Vascular inflammation and low‐density lipoproteins: is cholesterol the link? A lesson from the clinical trials

Abstract

For long time, the role of LDL and inflammation in the pathogenesis of atherosclerosis have been studied independently from each other and only more recently a common platform has been suggested. Accumulation of excess cholesterol due to the presence of increased circulating LDL promotes endothelium dysfunction and activation, which is associated with increased production of pro‐inflammatory cytokines, overexpression of adhesion molecules, chemokines and C‐reactive protein (CRP), increased generation of reactive oxygen species and reduction of nitric oxide levels and bioavailability. All these processes favour the progressive infiltration of inflammatory cells within the arterial wall where cholesterol accumulates, both extracellularly and intracellularly, and promotes vascular inflammation. According to this, lipid‐lowering therapies should improve inflammation and, indeed, statins decrease circulating inflammatory markers such as CRP and improve endothelial function and plaque burden. Pleiotropic activities have been proposed to explain this effect. However, mendelian randomization studies ruled out a direct role for CRP on coronary artery disease and studies with other lipid lowering drugs, such as ezetimibe showed that the beneficial effect of LDL‐cholesterol‐lowering therapies on systemic inflammatory status, as monitored by changes in CRP plasma levels, could be achieved, independently of the mechanism of action, only in patients presenting with baseline inflamed conditions. These observations strengthen the direct link between cholesterol and inflammation and indicate that decreasing LDL levels is one of the key goals for improving cardiovascular outcome.

Linked Articles

This article is part of a themed section on Targeting Inflammation to Reduce Cardiovascular Disease Risk. To view the other articles in this section visit http://onlinelibrary.wiley.com/doi/10.1111/bph.v174.22/issuetoc and http://onlinelibrary.wiley.com/doi/10.1111/bcp.v82.4/issuetoc

Abbreviations

CAD

coronary artery disease

CETP

cholesteryl ester transfer protein

CHD

coronary heart disease

CRP

C‐reactive protein

CV

cardiovascular

FMD

flow‐mediated dilatation

HMG‐CoA

3‐hydroxy‐3‐methylglutaryl‐CoA

hs‐CRP

high‐sensitivity CRP

IVUS

intravascular ultrasound

LDL‐C

LDL‐cholesterol

LDLR

LDL receptor

Lp(a)

lipoprotein(a)

NLRP3

nucleotide‐binding domain, leucine‐rich‐containing family, pyrin domain‐containing‐3

OxLDL

oxidized LDL

PAV

percent atheroma volume

PCSK9

proprotein convertase subtilisin/kexin type 9

SNP

single nucleotide polymorphism

TG

triglycerides

VLDL

very low density lipoprotein

Introduction

Inflammation and hypercholesterolemia are linked in a vicious cycle in which the excess of cholesterol that accumulates in the arterial wall induces an inflammatory response that, in turn, accelerates cholesterol deposition and amplifies inflammation.

The role of LDL and inflammation in the pathogenesis of atherosclerosis have been, for a long time, studied independently from each other and only more recently a common platform has been suggested. The role of cholesterol (the ‘cholesterol era’) and that of cholesterol‐carrying LDL (the ‘LDL era’) in the formation of atherosclerotic plaques has been extensively evaluated starting from the beginning of the 20th century and was validated by the discovery of the LDL receptor (Linton et al., 2000; Goldstein and Brown, 2015). In parallel, the concept of a role for inflammation as a key player in atherosclerosis development gained attention, as a result of studies showing that atherosclerotic plaques are characterized by the accumulation of inflammatory cells. This accumulation, by producing pro‐inflammatory cytokines, further promotes the entry of monocytes into the arterial wall, thus further propagating the inflammatory reaction (Ross, 1999; Libby, 2002; Libby, 2012).

The aim of this brief review is to describe the evidence linking cholesterol and LDL to inflammation and discuss the data from clinical trials with lipid‐lowering drugs suggesting that the beneficial impact on inflammation is proportional to the reduction of levels of LDL‐cholesterol (LDL‐C).

Cholesterol promotes inflammation

Accumulation of excess cholesterol within the arteries promotes endothelium dysfunction and activation, which results in increased production of pro‐inflammatory cytokines and reactive oxygen species, overexpression of adhesion molecules, chemokines and reduction of nitric oxide levels and bioavailability (van Diepen et al., 2013; Gimbrone and Garcia‐Cardena, 2016) (Figure 1). These processes contribute to the recruitment and infiltration of monocytes, which differentiate into macrophages and, following the uptake of modified‐LDL via scavenger receptors, become foam cells (Ross, 1999; van Diepen et al., 2013; Sorci‐Thomas and Thomas, 2016) (Figure 1). More recently, cholesterol has been directly linked to inflammation via the activation of the NLRP3 inflammasome (Grebe and Latz, 2013), which might favour the instauration and amplification of a local and systemic immuno‐inflammatory response (Tabas and Bornfeldt, 2016), characterized by the production of several pro‐inflammatory cytokines; among them, C‐reactive protein (CRP), IL‐6 and IL‐1 are well‐established markers of inflammation and their possible causal role in atherosclerosis has been widely investigated (Ridker, 2016). In addition, among genetic traits associated with LDL‐C levels, single nucleotide polymorphisms (SNPs) in the CELSR2/PSRC1/SORT1 locus and in the APOE/APOC1/TOMM40 locus have been associated also with inflammatory‐related phenotypes (Kocarnik et al., 2014), which further supports the strong relationship between cholesterol and inflammation.

Figure 1.

Link between cholesterol and inflammation within the arterial wall.

Cholesterol and CRP

CRP is a plasma protein belonging to the superfamily of pentraxins, proteins involved in acute phase responses. It is synthesized by the liver in response to IL‐6 and is a well‐established marker of inflammation (Pepys and Hirschfield, 2003) whose levels significantly increase in response to inflammatory stimuli (Gabay and Kushner, 1999). A large meta‐analysis from 54 long‐term prospective studies reported a continuous associations of plasma CRP levels with coronary heart disease (CHD) risk (Kaptoge et al., 2010), further supporting the concept that CRP could be a powerful risk biomarker for first and recurrent cardiovascular (CV) events (Koenig, 2013).

CRP is, however, synthesized in cells other than hepatocytes, including cells within the atherosclerotic plaques, as suggested by the co‐localization of CRP and oxidized LDL (OxLDL) and macrophages in atherectomy specimens from patients with stable or unstable angina and acute myocardial infarction (Meuwissen et al., 2006). Also endothelial cells exposed in vitro to modified LDL or to pro‐inflammatory cytokines produce CRP (Venugopal et al., 2005; Chu et al., 2013). The role of the CRP produced by the atherosclerotic plaque, however, is unknown, and it is unclear whether it contributes to CRP plasma levels in CVD patients. CRP has been proposed to play a causal role in atherosclerosis (Zhang et al., 1999; Pasceri et al., 2000; Cirillo et al., 2005). Data from clinical trials indicating that the greatest reduction in CV events in statin‐treated patients is observed in those achieving both LDL‐C and high‐sensitivity CRP (hs‐CRP) reduction (Joshi and Jacobson, 2010) support this hypothesis. However, several other studies, including Mendelian randomization studies, have downplayed the causal role of CRP in CHD (Zacho et al., 2008; Elliott et al., 2009; Wensley et al., 2011; Lane et al., 2014; Noveck et al., 2014) and the field is still open to discussion.

Cholesterol and inflammatory cytokines

Cytokines play a key role in inflammatory diseases and a link with hypercholesterolemia and atherosclerosis has emerged mainly for the IL‐6, IL‐1 and TNFα pathways.

IL‐6 is a pro‐inflammatory cytokine produced by several cell types in response to infections or other conditions, playing a critical role in the pathogenesis of rheumatoid arthritis (RA) (Liu et al., 2015) but also involved in atherogenesis (Schuett et al., 2009). In endothelial cells, modified LDL up‐regulate IL‐6 (Lubrano et al., 2015), which in turn induces the expression of macrophage scavenger receptors involved in the uptake of modified LDL, thus promoting the formation of foam cells (Schuett et al., 2009) and establishing an inflammatory cycle in the plaque. In humans, IL‐6 levels predict future CV risk and correlate with endothelial dysfunction and carotid intima‐media thickness (Ridker, 2016). In addition, carriers of the Asp358Ala SNP in the IL‐6 receptor gene (IL6R) have increased serum levels of IL‐6R (probably due to an increased shedding of the receptor) and a paradoxical increase of IL‐6, but reduced levels of the downstream mediators CRP and fibrinogen. This is suggestive of an attenuation of the IL‐6/IL‐6R axis signalling in carriers of this variant; indeed this IL6R SNP was associated with a decreased CHD risk (Sarwar et al., 2012; Swerdlow et al., 2012). However, data from clinical trials in patients with RA treated with therapies targeting the IL‐6/IL‐6R axis challenged this hypothesis. Indeed the treatment of RA patients with tocilizumab, a monoclonal antibody that blocks both membrane‐bound and circulating IL‐6R, increased LDL‐C levels, an effect observed also with tofacitinib, a JAK inhibitor that blocks intracellular signalling of several cytokines (including IL‐6) (Souto et al., 2015). In vitro tocilizumab reduced the levels of hepatic LDL receptor (LDLR) (Strang et al., 2013), which may result in impaired LDL catabolism and explain the increased LDL‐C plasma levels observed in RA patients treated with this biological agent (Kawashiri et al., 2011; Strang et al., 2013; McInnes et al., 2015). These observations further stress the interplay between cholesterol metabolism and inflammatory signals.

In addition, IL‐1 plays a relevant role in vascular inflammation and atherogenesis (Qamar and Rader, 2012). IL‐1 is able to induce its own production but also up‐regulates the expression of downstream mediators of inflammation such as IL‐6 and CRP (Signorelli et al., 2014). Cholesterol crystals contribute to the activation of IL‐1 dependent pathway by activating the NLRP3 inflammasome, which favours the cleavage and secretion of IL‐1 (Duewell et al., 2010; Rajamaki et al., 2010). Several studies in humans confirmed the involvement of IL‐1 in the development of atherosclerosis (Galea et al., 1996; Fearon and Fearon, 2008; Olofsson et al., 2009), and patients with IL‐1 polymorphisms resulting in higher levels of pro‐inflammatory cytokines were at increased risk for the presence of coronary artery disease (CAD) and CV events (Tsimikas et al., 2014). The central role of IL‐1 in inflammation has been confirmed by studies with agents targeting IL‐1 activity, such as anakinra (an IL‐1 receptor antagonist), which decreased CRP production in acute coronary syndrome patients (Abbate et al., 2010; Abbate et al., 2013; Morton et al., 2015), but failed to reduce the risk of recurrent ischaemic events, whereas it may prevent new‐onset heart failure (Abbate et al., 2015). Canakinumab, a human monoclonal antibody that neutralizes IL‐1ß, reduces CRP and IL‐6 in patients with T2DM and established CVD without affecting plasma cholesterol levels (Choudhury et al., 2016).

TNFα is a pro‐inflammatory cytokine, which contributes to the development of atherosclerosis by inducing endothelial dysfunction and initiating the inflammatory cascade inside the arterial wall (Ross, 1999). Although its increase during acute inflammation is protective, its persistence at high levels during chronic inflammation may results in alterations of both lipid and glucose metabolism with detrimental consequences (Popa et al., 2007). In fact, TNFα may interfere with cholesterol metabolism, by decreasing the secretion of apolipoproteins and reducing cholesterol catabolism and excretion, which results in decreased LDL‐C concentrations (Popa et al., 2007). In addition, TNFα alters the quality of lipoproteins by favouring the generation of pro‐atherogenic small dense LDL and OxLDL due to changes in sphingolipid content (Popa et al., 2007). TNFα also reduces HDL‐C levels and alters HDL composition (Popa et al., 2007). Some SNPs in the TNFα gene are associated with changes in LDL‐C levels; the C‐857T SNP on the TNFα promoter region was associated with higher LDL‐C levels (3.14 mmol·L⁻¹ in the T carriers [TT/CT genotypes] and 2.89 mmol·L⁻¹ in the non‐T carriers [CC genotype], P < 0.05) and increased frequency of carotid plaque in patients with type 2 diabetes mellitus (DM) (87% in the T carriers vs. 63% in the non‐T carriers, P = 0.0358) (Yamashina et al., 2007; Takahashi et al., 2010). Interestingly, when analysed according to statin treatment, LDL‐C levels were higher in the T carriers compared with the C carriers only in statin‐treated subjects, but not in statin‐untreated, and the reduction of LDL‐C levels achieved with statins was lower in the T carriers than in the C carriers (27.6% vs. 36.4%, P = 0.031) (Takahashi et al., 2010). Similarly, among asthmatic patients, the frequency of the promoter region −308G/A polymorphism was higher in subjects having metabolic syndrome and was associated with higher TNFα levels and higher LDL‐C levels in GA/AA genotypes than in GG genotype (3.13 vs. 2.55 mmol·L⁻¹, P = 0.029) (Yang et al., 2015). Despite these observations, the anti‐TNFα therapies seem to be neutral on lipid profile, as reported by several meta‐analyses which could not find major significant changes for LDL‐C or apolipoprotein B following therapy with TNFα antagonists (van Sijl et al., 2011; Daien et al., 2012; Di Minno et al., 2014).

These observations support the concept that hypercholesterolaemia promotes systemic and vascular inflammation through the induction of several mediators. Some of them could contribute to the amplification of the inflammatory response, others, such as CRP, mark the ongoing inflammatory response, while a direct effect of inflammatory cytokines on plasma cholesterol levels is debatable.

Inflammation alters lipid metabolism

Chronic inflammatory diseases (such as RA or systemic lupus erythematosus) are associated with increased CV risk (Haque et al., 2008) and patients present with quantitatively and qualitatively altered lipid and lipoproteins profile that include a reduction of total cholesterol, HDL‐C and apolipoprotein A‐I, and increased levels of small dense LDL, lipoprotein(a) [Lp(a)] and triglycerides (TG) (de Carvalho et al., 2008; Amezaga Urruela and Suarez‐Almazor, 2012; Ammirati et al., 2014; Montecucco et al., 2015). A more severe disease state is associated with more pronounced alterations in lipids and lipoproteins profile.

Inflammatory mediators such as IL‐6, IL‐1ß and TNFα may alter lipid metabolism (Khovidhunkit et al., 2004), by increasing very low density lipoprotein (VLDL) production and secretion by the liver, paralleled by a decreased clearance of TG‐rich lipoproteins, with the net effect of increasing serum TG levels (Khovidhunkit et al., 2004). As a consequence, the activity of the cholesteryl ester transfer protein (CETP), in the attempt to transfer TG from VLDL/LDL to LDL/HDL, increases; these TG‐enriched particles become the substrate of the hepatic lipase and lipoprotein lipase with the generation of small dense LDL and HDL as final products. Small dense LDL enter the arterial intima more easily (Diffenderfer and Schaefer, 2014), and are more prone to oxidation while small dense HDL possess a limited antioxidant and anti‐inflammatory activity (Welty, 2013). In addition, the presence of IL‐6 responsive elements present in the promoter of apo(a) gene contribute to the increased Lp(a) levels observed during inflammation. (Wade et al., 1993).

In addition, LDL levels and composition change during inflammation. On one hand, LDL‐C levels decrease as a consequence of increased LDLR expression, which however fosters the intracellular accumulation of cholesterol (Ruan et al., 2006; Ye et al., 2009) and might induce inflammasome activation. On the other hand, circulating LDL have an increased susceptibility to oxidation (Frostegard et al., 2005; Garcia‐Gomez et al., 2014), which explain the increased plasma levels of OxLDL in patients with chronic inflammatory disease (Ahmad et al., 2014; Nowak et al., 2016). OxLDL are more atherogenic, can amplify the inflammatory response but can also favour the accumulation of cholesterol in lysosomes, which results in increased cellular toxicity by favouring lysosome disruption, because of the presence of cholesterol crystals (Roma et al., 1992). Similarly, apart from a decrease in serum HDL, also HDL particle functions are altered during inflammation, resulting in the deterioration of most steps of the reverse cholesterol transport process and a reduced ability of ‘inflamed’ HDL to protect LDL from oxidation (Namiri‐Kalantari et al., 2015).

Altogether, these observations indicate a convincing link between inflammation and lipids in the process of atherosclerosis. Is there clinical evidence that inflammation can be modified by lipid‐lowering therapies?

Effects of statins on vascular inflammation

As LDL‐C levels are directly correlated with systemic inflammation, which is a key element in the pathogenesis of atherosclerosis (Ross, 1999; Viola and Soehnlein, 2015) targeting lipoprotein metabolism should represent a therapeutic option to reduce the burden of inflammation and of CHD. Indeed, it is well known that statins [3‐hydroxy‐3‐methylglutaryl‐coenzyme A (HMG‐CoA) reductase inhibitors] decrease levels of LDL‐C and reduce CAD. Statins inhibit the biosynthesis of cellular cholesterol in the liver, thus resulting in an increased expression of the LDLR in the hepatocytes, which in turn favours the increased catabolism of LDL from the circulation. By reducing LDL‐C levels, statins decrease the number of LDL particles that can infiltrate the vessel wall and thus limit atherosclerosis progression. Beyond this mechanism, several experimental studies have shown that statins exert additional effects, which, at the molecular level, relate to their ability to influence protein prenylation (Jasinska et al., 2007), thus in turn affecting different intracellular signalling pathways independently of their lipid‐lowering property. These additional functions are grouped under the umbrella of ‘pleiotropic effects of statins’, which have been extensively reviewed elsewhere (Bellosta et al., 2000; Mihos et al., 2014; Satoh et al., 2015). Among them, statins improve endothelial function and reduce platelet aggregation, increase the number and activity of endothelial progenitor cells, inhibit migration and proliferation of smooth muscle cells, stabilize coronary plaques and promote atheroma regression (Bellosta et al., 2000; Mihos et al., 2014; Satoh et al., 2015). Are the beneficial effects of statins on inflammation a consequence of their pleiotropic effects or the consequence of LDL cholesterol lowering?

A key effect of statins is the ability to decrease the levels of inflammatory markers including CRP (Table 1). In the primary prevention setting, lovastatin (20 to 40 mg) (AFCAPS/TexCAPS study) (Ridker et al., 2001) showed to reduce CRP levels by approximately 15%, after 1 year of treatment. The MIRACL and the REVERSAL studies showed a dose‐dependent effect of statins on CRP reduction, with the more aggressive therapy (atorvastatin) to be more effective than the standard therapy (pravastatin) (Kinlay et al., 2003; Nissen et al., 2004).

Table 1.

Effect of statins on CRP levels in clinical trials

(Ref.)	Drug	Time of intervention	Median baseline CRP level (mg·L−1)	CRP % change (P value)
(Ridker et al., 1999)	Pravastatin	5 years	2.3	−17.4 (P = 0.004)
(Ridker et al., 2001)	Lovastatin	1 year	1.6	−14.8 (P < 0.001)
(Albert et al., 2001)	Pravastatin	24 weeks	2.0	−14.2 (P < 0.001)
(Jialal et al., 2001)	Pravastatin	6 weeks	2.6	−20.3 (P < 0.025)
	Simvastatin			−22.8 (P < 0.025)
	Atorvastatin			−28.3 (P < 0.025)
(Kinlay et al., 2003)	Atorvastatin	16 weeks	11.5	−34 (P < 0.0001)
(Nissen et al., 2004)	Pravastatin	18 months	3.0	−5.2
	Atorvastatin			−36.4 (P < 0.001 vs. pravastatin)
(Ridker et al., 2005)	Atorvastatin	24 months	12.2	−89.3 (P < 0.001)
	Pravastatin		11.9	−82.4 (P < 0.001)
(Morrow et al., 2006)	Simvastatin	4 months	2.4	−29.2 (P < 0.0001)
(Ridker et al., 2008)	Rosuvastatin	48 months	4.2	−57.1 (P < 0.001)
(Emberson et al., 2011)	Simvastatin	5.0 years	3.07	−27 (P < 0.0001)
(Sever et al., 2013)	Atorvastatin	6 months	2.4 w/o events 3.0 w/ events	−25.8 (P < 0.02)
(Soedamah‐Muthu et al., 2015)	Atorvastatin	1 year	1.3	−9.8

Also in a post hoc analysis of a study with pravastatin in secondary prevention (CARE study), this statin (5 years of treatment) reduced CRP by approximately 17% (Ridker et al., 1999) (Table 1). In both the A to Z trial and the PROVE IT‐TIMI 22 study, the best outcomes were observed in patients who reached both an LDL‐C less than 70 mg·dL⁻¹ and an hs‐CRP less than 2.0 mg·L⁻¹, with even greater benefit in those individuals in which the hs‐CRP was less than 1.0 mg·L⁻¹ (Ridker et al., 2005; Morrow et al., 2006). This observation supported the concept of a ‘dual target therapy’, in which patients benefit from both LDL‐C and hs‐CRP lowering. These observations paved the road for the most important trial which to date examined the effect of statins on hs‐CRP and the resulting clinical outcomes, which is the JUPITER trial. This study investigated approximately 17 800 patients with median LDL‐C of 108 mg·dL⁻¹ but elevated hs‐CRP (>2.0 mg·L⁻¹) (Ridker et al., 2008) which were treated with rosuvastatin (20 mg) or placebo. Major CV events such as stroke, nonfatal myocardial infarction, revascularization, unstable angina or death from CV causes were reduced by 44% (P < 0.00001) in the treatment arm as was the case for both LDL‐C and CRP (Ridker et al., 2008). A major finding of the study was the observation that the greatest reduction in CV events was in the treatment group that achieved both LDL‐C less than 70 mg·dL⁻¹ and hs‐CRP less than 2 mg·L⁻¹ (65% reduction), compared with only a 33% risk reduction in patients that achieved one or neither target (P < 0.0001) (Ridker et al., 2009).

However, other clinical trials (ASCOT and CARDS) (Table 1) have reported different findings. Indeed the lowest risk for CV events was observed in statin‐treated subjects who achieved LDL‐C level below the median independent of on‐treatment CRP levels, suggesting that on‐therapy LDL‐C levels are the major determinant of the beneficial effects of statins (Sever et al., 2013; Soedamah‐Muthu et al., 2015).

It is important to note that baseline CRP levels in the CARDS trial were much lower compared with the JUPITER trial (1.4 vs. 4.3 mg·L⁻¹). Similarly, in the ASCOT trial the median level of CRP in patients without history of CV events (~90% of the studied population) was 2.4 mg·L⁻¹ (Ridker et al., 2008; Sever et al., 2013; Soedamah‐Muthu et al., 2015). In addition, the different characteristics of the subjects included in these studies may have influenced the results. In fact, the JUPITER study included apparently healthy subjects with baseline LDL‐C levels <130 mg·dL⁻¹ (but CRP ≥2 mg·L⁻¹) (Ridker et al., 2008), while CARDS enrolled type 2 diabetes patients (Soedamah‐Muthu et al., 2015) and ASCOT hypertensive patients with other ≥3 other CV risk factors (Sever et al., 2013).

Furthermore, a meta‐analysis, which included 23 studies with statins, reported that 89 ‐ 98% of the CRP reduction was directly related to the degree of LDL‐C reduction obtained (Kinlay, 2007). Thus, although individual statin studies have shown a poor relationship between CRP reduction and LDL‐C reduction, when data are analysed as aggregate, a strong correlation can be observed. The possible relationship between LDL‐C and CRP reductions in different clinical settings might be under‐estimated, as many authors reported a lack of correlation for individual data. Although this may be explained by large measurement error and large intra‐individual variations in CRP levels, we would like to offer a different interpretation that relies on individual variability to the cholesterol and lipid burden‐mediated inflammatory response in tissues that makes each of us unique as responder. This interpretation will also explain why CRP is a very good predictor of event, independent of LDL; because it captures the information derived from individual variability.

Indeed statin therapy was associated also with the reduction of IL‐6, TNF‐α and cell adhesion molecule levels in patients with CV risk factors (van de Ree et al., 2003; Ascer et al., 2004). In the MIRACL study, the reduction of IL‐6 achieved with atorvastatin therapy was associated with a relative reduction of the risk of stroke after an acute coronary syndrome (Kinlay et al., 2008).

Apart from the beneficial effects on inflammatory markers, statins also improve endothelial function in patients at high CV risk or with CAD (Fichtlscherer et al., 2006). Several meta‐analyses have evaluated the benefit of statins on endothelial function in different cohorts. The data indicate that statin therapy is associated with significant improvement in both peripheral and coronary endothelial function (Reriani et al., 2011). Of note, controversies exist among trials reporting the effects of statins on endothelial dysfunction in patients with diabetes mellitus and while the overall finding was that statins improve endothelial function also in diabetics, patients with a more compromised endothelial function are less likely to benefit from statin treatment (Zhang et al., 2012). A key gap in these meta‐analyses is that it is not known whether the benefit of statins on endothelial function is proportional to the reduction of LDL‐C levels.

Aside from affecting endothelial function, statins were also demonstrated to favour atherosclerosis regression, by reducing the percent atheroma volume (PAV), as determined by intravascular ultrasound (IVUS). This was clearly shown in the SATURN and ASTEROID trials (Nicholls et al., 2011), where average on‐treatment LDL‐C levels below 70 mg·dL⁻¹ for 24 months were associated with a significant PAV reduction. A pre‐specified post hoc analysis later indicated that CRP levels, but not LDL‐C levels, were associated with coronary atheroma regression and CV events. However, the absolute change in CRP was not prognostic of major CV events (Puri et al., 2013).

All these observations clearly point to a beneficial effect of statins in inflammation, which is reflected in a decrease of circulating inflammatory markers such as CRP and improvement of endothelial function and plaque burden.

Effects of LDL‐C‐lowering therapies other than statins on vascular inflammation and outcomes

While there is no doubt that statins improve systemic and vascular inflammation, the rationale for using a statin to decrease CRP and therefore to target a causal factor in vascular inflammation associated with atherosclerosis should be carefully considered. Indeed, several analyses have shown that CRP does not play a causal role in the pathogenesis of atherosclerosis (Zacho et al., 2008; Elliott et al., 2009; Wensley et al., 2011) and the acute manifestations of CAD. On the basis of these results, it is likely that the effect of statins on CRP could mirror the beneficial effect of this class of drugs on other atherogenic players, rather than a direct anti‐inflammatory/vasculoprotective effect mediated through CRP reduction. This leaves the room open for considering that statins might exert their anti‐inflammatory activities through the ability of reducing the pro‐inflammatory/pleiotropic effects of cholesterol‐rich lipoproteins. If this is the case, then pharmacological approaches aimed at reducing LDL‐C via mechanisms that are independent of HMG‐CoA reductase inhibition should demonstrate a benefit in terms of CRP reduction and vascular inflammation, which should be proportional to the reduction of LDL‐C achieved. This is the case for some but not all lipid lowering agents.

Ezetimibe, by inhibiting the cholesterol transport protein NPC1‐like 1, reduces the intestinal absorption of cholesterol and is utilized in clinical practice as monotherapy or as an adjunct to statins (Sudhop et al., 2009). This drug has been extensively studied for the ability not only to decrease plasma LDL‐C levels but also to decrease CRP and improve CV outcome. Two‐pooled analyses of randomized, placebo‐controlled trials of ezetimibe in hypercholesterolemic patients have demonstrated that patients treated with placebo had a 5% increase in CRP levels while those treated with ezetimibe had a 1% decrease, with a between‐treatment difference of 6%, that however was not statistically significant (Pearson et al., 2009). LDL‐C decreased significantly with ezetimibe (−18%) compared with placebo (+0.5%) (Pearson et al., 2009). When ezetimibe was added to a statin, CRP was significantly reduced compared with statin monotherapy (−12% vs. −1%), as it was LDL‐C (−27% vs. −3%) (Pearson et al., 2009). When data available from the different clinical trials with ezetimibe are investigated and the two arms (statins only or statins + ezetimibe) are compared, the reduction in CRP levels is proportional to the reduction observed in LDL‐C levels (Figure 2). The correlation between the changes of the two parameters is similar in the statins only and in the statin + ezetimibe group (Figure 2).

Figure 2.

Association between reduction in LDL‐C levels and reduction in CRP levels among clinical trials comparing statin monotherapy with ezetimibe/statin combination. The following studies were included in the figure: VYVA: Ballantyne, Am Heart J, 2005; VYMET: Robinson, Am J Cardiol, 2009; EXPLORER: Ballantyne, Am J Cardiol, 2007; ENHANCE: Kastelein, NEJM, 2008; SANDS: Fleg, J Am Coll Cardiol, 2008; A: Sager, Atherosclerosis, 2005; B: Goldberg, Mayo Clin Proc, 2004; C: Bays, Clin Ther, 2004; D: Ballantyne, Circulation, 2003; E: Gagne, Am J Cardiol, 2002.

Studies with other lipid lowering drugs, including those inhibiting the synthesis of apolipoprotein B containing lipoproteins, such as lomitapide or mipomersen (Norata et al., 2013), as well as those increasing LDL‐C metabolism, such as inhibitors of the proprotein convertase subtilisin/kexin type 9 (PCSK9) (Norata et al., 2014; Norata et al., 2016) or CETP inhibitors have not investigated in detail the effects on CRP reduction and the few available data do not show a robust correlation between the magnitude of LDL‐C reduction and that of CRP (Flaim et al., 2014; Stroes et al., 2014; Cannon et al., 2015; Hovingh et al., 2015; Raal et al., 2015a, b; Nicholls et al., 2016; Sahebkar et al., 2016). While it could appear surprising that PCSK9 inhibitors, which decrease LDL‐C levels up to 60% (Robinson et al., 2015; Sabatine et al., 2015), do not affect CRP levels, also in the placebo arm (statins and/or ezetimibe treated patients) of almost all clinical trials, no significant changes in CRP plasma levels were observed (Figure 3). These findings could be explained by noting that, in clinical trials with anti‐PCSK9 therapies, the median baseline levels of CRP were below 2 mg·L⁻¹ (also in those trials who tested statin‐intolerant patients), thus excluding patients with systemic inflammation from these studies and perhaps limiting the possibility to appreciate a beneficial effect of lipid‐lowering drugs in spite of LDL‐C reduction. Indeed when clinical trials with CRP levels above 2 mg·L⁻¹ are considered, CRP is reduced by lipid‐lowering therapies, independently of the mechanism of action (Figure 3). Furthermore, even in the SATURN study, rosuvastatin, in spite of a LDL‐C reduction of −43.5%, decreased CRP levels only in those patients with CRP baseline level of 2.3 mg·L⁻¹ (0.8 mg·L⁻¹ at follow‐up), while no effect or rather a small increase was reported in the group with CRP baseline level of 1.1 mg·L⁻¹ (1.6 mg·L⁻¹ at follow‐up) (Puri et al., 2013). This further suggests that the beneficial effect of LDL‐C‐lowering therapies on systemic inflammatory status, as monitored by changes in CRP plasma levels, is evident only in patients presenting with increased inflammatory conditions. Could these findings be transferred also to markers of vascular dysfunction and inflammation?

Figure 3.

Effects of lipid‐lowering therapies on plasma CRP levels reduction according to CRP baseline levels.

The PANACEA study showed that, in obese patients with metabolic syndrome, a low‐dose statin/ezetimibe combination (10/10 mg) not only resulted in a similar of LDL‐C reduction compared to high‐dose statin monotherapy (simvastatin 80 mg) but also in a comparable endothelial function as determined by both flow‐mediated dilatation (FMD) and peripheral arterial tonometry (EndoPAT) (Westerink et al., 2013). Similar observations were previously reported also in other studies. Stable CAD patients with dysglycemia that commonly are dyslipidemic but also present endothelial dysfunction and vascular inflammation, were treated with simvastatin 80 mg or ezetimibe/simvastatin 10/10 mg for 6 weeks. Similar LDL‐C reductions from the baseline were obtained as well as similar improvement of FMD (Settergren et al., 2008). Similarly, obese women with LDL‐C ≥100 mg·dL⁻¹ treated with simvastatin 80 mg or ezetimibe/simvastatin 10/10 mg for 8 weeks showed similar reductions of LDL‐C levels (−27% and −30%, respectively) and similar increases of FMD (+39% and +41%, respectively) (Garcia et al., 2016). These observations, suggesting that lipid‐lowering per se rather than pleiotropic properties of statins plays a key role in the improvement of endothelial function, were further supported by the results of a meta‐analysis including six trials with 213 participants which compared high‐dose statin versus low‐dose statin combined with ezetimibe (Ye et al., 2012). The two lipid‐lowering regimens induced similar reductions of LDL‐C (P = 0.12 between treatments) and CRP levels (P = 0.89) and similar increments of FMD (P = 0.68), suggesting similar beneficial effects. In line with these findings, we have reported that pravastatin and ezetimibe reduced LDL‐C levels similarly and increased FMD at a comparable extent in subjects with moderate hypercholesterolemia (Grigore et al., 2013).

Not all studies are in agreement with these findings. Treatment of dyslipidemic subjects without signs of CAD with simvastatin 40 mg or ezetimibe/simvastatin 10/10 mg for 4 weeks resulted in the comparable decreases in LDL‐C levels (−38.5% vs. −34.8%), while FMD was improved in subjects treated with simvastatin 40 mg but not in those who received the combination therapy (Liu et al., 2009). Patients with CAD treated with atorvastatin 40 mg (either de novo therapy or dose escalation from chronic 10 to 40 mg) had improvement in their endothelial function, whereas those treated with ezetimibe 10 mg alone or added to chronic simvastatin 10 mg did not show changes, despite LDL‐C levels decreased similarly with all considered therapeutic regimens (Fichtlscherer et al., 2006). Similar results were observed in patients with chronic heart failure, in which simvastatin 10 mg or ezetimibe 10 mg reduced LDL‐C levels to a similar extent but only simvastatin improved endothelial function (Landmesser et al., 2005).

Of note, the results from the GLAGOV trial, which was aimed at investigating the effects of PCSK9 inhibition in patients with angiographic coronary disease, showed a greater decrease in PAV after 76 weeks of treatment with evolocumab compared with statins alone (Nicholls et al., 2016). Baseline CRP levels were 1.6 mg·L⁻¹ and were unaffected by the treatment, further supporting a beneficial effect of LDL‐C lowering on atherosclerosis, which was independent of systemic inflammatory status (Nicholls et al., 2016).

In summary, most studies indicate that lipid‐lowering per se may play a key role in reducing vascular inflammation, although the well‐established pleiotropic effects of statins still leave the question open.

Conclusion

Historically, atherosclerosis has been seen as the consequence of impaired lipid metabolism, which promotes endothelial dysfunction and cholesterol deposition into the plaque, thus resulting in the recruitment of inflammatory cells. This response to injury becomes uncontrolled and generates the inflammatory response. On this basis, therapeutic approaches aimed at controlling either the excess of lipoproteins or the dampening the inflammatory response were tested. While most of the studies with anti‐inflammatory agents did not show a relevant benefit in terms of CV risk reduction, statins, by promoting LDL‐R expression and LDL‐C reduction, demonstrated a significant and robust benefit in terms of CV risk reduction. Later this benefit was, at least in part, ascribed to the ability of statins to reduce CRP. However, Mendelian randomization studies have clearly demonstrated that CRP is not a causal factor for atherosclerosis but rather a marker of systemic inflammation. The question of why statins by decreasing CRP levels, independently of LDL‐C reduction, improved CV risk has been involving the scientific community for the last 15 years. Lipid lowering treatments such as ezetimibe or anti‐PCSK9 monoclonal antibodies, which reduce LDL‐C through mechanisms independent of the inhibition of HMG‐CoA reductase, result in CRP reduction only in those patients with baseline levels above 2 mg·L⁻¹, which is the threshold indicating the presence of an inflammatory condition. On the contrary, in patients with CRP below 2 mg·L⁻¹, any type of LDL‐C‐lowering treatment, from statins to ezetimibe to anti‐PCSK9 antibodies, do not change CRP levels in spite of a substantial LDL‐C reduction, probably because these patients do not present a relevant systemic inflammation. Of note, even in patients where CRP is below 2 mg·L⁻¹ and not altered by the therapy, such as in the GLAGOV study, LDL‐C reduction with anti‐PCSK9 antibodies results in atherosclerotic plaque regression as determined by IVUS (Nicholls et al., 2016). As a further proof of concept directly linking LDL‐C with vascular impairment, data demonstrating that LDL‐C reduction via different approaches will translate into a decrease of CV events also when baseline CRP levels are below 2 mg·L⁻¹ (thus not expected to be modulated by the therapy) have been recently published (Sabatine et al., 2017). The FOURIER trial showed that, in more than 27 500 patients with atherosclerotic disease, the addition of evolocumab to an optimized regimen of lipid‐lowering therapy significantly reduced the risk of CV events as a consequence of LDL‐C reduction (from 92 to 30 mg·dL⁻¹), compared to that of patients maintaining the optimized regimen of lipid‐lowering therapy (Sabatine et al., 2017). Of note, the level of CRP was 1.7 mg·L⁻¹ (IQR 0.9–3.6) at baseline and 1.4 mg·L⁻¹ (IQR 0.7–3.1) after 48 weeks of treatment in both arms (Sabatine et al., 2017).

Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Southan et al., 2016), and are permanently archived in the Concise Guide to PHARMACOLOGY 2015/16 (Alexander et al., 2015a,b).

Conflict of interest

The authors has received research funding, and/or honoraria for advisory boards, consultancy or speaker bureau from Aegerion (ALC, GDN), Amgen (ALC, GDN), AstraZeneca (ALC), Eli Lilly (ALC), Genzyme (ALC), Mediolanum (ALC), Merck or MSD (ALC), Pfizer (ALC, GDN), Recordati (ALC, GDN), Rottapharm (ALC), Sanofi‐Regeneron (ALC, GDN) and Sigma‐Tau (ALC). A.P. reports no disclosures.

Catapano, A. L. , Pirillo, A. , and Norata, G. D. (2017) Vascular inflammation and low‐density lipoproteins: is cholesterol the link? A lesson from the clinical trials. British Journal of Pharmacology, 174: 3973–3985. doi: 10.1111/bph.13805.