Carbamylation of elastic fibers is a molecular substratum of aortic stiffness

Because of their long lifespan, matrix proteins of the vascular wall, such as elastin, are subjected to molecular aging characterized by non-enzymatic post-translational modifications, like carbamylation which results from the binding of cyanate (mainly derived from the dissociation of urea) to protein amino groups. While several studies have demonstrated a relationship between increased plasma concentrations of carbamylated proteins and the development of cardiovascular diseases, molecular mechanisms explaining the involvement of protein carbamylation in these pathological contexts remain to be fully elucidated. The aim of this work was to determine whether vascular elastic fibers could be carbamylated, and if so, what impact this phenomenon would have on the mechanical properties of the vascular wall. Our experiments showed that vascular elastin was carbamylated in vivo. Fiber morphology was unchanged after in vitro carbamylation, as well as its sensitivity to elastase degradation. In mice fed with cyanate-supplemented water in order to increase protein carbamylation within the aortic wall, an increased stiffness in elastic fibers was evidenced by atomic force microscopy, whereas no fragmentation of elastic fiber was observed. In addition, this increased stiffness was also associated with an increase in aortic pulse wave velocity in ApoE−/− mice. These results provide evidence for the carbamylation of elastic fibers which results in an increase in their stiffness at the molecular level. These alterations of vessel wall mechanical properties may contribute to aortic stiffness, suggesting a new role for carbamylation in cardiovascular diseases.


Results
Bovine aortic elastin is carbamylated in vivo. HCit was first quantified in total extracts of bovine (Bos taurus) aortas in order to determine the carbamylation rate of aorta proteins and its evolution with age. Mean HCit concentrations were about 0.46 mmol/mol Lys in 6-month-old bovines and progressively increased with age, reaching 0.92 mmol/mol Lys in 9-year-old animals ( Fig. 2A). Quantification of HCit in elastin extracted from the same samples showed that it was carbamylated despite its low number of free lysine residues. HCit concentrations were expressed as ratios to glutamate content in order to avoid any artificial overestimation due to the low lysine content, and ranged from 2.58 mmol/mol Glu in younger bovines to 4.70 mmol/mol Glu in the older animals. Calculation of Spearman's correlation coefficient showed a significant (p < 0.001) increase in elastin carbamylation rate with age (Fig. 2B).
Neither morphology of bovine elastic fibers nor their sensitivity to elastase are modified by in vitro carbamylation. Elastin extracted from bovine aorta was carbamylated in vitro by incubation with 100 mM NaCNO (or 100 mM NaCl in control conditions) for 24 h at 37 °C and then observed by scanning electron microscopy (SEM). SEM pictures showed elastin forming a network of intertwined fibers, without any morphological change after carbamylation (Fig. 3A). Sensitivity of carbamylated elastin to proteolysis was then studied by quantifying elastin peptides released after 8 and 18 h of incubation with pancreatic elastase (Fig. 3B). The concentrations of peptides released from carbamylated elastin were not significantly different in comparison with control elastin at both incubation times (8 h-incubation: 29.3 ± 6.0 μg/mL vs 30.8 ± 13.7 μg/mL, 18 h-incubation: 47.0 ± 13.1 μg/mL vs 41.2 ± 12.6 μg/mL), suggesting that carbamylation does not alter sensitivity of aortic elastin to elastase-driven proteolysis.
Enhancement of carbamylation in NaCNO-fed wild-type mice does not result in elastic fiber degradation. Wild-type (WT) mice received water supplemented with 1 mM NaCNO (or 1 mM NaCl in control conditions) for 3 weeks in order to enhance the carbamylation reaction. HCit concentrations measured in total aorta extracts revealed a significant ninefold increase (p < 0.01) of carbamylation rate of proteins in NaCNO-fed mice (2.11 ± 0.97 mmol/mol Lys vs 0.24 ± 0.02 mmol/mol Lys in control group) (Fig. 4A). Confocal microscopy analysis of aorta cross-sections did not show any morphological differences of aortic elastic fibers in NaCNO-fed mice in comparison with control mice (Fig. 4B). The evaluation of elastic fiber degradation rate based on the calculation of a rupture index (ratio of the number of breaking points to the total area of fibers) did not reveal any significant difference between the two mice groups after 3 weeks of NaCNO diet (Fig. 4C).
Carbamylation of murine elastic fibers results in an increase in their stiffness. Molecular stiffness of elastic fibers can be assessed by measuring the Young's modulus (YM), which is calculated after analysis of aortic cross-sections by AFM. In AFM pictures, the modulus is inversely correlated to the color intensity: thus, elastic fibers appear in white, which means that they are stiffer than the brown, inter-fiber spaces (Fig. 5A). After 3 weeks of feeding WT mice with 1 mM NaCNO, the YM of the elastic fibers was increased by + 102% (975 ± 240 kPa vs 482 ± 21 kPa, p < 0.01) when compared with controls (Fig. 5B). The stiffness of the inter-fiber spaces, which contains smooth muscle cells and other matrix proteins including collagens, was generally lower, but not significantly so, after carbamylation (32 ± 10 kPa for NaCNO-fed mice vs 50 ± 21 kPa for control mice).
In order to determine if the increase in stiffness observed at the molecular level had an impact on the stiffness of the whole vessel, aortic pulse wave velocity (aPWV) measurements were realized after 3 and 7 weeks of NaCNO administration (Fig. 5C). Similar aPWV values were found in the two groups at both analysis times (3 weeks: Figure 2. Evolution of carbamylation in bovine aorta with age. Homocitrulline (HCit) was quantified by LC-MS/MS in (A) total aorta extracts and in (B) aorta elastin of 21 bovine samples. Nonlinear regressions (95% confidence interval; shaded zones) and linear regressions (solid lines) of HCit concentrations over time were calculated. Spearman coefficients of correlation (r) showed a significant HCit increase with age (p < 0.05 for total extracts and p < 0.001 for elastin). A last set of experiments was undergone in order to determine if the carbamylation-induced increase in elastic fiber stiffness was also observed in ApoE −/− mice, which are considered a relevant murine model of accelerated vascular aging 23 . ApoE −/− mice were subjected to the same diet as described above (1 mM NaCNO or 1 mM NaCl in drinking water for 3 weeks), and aorta cross-sections were analyzed by AFM. Similar results to those of WT mice were obtained, i.e. an increase in the YM of the elastic fibers under carbamylating conditions. Values increased from 527 ± 256 kPa in the control group to 908 ± 210 kPa in the NaCNO-fed group (+ 72%, p < 0.01), whereas no significant difference was observed in YM values measured in the inter-fiber spaces (Fig. 6A).
In contrast with WT mice, whereas no difference of aPWV values was found after 3 weeks of NaCNO diet, a significant increase was noticed (+ 20%, p < 0.01) after 7 weeks (control group: 3.5 ± 0.3 m/s vs NaCNO-fed group: 4.2 ± 0.3 m/s), suggesting that carbamylation of vascular wall proteins is associated with a progressive increase in arterial stiffness (Fig. 6B).

Discussion
Elastic fibers play a key role in the mechanical properties of vessels. However, as there is virtually no renewal of elastin over the lifetime of adult humans, it may be hypothesized that their molecular aging has a major impact on vessel function and could favor many aspects of cardiovascular diseases. Among the phenomena thought to be responsible for molecular aging, enzymatic proteolysis and non-enzymatic post-translational modifications (NEPTMs) such as glycation and carbamylation are the most important. While there is some evidence for the glycation of elastic fibers in human arteries and its subsequent role in arterial stiffening 7,11 , ours is the first study, to our knowledge, to address the effect of carbamylation. The aims of the present study were thus to determine if vascular elastic fibers were prone to carbamylation, and if so, what the consequences on their properties would be. www.nature.com/scientificreports/ To date, the relationship between carbamylation and cardiovascular diseases has been evidenced mainly from basic studies suggesting the deleterious role of carbamylated lipoproteins and from clinical trials using carbamylated proteins as biomarkers in these contexts 17,18,24 . By contrast, very few studies have focused on vascular extracellular matrix (ECM) proteins which, because of their long half-lives, are expected to be particularly susceptible to carbamylation. While we have demonstrated that type I collagen was carbamylated at several specific sites which altered its conformation and functions 25 , the case of elastin is more tricky, the number of free lysine residues being very low since most are involved in cross-links like desmosine and isodesmosine. It is estimated that elastin lysine residues represent only about 4% of total amino acids, 16% of them being accessible  www.nature.com/scientificreports/ to modifications 26 . Despite this, our results show that bovine vascular elastin is indeed carbamylated in vivo, and that this rate is of the same magnitude as that of skin elastin 22 . After having shown that bovine vascular elastin was carbamylated in vivo, we then focused on the impact of carbamylation on the properties of the elastic fibers. In a first step, carbamylation-induced changes on fibers' morphology and resistance were evaluated. No morphological changes were shown by SEM after in vitro carbamylation of bovine elastin, which may be because the experiments were performed on cross-linked molecules. While carbamylation of tropoelastin may interfere with the process of fiber formation, as observed with carbamylated collagen which exhibits a fibrillogenesis defect 25 , elastogenesis is almost nonexistent after puberty in humans and carbamylation of mature and reticulated fibers is of greater importance than its interference with elastic fiber formation. The degradation assay, performed on bovine elastin with pancreatic elastase, did not demonstrate any modification of sensitivity of carbamylated elastin to proteolysis.
A NaCNO-fed murine model was then used to study elastic fibers' status in vivo in conditions of increased carbamylation, and did not show any alteration of their morphology. However, AFM analysis of aorta wall showed that a 3-week administration of NaCNO to WT mice led to a significant increase in elastin lamellae stiffness, whereas no difference was observed in the inter-fiber spaces which mainly contain collagens and smooth muscle cells. Indeed, AFM permitted a direct assessment of the impact of carbamylation on elastic fibers' stiffness by YM measurements. This method had already been used for the measurement of stiffness of different zones within atherosclerotic lesions 27 . These data provide evidence, for the first time to our knowledge, that carbamylation of some of the available lysine residues in elastic fibers is associated with a dramatic increase in their stiffness through a molecular mechanism that remains to be defined. Important differences in YM values between elastic lamellar and inter-fiber spaces were also noted, indicating that the elastic fibers were tenfold stiffer than the interfiber spaces, which is consistent with previous observations reported for sheep and human aorta 27,28 .
No data on carbamylation of elastic fibers are available in the literature, which precludes comparison of our data with other studies. One study on glycation of elastin reported opposite effects, i.e. a decrease in stiffness by glycation 29 . However, this in vitro study was performed on elastin extracted from arteries and glycated with very high, non-physiological, concentrations of glucose (i.e. 5-10 M). The authors also used a different method for measuring stiffness, further complicating comparison. Besides, glycation and carbamylation are two distinct mechanisms which may lead to opposite effects at the molecular level. For example, while we observed no effect of carbamylation on elastin sensitivity to elastase, glycation of alpha-elastin made it more resistant to neutrophil elastase digestion 30 . An important concept to be borne in mind is the possible competition between these two reactions. Indeed, the presence of glycated elastin in human arteries 7,31 , brought face to face with our data showing carbamylation of bovine elastin, suggests it is subjected to several NEPTMs in vivo. Given the limited number of free lysine residues available, it could be expected that these reactions compete for the modification of elastin, as we have recently shown for collagen 32 , but also contribute simultaneously to the modification of elastic fibers' properties.
To determine the impact of carbamylation under pathological (i.e. pro-atherosclerotic) conditions, these experiments were repeated in an ApoE −/− murine model. The vascular wall of ApoE −/− mice differs from that of WT mice due to the lack of apolipoprotein E gene expression, thereby promoting dyslipidemia, subintimal inflammation and atherosclerosis development. ApoE −/− mice develop atherosclerotic lesions as early as 10 weeks old and constitute an excellent model of accelerated vascular aging 23 . In this study, feeding ApoE −/− mice with NaCNO resulted in an increase in elastic fibers' stiffness at the molecular level, but also globally at the vessel level as indicated by the significant increase in aPWV after 7 weeks treatment. Unexpectedly, no variation in aPWV was observed in WT mice. One hypothesis to explain this result could be a delay in the development of arterial stiffness in WT compared with ApoE −/− mice, which are more prone to vascular changes because of their genetic background. Indeed, ApoE −/− mice have been described to develop vascular stiffness (demonstrated by the increase in aPWV) more rapidly than WT mice 33,34 . Further studies remain thus to be performed to better determine the consequences of increased elastic fiber stiffness on functional properties of arteries.
Such findings may be of added value in chronic renal failure, a disease that promotes carbamylation due to hyperuremia and whose complications often include cardiovascular events: it has been demonstrated that carbamylated lipoproteins are more atherogenic 21,24 and that uremic, pro-carbamylating conditions stimulate calcification processes which also contribute to arterial stiffening 35 . Our data offer a mechanistic explanation of how carbamylation of vascular ECM proteins may contribute, with these other mechanisms, to the progression of vascular diseases in uremic patients.
In conclusion, our study has shown, for the first time, that vascular elastic fibers are prone to carbamylation and that this process is associated with an increase in their stiffness at the molecular level. These alterations may have deep consequences on the mechanical properties of the vascular wall, which suggests that carbamylation, together with other NEPTMs, are crucial factors in the etiology of cardiovascular diseases.

Methods
All the procedures were carried out in accordance with proper guidelines and regulations.
Tissue samples and animal models. Bovine  www.nature.com/scientificreports/ room at a constant ambient temperature and with a 12-h light-dark cycle. All procedures involving mice were approved by the institutional animal care committee of the University of Reims Champagne-Ardenne (registration 56) and the veterinary services of the health and the production in accordance with French government policies (APAFIS#4433-2016030911228135). This study was carried out in compliance with the ARRIVE guidelines. Cyanate-consuming model: 6-week-old mice (WT and ApoE −/− ) were randomly assigned to two groups (n = 6 each): one group received drinking water supplemented with 1 mM NaCl (i.e. control conditions), and one group received drinking water supplemented with 1 mM NaCNO (NaCNO-fed group). Water was renewed twice a week during the course of the experiments (3 or 7 weeks).
Extraction protocols. Total tissue extracts. Bovine (100 mg) or mice (15 mg) aortas were homogenized with 800 µL of 0.5 M acetic acid in Lysing Matrix D™ tubes using the FastPrep-24 System (MP Biomedicals, Illkirch-Graffenstaden, France). After 6 cycles of homogenization (40 s, 6 m/s), samples underwent 10% (m/m) pepsin digestion for 24 h at 37 °C. After 3 additional cycles of homogenization under the same conditions, samples were centrifuged at 14,000g for 5 min, and supernatants were collected and stored at − 80 °C before HCit quantification by liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS).
Elastin extraction. Bovine aorta samples (200 mg) were ground in a ball mill under liquid nitrogen cooling (Retsch Technology, Eragny sur Oise, France), and elastin was extracted as described previously 22 . Briefly, the protocol consisted in different steps combining washing procedures with buffers and solvents (1 M NaCl, pure ethanol, chloroform:methanol (2:1 vol/vol), ether, acetone) and protein cleavages with cyanogen bromide or trypsin. At each step, a volume of 1.5 mL corresponding buffer was added, samples were centrifuged at 14,000g for 2 min, and the supernatant was carefully removed.
In vitro elastin carbamylation and degradation assay. In vitro elastin carbamylation. Elastin was carbamylated by incubating 2 g of extracted bovine elastin with 100 mM NaCNO in 150 mM phosphate buffer pH 7.4 for 24 h at + 37 °C under gentle stirring. A control experiment was carried out in parallel with NaCl instead of NaCNO. Samples were then washed several times with phosphate buffer and dried at 37 °C.
In vitro degradation assay. Carbamylated (NaCNO) or control elastin (200 µg) was incubated at a concentration of 1 mg/mL in 50 mM Tris buffer pH 7.0 under gentle stirring with 0.1 U/mL porcine pancreatic elastase (Merck Chimie, Fontenay-sous-bois, France) for 8 and 18 h at + 37 °C (n = 6). The reaction was stopped by adding 0.5% (v/v, final concentration) trifluoroacetic acid (Sigma, Machelen, Belgium). Samples were centrifuged at 10,000g for 10 min at + 4 °C, and supernatants were stored at − 80 °C until analysis. The quantification of released elastin peptides used the "Elastin Assay-Fastin™" (Biocolor, INTERCHIM, Montluçon, France). For this purpose, elastin fragments contained in the supernatant were precipitated in an "Elastin Precipitating Reagent buffer". After centrifugation at 10,000g for 10 min at + 4 °C, the pellet containing elastin fragments was incubated for 1h30 with 1 mL of Fastin Dye Reagent under orbital shaking under darkness. This mixture was centrifuged (10,000g, 10 min, + 4 °C) and the pellet was solubilized in 250 μL of Dye Dissociation Reagent buffer. Absorbance was measured at 513 nm, a calibration curve being established in parallel using a standard containing a known concentration of elastin peptides.
HCit quantification. HCit was quantified in tissue samples using a previously described liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) method 22,36 . Briefly, samples were subjected to acid hydrolysis with 6 M HCl for 18 h at 110 °C and hydrolysates were twice evaporated to dryness under a nitrogen stream. Dried samples were resuspended in 100 μL 125 mM ammonium formate containing 1 μM d7-citrulline and 65 μM d8-lysine [used as internal standards (ISs)] and filtered using Uptidisc PTFE Filters (4 mm, 0.45 μm; Interchim). Ten-fold diluted hydrolysates were subjected to LC-MS/MS analysis (API4000; ABSciex) to quantify HCit, Lys or Glu. Liquid chromatography was performed using a Kinetex HILIC Column (100 × 4.6 mm, 2.6 μm; Phenomenex) with 5 mM ammonium formate (pH 2.9) as mobile phase A and 100% acetonitrile as mobile phase B. The flow rate was constant at 0.9 mL/min during all separation steps. Details parameters for LC separation and MS/MS detection have been described elsewhere 22,36 . Results were expressed as ratios to lysine content in total aorta extracts and to glutamate content in elastin extracts.
Scanning electron microscopy. Dehydrated elastin pellets obtained after drying at 37 °C for 24 h were attached to an aluminum slide using of double-sided carbon tape and covered with a conductive layer (Au-Pd, cathodic evaporator JEOL JFC-1100, 1.2 kV, 8 mA, 10 min) to favor the interactions between the sample and the electron beam. Observation was performed using a field emission electron microscope with Schottky tip (JSM-7900F prime, JEOL Europe SAS, Croissy sur Seine, France) operating with a beam of 20 kV primary energy. www.nature.com/scientificreports/ section was averaged and the degradation rate of elastic fibers was expressed as a rupture index, corresponding to the ratio of the number of breaking points to the total area of the 2-PEF signal. For the PFQNM calibration, the standard supplier protocol was applied to get quantitative measurements of the YM. First, the deflection sensitivity was calibrated before use in the buffer by carrying out indentation ramps on a clean and hard sapphire surface. Then, the cantilever spring constant was calculated before and after each experiment following the thermal tuning method. The last step was to calibrate the curvature radius of the tip by using a standard titanium tipcheck sample. This curvature radius was confirmed by performing a test measurement of the YM of a calibrated known sample. A PeakForce frequency of 0.25 kHz was used in order to maximize the contact time between the tip and the sample and the PeakForce amplitude was set to 2 µm. The distance synchronization parameter was manually and constantly adjusted over time so that the turnaway point of each force curve was exactly at the (x,y) maximum position. Images were captured with a resolution of 256 pixels per line. Once the different AFM images acquired, for the YM calculation, the force curves were extracted from chosen areas in the PFQNM images and the conventional Derjaguin-Muller-Toporov (DMT) model was used to fit the linear part of the extension curve as it was identified as the best suited model according to the tip geometry and the properties of the samples. The YM at each point of the elastic fibers or of the inter-fiber spaces was calculated using a value of the Poisson ratio of 0.5 for our samples considered as incompressible. For each condition, at least 5000 force curves were treated to get the statistical values of the YM for the elastic fibers and the inter-fiber spaces. The analyses were performed at 3 different locations in each cross-section, for a total of 9 cross-sections obtained from 3 different mice.

Measurement of murine elastic fiber stiffness by atomic force microscopy.
Evaluation of murine arterial stiffness in vivo. Measurement of aPWV in WT and ApoE −/− mice, fed or not with 1 mM NaCNO, was performed using the Doppler Flow Velocity system (Indus Instruments, Houston, Texas) 37 . Doppler signals were measured at the level of thoracic and abdominal aortas by using focused 10 MHz and 20 MHz Doppler probes, respectively. Doppler signals were recorded and the distance between the two probes measured, allowing aPWV calculation using the following formula: aPWV (m/s) = distance (m)/Δt (s), where "distance" is the distance between the two probes and Δt is the averaged time between two ejection times.
Measurement of blood pressure. Arterial pressure was measured using a non-invasive blood pressure measurement system, i.e. a tail-cuff sphygmomanometer equipped with the photoplethysmogram Visitech BP-2000 (Visitech Systems, Apex, NC). Mice were accustomed to the procedure for 5 consecutive days before measurement and were monitored weekly during the 7 weeks of the study. For blood pressure measurements, the mice were placed in a warm (37 °C) restraint device with an appropriately sized cuff around the tail. In order to minimize stress-induced blood pressure fluctuations in the animals, all measurements were taken after a 5-min adaptation period. All measurements of diastolic and systolic pressures were taken between 9:00 a.m. and 12:00 a.m., and recordings were averaged from at least 5 consecutive readings using BP-2000 analysis software. Mean arterial pressure was calculated using following formula: [(2 × diastolic) + systolic]/3. www.nature.com/scientificreports/