15B 8.92 1.61 7.02 ANGPT1 ANGPT1 7.36 0.77 3.34 APOLD1 APOLD1 7.38 0.66 2.76 AR AR 8.75 0.42 1.92 BMP2 BMP2 4.69 0.95 4.81 CGA CGA 2.41 1.36 5.44 CRYBB2 CRYBB2 3.11 1.09 5.27 CSNK2A1 CSNK2A1 8.12 0.26 1.12 CXCR4 CXCR4 8.99 0.88 4.46 CYLD CYLD 7.06 0.43 2.34 CYP17A1 CYP17A1 3.01 1.42 5.38 CYP19A1 CYP19A1 2.21 1.20 4.88 CYP3A4 CYP3A4 2.30 1.09 4.72 EIF2A EIF2A 7.43 0.52 2.72 EIF5A EIF5A 10.89 0.30 1.18 EPHB2 EPHB2 5.42 0.72 2.98 ERG ERG 5.29 1.53 5.90 ESR1 ESR1 7.03 0.66 3.46 ESR2 ESR2 4.83 0.95 3.61 FADS1 FADS1 6.69 0.83 3.09 FASN FASN 10.57 0.90 4.10 FOXO3 FOXO3 8.92 0.33 1.35 FTO FTO 8.43 0.31 1.47 GNPDA2 GNPDA2 6.98 0.50 2.66 GOLIM4 GOLIM4 8.35 0.32 1.19 GPER1 GPER1 5.47 0.79 2.95 HLA_DRB3 HLA-DRB3 10.48 0.81 3.31

48

Appendix A: Descriptive Statistics for the Gene Expression of 99 Obesity-Related Genes

among the Three Tissue Samples (Continued)

HNF1A HNF1A 2.47 1.16 5.08 HNF1B HNF1B 5.74 0.78 4.34 HSD3B1 HSD3B1 2.15 1.57 6.33 HSPB8 HSPB8 10.24 0.89 4.37 IGF1 IGF1 9.05 0.82 3.59 IGFBP3 IGFBP3 8.65 0.96 4.24 IL1B IL1B 5.15 1.16 5.56 IL6 IL6 6.57 1.56 6.47 INS INS 1.39 1.20 4.40 INSIG2 INSIG2 5.81 0.43 1.86 INSR INSR 9.03 0.24 1.33 ITGB5 ITGB5 8.51 0.33 1.58 KLK4 KLK4 11.77 0.83 3.67 LEP LEP 1.86 1.09 4.20 LEPR LEPR 6.97 0.65 2.67 LGALS8 LGALS8 7.77 0.39 1.73 MAPK1 MAPK1 8.59 0.35 1.61 MAPKAPK2 MAPKAPK2 7.07 0.38 2.40 MC4R MC4R 1.75 1.28 5.40 MMP9 MMP9 5.59 1.49 6.59 MTCH2 MTCH2 8.83 0.29 1.52 NFKB1 NFKB1 6.09 0.65 3.50 NPC1 NPC1 6.55 0.31 1.18 NPY1R NPY1R 5.39 0.88 3.23 NPY5R NPY5R 2.92 1.18 4.90 OR2L13 OR2L13 2.37 1.26 5.08 PAH PAH 3.83 1.41 5.46 PCGEM1 PCGEM1 4.53 1.73 7.52 PCSK1 PCSK1 2.30 1.28 5.08 PI3 PI3 3.90 1.22 4.65 PIK3CB PIK3CB 6.79 0.47 1.86 PIK3CA PIK3CA 7.10 0.28 1.19 PIK3R1 PIK3R1 8.84 0.50 2.14 PLA2G2A PLA2G2A 10.73 1.61 7.18 POMC POMC 2.64 1.18 4.81 PRKAA2 PRKAA2 7.05 0.40 1.77 PSCA PSCA 6.53 1.68 7.61 PSPH PSPH 6.93 0.66 3.68 PTEN PTEN 8.98 0.29 1.44 RAB6A RAB6A 9.54 0.24 0.90 RHOA RHOA 9.48 0.32 1.29 SCD SCD 8.99 0.96 4.72 SHBG SHBG 3.08 1.18 5.01 SIRT1 SIRT1 7.11 0.32 1.69 SLC2A2 SLC2A2 1.64 1.39 5.08 SLC2A3 SLC2A3 2.72 1.11 4.51 SLC2A4 SLC2A4 6.76 0.79 3.47 SLC39A1 SLC39A1 8.44 0.32 1.78 SLC39A2 SLC39A2 2.94 1.33 5.63 SRD5A2 SRD5A2 7.91 0.71 2.90 STAT1 STAT1 9.85 0.47 1.65 STAT3 STAT3 10.91 0.38 1.60 SULT1A1 SULT1A1 6.95 0.70 3.46 TCF7L2 TCF7L2 8.66 0.27 1.51 TFF1 TFF1 4.15 1.01 4.43 TGFB1 TGFB1 8.50 0.53 1.93 TIMP3 TIMP3 10.83 0.55 2.33 TNF TNF 3.94 1.34 6.44

49

Appendix A: Descriptive Statistics for the Gene Expression of 99 Obesity-Related Genes

among the Three Tissue Samples (Continued)

TP53 TP53 7.31 0.47 2.13 VDR VDR 4.04 0.67 3.28 YY1 YY1 8.86 0.23 0.86