J C E M
O N L I N E
Hot Topics in Translational Endocrinology—Endocrine Research
Expression Profiles of Six Circulating MicroRNAs
Critical to Atherosclerosis in Patients With Subclinical
Hypothyroidism: A Clinical Study
Xinhuan Zhang,* Shanshan Shao,* Houfa Geng,* Yong Yu, Chenggang Wang,
Zhanfeng Liu, Chunxiao Yu, Xiuyun Jiang, Yangxin Deng, Ling Gao,
and Jiajun Zhao
Departments of Endocrinology (X.Z., S.S., H.G., C.Y., X.J., L.G., J.Z.), Sonography (Y.Y.), and Clinic
Laboratory (Z.L.), and Central Laboratory (L.G.), Shandong Provincial Hospital Affiliated to Shandong
University, and Shandong Clinical Medical Center of Endocrinology and Metabolism (X.Z., S.S., H.G.,
C.Y., X.J., J.Z.), Institute of Endocrinology and Metabolism, Shandong Academy of Clinical Medicine,
Jinan 250021, China; Department of Endocrinology (X.Z., Y.D.), the Affiliated Hospital of Taishan
Medical University, Tai’an 271000, China; Department of Endocrinology (H.G.), Affiliated Hospital of
Jining Medical University, Jining 272029, China; and Department of Preventive Medicine (C.W.),
Shandong University of Traditional Chinese Medicine, Jinan 250355, China
Context: Increasing evidence shows that subclinical hypothyroidism (SCH) is associated with atherosclerosis (ATH), but the association remains controversial. MicroRNAs (miRNAs) have been
proved to be involved in atherosclerosis and dyslipidemia as gene regulators.
Objective: The objective of the study was to determine the expression profiles of six serum miRNAs
critical to atherosclerosis in SCH patients and reanalyze the association between atherosclerosis
and SCH from a new perspective.
Outcomes, Design, and Participants: MicroRNA profiling analysis was performed by real-time PCR
in normal control subjects (NC; n ⫽ 22); patients with subclinical hypothyroidism alone (SCH; n ⫽
20); SCH patients plus atherosclerosis (SCH⫹ATH; n ⫽ 21); and patients with atherosclerosis but
without subclinical hypothyroidism (ATH; n ⫽ 22).
Results: MiR-21-5p was up-regulated in SCH, SCH⫹ATH, and ATH groups than in the NC group. In
addition, expression levels of miR-21-5p in SCH⫹ATH group were higher than in SCH alone and ATH
alone groups, respectively. Both miR-125a-5p and miR-126-3p showed a decreased trend from NC
to SCH and then to SCH⫹ATH or ATH subjects. MiR-221-3p and miR-222-3p were decreased in the
SCH⫹ATH and ATH groups compared with either the NC or SCH groups. No differences were found
in the levels of miR125a-5p, miR126-3p, miR221-3p, and miR222-3p between the ATH and
SCH⫹ATH group.
Conclusions: MiR-21-5p showed the most specific expression patterns in all patients with subclinical
hypothyroidism (SCH and SCH⫹ATH groups). Down-regulation of miR-125a-5p, miR-126-3p, miR221-3p, and miR-222-3p may be a manifestation of atherosclerosis either in SCH⫹ATH or in ATHalone patients. MiR-126-3p has the most specific expression patterns in all atherosclerosis patients
(SCH⫹ATH and ATH groups). (J Clin Endocrinol Metab 99: E766 –E774, 2014)
ISSN Print 0021-972X ISSN Online 1945-7197
Printed in U.S.A.
Copyright © 2014 by the Endocrine Society
Received March 12, 2013. Accepted January 31, 2014.
First Published Online February 11, 2014
E766
jcem.endojournals.org
* X.Z., S.S, and H.G. contributed equally to this study.
Abbreviations: ATH, atherosclerosis; AUC, area under the ROC curve; BMI, body mass
index; BP, blood pressure; CIMT, carotid IMT; Ct, threshold cycle; FT3, free T3; FT4, free T4;
IMT, intima-media thickness; LDL-c, low-density lipoprotein cholesterol; LSD, least significant differences; NC, normal control; RNase, ribonuclease; ROC, receiver-operating characteristic; SCH, subclinical hypothyroidism; TC, total cholesterol; TG, triglyceride; VSMC,
vascular smooth muscle cell.
J Clin Endocrinol Metab, May 2014, 99(5):E766 –E774
doi: 10.1210/jc.2013-1629
doi: 10.1210/jc.2013-1629
ubclinical hypothyroidism (SCH) is a disorder characterized by elevated serum TSH level above the upper limit of reference with normal free thyroid hormones
[free T3 (FT3) and free T4 (FT4)] (1). Recently several
studies have linked SCH to dyslipidemia (2), cardiovascular diseases (3, 4), and increased carotid artery atherosclerosis risk (5, 6). However, one study did not find
such associations (7). Therefore, the association between SCH and atherosclerosis and the possible mechanism of their connection need to be reanalyzed from
different perspectives.
MicroRNAs (miRNAs) are a class of endogenous, small
(21–23 nucleotides in length), noncoding RNAs that regulate
many cellular functions by inhibiting the translation of proteins from specific mRNAs or by promoting the degradation
of the mRNAs. Since discovered, miRNAs were found to
play significant roles in a wide spectrum of diseases,
including atherosclerosis and dyslipidemia (8 –10). Unlike
mRNAs, miRNAs have been found in serum, plasma, and
other body fluids in a stable form that is protected from endogenous ribonuclease (RNase) activity (11). Confirming
the expression patterns of atherosclerosis-specific serum
miRNAs in patients with SCH may help us reanalyze the
association between atherosclerosis and subclinical hypothyroidism from different perspectives.
We aimed to address the hypothesis that expression patterns of serum atherosclerosis-specific miRNAs in SCH patients could predict the presence of atherosclerosis at an early
S
jcem.endojournals.org
E767
stage. Six serum miRNAs that have been confirmed as critical
gene regulators involved in atherosclerosis and can be stable
determined in serum (10 –17) were selected in our study.
miR-126 governs vascular integrity and angiogenesis (12).
miR-125a inhibits the inflammatory response and contributes to vascular inflammation (13). Both miR-21 and miR210 are proangiogenesis-associated miRNAs involved in the
proliferation and apoptosis of vascular smooth muscle cells
(VSMCs) (14, 15). However, miR-221 and miR-222 are antiangiogenic gene-regulating miRNAs involved in VSMC
growth (16, 17).
Materials and Methods
Research subjects
SCH is defined as elevated TSH level above the normal upper limit
with normal levels of free T4 (1). Carotid artery atherosclerosis (ATH)
defined as thickness of maximum carotid intima-media is above 0.9
mm (18). All subjects recruited in our study were divided into four
groups. Healthy control subjects (NC; n ⫽ 22), SCH-alone patients
(SCH;n⫽20),newlydiagnosedSCHpatientsaccompaniedbycarotid
arteryATHsubjects(SCH⫹ATH;n⫽21),andpatientswithATHbut
without SCH (ATH; n ⫽ 22).
During the selection of the individuals recruited in our study, we
have screened about 1000 newly diagnosed subclinical hypothyroidism patients in the outpatient department of endocrinology
from Provincial Hospital affiliated to Shandong University and Affiliated Hospital of Taishan University from January 2010 to December 2012. Only 296 SCH patients completely met the exclusion
criteria (history of excessive drinking,
smoking, malignant tumor, ischemic heart
disease, hypertension, diabetes mellitus,
chronic liver diseases, and chronic renal
diseases). Among them, we first picked out
the 21 SCH⫹ATH patients, and then 20
SCH alone patients matched to the
SCH⫹ATH group for age, sex, blood pressure (BP), and body mass index (BMI) were
drawn from the 296 SCH patients. In addition, 22 NC subjects and 22 ATH alone
patients matched to the SCH⫹ATH group
for age, sex, BP, and BMI were drawn from
about 1200 individuals who receiving
health checkups at the physical examination center of Shandong provincial hospital during the same period. The study was
performed according to the Declaration of
Helsinki and was approved by the ethics
committee of the each hospital. Written informed consents were obtained from subjects in advance.
Figure 1. Multiple comparisons of baseline characteristics (TC, LDL-c, TG, and age) among NC,
SCH, SCH⫹ATH, and ATH subjects. Data are shown as the mean ⫾ SEM. P values were
generated by one-way ANOVA test followed by the LSD post hoc multiple comparisons test.
P ⬍ .05 was considered significant.
Measurement of
anthropometric and
biochemical parameters
See Supplement Online Data, published on The Endocrine Society’s Jour-
E768
Zhang et al
microRNAs in Subclinical Hypothyroidism
nals Online web site at http://jcem.endojournals.org, for
details.
Carotid artery ultrasound
All patients underwent a carotid intima-media thickness (CIMT)
examination using a color ultrasonic diagnostic apparatus equipped
with a 9-MHz linear-array transducer (Toshiba Aplio 500 ultrasound
scanner). Subjects resting in the supine position with head tilted backward while the study was performed at a temperature-controlled room
(25°C). Three different longitudinal-view observations of the bilateral
carotid artery were made at the following points: common carotid
artery (10 mm before the bulb), bulb (10 mm cranially to the start of
the bulb), and internal carotid artery (10 mm after the flow divider).
The CIMT was defined as the viewable distance between the bloodintimae and the media-adventitia interfaces on the wall of the artery.
Mean intima-media thickness (IMT; the mean of the three IMT measurementsoneachside)andthemaximumIMT(max-IMT;thehighest
IMT value among the six segments studied) were assessed (18, 19).
According to current sonographic criteria, we refer to max-IMT of 0.9
mm or less was normal; the max-IMT greater than 0.9 mm was considered indicative of thickened intima, whereas the max-IMT of 1.3
mm or greater indicative of atherosclerotic plaque (18, 19) (Supplemental Online Figures 1–3, respectively). The study protocol and criteria were based on current sonographic guidelines (18–20). Each scan
was made by the same investigator, who was blind to the patients’
clinical data.
Isolation of miRNAs
Total RNA-containing small RNAs was isolated from 200 ␮L
serum using the miRNeasy serum/plasma kit (QIAGEN: catalog
number 217184). Because of the lack of a stable endogenous
miRNA control for plasma or serum, a synthetic Caenorhabditis
elegans miR-39 miRNA mimic (cel-miR-39: catalog number
219610) was introduced as a spike-in control to monitor technical
variations in RNA recovery and the effectiveness for RT-PCR and
quantitative PCR. Previous studies have determined that celmiR-39 is suitable for normalization of circulating miRNA preparations (21). After addition of a denaturing solution, 3.5 ␮L of the
Table 1.
J Clin Endocrinol Metab, May 2014, 99(5):E766 –E774
miRNeasy serum/plasma spike-in control (1.6 ⫻ 108 copies/␮L
working solution) was added into the plasma samples to extract
again (recommend by QIAGEN). Total RNA containing miRNAs
was eluted in 14 ␮L of RNase-free water. The concentration of
RNA was quantified using a NanoDrop 1000 (NanoDrop).
Reverse transcribed to cDNA and quantitative
real-time PCR for specific miRNA
Serum RNA-containing small RNAs was reverse transcribed
to cDNA using a miScript II reverse transcriptase kit (QIAGEN;
catalog number 218160). RT-PCRs contained 50 ng total RNA,
5 ␮L 5⫻ miScript HiSpec buffer, 2.5 ␮L 10⫻ miScript nucleics
mix, 2.5 ␮L miScript reverse transcriptase mix, and valuable
RNase-free water up to 25 ␮L. The reactions were incubated at
37°C for 60 minutes, followed by inactivation of the reaction by
incubating at 95°C for 5 minutes.
Specific miRNA quantification was performed by SYBR
Green-based real-time PCR using a miScript SYBR Green PCR
kit (QIAGEN; catalog number 218073). Detailed information
about the six candidate miRNAs are presented in Supplemental
Table 1 and can be found at miRBase (http://microrna.sanger.
ac.uk). Each 20-␮L PCR mix contained 10 ␮L 2⫻ QuantiTect
SYBR Green PCR master mix, 2 ␮L 10⫻ miScript universal
primer, 2 ␮L 10⫻ QuantiTect miScript primer assay, 2.5 ␮L
RNase-free water, 3.5 ␮L cDNA as template. The reactions were
incubated at 95°C for 15 minutes, followed by 55 cycles of 94°C
for 15 seconds, and 55°C for 30 seconds and 70°C for 30 seconds. All reactions were run in duplicate. The threshold cycle
(Ct) was calculated using the second-derivative max method. Ct
values greater than 40 were treated as undetermined (22). The
relative expression of each miRNA after normalization to celmiR-39 is displayed as 2⫺[Ct (miRNA) ⫺ Ct (cel-miR-39)].
Statistical analysis
Data are expressed as the mean ⫾ SEM or raw numbers. The
normality of the variables was assessed using the Shapiro-Wilk test.
For data that were normally distributed, one-way ANOVA and the
Characteristics of Subjects Recruited in This Study
Gender (M/F)
Age, y
BMI, kg/m2
SBP, mm Hg
DBP, mm Hg
TC, mmol/L
TG, mmol/L
HDL-c, mmol/L
LDL-c, mmol/L
FT3, mmol/L
FT4, mmol/L
TSH, mIU/mL
mean-IMT, mm
max-IMT, mm
NC
SCH
SCHⴙATH
ATH
P Value
22 (5/17)
45.27 ⫾ 1.55
24.09 ⫾ 0.77
120.00 ⫾ 1.75
77.45 ⫾ 0.85
4.53 ⫾ 0.20
1.04 ⫾ 0.04
1.25 ⫾ 0.05
2.68 ⫾ 0.15
4.75 ⫾ 0.15
14.16 ⫾ 0.41
3.75 ⫾ 0.09
0.71 ⫾ 0.04
0.84 ⫾ 0.05
20 (3/17)
42.00 ⫾ 2.80
24.65 ⫾ 0.59
114.75 ⫾ 1.59
75.85 ⫾ 2.09
5.38 ⫾ 0.23
1.33 ⫾ 0.08
1.20 ⫾ 0.06
3.29 ⫾ 0.16
4.34 ⫾ 0. 17
14.48 ⫾ 0.39
7.86 ⫾ 0.60
0.72 ⫾ 0.07
0.87 ⫾ 0.02
21 (3/18)
47.95 ⫾ 2.35
25.14 ⫾ 0.65
120.14 ⫾ 3.01
80.00 ⫾ 1.94
6.25 ⫾ 0.27
1.42 ⫾ 0.10
1.32 ⫾ 0.07
3.84 ⫾ 0.22
4.25 ⫾ 0.23
13.96 ⫾ 0.50
8.44 ⫾ 0.84
0.93 ⫾ 0.14
1.35 ⫾ 0.16
22 (4/18)
50.41 ⫾ 1.47
25.90 ⫾ 0.57
122.36 ⫾ 1.69
77.45 ⫾ 1.68
5.75 ⫾ 0.19
1.36 ⫾ 0.12
1.33 ⫾ 0.07
3.29 ⫾ 0.18
4.82 ⫾ 0. 19
14.52 ⫾ 0.46
3.64 ⫾ 0.11
0.94 ⫾ 0.11
1.26 ⫾ 0.16
.884
.036a
.245
.082
.393
⬍.001b
.015a
.385
⬍.001b
.084
.783
NA
NA
NA
Abbreviations: DBP, diastolic blood pressure; HDL-c, high-density lipoprotein cholesterol; NA, not applicable; SBP, systolic blood pressure.
Continuous and categorical variables data were expressed as mean ⫾ SEM or real number of subjects, respectively. The statistical P value was
generated by the one-way ANOVA test or Kruskal-Wallis test. The ␹2 test was used to compare gender distribution.
a
P ⬍ .05 was considered significant.
b
P ⬍ .01 was considered significant.
doi: 10.1210/jc.2013-1629
least significant differences (LSD) post hoc multiple comparisons
test were applied. The Kruskal-Wallis test was performed to compare data that were not normally distributed. The ␹2 test was used
to compare gender distributions. Spearman’s correlations were performed to explore the relationships between miRNAs and TSH,
lipid parameters as well as CIMT. Cluster version 3.0 and Java
TreeView version 1.60 (http://rana.lbl.gov/eisensoftware.htm)
were used to perform an agglomerative hierarchical cluster analysis
based on the miRNA expression patterns as a panel. A squared
Euclidean distance measure method was involved in this analysis.
Receiver-operating characteristic (ROC) curves analysis was performed to assess the ability of six candidate miRNAs for detection
jcem.endojournals.org
E769
of all subclinical hypothyroidism patients or all atherosclerosis patients. All data were processed by SPSS software package for Windows version 13.0 (SPSS, Inc). All statistical tests were two sided,
and statistical significance was defined as P ⬍ .05.
Results
Significant differences in lipid parameters among
the four groups at baseline
As displayed in Figure 1, serum total cholesterol (TC),
low-density lipoprotein cholesterol (LDL-c), and triglyceride (TG) levels in the SCH,
SCH⫹ATH, and ATH groups were
all significantly higher than that in
NC group. In addition, serum TC
and LDL-c levels in the SCH⫹ATH
group were higher than in SCH
group, whereas no differences of serum TC, LDL-c, and TG levels were
found between the SCH⫹ATH and
ATH subjects. Age in the ATH group
was older than in the SCH group. No
significant differences were found in
other clinical factors, including gender distribution, BMI, BP, FT3, FT4,
and high-density lipoprotein cholesterol (Table 1).
Figure 2. A, Concentration of all RNA samples has no significance difference among the four
groups. B, The six miRNAs and cel-miR-39 served as the spike-in control all showed reliable Ct
values in all subjects. C, The relative expression levels of six candidate serum miRNAs (shown in
log2 scale) in the NC, SCH, SCH⫹ATH, and ATH groups. miR-39 is a synthetic Caenorhabditis
elegans miR-39 miRNA mimic (cel-miR-39) that served as the spike-in control to monitor the
effectiveness of RT-PCR and quantitative PCR. The horizontal lines indicate the mean. P values
were generated by one-way ANOVA test followed by the LSD post hoc multiple comparisons
test. P ⬍ .05 was considered significant.
All six miRNAs can be stably
detected in serum samples from
all subjects
Six miRNAs were stably detected
in serum samples of all patients. The
amplification curve (Supplemental
Online Figure 4) and melting curve
(Supplemental Online Figure 5) of
the six candidate miRNAs were presented. No significant differences
were discovered in the total RNA
concentration among the four groups
(Figure 2A; NC, 11.65 ⫾ 0.63 ng/L;
SCH, 11.64 ⫾ 0.67 ng/L; SCH⫹ATH,
11.45 ⫾ 0.62 ng/L, ATH, 11.55 ⫾ 0.64
ng/L).
In the quantitative PCR analysis
(Figure 2B), all six miRNAs and celmiR-39 showed reliable Ct values in
all subjects (Figure 2B). miR-210 expression was the least abundant,
with a median Ct of 36.95, whereas
miR-21-5p expression was the highest abundant, with a median Ct of
26.83. The spike-in control miRNA
E770
Zhang et al
microRNAs in Subclinical Hypothyroidism
J Clin Endocrinol Metab, May 2014, 99(5):E766 –E774
mimic (cel-miR-39) demonstrated both stable and moderate
abundance in all subjects and no differences among the four
groups, with a median Ct of 28.76.
creased in patients with atherosclerosis including the
SCH⫹ATH (⫺2.466 ⫾ 0.142 and ⫺2.636 ⫾ 0.128, respectively) and ATH groups (⫺2.323 ⫾ 0.197 and
⫺2.628 ⫾ 0.208, respectively) than in patients without
atherosclerosis including the NC (⫺1.276 ⫾ 0.160 and
⫺1.392 ⫾ 0.203, respectively) and SCH-alone groups
(⫺1.368 ⫾ 0.110 and ⫺1.563 ⫾ 0.125, respectively). No
significant differences were found between NC and SCH
subjects or between the SCH⫹ATH and ATH groups,
respectively.
Five of six candidate miRNAs have significantly
different expression levels among the four groups
As shown in Figure 2C, the relative expression of miR21-5p showed a significant up-regulation in the SCH
(1.842 ⫾ 0.221), SCH⫹ATH (2.447 ⫾ 0.102), and ATH
(1.654 ⫾ 0.171) groups compared with NC (0.912 ⫾
0.274), respectively, especially in the SCH⫹ATH group.
The increased miR-21-5p exceed in the ATH-alone and
SCH-alone groups, respectively. miR-210 showed a slight
but not significant stepwise elevation trend from NC
(⫺8.809 ⫾ 0.303) to SCH (⫺8.337 ⫾ 0.255) and then to
the SCH⫹ATH (⫺8.203 ⫾ 0.291) or ATH (⫺8.164 ⫾
0.287) group. Both miR-125a-5p and miR-126-3p
showed a significant stepwise decrease trend from the NC
(0.053 ⫾ 0.234 and 1.125 ⫾ 0.211, respectively) to SCH
(⫺0.732 ⫾ 0.206 and ⫺0.279 ⫾ 0.247, respectively) and
then to the SCH⫹ATH (⫺1.895 ⫾ 0.220 and ⫺1.257 ⫾
0.219, respectively) or ATH (⫺2.290 ⫾ 0.223 and
⫺1.232 ⫾ 0.289, respectively) group. No significant difference was found between the ATH and SCH⫹ATH subjects. Finally, both miR-221-3p and miR-222-3p were de-
Correlations between six candidate miRNAs and
TSH, lipid parameters as well as CIMT
As demonstrated in Table 2, Spearman’s correlation
analysis showed that TSH was positively correlated with
TC and LDL-c. TC, TG, and LDL-c were each positively
correlated with max-IMT and mean-IMT. With regard to
relative expressions of six candidate miRNAs in our study,
log2miR-21-5p was positively correlated with TSH and
max-IMT, whereas no associations were found between
log2miR-21-5p and blood lipid parameters as well as
mean-IMT. All the other down-regulated miRNAs
(log2miR-125a-5p, log2miR-126-3p, log2miR-221-3p,
and log2miR-222-3p) were negatively correlated with
blood lipids (TC, LDL-c, or TG) and CIMT. No correla-
Table 2. Spearman’s Correlations Between Relative Expression Levels of Six Candidate miRNAs (Shown in Log2
Scale) and TSH, Lipid Profiles, and CIMT in All Subjects
Parameter
TC
R
P
LDL-c
R
P
TG
R
P
Log2 miR-21-5p
R
P
Log2 miR-125a-5p
R
P
Log2 miR-126-3p
R
P
Log2 miR-210
R
P
Log2 miR-221-3p
R
P
Log2 miR-222-3p
R
P
TSH
TC
TG
LDL-c
Mean-IMT
Max-IMT
0.236
.029
1.000
NA
0.408
⬍.001
0.712
⬍.001
0.357
.001
0.454
⬍.001
0.278
.010
0.712
⬍.001
0.477
⬍.001
1.000
NA
0.402
⬍.001
0.472
⬍.001
0.145
.186
0.408
⬍.001
1.000
NA
0.477
⬍.001
0.222
.041
0.291
.007
0.414
⬍.001
0.016
.885
0.019
.866
0.052
.635
0.193
.077
0.228
.036
⫺0.015
.889
⫺0.306
.004
⫺0.177
.105
⫺0.241
.027
⫺0.492
⬍.001
⫺0.608
⬍.001
⫺0.094
.393
⫺0.365
.001
⫺0.224
.039
⫺0.283
.009
⫺0.479
⬍.001
⫺0.614
⬍.001
0.089
.417
0.119
.278
0.075
.496
0.165
.131
0.077
.482
0.149
.175
0.028
.800
⫺0.262
.015
⫺0.091
.408
0.205
.059
⫺0.505
⬍.001
⫺0.480
⬍.001
0.086
.436
⫺0.257
.017
⫺0.131
.231
⫺0.224
.039
⫺0.522
⬍.001
⫺0.512
⬍.001
Abbreviation: NA, not applicable. Correlations were presented as correlation coefficients (R) and significance (P) (P ⬍ .05 was considered significant).
doi: 10.1210/jc.2013-1629
jcem.endojournals.org
Figure 3. A majority of subjects of each group clustered together in the hierarchical clustering analysis
based on five differentially expressed serum miRNAs. Each row represents an individual serum miRNA, and
each column represents an individual subject. Red, black, and green colors indicate up-regulation,
unchanged expression, and down-regulation compared with spike-in-control, respectively. The color bar
displays different colors that correspond to the relative expression levels of miRNAs.
tions between log2miR-210 and TSH, lipid profile as well
as CIMT were observed. No correlations between the six
candidate miRNAs and clinical characteristics including
age, sex, BMI were found (P ⬎ .05).
Hierarchical clustering based on the five
differentially expressed serum miRNAs among
four groups could separate a majority of subjects
of each group
Hierarchical clustering analysis can find the linear combinations of miRNAs that maximize the probability of correctly
assigning subjects (23). The five miRNAs and their expression
levels were significantly different among the four groups were
included to generate a heat map diagram. Hierarchical clustering analysis based on the five differentially expressed serum
miRNAs could separate a majority of subjects of each group
(Figure 3): 16 of 22 NC subjects were clustered together with
two SCH patients mixed in it. Fifteen of 20 SCH-alone subjects
were clustered together with one ATH patient and three NC
subjects mixed in it. Sixteen of 21 SCH⫹ATH subjects were
clusteredtogetherwithonlytwoATHsubjectsmixedinit.Eighteen of 22 ATH subjects were clustered together with one SCH
patient and four SCH⫹ATH patients mixed in it. With regard
to variable clustering, miR-125a-5p was clustered with miR126-3p, and miR-221-3p was clustered with miR-222-3p,
respectively.
ROC curves of six candidate serum miRNAs for
detection of all subclinical hypothyroidism
patients or all atherosclerosis patients
ROC curves were constructed to estimate the sensitivity
and specificity of the six candidate serum miRNAs for
detection of all subclinical hypothyroidism patients and
E771
for detection of all atherosclerosis
patients from all subjects. The areas
under the ROC curves (AUCs) of the
six miRNAs are displayed in Figure
4, respectively. miR-21-5p (Figure
4A) had the largest AUC: 0.762
(95% confidence interval 0.658 –
0.865) among the six candidate
miRNAs for detection of all subclinical hypothyroidism patients including the SCH and SCH⫹ATH groups
from all subjects. miR-126-3p (Figure 4B) had the largest AUC: 0.888
(95% confidence interval 0.819 –
0.958) among the six candidate
miRNAs for detection of all atherosclerosis patients including the SCH⫹
ATH and ATH subjects.
Discussion
In this study, all patients recruited were newly diagnosed
SCH patients, never underwent any intervention, and a
different degree of the ATH may be accompanied by different changes of atherosclerosis-related parameters in
vivo. Results in our analysis showed that patients in the
SCH, SCH⫹ATH, and ATH groups differed from NC
individuals in the mean values of TG, TC, and LDL-c.
Moreover, TC and LDL-c were positively correlated with
TSH, indicating that TSH may contribute to increase the
circulating lipid levels. This result is in consistent with the
results of other studies that have assessed these associations (24, 25). With regard to the relationships between
CIMT and blood lipid, TC, LDL-c, and TG were all positively correlated with mean-IMT and max-IMT as expected. These data indicate that hyperlipidemia in patients
with subclinical hypothyroidism may be involved in the
process of atherosclerosis in SCH patients. This is supported by several previous studies (2– 4).
Current sonographic criteria of CIMT used to define
different degree of atherosclerosis were inconsistent (18,
26, 27). In our study, we defined CIMT of 0.9 mm or less
is normal, 0.9 mm less than CIMT less than 1.3 mm is
thickened CIMT, and CIMT of 1.3 mm or greater is
indicative of carotid plaque according to current sonographic criteria (18, 19). Although max-IMT was routinely treated as a representative of early stage atherosclerosis by guidelines, more recent literatures are paying
attention to the significance of mean-IMT in artery disease
(28 –30). So both mean-IMT and max-IMT were included
into our analysis like the recent similar reports (29, 30).
E772
Zhang et al
microRNAs in Subclinical Hypothyroidism
J Clin Endocrinol Metab, May 2014, 99(5):E766 –E774
correlations showed that the relative
expression level of miR-21-5p was
positively correlated with TSH and
no correlation between miR-21-5p
and TC, TG, or LDL-c were found,
suggesting that the up-regulation of
miR-21-5p in SCH may be chiefly
influenced by TSH rather than hyperlipidemia. Although no study focused on the significance of miR21-5p in thyroid disease can be
found to date, but a recent paper reported that thyroid dysfunction
could have a latent effect on circulating miRNA expression (32). Previous studies also suggested that the
connection between SCH and atherosclerosis can not be entirely explained by dyslipidemia (33, 34).
Furthermore, the expression level of
miR-21-5p in the ATH-alone group
was higher than in the NC group,
and miR-21-5p in the SCH⫹ATH
group was changed more dramatically than in the SCH-alone group.
The results suggested that the
changes of the expression level of
miR-21 also marked the occurrence
Figure 4. A, ROC curves for the ability of the six candidate serum miRNAs to differentiate the all
of atherosclerosis and may be inSCH patients (SCH⫹ATH and SCH) from the cases (NC and ATH). B, ROC curves for the ability of
volved in the atherosclerosis process
the six candidate serum miRNAs to differentiate the all atherosclerosis patients (SCH⫹ATH and
in patients with SCH. A previous
ATH) from the cases (NC and SCH). CI, confidence interval.
study has suggested that increased
miR-21-5p expression is involved in
Our quantitative PCR analysis confirmed previous reelevated stretch-induced human aortic smooth muscle cell
ports that circulating miRNAs can be reliably measured in
proliferation, and p27 and PDCD4 may be among its
circulation from humans (31). Five of the six atherosclerosisdownstream atherosclerosis-related target genes (14, 35).
specific miRNAs had significantly different expressions
Although miR-210 showed an elevated stepwise trend
among the four groups, and their levels were correlated with
from NC to SCH then to SCH⫹ATH or ATH, no signifblood lipids or TSH. All the levels of five differentially exicant difference was observed among the four groups. This
pressed serum miRNAs were correlated with max-IMT. The
results suggest that the five atherosclerosis-related serum result is not consistent with another study that focused on
miRNAs may be involved in the process of atherosclerosis in circulating miRNAs and atherosclerosis (9). A plausible
explanation for this inconsistency may be that carotid arour patients with atherosclerosis.
First, miR-21-5p was significantly up-regulated in the tery atherosclerosis indicates only the early stage of athSCH than in the NC groups when compared in subjects erosclerosis, and at this point critical gene regulators rewith normal CIMT. Meanwhile, expression level of miR- lated to smooth muscle cell proliferation have not changed
21-5p was higher in the SCH⫹ATH group than in ATH- dramatically.
miR-125a-5p showed a stepwise decrease trend from
alone group when compared in patients with thickened
NC
to SCH and then to SCH⫹ATH or ATH subjects, and
CIMT. The ROC curve analysis indicated that miR-21-5p
had the largest AUC for detection of all patients with sub- its relative expression level was negatively correlated with
clinical hypothyroidism. Suggesting that miR-21-5p may blood lipids and CIMT. These findings indicate that the
have specific expression patterns in all SCH patients in- decline of miR-125a-5p is closely related to dyslipidemia
cluding the SCH and SCH⫹ATH patients. Spearman’s and increased risk of atherosclerosis. miR-125a-5p regu-
doi: 10.1210/jc.2013-1629
lates the inflammatory response and lipid uptake in oxidized low-density lipoprotein-stimulated monocyte/macrophages and contributes to vascular inflammation (13).
We could find no research focused on circulating miR125a-5p in atherosclerosis disease, but other studies have
found that a cluster of specific miRNAs involved in the
vascular inflammation that is highly expressed in vitro are
down-regulated unexpectedly in the circulation of patients with atherosclerosis disease (8, 36).
miR-126-3p, an endothelial-specific miRNA, governs
vascular integrity and angiogenesis (12, 37). Our results
indicate that miR-126-3p was down-regulated in all patients with thickened CIMT than in patients with normal
CIMT. ROC curves indicate that miR-126-3p may be a
specific marker in detection of all atherosclerosis patients
including the SCH⫹ATH and ATH groups in all subjects.
The result indicated that injury of endothelial cells may be
the initial factor of any cause of atherosclerosis either
caused by SCH or by dyslipidemia. Fish et al (12) have
demonstrated that miR-126 regulates the response of endothelial cells to vascular endothelial growth factor. Harris et al (38) found that miR-126 inhibits vascular cell
adhesion molecule-1 expression, which is involved in leukocyte adhesion to endothelial cells. In addition, miR126-3p was correlated with blood lipids in our analysis,
consistent with another recent report (39). miR-221 and
miR-222 belong to the same family of antiangiogenic
gene-regulating miRNAs. P27 and P57 are two target
genes that are involved in VSMC growth (16). However,
miR-221 and miR-222 overexpression also indirectly reduces the expression of endothelial nitric oxide synthase,
which caused functional impairments, including inhibited
tube formation, endothelial cell migration, proliferation,
and angiogenesis (16, 17). Our data show that both miR221-3p and miR-222-3p decreased significantly in patients with ATH (SCH⫹ATH and ATH subjects) than in
patients without ATH (SCH and NC subjects), indicating
that the down-regulation of these two miRNAs may be
involved in the atherosclerosis process. However, no significant changes in their expression levels were found between NC and SCH subjects as well as between ATH and
SCH⫹ATH subjects, implying that the changes may be
mainly simultaneously with vascular proliferation and
marked the occurrence of ATH of any cause. These results
are consistent with a few but not all other reports (8, 9).
We were particularly interested in whether the patterns of
all examined miRNAs treated together as a panel could
provide more information than individual miRNAs. Hierarchical clustering analysis (23) based on five differentially expressed miRNAs could separate a majority of subjects of each group, indicating a better recognition of
subjects by miRNAs treated as a panel. The results sug-
jcem.endojournals.org
E773
gested that the five ATH-related miRNAs treated as a
panel featured a specific expression pattern and could correctly assign the subjects of each group.
No correlations between miRNAs and baseline characteristics including age, sex, and BMI were found, and
this is consistent with another recent report (40).
The strengths of this study include the repeated measurement of thyroid function to avoid labeling individuals
with transient TSH elevation as SCH; the hierarchical
clustering analysis including variable clustering and subject clustering based on the miRNA expression levels; and
the heat map generated by Cluster 3.0 and Java TreeView
1.6 software. The potential limitations of our study are the
small number of patients available due to the low prevalence of SCH.
In summary, the present study provides a first insight
into the significance of ATH-specific circulating miRNA
expression profiles in patients with SCH. We have shown
that five of the six ATH-related circulating miRNAs had
significantly different expressions among NC, SCH,
SCH⫹ATH, and ATH subjects. miR-21-5p showed the
most specific expression patterns in all patients with subclinical hypothyroidism, whereas miR-126-3p has the
most specific expression patterns in all atherosclerosis patients. However, prospective, large-scale studies are required to determine the potential value of circulating
miRNAs of ATH in patients with SCH.
Acknowledgments
Address all correspondence and requests for reprints to: Ling
Gao or Jiajun Zhao, Jiajun Zhao, Department of Endocrinology,
Shandong Provincial Hospital affiliated to Shandong University,
Institute of Endocrinology and Metabolic Diseases, Shandong
Academy of Clinical Medicine, 324 Jingwu Road, Jinan 250021
Shandong, China. E-mail: [email protected] or jjzhao@
medmail.com.cn.
This work was supported by the National Basic Research
Program (Grant 2012CB524900); National Natural Science
Foundation of China (Grant 81230018, 81100593, and
81101590); the Department of Science and Technology (Grant
2012GSF11824); Natural Science Foundation (Grants
ZR2009CM101 and BS2010YY049) of Shandong Province;
Jinan self-renovation plan of colleges and universities; and Scientific Research Institutes (Grant 200906012) of China.
Disclosure Summary: The authors have nothing to disclose.
References
1. Surks MI, Ortiz E, Daniels GH, et al. Subclinical thyroid disease:
scientific review and guidelines for diagnosis and management.
JAMA. 2004;291:228 –238.
2. Iqbal A, Jorde R, Figenschau Y. Serum lipid levels in relation to
E774
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
Zhang et al
microRNAs in Subclinical Hypothyroidism
serum thyroid-stimulating hormone and the effect of thyroxine
treatment on serum lipid levels in subjects with subclinical hypothyroidism: the Tromso Study. J Intern Med. 2006;260:53– 61.
Razvi S, Weaver JU, Vanderpump MP, Pearce SH. The incidence of
ischemic heart disease and mortality in people with subclinical hypothyroidism: reanalysis of the Whickham Survey cohort. J Clin
Endocrinol Metab. 2010;95:1734 –1740.
Ertugrul O, Ahmet U, Asim E, et al. Prevalence of subclinical hypothyroidism among patients with acute myocardial infarction.
ISRN Endocrinol. 2011;2011:810251.
Valentina VN, Marijan B, Chedo D, Branka K. Subclinical hypothyroidism and risk to carotid atherosclerosis. Arq Brasil Endocrinol Metab. 2011;55:475– 480.
McQuade C, Skugor M, Brennan DM, Hoar B, Stevenson C, Hoogwerf BJ. Hypothyroidism and moderate subclinical hypothyroidism
are associated with increased all-cause mortality independent of coronary heart disease risk factors: a PreCIS database study. Thyroid.
2011;21:837– 843.
Rodondi N, Newman AB, Vittinghoff E, et al. Subclinical hypothyroidism and the risk of heart failure, other cardiovascular events, and
death. Arch Intern Med. 2005;165:2460 –2466.
Fichtlscherer S, De Rosa S, Fox H, et al. Circulating microRNAs in
patients with coronary artery disease. Circ Res. 2010;107:677– 684.
Li T, Cao H, Zhuang J, et al. Identification of miR-130a, miR-27b
and miR-210 as serum biomarkers for atherosclerosis obliterans.
Clin Chim Acta. 2011;412:66 –70.
Hulsmans M, Sinnaeve P, Van der Schueren B, Mathieu C, Janssens
S, Holvoet P. Decreased miR-181a expression in monocytes of obese
patients is associated with the occurrence of metabolic syndrome
and coronary artery disease. J Clin Endocrinol Metab. 2012;97:
E1213–E1218.
Mitchell PS, Parkin RK, Kroh EM, et al. Circulating microRNAs as
stable blood-based markers for cancer detection. Proc Natl Acad Sci
USA. 2008;105:10513–10518.
Fish JE, Santoro MM, Morton SU, et al. miR-126 regulates angiogenic signaling and vascular integrity. Dev Cell. 2008;15:272–284.
Chen T, Huang Z, Wang L, et al. MicroRNA-125a-5p partly regulates the inflammatory response, lipid uptake, and ORP9 expression in oxLDL-stimulated monocyte/macrophages. Cardiovasc Res.
2009;83:131–139.
Baker AH. MicroRNA 21 “shapes” vascular smooth muscle behavior through regulating tropomyosin 1. Arteriosclerosis, Thromb
Vasc Biol. 2011;31:1941–1942.
Hu S, Huang M, Li Z, et al. MicroRNA-210 as a novel therapy for
treatment of ischemic heart disease. Circulation. 2010;122:S124 –
S131.
Li Y, Song YH, Li F, Yang T, Lu YW, Geng YJ. MicroRNA-221
regulates high glucose-induced endothelial dysfunction. Biochem
Biophys Res Commun. 2009;381:81– 83.
Dentelli P, Rosso A, Orso F, Olgasi C, Taverna D, Brizzi MF. microRNA-222 controls neovascularization by regulating signal transducer and activator of transcription 5A expression. Arterioscler
Thromb Vasc Biol. 2010;30:1562–1568.
Borhani NO, Mercuri M, Borhani PA, et al. Final outcome results
of the Multicenter Isradipine Diuretic Atherosclerosis Study (MIDAS). A randomized controlled trial. JAMA. 1996;276:785–791.
Doria A, Shoenfeld Y, Wu R, Gambari PF, et al. Risk factors for
subclinical atherosclerosis in a prospective cohort of patients with
systemic lupus erythematosus. Ann Rheum Dis. 2003;62:1071–
1077.
Mancia G, De Backer G, Dominiczak A, Cifkova R, et al. 2007
guidelines for the management of arterial hypertension: the Task
Force for the Management of Arterial Hypertension of the European
Society of Hypertension (ESH) and of the European Society of Cardiology (ESC). J Hypertens. 2007;25:1105–1187.
Kroh EM, Parkin RK, Mitchell PS, Tewari M. Analysis of circulating
J Clin Endocrinol Metab, May 2014, 99(5):E766 –E774
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
microRNA biomarkers in plasma and serum using quantitative reverse transcription-PCR (qRT-PCR). Methods. 2010;50:298 –301.
Wang GK, Zhu JQ, Zhang JT, et al. Circulating microRNA: a novel
potential biomarker for early diagnosis of acute myocardial infarction in humans. Eur Heart J. 2010;31:659 – 666.
Sturn A, Quackenbush J, Trajanoski Z. Genesis: cluster analysis of
microarray data. Bioinformatics. 2002;18:207–208.
Wang F, Tan Y, Wang C, et al. Thyroid-stimulating hormone levels
within the reference range are associated with serum lipid profiles
independent of thyroid hormones. J Clin Endocrinol Metab. 2012;
97:2724 –2731.
Tian L, Song Y, Xing M, et al. A novel role for thyroid-stimulating
hormone: up-regulation of hepatic 3-hydroxy-3-methyl-glutarylcoenzyme A reductase expression through the cyclic adenosine
monophosphate/protein kinase A/cyclic adenosine monophosphate-responsive element binding protein pathway. Hepatology.
2010;52:1401–1409.
Handa N, Matsumoto M, Maeda H, Hougaku H, Kamada T. Ischemic stroke events and carotid atherosclerosis. Results of the Osaka
Follow-up Study for Ultrasonographic Assessment of Carotid Atherosclerosis (the OSACA Study). Stroke. 1995;26:1781–1786.
Onut R, Balanescu AP, Constantinescu D, Calmac L, Marinescu M,
Dorobantu PM. Imaging atherosclerosis by carotid intima-media
thickness in vivo: how to, where and in whom? Maedica. 2012;7:
153–162.
Larsen JL, Colling CW, Ratanasuwan T, et al. Pancreas transplantation improves vascular disease in patients with type 1 diabetes.
Diabetes Care. 2004;27:1706 –1711.
Polak JF, Pencina MJ, Pencina KM, O’Donnell CJ, Wolf PA,
D’Agostino RB Sr. Carotid-wall intima-media thickness and cardiovascular events. N Engl J Med. 2011;365:213–221.
Irie Y, Katakami N, Kaneto H, et al. Maximum carotid intimamedia thickness improves the prediction ability of coronary artery
stenosis in type 2 diabetic patients without history of coronary artery
disease. Atherosclerosis. 2012;221:438 – 444.
Fang C, Zhu DX, Dong HJ, et al. Serum microRNAs are promising
novel biomarkers for diffuse large B cell lymphoma. Ann Hematol.
2012;91:553–559.
Liu R, Ma X, Xu L, et al. Differential microRNA expression in
peripheral blood mononuclear cells from Graves’ disease patients.
J Clin Endocrinol Metab. 2012;97:E968 –E972.
Taddei S. Impaired endothelium-dependent vasodilatation in subclinical hypothyroidism: beneficial effect of levothyroxine therapy.
J Clin Endocrinol Metab. 2003;88:3731–3737.
Shakoor SK, Aldibbiat A, Ingoe LE, et al. Endothelial progenitor
cells in subclinical hypothyroidism: the effect of thyroid hormone
replacement therapy. J Clin Endocrinol Metab. 2010;95:319 –322.
Song J, Hu B, Qu H, Bi C, Huang X, Zhang M. Mechanical stretch
modulates microRNA 21 expression, participating in proliferation
and apoptosis in cultured human aortic smooth muscle cells. PloS
One. 2012;7:e47657.
Urbich C, Kuehbacher A, Dimmeler S. Role of microRNAs in vascular diseases, inflammation, and angiogenesis. Cardiovasc Res.
2008;79:581–588.
Zou J, Li WQ, Li Q, et al. Two functional microRNA-126s repress
a novel target gene p21-activated kinase 1 to regulate vascular integrity in zebrafish. Circ Res. 2011;108:201–209.
Harris TA, Yamakuchi M, Ferlito M, Mendell JT, Lowenstein CJ.
MicroRNA-126 regulates endothelial expression of vascular cell
adhesion molecule 1. Proc Natl Acad Sci USA. 2008;105:1516 –
1521.
Sun X, Zhang M, Sanagawa A, et al. Circulating microRNA-126 in
patients with coronary artery disease: correlation with LDL cholesterol. Thromb J. 2012;10:16.
Corsten MF, Dennert R, Jochems S, et al. Circulating MicroRNA208b and MicroRNA-499 reflect myocardial damage in cardiovascular disease. Circ Cardiovasc Genet. 2010;3:499 –506.