American Journal of Epidemiology
© The Author 2013. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of
Public Health. All rights reserved. For permissions, please e-mail: [email protected].
Vol. 178, No. 8
DOI: 10.1093/aje/kwt136
Advance Access publication:
August 28, 2013
Original Contribution
Blood Levels of Saturated and Monounsaturated Fatty Acids as Markers of
De Novo Lipogenesis and Risk of Prostate Cancer
Jorge E. Chavarro*, Stacey A. Kenfield, Meir J. Stampfer, Massimo Loda, Hannia Campos,
Howard D. Sesso, and Jing Ma
* Correspondence to Dr. Jorge E. Chavarro, Department of Nutrition, Harvard School of Public Health, 665 Huntington Ave., Boston, MA 02115
(e-mail: [email protected]).
Initially submitted December 3, 2012; accepted for publication May 10, 2013.
De novo lipogenesis has been implicated in prostate carcinogenesis, and blood levels of specific saturated fatty
acids (SFAs) and monounsaturated fatty acids (MUFAs) could reflect activity of this pathway. We used gas chromatography to measure blood SFA and MUFA levels in prediagnostic samples from 476 incident prostate cancer cases
(1982–1995) in the Physicians’ Health Study and an equal number of controls matched on age and smoking status.
Five tagging polymorphisms in the fatty acid synthase (FASN) gene (rs1127678, rs6502051, rs4246444, rs12949488,
and rs8066956) were related to blood SFA and MUFA levels. Conditional logistic regression was used to estimate the
rate ratios, with 95% confidence intervals, of prostate cancer across quintiles of blood fatty acid levels. The polymorphisms rs6502051 and rs4246444 were associated with lower levels of 14:1n-5, 16:1n-7, and 18:1n-9. Blood levels of
16:1n-7 were associated with higher prostate cancer incidence, with rate ratios for men in increasing quintiles of 1.00,
1.40, 1.35, 1.44, and 1.97 (95% confidence interval: 1.27–3.06; Ptrend = 0.003). Furthermore, 16:1n-7 levels were positively related to incidence of high-grade (Gleason score ≥7) tumors (rate ratioQ5–Q1 = 3.92; 95% confidence interval:
1.72–8.94) but not low-grade tumors (rate ratioQ5–Q1 = 1.51; 95% confidence interval: 0.87–2.62) (Pheterogeneity = 0.02).
Higher activity of enzymes involved in de novo lipogenesis, as reflected in blood levels of 16:1n-7, could be involved in
the development of high-grade prostate cancer.
biomarkers; epidemiology; fatty acids; nutrition; prostate cancer
Abbreviations: CI, confidence interval; FASN, fatty acid synthase; MUFA, monounsaturated fatty acids; SFA, saturated fatty acids;
SNP, single-nucleotide polymorphism.
Prostate cancer is the most commonly diagnosed malignancy and the second highest contributor to cancer deaths in
men in the United States (1). There are few well-established
risk factors for prostate cancer other than age, family history,
and African ancestry. Therefore, the identification of risk factors for prostate cancer, particularly for clinically relevant disease, is important.
Blood fatty acid levels can serve as biomarkers of diet and
of metabolic processes that could be relevant in prostate carcinogenesis. For example, blood levels of many polyunsaturated
fatty acids and of trans fatty acids, which cannot be endogenously synthesized by humans, serve as biomarkers of intake
(2, 3) and have been associated with prostate cancer risk (4, 5).
On the other hand, saturated fatty acids (SFAs) and monounsaturated fatty acids (MUFAs) with an even-numbered carbon
chain length can be synthesized de novo in humans, and therefore circulating levels do not necessarily represent diet. Similar
to polyunsaturated and trans fatty acids, short-term feeding of
SFA-rich foods transiently increases blood levels of these fatty
acids (6–8). However, unlike polyunsaturated fatty acids and
trans fatty acids, long-term intake of SFAs and MUFAs does
not correlate well with their tissue levels (3). Thus, SFA and
MUFA levels in blood among free-living individuals are probably better understood as markers of de novo lipogenesis and
of the relative activity of the different enzymes involved in this
metabolic process (Figure 1).
1246
Am J Epidemiol. 2013;178(8):1246–1255
Blood SFA and MUFA Levels and Prostate Cancer 1247
was more than 99% complete for morbidity and 100% complete for mortality.
Acetyl CoA
ACC
Malonyl CoA
Selection of cases and controls
FASN
ELOVL6
14:0
Saturated fatty acids
SCD
Monounsaturated
fatty acids
14:1 n-5
ELOVL6
16:0
SCD
16:1 n-7
18:0
SCD
18:1 n-9
Figure 1. Main biosynthetic pathway of saturated and monounsaturated fatty acids. The main biosynthetic product of fatty acid synthase
(FASN) is 16:0, but saturated fatty acids of 12 and 14 carbon atoms
are also produced. ACC indicates acetyl-CoA carboxylase; ELOVL6,
fatty acid elongase; and SCD, stearoyl-CoA desaturase.
Increasing evidence suggests a role of de novo synthesis
of fatty acids in the development of prostate cancer, particularly of clinically relevant tumors. For example, fatty acid
synthase (FASN) is overexpressed in prostate cancer (9), and
its expression level is positively correlated to tumor grade
(9, 10). In addition, we have found that single-nucleotide
polymorphisms (SNPs) in FASN are related to lower prostate
cancer risk and lower disease-specific mortality rate (11). To
gain further insights into the role of fatty acid metabolism in
prostate cancer, we conducted a prospective study to examine
whether common polymorphisms in FASN were associated
with blood levels of several SFAs and MUFAs and whether
these fatty acids were related to prostate cancer risk in men
followed up for more than 20 years.
MATERIALS AND METHODS
Study population
This study is based on the Physician’s Health Study (12, 13),
a randomized trial of aspirin and β-carotene in the prevention
of heart disease and cancer among 22,071 male physicians
who were 40–84 years of age in 1982. Men were excluded
from the study if they had a history of myocardial infarction,
stroke, transient ischemic attack, unstable angina, cancer (except
nonmelanoma skin cancer), renal or liver disease, peptic ulcer,
or gout; had a contraindication to the use of aspirin; or were
users of aspirin, platelet-active medications, or vitamin A supplements. The aspirin component of the trial was terminated
early in 1988 because of the benefits of aspirin on myocardial
infarction (12). The β-carotene component of the trial was terminated as scheduled in 1995 (13). Written, informed consent
was obtained from each participant, and the study was approved
by the Human Research Committee at Brigham and Women’s
Hospital, Boston, Massachusetts.
Prerandomization blood specimens were obtained from
14,916 participants (68%), processed after overnight delivery, and stored at −82°C (14). The present report is restricted
to men diagnosed with prostate cancer after having provided
the baseline sample and their matched controls. Follow-up
Am J Epidemiol. 2013;178(8):1246–1255
Whenever a participant reported a diagnosis of prostate cancer, we requested hospital records and pathology reports for
review by study physicians from the Physicians’ Health Study
Endpoints Committee to confirm the diagnosis of prostate
cancer and determine the tumor stage and grade at diagnosis.
Histological grade was recorded according to the Gleason
scoring system. Tumor stage was recorded according to the
tumor-node-metastasis (TNM) staging system or was converted from the modified Whitmore-Jewett classification
scheme (for prostate cancer cases diagnosed during the early
years of Physicians’ Health Study follow-up). Using the risk
set sampling method, we selected a control subject for each
confirmed case among the men in the entire risk set who had
provided a blood sample and did not have a partial or total
prostatectomy or prostate cancer at the time of the case’s
diagnosis. Controls were individually matched to cases by
baseline age (±1 year for men ≤55 years of age and ±5 years
for men >55 years of age) and smoking status (current, former, or never). Of the 758 cases accrued through 1995 in the
entire study, 505 had provided a baseline blood sample that
could be used for the determination of fatty acid levels.
Cases and controls whose blood samples had been received
≥6 days after they were drawn were excluded from analyses,
leaving 476 cases and their matched controls.
Laboratory analyses
Blinded samples from cases and their matched controls
were processed and analyzed together to reduce any effect of
interassay variability. Fatty acids were extracted from whole
blood into isopropanol and hexane containing 50 mg of 2.6di-tert-butyl-p-cresol as an antioxidant. Fatty acids were transmethylated with methanol and sulfuric acid, as previously
described (2, 15, 16). After esterification, the samples were
evaporated, and the fatty acids were redissolved in iso-octane
and quantified by gas-liquid chromatography on a fused silica
capillary cis/trans column (SP2560, Supelco, Belafonte, Pennsylvania). Peak retention times were identified by injecting known
standards (Nu-Chek Prep, Inc., Elysian, Minnesota) and were
analyzed with ChemStation A.08.03 software (Agilent Technologies, Lexington, Massachusetts). The fatty acid levels in
each sample were expressed as the percentage of total fatty
acids. Coefficients of variation for all fatty acid peaks were
measured by analyzing blinded quality control samples randomly distributed throughout the study samples. The coefficients of variation were 9% for myristic acid (14:0), 0.9% for
palmitic acid (16:0), 1.3% for stearic acid (18:0), 15.7% for
myristoleic acid (14:1n-5), 2.9% for palmitoleic acid (16:1n7), and 0.3% for oleic acid (18:1n-9). Although we could not
directly assess whether long-term storage and freezing affected
the fatty acid measurements, n-3 and n-6 polyunsaturated fatty
acids have moderately high reliability coefficients (0.66 and
0.53, respectively) and minimal oxidation in serum samples
stored for up to 12 years at −80°C (17).
1248 Chavarro et al.
FASN SNPs and genotyping
RESULTS
Using the HapMap database (National Center for Biotechnology Information Build 35) and the Web-based Tagger
application (http://broad.harvard.edu/mpg/tagger/), we identified 5 SNPs that captured genetic variation (with R 2 >
0.80) within FASN and 5 kb upstream and downstream. SNPs
with a minor allele frequency less than 5% in the HapMap
CEU population (Utah residents with ancestry from northern
and western Europe) were excluded. DNA was extracted
from whole blood. Genotyping was performed with iPLEX
(Sequenom, Inc., San Diego, California) matrix-assisted laser
desorption/ionization–time-of-flight mass spectrometry technology at the Partners HealthCare Center for Personalized
Genetic Medicine, Boston, Massachusetts. All SNPs had genotype completion rates greater than 91%, which did not differ
between cases and controls.
At baseline, men subsequently diagnosed with prostate
cancer had higher blood levels of total MUFAs than controls
but had no other significant differences in blood levels of
other major fatty acid groups or other baseline characteristics
(Table 1). The median time between enrollment and prostate
cancer diagnosis was 9 years, ranging from 1 month to 13
years. Ninety-five percent of the cases (n = 451) were diagnosed at least 2 years after enrollment, and 85% (n = 404)
were diagnosed at least 5 years after enrollment. Most prostate cancer cases presented as localized and low-grade disease.
Two thirds of the cases were diagnosed after the widespread
use of prostate-specific antigen screening became routine.
Blood fatty acid levels were, in general, positively correlated with each other (Table 2). Correlations were strongest
for those most closely related within the de novo lipogenesis
pathway (14:0 and 14:1n-5; 16:0 and 16:1n-7). Stearic acid
(18:0) was the exception to this pattern, being inversely related
to other fatty acids.
Of the 5 FASN SNPs evaluated, 3 were associated with blood
levels of at least 1 SFA or MUFA. Men who were variant allele
homozygotes in rs6502051 (G>T) and rs4246444 (C>A) had
lower blood levels of myristoleic (14:1n-5), palmitoleic (16:
1n-7), and oleic (18:1n-9) acids than did the wild-type homozygotes. In addition, variant allele homozygotes in rs12949488
had significantly higher blood levels of myristic acid (14:0) than
did wild-type homozygotes (Table 3).
Blood levels of SFAs were not significantly related to prostate cancer incidence (Table 4). However, incidence of prostate cancer increased with increasing blood levels of MUFAs
(Table 5). Blood levels of myristoleic (14:1n-5), palmitoleic
(16:1n-7), and oleic (18:1n-9) acids were associated with higher
incidence of prostate cancer, and the association was strongest
for palmitoleic acid. Men in the highest quintile of blood palmitoleic acid levels were twice as likely to develop prostate
cancer as men in the bottom quintile of blood levels for this
fatty acid.
Because some men were diagnosed with prostate cancer
shortly after enrollment, we evaluated whether exclusion of
cases identified during the first 2 or 5 years of follow-up
affected the results (Web Table 1, available at http://aje.
oxfordjournals.org/). In addition, we examined whether the
associations between fatty acids and prostate cancer differed
significantly between cases diagnosed during the first and
the second halves of the follow-up period (Web Table 2).
The rate ratios comparing top with bottom quintiles of 16:1n7 were 1.87 (95% confidence interval (CI): 1.19, 2.92), 2.04
(95% CI: 1.26, 3.29), and 2.57 (95% CI: 1.34, 4.95) after
exclusion of the first 2, 5, and 9 years of follow-up, respectively. There was no evidence of heterogeneity according to
follow-up time (Pheterogeneity = 0.63).
We also assessed the ratio of palmitoleic to palmitic acid
(16:1n-7/16:0) as an index of stearoyl CoA desaturase activity and observed rate ratios of prostate cancer in increasing
quintiles of estimated stearoyl CoA desaturase activity of
1.00, 1.42 (95% CI: 0.91, 2.21), 1.34 (95% CI: 0.85, 2.12),
1.51 (95% CI: 0.97, 2.35), and 1.96 (95% CI: 1.26, 3.06)
(Ptrend = 0.004). This association was nearly identical to that
observed for palmitoleic acid. We also examined the ratio of
Statistical analyses
We calculated median values and proportions of the baseline characteristics of case and control subjects. To evaluate
whether these characteristics differed between cases and controls, categorical variables were tested with the McNemar test,
and continuous variables were tested with the Wilcoxon signrank test. We examined the relation between tagging SNPs in
FASN and blood levels of 14:0, 16:0, 18:0, 14:1n-5, 16:1n-7,
and 18:1n-9 by using linear regression models adjusted for
baseline age and smoking status—the matching variables for
selecting case-control pairs.
To estimate the association between blood levels of SFAs
and MUFAs and prostate cancer, we first divided cases and
controls into 5 groups according to quintiles of fatty acid levels
among the controls. We then used conditional logistic regression to estimate the rate ratio of prostate cancer (with 95%
confidence interval) in a given quintile of fatty acid level in
relation to the lowest quintile. We considered the potential confounding effects of baseline characteristics by adding to the
initial model terms for variables associated with prostate cancer
and fatty acid levels at P < 0.20 and by evaluating whether
the addition of these variables changed the initial fatty acid
estimates by >10%. Addition of the variables meeting these
criteria (height, body mass index, and blood fatty acids previously related to prostate cancer in this cohort (4, 5)) did not
change the initial fatty acid estimates substantially, and therefore further adjustment for these variables was not performed.
We fitted regression models in subgroups defined by tumor
stage and grade at diagnosis and estimated the significance
of differences in stratum-specific estimates by using polytomous logistic regression. Last, because we had found previously
that body mass index (weight (kg)/height2 (m2)) modified the
relation between FASN polymorphisms and prostate cancer
risk and death (11), we evaluated body mass index as an
effect modifier of the relation between fatty acids and prostate cancer risk. Tests for linear trend were conducted in all
models, with the median fatty acid levels in each quintile
used as a continuous variable. All statistical analyses were
performed in SAS version 9.1 (SAS Institute Inc., Cary,
North Carolina). Results were considered to be statistically
significant when P < 0.05 (2-tailed).
Am J Epidemiol. 2013;178(8):1246–1255
Blood SFA and MUFA Levels and Prostate Cancer 1249
Table 1. Clinical Characteristics of Prostate Cancer Cases and Control Subjects, Physicians’ Health Study, United States, 1982–1995
Cases (n = 476)
Median
Age at baseline, yearsb
Length of follow-up, years
Age at diagnosis, years
25th–75th Percentiles
58
53–64
9
7–11
67
62–72
Controls (n = 476)
%
Median
58
25th–75th Percentiles
%
P Valuea
53–63
Tumor stage at diagnosis (TNM)
T1/T2
82
T3
8
T4/N1/M1
7
Undetermined
4
Tumor grade at diagnosis
Gleason score <7
63
Gleason score = 7
25
Gleason score ≥8
11
Undetermined
2
PSA at diagnosis, ng/mL
<4
7
4–9.9
28
10–19.9
17
≥20
12
Missing
36
Date of diagnosis
Before October 1, 1990
33
On or after October 1, 1990
67
Smoking statusb
Current
8
8
Former
42
42
<8 hours
76
75
≥8 hours
20
21
Unknown
4
4
95
93
Time since last meal at blood draw
0.62
White/Caucasian
Height, m
Body mass index, kg/m2
1.78
24.4
1.75–1.83
1.78
23.1–25.8
24.2
0.51
1.73–1.83
0.12
22.8–25.8
0.13
Regular multivitamin use
21
24
0.47
Vigorous exercise twice per week or more
58
55
0.41
Alcohol use once per day or more
32
30
0.47
Blood fatty acids (of total fatty acids)
Total saturated fatty acids
32.2
31.1–33.2
32.2
31.1–33.2
Total monounsaturated fatty acids
20.7
19.2–22.3
20.2
18.7–21.9
0.66
0.01
Total n-6 polyunsaturated fatty acids
38.5
36.2–40.4
38.9
36.6–40.5
0.08
Total n-3 polyunsaturated fatty acids
5.2
4.6–5.9
5.3
4.7–6.1
0.06
Total trans fatty acids
1.8
1.6–2.3
1.8
1.5–2.2
0.26
Abbreviations: PSA, prostate-specific antigen; TNM, tumor-node-metastasis staging system.
a
P values were computed with the Wilcoxon sign-rank test for continuous variables and the McNemar’s test for categorical variables.
b
Cases and controls were individually matched on these variables.
Am J Epidemiol. 2013;178(8):1246–1255
16.6, 17.1
16.5, 17.2
16.4, 17.8
16.9
16.8
17.1
0.95, 1.07
1.01, 1.15
0.92, 1.24
1.01
1.08
1.08
1.81, 2.31
2.17, 2.77
1.94, 3.27
2.06
2.47*
2.61
9.90, 10.1
9.86, 10.1
9.69, 10.2
GG
GA
AA
rs8066956
Abbreviations: CI, confidence interval; SNP, single-nucleotide polymorphism.
* P < 0.05 (compared with wild-type homozygous).
a
Adjusted for baseline age and smoking status.
b
Values of 14:1n-5 are expressed as original value ×102.
9.99
9.97
9.93
19.3, 19.7
19.6, 20.0
19.1, 20.1
19.5
19.8*
19.6
0.45, 0.52
0.50, 0.58
0.46, 0.62
0.49
0.54*
0.54
GG
GA
AA
rs12949488
5.9
35.8
58.3
16.6, 17.1
16.4, 17.1
16.6, 18.0
16.9
16.8
17.3
0.96, 1.07
1.02, 1.16
0.82, 1.14
1.02
1.09
0.98
1.89, 2.38
2.06, 2.68
1.81, 3.16
2.13
2.37
2.48
9.89, 10.1
9.90, 10.1
9.74, 10.2
9.97
10.0
9.98
19.4, 19.8
19.5, 20.0
19.2, 20.3
19.6
19.7
19.7
0.46, 0.52
0.48, 0.56
0.51, 0.67
0.49
0.52
0.59*
CC
CA
AA
rs4246444
5.7
30.7
63.6
16.7, 17.3
16.5, 17.0
15.9, 17.0
17.0
16.7
16.4*
1.02, 1.14
0.95, 1.08
0.82, 1.06
1.08
1.02
0.94*
2.16, 2.68
1.77, 2.34
1.31, 2.34
2.42
2.06*
1.83*
9.91, 10.1
9.86, 10.1
9.80, 10.2
10.0
9.96
9.98
19.5, 19.9
19.3, 19.8
19.2, 20.0
19.7
19.6
19.6
0.50, 0.56
0.44, 0.52
0.41, 0.54
0.53
0.48*
0.48
GG
GT
TT
9.7
39.6
50.7
16.7, 17.4
16.6, 17.1
16.1, 16.9
17.1
16.9
16.5*
1.02, 1.17
0.98, 1.11
0.89, 1.06
1.10
1.04
0.98*
2.21, 2.87
1.90, 2.45
1.70, 2.42
2.54
2.17
2.06*
9.84, 10.1
9.86, 10.1
9.81, 10.1
9.95
9.95
9.94
19.5, 20.0
19.5, 19.9
19.2, 19.8
19.7
19.7
19.5
0.49, 0.57
0.46, 0.53
0.46, 0.55
0.53
0.50
0.50
16.5, 17.1
16.7, 17.3
15.9, 17.3
rs6502051
27.8
48.7
23.5
16:1 n,7
Mean
Mean
Mean
Mean
Mean
GG
GA
AA
We found that blood levels of palmitoleic acid (16:1n-7)
were related to prostate cancer incidence in this prospective
study. The association for palmitoleic acid was strongest for
Gleason ≥7 tumors, with a 4-fold greater incidence for men
in the top quintile of blood palmitoleic acid levels than for
men in the lowest quintile. Furthermore, we found that 2
tagging SNPs in FASN (rs6502051 and rs4246444) were
significantly related to lower blood levels of myristoleic (14:
1n-5), palmitoleic (16:1n-7), and oleic (18:1n-9) acids. These
results, together with our previous findings that the same 2
rs1127678
DISCUSSION
14:1 n,5b
stearic to palmitic acid (18:0/16:0) as an index of fatty acid
elongase (ELOVL6) activity and found rate ratios of 1.00,
0.94 (95% CI: 0.64, 1.38), 0.75 (95% CI: 0.50, 1.13), 0.71
(95% CI: 0.47, 1.09), and 0.73 (95% CI: 0.47, 1.14) (Ptrend =
0.08), which closely mirrored the rate ratios observed for
stearicid.
We found no appreciable differences in the association
between individual fatty acids and prostate cancer when stratified by stage at diagnosis (Table 6). However, when stratified
by grade (Table 6), the association between blood palmitoleic
acid levels and prostate cancer incidence was limited to highgrade tumors (Gleason score ≥7) (Pheterogeneity = 0.02). When
blood levels of myristoleic (14:1n-5), palmitoleic (16:1n-7),
and oleic (18:1n-9) acids were included simultaneously in the
same model, the relation of palmitoleic acid (16:1n-7) with
high-grade tumors became stronger. The adjusted rate ratios
of Gleason ≥7 prostate cancer comparing top with bottom
quintiles of fatty acids in this model were 0.91 (95% CI: 0.35,
2.39; Ptrend = 0.58) for myristoleic acid, 4.32 (95% CI: 1.59,
11.8; Ptrend = 0.001) for palmitoleic acid, and 1.04 (95% CI:
0.42, 2.55; Ptrend = 0.96) for oleic acid. The association with
palmitoleic acid persisted even after adjustment for height,
body mass index, fatty acids previously related to prostate cancer risk in this cohort (linoleic, long chain n-3, and 18:2trans), and fasting status. In this model, the adjusted rate ratio
of Gleason ≥7 prostate cancer comparing top with bottom
quintiles of palmitoleic acid was 3.73 (95% CI: 1.23, 11.3;
Ptrend = 0.009). There was no evidence that the relation
between palmitoleic acid and prostate cancer was modified
by body mass index (Pinteraction = 0.44).
18:0
With this sample size, P < 0.05 for all r ≥ |0.13|.
16:0
1.00
14:0
a
16.8
17.0
16.6
0.50
18:1 n-9
0.98, 1.10
0.95, 1.10
0.93, 1.25
1.00
16:1 n-7
1.04
1.03
1.09
0.49
1.94, 2.43
1.94, 2.55
1.95, 3.33
0.69
2.18
2.25
2.64
1.00
14:1 n-5
9.86, 10.0
9.91, 10.1
9.75, 10.2
−0.42
9.95
10.0
10.0
−0.50
19.5, 19.8
19.4, 19.8
19.4, 20.4
−0.34
19.6
19.6
19.9
1.00
0.47, 0.54
0.46, 0.54
0.47, 0.64
0.37
18:0
0.44
0.50
0.50
0.56
0.72
95% CI
0.59
0.69
95% CI
0.91
1.00 −0.34
95% CI
1.00 0.69 −0.34
16:0
95% CI
14:0
95% CI
18:1 n-9
%
16:1 n-7
Genotype
14:1 n-5
SNP ID
18:0
Adjusteda Fatty Acid Concentration
16:0
Table 3. Blood Fatty Acid Levels According to Genotype in Tagging Polymorphisms of FASN, Physicians’ Health Study, United States, 1982–1995
14:0
Mean
Fatty Acid
Fatty Acid
5.2
31.7
63.1
18:1 n,9
Table 2. Spearman Correlation Coefficients Between Whole-Blood
Levels of Saturated and Monounsaturated Fatty Acids Among the
Controls (n = 476),a Physicians’ Health Study, United States, 1982–
1995
95% CI
1250 Chavarro et al.
Am J Epidemiol. 2013;178(8):1246–1255
Blood SFA and MUFA Levels and Prostate Cancer 1251
Table 4. Adjusteda Rate Ratios of Prostate Cancer (With 95% Confidence Intervals) by Control Quintiles of WholeBlood Saturated and Monounsaturated Fatty Acid Levels, Physicians’ Health Study, United States, 1982–1995
Concentrationb
Cases
Controls
RR
95% CI
14:0
0.40
Q1
0.21
74
94
1.00
Q2
0.32
95
95
1.32
0.86, 2.03
Q3
0.42
95
94
1.38
0.89, 2.15
Q4
0.55
122
97
1.70
1.10, 2.64
Q5
0.84
90
96
1.30
0.81, 2.07
16:0
0.25
Q1
17.6
82
90
1.00
Q2
18.7
96
97
1.10
0.71, 1.68
Q3
19.6
85
96
1.00
0.64, 1.55
Q4
20.5
110
98
1.28
0.83, 1.96
Q5
21.9
103
95
1.25
0.80, 1.97
9.0
112
97
1.00
18:0
Q1
Ptrendc
0.10
Q2
9.6
90
96
0.81
0.55, 1.19
Q3
10.0
112
93
1.02
0.69, 1.50
Q4
10.4
87
97
0.76
0.50, 1.15
Q5
10.9
75
93
0.68
0.44, 1.03
Q1
30.0
85
95
1.00
Q2
31.4
115
93
1.36
0.92, 2.03
Q3
32.2
74
96
0.83
0.53, 1.31
Q4
32.9
106
97
1.22
0.80, 1.87
Q5
34.1
96
95
1.11
0.72, 1.72
Total SFA
0.75
Abbreviations: CI, confidence interval; RR, rate ratio; SFA, saturated fatty acids.
a
Adjusted for matching factors (age, smoking status at baseline, and length of follow-up).
b
Median concentration (percentage of total fatty acids) in each quintile.
c
Calculated with median fatty acid concentration in each quintile as a continuous variable.
SNPs that predict lower blood levels of MUFAs in the present study are also associated with lower risk of death due to
prostate cancer (11), strongly suggest that de novo lipogenesis, acting through palmitoleic acid, could be an important metabolic pathway involved in prostate carcinogenesis.
Our findings lend further support to the hypothesis that de
novo lipogenesis is important in prostate carcinogenesis. FASN
is overexpressed in prostate tumors as compared with normal
prostate tissue, both at the mRNA level and the protein level
(9, 10, 18, 19). In addition, the expression level of FASN in prostate tumors is positively correlated to tumor grade (9, 10,
20), is highest in metastatic tumors (10, 20), and is related to
disease-specific mortality rate among overweight men with
prostate cancer (11). Inhibition of FASN by chemical inhibitors (cerulenin and C75) or by RNA interference–selective
gene silencing leads to a rapid decline in fatty acid synthesis,
associated with growth inhibition and cell apoptosis (20–
25). The growth arrest response triggered by FASN inhibition has been attributed to altered lipid production (21, 26).
Similarly, selective inhibition of acetyl CoA carboxylase
Am J Epidemiol. 2013;178(8):1246–1255
(ACC1) with specific gene silencing by RNA interference in
human LNCaP cells (25) is correlated to a decline in fatty
acid synthesis, growth arrest, and induction of apoptosis associated with alteration of mitochondrial function. We have
also reported that SNPs in FASN are related to prostate cancer–
specific mortality rate (11), and the direction of the association between FASN SNPs and prostate cancer mortality rate
is in the same direction of their association with blood MUFAs
in the present study. Specifically, the rate ratios for prostate
cancer–specific death were 0.72 (95% CI: 0.55–0.95) for
rs4246444 and 0.81 (95% CI: 0.64–1.03) for rs6502051. In
addition, the rate ratio of advanced prostate cancer for
rs6502051 was 0.77 (95% CI: 0.59–0.99). These 2 tagging
SNPs were also associated with lower blood levels of 14:1n5, 16:1n-7, and 18:1n-9 in the present study.
The results presented in this article, together with our previous findings (11), suggest that there could be at least 3
non–mutually exclusive mechanisms linking FASN to prostate cancer. On one hand, rs8066956, which was related to
FASN expression in tumor tissue (11), might have a direct
1252 Chavarro et al.
Table 5. Adjusteda Rate Ratios of Prostate Cancer (With 95% Confidence Intervals) by Control Quintiles of WholeBlood Monounsaturated Fatty Acid Levels, Physicians’ Health Study, United States, 1982–1995
Concentrationb
Cases
Controls
RR
95% CI
14:1 n-5
Ptrendc
0.01
Q1
0.00
90
117
1.00
Q2
0.01
66
73
1.23
Q3
0.02
98
92
1.46
0.97, 2.20
Q4
0.03
104
98
1.52
1.00, 2.31
Q5
0.04
118
96
1.74
1.15, 2.65
Q1
0.58
69
94
1.00
Q2
0.78
93
96
1.40
0.88, 2.22
Q3
0.96
89
95
1.35
0.86, 2.11
Q4
1.19
96
95
1.44
0.93, 2.24
Q5
1.71
129
96
1.97
1.27, 3.06
0.80, 1.90
16:1 n-7
0.003
18:1 n-9
0.04
Q1
14.2
78
95
1.00
Q2
15.5
88
95
1.18
Q3
16.5
83
94
1.12
0.73, 1.73
Q4
17.7
116
96
1.53
1.00, 2.35
Q5
19.5
111
96
1.47
0.95, 2.26
Q1
17.5
73
93
1.00
Q2
19.0
79
97
1.10
0.71, 1.72
Q3
20.2
107
95
1.50
0.97, 2.31
Q4
21.5
96
94
1.37
0.88, 2.13
Q5
23.8
121
97
1.70
1.09, 2.63
0.76, 1.83
Total MUFA
0.01
Abbreviations: CI, confidence interval; MUFA, monounsaturated fatty acids; RR, rate ratio.
a
Adjusted for matching factors (age, smoking status at baseline, and length of follow-up).
b
Median concentration (percentage of total fatty acids) in each quintile.
c
Calculated with median fatty acid concentration in each quintile as a continuous variable.
local effect on disease. Other variants in FASN could be
operating at a systemic level. rs1127678 could increase incidence of clinically relevant disease by increasing body mass
index (11)—itself a risk factor for disease-specific death in
this cohort (27)—and its related metabolic abnormalities.
Modulation of circulating 16:1n-7 levels (representing primarily adipose tissue and hepatic lipogenesis) might be an
additional mechanism explaining the previously reported
associations of rs6502051 and rs4246444 with prostate cancer incidence and disease-specific mortality rate. Experimental data suggest that 16:1n-7 behaves as a hormone in vivo
(28) and stimulates cell proliferation in vitro (29). There are
some gaps in this picture, however, that require further study.
It is not clear whether the hormone-like behavior of this fatty
acid is present in humans (30, 31), nor is it known whether
its mitogenic effects are also present in prostate tissue in
vivo. We also lacked data on other genes involved in de
novo lipogenesis: acetyl CoA carboxylase (ACC1), fatty acid
elongase (ELOVL6), and stearoyl CoA desaturase (SCD1).
Moreover, the main metabolic product of FASN 16:0 was
not related to the tagging SNPs evaluated, and though it was
positively related to prostate cancer incidence, this association was not statistically significant. Further examination of this
metabolic pathway is needed to clarify its relation to prostate
cancer.
In agreement with our findings, most previous studies
have found no relation between levels of myristic (14:0) (32,
33), palmitic (16:0) (14, 33–37), or stearic (18:0) (33–38)
acids and prostate cancer risk. The exceptions have been
reports from the Alpha-Tocopherol, Beta-Carotene Cancer
Prevention (ATBC) trial (35) and Norway’s Janus serum bank
(38) that indicated higher prostate cancer risk with increased
myristic acid levels, reports from Janus (38) and the European
Prospective Investigation into Cancer and Nutrition (EPIC)
(32) that showed greater prostate cancer risk with elevated
palmitic acid levels, and a report from EPIC that indicated
higher risk associated with blood levels of stearic acid (32).
On the other hand, our results sharply contrast with the literature on circulating MUFA levels and prostate cancer risk.
Our results for palmitoleic acid (16:1n-7) are in agreement
Am J Epidemiol. 2013;178(8):1246–1255
0.69
Abbreviations: CI, confidence interval; RR, rate ratio.
a
Adjusted for matching factors (age, smoking status at baseline, and length of follow-up).
b
For men in the highest quintile of the specific fatty acid in comparison with men in the lowest quintile.
c
Calculated with median fatty acid concentration in each quintile as a continuous variable.
0.77
0.15
0.07
0.81, 3.69
0.79, 3.70
1.71
1.73
0.05
0.11
0.80, 2.30
0.95, 2.86
1.65
1.36
0.71
0.81
0.27
0.31
0.41, 5.36
0.47, 4.91
1.52
1.48
0.1
0.03
0.98, 2.62
1.6
0.85, 2.18
18:1 n-9
1.36
Am J Epidemiol. 2013;178(8):1246–1255
Total monounsaturated
0.78
0.02
<0.001
0.07
1.01, 4.32
1.72, 8.94
3.92
2.09
0.04
0.27
0.87, 2.62
0.97, 2.77
1.64
1.51
0.86
0.25
0.91
0.2
0.66, 7.40
0.40, 5.38
1.46
2.21
0.01
1.12, 2.95
1.82
0.01
1.17, 2.91
1.85
14:1 n-5
16:1 n-7
0.13
0.17
0.79, 3.57
1.68
0.51
0.52, 1.55
0.9
0.55
0.66
0.24, 2.28
0.75
0.63
0.73, 1.95
1.19
Monounsaturated fatty acids
Total saturated
0.19
0.97
0.42, 1.80
0.86
0.03
0.31, 0.89
0.53
0.34
0.86
0.34, 2.81
0.98
0.38, 0.97
18:0
0.6
0.05
0.58
0.7
0.39
0.28
0.70, 3.47
0.75, 3.53
1.63
1.56
0.59
0.61
0.62, 2.06
0.64, 1.99
1.13
1.13
0.11
0.46
0.82
0.25
0.16, 3.14
0.41, 4.95
1.42
0.7
0.76, 2.11
0.16
0.83, 2.26
1.27
14:0
1.37
0.18
Ptrendc
95% CIb
RRb
Ptrendc
95% CIb
RRb
Saturated fatty acids
16:0
Pheterogeneity
Ptrendc
95% CIb
Gleason ≥7 (n = 168 cases)
RRb
Ptrendc
95% CIb
RRb
Pheterogeneity
Tumor Grade (Gleason Score)
Gleason <7 (n = 298 cases)
T3/T4/M1 (n = 69 cases)
Tumor Stage (Tumor Node Metastasis System)
T1/T2 (n = 388 cases)
Fatty Acid
Table 6. Adjusteda Rate Ratios of Prostate Cancer (With 95% Confidence Intervals) Comparing Top and Bottom Quintiles of Blood Fatty Acids According to Tumor Stage and Grade,
Physicians’ Health Study, United States, 1982–1995
Blood SFA and MUFA Levels and Prostate Cancer 1253
with a positive association reported by Harvei and colleagues
(38) but not with the null results reported by 3 other prospective nested case-control studies (32, 34, 35). Likewise, the
positive association between oleic acid (18:1n-9) levels and
prostate cancer is consistent with a previous report from the
Physicians’ Health Study (14), but it contrasts with all other
previous nested case-control studies (32, 34, 35, 38). We are
unaware of previous studies relating levels of myristoleic acid
(14:1n-5) to prostate cancer. An important consideration in
comparing our results with previous studies is that we used
whole blood, whereas most previous studies have used plasma
(14, 32, 37), serum (35, 38), or erythrocyte (34, 39) levels.
However, we have found previously that whole-blood fatty
acid levels are highly correlated to each other (2) and that the
relation of dietary fatty acids with erythrocyte levels is not different from their relation with plasma levels (3), which suggests that it is possible to directly compare our results with the
existing literature. Another important consideration is that the
time between blood draws was generally longer in our study
than in the existing literature. This might be important, given
that the results for palmitoleic acid tended to be stronger with
longer follow-up time.
Our study has several strengths. First, blood samples were
collected before prostate cancer diagnosis. The prospective
design and high follow-up rates of the Physicians’ Health
Study cohort decrease the possibility that our findings could
be result of bias. Moreover, the results became stronger after
exclusion of the first 2, 5, or 9 years of follow-up, which suggests that reverse causation is not a plausible explanation for
our findings. The large number of cases allowed us to examine
these associations with sufficient statistical power for total
prostate cancer and for major clinical groupings of the disease. The most important limitation of our study is that residual and unmeasured factors associated with blood fatty acid
levels could be responsible for the observed associations.
Nevertheless, we evaluated several variables as potential confounders and found that adjustment for the few variables associated with fatty acid levels and prostate cancer, including all
the fatty acids previously associated with prostate cancer
risk in this cohort (4, 5), had minimal impact on the results.
In summary, we found that blood levels of palmitoleic (16:
1n-7) acid were associated with a higher incidence of prostate
cancer, particularly of high-grade disease. We also found that
2 tagging SNPs in FASN (rs6502051 and rs4246444) previously related to lower prostate cancer mortality rate (11) are
also associated with lower whole-blood levels of this fatty acid.
The findings presented in the present article, together with our
previous report and experimental data from others, strongly
suggest that de novo lipogenesis is involved in the origin of
clinically relevant prostate cancer. Further investigation of how
this metabolic pathway is involved in prostate carcinogenesis is
warranted.
ACKNOWLEDGMENTS
Author affiliations: Department of Nutrition, Harvard
School of Public Health, Boston, Massachusetts (Jorge E.
Chavarro, Meir J. Stampfer, Hannia Campos); Department
1254 Chavarro et al.
of Epidemiology, Harvard School of Public Health, Boston,
Massachusetts (Jorge E. Chavarro, Meir J. Stampfer); Channing Division of Network Medicine, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical
School, Boston, Massachusetts (Jorge E. Chavarro, Meir J.
Stampfer, Jing Ma); Department of Urology, School of Medicine, University of California San Francisco, San Francisco,
California (Stacey A. Kenfield); Department of Pathology,
Center for Molecular Oncologic Pathology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, Massachusetts
(Massimo Loda); Department of Medical Oncology, DanaFarber Cancer Institute, Harvard Medical School, Boston, Massachusetts (Massimo Loda); Division of Preventive Medicine,
Department of Medicine, Brigham and Women’s Hospital,
Harvard Medical School, Boston, Massachusetts (Howard D.
Sesso).
This work was supported by grant W81XWH-11-1-0529
from the US Department of Defense and grants CA42182,
CA58684, CA90598, CA141298, CA97193, CA34944,
CA40360, CA131945, P50CA90381, 1U54CA155626-01,
P30DK046200, HL26490, and HL34595 from the National
Institutes of Health.
Conflict of interest: None declared.
10.
11.
12.
13.
14.
15.
16.
17.
REFERENCES
1. American Cancer Society. Cancer Facts and Figures 2012.
Atlanta: American Cancer Society; 2012:19–20.
2. Baylin A, Kim MK, Donovan-Palmer A, et al. Fasting whole
blood as a biomarker of essential fatty acid intake in
epidemiologic studies: comparison with adipose tissue and
plasma. Am J Epidemiol. 2005;162(4):373–381.
3. Sun Q, Ma J, Campos H, et al. Comparison between plasma
and erythrocyte fatty acid content as biomarkers of fatty acid
intake in US women. Am J Clin Nutr. 2007;86(1):74–81.
4. Chavarro JE, Stampfer MJ, Li H, et al. A prospective study of
polyunsaturated fatty acid levels in blood and prostate cancer
risk. Cancer Epidemiol Biomarkers Prev. 2007;16(7):
1364–1370.
5. Chavarro JE, Stampfer MJ, Campos H, et al. A prospective
study of trans-fatty acid levels in blood and risk of prostate
cancer. Cancer Epidemiol Biomarkers Prev. 2008;17(1):
95–101.
6. Vega-Lopez S, Ausman LM, Jalbert SM, et al. Palm and
partially hydrogenated soybean oils adversely alter lipoprotein
profiles compared with soybean and canola oils in
moderately hyperlipidemic subjects. Am J Clin Nutr. 2006;84
(1):54–62.
7. Tholstrup T, Raff M, Basu S, et al. Effects of butter high in
ruminant trans and monounsaturated fatty acids on
lipoproteins, incorporation of fatty acids into lipid classes,
plasma C-reactive protein, oxidative stress, hemostatic
variables, and insulin in healthy young men. Am J Clin Nutr.
2006;83(2):237–243.
8. Lichtenstein AH, Erkkila AT, Lamarche B, et al. Influence of
hydrogenated fat and butter on CVD risk factors: remnant-like
particles, glucose and insulin, blood pressure and C-reactive
protein. Atherosclerosis. 2003;171(1):97–107.
9. Swinnen JV, Roskams T, Joniau S, et al. Overexpression
of fatty acid synthase is an early and common event in
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
the development of prostate cancer. Int J Cancer. 2002;
98(1):19–22.
Rossi S, Graner E, Febbo P, et al. Fatty acid synthase
expression defines distinct molecular signatures in prostate
cancer. Mol Cancer Res. 2003;1(10):707–715.
Nguyen PL, Ma J, Chavarro JE, et al. Fatty acid synthase
polymorphisms, tumor expression, body mass index, prostate
cancer risk, and survival. J Clin Oncol. 2010;28(25):
3958–3964.
Steering Committee of the Physicians’ Health Study Research
Group. Final report on the aspirin component of the ongoing
Physicians’ Health Study. N Engl J Med. 1989;321(3):
129–135.
Hennekens CH, Buring JE, Manson JE, et al. Lack of effect of
long-term supplementation with beta carotene on the incidence
of malignant neoplasms and cardiovascular disease. N Engl J
Med. 1996;334(18):1145–1149.
Gann PH, Hennekens CH, Sacks FM, et al. Prospective study
of plasma fatty acids and risk of prostate cancer. J Natl Cancer
Inst. 1994;86(4):281–286.
Zock PL, Gerristen J, Katan M. Partial conservation of the sn2 position of dietary triglicerides in fasting plasma lipids in
humans. Eur J Clin Invest. 1996;26(2):141–150.
Zock PL, Mensink RP, Harryvan J, et al. Fatty acids in
serum cholesteryl esters as quantitative biomarkers of
dietary intake in humans. Am J Epidemiol. 1997;145(12):
1114–1122.
Zeleniuch-Jacquotte A, Chajes V, Van Kappel AL, et al.
Reliability of fatty acid composition in human serum
phospholipids. Eur J Clin Nutr. 2000;54(5):367–372.
Swinnen JV, Vanderhoydonc F, Elgamal AA, et al. Selective
activation of the fatty acid synthesis pathway in human prostate
cancer. Int J Cancer. 2000;88(2):176–179.
Shah US, Dhir R, Gollin SM, et al. Fatty acid synthase gene
overexpression and copy number gain in prostate
adenocarcinoma. Hum Pathol. 2006;37(4):401–409.
Bandyopadhyay S, Pai SK, Watabe M, et al. FAS expression
inversely correlates with PTEN level in prostate cancer and a
PI 3-kinase inhibitor synergizes with FAS siRNA to induce
apoptosis. Oncogene. 2005;24(34):5389–5395.
Kuhajda FP, Jenner K, Wood FD, et al. Fatty acid synthesis: a
potential selective target for antineoplastic therapy. Proc Natl
Acad Sci U S A. 1994;91(14):6379–6383.
Pizer ES, Jackisch C, Wood FD, et al. Inhibition of fatty acid
synthesis induces programmed cell death in human breast
cancer cells. Cancer Res. 1996;56(12):2745–2747.
Zhou W, Simpson PJ, McFadden JM, et al. Fatty acid synthase
inhibition triggers apoptosis during S phase in human cancer
cells. Cancer Res. 2003;63(21):7330–7337.
De Schrijver E, Brusselmans K, Heyns W, et al. RNA
interference–mediated silencing of the fatty acid synthase gene
attenuates growth and induces morphological changes and
apoptosis of LNCaP prostate cancer cells. Cancer Res.
2003;63(13):3799–3804.
Brusselmans K, De Schrijver E, Verhoeven G, et al. RNA
interference–mediated silencing of the acetyl-CoAcarboxylase-alpha gene induces growth inhibition and
apoptosis of prostate cancer cells. Cancer Res. 2005;65(15):
6719–6725.
Li JN, Gorospe M, Chrest FJ, et al. Pharmacological inhibition
of fatty acid synthase activity produces both cytostatic and
cytotoxic effects modulated by p53. Cancer Res. 2001;61(4):
1493–1499.
Ma J, Li H, Giovannucci E, et al. Prediagnostic body-mass
index, plasma C-peptide concentration, and prostate cancer–
Am J Epidemiol. 2013;178(8):1246–1255
Blood SFA and MUFA Levels and Prostate Cancer 1255
28.
29.
30.
31.
32.
33.
specific mortality in men with prostate cancer: a long-term
survival analysis. Lancet Oncol. 2008;9(11):1039–1047.
Cao H, Gerhold K, Mayers JR, et al. Identification of a
lipokine, a lipid hormone linking adipose tissue to systemic
metabolism. Cell. 2008;134(6):933–944.
Koeberle A, Shindou H, Harayama T, et al. Palmitoleate is a
mitogen, formed upon stimulation with growth factors, and
converted to palmitoleoyl-phosphatidylinositol. J Biol Chem.
2012;287(32):27244–27254.
Gong J, Campos H, McGarvey S, et al. Adipose tissue
palmitoleic acid and obesity in humans: does it behave as a
lipokine? Am J Clin Nutr. 2011;93(1):186–191.
Hodson L, Karpe F. Is there something special about
palmitoleate? Curr Opin Clin Nutr Metab Care. 2013;
16(2):225–231.
Crowe FL, Allen NE, Appleby PN, et al. Fatty acid
composition of plasma phospholipids and risk of prostate
cancer in a case-control analysis nested within the European
Prospective Investigation into Cancer and Nutrition. Am J Clin
Nutr. 2008;88(5):1353–1363.
Jackson MD, Walker SP, Simpson-Smith CM, et al.
Associations of whole-blood fatty acids and dietary intakes
with prostate cancer in Jamaica. Cancer Causes Control.
2012;23(1):23–33.
Am J Epidemiol. 2013;178(8):1246–1255
34. Park SY, Wilkens LR, Henning SM, et al. Circulating fatty
acids and prostate cancer risk in a nested case-control study:
the Multiethnic Cohort. Cancer Causes Control. 2009;20(2):
211–223.
35. Mannisto S, Pietinen P, Virtanen MJ, et al. Fatty acids and risk
of prostate cancer in a nested case-control study in male
smokers. Cancer Epidemiol Biomarkers Prev. 2003;12(12):
1422–1428.
36. Shannon J, O’Malley J, Mori M, et al. Erythrocyte fatty acids
and prostate cancer risk: a comparison of methods.
Prostaglandins Leukot Essent Fatty Acids. 2010;83(3):
161–169.
37. Ukoli FA, Fowke JH, Akumabor P, et al. The association of
plasma fatty acids with prostate cancer risk in African
Americans and Africans. J Health Care Poor Underserved.
2010;21(1 suppl):127–147.
38. Harvei S, Bjerve KS, Tretli S, et al. Prediagnostic level of fatty
acids in serum phospholipids: omega-3 and omega-6 fatty
acids and the risk of prostate cancer. Int J Cancer. 1997;71(4):
545–551.
39. Shannon J, King IB, Moshofsky R, et al. Erythrocyte
fatty acids and breast cancer risk: a case-control study
in Shanghai, China. Am J Clin Nutr. 2007;85(4):
1090–1097.