Circulating Fatty Acid-Binding Protein 4 Concentration Predicts the Progression of Carotid Atherosclerosis in a General Population Without Medication

Abstract

Background: Fatty acid-binding protein 4 (FABP4), which is expressed in both adipocytes and macrophages, is secreted from the cells and acts as an adipokine. An elevated circulating FABP4 level is associated with insulin resistance and atherosclerosis.

Methods and Results: We investigated the causative association between FABP4 level and progression of atherosclerosis in subjects of the Tanno-Sobetsu Study, a population-based cohort. In 281 subjects without medication (male/female: 109/172) in the year 2010 or 2013, the carotid intima-media thickness (CIMT) assessed using carotid ultrasonography was significantly correlated with age, adiposity, blood pressure, renal dysfunction and levels of cholesterol, triglycerides, fasting glucose, HbA1c and FABP4 (r=0.331, P<0.001). Multiple regression analysis demonstrated that age, sex and FABP4 concentration were independent predictors of CIMT. A total of 78 (male/female: 29/49) of the 156 subjects in 2010 underwent carotid ultrasonography again in 2013. The change in CIMT each year during that 3-year period (mean±SD: 3.8±22.3 µm/year) was positively correlated with basal levels of high-sensitivity C-reactive protein (hsCRP) (r=0.231, P=0.046) and FABP4 (r=0.267, P=0.018) in 2010. After adjustment for age, sex and hsCRP level, the basal FABP4 level was independently associated with the change in CIMT per year.

Conclusions: FABP4 concentration is an independent predictor of the progression of carotid atherosclerosis.

Fatty acid-binding proteins (FABPs) are a family of intracellular lipid chaperones, approximately 14–15-kDa predominantly cytosolic proteins that can reversibly bind hydrophobic ligands such as saturated and unsaturated long-chain fatty acids.^1–3 FABPs have been proposed to facilitate the transport of lipids to specific compartments in the cell. Of them, FABP4, also known as adipocyte FABP (A-FABP) or aP2, is mainly expressed in adipocytes and macrophages and plays an important role in the development of obesity, insulin resistance (IR), type 2 diabetes mellitus (T2DM) and atherosclerosis.^1,4 We previously demonstrated that inhibition of FABP4 by a small molecule might be a novel therapeutic strategy against IR, T2DM and atherosclerosis.⁵

FABP4 is secreted from adipocytes in association with lipolysis via a non-classical secretion pathway,^6–8 though there are no typical secretory signal peptides in the sequence of FABP4.¹ It has recently been shown that FABP4 is also secreted from macrophages,⁹ but the mechanism remains unknown. Circulating FABP4 acts as an adipokine for the development of IR⁷ and atherosclerosis⁹ in experimental models. Furthermore, a recent study demonstrated the possibility of a new strategy to treat metabolic disease by targeting serum FABP4 with a monoclonal antibody to FABP4.¹⁰

Elevated circulating FABP4 levels are associated with obesity, IR, T2DM, hypertension, dyslipidemia, cardiac dysfunction, renal dysfunction, atherosclerosis and cardiovascular events.^6,11–18 Notably, an increased plasma level of FABP4 was shown to be independently associated with the presence of coronary artery disease.^19,20 Moreover, FABP4 has been found in human atherosclerotic plaques, and its presence was associated with the features of inflammatory and vulnerable plaques as high-risk phenotypes.^19,21

However, little is known about the causal link between circulating FABP4 and progression of atherosclerosis in humans. In the present study, we investigated the association of FABP4 level with the extent of atherosclerosis assessed by carotid intima-media thickness (CIMT) measured by carotid ultrasonography and its association with the progression of carotid atherosclerosis during a 3-year period in a general population of subjects who had not regularly taken any medications.

Methods

The present study consisted of 2 studies (Study 1 and Study 2) in the Tanno-Sobetsu Study, which had a population-based cohort design and was conducted in 2 rural towns, Tanno and Sobetsu, in Hokkaido, the northernmost island of Japan. All studies conformed to the principles outlined in the Declaration of Helsinki and were performed with the approval of the Ethical Committee of Sapporo Medical University. Written informed consent was received from all of the study subjects.

Study 1: FABP4 Level and CIMT

A total of 818 Japanese subjects (male/female: 353/465, mean age: 64±15 years) underwent annual examinations in Sobetsu Town in 2010 and/or 2013. All participants were invited to have a carotid ultrasonographic examination for assessment of CIMT, a marker of carotid atherosclerosis. Subjects who were being treated with any medications were excluded. A total of 281 subjects (male/female: 109/172) without medication who underwent carotid ultrasonography were enrolled in Study 1. The numbers of recruited subjects were 156 (male/female: 52/104, mean age: 61±14 years) in 2010 and 125 (male/female: 57/68, mean age: 63±17 years) in 2013. Among the 281 subjects without medication, past histories of heart disease and stroke in 8 and 3 subjects, respectively, were ascertained by a self-reported questionnaire.

Study 2: FABP4 Level and Change in CIMT

Of the 156 subjects without medication in 2010 in Study 1, 78 (male/female: 29/49) underwent carotid ultrasonography again in 2013. They were enrolled in Study 2. During that 3-year period, 15 of the 78 subjects started to take medications, including antihypertensive drugs (n=9), antidyslipidemic drugs (n=3), antiplatelet drugs (n=2) and antiarrhythmic drugs (n=1). The change in CIMT per year (µm/year) was calculated using the CIMT values from 2010 and 2013.

Measurements

Medical check-ups were performed between 06:00 hours and 09:00 hours after an overnight fast. After measuring anthropometric parameters, blood pressure (BP) was measured twice consecutively in the upper arm using an automated sphygmomanometer (HEM-907, Omron Co., Kyoto, Japan) with subjects in a seated resting position, and the average BP was used for analysis. Body mass index (BMI) was calculated as body weight (in kilograms) divided by the square of body height (in meters). Peripheral venous blood samples were obtained from study subjects after the physical examination for complete blood count and biochemical analyses. Samples of serum and plasma were analyzed immediately or stored at −80℃ until biochemical analyses.

The concentration of FABP4 was measured using a commercially available enzyme-linked immunosorbent assay kit for FABP4 (Biovendor R&D, Modrice, Czech Republic). The accuracy, precision and reproducibility of the kit have been described previously.⁶ The intra- and interassay coefficients of variation in the kit were <5%. According to the manufacturer’s protocol, no cross-reactivity of FABP4 with other FABP types occurs. Plasma glucose was determined by the glucose oxidase method. Fasting plasma insulin was measured by a chemiluminescent enzyme immunoassay method. Hemoglobin A1c (HbA1c) was determined by a latex coagulation method and expressed on the National Glycohemoglobin Standardization Program scale. Creatinine, blood urea nitrogen (BUN), uric acid, aspartate transaminase (AST), alanine aminotransferase (ALT), γ-glutamyl transpeptidase (γGTP) and lipid profiles, including total cholesterol, high-density lipoprotein cholesterol (HDL-C) and triglycerides, were determined by enzymatic methods. The low-density lipoprotein cholesterol (LDL-C) level was calculated by the Friedewald equation. B-type natriuretic peptide (BNP) was measured using an assay kit (Shionogi & Co., Osaka, Japan). High-sensitivity C-reactive protein (hsCRP) was measured by a nephelometry method. The homeostasis model assessment of IR (HOMA-R), an index of IR, was calculated by the previously reported formula: HOMA-R=insulin (μU/ml)×glucose (mg/dL)/405. As an index of renal function, estimated glomerular filtration rate (eGFR) was calculated by an equation for Japanese: eGFR (mL/min/1.73 m²)=194×creatinine^(−1.094)×age^(−0.287)×0.739 (if female).

Carotid Ultrasonography

Carotid ultrasonographic examinations were performed by 3 experienced examiners certified by the Japan Society of Ultrasonics in Medicine, who were blinded to the clinical data, using a Vivid 9 (GE Health Care, Tokyo, Japan) equipped with a multifrequency 4–10-MHz linear-array transducer. Mean CIMT of the far wall of both of the common carotid arteries was measured using commercially available semi-automated edge-detection software (IMT Option, GE Medical System, Milwaukee, WI, USA). The region of interest was placed from the beginning of the carotid bulbs to a 2-cm proximal site in each common carotid artery. In our laboratory, coefficients of variance of intra- and interoperator variability for semi-automated CIMT measurement (3.5% and 8.0%, respectively) were significantly lower than those for manual CIMT measurement (14.1% and 16.7%, respectively).²² The average value of CIMT in both of the common carotid arteries was used for analysis because of a significant positive correlation between the left and right CIMT values (n=281, r=0.670, P<0.001).

Statistical Analysis

Numeric variables are expressed as mean±SD for normal distributions or median (interquartile range) for skewed variables. The distribution of each parameter was tested for its normality using the Shapiro-Wilk W test, and non-normally distributed parameters were logarithmically transformed for correlation and regression analyses. Comparison between groups was done with the chi-square test, Wilcoxon signed-rank test for paired samples and the Mann-Whitney U test for unpaired samples. The correlation between 2 variables was evaluated using Pearson’s correlation coefficient. Stepwise and subsequent multivariate regression analyses were performed to identify independent determinants of FABP4, CIMT and change in CIMT using the variables with a significant correlation as independent predictors after consideration of multicollinearity, showing the t-ratio calculated as the ratio of unstandardized regression coefficient and standard error (SE) of the unstandardized regression coefficient, the standardized regression coefficient (β), the percentage of variance in the object variables that the selected independent predictors explained (R²) and Akaike’s Information Criterion (AIC). Among the candidate models, the best-fit model using AIC for each dependent variable was selected. P<0.05 was considered statistically significant. All data were analyzed using JMP 9 for Macintosh (SAS Institute, Cary, NC, USA).

Results

Study 1

The characteristics of the 281 recruited subjects with no medication (male/female: 109/172) are shown in Table 1. Mean age, BMI and waist circumference (WC) of the recruited subjects were 62±15 years, 22.8±3.2 kg/m² and 82.2±10.6 cm, respectively. Male subjects had significantly larger BMI and WC, significantly higher frequency of habitual smoking, and significantly higher levels of systolic and diastolic BPs, triglycerides, fasting glucose, HOMA-R, HbA1c, BUN, creatinine, uric acid, AST, ALT, γGTP and CIMT and lower levels of total cholesterol, HDL-C, BNP and FABP4 than did female subjects. No significant difference in age, pulse rate or level of LDL-C, eGFR, or hsCRP was found between the male and female subjects. Hypertension (systolic BP ≥140 mmHg or diastolic BP ≥90 mmHg), hyperlipidemia (LDL-C ≥140 mg/dL, HDL-C <40 mg/dL or triglycerides ≥150 mg/dL) and DM (HbA1c ≥6.5% and fasting glucose ≥126 mg/dL) were found in 91, 110, and 4 subjects, respectively.

Table 1. Characteristics of the Study 1 Subjects Without Medication (n=281)

	Total	Male	Female	P value
n	281	109	172
Age (years)	62±15	63±16	61±15	0.399
BMI (kg/m2)	22.8±3.2	23.9±2.9	22.1±3.2	<0.001
WC (cm)	82.2±10.6	86.3±8.9	79.6±10.8	<0.001
Systolic BP (mmHg)	129±20	135±18	126±21	0.015
Diastolic BP (mmHg)	74±12	79±11	72±11	0.001
Pulse rate (beats/min)	71±11	70±13	72±10	0.512
Habitual smoking	49 (17.4)	26 (23.9)	23 (13.4)	0.024
Biochemical data
TC (mg/dL)	204±34	196±32	209±34	0.002
LDL-C (mg/dL)	122±29	120±27	123±30	0.291
HDL-C (mg/dL)	66±18	58±17	70±16	<0.001
Triglycerides (mg/dL)	92 (65–125)	103 (74–145)	85 (61–116)	<0.001
Fasting glucose (mg/dL)	95 (90–103)	97 (90–105)	91 (86–98)	<0.001
Insulin (μU/mL)	5.3 (4.4–5.9)	5.4 (4.9–5.9)	5.2 (4.2–5.9)	0.189
HOMA-R	1.22 (0.97–1.48)	1.26 (1.07–1.58)	1.16 (0.92–1.41)	0.024
HbA1c (%)	5.2±0.4	5.3±0.5	5.2±0.4	0.002
BUN (mg/dL)	15±5	16±6	14±5	0.003
Creatinine (mg/dL)	0.8±0.2	0.9±0.3	0.7±0.1	<0.001
eGFR (mL/min/1.73 m2)	72.6±15.8	73.3±16.6	72.1±15.3	0.524
Uric acid (mg/dL)	5.1±1.3	6.0±1.2	4.6±1.0	<0.001
AST (IU/L)	22 (19–26)	23 (20–27)	21 (18–25)	0.049
ALT (IU/L)	18 (14–23)	21 (16–30)	16 (13–21)	<0.001
γGTP (IU/L)	20 (15–31)	27 (19–44)	18 (14–24)	<0.001
BNP (pg/mL)	17 (11–31)	15 (9–30)	19 (12–31)	0.024
hsCRP (mg/dL)	0.03 (0.01–0.06)	0.03 (0.02–0.06)	0.03 (0.01–0.06)	0.435
FABP4 (ng/ml)	10.8 (7.8–14.3)	9.3 (7.3–13.0)	11.5 (8.4–15.4)	0.001
Carotid ultrasonographic data
CIMT (mm)	0.69±0.13	0.71±0.14	0.67±0.13	0.011

Variables are expressed as number (%), mean±SD or median (interquartile range). AST, aspartate transaminase; ALT, alanine transaminase; BMI, body mass index; BNP, B-type natriuretic peptide; BP, blood pressure; BUN, blood urea nitrogen; CIMT, carotid intima-media thickness; eGFR, estimated glomerular filtration rate; FABP4, fatty acid-binding protein 4; γGTP, γ-glutamyl transpeptidase; hsCRP, high-sensitivity C-reactive protein; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; TC, total cholesterol; WC, waist circumference.

There was no significant difference between the level of FABP4 or CIMT in male subjects with and without a smoking habit. Levels of FABP4 and CIMT were significantly lower in female subjects with a smoking habit than in those without a smoking habit.

The logarithmically transformed serum FABP4 level was positively correlated with age, BMI, WC, systolic BP, and levels of fasting glucose, insulin, HOMA-R, HbA1c, BUN, creatinine, AST, ALT, BNP and hsCRP, and was negatively correlated with levels of HDL-C and eGFR (Table 2). A significantly positive correlation was found between FABP4 level and CIMT (r=0.331, P<0.001) (Figure 1). Multiple regression analysis showed that the FABP4 level was independently associated with sex, WC and levels of eGFR, AST and CIMT, explaining a total of 47.8% of the variance in this measure (R²=0.478) (Table 2).

Table 2. Correlation and Multiple Regression Analyses for FABP4 and CIMT (Study 1: n=281)

	Log FABP4				CIMT
	Correlation		Multiple regression		Correlation		Multiple regression
	r	P value	β	P value	r	P value	β	P value
Age	0.405	<0.001	0.059	0.400	0.677	<0.001	0.627	<0.001
Sex (male)	–		−0.377	<0.001	–		0.141	0.002
BMI	0.459	<0.001	NS		0.239	<0.001	NS
WC	0.462	<0.001	0.481	<0.001	0.233	<0.001	NS
Systolic BP	0.261	0.001	NS		0.445	<0.001	NS
Diastolic BP	0.124	0.124	–		0.126	0.116	–
Pulse rate	0.144	0.160	–		−0.010	0.924	–
Smoking habit	–		NS		–		NS
TC	0.033	0.578	–		0.051	0.393	–
LDL-C	0.098	0.101	–		0.150	0.012	NS
HDL-C	−0.140	0.019	NS		−0.243	<0.001	NS
Log triglycerides	0.087	0.144	–		0.183	0.002	NS
Log fasting glucose	0.143	0.017	NS		0.204	0.001	NS
Log insulin	0.192	0.001	NS		0.029	0.626	–
Log HOMA-R	0.210	<0.001	NS		0.078	0.193	–
HbA1c	0.258	<0.001	NS		0.259	<0.001	NS
BUN	0.236	<0.001	NS		0.251	<0.001	NS
Creatinine	0.204	0.001	NS		0.279	<0.001	NS
eGFR	−0.417	<0.001	−0.212	<0.001	−0.406	<0.001	NS
Uric acid	0.084	0.160	–		0.168	0.005	NS
Log AST	0.204	0.001	0.131	0.004	0.151	0.011	NS
Log ALT	0.142	0.017	NS		−0.011	0.850	–
Log γGTP	−0.001	0.990	–		−0.111	0.064	–
Log BNP	0.240	<0.001	NS		0.354	<0.001	NS
Log hsCRP	0.163	0.007	NS		0.080	0.188	–
Log FABP4	–	–	–		0.331	<0.001	0.105	0.032
CIMT	0.331	<0.001	0.130	0.030	–		–
			R2=0.478				R2=0.481

NS, not selected. Other abbreviations as in Table 1.

Figure 1.

Correlation between level of fatty acid-binding protein 4 (FABP4) and the extent of carotid atherosclerosis. Carotid intima-media thickness (CIMT) plotted against logarithmically transformed level of FABP4 in each subject of Study 1 (n=281). Open circle and broken regression line: males (n=109); closed circle and solid regression line: females (n=172).

On the other hand, CIMT was positively correlated with age (r=0.677, P<0.001), BMI, WC, systolic BP, and levels of LDL-C, triglycerides, fasting glucose, HbA1c, BUN, creatinine, uric acid, AST, BNP and FABP4, and was negatively correlated with levels of HDL-C and eGFR (Table 2). After adjustment for age and sex, CIMT was independently associated with the FABP4 level (β=0.105, P=0.032), explaining a total of 48.1% of the variance in this measure (R²=0.481) (Table 2).

Study 2

Among the 156 subjects without medication enrolled in Study 1 in 2010, 78 subjects were recruited for Study 2. There was no significant difference in basal parameters except for total cholesterol and LDL-C, between the 78 recruited subjects and 78 non-recruited subjects (Table 3). The characteristics of the 78 recruited subjects who underwent carotid ultrasonography in both 2010 and 2013 are shown in Table 4. Mean age, BMI and WC of the recruited subjects were 62±11 years, 22.7±3.2 kg/m² and 80.9±10.1 cm, respectively. As in Study 1, male subjects had significantly larger WC and had higher levels of diastolic BP, triglycerides, fasting glucose, creatinine, eGFR, uric acid and γGTP and lower levels of total cholesterol and HDL-C than did female subjects.

Table 3. Characteristics of the Recruited and Non-Recruited Subjects (Study 2)

	Recruited	Non-recruited	P value
n (male/female)	78 (29/49)	78 (23/55)	0.308
Age (years)	62±11	60±17	0.380
BMI (kg/m2)	22.7±3.2	22.3±2.8	0.438
WC (cm)	80.9±10.1	79.8±10.3	0.497
Systolic BP (mmHg)	128±18	130±22	0.443
Diastolic BP (mmHg)	74±11	75±12	0.511
Pulse rate (beats/min)	70±11	71±10	0.845
Habitual smoking	11 (14.1)	14 (17.9)	0.513
Biochemical data
TC (mg/dL)	215±30	202±29	0.006
LDL-C (mg/dL)	126±28	115±26	0.012
HDL-C (mg/dL)	67±19	66±18	0.639
Triglycerides (mg/dL)	101 (69–127)	90 (63–123)	0.166
Fasting glucose (mg/dL)	93 (89–99)	92 (87–100)	0.859
Insulin (μU/mL)	5.4 (3.5–7.6)	4.6 (3.0–6.3)	0.117
HOMA-R	1.29 (0.78–1.85)	1.02 (0.68–1.55)	0.218
HbA1c (%)	5.1±0.2	5.1±0.3	0.979
BUN (mg/dL)	14±4	14±4	0.682
Creatinine (mg/dL)	0.7±0.1	0.7±0.1	0.202
eGFR (mL/min/1.73 m2)	73.0±12.7	76.8±16.3	0.102
Uric acid (mg/dL)	4.9±1.2	5.0±1.2	0.814
AST (IU/L)	22 (19–26)	22 (19–26)	0.798
ALT (IU/L)	19 (15–23)	17 (13–24)	0.386
γGTP (IU/L)	20 (16–32)	19 (15–28)	0.322
BNP (pg/mL)	15 (11–28)	20 (12–33)	0.128
hsCRP (mg/dL)	0.02 (0.02–0.05)	0.03 (0.02–0.05)	0.773
FABP4 (ng/mL)	10.7 (7.8–13.2)	10.1 (7.6–14.0)	0.836
Carotid ultrasonographic data
CIMT (mm)	0.70±0.13	0.68±0.12	0.539

Variables are expressed as number (%), mean±SD or median (interquartile range). Abbreviations as in Table 1.

Table 4. Characteristics of the Study 2 Subjects Without Medication (n=78)

	Total	Male	Female	P value
n	78	29	49
Age (years)	62±11	62±10	62±11	0.990
BMI (kg/m2)	22.7±3.2	23.5±3.2	22.2±3.1	0.080
WC (cm)	80.9±10.1	85.4±8.6	78.3±10.1	0.003
Systolic BP (mmHg)	128±18	132±17	125±19	0.110
Diastolic BP (mmHg)	74±11	78±11	72±11	0.025
Pulse rate (beats/min)	70±11	69±14	72±10	0.404
Habitual smoking	11 (14.1)	6 (20.7)	5 (10.2)	0.199
Biochemical data
TC (mg/dL)	215±30	204±27	221±30	0.019
LDL-C (mg/dL)	126±28	122±28	129±27	0.331
HDL-C (mg/dL)	67±19	59±21	73±16	0.001
Triglycerides (mg/dL)	101 (69–127)	117 (84–163)	92 (65–113)	0.001
Fasting glucose (mg/dL)	93 (89–99)	94 (91–104)	92 (88–97)	0.033
Insulin (μU/mL)	5.4 (3.5–7.6)	5.4 (3.5–7.5)	5.8 (3.4–7.6)	0.871
HOMA-R	1.29 (0.78–1.85)	1.24 (0.79–1.90)	1.34 (0.75–1.83)	0.871
HbA1c (%)	5.1±0.2	5.1±0.3	5.0±0.2	0.150
BUN (mg/dL)	14±4	14±3	15±4	0.636
Creatinine (mg/dL)	0.7±0.1	0.8±0.1	0.7±0.1	<0.001
eGFR (mL/min/1.73 m2)	73.0±12.7	77.9±11.6	70.1±12.6	0.008
Uric acid (mg/dL)	4.9±1.2	5.7±1.1	4.4±1.0	<0.001
AST (IU/L)	22 (19–26)	24 (20–26)	22 (19–25)	0.597
ALT (IU/L)	19 (15–23)	20 (17–25)	18 (14–22)	0.201
γGTP (IU/L)	20 (16–32)	24 (19–36)	18 (14–28)	0.007
BNP (pg/mL)	15 (11–28)	12 (6–22)	18 (12–30)	0.413
hsCRP (mg/dL)	0.02 (0.02–0.05)	0.02 (0.02–0.05)	0.03 (0.01–0.06)	0.818
FABP4 (ng/mL)	10.7 (7.8–13.2)	9.2 (7.7–11.4)	11.2 (8.1–13.8)	0.064
Carotid ultrasonographic data
CIMT (mm)	0.70±0.13	0.72±0.12	0.69±0.13	0.323

Variables are expressed as number (%), mean±SD or medians (interquartile range). Abbreviations as in Table 1.

The distribution of changes in CIMT per year during the 3-year period is shown in Figure 2A, and the mean value was 3.8±22.3 (median: 1.7) µm/year. Progression and regression of CIMT were found in 43 (55.1%) and 35 (44.9%) of the subjects, respectively. There was no significant difference in CIMT per year between the male and female subjects (3.8±25.5 vs. 3.7±20.5 µm/year, P=0.992). No significant difference was found between changes in CIMT in subjects with and without a smoking habit (9.5±16.5 vs. 2.8±23.1 µm/year, P=0.251). The change in CIMT per year was positively correlated with logarithmically transformed levels of hsCRP (r=0.299, P=0.009) (Figure 2B) and FABP4 (r=0.267, P=0.018) (Figure 2C). No significant correlation was found between the change in CIMT per year and other basal parameters (Table 5).

Figure 2.

Progression of carotid atherosclerosis and the level of fatty acid-binding protein 4 (FABP4). (A) Histogram of change in carotid intima-media thickness (CIMT) per year (µm/year) in Study 2 (n=78). Open bar: males (n=29), closed bar: females (n=49). (B,C) Change in CIMT per year plotted against logarithmically transformed levels of high-sensitivity C-reactive protein (hsCRP) (B) and FABP4 (C) in each subject of Study 2 (n=78). Open circle: males (n=29), closed circle: females (n=49).

Table 5. Correlation Between Change in CIMT and Basal Clinical Variables (Study 2: n=78)

	r	P value
Age	−0.018	0.088
BMI	0.068	0.554
WC	0.057	0.621
Systolic BP	0.049	0.672
Diastolic BP	0.064	0.576
Pulse rate	0.060	0.691
TC	0.123	0.284
LDL-C	0.137	0.254
HDL-C	0.001	0.991
Log triglycerides	0.003	0.978
Log fasting glucose	0.181	0.113
Log insulin	0.055	0.634
Log HOMA-R	0.085	0.462
HbA1c	0.198	0.082
BUN	0.203	0.075
Creatinine	−0.067	0.562
eGFR	0.082	0.478
Uric acid	0.099	0.821
Log AST	−0.026	0.821
Log ALT	0.018	0.877
Log γGTP	0.017	0.886
Log BNP	−0.220	0.055
Log hsCRP	0.299	0.009
Log FABP4	0.267	0.018
CIMT	0.014	0.904

Abbreviations as in Table 1.

No significant change between the years 2010 and 2013 was found in the level of hsCRP (0.02 [0.02–0.05] vs. 0.02 [0.01–0.05] mg/dL, P=0.148) or FABP4 (median [IQR]: 10.7 [7.8–13.2] vs. 11.0 [7.3–15.3] ng/mL, P=0.112). The change in the level of FABP4 (r=−0.062, P=0.589) or hsCRP (r=−0.168, P=0.150) during the 3-year period was not significantly correlated with the change in CIMT.

Multiple regression analysis showed that the levels of FABP4 and hsCRP were independently associated with CIMT after adjustment for age and sex, explaining a total of 14.7% of the variance in this measure (R²=0.147) (Table 6).

Table 6. Multiple Regression Analysis for Change in CIMT (μm/year) (Study 2: n=78)

	Regression coefficient	SE	Standardized regression coefficient (β)	t	P value
Age	−0.308	0.242	−0.691	−0.20	0.208
Sex (male)	2.095	5.166	0.046	0.41	0.686
Log hsCRP	6.242	2.743	0.261	2.28	0.026
Log FABP4	15.713	7.508	0.250	2.09	0.040

R²=0.147. Abbreviations as in Table 1.

Discussion

The present study showed for the first time that the serum FABP4 concentration was an independent predictor of the progression of carotid atherosclerosis assessed by the change in CIMT per year during a 3-year period in a general population of subjects who had not taken any relevant medications. FABP4 is secreted from adipocytes and macrophages despite a lack of secretory signal peptides in the sequence of FABP4.^7–9 Previous studies using in vitro and in vivo experiments showed that FABP4 acts as an adipokine for the development of hepatic IR through increased hepatic glucose production,⁷ suppression of cardiomyocyte contraction,²³ and development of atherosclerosis through induction of inflammatory responses, inhibition of endothelial nitric oxide synthase activity in endothelial cells and proliferation and migration of vascular smooth muscle cells.⁹ Furthermore, a recent study demonstrated the possibility of a new strategy to treat metabolic disease by targeting serum FABP4 with a monoclonal antibody.¹⁰ Circulating FABP4 may directly promote the progression of atherosclerosis in humans.

Elevated serum FABP4 level is associated with atherosclerosis as assessed by several examinations, including cardio-ankle vascular index,²⁴ carotid-femoral pulse wave velocity,²⁵ CIMT^15,26,27 and plaque volume in the coronary artery measured by intravascular ultrasound.²⁸ In the present study, we confirmed that the serum FABP4 level was positively and independently correlated with CIMT, a morphological index of atherosclerosis. Several drugs, including angiotensin II receptor blockers,^24,29 a statin,³⁰ omega-3 fatty acid ethyl esters,³¹ a dipeptidyl peptidase-4 inhibitor,³² a sodium glucose cotransporter 2 inhibitor³³ and a thiazolidinedione,³⁴ have been reported to modulate the circulating FABP4 level. The serum FABP4 level has also been reported to predict long-term cardiovascular events.^16–18 Reduction of the FABP4 level would be a novel therapeutic strategy for preventing the development of atherosclerosis and reducing the incidence of atherosclerotic cardiovascular disease. However, it is unclear whether FABP4 acts by an intracellular signaling mechanism via unidentified receptors, though it is a potential adipokine, an adipocyte-derived bioactive molecule.^7,9,35,36 It has also been reported that extracellular FABP4 is partially internalized into cells,³⁶ though the mechanism and significance of internalization remain unclear. Further understanding of the mechanism of FABP4’s action may enable the development of new therapeutic strategies for atherosclerotic cardiovascular disease, such as neutralization of FABP4 and/or blockade of the FABP4 receptor, if any.

Treatment with several drugs for dyslipidemia, hypertension and DM has been shown to prevent the progression of atherosclerosis, indicated by regression of the CIMT.³⁷ Conversely, it has been reported that the average increase in CIMT per year is 30–40 µm in patients with T2DM.^38,39 However, little is known about the average change in CIMT in a general population without medications. A predictive increase in CIMT per year in healthy subjects has been reported to be approximately 8–9 µm according to the regression coefficient with age in cross-sectional analyses,^40,41 because there was a strong positive correlation between age and CIMT. In the present study, the distribution of the change in CIMT during the 3-year period was relatively large, and progression and regression of CIMT were found in 43 (55.1%) and 35 (44.9%), respectively, of the 78 studied subjects (Figure 2A). The average value of the change in CIMT was 3.8±22.3 (median: 1.7) µm/year and was smaller than the predictive value, suggesting a variety of alterations in CIMT caused by several factors, including modification of lifestyle, foods, salt intake, smoking habit and other factors, over a certain period of time in a general population without medications. The present study revealed that the basal levels of hsCRP and FABP4, but not the change in hsCRP or FABP4 during the 3-year period, were independent predictors of the change in CIMT during the 3-year period after adjustment for age and sex.

Study Limitations

First, the number of patients enrolled, especially in Study 2, was small, and the possibility of type 1 or type 2 errors in statistical tests cannot be excluded. Second, only CIMT including plaque in the common carotid arteries was evaluated in the present study, and the use of different methods to measure CIMT or different parameters may yield different results. Third, errors of measurement of CIMT at 2 time points may affect the change in CIMT per year, though the coefficients of variance of intra- and interoperator variability for semi-automated CIMT measurement (3.5% and 8.0%, respectively) in our laboratory were better than those for manual CIMT measurement (14.1% and 16.7%, respectively).²² Fourth, some medications were started in 15 of the 78 subjects in Study 2 and might have modulated the progression and regression of atherosclerosis, though the results of analyses were similar when those 15 subjects were excluded (data not shown). Lastly, because the recruited subjects were only Japanese, it is unclear whether the present findings can be generalized to other ethnicities.

Conclusions

FABP4 concentration is an independent predictor of atherosclerosis assessed by CIMT and the progression of carotid atherosclerosis. A further understanding of the mechanism underlying the link between circulating FABP4 and progression of atherosclerosis may enable the development of new therapeutic strategies for cardiovascular and metabolic diseases.

Acknowledgments

M.F. was supported by grants from JSPS KAKENHI, AMED Translational Research Network Program, Uehara Memorial Foundation, SENSHIN Medical Research Foundation, Japan Diabetes Foundation, Takeda Medical Research Foundation, Ono Medical Research Foundation, Takeda Science Foundation, Akiyama Life Science Foundation, Yamaguchi Endocrine Research Foundation, Naito Foundation Natural Science Scholarship, Suhara Memorial Foundation, Kondou Kinen Medical Foundation and Terumo Foundation for Life Science and Arts.

Conflict of Interest

The authors declare no conflict of interest.

References

• 1.   Furuhashi M, Hotamisligil GS. Fatty acid-binding proteins: Role in metabolic diseases and potential as drug targets. Nat Rev Drug Discov 2008; 7: 489–503.
• 2.   Furuhashi M, Ishimura S, Ota H, Miura T. Lipid chaperones and metabolic inflammation. Int J Inflam 2011; 2011: 642612.
• 3.   Furuhashi M, Saitoh S, Shimamoto K, Miura T. Fatty acid-binding protein 4 (FABP4): Pathophysiological insights and potent clinical biomarker of metabolic and cardiovascular diseases. Clin Med Insights Cardiol 2014; 8: 23–33.
• 4.   Furuhashi M, Fucho R, Gorgun CZ, Tuncman G, Cao H, Hotamisligil GS. Adipocyte/macrophage fatty acid-binding proteins contribute to metabolic deterioration through actions in both macrophages and adipocytes in mice. J Clin Invest 2008; 118: 2640–2650.
• 5.   Furuhashi M, Tuncman G, Gorgun CZ, Makowski L, Atsumi G, Vaillancourt E, et al. Treatment of diabetes and atherosclerosis by inhibiting fatty-acid-binding protein aP2. Nature 2007; 447: 959–965.
• 6.   Xu A, Wang Y, Xu JY, Stejskal D, Tam S, Zhang J, et al. Adipocyte fatty acid-binding protein is a plasma biomarker closely associated with obesity and metabolic syndrome. Clin Chem 2006; 52: 405–413.
• 7.   Cao H, Sekiya M, Ertunc ME, Burak MF, Mayers JR, White A, et al. Adipocyte lipid chaperone AP2 is a secreted adipokine regulating hepatic glucose production. Cell Metab 2013; 17: 768–778.
• 8.   Mita T, Furuhashi M, Hiramitsu S, Ishii J, Hoshina K, Ishimura S, et al. FABP4 is secreted from adipocytes by adenyl cyclase-PKA- and guanylyl cyclase-PKG-dependent lipolytic mechanisms. Obesity (Silver Spring) 2015; 23: 359–367.
• 9.   Furuhashi M, Fuseya T, Murata M, Hoshina K, Ishimura S, Mita T, et al. Local production of fatty acid-binding protein 4 in epicardial/perivascular fat and macrophages is linked to coronary atherosclerosis. Arterioscler Thromb Vasc Biol 2016; 36: 825–834.
• 10.   Burak MF, Inouye KE, White A, Lee A, Tuncman G, Calay ES, et al. Development of a therapeutic monoclonal antibody that targets secreted fatty acid-binding protein aP2 to treat type 2 diabetes. Sci Transl Med 2015; 7: 319ra205.
• 11.   Ishimura S, Furuhashi M, Watanabe Y, Hoshina K, Fuseya T, Mita T, et al. Circulating levels of fatty acid-binding protein family and metabolic phenotype in the general population. PLoS One 2013; 8: e81318.
• 12.   Ota H, Furuhashi M, Ishimura S, Koyama M, Okazaki Y, Mita T, et al. Elevation of fatty acid-binding protein 4 is predisposed by family history of hypertension and contributes to blood pressure elevation. Am J Hypertens 2012; 25: 1124–1130.
• 13.   Furuhashi M, Omori A, Matsumoto M, Kataoka Y, Tanaka M, Moniwa N, et al. Independent link between levels of proprotein convertase subtilisin/kexin type 9 and FABP4 in a general population without medication. Am J Cardiol 2016; 118: 198–203.
• 14.   Fuseya T, Furuhashi M, Yuda S, Muranaka A, Kawamukai M, Mita T, et al. Elevation of circulating fatty acid-binding protein 4 is independently associated with left ventricular diastolic dysfunction in a general population. Cardiovasc Diabetol 2014; 13: 126.
• 15.   Yeung DC, Xu A, Cheung CW, Wat NM, Yau MH, Fong CH, et al. Serum adipocyte fatty acid-binding protein levels were independently associated with carotid atherosclerosis. Arterioscler Thromb Vasc Biol 2007; 27: 1796–1802.
• 16.   Furuhashi M, Ishimura S, Ota H, Hayashi M, Nishitani T, Tanaka M, et al. Serum fatty acid-binding protein 4 is a predictor of cardiovascular events in end-stage renal disease. PLoS One 2011; 6: e27356.
• 17.   von Eynatten M, Breitling LP, Roos M, Baumann M, Rothenbacher D, Brenner H. Circulating adipocyte fatty acid-binding protein levels and cardiovascular morbidity and mortality in patients with coronary heart disease: A 10-year prospective study. Arterioscler Thromb Vasc Biol 2012; 32: 2327–2335.
• 18.   Chow WS, Tso AW, Xu A, Yuen MM, Fong CH, Lam TH, et al. Elevated circulating adipocyte-fatty acid binding protein levels predict incident cardiovascular events in a community-based cohort: A 12-year prospective study. J Am Heart Assoc 2013; 2: e004176.
• 19.   Agardh HE, Folkersen L, Ekstrand J, Marcus D, Swedenborg J, Hedin U, et al. Expression of fatty acid-binding protein 4/aP2 is correlated with plaque instability in carotid atherosclerosis. J Intern Med 2011; 269: 200–210.
• 20.   Doi M, Miyoshi T, Hirohata S, Nakamura K, Usui S, Takeda K, et al. Association of increased plasma adipocyte fatty acid-binding protein with coronary artery disease in non-elderly men. Cardiovasc Diabetol 2011; 10: 44.
• 21.   Peeters W, de Kleijn DP, Vink A, van de Weg S, Schoneveld AH, Sze SK, et al. Adipocyte fatty acid binding protein in atherosclerotic plaques is associated with local vulnerability and is predictive for the occurrence of adverse cardiovascular events. Eur Heart J 2011; 32: 1758–1768.
• 22.   Kaneko R, Yuda S, Osawa H, Sakai E, Yokoya R, Fujita M, et al. Usefulness of the semi-automated edge-detection software for measurement of carotid intima-media thickness. Jpn J Med Ultrasound Technol 2011; 36: 343–348.
• 23.   Lamounier-Zepter V, Look C, Alvarez J, Christ T, Ravens U, Schunck WH, et al. Adipocyte fatty acid-binding protein suppresses cardiomyocyte contraction: A new link between obesity and heart disease. Circ Res 2009; 105: 326–334.
• 24.   Miyoshi T, Doi M, Hirohata S, Kamikawa S, Usui S, Ogawa H, et al. Olmesartan reduces arterial stiffness and serum adipocyte fatty acid-binding protein in hypertensive patients. Heart Vessels 2011; 26: 408–413.
• 25.   Tsai JP, Wang JH, Lee CJ, Chen YC, Hsu BG. Positive correlation of serum adipocyte fatty acid binding protein levels with carotid-femoral pulse wave velocity in geriatric population. BMC Geriatr 2015; 15: 88.
• 26.   Hao Y, Ma X, Luo Y, Shen Y, Dou J, Pan X, et al. Serum adipocyte fatty acid binding protein levels are positively associated with subclinical atherosclerosis in Chinese pre- and postmenopausal women with normal glucose tolerance. J Clin Endocrinol Metab 2014; 99: 4321–4327.
• 27.   Furuhashi M, Ogura M, Matsumoto M, Yuda S, Muranaka A, Kawamukai M, et al. Serum FABP5 concentration is a potential biomarker for residual risk of atherosclerosis in relation to cholesterol efflux from macrophages. Sci Rep 2017; 7: 217.
• 28.   Miyoshi T, Onoue G, Hirohata A, Hirohata S, Usui S, Hina K, et al. Serum adipocyte fatty acid-binding protein is independently associated with coronary atherosclerotic burden measured by intravascular ultrasound. Atherosclerosis 2010; 211: 164–169.
• 29.   Furuhashi M, Mita T, Moniwa N, Hoshina K, Ishimura S, Fuseya T, et al. Angiotensin II receptor blockers decrease serum concentration of fatty acid-binding protein 4 in patients with hypertension. Hypertens Res 2015; 38: 252–259.
• 30.   Karpisek M, Stejskal D, Kotolova H, Kollar P, Janoutova G, Ochmanova R, et al. Treatment with atorvastatin reduces serum adipocyte-fatty acid binding protein value in patients with hyperlipidaemia. Eur J Clin Invest 2007; 37: 637–642.
• 31.   Furuhashi M, Hiramitsu S, Mita T, Omori A, Fuseya T, Ishimura S, et al. Reduction of circulating FABP4 level by treatment with omega-3 fatty acid ethyl esters. Lipids Health Dis 2016; 15: 5.
• 32.   Furuhashi M, Hiramitsu S, Mita T, Fuseya T, Ishimura S, Omori A, et al. Reduction of serum FABP4 level by sitagliptin, a DPP-4 inhibitor, in patients with type 2 diabetes mellitus. J Lipid Res 2015; 56: 2372–2380.
• 33.   Furuhashi M, Matsumoto M, Hiramitsu S, Omori A, Tanaka M, Moniwa N, et al. Possible increase in serum FABP4 level despite adiposity reduction by canagliflozin, an SGLT2 inhibitor. PLoS One 2016; 11: e0154482.
• 34.   Cabre A, Lazaro I, Girona J, Manzanares JM, Marimon F, Plana N, et al. Fatty acid binding protein 4 is increased in metabolic syndrome and with thiazolidinedione treatment in diabetic patients. Atherosclerosis 2007; 195: e150–e158.
• 35.   Yamamoto T, Furuhashi M, Sugaya T, Oikawa T, Matsumoto M, Funahashi Y, et al. Transcriptome and metabolome analyses in exogenous FABP4- and FABP5-treated adipose-derived stem cells. PLoS One 2016; 11: e0167825.
• 36.   Fuseya T, Furuhashi M, Matsumoto M, Watanabe Y, Hoshina K, Mita T, et al. Ectopic fatty acid-binding protein 4 expression in the vascular endothelium is involved in neointima formation after vascular injury. J Am Heart Assoc 2017; 6: e006377.
• 37.   de Groot E, van Leuven SI, Duivenvoorden R, Meuwese MC, Akdim F, Bots ML, et al. Measurement of carotid intima-media thickness to assess progression and regression of atherosclerosis. Nat Clin Pract Cardiovasc Med 2008; 5: 280–288.
• 38.   Yamasaki Y, Kodama M, Nishizawa H, Sakamoto K, Matsuhisa M, Kajimoto Y, et al. Carotid intima-media thickness in Japanese type 2 diabetic subjects: Predictors of progression and relationship with incident coronary heart disease. Diabetes Care 2000; 23: 1310–1315.
• 39.   Yokoyama H, Katakami N, Yamasaki Y. Recent advances of intervention to inhibit progression of carotid intima-media thickness in patients with type 2 diabetes mellitus. Stroke 2006; 37: 2420–2427.
• 40.   Handa N, Matsumoto M, Maeda H, Hougaku H, Ogawa S, Fukunaga R, et al. Ultrasonic evaluation of early carotid atherosclerosis. Stroke 1990; 21: 1567–1572.
• 41.   Yamasaki Y, Kawamori R, Matsushima H, Nishizawa H, Kodama M, Kajimoto Y, et al. Atherosclerosis in carotid artery of young IDDM patients monitored by ultrasound high-resolution B-mode imaging. Diabetes 1994; 43: 634–639.