Relationship between Plasma Lipidomics and Carotid Atherosclerotic Plaque

Background: The rapidly growing discipline of lipidomics allows the study of a wide spectrum of lipid species in human plasma and provides new insights into the pathogenesis of atherosclerosis. Objective: To explore the roles of plasma lipid components in relation to atherosclerotic plaques. Methods: Ten non-symptomatic persons (age 62.42±3.45) with 2 or more carotid plaques (soft or mixed) and ten apparently healthy controls without any carotid plaque (age 62.82±4.38), were randomly selected as cases and controls from a community-based sample of 1312 persons. Ultra-performance liquid chromatography coupled with electrospray ionization quadrupole mass spectrometry (UPLC-ESI-QTOF MS)-based lipidomics was used to measure lipid components. The difference of lipid species between cases and controls was analyzed by t-test and general linear regression. Results: Among 53 lipid components, signicantly elevated plasma levels of one novel individual ceramide species (ceramide 22: 0) were observed in cases compared to the controls (0.59±0.18 versus 0.41±0.10μg/mL, p values<0.05). After adjusting for confounding factors, such as total cholesterol and low-density lipoprotein cholesterol, the results remained signicant (p<0.05). 22:0 ceramide was signicantly negatively correlated with vegetable intake (r=-0.74, p=0.014). Conclusion: High level of plasma ceramide 22:0 may be a novel risk factor for carotid atherosclerosis in human which independent of cholesterol and low-density lipoprotein cholesterol, it deserves future studies with larger sample size to conrm. 12 phosphorylethanolamine, 7 sphingomyelin, 26 phosphatidylcholine, and 1 phosphatidylinositol. We also calculated ratio of ceramides C22:0/C16:0, C24:0/C16:0, C24:0/C24:1, C22:0/C24:1, which were reported by other previous studies 21-23


Introduction
Atherosclerosis (AS) remains one of the major health problems globally, yet its underlying mechanisms are still unclear. Known risk factors for AS include hypercholesterolemia, hypertension, diabetes, smoking, obesity, family history, and unhealthy dietary habits [1][2][3][4] . However, the mechanisms by which these risk factors lead to the development of AS are less well understood.
There are several proposed mechanisms including lipid deposition theory, in ammatory theory, and stress (mechanical) damage theory 5,6 . These hypotheses involve a complex series of reactions of biomolecules, including low-density lipoprotein, oxidized low-density lipoprotein, cytokines (C-reactive protein, intercellular adhesion molecule-1, colonystimulating factor-1, tumor necrosis factor, Interleukin-1, 6, 10, 18), nitric oxide, and signaling molecules 7-10 , all of which play an important role in the development of atherosclerosis.
Whilst it is well known that abnormal lipid metabolism plays a crucial role in the development of atherosclerosis, previous studies have focused on atherosclerosis plaques that have already occurred 11 . Inside human carotid plaques, the sphingolipids and particularly glucosylceramide are associated with and are possible inducers of plaque in ammation and instability 11 . Another recent study found that elevated plasma levels of C16:0 and C24:1 ceramide, correlating with immune activation and in ammation, were associated with antiretroviral therapy use and progression of carotid artery atherosclerosis in 2 HIV cohorts 12 . The speci c roles of biomolecules in early stages of AS development in general population, however, remained unknown 13,14 . The study of a wide spectrum of lipid species in human plasma may provide potentially new insights into the pathogenesis of atherosclerosis. Therefore, the present study aims to explore the role of plasma lipid components as potential risk factors of atherosclerotic plaque development in patients with carotid atherosclerotic plaques and their healthy controls.

Study population
The ow chart of carotid plaque group and control group were shown in Figure 1. In 2018, 10-year Follow-up Study of Subclinical Atherosclerosis Cohort were performed in 11 communities in Beijing in 2012 15 . Among 1312 participants, 250 were <69 years old, free of diseases (angina/myocardial infarction, stroke, and other symptomatic diseases, such as cancer, liver disease, kidney disease, respiratory disease, connective tissue disease, thyropathy, etc.), abnormal ECG, surgery in recent two weeks, cardiovascular surgery in recent 10 years, lipid lowering drug in recent 2 weeks, but having necessary analysis data, in which 161 had carotid plaque, and 89 were free of carotid plaque. Among these 161 with carotid plaque, 85 had 2 or more soft/mixed plaque, in which ten participants were randomly (using SURVEYSELECT procedure in SAS® 9.4 software) selected as carotid plaque group (cases). Among these 89 free of carotid plaque, ten were randomly selected as control group (controls).

Study protocol
Data on lifestyle factors and medical history were obtained using a standard questionnaire by an interview in 2012 15,16 .
Smoking was de ned as at least 1 cigarette per day for more than 1 year. The frequency and weight of food were collected in the validated questionnaire (Supplement A in File S1). Body Mass Index (BMI) = weight (kg) / height 2 (m2). Blood pressure was measured on the right arm by a sphygmomanometer with two consecutive measurements that were at least 30s apart. The average of the two readings was used for analysis. Laboratory tests include fasting blood glucose (FBG), serum total cholesterol (TC), triglycerides (TG), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), interleukin-6 (IL-6), interleukin-10 (IL-10), high-sensitivity C-reactive protein (CRP), P-selectin, and soluble intercellular adhesion molecule-1(s-ICAM-1) 8,9 . In the morning, an empty 12 h venous blood sample was taken for testing, and all tests were performed in the central laboratory.
At the same time (2012), carotid plaque measurements were performed using a standardized protocol as described previously 15,17 . The participant was placed in the supine position with the head tilted back and to the opposite side of the inspection. The probe was placed in the running area of the corresponding body surface of the blood vessel to be examined, and scanning was performed from the proximal end to the distal side. Plaques were de ned as more than or equal to 1.3 mm of local intima thickening or more than or equal to 1.3 mm of local bulging into the lumen. The examination included bilateral carotid artery and carotid sinus. According to the ultrasonic echo level, the plaque was divided into: 1 = hard plaque (high echo); 2 = mixed plaque (mixed echo); 3 = soft plaque (low echo). The mixed/soft plaques usually occur newly and later than hard plaques. At same time, bigger differences of lipid molecules and statistic power were expected to be observed between severer patients and healthy controls. Thus, in the present study, we selected those with 2 or more mixed/soft plaque as cases with atherosclerosis. The repeatability evaluation of carotid ultrasound examination were shown in previous published papers 15,18 .
Blood samples were collected in 2012. Then the plasma samples were separated from blood samples and stored at -80°C. In 2018, the plasma samples of the ten randomly selected persons with 2 or more soft or mixed carotid plaques and ten healthy controls were taken out and thawed at 4 °C. Lipid species of these plasma samples were measured using Ultraperformance liquid chromatography coupled with electrospray ionization quadrupole mass spectrometry (UPLC-ESI-QTOF MS) system in Tianjin University. The application studies of this system were shown in previous published papers 19,20 . Firstly, 10 μl of plasma was added to 10 μl of mixed standard solution (containing 2 μg/mL 19:0-19:0 PC, 10 μg/mL 17:0-17:0 PE, 15 μg/mL 12:0 SM, 2μg/mL 19:0 Lyso PC, or 10μg/mL 19:0 ceramide dissolved in isopropanol-acetonitrile solution). 10μl 0.9% NaCl and 100μl chloroform-methanol (2:1) extract were then added to the mixture. The mixture was vortexed for 60s, left to settle at 4 °C for 30 min, and then centrifuged at 13,000 g/min for 3 min. A 1 mL syringe was used to transfer 50 μl of the lower layer to a 1.5 mL EP tube, which was blown dry with nitrogen, reconstituted with 25 μl of acetonitrile-isopropanol (1:1), and vortexed for 60 s. The lipids were detected by Eksigent LC100 and AB SCIEX Triple TOF 5600 instruments. The conditions of Eksigent LC100 were as follows: mobile phase A was 10 mM ammonium acetate + 0.1% formic acid + 99.9% water, mobile phase B phase was 10 mM ammonium acetate + 0.1% formic acid + 49.95% acetonitrile + 49.95% isopropanol, the column temperature was 40 ° C, the injection chamber temperature was 4 ° C, the ow rate was 0.4 mL / min, the injection volume was 2 μl. The gradient elution conditions were shown in Supplement B in File S1. The AB SCIEX Triple TOF 5600 instrument was measured in negative ion mode and the test parameters were shown in Supplement C in File S1.

Quality Control
First, calibration standards of known concentrations were used in measurement of lipid molecules. Second, based on the research protocol, a comprehensive manual was developed, including standard operating procedures for questionnaire, height, weight, blood pressure, blood collection, specimen transportation, and laboratory testing. Third, the survey database was built using EpiData 3.1 (EpiData Association, Odense, Denmark). Double entry was performed independently by two people and all raw data were con rmed to be completely consistent. Finally, data veri cation was performed, and all issues were resolved before locking the database.

Statistical Analysis
Statistical analysis was performed using SAS® 9.4 software (SAS Institute Inc., Cary, NC). T-test or Wilcoxon rank test was conducted to determine the differences of the mean or median values of continuous variables between the groups. The normality of the distribution of variables was assessed using the Shapiroe-Wilk test. Chi-square test or Fisher's exact test was conducted to determine the differences of the percentages of categorical variable between the groups. Confounding factors were controlled using two general linear models: model 1) for adjustment of total cholesterol (TC) and low-density lipoprotein cholesterol (LDL-c); and model 2) for adjustment of interleukin-6(IL-6).
Model 1: Concentration of each lipid component=a+ b1*group +b2*TC+b3*LDL-c; Model 2: Concentration of each lipid component=a+ b1*group +b2*IL-6 Here group was case/control (case=1; control=0). In order to reduce the effect of multiple comparison, we limited the models to those lipid components which showed signi cant association with carotid plague by t-test. Correlations between ceramide and risk factors/dietary intake were assessed by Spearman Correlation.
The signi cance level of all analyses was set at P < 0.05. Due to the exploratory study design, p cut-point value adjustment for multiple comparison were not performed.

Comparison of Key Factors between patients and healthy controls
Between carotid plaque group and healthy control group, no signi cant differences were found for age, sex, smoking frequency, BMI, systolic blood pressure, diastolic blood pressure, fasting blood glucose, triglyceride, high-density lipoprotein cholesterol, interleukin-6, interleukin-10, C-reactive protein, P-selectin, or soluble intercellular adhesion molecule-1 (s-ICAM-1). However, higher total cholesterol and low-density lipoprotein cholesterol levels were signi cantly higher in the carotid plaque group than the healthy control group (p<0.05), and thus were adjusted to assess the relationship of lipids index with atherosclerosis plaque ( Table 1) Because the interleukin-6 was not signi cantly different between cases and controls, it may not be a confounder and adjustment for interleukin-6 did not change the signi cance (data not shown).
The Relationship between 22:0 Ceramide and Cardiovascular Risk Factors (Table 3) In order to explain why 22:0 ceramide was associated with atherosclerosis, we explore the relationship between 22:0 ceramide and risk factors of cardiovascular disease using Spearman Correlation. 22:0 ceramide was signi cantly negatively correlated with vegetable intake (r=-0.74, p=0.014). The other correlations did not reach a signi cant level (Table  3).

Discussion
In this study, ten non-symptomatic persons with carotid plaque and ten clinical healthy controls were randomly selected from community residents to analyze the association between blood lipid species and carotid atherosclerosis (AS). Among the 53 individual lipid species examined, after adjusting for confounding factors, 22:0 ceramide in carotid plaque group was signi cantly higher than the healthy controls.
Ceramide is a kind of sphingolipids which is made up of long-chain bases of sphingosine and fatty acids. For a long time, it was believed that the only function of ceramide was to combine with choline phosphate and hence forming sphingomyelin, which constitutes the lipid bilayer structure of mammalian cell membrane. A recent study found that higher levels of C16:0, C22:0, and C24:0 ceramides were associated with higher risk of ischemic stroke 24 . However, the mechanism is unclear. Whereas this study provided a preliminary scienti c basis for the atherogenic effect of ceramide using a population epidemiology approach. We identi ed 22:0 ceramide molecular signi cantly associated with mixed or soft carotid atherosclerosis plaques. Moreover, there was a tendency for higher ceramide (24:0) in plaque subjects.
Ischemic stroke due to artery embolism is mostly due to carotid plaque formation 17 . Thus the impact of ceramide on carotid plaque may play key role in pathophysiology of ischemic stroke due to artery-artery thromboembolism.
More important was that the associations of 22:0 ceramide with carotid plaque were independent of total cholesterol and LDL-c. In the present study, total cholesterol and LDL-c were the rst and second strongest macro lipid associated with atherosclerosis, in line with current literatures 25,26 . It was previously shown that ceramides (micro lipid) was signi cantly associated with LDL aggregation, and signi cantly associated with mortality coronary artery disease (CAD) 27 . Thus, ceramide may have independent extra effect on the development of atherosclerosis outside of the effect of LDL and total cholesterol.
It is worth highlighting that ceramide showed a signi cant negative correlation with vegetable intake (p=0.01). Previously, it was shown that sphingomyelin (SM) levels in LDL were reduced with healthy, vegetable containing diet suggesting that LDL with lower amount of SM is not a good substrate for sphingomyelinase in arterial wall and thereby affect ceramide formation 27 . The observed association between vegetable intake and ceramide provides a potential explanation on the previous nding that vegetable intake reduce risk of subclinical carotid atherosclerosis 28 . If our results were con rmed by larger studies in the future, population level strategies to incentivize higher vegetable intake may lead to reduction of ceramide levels and hence prevent atherosclerosis at population level.
There are some limitations. Firstly, the sample size was small. Thus, we performed a retrospective power-analysis. When type error is set as 0.05, the statistic power is calculated as 0.734 to check out the difference of 22:0 ceramide in patient and control group using PASS11 software. Therefore, detected associations between 22:0 ceramide and the development of AS still needs to be veri ed in larger sample populations. If this association can be con rmed in a large sample cohort, it will promote the understanding of 22:0 ceramide as a new risk factor as well as an important therapeutic target for cardiovascular diseases. Secondly, this study measured 22:0 ceramide and atherosclerotic plaque in a time section. If there is a causal relationship still needs to be con rmed by cohort studies. Thirdly, blood samples were collected in 2012 and plasma samples were stored at -80°C, and were thawn in 2018 to be used for lipidomics. Some lipid molecules were prone to degrade during storage. However, the degree of degradation may be limited because the lipid molecules, such as ceramide, were at same order of magnitude in our study as another published study 29 . Moreover, because the blood samples were stored at same condition and were tested by same methods, the degradation (if any) is likely to be similar in both cases and controls. Thus, whilst limited degradation of sample may weaken the strength of the association, it is unlikely to have impact on the direction of results. Fourthly, we used whole plasma. Because different lipoprotein fractions contain varying amounts of ceramide especially LDL and HDL, when using whole plasma and especially here with small number of samples the plasma values are likely to be diluted 30 . These limitations may weaken the association but very unlikely to change the associations observed and thus study conclusion. Fifthly, we did not match lipid pro le (total cholesterol, LDL-c, etc.).
In summary, the plasma 22:0 ceramide was signi cantly associated with atherosclerotic plaque in the present study. The result suggested that 22:0 ceramide might be a new risk factor of atherosclerosis, but further validation is needed using larger sample cohorts.   Figure 1 Flow chart of carotid plaque cases and controls

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. S1.docx