AERNOUD T.L. FIOLET USING the AVAILABLE to ENABLE the SUSTAINABLE
Colchicine in Coronary Disease. Using the available to enable the sustainable. PhD Thesis with summary in Dutch. Utrecht University Cover design: James Jardine | www.jamesjardine.nl Layout: James Jardine | www.jamesjardine.nl Print: Ridderprint | www.ridderprint.nl ISBN: 978-94-93108-30-1 Copyright © Aernoud T.L. Fiolet. All rights reserved. No part of this thesis may be reproduced, stores or transmitted in any way or by any means without the prior permission of the author, or when applicable, of the publishers of the scientific papers.
Colchicine in coronary disease Using the available to enable the sustainable Colchicine in coronairlijden Met beschikbare middelen het bestendige mogelijk maken (met een samenvatting in het Nederlands) Proefschrift ter verkrijging van de graad van doctor aan de Universiteit Utrecht op gezag van de rector magnificus, prof.dr. H.R.B.M. Kummeling, ingevolge het besluit van het college voor promoties in het openbaar te verdedigen op donderdag 24 februari des middags te 12:15 uur door Aernoud Theodore Llewelyn Fiolet geboren op 5 april 1989 te Geleen
Promotoren: Prof. dr. D.E. Grobbee Prof. dr. P.A.F.M. Doevendans Prof. dr. J.H. Cornel Copromotor: Dr. A. Mosterd Het onderzoek beschreven in dit proefschrift werd (mede) mogelijk gemaakt met financiële steun van de Nederlandse organisatie voor gezondheidsonderzoek en zorginnovatie (ZonMw) in combinatie met de Hartstichting, de Withering Stichting Nederland, de Werkgroep Cardiologische Centra Nederland (WCN) en het Meander Medisch Centrum. De druk van dit proefschrift werd mede mogelijk gemaakt door financiële bijdragen van de Hartstichting en de Stichting Wetenschappelijk Onderzoek Hart en Vaatziekten Amersfoort. Daarnaast werd de druk mogelijk gemaakt door een bijdrage van IQVIA.
Beoordelingscommissie: prof. dr. ir. H.M. den Ruijter (voorzitter) prof. dr. S.A.J. Chamuleau prof. dr. A.W. Hoes prof. dr. H.G.M. Leufkens prof. dr. E. Lutgens Paranimfen: Drs. Tim R. van der Hoeven Drs. Max J.M. Silvis
Opgedragen aan mijn ouders en aan Isabella, Fiene en Lou; in de hoop dat ook zij zich een leven lang mogen verwonderen
Chapter 1 General introduction and thesis outline 11 PART I: The role of crystal-induced inflammation in atherosclerosis Chapter 2 Viewing Atherosclerosis through a Crystal Lens: How the Evolving Structure of Cholesterol Crystals in Atherosclerotic Plaque Alters its Stability Journal of Clinical Lipidology 2020;14:619-630 51 Chapter 3 Colchicine in Stable Coronary Artery Disease Clinical Therapeutics 2019;41:30-40 73 PART II: Efficacy and safety of colchicine in coronary disease Chapter 4 The effect of low-dose colchicine in patients with stable coronary artery disease: The LoDoCo2 trial rationale, design, and baseline characteristics American Heart Journal 2019;218:46-56 93 Chapter 5 Colchicine in Patients with Chronic Coronary Disease New England Journal of Medicine 2020;383:1838-1847 117 Chapter 6 Reply. Colchicine in Patients with Chronic Coronary Disease New England Journal of Medicine 2021;384:776-779 161 Chapter 7 The efficacy and safety of low-dose colchicine in patients with coronary disease: a systematic review and meta-analysis of randomised trials European Heart Journal 2021;42:2765-2775 173 Chapter 8 Colchicine in chronic coronary disease in relation to history of acute coronary syndrome Journal of the American College of Cardiology 2021;78:859-866 215 PART III: Insights in mechanism and tolerance Chapter 9 Short-term effect of low-dose colchicine on inflammatory biomarkers, lipids, blood count and renal function in chronic coronary artery disease and elevated high-sensitivity C-reactive protein PLOS ONE 2020;15:e0237665 233 TABLE OF CONTENTS
Chapter 10 Colchicine Attenuates Inflammation Beyond the Inflammasome in Chronic Coronary Artery Disease. A LoDoCo2 Proteomic Substudy Circulation 2020;142:1996-1998 257 Chapter 11 Colchicine reduces extracellular vesicle NLRP3 inflammasome protein levels in vitro and in vivo in chronic coronary disease: a LoDoCo2 substudy Atherosclerosis 2021;334:93-100 267 Chapter 12 Predictors of early intolerance to low-dose colchicine in patients with coronary disease In preparation 291 PART IV: Improving pragmatic trial conduct Chapter 13 Text-mining in electronic healthcare records can be used as efficient tool for screening and data collection in cardiovascular trials: a multicenter validation study Journal of Clinical Epidemiology 2021;132:97-105 313 Chapter 14 Accuracy of using routinely collected electronic healthcare data to identify major cardiovascular endpoints in randomized controlled trials: a multicenter validation study In preparation 337 PART V: General discussion and summary Chapter 15 Clinical implications and future perspectives 359 Chapter 16 Summary 381 Appendices Summary in Dutch 391 List of publications 399 Acknowledgements 405 About the author 419
01
General introduction and thesis outline Parts of this chapter have been published before.1,2,3 1. Fiolet ATL, Nidorf SM, Cornel JH. Colchicine for secondary prevention in coronary disease. European Heart Journal 2021;42:1060-1061 2. Fiolet ATL, Opstal TSJ, Silvis MJM, Cornel JH, Mosterd, A. Targeting residual inflammatory risk in coronary disease: to catch a monkey by its tail. Netherlands Heart Journal, in print 3. Fiolet ATL, Opstal TSJ, Thompson PL, Cornel JH. Colchicine in Acute Coronary Syndrome: When to Commence? American College of Cardiology. http://www.acc.org. Sept 15, 2021.
13 General introduction and thesis outline GENERAL INTRODUCTION This general introduction describes the scientific prelude that incited the ideas and hypotheses investigated in this thesis. Firstly, it describes the role of inflammation in medicine and the particular mechanisms of inflammation that contribute to atherogenesis and atherotrombosis. Secondly, it describes the available clinical data on the potential to modulate these processes. Lastly, it discusses the contemporary challenges encountered in drug research and postulates methodological alternatives. INFLAMMATION IN MEDICINE Aulus Cornelius Celsus (c. 25 BC–c. 50 AD) is credited with the first proper description of inflammation, coining the tetrad of calor (warmth), dolor (pain), tumor (swelling), and rubor (redness and hyperaemia).1 After this first description by Celsus the term inflammation has undergone some form of devaluation to an idle term used by many in widely divergent diseases. Especially in cardiovascular medicine, where it unfruitfully has been announced as the “next clinical breakthrough” for several decades. However, the dissolving of the term inflammation is unjustified, although the term indeed can be considered somewhat undescriptive. No argument has to be made for the quintessential role of inflammation in atherothrombosis. Inflammatory mechanisms play pivotal roles in all stages of atherosclerosis and atherothrombosis.2,3 (Figure 1) Firstly, we distinguish inflammation in endothelial dysfunction and in response to pathological metabolic states (obesity, hyperglycemia, hyperlipidemia). Secondly, inflammatory mechanisms initiate the formation of a fatty streak and the progression to fibroatheroma. Thirdly, thromboembolic complications may occur after inflammation-initiated plaque destabilisation. And finally, initiated by the ischemic sequelae of the aforementioned processes, many inflammatory responses occur in the damaged cardiomyocytes.4,5 These various physiological and pathophysiological processes are all described by the term “inflammation.” Inflammation—the biological response to repel pathogens and/or (subsequent) tissue damage—encompasses both our innate and our adaptive immunological response. Both elements play a similarly important role in the
14 Chapter 1 initiation and development of atherothrombotic plaque. Moreover, both play a crucial role in the metamorphosis of a chronic-stable disease to an acute-unstable disease. Inflammation in atherothrombosis is complex, and to understand how to modulate such mechanisms, one first needs to establish the basic elements of the immunological response that follow the many vascular threats experienced during lifetime. The understanding of this interplay of protective mechanisms has been slowly unravelled by years of experimental and clinical research. Paradoxically, the more actors that are identified, the more limited our actual understanding seems. However, when accurately aligning the windows of insights offered to us by preceding scientists, we may still glimpse ways to intervene—albeit these insights always come with new biological and methodological challenges. We then may come to realise, freely paraphrasing Sir William Osler, that he who understands inflammation, understands medicine. THE BIOLOGICAL PERSPECTIVE: BASIC ELEMENTS OF THE IMMUNOLOGICAL RESPONSE IN ATHEROTHROMBOSIS Endothelial dysfunction A hallmark in the origin of atherothrombosis is endothelial dysfunction, which comprises increased endothelial permeability to lipoproteins, leukocyte adherence, a prothrombotic state, and decreased vasomotor function.4,6,7 These alterations in endothelial state first occur in predisposing regions. These are the regions that have calibre changes or bifurcations that lead to the disruption of the laminar blood flow, resulting in a more turbulent flow with lower shear stress encountered by the endothelium.8 Hypertension, oxidative stress, and hypercholesterolaemia further induce the endothelial changes, which are mediated by transcriptional factors such as nuclear factor kB.9 Although endothelial dysfunction denotes impairment of any of its properties, its inability to adequately maintain vascular tone is used clinically as an objective measure and is associated with future development of atherothrombosis.10 Initiation of the fatty streak and fibroatheroma Following the increased permeability of the endothelium, ApoB-containing lipoproteins will enter the extracellular matrix of the intima. Most notably are
15 General introduction and thesis outline low-density lipoproteins (“LDL”).11 On entering the intima, oxidation occurs. The oxidised LDL acts as a danger-associated molecular-pattern for Toll-like receptors of the endothelial cell.12 In an attempt to avert these undesirable compounds, the immune system is activated in various ways. This is accomplished in the first place by increasing the adhesiveness of the endothelium by an increased expression of “vascular cell adhesion molecule – 1” (VCAM-1)ontheintimalsideoftheendothelialcell,whichisregulatedbycytokines.13 Such upregulated expression is seen within a week of initiating an atherogenic diet in an experimental setting.14 This adhesion molecule will most importantly interact with the “very late activation antigen 4”, a so-called integrin. These integrins can be found on the outer side of monocytes rolling down the endothelium and are activated by chemokines expressed by the endothelium. The most important chemokine responsible for the attraction of these monocytes is “monocyte chemoattractant protein 1”. This specific chemokine is mainly expressed in response to oxidised LDL in the intima and not so much by native LDL.15,16 The receptor for monocyte chemoattractant protein 1 is CCR2.17 The importance of this receptor and the link between the oxidised LDL and monocyte attraction via the monocyte attractant protein 1 is evidenced by the upregulation of CCR in hypercholesterolaemia. This upregulation increases the sensitivity of the “monocyte-to-monocyte attractant protein 1”.18 This chemokine is localised in the subendothelial space and leads the monocyte from the endothelial surface into the intima.19 Thus, by increasing the expression of adhesion molecules on the endothelium and by activating the integrins on the monocyte, the rolling monocyte is halted and firm adhesion to the endothelium occurs, after which chemokines introduce the monocyte into the intima.20 Monocytes, which become macrophages under the influence of “monocyte-colony stimulating factor” produced by endothelial cells and bone marrow, have multiple purposes.21,22 They serve as scavenger cells, antigen-presenting cells, and secretory cells, with an extremely diverse spectrum of secretion. Conventionally, after scavenging the injured site, macrophages migrate from the injury. During a longer stay in a lesion, the macrophage differentiates into specific phenotypes. The M1 phenotype macrophage is stimulated by pro-inflammatory cytokines and oxidised LDL and expresses multiple pro-inflammatory cytokines itself, maintaining the pro-inflammatory state. Its effect are counterbalanced by M2 macrophages, which have anti-inflammatory effects.23 M1 macrophages are dominant in the shoulder regions of the growing plaque, while M2 macrophages are found in the
16 Chapter 1 adventitia.24,25 This classic dichotomous subdivision of macrophage phenotypes is probably too simplified, as a more diverse continuum of cell type differentiation may be found in the plaque.26 The CD36 receptor is the most notable scavenger receptor for binding oxidised LDL in the intima. Lipid accumulation in the macrophage does not suppress CD36 expression, making room for unrestricted lipid accumulation.27 Engulfing the lipids, large numbers of macrophages develop lipid droplets, giving them a foamy appearance (“foam cells”).28 The large number of macrophages filled with lipid droplets finally perish and release their intracellular contents. This occurs mainly due to apoptosis and secondary necrosis from the toxicity of the oxidised LDL and hypoxia. Without effective efferocytosis—the immunological phenomenon where apoptotic cells are cleared out—large pools with cellular debris form.29–31 Summarizing, the macrophage thus plays a distinct role in the initiation and morphological changes of atherosclerotic plaque.32 The learned innate immune system Classically, the innate immune system is said to operate without an “immunological memory.” However, for longer, observations have suggested the existence of “trained innate immunity.”33 After genetically eliminating the adaptive immune system of mice that were administered a vaccination for tuberculosis and introduced to diverse second inflammatory stimuli at a later stage, all demonstrated augmented monocyte production. This proves that the innate immune system is able to increase its response after second similar and different threats, without the involvement of the adaptive immune system.34 The change to a pro-inflammatory monocyte occurs via DNA methylation or “epigenetic modification.” Not only exogenous pathogens, such as bacteria, can induce this mechanism of “trained innate immunity”, but endogenous antigens may also introduce this epigenetic reprogramming of the monocyte. This indicates the importance of “trained innate immunity,” in atherosclerosis as well.35 Indeed, oxidised LDL acting as the first stimulus to such monocytes produces highly pro-atherogenic monocytes.36 The controversial association between traces of microorganisms (such as Chlamydia pneumoniae or viruses) in atherosclerotic plaques, and their role in the development of these plaques may also relate to this mechanism.35
17 General introduction and thesis outline Figure 1, A and B. Development of the atherosclerotic lesion. The normal muscular artery and the cell changes that occur during disease progression to thrombosis are shown. Panel A shows the normal artery that contains three layers. The inner layer, the tunica intima, is lined with a monolayer of endothelial cells that is in contact with the blood over the basement membrane. In contrast to many animal species used for atherosclerosis experiments, the human intima contains resident smooth muscle cells. The middle layer, or tunica media, contains smooth muscle cells embedded in a complex extracellular matrix. Arteries affected by obstructive atherosclerosis generally have the structure of muscular arteries. The arteries often studied in experimental atherosclerosis are elastic arteries, which have clearly demarcated laminae in the tunica media, where layers of elastin lie between the strata of smooth muscle cells. The adventitia, the outer layer of arteries, contains mast cells, nerve endings, and microvessels. Panel B shows the initial steps of atherosclerosis that include the adhesion of blood leukocytes to the activated endothelial monolayer, directed migration of the bound leukocytes into the intima, maturation of monocytes (the most numerous of the leukocytes recruited) into macrophages, and their uptake of lipids, yielding foam cells. Adapted from Libby et al.37
18 Chapter 1 Figure 1, C and D. Development of the atherosclerotic lesion. Panel C shows lesion progression that involves the migration of smooth muscle cells from the media to the intima, the proliferation of resident intimal smooth muscle cells and media-derived smooth muscle cells, and the heightened synthesis of extracellular matrix macromolecules, such as collagen, elastin, and proteoglycans. Plaque macrophages and smooth muscle cells can die in advanced lesions, some by apoptosis. Extracellular lipids derived from dead and dying cells can accumulate in the central region of plaque, often denoted the lipid or necrotic core. Advanced plaques also contain cholesterol crystals and microvessels. Panel D shows how thrombosis, the ultimate complication of atherosclerosis, often complicates the physical disruption of the atherosclerotic plaque. Panel D also shows a fracture of the plaque's fibrous cap, which has enabled blood coagulation components to come into contact with tissue factors in the plaque's interior, triggering the thrombus that extends into the vessel lumen, where it can impede blood flow. Adapted from Libby et al.37
19 General introduction and thesis outline SMALL-SIZED BUT CRUCIAL ACTORS: CYTOKINES Not surprisingly, inflammatory mediators that play important roles in autoimmunological pathologies and proliferative conditions are also seen at higher concentrations in and around atherothrombotic lesions. There are several families of inflammatory mediators all with both proinflammatory and anti-inflammatory subsets: chemokines, growth factors, and most notably cytokines. More than 60 different cytokines have been described. Divided into several families, the interleukin family encompasses the largest subset of cytokines, next to tumour necrosis factor and interferons.38 Cytokines enable communication between leukocytes as well as endothelium and vascular smooth muscle cells in autocrine, paracrine, and endocrine ways.39 In vivo and in vitro data have revealed cytokines with pro- and anti-inflammatory properties, in balanced equilibrium in physiological circumstances. Important proinflammatory cytokines include interleukin – 1, interleukin – 6, tumour necrosis factor α, and interferon - γ. Important anti-inflammatory cytokines include interleukin – 10 and transforming growth factor β. They are regarded as antiinflammatory and are secreted by M2 macrophages. They induce atheroregression using complex and not fully understood mechanisms.40 Interleukin – 6 Two cytokines, interleukin – 1 and interleukin – 6, have been shown to play a particularly important role in atherothrombosis. Interleukin – 6 levels are increased relative to the systemic concentration around ruptured plaques, and human atherosclerotic plaques have increased gene expression for interleukin – 6.41,42 Interleukin – 6 plays a pivotal role in immunology, since it has a direct effect on leukocyte recruitment by endothelial cells and amplifies the inflammatory response by inducing chemokine release.43 The observation that interleukin – 6 is necessary for the recruitment of inflammatory cells in atherosclerotic plaque was confirmed in several atherosclerosis mouse models. Exogenous elevated levels of interleukin – 6 by the administration of recombinant interleukin – 6 in obese mice increases atherosclerotic plaques and increases other inflammatory mediators, such as tumour necrosis factor-α and interleukin – 1. In addition, blocking trans-signaling of interleukin – 6 in mice reduces monocyte infiltration and atherosclerotic plaque progression.44 However, interleukin – 6
20 Chapter 1 deficient atherosclerotic-prone mice models have lower inflammatory cell influx but increased lipid deposition and atherosclerotic plaque formation, underscoring an initially protective role of the inflammatory response in lipid accumulation.45,46 Complementary to the animal studies are epidemiological observations of interleukin – 6. Elevated concentrations of interleukin – 6 are associated with an increased risk of major adverse cardiovascular events, even when corrected for traditional cardiovascular risk factors known to influence circulating interleukin – 6, such as adiposity and smoking.47,48 Interleukin – 1 In similar ways to interleukin – 6, a strong association between interleukin – 1 and atherosclerosis exists. Interleukin – 1 increases endothelial adhesion molecules, as well as cytokine and chemokine expression, and thereby helps with the recruitment of leukocytes.49 The interleukin – 1 family consists of 11 cytokines. They are expressed by all cells of the innate immune system. Auto-inflammatory and auto-immune diseases, as well as degenerative and infectious diseases, all are strongly driven by interleukin – 1 communication. Interleukin – 1 has an effect on the hypothalamus– pituary–adrenal axis (resulting in fever and increased cortisol release), the acutephase response in the liver (mediated via interleukin – 6), induction of endothelial adhesion molecules, and the lifespan of neutrophils and macrophages. In addition, interleukin – 1 has self-promoting properties, amplifying its own release.50 The three most-studied members of the family are interleukin – 1α, interleukin – 1β, and interleukin – 1 receptor antagonist. These three cytokines compete in binding to the interleukin – 1 receptor.51 The interleukin – 1α precursor is present in epithelial cells and is released after cell death. It is fully active, acting in the early response in sterile inflammation.52 The membrane form of active interleukin – 1α is found in activated monocytes.50,53 Interleukin – 1β is the main form of circulating interleukin – 1 and it is produced by hematopoietic cells, such as monocytes/macrophages. The precursor of interleukin – 1β is not active until cleaved by caspase 1, after which the active cytokine will be released in the extracellular compartment. Although interleukin – 1α and interleukin – 1β differ in source and release mechanism, both bind to the interleukin – 1 receptor and thus initiate similar inflammatory cascades.51
21 General introduction and thesis outline Experimental evidence for the role of interleukin – 1 in atherosclerosis is well established. Interleukin – 1-deficient hypercholesterolemic mice models have decreased atherosclerotic lesion size, lower blood pressure levels, and lower levels of endothelial dysfunction. In particular, these findings were more frequently observed in those that are deprived of vessel wall interleukin – 1 signalling than of those that are deprived of bone marrow interleukin – 1.54 Carotid atherosclerotic lesions are reduced in animals for which the interleukin – 1 receptor antagonist is exogenously increased, and the expression of adhesion molecules and monocyte chemotactic protein 1 are all reduced in mice lacking interleukin – 1β.55,56 Interleukin – 1β and the interleukin – 1 receptor antagonist are both present in human atherosclerotic lesions.57 Direct evidence of interleukin – 1b in human atherosclerotic lesions and prognostic information is scarce due to difficulties encountered when measuring levels of the cytokine. The interleukin – 1 receptor antagonist can be measured with greater precision.51 Claims have been made that increased levels of interleukin – 1 receptor antagonist may correlate with increased interleukin – 1β activity, making it usable as surrogate marker, though this is controversial.58 Increased levels of circulating interleukin – 1 receptor antagonist are indeed associated with an increased risk of cardiovascular disease.59 CAUSALITY IN INFLAMMATORY MEDIATORS AND ATHEROTHROMBOSIS Similar to interleukin – 6, there is evidence of an increased risk of cardiovascular disease when interleukin – 1–binding cytokine levels are increased. These inflammatory mediators—like many inflammatory mediators—lead to downstream production of several acute-phase protein such as pentraxins (e.g., pentraxin 3 and C – Reactive protein) and serum amyloid A proteins.60 (Figure 2) Increased levels of these acute phase proteins may thus be the direct consequence of increased atherosclerotic inflammation mediated by interleukin – 6 and interleukin – 1. However, many other pathophysiological processes can cause a similar increased expression. Many of these pathophysiological processes have equal important roles in the development of atherothrombosis and consequent complications. Thus, observing an association of elevated levels of circulating inflammatory biomarkers and the risk for cardiovascular disease does not necessary provide new mechanistic insights, but may only affirm established
22 Chapter 1 Figure 2. The cytokine cascade. Activated immune cells in the plaque produce inflammatory cytokines (interferon-γ, interleukin – 1, and tumour necrosis factor), which induce the production of substantial amounts of interleukin-6. These cytokines are also produced in various tissues in response to infection, as well as in the adipose tissue of patients with metabolic syndrome. Interleukin – 6, in turn, stimulates the production of large numbers of acute-phase reactants, including C-Reactive protein (CRP), serum amyloid A, and fibrinogen, especially in the liver. Although cytokines at all steps have important biological effects, their amplification at each step of the cascade makes the measurement of downstream mediators, such as CRP, particularly useful for clinical diagnosis. Adapted from Hansson.3
23 General introduction and thesis outline biological relationships viewed from a different perspective. In this regard one should realise that vascular threats come from many behavioural and biological pathophysiological concepts. In most cases, one should therefore see a biomarker primarily as just yet one of many links, rather than as the missing link.61 In animal experiments, the roles of such cytokines are investigated using transgenic or gene-deficient mouse models, in which certain receptors, cytokines, or cell lines are absent (the “knockout model”) or over-expressed. The effects of under- or over-expression of specific components that are thought to be of immunological importance in atherosclerosis in humans are investigated in population-wide geneassociation studies. In this sort of research, known genetic variations causing loss of function or over-expression caused by single-nucleotide polymorphisms are used as “nature’s own experiment.” When investigating the effects of certain types of (immunological) exposure in a conventional observation study design, causality as origin of the association is often hard to prove since the exposure might merely be the derivate of a different, yet unknown, causal factor. However, exactly those studies that use these genetic polymorphisms (or: “Mendelian randomisation” studies) remove confounding by any acquired disease-influencing factors and supply the strongest evidence of a possible causal relationship between the factor and disease outcomes.62 The acute phase protein C – Reactive protein is an excellent example of such a marker of inflammation with an unmistakably strong association with the risk of cardiovascular disease. Findings from retrospective and prospective cohort studies all confirmed an increase in levels of circulating C – Reactive protein as risk factor for cardiovascular disease.63,64 Those with the genetic polymorphism known to lead to increased levels of circulating C – Reactive protein do not however have an increased risk of cardiovascular disease.65,66 Ergo, C – reactive protein is a powerful marker for disease, but lacks a causal role, making it not a suitable target for treatment. As discussed above, interleukin – 6 plays an important role in the increased release of C – reactive protein. Population based studies on nucleotide polymorphisms leading to anomalous interleukin – 6 levels show that an increased expression of interleukin – 6 due to genetic polymorphisms is directly associated with the lifetime risk of cardiovascular disease. This proves the case for a causal role of interleukin – 6 in atherothrombosis.67
24 Chapter 1 THE SHARPLY SPIKED LINKING COMPONENTS: CHOLESTEROL CRYSTALS As we now know, there certainly is a great pool of evidence for the causal and essential role of these cytokines in many inflammatory diseases—and as such they are a conditio sine qua non for atherothrombosis. The next crucial mechanism is that of plaque destabilisation. The relation of cholesterol, inflammation and plaque destabilisation is described below. Plaque destabilisation Three different mechanism can compromise the integrity of the cap covering the lipid core of the atheroma: erosion, rupture, and haemorrhage. Data from large histopathological series indicate a contemporary change in ratio in favour of the former.68,69 These phenomena and the subsequent thrombotic sequelae may cause partial or complete luminal obstruction leading to distal ischaemia. This will contribute to tissue injury, clinical symptomatology and eventually tissue damage with functional impairment of the corresponding organ. The haemostasis, aimed to be of a protective effect of the local vascular damage, thus comes at high costs when it compromises luminal patency and distal tissue viability. The cap of the fibroatheroma is formed after relocalisation of smooth vascular muscle cells and consists of collagen structures. (Figure 1) Classically, ruptured caps are often thin (the “thin-capped fibroatheroma”) and show high levels of matrix metalloproteinases, that have collagenase properties.70,71 An important role in the loss of cap integrity has thus been assigned to the presence of these enzymes.72 The matrix- metalloproteinases are unique in their interstitial collagenase activity and are produced in plaques by macrophages. In vulnerable or ruptured plaques predominantly matrix metalloproteinases 1, 8, and 18 are found. The expression of these matrix metalloproteinases by macrophages is regulated by, among others, T‑cells and increased in regions with low shear stress.73,74 The degraded, thin cap will further disintegrate in these circumstances, ultimately initiating the unfortunate thrombotic cascade.
25 General introduction and thesis outline First role of cholesterol crystals The components that lead to the activation of interleukin – 1 and interleukin – 6 are of equal importance in the initiation and proliferation of the inflammatory response, which can promote atherogenesis and degradation of the atherosclerotic cap. Pro– interleukin – 1β is cleaved by caspase 1 to become the active form of interleukin – 1β. Caspase 1 activity depends on a caspase 1–activating complex. Such complexes are inflammasomes, which are comprised of (a) caspase 1 or caspase 5, (b) an apoptosis-associated speck-like protein containing a C-terminal caspase-recruitment domain adaptor protein (ASC) and (c) members of cytoplasmic nucleotide-binding oligomerisation domain (NOD)-like receptors. (Figure 3) The combination of these proteins and receptors is abbreviated as “NALP” or “NLRP” inflammasome. Each inflammasome is activated by different stimuli. The NLRP3 inflammasome is the best characterised inflammasome.75,76 The NLRP3 inflammasome can be activated by physiochemical stressors, microbiologic antigens, cell debris, and danger-associated molecular patterns of any kind.77–80 Crystalline structures are one of the most striking activators of the NALP3containing inflammasome (i.e. the NLRP3 inflammasome). Crystals are solid particles formed from ions or molecules that aggregate and are arranged in a regular, ordered structure. If these crystals tend to stick together, they can form calculi or stones. Common clinical examples include gallstones (cholelithiasis), kidney stones or ureter stones (nephrolithiasis or uretolithiasis, respectively), and salivary stones (sialothiasis).81 A typical example of crystallised structures causing an inflammatory response is seen in gout, where hyperuricaemia leads to the precipitation of uric acid crystals, most prominently in the distal extremities, due to their relatively low temperatures. The inflammatory response following crystal deposition is marked by the release of activated interleukin – 1 and is dependent on the NLRP3 inflammasome.82 The crystalline structure of silica crystals, aluminium salts, and asbestos all initiate similar NLRP3 inflammasome–mediated interleukin – 1 release.83,84
26 Chapter 1 Figure 3. NLR family pyrin domain containing 3 (NLRP3) inflammasome. NLR family pyrin domain containing 3 (NLRP3) was previously known as NACHT, LRR and PYD domains-containing protein 3 (NALP3). It is the major component of the NLRP3 inflammasome. NLRP3 consists of three components: NACHT, LRR and PYD. NACHT, LRR, and PYD are abbreviations for the following: (a) NACHT – which on itself is composed of the abbreviations of NAIP (neuronal apoptosis inhibitor protein), C2TA (class 2 transcription activator, of the major histabilility compex), HET-E (heterokaryon incompatibility) and TP1 (telomerase-associated protein 1); (b) LRR – "leucine-rich repeat" which is synonymous with NLR (nucleotide-binding domain, leucinerich repeat); (c) PYD – "PYRIN domain," after the pyrin proteins. Adapted from Silvis et al. 80 Crystalline threats in atherothrombosis exist in the form of crystallised cholesterol. The solubility and transition from liquid to a solid cholesterol crystal state is determined by its concentration and physical factors (temperature, acidity, blood pressure).85 Differences in these factor may lead to the formation of microcrystals with needle or dendrite morphologies.86 Their presence in atherosclerosis was initially considered to be a phenomenon in advanced lesions, but new conservation techniques made clear they occur as early as two weeks after the initiation of an atherogenic diet in a mouse model.87,88 The crystals are found both in the subendothelium and in the necrotic core, mostly in the extracellular matrix.89 The cholesterol crystals in the arterial vessel wall evoke NLRP3 inflammasome–
27 General introduction and thesis outline mediated interleukin – 1 release by macrophages identical to other crystal-induced inflammatory responses.90 The direct relationship with the development of atherosclerosis has been shown using NLRP3, ASC, or interleukin – 1 knockout mice models. These mice have lower inflammatory responses after the development of cholesterol crystals and smaller aortic atherosclerotic lesions.87 Summarising, the crystallisation of cholesterol plays a pivotal role in the connection between hypercholesterolaemia and the inflammatory response in the developing atherosclerotic lesion. The interplay between cholesterol, inflammation and plaque progression is mediated by a unifying crystal-specific inflammasome pathway. The second role of cholesterol crystals The second local threat to the thinned cap is the puncturing effects of cholesterol crystals. As mentioned above, these effects have sometimes been overlooked due to the methods of conservation and preparation prior to the pathophysiological examination of plaques, since the cholesterol dissipates in the lipid-dissolving ethanol.91 Alternative methods of conservation have provided us with new, detailed information on the atherothrombotic complications associated with, preceded by—and possibly initiated by—cholesterol crystals. The cholesterol crystallisation causes an increase in volume of the plaque. Results of in vitro studies show that the thin cholesterol crystals perforate biological membranes.92 Indeed, ex vivo case-control studies using scanning electron microscope techniques and vacuum dehydration rather than ethanol dehydration showed that crystals perforated the plaque and intima in patients with acute coronary syndromes but not in those without acute coronary syndromes.91 The amount of cholesterol crystals is directly related to the serum concentration of cholesterol.93 In vivo studies in patients with acute coronary syndrome using optical coherence tomography show a higher incidences of thrombosis and plaque rupture in lesions with cholesterol crystals, and their presence is more often seen in those with STelevated myocardial infarction.94,95 With the same technique, cholesterol crystals were observed in a more superficial position in the plaque of patients with acute coronary syndromes compared to the position in those with stable coronary artery disease.96 Connecting crystals, cytokines, and crevices in the cap Recapitulating, the cholesterol crystal is responsible for two separate mechanisms in the atherothrombotic plaque. Firstly, it is an essential part of the extensive
28 Chapter 1 immunological response in the development of the atherothrombotic plaque—by initiating the cytokine effects of interleukin – 1β, mediated by the crystal-induced activation of the NLRP3 inflammasome. Secondly, the biomechanical effect of their crystallised form on the—possibly already weakened—cap of the atheroma may lead to perforation of the endothelium of the cap, leading to erosion or rupture, inevitably disrupting vascular homeostasis, causing distal thromboembolic complications. THE CLINICAL PERSPECTIVE: TRANSLATING INFLAMMATION FROM CYTOKINES AND CRYSTALS TO CLINICAL CARE The magnitude of inflammation as residual risk To translate the clinical consequences of the inflammatory routes in atherosclerosis, contemporary studies focusing on lowering LDL cholesterol levels are used. Findings from these studies provided insight into the magnitude of the risk that remains after optimal treatment of dyslipidaemia. For example, in the Further Cardiovascular Outcomes Research with PCSK9 Inhibition in Subjects with Elevated Risk (FOURIER) trial, intensive lipid lowering was attained with statins and the proprotein convertase subtilisin/kexin type 9 inhibiting molecular antibody evolocumab. Patients achieved a median LDL cholesterol level of 0.78 mmoles/liter. Even with this state-of-the-art lipid-lowering strategy, 9.8% of patients developed a major adverse cardiovascular event after a median follow-up period of just over 2 years.97 The most important independent contributors to this risk are hypertriglyceridemia, residual thrombotic risk, diabetes-associated morbidity, and low-grade inflammation.98 As mentioned above, high-sensitivity C – Reactive Protein (hsCRP) is a robust prognostic risk marker, albeit not causally associated with the disease. Findings from patient-level meta-analyses demonstrated that each standard deviation increment in log-normalised hsCRP is associated with a 37% increase in the relative risk of coronary heart disease. This risk is similar to the risk associated with an increase in systolic blood pressure (35% relative risk increase per every standard deviation increment) and twice as high as the risk associated with an increase in total cholesterol (16% relative risk increase per every standard deviation increment).63
29 General introduction and thesis outline Data on the pleiotropic effects of statins can be used to estimate the proportion of the residual inflammatory risk. In the Pravastatin or Atorvastatin Evaluation and Infection Therapy–Thrombolysis in Myocardial Infarction 22 (PROVE IT–TIMI 22) study and in the Improved Reduction of Outcomes: Vytorin Efficacy International Trial (IMPROVE IT), approximately one-third of patients had an hsCRP level equal or above 2 mg/liter after achieving an LDL cholesterol level below 1.8 mmoles/liter. Depending on treatment intensity, statin treatment could reduce median hsCRP by a third, with a greater reduction in hsCRP level associated with a greater relative risk reduction of cardiovascular events.99,100 In addition, the risk reduction following lowering hsCRP in these patients occurred irrespective of the change in LDL cholesterol, emphasising independence of the inflammatory pathway.101 The abovementioned findings further shaped the clinical role of the “inflammation hypothesis” in atherosclerotic disease. Subsequently, multiple broad-acting and targeted agents were introduced to investigate whether anti-inflammatory treatment would truly translate into beneficial effects for patients with atherosclerosis. A frustrating start looking for therapeutic interventions with antiinflammatory agents Nonsteroidal anti-inflammatory drugs (NSAIDs) are widely used antiinflammatory drugs. Their anti-inflammatory properties are not suitable for addressing inflammation in atherosclerosis, since they are consistently associated with an increased risk of major coronary events, with the exception of low-dose acetylsalicylic acid.102 This is a consequence of the dose-dependent pharmacodynamic properties of acetylsalicylic acid. Acetylsalicylic acid inhibits cyclooxygenase 1 (COX-1) and COX-2. COX-1 inhibition reduces thromboxane A2–induced platelet aggregation at low doses, accounting for its atherothrombotic protective effects. At higher doses, the anti-inflammatory properties driven by COX-2 inhibition arise.103 Many novel NSAIDs are designed to selectively inhibit only COX-2. COX-2 inhibition also leads to undesired cardiovascular effects, such as prostaglandin E2–mediated sodium and water retention, vascular endothelium prostacyclin–mediated platelet activation, and vasoconstriction, increasing the risk of major cardiovascular events.104 Glucocorticoids play an integral role in the management of many inflammatory conditions. However, many unfavourable side effects, such as hypertension, impaired glucose tolerance, and obesity following long-term treatment with corticosteroids render them unsuitable to dampen the inflammatory component of atherosclerosis.105
30 Chapter 1 One of the first inflammatory targets in coronary disease risk was the inhibition of lipoprotein-associated phospholipase A2 (Lp-PLA2). Lp-PLA2 is bound to LDL cholesterol and plays a role in oxidative modification within the vascular wall, and it increases vascular inflammation in atherosclerosis. The Lp-PLA2 inhibitor darapladip was tested in two major trials in almost 20,000 patients with chronic coronary disease and recent acute coronary syndrome. Both trials did not demonstrate a reduction in the risk of major cardiovascular events.106,107 A variant of the compound inhibiting the secretory form of PLA2, varespladib, was associated with a higher rate of recurrent myocardial infarction.108 Parallel to these efforts, inhibition of the broad p38 mitogen-activated protein (MAP) kinases system was proposed. The p38 MAP kinase system takes part in various intracellular signaling routes and is active in endothelial cells, smooth muscle cells and leucocytes. The p38 MAP kinase system inhibitor losmapimod however did not show any clinical effect.109 Methotrexate is a broad-acting immunomodulating drug that inhibits DNA synthesis by competing with folate synthesis and by inhibiting T-cell adhesion molecules and T-cell activity. The drug is used in a wide array of autoimmune and oncological conditions. The ability of methotrexate to dampen atherosclerotic inflammation was studied in over 4,500 patients with chronic coronary disease and type 2 diabetes or metabolic syndrome. The trial was ended prematurely for reasons of futility, as accruing more data would unlikely be able to demonstrate any signal of clinical benefit.110 One of the common denominators in these trials was the absence of any evident biochemical response to the treatment. These data strengthened the hypothesis that effective inhibition of the inflammatory pathways in atherosclerosis should comprise targeted and detectable interleukin – 1 and interleukin – 6 inhibition, confirming the earlier experimental findings. Promising therapeutic developments The Canakinumab Antiinflammatory Thrombosis Outcome Study (CANTOS) was the first to prove that modulating the inflammatory pathway in atherosclerosis reduces major adverse cardiovascular events in patients with a recent history of myocardial infarction and an hsCRP level equal or above 2 mg/liter. The CANTOS trial randomised 10,061 patients to receive three subcutaneous doses of canakinumab, a selective interleukin – 1β inhibitor, or placebo. Canakinumab (150 mg) reduced the risk for the composite end point of nonfatal myocardial infarction, nonfatal
31 General introduction and thesis outline stroke, or cardiovascular death by 15% and reduced the risk of the composite end point of nonfatal myocardial infarction, nonfatal stroke, cardiovascular death or urgent hospitalisation for angina leading to urgent revascularisation by 17%.111 Although these results were ground breaking, canakinumab is a costly drug, and the findings from the trial have not led to registration of the drug for secondary prevention in cardiovascular disease. Colchicine is a well-known anti-inflammatory drug used to treat one of the most distinct types of crystalloid diseases: gout. It has a strong inhibitory effect on the NLPR3 inflammasome and reduces expression of interleukin – 1 and interleukin – 6.112,113 However, the drug has a plethora of cell interactions, which all could contribute to an atheroprotective effect.114,115 Neutrophil mobilisation and recruitment are diminished as colchicine inhibits endothelial adhesiveness.116 Furthermore, leukocyte-platelet aggregation seems to be supressed during colchicine use.117,118 Thus, the strong effects of colchicine on the crystal-induced inflammatory pathway is most likely complemented by additional inflammation modulatory effects. THE METHODOLOGICAL PERSPECTIVE: HOW TO FACILITATE DRUG REPURPOSING AND IMPROVE PRAGMATIC TRIAL CONDUCT Although colchicine may have potential to modulate the broad inflammatory pathway driving atherosclerosis and its thrombotic complications, it can be seen as an unusual drug from the perspective of the conventional routes of clinical research. To understand this, the structure of conventional drug research and development needs an explanation. Conventional routes of drug research The majority of clinical research is conducted in a structure where a pharmaceutical company sponsors a commercial or academic research organisation. These institutions execute the clinical experiment to the highest standards of scientific integrity in an ideally completely autonomous manner—a concept guarded by, among others, blind treatment allocation, data monitoring boards, and independence of the academics to the sponsor when conducting pre-specified analyses. The conditions and requirements for such research that – in case of positive results – could lead to clinical implementation, are set by legislation and national
32 Chapter 1 or international drug authorities. These requirements are strongly focused on reducing the risk of false-positive results and are driven by the principal goal of ensuring the safety of participants. This makes these phases of research costly. The estimated costs for the complete route of drug research until market authorisation can reach from 800 million up to 1.5 billion US dollars in the current era.119 (Figure 4) N NME cost (US$ billions) 2.0 1.5 1.0 0 0.5 2010 1990 1970 1950 a y = 1E–113e0.1335x R2 = 0.98 Log (NME cost) 8.0 6.0 4.0 0 2.0 1970 1950 b y = 0.1335x–260 R2 = 0.98 Nature Reviews | Drug Discovery 2010 1990 Log (NME cost) 8.0 6.0 4.0 0 2.0 2010 1990 1970 1950 b y = 0.1335x–260 R2 = 0.98 Figure 4. The exponential costs of new drugs. Panel A shows a plot of twelve independent estimates of the cost of a new molecular entity (NME) spanning 48 years, reaching 1.5 billion US dollars. Panel B shows the same data plotted on a logarithmic scale. The exponent in the line equation in Panel A and the gradient of the line in Panel B show that the cost per NME has grown at an annual compound rate of 13.35% since the late 1950s. Adapted from Munos.119 The cost of research includes the development of the compound, production, and exploratory preclinical research (phase I and II clinical research) with limited participants. Large clinical outcome trials (phase III clinical research) are more costly endeavours. Median costs of conducting such trials are estimated at 21 million US dollars per drug, but can reach over 75 million US dollars.120 (Figure 5) When the incidence of the primary outcome measure in the final efficacy evaluation
33 General introduction and thesis outline is low, the prerequisite sample size or follow-up time further increases operational costs. Thus, the provision of return on investment for the sponsor is crucial for embarking large clinical outcome trials. Nature Reviews | Drug Discovery Cost (US$ millions) 75 50 25 0 Phase IP Phase II Phase III Median Mean 3.4 8.6 21.4 Nature Reviews | Drug Discovery Outsourcing Grants/ contracts Personnel Other expenses 22% 21% 21% 37% Figure 5. Trial costs and components. The upper panel shows the distribution of the costs of 726 studies conducted in patients from 2010–2015. Medians and means are indicated with diamonds and lines. Boxes indicate the 25th and 75th percentiles, and whiskers indicate the 10th and 90th percentiles. Phase IP = phase I study involving patients. The lower panel shows the proportion of various components that make up expenses in phase III clinical trials. Data are derived from 273 studies conducted in patients from 2010–2015. Adapted from Martin et al.120
34 Chapter 1 Conventionally, this is guaranteed for the sponsor by the ability to register as the market authorisation holder and by patents that protect intellectual property. Patents have a limited time span, to allow for widespread production after a predefined period of time. The last step in drug development should be, when patent protection ends, the generic pharmaceutical company begins large-scale production. This should decrease costs and, finally, the burden to the health care system. Drug repurposing From a clinical and biological perspective, it is highly likely that compounds can serve multiple therapeutic purposes. In fact, medical history has taught us that some chemical compounds or drugs are even more effective in their second than their first life.121 (Table 1) Often such discoveries are impressive cases of serendipity: an eponym derived from the Persian tale “The Three Princes of Serendip” (Serindip, being an old name for Sri Lanka), where the main characters stumble upon all sorts of unplanned fortunate discoveries.122,123 An example par excellence of serendipity in medicine is our introduction to antibiotics. Alexander Fleming failed to disinfect the cultures of bacteria he was investigating. When he returned from his vacation he realised that the cultures that were contaminated with Penicillium moulds all had perished; by accident he discovered the Gram-positive bacteria growth inhibiting properties of the fungus, introding Penicillin as the worlds first antibiotic.121 Another example is the observation of the strong leukopenic effects of mustard gas, used as chemical warfare agent in World War I and II. This particular characteristic of nitrogen mustard led to the invention of the first chemotherapeutic drug. The cytotoxic effects of nitrogen mustard were used to its advantage in the treatment of excesses of leukocytes such as seen in lymphoma and leukemia.124 This also gave rise to the first alkylating agents (mechloretamine) and later to studies on interacting with folic acid, leading to the development of methotrexate. Lastly, highly appealing to the imagination as an example of serendipity, are the anecdotal records on trials of sildenafil, which was originally designed as an anti-anginal drug. Trial participants reportedly did not return study medication due to unexpected "desirable side effects” of the drug. It now has widespread use in erectile dysfunction, and is wellknown by its brand name Viagra.123
www.ridderprint.nlRkJQdWJsaXNoZXIy MTk4NDMw