Background
Atherosclerosis is a chronic, inflammatory disease that develops from early childhood onwards and progresses silently for decades before evolving into symptomatic cardiovascular disease (CVD) [
1]. The initiation and progression of atherosclerosis depends on profound dynamic modifications in arterial biology [
2]. A crucial multi-step biological mechanism that is considered a fundamental early promotor of atherosclerosis is an enhanced homing and adherence of leukocytes to the vascular endothelium and their ensuing migration through the arterial wall into its intima where they instigate the migration of smooth muscle cells (SMCs) and the evolution of foam cells and lipid deposits [
3].
The multi-step mechanism is induced by atherogenic stimuli (i.e. lipoproteins, inflammatory cytokines) and mediated by various endothelial cell adhesion molecules (CAMs) that are expressed on the surface of endothelial cells [
3‐
5]. Among identified endothelial CAMs, the biological properties and expression of circulating levels of E-selectin, P-selectin, intercellular adhesion molecule (ICAM-1) and vascular cell adhesion molecule (VCAM-1) are well described but remain to be fully studied in various populations [
4]. To date, research on endothelial CAMs has mainly been performed in middle to older aged populations and populations with an increased CVD risk burden. Yet, given the early life origins of atherosclerosis, their role in younger, asymptomatic populations warrants exploration.
As in young adults clinical CVD endpoints rarely occur, proxies for detecting subclinical atherosclerosis are required. A high-end modality that may serve this purpose is cardiovascular magnetic resonance imaging (CMR) [
6]. In addition to the completely non-invasive obtainment of indices, CMR has a large anatomical coverage and superior soft tissue contrast as well as the ability to combine arterial morphology and function in a single examination and depict tissues deep inside the body. As such, CMR allows for accurate evaluation of the various stages of aortic atherosclerosis [
6]. Indeed, various CMR-derived indices such as aortic wall thickness and pulse wave velocity (PWV), have shown to reliably reflect atherosclerosis burden in asymptomatic populations [
6]. In young populations, specifically the aorta is an important artery given that it is prone to early atherosclerotic alterations at a young age [
1].
In view of the above, the aim of the current study was to explore the relation between circulating endothelial CAMs and CMR-derived indicators of arterial wall alterations in a community-based random sample of young adults in order to improve our understanding of the pathophysiology underlying the relatively early stages of atherosclerosis.
Discussion
This study expands current knowledge on the relation of endothelial CAMs with arterial wall alterations in young adults from the general population by showing that already in a young population an increase in circulating P-selectin and E-selectin relate to an increase in CMR-derived aortic characteristics, possibly with a important role for smoking. Our results suggest that upregulation of P-selectin, and to a lesser extent, E-selectin may mirror atherogenic inflammatory alterations in the vascular bed. This study may contribute to an improved understanding of the biology and determinants of early atherosclerosis and thus may possibly aid in developing effective interventions when atherosclerosis is still, at least partially, reversible.
Soluble forms of CAMs can be detected in the circulation due to their release from the endothelium via shedding or proteolytic cleavage [
14,
15]. Although their biological role is not yet fully identified, soluble CAMs appear to reliably mirror increased expression of membrane-bound CAMs and reflect the inflammatory component of atherosclerosis [
14,
15]. Although studies report discordant results, CAMs seem involved in CVD pathophysiology. Their levels rise in relation to various CVD risk factors. Additionally, they have been related to morphological and functional measures of atherosclerosis as well as to an unfavourable CVD prognosis in various populations [
4]. For example, positive relations between all four CAMs and age, BMI, blood pressure and lipid levels have been reported [
16]. Additionally, for P-selectin, E-selectin and ICAM-1, higher levels have been observed in smokers as compared to non-smokers. [
16‐
18]. Moreover, studies have reported positive relations of P-selectin and E-selectin with carotid intima-media thickness (CIMT), arterial stiffness, plaque burden and presence of clinically overt CVD in various low and high-risk populations [
14]. For VCAM-1 and ICAM-1 relations have been observed with CIMT, plaques and clinically overt CVD [
14]. Yet, others failed to observe such relations [
19].
In this study, both the crude and multivariable model showed that P-selectin and E-selectin positively related to aortic wall thickness and aortic stiffness, respectively. This indicates that the a priori selected confounding variables, known to be risk factors for atherosclerosis and related to the CAMs, exerted little effect on the observed associations, assuming these variables are not in the causal pathway between CAMs and aortic characteristics. However, the relation of P-selectin with aortic wall area was significant in the crude model but lost its significance in the multivariable model, with a substantial change in regression coefficient (> 30%), implying that confounding biased the relation of P-selectin with aortic wall area by increasing the effect of the association.
Interestingly, when current/former smokers were compared to never smokers, the significant associations only remained significant in the current/former smoking population. Although these results have to be interpreted with care given the relatively small study population, they suggest effect modification of the association between CAMs and aortic characteristics by smoking. In addition to being an established risk factor for atherosclerosis, smoking may prompt vascular inflammation as it increases endothelial adhesion and yields pathological production of endothelial vasoactive substances, including endothelial generation and expression of adhesion molecules [
20‐
23]. Indeed, studies have documented a larger variation in CAMs, including P-selectin and E-selectin, in smokers as compared to non-smokers [
17,
20‐
23]. This increase in CAMs may facilitate the recruitment of leukocytes and other inflammatory mediators to locations of arterial damage and as such, may inflict atherogenic vascular inflammation [
18,
22,
23].
In this study, VCAM-1 and ICAM-1 did not relate to the selected aortic characteristics. This may be due to the varying expression and release of CAMs, depending on the sources and arterial beds they are derived from as well as to intrinsic differences in the biological function of endothelial CAMs, the (CVD risk) factors affecting their levels and the shedding process. For example, whereas the selectins are mostly involved in recruitment and rolling of leukocytes, VCAM-1 and ICAM-1 mainly promote binding and transmigration of leukocytes into the arterial intima, where they exert their noxious effects [
18]. Furthermore, VCAM-1 appears to be specifically upregulated in regions of intimal neovascularization and advanced atherosclerotic lesions, which may clarify the lack of an association with relatively early arterial wall alterations [
24]. Moreover, whereas E-selectin is derived solely from endothelial cells, the other CAMs are derived from multiple sources such as platelets (P-selectin), leukocytes (ICAM-1) and cell types of non-vascular origin (VCAM-1), which may explain the strong association of E-selectin with aortic PWV [
25]. Additionally, heterogeneity in studied population and used methodology also yield discrepancies in results across studies and may explain why we did not observe significant associations of VCAM and ICAM with the studied aortic characteristics.
Reducing CVD risk factors in the young may possibly reverse the atherogenic process. Prior intervention studies indeed observed that interventions, such as a diet, smoking cessation or antihypertensive medication use, did not only reduce lipid levels, weight and blood pressure, but also induced a decline in levels of the related CAMs [
18,
21,
26,
27]. Unfortunately, to date, the involvement of circulating endothelial CAMs in atherosclerosis pertains to be complex, multifaceted and not yet fully understood. CVD risk factors, atherogenic stimuli, inflammatory mediators and the endothelium all appear to be tangled in a vicious circle that induces harmful arterial wall alterations. This study indicates that P-selectin, E-selectin and smoking play a central role already in the early stages of this vicious circle. Nevertheless, the ambiguous results across studies seem to advocate that, in addition to developing a standardized methodology for these type of studies, further unravelling of the mechanisms, biology and determinants of circulating endothelial CAMs is merited before hard conclusions can be drawn. Therefore, we believe that our results warrant further exploration in larger cohorts.
Strengths of this study are our random selection of young adults from the general population, the use of promising, highly innovative imaging based indicators of arterial wall alterations and the measurement of multiple endothelial biomarkers. Moreover, in young individuals, confounding effects of lifestyle on the levels of the endothelial biomarkers are limited as compared to middle-aged to older individuals. However, limitations of this study also merit attention. Although this is one of the largest studies comparing biomarkers with CMR-derived indices of arterial wall alterations, our limited sample size forced us to interpret our results with care. Chance as well as residual confounding may have played a role, yet the latter can never be excluded in cohort studies like this. Also, due to the limited spatial resolution of MRI, arterial wall alterations that may have been present but are yet too subtle to be detected with CMR (i.e. intimal thickening) may have remained undetected. Technical improvements such as calibration of imaging protocols, standardization and automation of quantification methods as well as a refinement in spatial resolution are warranted. Moreover, although recruitment was random, sampling bias may have influenced our results; health-conscious individuals may have been more eager to participate, which may have induced an underestimation of our results. In addition, our study population comprised Caucasian, young adults with a relatively large percentage of current/former smokers. As such, the features of our study population restrict the generalizability of our results to populations that are similar to our study population. Also, the cross-sectional nature of our study prevented us from making inferences on causality of the studied relations. Moreover, we assessed only 4 endothelial biomarkers, yet other unknown molecules and pathways are undeniably involved in the multi-step mechanism.