The role of non-enzymatic post translational protein modifications in uremic vascular calcification

Abstract

Considerable technological advances have enabled the identification and linkage of non-enzymatic post-translationally modified proteins to the pathogenesis of cardiovascular disease (CVD) in patients with kidney failure. Through processes such as the non-enzymatic carbamylation reaction as well as the formation of advanced glycation end products, we now know that protein modifications are invariably associated with the development of CVD beyond a mere epiphenomenon and this has become an important focus of nephrology research in recent years. While the specific mechanisms by which protein modifications occurring in kidney failure may contribute to CVD are diverse and include pathways such as inflammation and fibrosis, vascular calcification has emerged as a distinct pathological sequelae of protein modifications. In this review, we consider the biological mechanisms and clinical relevance of protein carbamylation and advanced glycation end products in CVD development with a focus on vascular calcification.

INTRODUCTION

Vascular calcification is a highly complex, regulated, and active cell-mediated process.¹ The development of vascular calcification has been observed in both healthy aging as well as pathologic states such as atherothrombotic cardiovascular (CV) disease, diabetes mellitus (DM), and chronic kidney disease (CKD).² Accelerated calcification of a vessel wall can result in serious clinical consequences including increased wall stiffness, elevated pulse pressure, and the development of hypertension. Together, these vascular alterations can lead to left ventricular hypertrophy, myocardial stress, and decreased coronary perfusion. In a recent meta-analysis that included over 200,000 patients from 30 studies that were followed for an average of 10 years, the odds ratio (OR) for CV mortality based on the presence of vascular calcification versus no calcification was 3.94 (95% CI: 2.39–6.50).³

CVD is the leading cause of mortality in patients with CKD and end-stage renal disease (ESRD), and vascular calcification has been shown to be an independent predictor of morbidity and mortality in this population. Subgroup analysis from the aforementioned meta-analysis, showed the OR for any CV event was significantly higher in individuals with CKD plus vascular calcification versus no calcification (OR 6.22, 95% CI: 2.73–14.14). Thus, the high prevalence of vascular calcification in patients with CKD may be a significant contributor to their increased risk for adverse CV events.³ A better understanding of the pathogenesis of vascular calcification is critical to developing better therapies for the CKD population at such high CV risk. Extensive research has shown that mineral abnormalities, oxidative stress, inflammation, and deficiency of endogenous calcification inhibitors such as Klotho, matrix Gla-protein (MGP), osteopontin, osteoprotegerin, and fetuin-A can promote the development of arterial calcification in patients with CKD.^4,5 However, recent advances in areas such as mass spectrometry and bioinformatics have contributed to a growing understanding that pathologic non-enzymatic post-translational modification (NEPTM) of proteins may further and distinctly contribute to the substantial excess CV risk in CKD, possibly through calcification pathways. For example, high urea levels facilitate the NEPTM of proteins through a process called protein carbamylation, pathologically altering their structure and function in a variety of pathways including vascular calcification.^6,7 Additionally, as DM remains the most common cause of CKD, advanced glycation end products, resulting in part from hyperglycemia induced NEPTMs, have been implicated in the novel pathophysiology of uremia including vascular calcification.^8,9 This review considers contemporary data on the role of carbamylated proteins and advanced glycation end products in the development of CVD with a focus on vascular calcification.

PROTEIN CARBAMYLATION AND CARDIOVASCULAR DISEASE

Considerable evidence now suggests that a major consequence of chronically elevated urea levels in patients with CKD is the increased chemical modification of proteins through a process known as carbamylation (or “carbamoylation” in some reports, see below).^6,10,11 Carbamylation is a non-enzymatic, post translational, spontaneous reaction of cyanate to protein or amino acid functional groups (Figure 1). Because cyanate is a reactive dissociation product of urea, when kidney function and thus urea excretion decline, systemic levels of cyanate and thus protein carbamylation, naturally increase. It should also be noted that free amino acids (AAs) compete with proteins for reaction with cyanate, in essence shielding proteins from carbamylation, and AA deficiencies from a variety of causes can exacerbate carbamylation burden.¹² Additionally, Wang demonstrated that beyond renal insufficiency and azotemia, inflammation can create a unique source of cyanate and carbamylation via myeloperoxidase (MPO) activity.^13,14 MPO, mainly stored in granules of neutrophils, monocytes, and certain tissue macrophages, can react with hydrogen peroxide and environmentally derived thiocyanate to generate cyanate and thus promote the carbamylation reaction to occur (Figure 1). Thiocyanate derived from foods, smoke exposure, or even air pollution,¹⁵ can be oxidized through the MPO catalyzed reaction with hydrogen peroxide yielding an alternative pathway to cyanate. Given the marked increased levels of MPO at sites of inflammation, it was demonstrated MPO could further promote carbamylation independent of urea load at sites of inflammation.^13,16

Figure 1.

The carbamylation reaction.

The net result of these reactions, which we will call “carbamylation”, is the addition of a “carbamoyl” moiety (-CONH₂) to a functional group (Figure 1). It must be noted that several authorities accurately state that the reaction in question is more correctly deemed “carbamoylation” and, in fact, “carbamylation” denotes a different chemical reaction (the reversible interaction of CO2 with a and £-amino groups of proteins).¹⁷ However, over the past decades, the biomedical literature has skewed towards calling the reaction involving a carbamoyl moiety, “carbamylation”, and we will continue in this manner for convention. Such reactions can alter the charge, structure, and thus functional properties of various enzymes, hormones and other proteins, ultimately contributing to the failure of normal physiological processes as kidney failure ensues.⁶

The protein carbamylation process

Under physiological conditions, urea dissociates into cyanate and its tautomer, isocyanate.⁶ The normal cyanate to urea ratio typically averages less than 1:100, thereby favoring urea. However, if urea levels increase as seen in conditions such as CKD, so can cyanate. Isocyanate is the active form of cyanate and was initially discovered to be a highly reactive electrophile that quickly reacts with nucleophilic groups such as primary amines and free sulfhydryls. It was later found that cyanate can also produce irreversible modifications of primary amines and reversible modifications of thiols, hydroxyls, phenols and imidazole groups.^10,18–20

In normal health, a basal level of carbamylation appears to occur as demonstrated from animal and human studies.^13,21 The notion that the reaction occurs under physiological conditions at a baseline level and can thus naturally accumulate on proteins with long half-lives, was the basis for investigating the age-related kinetics of carbamylation accumulation.²² Interestingly, the accumulation rate of carbamylation over time in the skin of different mammalian species with different life expectancies was inversely correlated with longevity. Taken in the correct context, this suggests that protein carbamylation may be considered a hallmark of aging.²² However, as kidney function declines, both urea levels and cyanate generation increase creating an environment that promotes protein carbamylation. In health, the plasma concentration of isocyanate is ~45 nmol/L, and in uremic patients it reaches 140 nmol/L.²³ These relatively low levels in comparison to other biomolecules can be accounted for by the constant generation of cyanate via urea dissociation and its rapid consumption through binding to neighboring proteins and amino acids. Evidence for this is reflected by the high serum concentrations of many carbamylated free amino acids in patients with ESRD that exceed the concentrations of their unmodified precursors.²⁴

Emerging links between protein carbamylation with cardiovascular disease

Accumulating observational data have now linked protein carbamylation to the development of CVD beyond a mere epiphenomenon. Clinical studies have demonstrated a strong and independent association between serum carbamylated protein levels and CV and all-cause mortality in ESRD patients on hemodialysis as well as individuals with preserved kidney function.^{12,13,25–28} These studies have employed mass spectrometry to quantify validated markers of total carbamylation burden in patients followed prospectively for clinical outcomes (e.g. carbamylated albumin levels [C-Alb] offer time averaged carbamylation assessment analogous to hemoglobin (Hb)A1C in DM; or protein bound homocitrulline levels which measure the amount of carbamylated lysine residues in a given bio-sample).^{12,13,25–28}

In a case-control study involving age- and gender-matched cases (n=150) with known CVD compared to control patients (n=300), plasma protein-bound homocitrulline, measured using mass spectrometry, predicted increased risk of coronary artery disease (CAD), future myocardial infarction (MI), stroke, and death.¹³ Moreover, higher levels of protein-bound homocitrulline were associated with increased risk of prevalent CVD. Subjects in the highest quartile of carbamylation had a 7–8-fold higher risk of CVD as compared to the lowest quartile. The same report went on to confirm these findings in a replication study stemming from the same cohort (n=550), using the outcome of a major adverse cardiac event (MACE) over a 3-year period including subjects undergoing a revascularization procedure, or experiencing nonfatal MI, stroke, or death. Subjects who experienced a MACE had significantly higher carbamylation levels, and these findings remained significant following adjustment for traditional CVD risk factors, eGFR, and inflammatory markers. In another study that examined 96 patients with CKD, elevated carbamylated LDL concentrations were significant predictors for CV events and all-cause mortality.²⁹

Notably, similar findings have been found in ESRD patients. Koeth et al. measured serum protein bound homocitrulline in a group of 347 patients undergoing maintenance hemodialysis with 5 years of follow up.²⁶ After multivariable adjustments, carbamylation levels in the highest tertile conferred a more than double risk of death compared to patients with carbamylation levels in the middle or lowest tertiles (adjusted hazard ratio [HR] 2.4; 95% confidence interval [CI], 1.5–3.9; and adjusted HR, 2.3; 95%CI, 1.5–3.7; respectively). Nearly simultaneously, we reported similar findings using 2 distinct cohorts of hemodialysis patients.¹² The first cohort consisted of 187 incident dialysis patients followed for up to 1 year. Like the Koeth et al study, we found a risk ratio 3.23; 95% CI, 1.74 to 6.00 for the top tertile of carbamylation (measured using C-Alb) compared to the lowest. In the largest cohort in which carbamylation has been studied, we employed data and blood samples from the randomized controlled “German Diabetes and Dialysis” study which investigated the benefits of the cholesterol lowering drug atorvastatin in diabetic dialysis patients (n= 1,161).³⁰ In this group of prevalent diabetic hemodialysis subjects, we again found a significant risk of death among individuals with the highest tertile of C-Alb (HR 2.25; 95%CI, 1.42–3.56) even after multivariable adjustments. Thus, in well conducted studies by different research teams, with replication, and using different assays, carbamylation has been linked to mortality in patients with preserved kidney function as well as ESRD.

Protein carbamylation and the development of arterial calcification: A potential mechanism to explain epidemiological observations?

The epidemiological findings described above denote compelling associations but still lack mechanistic insights. However, many groups have pursued more targeted clinical observations along with fundamental investigations to better elucidate the mechanistic underpinnings of the perceived increased risks associated with carbamylation load. Multiple experimental studies suggest carbamylation can accelerate the biochemical events of atherosclerosis, thrombosis, and vascular calcification (below) via its effects on lipoproteins,^31–34 collagen,³⁵ fibrin,³⁶ mitochondrial proteins,³⁷ and proteoglycans and fibronectin.³⁸

In a recent breakthrough study by Mori, et al., the investigators showed that protein carbamylation can directly exacerbate the development of human vascular smooth muscle cell (VSMC) calcification.⁷ Protein carbamylation was found to accelerate VSMC calcification by suppressing ectonucleotide pyrophosphate/phosphodiesterase 1 (ENPP1), a critical enzyme involved in pyrophosphate generation that functions as a potent inhibitor of ectopic calcification. The decreased expression of ENPP1 correlated with the extent of calcification development in VSMCs exposed to urea. Interestingly, ENPP1 itself was not carbamylated, but several mitochondrial, cytoskeletal, and associated proteins were found to be carbamylated in urea-treated VSMCs and cyanate-treated rat aorta. Protein carbamylation induced mitochondrial dysfunction resulting in oxidative stress in VSMCs exposed to urea or cyanate treatment. Similarly, mitochondrial dysfunction was also found in aortas from rats fed diet high in urea.

These data highlight the critical role of protein carbamylation in disrupting mitochondrial bioenergetics and oxidative stress under uremic conditions, with the consequence of calcifying effects in arteries. In the study above by Mori, et al., importantly, the investigators found that ENPP1 knockdown in VMSCs abolished the effects of carbamylation on ectopic calcification. Further studies are warranted to further delineate the mechanisms by which carbamylation promotes the development of arterial calcification and the signaling pathways responsible for suppressing ENPP1 expression. Additionally, it is currently unknown whether critical endogenous or circulating regulators of arterial calcification such as Klotho, FGF-23, MGP, or fetuin-A are carbamylated and if so, how this may influence their function. Moreover, to date, there have been no human studies examining plasma carbamylated protein levels as clinical predictors of arterial calcification or arterial stiffness, such as with pulse wave velocity (PWV) or augmentation index (AI). Additionally, no studies have yet examined the association of carbamylated protein levels with spiral computer tomography assessment of arterial calcification. These are all areas of future research that will be most informative when undertaken.

Similar to VSMCs, the endothelium can play an important role in the development of arterial calcification and atherosclerosis, and increasing evidence has implicated lipoprotein carbamylation as an notable mediator of endothelial dysfunction.^29,39 Carbamylated LDL has been shown to impair endothelium-dependent relaxation by stimulating endothelial nitric oxide synthase (eNOS) uncoupling, while increasing reactive oxygen species (ROS) production by activating NADPH-oxidase. Additionally, carbamylated LDL interacts with endothelial lectin-like-oxidized LDL receptor-1 (LOX-1), and its activation results in endothelial cytotoxicity.^29,40 There is evidence that the endothelium functions as a source of osteoprogenitor cells in vascular calcification development and it is currently unknown whether this process can be promoted by carbamylated proteins.³⁹

Carbamylated LDL has been reported to have several additional pro-atherosclerotic biological activities: Carbamylated LDL is recognized by macrophage scavenger receptors and this facilitates cholesterol accumulation and macrophage foam-cell formation, as well as pro-inflammatory signaling.¹³ Additionally, carbamylated LDL promotes adhesion of monocytes to endothelial cells and induces endothelial cell apoptosis by activating the mitogen-activated protein kinase pathway.^40,41 There appears to be overlapping mechanisms of carbamylated LDL on VSMCs in atherogenesis that are shared with the development of VSMC calcification. For example, carbamylated LDL is a potent stimulator of VSMC proliferation and activation of pro-apoptotic pathways.^13,42 Both these VSMC events are key pathogenic processes involved in the initiation and progression of vascular calcification and may represent other mechanisms by which carbamylated proteins promote the calcific response in uremia.^5,43 Similarly, carbamylation of high-density lipoprotein (HDL) produces a dysfunctional form contributing to failure of atheroprotective functions.^13,44,45

This evidence has translated into more specific clinical observations as well. For example, in a study by Spinelli, et al., the investigators assessed the association between anti-carbamylated protein (anti-CarP) antibodies with measures of subclinical atherosclerosis and arterial stiffness in 50 patients with rheumatoid arthritis who carry a high risk of CV morbidity and mortality.⁴⁶ The authors found that anti-CarP antibodies were significantly associated with brachial artery flow-mediated dilatation (FMD), carotid intima-media thickness (IMT) and cardio-ankle vascular index (CAVI) suggesting a specific vascular link to the protein modification that could also account for mortality observations.

Analogous to the assessment of glycated hemoglobin for the assessment of DM control, measurement of long lived proteins for carbamylation levels offers a time-averaged record of carbamylation burden from inputs such as urea concentrations, amino acid deficiency, and inflammatory state.⁶ Given carbamylation’s emerging role in the development of arterial calcification and CVD, further studies are warranted to determine whether carbamylated proteins could be used as biomarkers for assessing the extent of arterial calcification. This could have significant bearing to help inform early preventative and management strategies before overt CV complications of kidney failure develop.

ADVANCED GLYCATION END PRODUCTS AND VASCULAR CALCIFICATION

A multitude of studies have demonstrated that hyperglycemia, and insulin intolerance and resistance in DM are associated with progression of CVD.⁴⁷ Prospective studies have shown that intensive glycemic control significantly lowers the incidence and progression of CAD long-term compared to standard therapy.^48–50 Two landmark studies, the DCCT (Diabetes Control and Complications Trial) and the UKPDS (United Kingdom Prospective Diabetes Study) demonstrated reductions in both diabetic microvascular and macrovascular complications in patients with type 1 and type 2 DM which correlated with duration of intensive glycemic control.⁴⁷ These data support the notion that long-term hyperglycemia is imperative to the development and progression of diabetic vascular complications. Additionally, this has given rise to the concept of “metabolic memory” that is perhaps best reflected by animal studies demonstrating continued progression of diabetic retinopathy or sustained increases in markers of oxidative stress despite correction of hyperglycemia.^51,52 While the mechanisms underlying metabolic memory remain incompletely understood, accumulating data have implicated advanced glycation end products (AGEs) as important mediators in the development of CVD.

Pathophysiology of AGEs

AGEs are a heterogenous class of molecules or compounds that are derived from the non-enzymatic glycation of proteins, lipids or nucleic acids.^53,54 Similar to protein carbamylation, effects of AGEs include alteration of the normal function of proteins and lipoproteins via several mechanisms including, change in molecular conformation and clearance, alteration of enzyme activity, and interference with receptor recognition.⁵⁵ Additionally, AGEs can contribute to a variety of complications by engaging a multiligand receptor for AGEs (RAGE, detailed below). AGEs can accumulate in virtually every tissue type, including the vasculature, kidney, and eye contributing to sequelae such as CVD, nephropathy, and retinopathy, respectively. Production of AGEs was first described in 1912 in what is known as the Maillard reaction named after French scientist, Louis Camille Maillard (1878–1936).^56–59 This complex reaction involves a series of steps in which reducing sugars undergo non-enzymatic reactions that produce reactive carbonyl compounds and subsequent glycoxidation of proteins, lipids and nucleic acids. Work on the Maillard reaction initially focused on food and beverages, and was later found to occur in virtually all heat processed and stored foods.⁶⁰

AGE formation via the Maillard reaction stems from Schiff bases and the Amadori product, a 1-amino-1-deoxyketose, generated through the interaction between the carbonyl group of a reducing sugar such as glucose, and amino groups on proteins, lipids and nucleic acids. After formation of an unstable Schiff base, molecular rearrangement (Amadori rearrangement) leads to the formation of Amadori products, characterized by their stable keto-amine bond. The formation of Amadori products is not terminal, however, and in later steps, this cumulative and irreversible reaction evolves with increasing exposure to modifying agents and is amplified by oxidative processes, generating a variety of intermediate carbonyl compounds that may in turn also bind to proteins.⁶¹ Amadori compounds can undergo dehydration, oxidation, cyclization, and condensation reactions that lead to the generation of protein-bound AGEs. As glycation and oxidation occur simultaneously and share common steps, the combination of these related pathways is often globally referred to as ‘glycoxidation’, whereas glycation only refers to the reactions from the binding of sugars to proteins.^61,62 Both reactions lead to the formation of stable, generally fluorescent, compounds considered “AGEs”, together with products of other oxidative pathways such as advanced lipoxidation end products (ALEs)⁶³. These terminal products are analogous to the Maillard products found and described in food, but appear to a lesser extent in vivo.⁶¹

Formation of AGEs is enhanced in the presence of chronic hyperglycemia and is proportional to the availability of substrate (i.e. monosaccharides). As with other NEPTMs like carbamylation, AGE formation is influenced by the rate of protein turnover. Proteins with long half-lives that contain significant lysine and arginine content (for example collagen and elastin) are particularly susceptible to glycation. While AGE formation occurs normally during healthy aging, this process is accelerated in disease states including hyperglycemia, oxidative stress and inflammation, as well as tobacco smoking, exposure to transitional metals, and reducing agents.^56,64,65 For example, during periods of oxidative stress, reducing sugars, amino acids and lipids undergo auto-oxidation to generate additional reactive carbonyl compounds and increase production of AGEs.

It has been postulated that the development of hyperglycemia in the early stages of DM leads to a proportional increase in AGE formation and oxidative stress. This leads to glycation of mitochondrial respiratory chain proteins and mitochondrial DNA damage, resulting in a self-perpetuating cycle of AGE formation and oxidative stress that can then occur independently of hyperglycemia.⁶⁶ Alternatively, glycated albumin can disrupt normal glucose metabolism in muscle and adipocytes, thereby leading to reduced glucose uptake mediated by insulin, driving hyperglycemic states.⁶⁷ Thus, it appears that AGEs both independently and synergistically with DM, can contribute to the progression of complications in this patient population. Given such relevance, many AGEs have been identified and are used as biomarkers in research studies. For example, N-(Carboxymethyl)lysine (CML) is commonly used as an AGE marker in multiple disciplines from clinical outcome studies to food analysis.^68–70

AGEs and cardiovascular disease

Considerable evidence has linked AGEs to the development of macrovascular and microvascular diseases which are responsible for serious complications in DM.^67,71,72 Increased circulating AGE levels have been associated with various cardiac pathologies, including systolic and diastolic dysfunction, arrhythmias, and CAD in patients with diabetes.^73–75 In the Invecchiare in Chianti study that included 1113 adults, participants with plasma CML (assessed by ELISA) in the highest tertile had greater all-cause (HR 1.84, 95% CI 1.3–2.6, P<0.01) and CVD (HR 2.11, 95% CI 1.27–3.49, P<0.01) mortality compared to the lower two tertiles after adjustment for confounders. This association remained significant in adults without DM.⁷⁶ In a study that examined 98 patients with type 2 DM and 61 age-matched controls, investigators found that serum pentosidine levels (another AGE biomarker, assessed by ELISA) were significantly correlated with arterial stiffness (heart brachial PWV (r=0.304, P<0.01) and carotid intima-media wall thickness (IMT; r=0.300, P<0.01).⁷⁷ Yozgatli et al., showed in a prospective cohort study of 563 participants that increased tissue AGE levels as inferred by skin autofluorescence, a noninvasive technique for evaluating tissue content of several types of fluorescent AGEs, was associated with macrovascular events (HR 1.53, P<0.001 and HR 1.28, P=0.03) after correction for UKPDS score including CAD, peripheral vascular disease, and cerebrovascular disease.⁷⁸

AGE receptors and cardiovascular disease

AGEs can exert receptor-dependent and receptor-independent effects. They bind to a variety of extracellular and intracellular receptors of different cell types, including VSMCs, endothelial cells, macrophages, and adipocytes.^53,79–81 Scavenger receptors (such as macrophage scavenger receptor-AI, macrophage scavenger receptor-AII, CD68 and CD36) are responsible for the removal of AGEs.⁸² RAGE (receptor for advanced glycation end products), a member of the immunoglobulin multiligand receptor family, has been implicated for many biological effects of AGEs.⁸³ RAGE is composed of a multiligand binding extracellular domain, a membrane spanning domain, and an intracellular carboxy-terminal domain.⁸⁴ Extracellular metalloproteinases can cleave the cytosolic portion of cell surface RAGE on endothelial cells leading to a circulating receptor (sRAGE).⁸⁵ Multiple alternative splice forms of RAGE exist leading to isoforms with partial functionality, including N-truncated RAGE, dominant negative RAGE, and endogenous secreted RAGE (esRAGE). AGE- RAGE interactions can trigger complex intracellular signaling cascades as noted in Figure 2 and RAGE activation has been implicated in inflammatory responses, apoptosis, prothrombic activity, expression of adhesion molecules, and oxidative stress.⁸⁶

Figure 2:

Adverse effects of advanced glycation end products (AGEs) in vasculature. Receptor for advanced glycation end products (RAGE); Low density lipoproteins (LDL); High-density lipoproteins (HDL); insulin-like growth factor-1 (IGF-1); platelet derived growth factor (PDGF); vascular smooth muscle cells (VSMCs).

Researchers have hypothesized that sRAGE may actually neutralize the action of AGEs exerting a protective role against the development of CVD and that some ligand to receptor ratio may be informative. Indeed, in chronic pro-inflammatory conditions such as atherosclerosis, plasma S100A12 (a RAGE ligand) levels are increased while plasma sRAGE levels are decreased (both assessed by ELISA), and these alterations associate with risk of CVD in diabetic and non-diabetic patients.⁸⁷ In another study that examined 206 (171 nondiabetic) patients with ESRD, the investigators found that patients with plasma esRAGE in the lowest tertile had a significantly higher incidence of CV death compared to those with levels in the highest (HR 0.4, 95% CI 0.18–0.89) and middle tertile (HR 0.26, 95% CI 0.10–0.66).⁸⁸ Koyama, et al. showed that plasma esRAGE (assessed by ELISA) was inversely associated with carotid or femoral atherosclerosis as assessed by IMT in a cohort of 203 type diabetic and 134 nondiabetic participants.⁸⁹ Overall, higher s100A12 and lower sRAGE levels were also associated with inflammation, CVD, and mortality in CKD patients; although this has not been a consistent finding.^90–92

Why the associations between AGEs, sRAGE and esRAGE with adverse outcomes have been conflicting, remains unclear and controversial. Potential contributors to these variations include genetic and racial factors affecting levels of the AGE receptors.^93,94 Also, importantly, recent work has shown that when measured using standard ELISA technique, CML, sRAGE and esRAGE lack specificity which could contribute to variability in findings between studies.⁹⁵ The ELISA technique lacks standardization highlighting the need for more specific analytical techniques such as mass spectrometry in future studies in order to help draw reproducible clinical conclusions.⁹⁵

AGEs, AGE receptors, and vascular calcification

Accumulating data have now linked AGEs to the development and progression of vascular calcification. In a study of 122 Japanese type 2 diabetic patients, the investigators found that AGEs as assessed by skin autofluorescence was positively correlated with coronary artery calcium (CAC) scores.⁹⁶ Another study that examined vascular deposition of AGEs in radial arteries of 54 CKD patients (33 hemodialyzed, 21 pre-dialysis), found that vascular content of AGEs by immunohistochemistry was positively correlated with arterial calcification assessed on histologic examination.⁸ While there is some evidence to suggest that AGE levels may be useful as a biomarker to assess for the presence and severity of CAD, additional studies will be needed to assess its clinical utility as a biomarker for vascular calcification.⁹⁷

Emerging data also provide mechanistic insights to this phenomenon by suggesting AGEs can directly stimulate the development of vascular calcification. Ren, et al. showed that rat aortic VSMCs incubated with AGEs enhanced VSMC calcification, in vitro and these effects were time- and dose- dependent.⁹⁸ RAGE expression was also increased together with upregulation of bone proteins and enzymes, including osteopontin and alkaline phosphatase (ALP) in VSMCs exposed to AGE. Tanikawa, et al. showed that treatment of VSMCs with AGEs enhanced the development of calcification, increased expression of bone markers, Runx2, ALP, and osteocalcin, and these effects were inhibited by anti-RAGE antibody.⁹⁹

In a recent study by Belmokhtar, et al, the investigators found an increase in serum CML and S100 protein concentrations in CKD apoe^−/− mice that developed significant vascular calcification.⁹ Stimulation of VSMCs with Pi or S100A12 induced mineralization and osteoblast transformation, and upregulation of Pi co-transporter, Pit1, in vitro. These changes were inhibited by phosphonoformic acid (Pi cotransporters inhibitor) and Ager (RAGE) deletion as discussed in further detail below. Koike, et al. showed that treatment of rat VSMCs with AGEs markedly increased apoptosis and NAD(P)H oxidase-derived oxidative stress, in vitro.¹⁰⁰ Pyruvate dehydrogenase kinase 4 (PDK4) is a critical mitochondrial matrix enzyme and its hyperactivation has been reported in calcified vessels in patients with DM.¹⁰¹ Ma, et al. showed that CML can accelerate the development of VSMC calcification through PDK4 activation and these effects were attenuated by either inhibition of PDK4 expression or RAGE blockade, in vitro¹⁰² In a recent study by Zhu, et al., the investigators presented evidence that AGEs accelerate the development of vascular calcification partially through the HIF-1α/PDK4 pathway and suppress glucose metabolism.¹⁰³ The investigators found that HIF-1α translocation and gene expression were promoted by AGEs, and HIF-1α stabilization and nuclear translocation increased PDK4 expression in VSMCs, in vitro. Additionally, AGEs suppressed glucose metabolism. Knockdown of PDK4 by interference RNA attenuated AGEs-induced VSMC calcification and enhanced AGEs-down-regulated glycolysis.

A number of studies have also shown therapeutic benefit of RAGE deletion in ameliorating the development of vascular calcification: In the study by Belmokhtar, et al, described above, development of increased vascular calcification was abrogated in CKD apoe^−/− ager (RAGE) ^−/− knockout mice and in VSMCs with ager deletion exposed to high phosphate or S100A12.¹⁰⁰ Wei, et al. showed that diabetic rats treated with nicotine developed significant aortic calcification and elevated levels of aortic AGEs and RAGE expression, and reduced activity of Cu/Zn superoxide dismutase (SOD) activity.¹⁰⁴ Enhanced Nox1, RAGE expression and production of intracellular superoxide anions, and reduced Cu/Zn SOD expression induced by ROS was attenuated by anti-RAGE antibody as well as ROS inhibition in VSMCs, in vitro. Similarly, in an elegant study that examined the role of RAGE in the development of uremic atherosclerosis, anti-RAGE antibody was injected into apoe^−/− mice that underwent 5/6 nephrectomy.¹⁰⁵ The investigators found that aortic plaque area fraction was reduced by 59% in anti-RAGE antibody treated mice compared to placebo antibody treated mice. Taken together, these results suggest that pharmacologic inhibitors of RAGE may be a therapeutic strategy for targeting vascular calcification and atherosclerosis in CKD, and provides strong rationale for further studies.

CONCLUSION

Substantial technological advances evolved from traditional analytical chemistry, and molecular and cellular techniques, have enabled the rigorous study of protein modifications such as protein carbamylation and glycation, reshaping our understanding of CVD in CKD. At present, we remain in the early scientific discovery phase in our understanding of non-enzymatic post-translational protein alterations and the development of arterial calcification. However, emerging experimental data have created clear links between these protein changes and the development of vascular calcification and further studies are critically needed. This rapidly evolving field has a range of potential clinical applications including biomarker discovery, CVD risk stratification, and the development of new diagnostic and therapeutic strategies built on an improved understanding of biological mechanisms underlying the development of vascular calcification in CKD.

Clinical Summary.

• The non-enzymatic post-translational modification (NEPTM) of proteins that can result from processes such as carbamylation and glycation, yielding carbamylated proteins or advanced glycation end products, respectively, are abundant in CKD and can directly stimulate the development of vascular calcification.

• Further studies are critically needed to better understand the role of carbamylated proteins and advanced glycation end products in the development of vascular calcification, and whether interventions aimed at reducing these protein modifications translate into reduced vascular calcification burden and meaningful clinical improvements in CKD.