Renal denervation restores biomechanics of carotid arteries in a rat model of hypertension

The prevalence of hypertension increases with aging and is associated with increased arterial stiffness. Resistant hypertension is presented when drug treatments fail to regulate a sustained increased blood pressure. Given that the mechanisms between the sympathetic nervous system and the kidney play an important role in blood regulation, renal denervation (RDN) has emerged as a therapeutic potential in resistant hypertension. In this study, we investigated the effects of RDN on the biomechanical response and microstructure of elastic arteries. Common carotid arteries (CCA) were excised from 3-, 8- and 8-month-old denervated rats, and subjected to biaxial extension-inflation test. Our results showed that hypertension developed in the 8-month-old rats. The sustained elevated blood pressure resulted in arterial remodeling which was manifested as a significant stress increase in both axial and circumferential directions after 8 months. RDN had a favorable impact on CCAs with a restoration of stresses in values similar to control arteries at 3 months. After biomechanical testing, arteries were imaged under a multi-photon microscope to identify microstructural changes in extracellular matrix (ECM). Quantification of multi-photon images showed no significant alterations of the main ECM components, elastic and collagen fibers, indicating that arteries remained intact after RDN. Regardless of the experimental group, our microstructural analysis of the multi-photon images revealed that reorientation of the collagen fibers might be the main microstructural mechanism taking place during pressurization with their straightening happening during axial stretching.


Introduction
Arterial stiffness of elastic arteries depends mainly on the brous extracellular matrix (ECM) components, namely collagen and elastic bers, and their organization.Elastic and collagen bers are the two major ECM constituents in large elastic arteries and the primary load-bearing ECM components of the arterial wall.Elastic bers provide arteries with the ability to accommodate pulsatile blood ow and collagen bers primarily provide structural integrity and distensibility.Collagen has a half-life of ~ 22 days in arteries (1).In contrast, elastin is only produced in early development and childhood, has a very low turnover rate, and provides elasticity to the tissue through the lifetime of the individual (2).Hypertension is the most common cardiovascular disease affecting around 1 billion people worldwide and its prevalence increases with age (3).Hypertension has long been associated with arterial stiffening (4,5).Vascular cells sense the elevated blood pressure and respond with an abnormal metabolic activity and ECM protein production which results to an accelerated vascular stiffening (6).Remodeling of vascular ECM in hypertension is manifested with collagen deposition and elastin fragmentation which alters the collagen-to-elastin ratio and results in arterial wall thickening, while the expression of proteoglycans and glycoproteins is also upregulated (5,(7)(8)(9).Several studies have reported stiffening of the arterial wall of systemic arteries in hypertension in both clinical and animal studies (10)(11)(12)(13).Therefore, there is a strong connection between arterial stiffness and hypertension.
Resistant hypertension manifests when the conventional antihypertensive therapy fails to control the blood pressure.This persistent and prolonged high blood pressure carries a mortality risk to resistant hypertensive patients who are more likely to develop irreversible damage in end-organs such as the kidneys and brain due to cardiovascular implications (14).Several studies have reported the role of the renal sympathetic nervous system, which is comprised of both afferent and efferent nerves, as a regulatory mechanism for the blood pressure through sodium, renin and renal blood ow modulation (15)(16)(17).
Previous studies provided a wealth of evidence supporting a key role of increased sympathetic out ow in hypertension.Renal denervation (RDN) has been emerged as a potential therapeutic approach to resistant hypertension by selectively removing the renal nerves and therefore attenuating renal sympathetic out ow. Clinical trials have been conducted to study the effects of RDN in resistant hypertensive patients, although inconsistent outcomes failed to quantitatively assess the effects of the procedure (18).However, to the best of our knowledge, none of previous studies investigated the effect of RDN on microstructure and biomechanical response of elastic arteries.
This study focuses on understanding the interplay between sympathetic out ow and vascular remodeling.A rat model was used to study the remodeling and mechanical response of elastic arteries in hypertension and RDN.Common carotid arteries (CCA) were harvested from 3-, 8-and 8-month-old denervated rats and subjected to biaxial in ation-extension testing; circumferential and axial stresses as well as tangent modulus were calculated.Arterial ECM were investigated through multiphoton microscopy and microstructural parameters (waviness and orientation) of collagen and elastic bers were extracted.
The combined results from mechanical testing and multiphoton imaging allowed us to assess the structure and function of CCA in hypertension and RDN.

Animals
Male Sprague-Dawley rats aged 3-, and 8-months old were purchased from Envigo (Indianapolis, IN, USA).Rats were pairhoused prior to surgical intervention and were individually housed following survival surgery as described below, in accordance with Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines.Animals were housed in a temperature (range 68-79°F) and humidity (range 30-70%) controlled facility under a 12-h light-dark cycle and were allowed tap water and standard irradiated normal salt rodent diet (NS; Envigo Teklad, WI, Teklad Global Diet #2918, 18% protein, 5% crude fat, 5% ber, total potassium (K + ) content 0.6%, total NaCl content 0.6% [174 mEq Na + /kg]) ad libitum.In all studies rats were randomly assigned to experimental groups.All animal protocols were approved by the Institutional Animal Care and Use Committee under protocol number PROTO201800201 in accordance with the guidelines of the Boston University Chobanian & Avedisian School of Medicine and the National Institutes of Health Guide for the Care and Use of Laboratory Animals.All possible steps were taken to minimize pain and suffering, and euthanasia was conducted in accordance with approved protocols.

Surgical Procedures
Acute femoral vein, artery, and bladder cannulation Rats were anesthetized (sodium brevital, 20 mg/kg IP, supplemented with 10 mg/kg IV as required) and the left femoral vein, left femoral artery, and bladder were cannulated with PE-50 (vein and artery) or PE-240 (bladder) tubing to deliver intravenous infusions, measure arterial blood pressure, and collect urine, respectively.Rats were then gently placed in a Plexiglas holder and received an intravenous infusion of isotonic saline (20 µL/min) during a 2-h recovery period prior to experimentation for the stabilization of prior to experimentation for the stabilization of cardiovascular and renal excretory functions and to reach full consciousness.Mean arterial pressure (MAP) and heart rate (HR) were recorded continuously via the femoral artery cannula using BIOPAC data acquisition software (MP150 and AcqKnowledge 3.8.2;BIOPAC Systems) in conjunction with an external pressure transducer (P23XL; Viggo Spectramed) (19)(20)(21)(22)(23) .

Bilateral Renal Denervation (RDNX)
In 8-month-old male rats, RDNX was performed for 14 days prior to acute blood pressure measurement.In these animals, standard techniques were used to remove the in uence of both the afferent and efferent renal sympathetic nerve bers (21,(24)(25)(26).In brief, under sodium pentobarbital anesthesia (30 mg/kg IP), each kidney was exposed by a dorsal ank incision.
Using a dissecting microscope, the renal vein and artery were dissected from the surrounding fascia and all visible renal nerve bundles were removed.After dissection, the renal artery was coated with a 10% phenol solution in ethanol to ensure the destruction of any remaining renal nerve bers.In the sham denervation group, each kidney was exposed, and the renal artery and vein were visualized prior to suturing.The effectiveness of RDNX was con rmed at the end of the blood pressure measurement study by ELISA analysis of NE content in kidney tissue as per manufacturer's instructions (ELISA IB89537, IBL America, MN) (26).In intact 8 months old rats renal NE content was 832 ± 34 pg/mg and was reduced to 58 ± 18 pg/mg following renal denervation.

Cardiovascular function
After a 2-h recovery period, baseline MAP was continuously recorded in naive conscious rats via the femoral artery cannula during a 1-h blood pressure measurement period.Reported MAP values represent the average MAP during the entire 1-h period (19,27).Hypertension was developed in the 8 months group with a MAP of 136.2 ± 0.8 mmHg, which is signi cantly higher than the 3 months group with a MAP of 124.8 ± 0.8 mmHg (p < 0.05; Figure S1A).RDN attenuated hypertension in 8 months old rats with a MAP of 126.8 ± 1.2 mmHg, a signi cant decrease from the 8-month group (p < 0.05).Throughout this study, the 3 months old rats are denoted as "control" (CTL), the 8 months old rats that developed hypertension are denoted as "hypertensive" (HT) and the 8 months old rats that underwent renal denervation are denoted as "denervated" (RDN).Nine CCAs were used in each experimental group (n = 9).For some animals, both the left and right CCA were used.
Biaxial extension-in ation testing: The mechanical testing protocol was described in details in our previous study (28).Brie y, the collected arterial samples were cleaned from surrounding tissue and fat, and cannulated on a pressure myograph (110P DMT, Danish Myo Technology, Denmark) to perform the biaxial extension-in ation test.First, in vivo axial stretch ratio was estimated as the axial stretch at which the axial force remains constant during pressurization.Prior to testing, arteries were preconditioned to ensure consistent behavior during mechanical testing (29) by stretching axially to their estimated in vivo axial stretch ratio, and in ating from 0 to 140 mmHg three times.The mechanical testing protocol included a stepwise in ation from 0 to 120 mmHg (increment of 10 mmHg and a loading rate of 5 mmHg/sec) at three different axial stretch ratios: in vivo stretch ratio, and 5% above and below this ratio.Arteries were fully immersed into Krebs-Ringer biocarbonate buffer solution at room temperature (22 o C) throughout the testing.Deformed outer diameter, axial force and applied pressure were collected continuously.After the testing, the unloaded arterial dimensions were measured by imaging thin arterial rings (~ 0.5 mm in length; Figures S1B-D).Images were imported into FIJI (http:/Fiji.sc/Fiji,Ashburn, VA) and the outer and inner circumference were manually traced.
Assuming that the artery behaves as a thin-walled cylinder, the mean biaxial Cauchy stresses were calculated as: and ( where is the transmural pressure, is the wall thickness, is the deformed inner radius and is the axial force.
Subscripts and correspond to circumferential and axial direction, respectively.The arterial wall was assumed to be incompressible; thus, the deformed inner diameter, was calculated as: 2 where is the deformed outer diameter, is the undeformed inner and outer diameter and is the axial stretch ratio.

Multi-photon imaging
The underlying microstructure of the arterial ECM was investigated using a Carl Zeiss LSM 710 NLO multi-photon microscope system equipped with a tunable femtosecond IR pulse laser (Coherent Chameleon Vision-S).Samples were mounted into a custom-made tissue stretching-in ation chamber which allows the biaxial deformation of the artery while being imaged under a microscope (30).Each artery was in ated with phosphate buffered saline (PBS) using a pressure gauge while being immersed into a chamber lled with PBS at a room temperature.Samples were stretched to their in vivo stretch ratio and imaged under different pressure values (0, 30, 60, 90 and 120 mmHg).An excitation wavelength of 810 nm was used to generate two-photon excitation of uorescence (500/550 nm) from elastin and second harmonic generation (395/415 nm) from collagen (30,31).Arterial samples were imaged to a depth of ~ 50-60 µm with a 1 µm spacing and a eld of view of µm using a water immersion objective (20, NA 1.0, W Plan-Apochromat, Zeiss).Z-stacks of elastin and collagen images were acquired at every pressure level.

Straightness parameter of collagen bers
The straightness parameter, of collagen bers was calculated using CT-FIRE (32), which extracts structural parameters of ber-like elements.Pre-processing of the image includes the application of a curvelet transform-based denoising lter.Fibers are traced by applying the Euclidean distance transform to the image and to identify nucleation points for bers.Every image of the collagen z-stack was imported to CT-FIRE and was calculated for the individual traced bers in the image.A histogram of was generated for every image; all histograms of the collagen stack were combined to a three-dimensional (3-D) bar plot.
This procedure was repeated for all pressure levels and axial stretches.
Orientation of collagen and elastic bers: The acquired Z-stacks of elastic and collagen ber images were imported to FIJI (http:/Fiji.sc/Fiji,Ashburn, VA) and the ber orientation was extracted for every image in the stack using the Directionality plug-in.The program uses Fast Fourier Transform (FFT) to obtain the spatial frequencies of the image.The direction of the bers is extracted by calculating the angle and intensity of each spatial frequency in the FFT power spectra using edgedetector lters.The distribution functions of the stack were averaged and plotted for every pressure.On the other hand, to show the transmural variation in orientation for elastic bers in the adventitial and outer medial layer, the distribution functions at each imaging depth were combined into a two-dimensional (2-D) surface plot.

Histology
After mechanical testing and imaging, an arterial ring of ~ 0.5 mm in length was cut from the center of the artery, xed in 4% paraformaldehyde (PFA) and stored in 70% ethanol.Samples were then imbedded in para n and stained with Movat's Pentachrome (MOV) to identify collagen, elastin and smooth muscle cells.Histology slides were imaged using an optical microscope with a 40 objective (Olympus VS120 Automated Slide Scanner) at the Micro/Nano Imaging (MNI) Facility at the Biomedical Engineering Core Facilities, Boston University.

Statistical analysis
Average experimental data from mechanical testing are presented as mean ± standard error of the mean (SEM) for the biomechanical analysis, and as mean ± standard deviation (SD) for the multi-photon image analysis.A one-way analysis of variance (ANOVA) was used to access statistical differences between the three groups.Post hoc analysis performed using Bonferroni test.A value of p < 0.05 was considered signi cant.All statistical analyses were performed using IBM SPSS (Version 27.0).

Results
Figure 1A shows the axial force at the in vivo and ± 5% axial stretch ratio for a representative sample.When the artery was stretched at its estimated in vivo axial stretch ratio, the axial force remained constant under pressurization; when the axial stretch ratio increased 5%, the axial force also increased, whereas for a 5% decrease in axial stretch ratio, the axial force decreased with pressure.Figure 1B shows the estimated in vivo axial stretch ratio of CCAs for the three experimental groups.Hypertension had a signi cant impact on in vivo stretch ratio which increased from 1.72 ± 0.07 to 2.04 ± 0.08 (p < 0.05) for the 3-and 8-month groups, while RDN resulted in a decrease of the in vivo ratio to 1.75 ± 0.02 (p < 0.05).
The pressure-diameter and axial force-pressure data are given in Figs.1C-F along with the calculated Cauchy stresses in the circumferential and axial directions.The axial force needed to maintain CCAs at their in vivo stretch (Fig. 1C) increased with hypertension for the applied pressure range.After RDN, the axial force decreased and reached values similar to the 3-month group.On the other hand, outer diameter of CCAs during in ation seemed to be affected only due to hypertension, not RDN (Fig. 1D).Speci cally, deformed outer diameter showed a shift to the right in hypertensive arteries which remained after RDN.
The axial stress -stretch curve of the hypertensive CCAs was characterized by an upward shift, while the circumferential stress -stretch curve exhibited a leftward shift when compared to control group (Fig. 1E and 1F).After RDN, both circumferential and axial stress curves restored to values similar to the control arteries.
Figure 2A shows the progression of axial and circumferential stresses for the three groups as a function of pressure.In all cases except control arteries at 120 mmHg, axial stress was signi cantly higher than the circumferential one in all groups (p < 0.05).Changes between groups started to appear at 60 mmHg.At 90 mmHg, both axial and circumferential stresses decreased after RDN, a trend that continued at 120 mmHg as well.
Figure 2B-D shows the biaxial stretch ratios and stresses in the circumferential and axial directions, and the circumferential tangent modulus at MAP.During pressurization, the axial stretch remains at in vivo stretch ratio while the circumferential stretch increases.In the control group, the deformation of the arterial wall reaches a close to equi-biaxial stretch state at MAP, i.e., the circumferential and axial stretches are almost the same.However, stretching appeared to be non equi-biaxial in HT and RDN groups with a signi cantly higher in vivo than circumferential stretch ratio (p < 0.05; Fig. 2B).Hypertension resulted in a signi cantly higher stress level from 84.6 ± 8. The structural integrity of the arteries was examined by histology and multi-photon imaging.Histology images of the CCAs are shown in Fig. 3.No signi cant alterations of the main arterial ECM components, elastic and collagen bers, were observed with HT or RDN.A qualitative analysis of histology images did not show visible disruptions of the brous ECM components with HT and after RDN.
Representative multi-photon images of the adventitial collagen bers of CCAs for the three experimental groups are given in Fig. 4. CCAs were axially stretched to their estimated in vivo stretch ratio and in ated at four different pressure values (0, 30, 60, 90 and 120 mmHg).Thicker collagen bundles were observed in all groups, and they were more visible at higher pressures while the ber bundles are straightened.The corresponding 3D histograms of the straightness parameter, of collagen bers are given in Fig. 5 as a function of the imaging depth and pressure.In these gures, corresponds to straight bers whereas corresponds to undulated bers.Overall, a subtle straightening of the bers with pressurization was observed.Histograms of followed the same trend across the three groups with no visible changes within the pressure range.
To further understand the deformation of collagen bers, multi-photon images of CCAs were taken as a function of axial stretch at zero pressure.Representative images and the corresponding histograms of are given in Figs. 6 and 7, respectively.Undulated collagen bers are mainly present at lower axial stretches, while an increase of axial stretch ratio above 1.4 resulted in gradual straightening.This behavior was re ected in Fig. 7 with right-skewed histograms which showed smaller range and higher frequency of straighter bers ( ) for axial stretch ratios greater than 1.4.The histograms of for the HT and RDN groups are given in the supplementary material (Figures S2 and S3).
Since histograms followed a skewed distribution and there were no major changes of as a function of imaging depth (Figs. 5 and 7), the median value was extracted from every histogram.Line plots of were generated by calculating the mean value of the median values for each collagen stack and plotted as a function of the pressure and axial stretch in Fig. 8A and 8B, respectively.As the arteries were stretched in the axial direction, collagen bers showed an increasing trend of straightening which plateaued above 1.8 stretch (Fig. 8A).On the other hand, when arteries were xed at their in vivo stretch ratio remain almost constant with pressurization with no signi cant variations, regardless of the experimental group (Fig. 8B Figures 9 and 10 show collagen bers orientation as a function axial stretch ratio and pressure, respectively.Since there were no differences in the collagen ber orientation with axial stretch ratio between the three groups, only histograms for the control arteries are shown in Fig. 9.The histograms of collagen orientation for the HT and RDN groups are given in the supplementary material (Figure S4 and S5, for HT and RDN, respectively).The orientation of collagen bers seemed to be varied mostly ±45 o from the axial direction at low stretch ratios and aligned towards the axial direction with further axial stretching (Fig. 9).On the other hand, pressurization of the arteries resulted in a shifting of orientation from a bell-shaped histogram around 0 o (axial direction) towards a multimodal histogram with multiples peaks which ranged from − 45 o to 45 o (Fig. 10).For the denervated arteries, orientation started as a bimodal histogram with two peaks around axial direction, while those two peaks became more distinct and apart at higher pressures.
Figure 11 shows representative multi-photon images of the elastic bers under a pressure of 90 mmHg.The three columns show images at three different transmural sites: adventitia (left column), interface between adventitia and media (denoted as A/M interface; middle column), and outer media (right column).Sparse elastic bers were present in adventitia in all groups, with a disperse orientation.At the interface, elastic bers started to form a denser network.In the outer layers of media, elastic bers showed a preferred orientation towards the circumferential direction.Representative orientation of the elastic bers is given in Fig. 12 as a function of the imaging depth and pressure (60 mmHg, 90 mmHg and 120 mmHg).Adventitial elastic bers in CCA from the control group oriented around the axial direction throughout the applied pressure.Elastic bers of the hypertensive CCA were distributed at ± 45 o for the rst adventitial layers and then appeared to change their orientation to the axial direction.After RDN, adventitial elastic bers were randomly distributed from 45 o to -90 o .An abrupt change of the elastic bers orientation from axial to circumferential was observed in all samples (red line).After careful examination of this transitional area in the multi-photon stack of images, it was identi ed as the interface between adventitia and media.
To further understand the differences in elastic ber distributions in the arterial wall, the ratio of circumferential-to-axial ( ) oriented elastic bers was obtained and given in Fig. 13.The top row shows the for three different pressures (60, 90 and 120 mmHg) and the bottom row shows the for the three groups.Signi cance bars between media and adventitia are shown only in the top row gures.As it was noticed before, the elastic bers show a preferred orientation towards the axial direction in adventitia with a , and change to preferred circumferential distribution in the outer media.At the physiological pressure of 120 mmHg, in media decreased signi cantly from 0.99 ± 0.28 to 0.82 ± 0.2 and 0.89 ± 0.39 for the HT and RDN when compared to control group (p < 0.05) indicating an orientation shifting towards the axial direction.Differences of the were observed as a function of pressure for the control and HT groups.Speci cally, ratio increased signi cantly in the media for the control group from 0.83 ± 0.12 at 60 mmHg and 0.87 ± 0.2 at 90 mmHg to 0.99 ± 0.28 at 120 mmHg, and in the adventitia from 0.56 ± 0.14 at 60 mmHg and 0.57 ± 0.16 at 90 mmHg to 0.76 ± 0.28 at 120 mmHg (p < 0.05; Fig. 13B).For the HT group, ratio increased signi cantly only in the adventitia from 0.57 ± 0.17 at 60 mmHg to 0.66 ± 0.2 at 120 mmHg (p < 0.05; Fig. 13B).

Discussion
Sympathetic nervous system and the kidney play an important role in regulating blood pressure.In this study, we investigated the biomechanical response of CCAs in hypertension and after RDN using a rat model.Our results suggest a sustained elevated blood pressure resulted in arterial remodeling manifested by signi cantly increased circumferential and axial stresses in carotid arteries.RDN of 8 months old rats effectively reversed the biomechanical response of carotid arteries with no obvious structural remodeling in the ECM.Mean blood pressure and biaxial stresses were restored to normotensive values, with signi cant changes in stresses, in vivo stretch ratio, and axial force.Microstructural imaging and quanti cation revealed intact ECM in the denervated CCAs.
In the present study, mechanical properties of carotid arteries from the hypertensive group were impacted with increased biaxial stresses and tangent modulus (Fig. 2C and 2D).The positive effect of RDN was seen not only on hemodynamics as a decrease in MAP, but also on restoration of arterial biomechanical response.After RDN, passive wall biaxial stresses and A arterial stiffness were restored to similar ranges for the 3-month group (Fig. 2C and 2D).Clinical studies have shown a decrease in blood pressure, as well as reverse arterial stiffening following RDN measured by pulse wave velocity (PWV) (33)(34)(35).In these studies, PWV measurement at 1, 3, 6 and 12 months after RDN showed a signi cant decrease that started one day after renal denervation, while the PWV reduction sustained after a 12-month follow-up (33,35).Given that the early changes in PWV after RDN were independent of blood pressure, the authors suggested that the decrease in PWV might be associated with a modi cation of sympathetic tone (33).
One interesting nding in this study is the effect of hypertension and RDN in the axial biomechanics.Hypertension resulted in increased in vivo stretch ratio and axial force, while RDN restored the values of both metrics (Figs.1B and 1C).Remodeling in axial direction of an elastic artery is one of the rst structural changes detected during the early stages of postnatal growth and development, manifested by an increased axial stretching due to a body length growth, and a nonuniform axial wall thickening (36,37).Arteries remain prestretched in the axial direction with maturity, and exhibit minimal axial deformation during cardiac cycle, which is energetically favorable and facilitates the propagation of pulse pressure (38,39).Organ culture studies reported that arteries adapted to their biomechanical environment by increasing their length when subjected to sustained increased axial stretch or loading (40)(41)(42).Morphological changes in axial con guration were noticed with aging as well as hypertension, and include an elongation of arterial segments due to a deposition of stressed collagen which resulted in a decreased in vivo stretch ratio; however, studies on axial force resulted in contradictory ndings (43)(44)(45)(46)(47)(48)). Nevertheless, limited information is available on axial remodeling and biomechanical response of elastic arteries due to the lack of methods to accurately measure and obtain in vivo values of axial prestretch and stress (48).The arteries are unlikely to grow in the axial direction from the 3-to 8 month's groups; however, the axial force needed to maintain arteries in their in vivo length increased (Fig. 2C).Another interesting nding of this study is the difference in biaxial stress/stretch state of the arteries under physiological loading among the three groups.At MAP, our results indicate that the 3-month control group were subjected to a nearly equibiaxal stretching, i.e., the in vivo axial stretch and the circumferential stretch are the same, while the 8-month hypertensive and RDN arteries were subjected to a non-equibiaxial stretching with a signi cantly higher axial than circumferential stress/stretch (Fig. 2B).The biaxial stress state will likely play an important role in regulating cellular behavior and ECM structural organization in physiology and diseases.Understanding the altered biaxial stress/stretch state in the 8-months groups may potentially help us to explain the biomechanical changes in the axial direction.
While thickening of the arterial wall in hypertension is often seen as a compensatory mechanism to restore the increased wall stresses towards normal, homeostatic values (49), in our study, the sustained elevated blood pressure led to an enlargement of the arteries but with no signi cant changes in total thickness (Figure S1).The lack of signi cant changes in wall thickness in the present study (Figure S1D) could be explained by the time course of arterial remodeling.Previous studies have shown a temporal progression of changes in matrix synthesis and mechanical properties in hypertension.Speci cally, thickness of the arterial wall showed an increasing trend which peaked 2 weeks after inducing hypertension and started decreasing at 42 days, while stiffness increased during the rst week and reached a plateau after (45,50).These ndings correlate well with an immunohistochemical analysis of porcine aorta during hypertension which showed increased collagen production at 2 weeks of hypertension; however, collagen degradation resulted in a decrease in collagen content at 6 and 8 weeks (51).It should be noted that in these studies, hypertension was induced by constricted the aorta or renal arteries, and changes were monitored for a speci c time window after hypertension, ranging from 0 days to 16 weeks, whereas in our study we used 3 and 8 months old hypertensive rats with no information about the exact starting time point for hypertension.
The structural organization of the brous ECM components, i.e. collagen and elastic bers, has long been shown to play an important role in determining the mechanical behavior of elastic arteries.Studies investigating the relationship between the arterial microstructure and its mechanical response showed that the arterial wall exhibits a nonlinear and anisotropic behavior under loading with the elastic bers being the primary source of elasticity, and collagen ber waviness and gradual recruitment contributing to nonlinearity (52)(53)(54)(55).Previous studies have further investigated the layer-speci c microstructure and its correlation to arterial mechanical properties (31,56,57).These studies showed sequential engagement of elastic and collagen bers under biaxial loading.In our study, structural information (waviness and orientation) of collagen and elastic bers in adventitial and outer media was extracted during biaxial in ation and extension.Quanti cation of adventitial collagen bers microstructure showed no change in straightness parameter during pressurization from 0 to 120 mmHg while arteries were deformed to their in vivo stretch ratio (Fig. 8B); however, incremental stretching in the axial direction revealed a gradual straightening of collagen bers which reached a plateau for axial ratios greater than 1.6 (Fig. 8A).Given that the in vivo stretch ratio in the present study is higher than 1.7 in all groups, our results show that the unfolding of the adventitial collagen bers takes place during the axial stretching of the arteries, and thus possibly contribute to macroscopic deformation in the axial direction.Interestingly, pressurization resulted in a reorientation of adventitial collagen bers from the axial direction to more dispersed or two ber families in the diagonal direction (± 45 o ; Fig. 10).These ndings add to the existing knowledge about collagen orientation under loading and indicate that collagen bers straighten during axial stretching, whereas reorientation may be the main microstructural mechanism that takes place during pressurization.
Quanti cation of multi-photon images showed no signi cant changes in straightness parameter and orientation between control and hypertensive arteries within a physiological pressure range (Figs.8B and 10).Despite that the microstructural quanti cations from this study cannot be directly related to the increase in stresses and tangent modulus in the hypertensive group, compromised quality of the existing ECM components remains to be a possibility.Murtada et al. (58) have shown that elastic bers in mice thoracic aorta remained intact with no increase in the amount of elastin after angiotensin II induced hypertension.However, elastic bers were less effective and appeared to have a reduced ability to store elastic energy.Besides elastic bers, alterations in collagen subtypes proportions were also associated with arterial stiffness, regardless of changes in collagen amount.An altered collagen type-I turnover was found to be correlated with arterial stiffness in hypertensive subjects with an increased collagen type-I degradation (59).In addition, arterial stiffness was correlated with an upregulation of collagen type-V in arteries from hypertensive rats while the amount of collagen type-I decreased (60).The effect of RDN on ECM remodeling has been investigated in a few studies but its connection to arterial stiffness remains largely unknown.Studies performed histology analysis of arterial samples after RDN and reported a decreased expression of Angiotensin II receptor (61), reduced vascular brosis (62) and inhibition of NOX activity which indicated an attenuated oxidative stress (63), all of which have been known to be associated with arterial stiffness (64-66).Our multi-photon and histology images of collagen and elastic bers suggested that RDN resulted in no structural changes, leaving arterial tissues seemed intact at the microscopic level (Figs. 3, 4 and 11).The reversed biomechanical response after RDN (stresses and stiffness) were likely related to alterations in ECM composition of arterial wall resulted from combined neurological and hemodynamical modi cations.Future studies focusing on ber mechanics and biochemical compositional analysis of ECM constituents may yield insightful ndings that advance our understandings on vascular mechanics and microstructural composition.

Limitations
Multi-photon microscopy has emerged as a promising imaging technique since it requires minimal tissue preparation, however the imaging depth in intact arterial tissue is restricted to ~ 50-60 µm due to light scattering (30,31,67).Given that the adventitia thickness is usually 1/3 of the total thickness of the arterial wall (30,68,69), we were able to image the adventitia and only the outer layers of media.Generalizations of our results to the entire medial layer should be made cautiously.In addition, quanti cation of collagen and elastic ber structure (waviness and orientation) was performed in the two-dimensional imaging plane.However, ECM bers usually cross planes and thus a three dimensional (3D) reconstruction of the collagen bers is more representative than the (2D) analysis that was followed in this study (70).The MAP in all groups was higher than 120 mmHg, which was the highest applied pressure during imaging.However, since arteries already reach stiffening at 120 mmHg, we do not expect much noticeable structural changes at higher pressures.Moreover, vascular smooth muscle cells (VSMC) play an important role in regulating vascular properties during physiological and pathophysiological conditions.Speci cally in hypertension, VSMC differentiate their phenotype from a contractile to a proliferative and synthetic, which affect the arterial response to loading (71).However, the main objective of this study was to investigate the passive mechanical properties and microstructure of the arterial wall in hypertension and RDN, hence we did not take into consideration the contribution of VSMC.While it has been reported that the renal nerves fully reinnervate            Representative orientation maps of the elastic bers at 60, 90 and 120 mmHg as a function of the imaging depth for the control (CTL), hypertensive (HT) and denervated (RDN) group.Here 0 o and ± 90 o correspond to the axial and circumferential direction, respectively, and 0 μm depth corresponds to the outer surface of the adventitia.Scale bars represent the amount (%) of elastic bers oriented in each direction.In all gures adventitia-media interface was visually identi ed and marked with a red line.

Figures Figure 1 A
Figures

Figure 2 A
Figure 2

Figure 3 Representative
Figure 3

Figure 5 Three
Figure 5

Figure 7 Three
Figure 7

Figure 11 Representative
Figure 11