This article is mentioned in:

Introduction

Atherosclerosis (AS) of the carotid arteries is a major cause of stroke and transient ischemic attack, which remains the leading cause of mortality in China and is currently the most common cause of mortality worldwide (1,2). The symptoms of AS are not always detected by patients until complications arise and the therapeutic methods to treat AS are poor; therefore, it is particularly important to establish an appropriate, rigorous and efficient detection strategy for early-phase AS, in order to limit disease progression before it results in clinical consequences (3). There are two prevalent methods for diagnosing AS: Doppler ultrasound and computed tomographic angiography (CTA), which have become the standard methods for diagnosing AS (4,5). However, neither of these methods identify the early asymptomatic pathological changes of AS and ultrasound only assesses plaque in the carotid artery. Therefore, it is paramount to develop non-invasive methods for diagnosing high-risk, asymptomatic individuals before the onset of clinical events or symptoms (6,7). Furthermore, data supporting the routine use of Doppler ultrasound to screen for carotid stenosis in an asymptomatic population is considered to be weak, as there is a low overall prevalence of treatable AS in the general asymptomatic population (8). It is widely hypothesized that the next generation of screening tests will be based on molecular biomarkers (9).

DNA methylation, a stable epigenetic mark, is a non-traditional, heritable factor involved in gene transcription regulation (10). DNA methylation occurs at position 5′ of CpG dinucleotides; certain regions of DNA are rich in CpG and are termed CpG islands (GCIs), which predominantly locate in the promoter region (11). Changes in patterns of DNA methylation may lead to inappropriate gene expression, and contribute to the development and progression of disease (12). In addition, previous studies have demonstrated aberrant DNA methylation in AS (13). Alterations in DNA methylation have been identified as an early event in AS; Zhao et al (14) found that aberrant methylation may represent an early biomarker for AS, and demonstrated clinical correlations with carotid intima-media thickness (14). Current research primarily focuses on the effect of DNA methylation leading to AS and, to the best of our knowledge, there are few studies that investigate the role of DNA methylation on the early diagnosis of AS. Detection of DNA methylation, at candidate loci in AS, suggests that AS-specific methylation changes could be applied diagnostically. However, DNA methylation affects various factors, including age, living environment, diet and other factors limit the use of methylation changes for diagnosis. Previous studies have shown that simultaneous measure various differentially methylated loci may improve diagnostic use (15). Furthermore, a major limitation towards further developments of DNA methylation in clinical use may be that the majority of studies focus on single genes (16), whereas AS is associated with multiple factors, including genetic and epigenetic alterations (17). Measuring the methylation of individual genes was not as specific or sensitive as using a panel of epigenetic markers for the early detection of breast cancer (18).

In the present study, DNA methylation was profiled in AS and healthy groups of 300 patients using nested methylation-specific polymerase chain reaction (PCR; NT-MSP). Based in previous unpublished data, a biomarker panel capable of differentiating AS patients from healthy individuals, irrespective of AS histology, was identified. The present study provides insight into AS etiology, presents validated tissue-based diagnostic biomarkers, and supplies a framework for the development of DNA methylation-based molecular diagnostics for AS detection in patients.

Materials and methods

Ethics statement

The present study was reviewed and approved by the Ethics Committee of General Hospital of Ningxia Medical University (Yinchuan, China). Written informed consent was obtained from the participants before the collection of any samples and the specimens were irreversibly de-identified.

Inclusion and exclusion criteria

Patients with AS were diagnosed by Doppler ultrasound examination of the carotid artery. The healthy control group comprised of age- and gender-matched healthy individuals that were not exhibiting AS risk factors. The exclusion criteria were as follows: Individuals with moderate or severe valvular heart disease, chronic renal failure or diabetes; smokers; obese individuals; and those exhibiting disordered lipid metabolism.

NT-MSP analysis of DNA methylation

The following candidate genes formed the panel for early detection of AS: ATP binding cassette subfamily A member 1 (ABCA1), acetyl-CoA acetyltransferase 1 (ACAT1), peroxisome proliferator activated receptor α (PPARA), platelet derived growth factor subunit B (PDGFB), arachidonate 5-lipoxygenase (ALOX5), LDL receptor related protein 1 (LRP1), lecithin-cholesterol acyltransferase (LCAT), TIMP metallopeptidase inhibitor 1 (TIMP1), matrix metallopeptidase 9 (MMP9), intercellular adhesion molecule 1 (ICAM1), vascular endothelial growth factor A (VEGFA) and nuclear factor of κ light polypeptide gene enhancer in B-cells 1 (NFKB1). These genes contained CGIs, were expressed in peripheral blood cells, and were correlated with the occurrence and development of AS (19–24). DNA methylation of these candidate genes was examined in two independent test sets (total of 300 peripheral blood samples) and genomic DNA was isolated from the peripheral blood mononuclear cells using the Wizard Genomic DNA Purification kit (Promega Corporation, Madison, WI, USA). An EZ DNA Methylation-Gold™ kit (Zymo Research Corporation, Irvine, CA, USA) was used to detect the DNA methylation patterns, and integrates the DNA denaturation and bisulfite conversion processes into one step. The DNA was modified by sodium bisulfite (Zymo Research Corporation, Irvine, CA, USA) in which unmethylated cytosine residues were converted to thymine, whereas methylated cytosine residues were retained as cytosine; this difference was then utilized to specifically amplify either methylated or unmethylated DNA. NT-MSP, which consists of a two-step PCR amplification, was then used to detect methylation (25). The first step of NT-MSP uses an outer primer pair, which does not contain any CpGs. The second step of PCR was conducted with conventional PCR primers serving as inner primers. The primers for NT-MSP amplification were designed according to the bioinformatics program, MethPrimer (www.urogene.org/methprimer/index.html) and are presented in Table I. PCR products were gel purified with an agarose gel DNA fragment recovery kit, according to the manufacturer's instructions, and were sequenced by Invitrogen (Thermo Fisher Scientific, Inc., Waltham, MA, USA). To reduce mispriming and to increase efficiency, touchdown (TD) PCR was used for amplification. Samples were subjected to 30 cycles in a TD program (94°C for 5 min, 94°C for 30 sec; 56°C for 30 sec and 72°C for 1 min, followed by a decrease of 0.5°C in the annealing temperature every second cycle). After completion of the TD program, 20 cycles were subsequently run (94°C for 30 sec, 56°C for 30 sec and 72°C for 60 sec), terminating with a 7-min extension at 72°C. The PCR products were separated by electrophoresis through a 2% agarose gel (Borunlaite Science & Technology, Beijing, China) containing ethidium bromide (Tokyo Chemical Industry Co., Ltd., Shanghai, China) for 30 min at 100 V. The DNA bands were visualized using ultraviolet light. The following formula was used: Methylation %= methylation / (methylation + unmethylation) ×100%.

Table I Primer sequences for DNA methylation analysis.

Table I

Primer sequences for DNA methylation analysis.

Gene	Primer sequence (5′-3′)
	Forward	Reverse
ABCA1
Outer	GAAAGTAGGATTTAGAGGAAGTAAAT	AATTCAATCACTCAACAAAAAACAC
Methylation	AATTTTATTGGTGTTTTTGGTTGTC	ATATCTTAAAATCCGCGATCTACG
Unmethylation	AATTTTATTGGTGTTTTTGGTTGTT	CAAATATCTTAAAATCCACAATCTA
ACAT1
Outer	AGGAAGATTTAGAGATTTAGAAGTAA	TCAATAATCACTCAAACACACAA
Methylation	TTAATTAGTTTGGTTTTTGGTCGTT	CTTATATACCAAATGCGAACT
Unmethylation	TTAATTATGTGTGTTTGGTTTTTTG	TATCCAAATTAACCAAACTCCATA
PDGFB
Outer	TTTTTTTGTTTTGAAATTTTGGTTAA	CAAAATCCCAACAAAAATCTCC
Methylation	TTTGGAAATTAATGATAAGTTAGGC	AAGCATCATAAAAAAAACGCATC
Unmethylation	TTGGAAATTAATGATAAGTTAGGTGA	AAACATCATAAAAAAAACACATCA
ALOX5
Outer	TTATATTTTCGTCGTTTTACGTACG	AATATAAAAAAAATTTCGCGCG
Methylation	TATATTTTTGTTGTTTTATGTATGG	AAATATAAAAAAAATTACACACT
Unmethylation	AGATGTTATAGGGATTTTGTTGTTT	TCTAAAACACAAAAACTCTCAAC
LRP1
Outer	TTTGTTGTTTAGGTTGGAGTGTAGT	TAAAAAAATCACTTTTAAAAATTC
Methylation	TGTTTAGGTTGGAGTGTAGTGGTAC	AAAAAATCACTTTTCAAAATTCGA
Unmethylation	TGTTTAGGTTGGAGTGTAGTGGTAT	AAAAATCACTTTTCAAAATTCAAA
LCAT
Outer	GAGTTTTTTGTTTTGTTTTGGTTTC	AATCTTAATCAATCTATCTACCGC
Methylation	GAGTTTTTTGTTTTGTTTTGGTTTT	ATCTTAATCAATCCATCTACCACA
Unmethylation	TTTGGTTGTTTTTTTGTATTATTTG	TCAATCTTAATCAACTCATCTACC
TIMP1
Outer	TTTTTTTTAATGTTTATTTATTTATT	CCAAACACTCACAATTTATCTCAC
Methylation	TTAGGTAGTTTTTGTTTGAATTTCG	ATAAATACCGTTCTTATTCCCGTT
Unmethylation	TAGGTAGTTTTTGTTTGAATTTTGG	ATAAATACCAATTCTTATTCCCATT
MMP9
Outer	GGTTATATAGTTGGAAATGGTAGAGT	ATTCAAACAATTCTCTACCTCAAC
Methylation	TTTTAAATATAGTTTATTGGGTCGG	TTTAATAAAAATAAAATTTCACTA
Unmethylation	TTTTTTAAATATAGTTTATTGGGTTGG	AATAAAAATAAATTTCACTACATT
ICAM1
Outer	GTAAGAGTTTAGTGGAATTCGTTCG	GAATCACTAACCATCCAAAAACG
Methylation	TAAGAGTTTAGTGGAATTTGTTTGA	AAATCACTAACCATCCAAAAACAC
Unmethylation	TAAGAGTTTAGTGGAATTTGTTTGA	CCTAAAACTTTCCTATTATAAAAC
VEGFA
Outer	TAGTTAGAGTCGGGGTGTGTAGAC	GAAAAACCGAACAAAAACGAA
Methylation	TAGTTAGAGTTGGGGTGTGTAGATG	AACAAAAAACCAACAAAAACAAA
Unmethylation	GGAGTAAATTTTTTTTATTTTTTTT	ATTGTGGTTTTTGGTTTAGTTTTG
PPARA
Outer	AGTATAGTGGTATAGGGTATTGGTAG	TAAAACTCTACAAAACAAAAAAAA
Methylation	TAGTGGTAGGTATAGTTGGTAGCGG	ACCAATAACGAAAATAAAAAAAC
Unmethylation	TAGTGGTAGGTATAGTTGGTAGTGG	CAATAACAAAATAAAAAAACACC
NFKB1
Outer	TATAGATGAGTTTTATTTATTTGGTA	AAACTCTAACTCCTAACAAAAC
Methylation	TTGATTGGGTTCGGTAGGC	GCACTTCTAAAAAGCTATACGCC
Unmethylation	TTATTGATTGGGTTTGGTAGGT	CCCACACTTCTAACACTATACACC

Reverse transcription-quantitative PCR (RT-qPCR) for mRNA

The total RNA was extracted from the peripheral blood mononuclear cells using Invitrogen TRIzol reagent (Thermo Fisher Scientific, Inc.). Primer Premier 5 software was used to design the primers (Table II). The expression levels of candidate genes were normalized to those of glyceraldehyde-3-phosphate dehydrogenase (GAPDH): Forward, 5′-AGAAGGCTGGGGCTCATTTG-3′, and reverse, 5′-AGGGGCCATCCACAGTCTTC-3′. RNA was reverse transcribed using the PrimeScript Real-Time PCR reagent kit (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions. Total RNA then underwent cDNA synthesis and PCR amplification using the PrimeScript Real-Time PCR reagent kit according to the manufacturer's instructions. The qPCR reaction was performed following the manufacturer's instructions, as follows: Denaturation at 94°C for 10 min, 50 cycles at 94°C for 15 sec, annealing at 53°C for 30 sec and extension at 72°C for 30 sec. The RNA level of every gene was acquired from the value of the quantitation cycle (Cq) of the qPCR normalized to that of GAPDH using the following formula: ΔCq = CqGAPDH − Cqgene. The final results were expressed as the N-fold differences in the target gene expression and was relative to the calibrator, termed Ntarget, which was determined as follows: Ntarget =2ΔCq(sample)−ΔCq(calibrator), where ΔCq values of the calibrator and sample were determined by subtracting the Cq value of the target gene from the Cq value of GAPDH.

Table II Primers used in quantitative polymerase chain reaction for mRNA analysis.

Table II

Primers used in quantitative polymerase chain reaction for mRNA analysis.

Gene	Primer	Sequence (5′-3′)	Product size (bp)
ACAT1	Left	ACGCTGCTGTAGAACCTATTGA	116
	Right	GGCTTCATTTACTTCCCACATT
ABCA1	Left	ATCAAGGGCATCGTGTATGAG	227
	Right	AGGATTGTCACCACAGCAAAC
PPARA	Left	ATCACGGAACACGCTTTCAC	100
	Right	CGATGTTCAATGCTCCACTG
TIMP1	Left	ATACTTCCACAGGTCCCACAAC	194
	Right	GGATGGATAAACAGGGAAACAC
ALOX5	Left	AGTTCCAGCAAGGGAACATT	204
	Right	CATCCGAAGGGAGGAAAATAG

Statistical analysis

The patient characteristics were compared with those of the control subjects using the χ2 test and the independent-samples t-test was used to detect significantly different methylation levels between the AS patients and matched controls. Receiver-operating-characteristics (ROC) curves were calculated to evaluate the diagnostic performance of different marker combinations. Furthermore, sensitivity and specificity analyses were performed to assess different marker combinations for AS detection. P<0.05 was considered to indicate a statistically significant difference.

Results

Doppler ultrasound of carotid artery samples

All participants in the test and validation sets were verified by Doppler ultrasound. Images of carotid arteries of the healthy control subjects are presented in Fig. 1A, the samples demonstrate intima-media with no plaques and normal blood flow. Conversely, Doppler flow imaging of the patients with AS demonstrated anterior wall filling defects in the carotid artery, as shown in Fig. 1B and C. Furthermore, the plaque area was 0.23×2.23 cm in Fig. 1B and 0.47×3.59 cm in Fig. 1C.

Figure 1 Carotid artery plaques detected by Doppler ultrasound. (A) Carotid artery in the healthy control group. (B) Carotid atherosclerosis plaque sonograms from patients with (B) minor (plague area, 0.23×2.23 cm) and (C) severe (plague area, 0.47×3.59 cm) lesions. Sonograms were performed on the right side of the carotid artery to atherosclerosis plaques. Plaques are demonstrated by the red arrows.

Subject recruitment and collection of blood samples

A total of 300 peripheral blood samples were assessed in this case-control study, which included samples from AS patients and healthy control subjects. Between June 2011 and March 2012 150 patients and 150 healthy subjects (who had undergone Doppler ultrasound) were enrolled from the General Hospital of Ningxia Medical University for the test study. All participants were human immunodeficiency virus-, hepatitis B virus-, and hepatitis B virus-negative and did not exhibit inflammation or liver and kidney diseases. Furthermore, they did not have any previous history of cancer and were not pregnant. AS is often accompanied by lipid abnormalities; in the present study, AS patients demonstrated higher levels of the following factors: Total cholesterol (TC; P=0.0082), total triglycerides (TG; P=0.0091), high-density lipoprotein (HDL; P=0.0063), apolipoprotein A (P=0.0895) and apolipoprotein B (P=0.0739) compared with the matched control subjects, however, the low-density lipoprotein (LDL) is decreased. The two groups were matched according to gender, age, smoking habit and alcohol consumption; Table III presents the clinical profiles of the patients.

Table III Demographic characteristics of study subjects.

Table III

Demographic characteristics of study subjects.

Characteristic	Groups		P-value
	Matched control (%)	AS (%)
Gender ratio (male/female)	70/80	60/90	0.1359
Age (years)
30–40	63	18
41–49	55	50	0.3791
51–60	32	87
Alcohol	25	45	0.2852
Smoking	48	55	0.5721
Total triglyceride	6 (4)	66 (44)	0.0091
Total cholesterol	12 (8)	39 (26)	0.0082
High-density lipoprotein	6 (4)	88 (72)	0.0063
Low-density lipoprotein	45 (30)	25 (41)	0.0489
Apolipoprotein A	81 (54)	99 (66)	0.0894
Apolipoprotein B	24 (36)	75 (51)	0.0739

[i] P<0.05 was considered to indicate a statistically significant difference. AS vs. controls (n=150). AS, atherosclerosis.

Blood lipid levels

The 300 samples were evaluated using an automatic biochemical analyzer and the blood lipids levels are presented in Fig. 2. Compared with the healthy control group, the concentrations of TG and TC increased by 2.4- and 1.2-fold, respectively in the AS patients (P<0.01 and P<0.05), while, the HDL level was reduced by 43.8% in the AS group (P<0.05).

Figure 2 Blood lipid concentrations. *P<0.05, **P<0.01 vs. control group. TG, triglyceride; TC, total cholesterol; HDL, high-density lipoprotein; LDL, low-density lipoprotein.

Methylation frequencies of target genes in the AS and matched control subjects in the test set

DNA methylation of ABCA1, ACAT1, PPARA, PDGFB, ALOX5, LRP1, LCAT, TIMP1, MMP9, ICAM1, VEGFA and NFKB1 were initially assessed by NT-MSP (Fig. 3; Table IV). The figure demonstrates the patterns of hyper- and hypomethylated genes in the AS patients and matched control subjects. Methylation frequencies of ACAT1 (P<0.01), ALOX5 (P<0.05) and PPARA (P<0.05) were decreased whereas TIMP1 (P<0.01) and ABCA1 (P<0.01) were significantly increased in AS compared the matched controls. However, no statistically significant differences were identified between the remaining genes (PDGFB, LRP1, LCAT, MMP9, ICAM1, VEGFA and NFKB1) in the AS patients and matched control subjects, thus, these genes were excluded. The methylation of TIMP1, ACAT1 (P<0.01), ABCA1, ALOX5 and PPARA were significantly different in AS; thus, these genes may be used as a clinical tool for the early diagnosis of AS.

Figure 3 Methylation of AS-associated genes. Nested touchdown methylation-specific polymerase chain reaction was performed following bisulfite modification of DNA from samples of peripheral blood from the healthy and AS groups. Lane M, methylation; lane U, unmethylation; lane A, healthy group; lane B, AS group. AS, atherosclerosis.

Table IV Methylation levels of AS-associated genes in the two groups.

Table IV

Methylation levels of AS-associated genes in the two groups.

Gene	Healthy group	AS group	P-value
ACAT1	0.6103±0.01082	0.4357±0.02562	0.0024a
ICAM1	0.2812±0.02119	0.2530±0.02335	0.3722
ALOX5	0.3091±0.01589	0.2770±0.00573	0.0325b
LRP1	0.6977±0.01104	0.6893±0.01261	0.6213
MMP9	0.6246±0.02744	0.5585±0.02754	0.0944
TIMP1	0.5103±0.00276	0.5924±0.00782	0.0031a
VEGFA	0.5607±0.00841	0.5565±0.01386	0.8073
PDGDB	0.5493±0.01008	0.5217±0.01175	0.0812
ABCA1	0.4855±0.006517	0.6708±0.01609	0.0011a
LCAT	0.7087±0.01502	0.6827±0.01682	0.2544
PPARA	0.9944±0.00215	0.9786±0.00147	0.0266b
NFKB1	0.5709±0.01952	0.5631±0.007049	0.7090

a P<0.01

b P<0.05 vs. healthy group. Values are presented as the mean + standard deviation. AS, atherosclerosis.

Aberrant methylation profiles in AS

Among the 12 genes, the methylation frequencies of five genes, TIMP1, ACAT1, ABCA1, ALOX5 and PPARA, were significantly different between the AS patients and the matched control subjects. ROC curves were constructed for each of the five genes to classify AS patients and the matched control subjects. The area under the curve (AUC) of the ROC curve for TIMP1 was 0.891 (P<0.0001), which was the largest among the five genes. The sensitivity and specificity of TIMP1 were 78 and 84%, respectively, in the diagnosis of AS (Table V). When used separately to diagnose AS, the sensitivity of each gene ranged from 44 to 78% and the specificity ranged from 60 to 84%, the AUC of the ROC curves for the other four genes (ACAT1, ABCA1, ALOX5 and PPARA) ranged from 0.547 to 0.830.

Table V Diagnostic performance of five genes in AS group versus the matched controls in the test set.

Table V

Diagnostic performance of five genes in AS group versus the matched controls in the test set.

Gene	AS group		Matched control group		AUC	PPV	NPV	P-value
	Sensitivity (%)	Pos./total	Specificity (%)	Pos./total
TIMP1	78	117/150	84	24/150	0.891	0.83	0.79	<0.0001
ACAT1	68	102/150	84	24/150	0.830	0.81	0.72	<0.0001
ABCA1	62	90/150	78	33/150	0.711	0.74	0.67	<0.0001
ALOX5	48	72/150	60	60/150	0.547	0.55	0.54	0.0325
PPARA	44	66/150	72	42/150	0.535	0.61	0.56	0.0380

[i] AS, atherosclerosis; AUC, area under the curve; Pos., presented positive; PPV, positive predictive value; NPV, negative predictive value. n=150.

Combining various markers is a common strategy to improve diagnostic sensitivity when investigating clinical biomarkers. Based on the ROC curve for each gene, the most sensitive for the diagnosis of AS was TIMP1, which was found to be methylated in 117 of the 150 cases of AS patients, displaying a sensitivity of 78%; TIMP1 was also identified in 126 of the 150 cases of matched controls, displaying a specificity of 84% (Table V). The more frequently methylated gene was ACAT1 (68%; 102/150), followed by ABCA1 (62%; 93/150), ALOX5 (48%; 72/150) and PPARA (44%; 66/150). The AUC of ALOX5 (0.547) and PPARA (0.535) was the smallest among the five genes, and the methylation frequencies of ALOX5 and PPARA demonstrated no significant difference between the AS and matched control groups, ranging from 44 to 48% and from 60 to 72% in the in AS and matched control groups, respectively. Three genes, TIMP1, ACAT1 and ABCA1, were highly specifically methylated in AS; therefore, these constituted the first combination (P<0.01), which was the highest among the five genes for the diagnosis of AS. According to this analysis, combinations of different markers were examined to maximize specificity and sensitivity (Table VI). The combination of TIMP1, ACAT1 improved the sensitivity to 84%, which is higher than TIMP1 or ACAT1 alone, while the specificity dropped to 78%. Using this method, a total of three panels of genes were analyzed. The sensitivity and specificity of the three-gene panel (TIMP1, ACAT1 and ABCA1) was 88 and 90%, respectively and thus was identified to be the optimal combination of markers in the present study.

Table VI Diagnostic performance of different panels in the AS and matched groups.

Table VI

Diagnostic performance of different panels in the AS and matched groups.

Genes	AS group		Matched control group		PPV	NPV
	Sensitivity (%)	Pos./total	Specificity (%)	Pos./total
ACAT1, TIMP1	84	126/150	78	33/150	0.79	0.81
ABCA1, TIMP1	76	114/150	94	9/150	0.93	0.80
ACAT1, ABCA1	76	114/150	74	39/150	0.75	0.76
ACAT1, ABCA1, TIMP1	88	132/150	90	15/150	0.90	0.88

[i] AS, atherosclerosis; Pos., presented positive; PPV, positive predictive value; NPV, negative predictive value. n=150.

Expression levels of TIMP1, ACAT1, ABCA1, ALOX5 and PPARA in the AS and matched control groups

DNA methylation is often correlated with changes in the accessibility of DNA to transcriptional activators, which is crucial in the regulation of gene expression (26). RT-qPCR analysis of these five genes (TIMP1, ACAT1, ABCA1, ALOX5 and PPARA) was performed in the AS and matched control groups on samples of peripheral blood for which total RNA was available. Compared with the healthy group, the expression levels of PPARA, ALOX5 and ACAT1 were increased by 3.94-, 3.38- and 1.54-fold in the AS group, respectively (Table VII). These data suggest that these AS-associated genes were markedly altered in AS patents.

Table VII AS-associated gene mRNA expression in the two groups.

Table VII

AS-associated gene mRNA expression in the two groups.

Gene	Healthy group	AS group	P-value
ACAT1	0.1027±0.02326	0.1585±0.03378	0.0348a
ABCA1	0.4492±0.12360	0.1046±0.02010	0.0156a
ALOX5	0.07529±0.01171	0.2545±0.05715	0.0040b
TIMP1	0.1477±0.01251	0.05062±0.00502	0.0001b
PPARA	0.0861±0.02322	0.33960±0.06839	0.0025b

a P<0.05

b P<0.01 vs. healthy group. Values are presented as the mean + standard deviation. AS, atherosclerosis.

Analysis of sensitivity and specificity of methylation profiles

The novel panel, consisting of the genes TIMP1, ACAT1 and ABCA1, initially exhibited a combined sensitivity of 88% for the AS group and specificity of 90% for the matched control group, the performance of the novel panel was higher when compared with that of the individual genes for the detection of AS. The novel epigenetic (epi)-panel was compared with the previously used Doppler ultrasound method to identify AS in the validation data analysis. The validation set consisted of 100 samples that were also obtained from the General Hospital of Ningxia Medical University, including 50 patients and 50 healthy individuals, who had been excluded from the AS group via Doppler ultrasound diagnosis.

To study the correlation between methylation levels obtained by NT-MSP and Doppler ultrasound (which is a diagnostic criteria for AS) samples of the validation set were reexamined by qualitative NT-MSP. In order to increase the analytical sensitivity and to yield a higher reproducibility, primer binding sites in the test set and the validation set were located in the same genomic promoter region. By contrasting Doppler ultrasound with TIMP1, ACAT1 and ABCA1 promoter methylation, a significant, positive correlation was identified. The combination of TIMP1, ACAT1 and ABCA1 methylation revealed a sensitivity of 88% for the detection of AS and a specificity of 70% in healthy subjects, respectively. The novel panel also revealed a coincidence rate of 79%, which refers to consistency between the results of the novel panel and the Doppler ultrasound.

Discussion

The incidence of AS is increasing and represents a significant health issue, thus, there is a clear requirement for the development of prognostic markers in AS to provide risk-adjusted treatment and surveillance management. DNA methylation biomarkers have obvious applications in diagnostics (27). In the present study, the performance of a blood-based PCR assay for methylated DNA of 12 potential biomarkers was determined in two independent test sets with a total of 300 samples. The data confirmed the high performance of certain previously identified DNA methylation markers. Furthermore, the methylation frequencies of ACAT1, ABCA1, PPARA, TIMP1 and ALOX5 of AS peripheral blood were identified to be significantly changed when compared with that of the matched controls, the novel epi-panel consisting of TIMP1, ACAT1 and ABCA1 exhibited a higher sensitivity and specificity than any of the individual biomarkers.

These genes have previously been reported to be significantly involved in preventing or promoting AS. Aberrant methylation of ABCA1 and ACAT1 has been identified in the progression of AS (28). In addition, TIMP1 expression has been associated with a negative prognosis in AS (29). To the best of our knowledge the present study is the first to demonstrate TIMP1 methylation in AS, and that methylation of ACAT1 and ABCA1 has never previously been reported in AS.

Methylation changes in AS are often heterogeneous and, as yet, to the best of our knowledge, no single gene has been identified to be methylated in every AS patient specimen. Furthermore, in the majority of studies investigating methylation levels of only single genes the sensitivity was low. Therefore, it is considered to be advantageous to use a panel of genes for disease screening procedures. As a result of the comparative analysis, a biomarker panel with the best values for AS specificity was defined first in a test set. TIMP1, ACAT1 and ABCA1 proved to be candidate genes, showing a significant specificity of the methylation pattern in the matched control and AS groups. Furthermore, TIMP1, ACAT1 and ABCA1 methylation achieved 88% sensitivity, with a high specificity of 90% in the test set, this indicated that the panel had the potential to perform as effectively, potentially more effectively, than the single markers, indicating that the methylation of three genes may be useful when screening high risk individuals.

In an independent validation set, promoter methylation of the TIMP1, ACAT1 and ABCA1 combination enabled significant discrimination of AS from various control conditions. ROC analysis of three-gene panel methylation revealed a sensitivity of 88% with a specificity of 70%. Combining these with other detection methods may provide a robust epi-panel with a high sensitivity for AS detection. Compared with ultrasound, a reliable biomarker panel for the early detection of AS, revealed that the coincidence rate was 79%; thus, TIMP1, ACAT1 and ABCA1 methylation analyses may be used as a supplement to mammography, which has a low sensitivity in the early detection of AS.

In conclusion, methods for AS screening should be easy to perform, non-invasive and provide a benefit to patients. Blood-based biomarkers fulfill these three requirements. Quantifying promoter methylation of AS-associated genes in peripheral blood is a rapidly growing research topic for early AS detection. In the current study, 12 potential biomarkers were evaluated in two independent test sets, and promoter methylation of the TIMP1, ACAT1 and ABCA1 gene combination exhibited a high sensitivity and specificity in peripheral blood DNA from patients with AS. It is proposed that such a blood-based screening method would be convenient for the patient and reduce costs for health care providers.

Acknowledgments

The present study was supported by grants from the National Natural Science Foundation of China (grant nos. 81160044, 81260105, 81360073, 81200118 and 81260063).

References