Endothelial dysfunction in RA, a rate limiting factor in atherogenesis

The vascular endothelium plays a central role in blood vessel homeostasis. Under normal physiological circumstances, the endothelium has a quiescent phenotype that modulates permeability, coagulation and vasodilation. However, when the endothelium is exposed to inflammatory mediators derived from a local infection, the endothelium becomes activated to a state that expresses chemokines and adhesive receptors that together facilitate the recruitment of pro-inflammatory leucocytes. After clearance of the infection, the leucocytes leave and the endothelium readopts its quiescent state.

Irrespective of its cause, atherosclerosis is initiated by a chronic systemic inflammatory activation of the endothelium that leads to a state of endothelial cell dysfunction.6 As a result of damage to the endothelial barrier, lipoproteins and leucocytes accumulate in the subendothelial space and lead to the formation of the early fatty streaks that can subsequently progress to more advanced atherosclerotic plaques. At first sight, endothelial dysfunction as a common denominator in diabetes, hypertension, obesity, renal failure, hypercholesterolaemia and also RA is remarkable, as the pathophysiological basis of these diseases is quite dissimilar. Currently, a likely explanation of this mechanistic convergence centres on the regulation of a single endothelial cell-specific enzyme, endothelial nitric oxide synthase (eNOS). In the presence of oxygen and sufficient amounts of the cofactors calmodulin and tetrahydrobiopterin, eNOS converts l-arginine into l-citrulline, resulting in the generation of a single molecule of nitric oxide (NO). NO is a reactive molecule that can react with a wide range of biological molecules and can rapidly access different compartments of endothelial cells. Therefore, NO is well suited for a role in coordinating cellular processes via post-translational modification of regulatory proteins.7 In fact, it has been shown that in response to normal shear stress, NO is constitutively produced to maintain the quiescent, non-inflammatory endothelial phenotype. This role for eNOS could be readily demonstrated in rat models of chronic eNOS inhibition in which a marked influx of mononuclear cells in the intimal and medial regions of intrarenal arteries was observed.8 The molecular mechanism underlying NO-dependent endothelial quiescence partly depends on the capacity of NO to react with sulphur groups on cysteine residues in proteins with an acid–cysteine-based motif. Through the process of S-nitrosylation, NO can reduce the activity of pro-inflammatory transcription factors such as nuclear factor kappa B (NF-κB), activator protein 1 and DNA methyltransferases, enabling the endothelium to revert to a quiescent state.9

Most atherosclerotic risk factors accelerate disease progression by augmenting the production of reactive oxygen species (ROS) to the extent that it exceeds the neutralising capacity of cellular enzymatic and non-enzymatic antioxidants. In the endothelium, enzymatic systems that are capable of producing ROS such as xanthine oxidase and nicotinamide adenine dinucleotide phosphate, reduced form oxidase are activated by pro-inflammatory cytokines.10 As a consequence of this ROS overproduction, the cells enter a state of “oxidative stress” in which protein function in the cell is chemically modified. In endothelial cells this can lead to an inactivation of the inhibitor of NF-κB isoform alpha and, consequently to NF-κB activation.11 eNOS is thought to play a central role in this inflammatory activation by regulating redox signalling.12 Chronic redox signalling could lead to the oxidation of the essential eNOS cofactor, tetrahydrobiopterin,11 while concomitantly decreasing the regeneration of tetrahydrobiopterin by dihydrofolate reductase.13 As a consequence, the oxidation–substrate coupling of eNOS becomes disturbed, converting the enzyme from a NO into a superoxide-producing protein, further amplifying pro-inflammatory signalling. This so-called “uncoupled state” of eNOS is evolutionarily conserved and may be part of an integrated and controlled inflammatory signalling pathway that evolved for host defence.12 However, this powerful pro-inflammatory switch may, at the same time, act as the Achilles heel of the endothelium. Under conditions of chronic activation, eNOS uncoupling is likely to play a central role in the development of atherosclerosis. Support for this notion can be derived from the fact that repeated tumour necrosis factor alpha (TNFα) injections in rats lead to endothelial cell dysfunction in a nicotinamide adenine dinucleotide phosphate, reduced form oxidase-dependent pathway manner.14 In RA, the chronic release of pro-inflammatory cytokines such as IL-1α, IL-1β, IL-6 and TNFα from the synovial tissue may directly relate to systemic endothelial dysfunction.15 Indeed, 12 weeks of anti-TNFα treatment in patients with RA markedly improved endothelial-dependent vasodilation of the brachial artery.16

Endothelial integrity: a balance between injury and repair

Although the endothelium is equipped with potent anti-oxidant and anti-inflammatory systems, prolonged and repeated exposure of the endothelium to oxidative stress ultimately exhausts these protective mechanisms. As a result of this, the endothelium not only becomes dysfunctional, but this process can also lead endothelial cells to progress towards apoptosis and detachment.17 Also in patients with RA, elevated levels of mature endothelial cells have been measured in the peripheral blood.18 In the classic view, mature endothelial cells adjacent to the region of denudation proliferate to repair the damaged endothelium. However, the proliferative capacity of endothelial cells is limited either by the shortening of telomeres that is associated with chromosomal replication or by exposure to environmental factors that lead to premature, stress-induced cellular senescence. Using a computational model, it was estimated that under conditions of increased cellular turnover, as a consequence of oxidative stress, critical telomere shortening would lead to systemic replicative senescence of the endothelium in an individual by the age of 65 years.19 Moreover, telomere erosion itself is accelerated by chronic exposure of endothelial cells to oxidative stress.20 Assuming that the observations with cultured endothelial cells are representative of vascular endothelial cells, increased endothelial cell turnover and premature replicative senescence in patients at risk of cardiovascular disease21 make it highly unlikely that replicating, adjacent mature endothelial cells possess sufficient reparative capacity to repair the chronically damaged endothelium throughout the lifespan of an individual.

However, an alternative cellular source to facilitate the rejuvenation of dysfunctional endothelium was identified in the form of circulating endothelial progenitor cells (EPC).22 These cells express the early haematopoietic stem cell marker, CD34,23 and can differentiate into cells with endothelial cell characteristics in vitro. Platelets may play a central role in the homing and differentiation of EPC at sites of vascular injury in vivo.24 EPC have been shown both in animal models and humans to contribute to neovascularisation and re-endothelialisation, and evidence is accumulating for an essential role of these progenitor cells in endothelial maintenance and repair.25

The mobilisation of stem cells to the circulation is normally promoted by the release of angiogenic factors and haematopoietic cytokines in response to hypoxia and tissue injury.26 Vascular endothelial growth factor (VEGF) is one of the most potent mobilising angiogenic factors. VEGF interacts with its receptors, VEGFR2 and VEGFR1, which are expressed on endothelial and haematopoietic stem cells in the bone marrow. This interaction elicits a chemotactic response that facilitates the entry of stem cells from a quiescent niche into a permissive niche in the bone marrow, where they proliferate, differentiate and are mobilised into the circulation.27 This process is dependent on the secretion of matrix metalloproteinase 9 by haematopoietic and stromal cells in the bone marrow.28 Interestingly, mice lacking eNOS fail to upregulate matrix metalloproteinase 9 and are incapable of EPC mobilisation.29 The implication of this finding is that the recruitment of EPC may thus be impaired in patients with impaired NO bioavailability. Indeed, several studies have provided data for a reduced number and perturbed functioning of circulating EPC in patients with cardiovascular risk factors, such as atherosclerotic disease,30 type 2 diabetes31 and coronary artery disease.32 33

Consistent with other patient groups with cardiovascular risk, Grisar et al reported that the number of circulating EPC in patients with RA were reduced compared with healthy individuals and correlated inversely with disease activity.34 In a follow-up study in RA patients on glucocorticoid therapy, evidence was provided for a role of TNFα in EPC depletion.35 Another study demonstrated that low-grade inflammation in patients with RA is associated with a reduced number of EPC and is directly correlated with endothelial dysfunction.36 Next to elevated levels of TNFα, endothelial dysfunction can also be driven by exposure of endothelial cells to elevated plasma levels of the endogenous eNOS inhibitor asymmetric dimethyl-l-arginine.37 Interestingly, elevated levels of asymmetric dimethyl-l-arginine in patients with RA who were free of either cardiovascular disease or risk factors were shown to be associated with reduced numbers of EPC and increased carotid atherosclerosis.38 Those studies support a direct link between endothelial dysfunction, impairment of NO-dependent EPC mobilisation and/or survival and the promotion of atherogenesis in RA.

Myeloid progenitor cell dysfunction in RA suggests impaired collateral formation

Vascular maintenance and repair not only requires re-endothelialisation and angiogenesis but also arteriogenesis. In this process, pre-existing small-calibre arterioles remodel in response to elevated shear stress. CD14+ monocytes have been well established to play a critical role in arteriogenesis.39 In contrast to the differentiation to a pro-inflammatory macrophage, once recruited to the remodelling arteriole and in response to environmental cues, CD14+ monocytes are also capable of adopting a pro-arteriogenic phenotype. This enables the cells to produce proteases that dissociate matrix structures and secrete growth factors that amplify the cellular components of the vessel. Unlike their pro-inflammatory counterparts, these regenerative monocytes secrete angiogenic growth factors that most likely support the proliferation of endothelial cells.40

Interestingly, CD14+ monocytes are phenotypically highly similar to myeloid or early-outgrowth EPC (mEPC) that can be obtained when peripheral blood mononuclear cells are cultured for several days in VEGF-rich media. Upon transplantation, mEPC augment reperfusion and collateral formation in models of tissue ischaemia.41 Moreover, early trials in patients with acute myocardial infarction suggest that these cell preparations reduce infarct size and improve cardiac contractile function.42 Many laboratories, including our own, have demonstrated that most risk factors for cardiovascular disease are strongly associated with mEPC dysfunction.43 For example, the number of mEPC that can be obtained from the peripheral blood of patients with type 1 diabetes is reduced by approximately 40%.44 Furthermore, the functionality of these mEPC is impaired, as evidenced by a perturbed adhesion to endothelial cells, reduced incorporation into endothelial tubular structures, decreased release of pro-angiogenic factors in concert with an increased pro-inflammatory cytokine production. The molecular mechanisms underlying the impairment of mEPC function are currently under active investigation. Although they may differ, depending on the present metabolic and haemodynamic risk factors, the reduced bioavailability of NO45 and elevated oxidative stress again appears to be a common theme.46

Recently, in studies aimed at elucidating the nature of mEPC dysfunction in type 1 diabetes, we obtained evidence for a novel mechanism that appears to be associated with mEPC dysfunction. When comparing messenger RNA profiles of mEPC derived from healthy controls versus type 1 diabetes patients using microarray technology, we observed that diabetic mRNA profiles displayed an altered, pro-atherogenic phenotype. This suggested that mEPC exposed to an inflammatory diabetic environment might be “skewed” towards a pro-inflammatory, myeloid differentiation pathway.47 Consequently, in chronically elevated oxidative stress, the “healing” mEPC can turn into a pro-inflammatory, antigen-presenting cell that may contribute to neointma formation and atherogenesis. Several recent studies have reported that reduced mEPC yields34 36 48 and function36 have also been observed in patients with RA. Importantly, similar to the improved eNOS-mediated NO production and reduction of endothelial dysfunction, anti-TNFα therapy increased the number of viable mEPC that could be cultured from the blood of patients with RA and improved the migratory function of the cells.49 These observations strongly suggest that mediators of chronic inflammation in RA may affect the phenotype of circulating CD14+ monocytes. Indeed, in patients with RA, significantly more circulating CD14+ monocytes have been shown to co-express markers that are associated with more mature inflammatory phenotypes, such as Toll-like receptor 2 and CD16.50 51 As such, elevated numbers of circulating CD14+/CD16+ cells have been reported to be associated with increased plasma concentrations of cytokines such as TNFα and coronary atherosclerosis.52 The resultant pro-inflammatory skew of these circulating monocytes could potentially serve as a double-edged sword, as it may not only be atherogenic, but may also impair the capacity of monocytes to adopt a pro-angiogenic (mEPC) phenotype in remodelling vessels.

Conclusion

Next to chronic endothelial cell activation, premature atherosclerosis in patients with RA may also involve the inability of circulating cells to repair and regenerate damaged portions of the vasculature. The long-term reduction of the number of immature CD34+ vascular progenitor cells may hamper the rejuvenation of the endothelium by new endothelial (progenitor) cells. The observation that potent EPC mobilisers such as erythropoietin, basic fibroblast growth factor and VEGF are elevated in patients with RA but do not translate into elevated levels of EPC34 further supports the concept that bone marrow endothelial cell dysfunction may be an important causal factor in the depletion of circulating EPC. Nevertheless, it was also demonstrated that EPC traffic to the RA synovium53 most likely to support the processes that lead to increased vascularity of the inflamed tissue. As a consequence, EPC recruitment to the synovium could also contribute to the loss of EPC numbers in the circulation, thereby reducing the potential to support long-term systemic endothelial integrity.54

Finally, chronic inflammation in RA may also affect collateral formation and remodelling by changing the subset composition of circulating CD14+ myeloid cells. Although little is known regarding the impact of RA on collateral formation, large-artery remodelling in patients with long-standing RA has recently been reported55 and may well involve a role for circulating myeloid cells.56 The favourable outcome of anti-TNFα therapy in RA may not only be explained by the correction of endothelial cell dysfunction but also by the restoration of the vascular regenerative potential of RA patients.