Endothelial Cell Dysfunction and Increased Cardiovascular Risk in Patients With Chronic Kidney Disease

Abstract

The endothelium is considered to be the gatekeeper of the vessel wall, maintaining and regulating vascular integrity. In patients with chronic kidney disease, protective endothelial cell functions are impaired due to the proinflammatory, prothrombotic and uremic environment caused by the decline in kidney function, adding to the increase in cardiovascular complications in this vulnerable patient population. In this review, we discuss endothelial cell functioning in healthy conditions and the contribution of endothelial cell dysfunction to cardiovascular disease. Further, we summarize the phenotypic changes of the endothelium in chronic kidney disease patients and the relation of endothelial cell dysfunction to cardiovascular risk in chronic kidney disease. We also review the mechanisms that underlie endothelial changes in chronic kidney disease and consider potential pharmacological interventions that can ameliorate endothelial health.

Increased Cardiovascular Risk in Chronic Kidney Disease

Chronic kidney disease (CKD) is defined by kidney damage or a reduced kidney filtration function (glomerular filtration rate <60 mL/minute per 1.73 m²) for a period beyond 3 months that affects health.¹ With a global prevalence of ≈13.4%,² CKD imposes a serious burden on our socioeconomic and health care system. CKD is a progressive disease, classified into stages 1 to 5 based on the reduction in kidney function (Figure 1A).¹ In end-stage kidney disease (CKD stage 5: glomerular filtration rate <15 mL/minute per 1.73 m²), patients require kidney support therapy as dialysis or kidney transplantation to replace the failing kidney function.

Of note, patients with moderate to advanced CKD are at increased cardiovascular risk compared with the general population and patients with mild CKD, with lower estimated glomerular filtration rate and higher albuminuria identified as risk factors for all-cause and cardiovascular mortality, independent of traditional cardiovascular risk factors.^5,6 Overall, 33.3% to 37.1% of CKD3(a/b) patients and 39.9% of CKD4 patients die from cardiovascular diseases (CVD) compared with 21.7% to 26% of the general population (Figure 1B).³ For patients with end-stage kidney disease aged 25 to 34, annual mortality is even increased 500 to 1000 times compared with similarly aged controls with healthy kidney function and is comparable to 85-year olds in the general population, underscoring the high cardiovascular burden in CKD.⁶

Figure 1.

Chronic kidney disease (CKD) prevalence and cardiovascular risk in CKD. A, CKD classification and prevalence. *Compared with young adult level. ^a/bFigure based on data by ^aInker et al¹ and ^bHill et al.² B, Percentage of cardiovascular deaths according to CKD stage (age and sex adjusted). Figure based on data by Thompson et al.³ C, Cardiovascular causes of death in the general population compared with dialysis patients (CKD stage 5D). Figure based on data by Roberts et al.⁴ CVD indicates cardiovascular disease; GFR, glomerular filtration rate; and PU, proteinuria.

In the general population, myocardial infarction and cerebrovascular events are the most important cardiovascular causes of death, underlying ≈75% of all CVD-related deaths (Figure 1C).⁴ Also, in CKD5 patients on dialysis (CKD5D), these CVD types remain important as they account for 45% (myocardial infarction) and 13% (cerebrovascular events) of deaths by cardiovascular causes, though with a joint responsibility of ≈58% of all CVD-associated deaths, they are reduced in relative importance compared with the general population. Instead, CKD5D patients show a relative increase in sudden cardiac death and heart failure, being responsible for 28% and 9% of all CVD-associated deaths compared with 2% and 7% as observed in the general population (Figure 1C).⁴ This reveals a high increase especially in the risk of sudden cardiac death in advanced CKD.⁴

Atherosclerosis and Myocardial Infarction

The main underlying cause of both myocardial infarction and stroke is atherosclerosis, a lipid-driven inflammatory disease of medium to large-sized arteries triggering the development of atherosclerotic lesions.⁷ These lesions gradually grow over time and can ultimately restrict blood flow or trigger thrombosis through plaque rupture or erosion.^7,8 Patients with CKD3-5D show an increased prevalence of subclinical atherosclerotic lesions compared with the general population, with a larger increase in more advanced CKD stages after adjustment for sex, age and diabetes.^9,10 Furthermore, compared with patients without CKD progression, patients with CKD progression over 24 months displayed more frequently also a progression of atherosclerotic lesions as detected by ultrasound.^10,11 After acute myocardial infarction, patients with CKD show a reduced survival over time compared with non-CKD patients, with an increased risk of death as well as non-fatal cardiac events with increasing CKD stage.¹²

Sudden Cardiac Death and Uremic Cardiomyopathy

In the general population, coronary heart disease is responsible for 80% of sudden cardiac deaths.¹³ The disproportional increase in sudden cardiac death in patients with advanced CKD suggests differences in its pathophysiology and causes as kidney function declines. Left ventricular hypertrophy is significantly associated with increased risk of sudden cardiac death in the general population¹⁴ and can be caused by cardiac preload (intravascular volume overload), cardiac afterload (pressure overload), or afterload/preload-independent factors.¹⁵ Patients with CKD present more frequently with left ventricular hypertrophy, with a prevalence of up to 40% and even 75% in patients in CKD5D.¹⁶ Together with cardiac fibrosis, left ventricular hypertrophy is one of the hallmarks of uremic cardiomyopathy and may trigger cardiac electrical disturbances and lethal arrhythmias.¹⁵

Endothelial (Cell) Dysfunction as Contributor to Cardiovascular Risk

A main contributor to increased cardiovascular risk is endothelial cell dysfunction, which encompasses a whole array of maladaptive alterations in the endothelial cell functional phenotype associated with increased cardiovascular risk. This term was suggested in an excellent review by Gimbrone et al¹⁷ to provide a distinction from the more narrow term “endothelial dysfunction,” which typically has been used to refer to endothelial abnormalities triggering a reduction in nitric oxide bioavailability and associated vascular relaxation. Significant endothelial heterogeneity exists across the vascular tree, for example, when comparing arteries versus veins, as well as the macrovasculature (including the large elastic as well as muscular conduit arteries) versus the microvasculature (including the capillaries, arterioles, and venules) and both endothelial cell dysfunction at the macro- and microvascular level contributes to increased cardiovascular risk. In this review, we discuss the contribution of endothelial cell dysfunction to CVD with a special focus on patients with CKD. We review findings on molecular mechanisms underlying endothelial cell dysfunction in CKD as well as discuss the impact of pharmacological interventions.

Endothelial Cell Dysfunction as Contributor to Cardiovascular Risk

Endothelial Cell Dysfunction and Atherosclerotic Risk

The endothelial cell layer of the vasculature provides a semipermeable barrier enabling a regulated exchange of fluids, molecules, and cells and plays an important role in maintaining vascular health (Figure 2). Macrovascular endothelial cell dysfunction is an early event in the development of atherosclerotic lesions. On the one hand, it is influenced by hemodynamic factors: in atherosclerosis-resistant areas of the arteries, a laminar blood flow contributes to a protective endothelial cell phenotype. However, atherosclerosis-prone regions of the arterial vasculature are exposed to a disturbed, oscillatory blood flow and associated low time-averaged shear stress, which induce oxidative stress, endothelial phenotypic changes, and cell junction alterations as well as endothelial cell turnover (as discussed in more detail in the review by Gimbrone et al¹⁷). Furthermore, inflammatory triggers as proinflammatory cytokines, oxLDL (oxidized low-density protein) as well as different cardiovascular risk factors as metabolic disturbances and smoking, contribute to endothelial cell dysfunction. Also, excessive stretch on blood vessels can trigger endothelial permeability, inflammatory responses and oxidative stress.¹⁸ Combined, this triggers proinflammatory signaling in endothelial cells with an upregulation of proinflammatory cytokines (eg, IL [interleukin]-1, IL-8), chemokines (eg, C-C motif chemokine ligand 2), and endothelial-leukocyte adhesion molecules (VCAM-1 [vascular cell adhesion molecule 1], ICAM-1 [intercellular adhesion molecule 1], P-selectin), reduces endothelial production of atheroprotective nitric oxide and increases endothelial permeability. As a result, inflammatory leukocytes are recruited, adhere to the inflamed endothelium and infiltrate into the vascular wall, where they together with accumulated lipids contribute to the development and progression of atherosclerotic lesions.⁷ The atheroprotective phenotype of the endothelium is regulated by master transcription factors as KLF (Kruppel-like factor)-2, KLF-4, and NRF (nuclear factor erythroid 2-related factor)-2, whereas NF-κB (nuclear factor-κB) is a key transcription factor driving endothelial inflammation.¹⁷ Furthermore, endothelial cells demonstrate a de-differentiation and increased heterogeneity during atherosclerosis progression, with also signs of endothelial-to-mesenchymal transition. Endothelial-to-mesenchymal transition is characterized by the acquirance of mesenchymal cell functions as ECM (extracellular matrix) production and is mainly driven by the transcription factors Snail, Slug, and Twist1. Its extent has been associated with the severity of atherosclerosis plaques in human arteries,^19,20 and animal studies investigating key regulators of endothelial-to-mesenchymal transition suggested an important role in plaque progression²⁰ and calcification,²¹ as discussed in detail in an excellent review by Souilhol et al.²²

Figure 2.

Endothelial phenotype in healthy conditions in relation to thrombosis, vasoreactivity and inflammation. Vasodilatory, anti-inflammatory, and antithrombotic characteristics of healthy endothelium. For more information, see text. ADP indicates adenosine diphosphate; AMP, adenosine monophosphate; ATP, adenosine triphosphate; ATIII, antithrombin III; cAMP, cyclic adenosine monophosphate; CD39, ectonucleoside triphosphate diphosphohydrolase-1; CD73, ecto-5’-nucleotidase; cGMP, cyclic guanosine monophosphate; eNOS, endothelial nitric oxide synthase; EPCR, endothelial protein C receptor; ER, endoplasmic reticulum; HDL, high-density lipoprotein; LDL, low-density lipoprotein; NO, nitric oxide; NRF-2, nuclear factor erythroid 2-related factor 2; PG, proteoglycans; TFPI, tissue factor pathway inhibitor; and ZO, zonula occludens.

Of note, endothelial cell dysfunction does not only contribute to plaque initiation, progression, and destabilization with subsequent plaque rupture but also to atherosclerotic plaque erosion, which is expected to be responsible for one-third of acute coronary syndromes. Here, chronic low-grade endothelial activation by for example TLR (toll-like receptor)-2 ligands combined with endothelial cell apoptosis and catabolism of basement membrane components can trigger endothelial cell detachment with subsequent thrombus formation on the denuded area.⁸

Endothelial Cell Dysfunction and Thrombotic Risk

A healthy, functional endothelial layer at macro- and microvascular levels is crucial in the regulation of hemostasis and interferes both at the level of primary as well as secondary hemostasis to prevent unwanted platelet activation and coagulation (Figure 2). At the level of primary hemostasis, the endothelium elicits strong platelet inhibition by the continuous secretion of nitric oxide (NO) and prostacyclin causing an increase in intraplatelet cGMP (cyclic guanosine monophosphate) and cAMP (cyclic adenosine monophosphate) levels, respectively.²³ To prevent platelet activation by extracellular ATP and ADP, the endothelial layer expresses CD39 and CD73, ectonucleases that convert ATP and ADP to adenosine, a platelet inhibitor that by increasing platelet cAMP levels elevates the platelet’s activation threshold.²⁴ In addition, the endothelial glycocalyx repels platelets by its negative charge and as such aids in the prevention of platelet adhesion.^25,26 At the level of secondary hemostasis, within the glycocalyx, heparan sulfate proteoglycans bind and promote the activity of ATIII (antithrombin III), a potent inhibitor of multiple coagulation factors including thrombin, FIXa, FXa, FXIa, and FXIIa.^26,27 Furthermore, endothelial cells express TFPI (tissue factor pathway inhibitor), a serine protease that—as its name suggests—interferes with TF (tissue factor)-induced coagulation and thereby limits the activity of the extrinsic pathway.²⁷ Next to TFPI, endothelial cells constitutively express thrombomodulin, a membrane-bound protein that captures thrombin from the circulation and upon binding increases the affinity of thrombin for the anticoagulant protein C. Together with protein S, activated protein C disables FVa and FVIIIa.²⁷

Upon disturbances in lipid metabolism, inflammation, oxidative stress, and pathophysiological shear stress, endothelial cell dysfunction develops, characterized by a diminishment of antithrombotic and anti-inflammatory properties and a degradation of the glycocalyx. Concomitantly, the dysfunctional endothelium takes on proinflammatory and prothrombotic characteristics (Figures 3 and 4).²⁷ As a result, the production of NO and prostacyclin decreases, while the secretion of prothrombotic and proinflammatory molecules like vWF (von Willebrand factor) and C-C motif chemokine ligand 2 increases.²⁷ Furthermore, the expression of thrombomodulin is strongly downregulated upon endothelial cell dysfunction resulting in a downregulation of protein C activation, whilst the expression of TF is upregulated favoring the activation of coagulation.²⁷ Already during the early stages of atherogenesis, neutrophil extracellular traps are implicated in endothelial cell dysfunction and fuel the thromboinflammatory response.²⁸

Figure 3.

Endothelial phenotype in chronic kidney disease (CKD) conditions in relation to thrombosis. Endothelial phenotype in CKD conditions in terms of thrombosis. For more information, see text. cLDL indicates carbamylated low-density lipoprotein; eNOS, endothelial nitric oxide synthase; ER, endoplasmic reticulum; LOX-1, lectin-type oxidized low-density lipoprotein receptor 1; NO, nitric oxide; oxLDL, oxidized low-density lipoprotein; PAI-1, plasminogen activator inhibitor-1; TF, tissue factor; and vWF, von Willebrand factor.

Figure 4.

Endothelial phenotype in chronic kidney disease (CKD) conditions in relation to inflammation and atherosclerosis. Endothelial phenotype in CKD conditions in terms of proinflammatory and proatherosclerotic characteristics. For more information, see text. ATF indicates activating transcription factor; cLDL, carbamylated low-density lipoprotein; CREB, cyclic adenosine monophosphate response element-binding protein; eNOS, endothelial nitric oxide synthase; ER, endoplasmic reticulum; E-selectin, endothelial-leukocyte adhesion molecule 1; HCO₃⁻, bicarbonate; H₂O₂, hydrogen peroxide; ICAM-1, intercellular adhesion molecule 1; LOX-1, lectin-type oxidized low-density lipoprotein receptor 1; NADPH, nicotinamide adenine dinucleotide phosphate; NET, neutrophil extracellular trap; NF-κB, nuclear factor kappa-light-chain-enhancer of activated B cells; NO, nitric oxide; O₂, oxygen; O₂^*−, superoxide; oxHDL, oxidized high-density lipoprotein; oxLDL, oxidized low-density lipoprotein; P-selectin, granule membrane protein 140; ROS, reactive oxygen species; SDMA, symmetric dimethylarginine; TLR2, toll-like receptor 2; VCAM-1, vascular cell adhesion molecule 1; and ZO, zonula occludens.

All in all, these prothrombotic and proinflammatory responses of the dysfunctional endothelium spiral out of control, creating a vicious circle in which endothelial cell dysfunction progresses and vascular integrity is lost, resulting in a strongly increased thrombotic risk.

Endothelial Dysfunction, Reduced Vasorelaxation, Increased Vascular Stiffness, and Cardiovascular Risk

Aging as well as pathophysiological remodeling of the vascular wall by cardiovascular risk factors (eg, hypertension, diabetes, kidney disease) induce arterial stiffening, which reduces arterial compliance and increases pulsative shear and pressure on the vasculature. On structural level, arterial stiffness is characterized by collagen deposition and elastin degradation in the ECM.²⁹ Furthermore, vascular smooth muscle cell (VSMC) tone and endothelial dysfunction impact vascular reactivity, with an impaired relaxation capacity of VSMCs as well as a reduced endothelium-dependent vasorelaxation contributing to arterial stiffness. Endothelial cells play an important role in vasorelaxation through NO production by eNOS (endothelial NO synthase; Figure 2), with reduced production and/or bioavailability of NO or other vasodilatory substances reducing the vascular relaxation capacity (Figure 5).

Figure 5.

Endothelial phenotype in chronic kidney disease (CKD) conditions in relation to vasoreactivity. Endothelial phenotype in CKD conditions in terms of vasodilation/vasoconstriction. For more information, see text. eNOS indicates endothelial nitric oxide synthase; ER, endoplasmic reticulum; HCO₃⁻, bicarbonate; HDL, high-density lipoprotein; H₂O₂, hydrogen peroxide; LOX-1, lectin-type oxidized low-density lipoprotein receptor 1; NADPH, nicotinamide adenine dinucleotide phosphate; NO, nitric oxide; NOX, NADPH oxidase; O₂, oxygen; O₂^*−, superoxide; SDMA, symmetric dimethylarginine; TLR2, toll-like receptor 2; and TNFɑ, tumor necrosis factor ɑ.

The gold standard for quantifying arterial stiffness is the analysis of blood flow rate, measured as pulse wave velocity (PWV, analyzed as carotid-femoral PWV or brachial-ankle PWV). Alternatively, the shape of the arterial pressure waveform (pulse wave analysis) provides insights into arterial stiffness through quantification of the augmentation index. Endothelial function in terms of vasodilatory responses can be measured either invasively or noninvasively within the epicardial coronary arteries, the peripheral conduit arteries (analyzing flow-mediated dilation [FMD] of the brachial artery as gold standard) or within the coronary or peripheral microvasculature (Figure 6). Peripheral and coronary endothelial dysfunction have been reported to correlate with each other,^30,32 although others have reported on rather modest correlations between peripheral and coronary endothelial dysfunction and suggested a potential reflection of different pathologies or vascular beds.³³ Furthermore, whereas PWV and the augmentation index were shown to correlate well, this was not always the case for the augmentation index and the “reactive hyperemia index” as noninvasive endothelial functional readout in the periphery, potentially due to the impact of other factors than endothelial function on vascular stiffness.³⁴

Figure 6.

Overview of endothelium-dependent vasoreactivity analysis in patients. For a more in-depth discussion of these methods, we refer to 2 excellent reviews.^30,31 CBF indicates coronary blood flow; FMD, flow-mediated dilation; and PAT, peripheral arterial tonometry.

Vascular stiffness is an early prognostic marker of coronary heart disease.³⁵ Vascular stiffness increases the load on the heart; it urges a greater energy demand for cardiac ejection and triggers cardiac hypertrophy.²⁹ With increased arterial stiffness being associated with a reduced Windkessel function (ie, a reduction in the elastic buffering capacity required to dampen blood pressure fluctuations), arterial stiffness also imposes an increased pulsatility on the microcirculation and enhanced cyclical stretch on endothelial cells. Recently, it was shown that the latter triggers the secretion of protein GAS6 (growth arrest specific 6) by endothelial cells with subsequent proinflammatory GAS6/Axl signaling in monocytes and their transformation to macrophages and dendritic cells. As shown in a mouse model of Angiotensin II-induced chronic aortic remodeling, this then contributed to vascular and kidney inflammation,^36,37 linking arterial stiffness to further end-organ damage.

As endothelial dysfunction is an early event in CVD that precedes macrovascular complications, also endothelial functional analyses in terms of vascular tone regulation have been intensively studied for a predictive capacity of cardiovascular risk beyond traditional cardiovascular risk factors. In multiple studies, macro- and microvascular endothelial dysfunction could independently predict cardiovascular events in patients at risk for coronary artery disease.³¹ For example, in relation to endothelial reactivity of the conduit arteries (macrovasculature), medium to low endothelial function recorded by brachial artery FMD was associated with increased risk of cardiac events in patients with peripheral artery disease undergoing vascular surgery.³⁸ Also in patients with chronic heart failure, endothelial dysfunction analyzed by brachial and radial FMD was associated with future cardiac events or mortality.^39–41 Others showed healthy endothelial function recorded by brachial FMD in apparently healthy individuals to result in a significantly higher survival over a 5-year follow-up, even after adjustment for traditional risk factors.^42,43 Thus, endothelial cell dysfunction can add important information on cardiovascular risk beyond traditional risk factors.³⁰ Of note, this value of endothelial dysfunction analysis for cardiovascular risk prediction seems to be influenced by the analyzed patient cohort, since others could not confirm a cardiovascular risk prediction value for FMD-based analysis of endothelial dysfunction in an elderly cohort, potentially due to reduced arterial compliance (i.e. the vascular ability to expand upon pressure increases) in elderly subjects.⁴⁴ Nonetheless, this study did report an association of peripheral microvascular endothelial dysfunction with increased cardiovascular risk, as discussed in the next paragraph in more detail.

Microvascular Dysfunction and Cardiovascular Risk

Whereas larger arteries are predominantly affected by atherosclerotic changes, dysfunction of the microvasculature—the network of arterioles, capillaries, and venules enabling tissue perfusion—is of a different nature. Microvascular dysfunction (MVD) can develop at the level of the coronary microcirculation as well as in the periphery in the absence or presence of obstructive artery disease of the larger vessels. The development of MVD is multifactorial and can be the result of functional or structural changes or a combination thereof, depending on the underlying disease. In terms of functional changes, impaired vasodilatory responses—at least partially endothelium-dependent—underlie MVD. Furthermore, increases in the levels of vasoconstrictive substances in combination with an enhanced responsiveness toward these stimuli have been implicated in MVD, contributing to the occurrence of vascular spasm. Structural changes associated with MVD encompass luminal narrowing of the microvasculature due to adverse vascular remodeling and perivascular fibrosis; microvascular compression; microvascular rarefaction resulting in a loss of coherent microvascular trees (capillaries, small arterioles, and venules); and microembolization of atherosclerotic and thrombotic material.^45–48

As for macrovascular endothelial dysfunction, also microvascular endothelial dysfunction has been associated with cardiovascular risk.⁴⁹ This was for example the case in a population-based prospective study, in which peripheral microvascular endothelial dysfunction—but not FMD-mediated endothelial function analysis of conduit arteries—correlated with cardiovascular events (myocardial infarction, stroke or death) in elderly patients during 5 years of follow-up, even beyond major cardiovascular risk factors from the Framingham risk score.⁴⁴ Along the same line, peripheral microvascular endothelial dysfunction predicted ischemic heart disease, and even performed better in future cardiovascular risk prediction in nonobstructive coronary artery disease compared with other risk scores.⁵⁰ Also, retinal arteriolar endothelial dysfunction independently predicted MACE in patients with or at high risk of coronary artery disease.⁵¹

Also in relation to heart failure, both patients with heart failure with reduced left ventricular ejection fraction as well as patients with heart failure with preserved ejection fraction present with microvascular dysfunction. In failure with reduced left ventricular ejection fraction, peripheral microvascular endothelial function analysis was associated with an increased rate of HF-related events.⁵² Patients with heart failure with preserved ejection fraction present with both large vessel as well as microvascular dysfunction.⁵³ Furthermore, they display a reduction in cardiac microvascular density⁵⁴ as well as an increase in markers of endothelial cell dysfunction in myocardial biopsies, including the upregulation of endothelial adhesion molecules (E-selectin, ICAM-1), pro-oxidative regulators (NOX-2) and eNOS uncoupling.⁵⁵ Coronary microvascular dysfunction has been proposed to contribute to cardiac wall stiffening and diastolic dysfunction in heart failure with preserved ejection fraction patients. However, whether this involves a real causal relation or rather a non-causal association between coronary microvascular dysfunction and heart failure with preserved ejection fraction remains debated.^56,57 Of note, coronary microvascular dysfunction may also have a pathophysiological and prognostic role in other types of CVD, for which we refer to an excellent recent review by Del Buono et al.⁴⁶

Endothelial Cell Dysfunction and Vascular Aging

Endothelial cell dysfunction with reduced endothelial vasodilation capacity, increased inflammation and permeability and enhanced prothrombotic properties, all described above, is an important hallmark of vascular aging. Endothelial cell dysfunction is not only observed in the macrocirculation but also contributes to microvascular dysfunction upon aging. Vascular aging is furthermore characterized by functional and structural changes of the vascular wall and adventitia by processes of e.g. inflammation, vascular calcification and ECM remodeling, which further contribute to increased vascular stiffness and cardiovascular risk.⁵⁸

Increased Endothelial Cell Dysfunction in CKD

Endothelial cell dysfunction in the context of CKD has been extensively investigated both in animal studies as well as in patients (Table 1). Overall, these studies revealed a reduced endothelium-dependent vasodilation, endothelial glycocalyx damage, an increased endothelial permeability as well as increased proinflammatory and prothrombotic properties. Along with macrovascular endothelial cell dysfunction, CKD is associated with microvascular dysfunction and rarefaction. Also, CKD patients increasingly present with medial calcification, which has been associated with increased arterial stiffness.^77,78 Combined with aging-associated characteristics of the vascular wall and the immune system,⁷⁹ these pathophysiological changes underlie the vision of CKD as a process of increased vascular aging.⁸⁰

Table 1.

Endothelial (Cell) Dysfunction Analyses in Patients With CKD

Insights From Animal Studies

Animal models of CKD have provided us with many valuable insights in CKD-induced endothelial cell dysfunction as they allow the in vivo assessment of endothelial (cell) function. Induction of CKD in rats has been shown to lead to a reduced endothelium-dependent relaxation,⁸¹ an increased endothelial permeability of the thoracic aorta,⁸² and an increase in microvascular permeability that was coupled to a loss of the endothelial surface layer in a rat model of proteinuric kidney disease.⁸³ Furthermore, extensive damage to the glycocalyx was shown in 5/6-nephrectomized rats, indicated by a decreased thickness of the aortic glycocalyx and increased levels of the glycocalyx component syndecan-1 in plasma.⁸⁴ Also, in mouse models of CKD, endothelial cell dysfunction has been demonstrated at macro- and microvascular level, illustrated by a significant reduction in glycocalyx thickness and density, an upregulated expression of glycocalyx components, decreased endothelium-dependent relaxation in response to acetylcholine of aortic rings and an increased endothelial expression of ICAM-1 and VCAM-1.^85–87 Further, using microcomputed tomography imaging of the preglomerular arteries, Ehling et al⁴⁷ showed that kidney dysfunction resulted in a reduction in vessel diameter and vessel branching as well as an increase in vessel tortuosity in 3 murine models of progressive kidney disease and kidney fibrosis. In combination with an Apolipoprotein E-deficient background, CKD induction furthermore resulted in a higher plaque burden, indicating a higher atherosclerotic risk in CKD.⁸⁵

Insights From Patient Studies

Reduced Macrovascular Endothelium-Dependent Vasodilation in CKD

In terms of functional responses of the peripheral macrovasculature as assessed by FMD, multiple studies have shown that CKD patients have a significant impairment in endothelium-dependent vasodilation,^62–64 which increases with increasing CKD severity,^64,88 independently predicts cardiovascular outcome⁶⁴ and correlates with left ventricular hypertrophy.⁶⁵ Also, endothelium-independent vasodilation as measured by brachial FMD analysis upon nitroglycerin injection was significantly reduced in patients with advanced CKD—but not in patients with coronary artery disease and healthy kidney function—with vascular calcification suggested as potential underlying mechanism.⁸⁹ Additionally, multiple studies showed that as CKD progresses arterial stiffness increases,^67,68,70 although others could not confirm a correlation between PWV and kidney function after adjustment of covariates (including, but not limited to, age, sex, comorbidities, and medication use).⁶⁹

MVD and Rarefaction in CKD

CKD has been associated with an increased dysfunction of both the peripheral as well as coronary microvasculature and as CKD advances the structural and functional capillary density declines.^49,59,90 In human myocardial specimens of CKD patients, microvascular density was found to be reduced by on average 32%⁴⁹; it could already be observed in CKD3-4 and was accompanied by increased endothelial-to-mesenchymal transition and myocardial fibrosis.^49,91 Although CKD patients often present with comorbidities that strongly contribute to the development of MVD (eg, hypertension and diabetes), CKD-induced MVD—or uremic angiopathy—extends beyond these factors, with the resulting dysfunction of the microvasculature further adding to the progression of CKD. Evidence is derived from animal models of CKD as well as from CKD patient studies in which it was shown that CKD-induced microvascular rarefaction occurs in a heterogeneous pattern resulting in the development of large avascular areas.^48,49

Venous and finger plethysmography measurements indicated a compromised endothelium-dependent vasodilatory response in nondialysis-dependent CKD,^61,66 although Wang et al⁹² did not observe a correlation between estimated GFR and the reactive hyperemia index as measured by peripheral arterial tonometry. Focusing on coronary microvascular dysfunction, a recent systematic review with meta-analysis indicated CKD patients to have a significantly lower coronary flow reserve compared with non-CKD patients.⁵⁹ As in the general population, reduced coronary flow reserve as readout of coronary microvascular dysfunction is predictive for cardiovascular risk in CKD patients, even after adjustment for traditional cardiovascular risk factors.^60,93 Restoration of kidney function upon kidney transplantation partially ameliorated microvascular function as assessed by analysis of the sublingual microcirculation and the coronary flow reserve, suggesting that to some extent microvascular changes in CKD are reversible.^94,95

Endothelial Glycocalyx Damage in CKD

In patients with CKD, increased levels of the glycocalyx components syndecan-1 and hyaluronan in combination with elevated hyaluronidase activity have been detected in serum, with highest levels observed in CKD patients on dialysis.^71,84,96 Along the same line, treatment of human endothelial cells in vitro with uremic serum from pediatric CKD patients or adult hemodialysis patients led to a decrease in glycocalyx height and stiffening of the endothelial cells.⁹⁷ Glycocalyx damage was confirmed in dermal biopsies of CKD patients, in which a severe reduction in staining with ulex europaeus agglutinin-1 (a lectin binding glycoproteins and glycolipids) was observed in biopsies obtained from nondialysis-dependent CKD patients as well as hemodialysis patients.⁹⁸ Measurements of the perfused boundary region, as indicator of the thickness of the glycocalyx, are less consistent among studies. While Vlahu et al⁹⁶ observed an increase in the perfused boundary region and perfused diameter, Liew et al⁷¹ detected no significant difference in the perfused boundary region. Three months after kidney transplantation, markers of the endothelial glycocalyx improved as the perfused boundary region decreased and serum syndecan-1, sVCAM-1, and vWF levels were reduced in comparison to pretransplant levels.⁹⁹ Serum hyaluronan levels remained unchanged.⁹⁹

Soluble Biomarkers of Endothelial Health in CKD

Systemic biomarkers have been used to provide insights into the extent of endothelial cell dysfunction in patients with increased cardiovascular risk (Table 1). Dysfunctional endothelium typically produces less NO through reduced production or activity of eNOS, but the short half-life of NO makes it a difficult readout for accurately assessing endothelial cell function. Alternative systemic biomarkers that have been used when studying endothelial cell dysfunction include shedded forms of inflammation-induced cell surface adhesion receptors (eg, soluble forms of VCAM-1, ICAM-1, P-selectin, and E-selectin, with the latter regarded as most selective for endothelial cells).¹⁰⁰ With the endothelium playing an important regulating role in hemostasis, also molecules involved in platelet activation and coagulation are often quantified to assess the thrombotic responses of a potentially dysfunctional endothelium. This includes systemic levels of vWF, soluble thrombomodulin, tPA (tissue-type plasminogen activator), and PAI-1 (plasminogen activation inhibitor-1; a fibrinolysis inhibitor).^100,101 However, it should be noted that most of these systemic biomarkers are not specific for inflamed or damaged endothelium and only provide indirect insights into potential endothelial cell dysfunction.

Specifically in nondialysis-dependent patients in CKD stage 3-5, plasma and serum levels of asymmetric dimethylarginine (an eNOS inhibitor), L-arginine (the precursor of NO), sVCAM-1, sE-selectin, sP-selectin, s-thrombomodulin, tPA, uPA (urokinase-type plasminogen activator), TF and vWF were elevated (Table 1).^{71–73,75,76,102} These markers of endothelial activation associated significantly with albuminuria and a decreased estimated GFR as markers for CKD severity.^62,71 Even a mild impairment in kidney function associated independently with plasma levels of vWF and sVCAM-1 as markers of endothelial activation.⁷⁴ The effects of CKD on systemic levels of PAI-1 and sICAM-1 are less clear, varying from unchanged^62,74,76 to increased levels in nondialysis CKD patients.^73,75

Mechanistic Insights Into Endothelial Cell Dysfunction in CKD

A diversity of pathophysiological processes has been identified to contribute to endothelial cell dysfunction in CKD. These include chronic low-grade inflammation, increased oxidative stress, the accumulation of uremic toxins and associated posttranslational modifications of proteins and lipoprotein particles, as well as metabolic acidosis and high sympathetic activation (Figures 3 through 5). Furthermore, patients with CKD display signs of advanced vascular aging and present with vascular calcification. In the following paragraphs, the interaction of each of these different factors with endothelial cell dysfunction in CKD is being discussed. Effects of traditional cardiovascular risk factors such as hypertension and diabetes as frequent comorbidities and causes of CKD, are not discussed.

Chronic Low-Grade Inflammation

In the damaged kidney, unresolved low-grade inflammation triggers resident kidney cells to produce proinflammatory cytokines and chemokines and induce the deposition of ECM contributing to tubulointerstitial fibrosis. Thereby, chronic low-grade inflammation is an important driver of CKD progression.⁶⁰ Furthermore, the reduced filtration function of the kidneys in CKD leads to a systemic uremic milieu, which triggers systemic, chronic low-grade inflammation by altering the expression and release of proinflammatory mediators, such as cytokines, in the blood.¹⁰³ Hallmarks of low-grade inflammation in CKD are elevated systemic levels of the proinflammatory cytokines IL-1β, IL-6, TNF (tumor necrosis factor)-α and hsCRP (high-sensitivity C-reactive protein), which accumulate in blood due to an increased release as well as a reduced kidney clearance.^104,105

Systemic proinflammatory biomarkers are associated with CKD progression¹⁰⁶ as well as with cardiovascular risk in patients with CKD.¹⁰⁷ For example, serum levels of hsCRP could predict cardiovascular events in CKD patients,^108,109 with an adjusted hazard risk of 1.94 for cardiovascular mortality over a median follow-up of 10 years for CKD3-4 patients with high CRP (≥3 mg/L) compared with low CRP (<3 mg/L).¹⁰⁹ Also, systemic levels of IL-6 correlated with a poor outcome in CKD, and both IL-6 and TNF-α could predict all-cause mortality in CKD5D patients after adjustment for age, sex, diabetes, smoking, baseline kidney function, protein energy wasting as well as other inflammation biomarkers as hsCRP (high-sensitivity C-reactive protein).^104,107 On the other hand, also the anti-inflammatory IL-10 is upregulated in patients with CKD through increased secretion by uremic monocytes and reduced kidney clearance, with IL-10 counteracting the activated inflammatory response.¹⁰⁴ Increased IL-10 levels were found to be positively correlated with cardiovascular events in CKD patients at follow-up, with increased IL-10 levels interpreted as an increased compensatory effect in an overall proinflammatory state in CKD patients.¹¹⁰

Antibody-mediated blockade of IL-1β in the CANTOS trial (Canakinumab Anti-Inflammatory Thrombosis Outcomes Study) reduced the occurrence of major adverse cardiovascular events in patients with moderate CKD (CKD3a) patients over a mean follow-up of 3.7 years.¹¹¹ Blocking IL-6 for 24 weeks in patients with CKD3-5D and residual inflammatory risk in the RESCUE trial significantly reduced hsCRP as systemic inflammation marker, with the ZEUS study (Zlitivekimab Cardiovascular Outcomes Study) currently evaluating a potential benefit on cardiovascular outcome in CKD3-4 patients with high residual inflammatory risk.¹¹²

Chronic inflammation as well as the abovementioned cytokines are linked with endothelial inflammation and endothelial cell dysfunction. For example, IL-6 is known to trigger CRP production, with both IL-6¹¹³ and CRP¹¹⁴ decreasing endothelial production and/or activation of eNOS and thereby NO as crucial molecule in the maintenance of vascular health. IL-6 also actively damages the endothelial barrier and increases endothelial permeability by reducing endothelial expression of the adherens junction protein VE-Cadherin via the trans-signaling pathway¹¹⁵ as well as the junctional localization of the tight junction protein ZO-1 via STAT3 phosphorylation.¹¹⁶ TNF-α reduced endothelial-dependent vasodilation of mouse aortas by increasing superoxide production by activation of the NADPH oxidases,¹¹⁷ and both increased TNF-α and reduced NO could be causally linked to endothelial apoptosis in aged arteries of rats.¹¹⁸

In addition to elevated cytokine levels, patients with CKD display increased levels of danger-associated molecular pattern molecules such as HMGB1, S100 proteins and advanced glycation endproducts (AGEs), which trigger proinflammatory responses over RAGE (receptor for advanced glycation endproduct), TLRs, and NLRP3 inflammasome activation.^119,120 Furthermore, neutrophil extracellular traps as produced by activated neutrophils induce endothelial inflammation and enhanced TF expression and activity, thereby promoting coagulation.¹²¹ Other contributors to a chronic inflammatory state in CKD are a gut dysbiosis⁷⁹ as well as uremic toxins (discussed in more detail below).

A low-grade inflammatory state is not only observed in CKD, but also in other chronic diseases as well as upon aging, and is therefore referred to as inflammageing.⁷⁹ Overall, low-grade inflammation induces endothelial cell dysfunction as well as oxidative stress, which in turn amplifies the inflammatory state, as further discussed below.

Oxidative Stress

Oxidative stress is defined as the imbalance between pro-oxidants and antioxidants. An increase in pro-oxidants and the generation of reactive oxygen species (ROS) and reactive nitrogen species affects the metabolism of cells and can trigger severe cell damage and apoptosis.¹²² It can be caused by mitochondrial dysfunction resulting in increased superoxide levels, increased activity of NOX (NADPH oxidase) leading to increased hydrogen peroxide levels as well as through eNOS uncoupling resulting in peroxynitrite production.¹²²

In the vasculature, NOX is the major contributor to the generation of ROS.¹²³ Of the 7 NOX isoforms, endothelial cells express 4: NOX-1 (NADPH oxidase 1), NOX-2, NOX-4, and NOX-5, with NOX-4 being the most prominently expressed subtype in endothelial cells.^123,124 Especially NOX-2 and NOX-4 are often linked to the initiation and progression of cardiovascular complications.¹²⁴ Extended NOX activation, as e.g. through stimulation by inflammatory mediators such as TNF-α, increases ROS production in endothelial cells and thereby mediates NF-κB signaling, enhancing vascular inflammation and inducing a vicious circle of inflammation and oxidative stress.^125,126 Also, a range of uremic toxins has been described to trigger inflammation and by prolonged NOX activation to lead to oxidative stress in endothelial cells.¹²⁷ For an extensive overview on the different NOX isoforms in CVD, we refer to a detailed review by Zhang et al.¹²⁸

Lately, ROS-induced ROS release, a form of intracellular communication between ROS derived from NOX enzymes and mitochondria, has been presented as a feed-forward mechanism to maintain and amplify ROS signaling.¹²³ In the context of CKD with increased systemic AGE levels, endothelial cells stimulated with AGEs showed increased levels of NOX-2, cytosolic and mitochondrial ROS but decreased levels of mitochondrial sirtuin-3, with analysis of sirtuin-3 blockade suggesting a role for mitochondrial ROS in cytoplasmic ROS production.¹²⁹

In addition, uncoupling of eNOS is induced through CKD-mediated posttranslational modifications of LDL (low-density lipoprotein) and HDL (high-density lipoprotein), resulting in increased endothelial ROS production^130–132 (posttranslational modifications described in more detail below). Finally, a reduction in antioxidative mechanisms could increase the overall oxidative stress level. For example, patients with advanced CKD display a reduction in NRF-2 (nuclear factor erythroid 2-related factor-2), a cellular protector from oxidative stress.^133,134 Inducing the NRF-2 pathway in endothelial cells could therefore be a novel therapeutic approach treating inflammatory diseases such as atherosclerosis by protecting the endothelium for oxidative damage as shown in.¹³⁵

Based on these disturbances in the pro-/antioxidative balance also in the context of CKD,¹²² oxidative stress is believed to be an important contributor to cardiovascular morbidity in the general population as well as in CKD patients.^122,136

Uremic Toxins

Uremic toxins are broadly defined as substances of organic or inorganic origin that accumulate in the circulation due to kidney function decline and/or increased production, with harmful effects on the body. Currently >140 of such solutes have been identified.¹³⁷ Overall, the uremic milieu triggers proinflammatory effects (eg, VCAM-1 and C-C motif chemokine ligand 2 expression), NOX expression and ROS production, as well as reduced antioxidant enzyme activity in endothelial cells, as demonstrated upon incubation of endothelial cells with uremic serum in vitro (Table 2).^127,145 Furthermore, uremic serum collected from patients with CKD gradually reduces the endothelial glycocalyx height along with increasing CKD stage and increases the stiffness of the glycocalyx as well as of the actin-rich cortex beneath the plasma membrane in vitro. This, as well as uremic serum-induced reduction of eNOS and NO production in vitro, could be counteracted by blocking the mineralocorticoid receptors and the epithelial Na⁺ channel as its downstream target.¹⁴⁶

Table 2.

Impact of Uremic Toxins on Signaling Pathways in Endothelial Cells with Negative Impact on Endothelial Function

The recent systematic review by Harlacher et al¹²⁷ described in detail known uremic toxins in relation to detrimental effects on the endothelium. This revealed that uremic toxins as p-cresylsulfate, indoxylsulfate, cyanate, AGEs, asymmetric dimethylarginine and uric acid induce oxidative stress (ROS production, NOX activation) and inflammation and promote the adhesion of inflammatory leukocytes to the endothelium.^127,145 Furthermore, a subset of these uremic toxins reduces the proliferative capacity of endothelial cells and is able to trigger cell death. Also, cyanate enhances the prothrombotic effects of the endothelium by triggering the expression of TF and PAI-1.¹²⁷ As underlying mechanisms, these uremic toxins activate MAPK [mitogen-activated protein kinase]/NF-κB, RAGE, CREB (cAMP response element-binding protein)/ATF1 (AMP-dependent transcription factor 1) and AhR (aryl hydrocarbon receptor)-dependent pathways in endothelial cells (Table 2),^127,145 which are known to induce among others oxidative stress and inflammation. Indoxylsulfate also upregulates the endothelial expression of solute carrier family 22 member 6 (OAT1 [organic anion transporter 1]), a membrane transporter molecule mediating cellular uptake of p-cresylsulfate and indoxylsulfate.¹⁴⁵

Furthermore, AGEs were shown to enhance endothelial permeability and induce endothelial senescence as indicated by senescence-associated β-galactosidase staining and expression of the senescence-associated proteins p53, p21, and p16.¹⁵² p-Cresylsulfate, indoxylsulfate, cyanate, AGEs, and uric acid reduced the expression and activity of eNOS, thereby decreasing NO bioavailability and increasing vascular stiffness.¹²⁷ Along the same line, the uremic toxin kynurenine triggered an increase of superoxide production in the vasculature at least partly via AhR-dependent signaling in endothelial cells, thereby reducing NO-mediated vasorelaxation.¹⁴⁴ In patients with CKD, the plasma concentration of kynurenine positively correlated with sICAM-1, sVCAM-1, vWF, and thrombomodullin as markers of endothelial cell dysfunction.¹⁵³

Combined, this suggests an important contribution of uremic toxins to cardiovascular risk in patients with CKD. Meta-analysis studies concluded on a significant association of the uremic toxin p-cresylsulfate with cardiovascular risk in patients with CKD,¹⁵⁴ whereas indoxylsulfate and asymmetric dimethylarginine correlated with overall mortality but not cardiovascular mortality in CKD.^154,155

Posttranslational Modifications

Uremic toxin accumulation and oxidative stress in CKD do not only trigger proinflammatory signaling but also induce posttranslational modifications, which may alter the function of the targeted proteins as well as of lipoprotein particles. For example, triggered by increased urea concentrations in patients with CKD, the intracellular sorting receptor sortilin is carbamylated in CKD. Carbamylated sortilin promotes VSMC calcification in vitro and is associated with increased coronary artery calcification in CKD patients.¹⁵⁶ Also, PTMs have been identified to negatively affect lipoprotein particle function in CKD, with a negative impact on endothelial health. Both LDL and HDL particles are oxidized and carbamylated in patients with CKD,¹⁵⁷ triggered by increased oxidative stress and plasma urea concentrations in CKD, respectively. oxLDL is well known for its proinflammatory effects in both CVD and CKD patients.^157,158 Carbamylated LDL, but not native LDL, impaired endothelium-dependent vascular relaxation and enhanced ROS production through NADPH-oxidase activation and eNOS uncoupling through the LOX-1 receptor.¹³⁰ Carbamylated LDL was also shown to induce autophagy, cell death, and DNA fragmentation in endothelial cells.¹⁵⁹ Furthermore, it enhanced thrombin generation and injury-induced thrombus formation in a mouse model, as well as increased the production of TF and PAI-1 in endothelial cells through LOX-1.¹⁶⁰

Whereas HDL exerts anti-inflammatory and proproliferative effects in endothelial cells,^130,161 oxHDL (oxidized HDL) triggered NOX2-mediated ROS production as well as proinflammatory NF-κB signaling and cytokine expression in endothelial cells through LOX-1.¹³¹ Along this line, carbamylated HDL reduced endothelial migration and proliferation.¹⁶¹ Also, HDL from patients with CKD showed an enrichment in the proinflammatory protein SAA (serum amyloid A) as well as in the uremic toxin SDMA. SDMA-enriched HDL and CKD-HDL increased ROS through NADPH oxidase activation and reduced endothelial NO production via TLR2 in vitro. In contrast to HDL from healthy donors, SDMA-enriched HDL and CKD-HDL did not support endothelial repair after injury of the carotid artery in a mouse model.¹³²

In summary, CKD-induced alterations of lipoprotein particles make LDL an even more noxious particle and convert HDL from a protective to a damaging lipoprotein particle. For more details and additional alterations in lipoprotein particles in CKD, we refer to a recent detailed review by Noels et al.¹⁵⁷

Metabolic Acidosis

With a prevalence of 39% in predialysis patients with a glomerular filtration rate <20, chronic metabolic acidosis is a common complication in patients with advanced CKD,¹⁶² although it is consistently underdiagnosed and undertreated.¹⁶³ It is caused by a reduced excretion of metabolically produced acids, leading to decreased systemic bicarbonate levels. Metabolic acidosis in CKD has been associated with CKD progression as well as with an increased risk of adverse cardiovascular events, including heart failure.^164,165 On molecular level, chronic metabolic acidosis has been shown to induce ammoniagenesis and to increase the production of angiotensin II, aldosterone, and endothelin-1, in order to enhance net acid excretion.¹⁶⁶ However, sustained upregulation of these mediators exerts proinflammatory and profibrotic effects on the kidney, thus again contributing to CKD progression. Furthermore, these mediators exert proinflammatory and vasoconstrictive effects on the endothelium.^167,168 Also, in vitro studies revealed acidosis to trigger proinflammatory NF-κB signaling, endoplasmic reticulum stress, and the unfolded protein response in the endothelium via the proton-sensing receptor GPR4.^169,170 Specifically extracellular acidification inhibited store-operated Ca²⁺ entry through divalent cation channels and thereby interfered with agonist-mediated production of the protective factors NO and prostaglandin I₂ by endothelial cells.¹⁷¹ A pilot study identified an improvement of endothelial function in CKD stage 3b-4 patients upon sodium bicarbonate treatment for 6 weeks, as detected by a 1.8% increase in brachial artery flow-mediated dilation. Overall effects on cardiovascular outcome were not examined.¹⁷²

Sympathetic Nerve Activity

Sympathetic nerve activity—for example, as measured by plasma levels of norepinephrine or catecholamines—increases along with kidney function decline, potentially triggered by increased renin-angiotensin-aldosterone system signaling, reduced NO bioavailability, and increased oxidative stress, among other factors.^173,174 Increased sympathetic nerve activity contributes to hypertension, but also independent of blood pressure effects, high sympathetic nerve activity is associated with CKD progression as well as with increased cardiovascular risk in both predialysis and dialysis CKD patients.^173,175,176 Sympathetic nerve activation reduces endothelial-dependent vasodilation and increases vascular stiffness, as discussed in detail by Kaur et al.¹⁷³ In animal models, blocking sympathetic nerve activity reduced microvascular rarefaction.⁴⁹ Furthermore, in vitro studies showed that high levels of catecholamine neurotransmitters trigger endothelial adrenergic receptors, leading to endothelial permeability and glycocalyx loss in endothelial cells.¹⁷⁷ Combined, these findings suggest also a contribution of increased sympathetic nerve activity to endothelial dysfunction in CKD.

Vascular Aging

Although aging is a natural process, in case of CKD, it is accelerated. An indicator for cellular aging and subsequent decline in function is telomere length. A reduction in telomere length can drive endothelial cells into senescence, characterized by a stable arrest in cell growth and a proinflammatory phenotype.¹⁷⁸ Although telomere shortening has been observed in CKD independent of age,¹⁷⁹ a recent meta-analysis indicated a paradoxical association between CKD and telomere length. As such, the authors postulated that the shortening of telomeres associated with a declining kidney function is likely offset by cellular telomere reparative mechanisms in patients who survive longer with CKD,¹⁸⁰ but more research is needed. Moreover, to which extent telomere shortening occurs in the endothelial layer as a consequence of a reduced kidney function is unclear.

Premature aging takes place partially due to systemic inflammation (“inflammageing”).⁷⁹ The other way around, senescent cells can develop a senescence-associated secretory phenotype to release, among others, proinflammatory cytokines, growth factors and soluble receptors contributing to local as well as systemic inflammation, which accelerates tissue damage in patients with CKD.⁸⁰

Furthermore, patients with CKD display a reduction in Klotho due to impaired kidney function. Klotho is a protective protein with antioxidant, antiapoptotic, and antisenescent effects, also toward endothelial cells,¹⁸¹ and its depletion is a crucial contributor to premature vascular aging in CKD.^80,182,183 Animal studies by Shi et al linked a reduction in Klotho to a reduction in autophagy. Early αKlotho administration increased autophagic flux induction upon acute kidney injury and protected from kidney damage progression into CKD, suggesting the administration of Klotho as a potential therapeutic treatment after acute kidney injury to reverse kidney failure.¹⁸⁴ A disturbed autophagic flux is additionally induced by uremic toxins such as indoxylsulfate, p-cresyl sulfate, and indole acetic acid, leading to the accumulation of oxidized proteins and organelles and increasing the sensitivity of endothelial cells toward oxidative stress.¹⁸⁵

A depletion of Klotho in aortic endothelial and smooth muscle cells is accompanied by a significant reduction in SIRT1 (sirtuin-1). Similar to Klotho, SIRT1 is anti-inflammatory, antioxidative, antiapoptotic, and antisenescent; it inhibits the activation of NADPH oxidases and prohibits the production of ROS in endothelial cells.¹⁸⁶ Blocking of SIRT1 triggered a proinflammatory phenotype illustrated by an increased ROS production in aortic endothelial cells and reduced endothelial-dependent vascular relaxation through impaired NO production.^187,188 Beyond counteracting inflammation, oxidative stress and senescence, SIRT1 also protects from CKD-associated fibrosis and vascular calcification, suggesting SIRT1 as a potential future therapeutic target for CKD.¹⁸⁶

Smooth Muscle—Endothelium Interaction and Vascular Calcification

Within the vasculature, endothelial cells and VSMCs can bidirectionally communicate.¹⁸⁹ In relation to VSMC communication to endothelial cells, endothelial cells cocultured with VSMCs express increased levels of MMP-2 and MMP-9.¹⁹⁰ Synthetic VSMCs produce proinflammatory IL-1β and IL-6, which induced NF-κB activation and E-selectin expression in cocultured endothelial cells.¹⁹¹ Also, mechanical stress-induced microparticle production by VSMCs induced proinflammatory responses in endothelial cells.¹⁹² Despite these findings, the overall contribution of VSMC changes to endothelial cell dysfunction in CVD remains unclear. This is also true in the context of CKD. CKD patients frequently display medial vascular calcification, as for example identified in 88% of dialysis patients aged 20 to 30 years old.¹⁹³ Medial vascular calcification is associated with increased vascular stiffness as well as cardiovascular mortality in CKD patients.^194,195 Whereas the impact of vascular calcification on endothelial function has not been studied to our knowledge, a dysfunctional endothelial cell layer does contribute to medial calcification. For example, NO produced by endothelial cells counteracts VSMC calcification.¹⁹⁶ Also, inhibition of eNOS-mediated NO production by L-NAME increases warfarin-induced medial calcification in rats,¹⁹⁷ with warfarin triggering vascular calcification by interfering with the activation of the calcification inhibitor matrix gla protein. As discussed above, in CKD, endothelial cells show a reduced eNOS expression and activation resulting in a reduced NO bioavailability, for example, triggered by uremic toxins as well as hyper- and hypophosphatemia.¹²⁷ Further, uremic toxins can trigger endothelial inflammation,¹²⁷ with inflammatory mediators such as TNF-α and IL-1β reported to be able to sensitize endothelial cells to BMP-induced osteogenic differentiation into osteoprogenitor cells, which could contribute to vascular calcification.¹⁹⁸

Impact of Pharmacological Interventions on Endothelial Cell Dysfunction in CKD

Considering the function of the endothelium as gatekeeper of vascular health, maintaining its integrity by pharmacological intervention could contribute to alleviating the cardiovascular risk of patients with CKD. Drugs administered to patients with CKD to treat comorbidities as hypertension, hyperlipidemia, and diabetes have been extensively examined in light of endothelial cell function. Consequently, direct and indirect endothelial protective effects of antihypertensive, lipid-lowering (statins), and antihyperglycemic drugs have been well documented.¹⁹⁹ For an extensive review of the mechanisms responsible for these beneficial effects on the endothelium, we refer to the recent review by Xu et al.¹⁹⁹ Also in the context of CKD, endothelial protective effects of antihypertensive drugs with or without statin add-on therapy have been observed,^200,201 although it needs to be noted that ACE inhibitors—but not angiotensin receptor blockers—increased proinflammatory asymmetric dimethylarginine levels in hemodialysis patients.²⁰²

In recent years, SGLT2 (sodium-glucose cotransporter-2) inhibitors received much attention due to their cardiovascular and kidney protective effects. Also, in patients with CKD, SGLT2 inhibitors were associated with an improvement in cardiovascular and kidney health, irrespective of diabetes status.²⁰³ In terms of endothelial protective effects, a recent meta-analysis showed that treatment with the SGLT2 inhibitor dapagliflozin resulted in an improved FMD in patients with type 2 diabetes.²⁰⁴ The clinical trial PROCEED is currently ongoing to determine whether SGLT2 inhibitors are also capable of improving endothelial function in patients with diabetes and CKD.²⁰⁵ Interestingly, treatment of human cardiac microvascular endothelial cells with the SGLT2 inhibitor empagliflozin could counteract the increase in oxidative stress and the reduction in endothelial nitric oxide bioavailability caused by uremic serum exposure, but the underlying mechanisms remain unclear.²⁰⁶

Mineralocorticoid receptor antagonists, being potassium-sparing diuretics, have been shown to have endothelial protective effects as well. In patients with stable mild to moderate chronic heart failure, spironolactone treatment improved endothelium-dependent vasodilation and increased NO bioactivity.²⁰⁷ Also in the context of CKD, endothelial protective effects of mineralocorticoid receptor antagonists have been observed. In a small cohort of stable chronic hemodialysis patients, spironolactone treatment for 4 months resulted in an improvement of endothelial function as assessed by venous occlusion plethysmography.²⁰⁸ In animal models of kidney dysfunction, spironolactone and finerenone ameliorated endothelial dysfunction by enhancing NO bioavailability and reducing oxidative stress.^129,209,210

In terms of drugs targeting inflammation, blocking IL-1α/β with rilonacept for 12 weeks in patients with CKD3-4 improved FMD of the brachial artery and reduced systemic levels of hsCRP as well as endothelial expression of NADPH oxidase.²¹¹ Allopurinol, a xanthine oxidase inhibitor used to treat hyperuricemia, has contrasting results on its endothelial effects, with studies reporting no effect^212,213 to studies showing an improvement in endothelial function in terms of vasodilatory responses in patients with CKD treated with allopurinol for 8 weeks or 9 months.^214,215 Similarly, conflicting results have been published on the endothelial protective effects of Vitamin D. Whereas clinical trials have shown improvements in endothelial function upon Vitamin D supplementation,^216,217 other trials observed no change.^218–220

In recent years, many other clinical trials have been initiated investigating the effect of a wide range of pharmacological interventions or dietary supplements and/or adaptations (eg, endothelin receptor antagonism, antioxidant molecule mitoQ, low-AGE diet, plant-derived supplements or prebiotics) on endothelial function in CKD; however, outcomes were not always clear or were not yet reported (based on a search of the www.clinicaltrials.gov database for clinical trials in CKD measuring endothelial function). As the effects of CKD on the endothelium are multifactorial and the CKD patient population is very heterogeneous, protecting and maintaining endothelial integrity might require an early and multifactorial approach as well.

Conclusions

A healthy endothelial layer is a crucial gatekeeper counteracting CVD development. Patients with CKD display an impaired endothelial protective function due to the proinflammatory, prothrombotic, and uremic environment caused by their decline in kidney function, which contributes to the increased cardiovascular risk of these patients. Over the past decade, studies have started to reveal cellular and molecular mechanisms that underlie endothelial cell dysfunction in CKD, identifying a role for detrimental inflammatory and uremic mediators that are upregulated in CKD in contrast to a downregulation of protective factors. Clinical trials evaluating the effect of selected pharmacological interventions on endothelial function specifically in patients with CKD have been initiated and are ongoing, for example with a focus on targeting reduced endothelial nitric oxide bioavailability as well as increased inflammation and oxidative stress, and are expected to provide additional insights on patient level in the upcoming years. Additionally, preclinical and clinical studies should further support the development of new therapeutic options by unraveling novel disease mechanisms of increased cardiovascular risk specifically in this CKD population. This should also include a further focus on the level of immune-thrombosis interactions with the endothelium as gatekeeper of cardiovascular health. Overall, these efforts should support further cardiovascular risk reduction in this specific vulnerable patient population.

Article Information

Author Contributions

C.C.F.M.J. Baaten, S. Vondenhoff, and H. Noels performed literature research and wrote the article. S. Vondenhoff made the figures and tables.

Sources of Funding

This work was supported by the Dutch Heart Foundation (2020T020 to C.C.F.M.J. Baaten); the START-Program of the Faculty of Medicine of the RWTH Aachen University (105/20 to C.C.F.M.J. Baaten and H. Noels); a grant from the Interdisciplinary Centre for Clinical Research within the faculty of Medicine at the RWTH Aachen University (PTD 1-12 to H. Noels); by the German Research Foundation (DFG) Project-ID 322900939–SFB/TRR219 (M-05 to H. Noels) and Project-ID 403224013—SFB 1382 (A-04 to H. Noels); and the German Center for Cardiovascular Research (DZHK-B23 to H. Noels). Further funding was provided by the “Else Kröner-Fresenius-Stiftung” (Project 2020_EKEA.60 to H. Noels).

Disclosures

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.