Biaxial deformation of collagen and elastin fibers in coronary adventitia

Abstract

The microstructural deformation-mechanical loading relation of the blood vessel wall is essential for understanding the overall mechanical behavior of vascular tissue in health and disease. We employed simultaneous mechanical loading-imaging to quantify in situ deformation of individual collagen and elastin fibers on unstained fresh porcine coronary adventitia under a combination of vessel inflation and axial extension loading. Specifically, the specimens were imaged under biaxial loads to study microscopic deformation-loading behavior of fibers in conjunction with morphometric measurements at the zero-stress state. Collagen fibers largely orientate in the longitudinal direction, while elastin fibers have major orientation parallel to collagen, but with additional orientation angles in each sublayer of the adventitia. With an increase of biaxial load, collagen fibers were uniformly stretched to the loading direction, while elastin fibers gradually formed a network in sublayers, which strongly depended on the initial arrangement. The waviness of collagen decreased more rapidly at a circumferential stretch ratio of λ_θ = 1.0 than at λ_θ = 1.5, while most collagen became straightened at λ_θ = 1.8. These microscopic deformations imply that the longitudinally stiffer adventitia is a direct result of initial fiber alignment, and the overall mechanical behavior of the tissue is highly dependent on the corresponding microscopic deformation of fibers. The microstructural deformation-loading relation will serve as a foundation for micromechanical models of the vessel wall.

vascular tissue is a nonhomogeneous material, consisting of elastin and collagen fibers, smooth muscle cells, and ground substance with different load-bearing contributions (33, 34). The dense and tangled collagen and elastin fibers are the major mechanical components (33), while the smooth muscle cells and ground substance provide negligible support for the passive properties of blood vessels (6, 8, 34, 38). Specifically, elastin fibers sustain the major stresses at low strain level and reflect a linear deformation, while collagen fibers are recruited at large extension of the tissue and are responsible for the nonlinear behavior of vascular tissue (3, 34, 44). Moreover, the material properties, geometries, spatial distribution, and deformation of each component of the vascular tissue are intimately related and constitute the overall mechanical response to load (3, 16, 17, 26, 34, 39, 44).

The mechanical properties of individual layers (intima-media and adventitia) of blood vessels have been studied by both experimental (15, 31, 32) and modeling methods (18, 22, 42) to distinguish layer heterogeneity of vascular tissues. It is generally accepted that the media bears the majority of the load at physiological state, while the adventitia functions to protect the vessel from overstretch and rupture at high pressures (18, 20, 31). Although this hypothesis accounts for the circumferential mechanical behavior of blood vessels, it does not encompass the longitudinal direction (31). Previous studies showed that coronary adventitia is stiffer than the media in the longitudinal direction, while the media takes up most of loads in the circumferential direction under physiological conditions (31, 37). The microstructural basis for these observations, however, has yet to be elucidated. Furthermore, several studies have shown that axial strain level of blood vessels is highly associated with extracellular matrix reorganization and cell proliferation (23, 27, 40). Along these lines, it has been shown that axial stretch primarily affects axial remodeling, while luminal pressure largely affects circumferential remodeling (27). Therefore, it is important to resolve the deformation pattern of single fibers under biaxial loads to better understand vascular tissue function. To our knowledge, there are no quantitative data of in situ deformation of individual fibers of coronary arteries under biaxial loads.

The morphometric studies of vascular tissues have developed widely in the past decade with the emergence of noninvasive, nonlinear optical microscopy techniques (1–3, 7, 13, 25, 44, 45). These optical tools provide the specific geometries of microstructure, such as orientation, dimension, and distribution of fibers, which are essential to explain the overall mechanical behaviors and to predict the mechanical properties of vessels. The direct studies of microstructural deformation which reflect the local environment of vascular tissue under physiological state, however, have been generally limited due to difficulties in implementations of simultaneous mechanical loading-imaging of fresh, unfixed, and unstained tissues at the microscopic level. Our laboratory recently quantified the geometrical parameters (orientation, dimension, and waviness) of elastin and collagen fibers under in situ deformation using a simultaneous mechanical loading-imaging technique to provide the relations between fiber geometrical parameters and vessel distension pressure (3). Although this work revealed the layered structure of inner coronary adventitia and the loading-structure relation of collagen and elastin fibers, the loading was uniaxial (vessel distension) and did not shed light on the axial deformation of individual fibers.

The goal of the present study is to understand the biaxial (circumferential and axial stretch) microscopic deformation of individual fibers and to relate the microscopic structure to tissue-level response of coronary artery. Specifically, we used a simultaneous mechanical loading-imaging technique to quantify ex vivo deformation of individual fiber (elastin and collagen) of fresh, unstained coronary adventitia under both distention and axial stretch. The microstructural deformation of the fibers was related to the macroscopic mechanical response of the adventitia. This study clarifies the overall mechanical behaviors of coronary adventitia based on the microstructure under physiological loadings and provides seminal data for vascular tissue mechanics.

MATERIALS AND METHODS

Sample Preparation and Mechanical Testing

Ten hearts (average weight of 209 ± 34 g) of healthy pigs were harvested at a local slaughterhouse and transported immediately to the laboratory in 4°C physiological solution (0.9% NaCl). The left anterior descending arteries were dissected away from their emergence at the aortic ostia, and the adjacent tissue around the segments was dissected carefully in the physiological solution. The arteries were inverted, and the exposed intima-media layers were then peeled off carefully without damaging the adventitia, since an external elastic lamina structurally separates the media and adventitia layers (33). A straight segment of adventitia was inverted back and cut ∼2 cm in length and ligated at every branch bifurcation to prevent leakage.

Five segments were tested on an established motor-controlled biaxial machine (21, 31) to determine the mechanical response of the adventitia. Three axial stretch ratios (λ_z = L/L₀, with initial length L₀ at the no-load state as the reference, and L is length of the loaded vessel) were considered: λ_z = 1.0, 1.3, and 1.5, and the transmural pressure was gradually varied from 0 to 200 mmHg at each λ_z. Before the mechanical test, the segment was first preconditioned at various biaxial loadings several times to obtain reproducible mechanical data. After the mechanical tests, a ∼1-mm-long ring was cut from the middle of the segment to measure the no-load state and zero-stress state (ZSS) of the adventitia (12, 31). The ring was placed in a physiological solution and initially photographed in the no-load state. The ring was then cut radially with scissors that caused them to open into sectors and gradually approach a constant opening angle to reveal the ZSS. The cross section of each sector was photographed 30 min after the radial cut. An image processing program (ImageJ) was used for morphological measurements of inner, outer circumference, and cross-sectional area of the adventitia from the images.

The adventitia was assumed to be a thin cylindrical shell with uniform deformation through the wall, since the wall thickness (0.17 ± 0.07 mm) was ∼7% of the outer diameter (2.41 ± 0.43 mm). The circumferential Cauchy stress (i.e., wall tension as given by Laplace's law divided by wall thickness) of the adventitia was determined by σ_θθ = Pr_i/h, where P is distension pressure, r_i = $\sqrt{r_o^2-A_o/\pi {\lambda }_z}$ is the inner radius in loaded state, r_o is the outer radius in the loaded state, and A_o is the wall area in a no-load state. The axial Cauchy stress was computed by σ_zz = Pr_i²/h(r_o + r_i) + F/π(r_o² − r_i²), where h = r_o − r_i is the wall thickness in loaded state, and F is the axial force.

Multiphoton Microscopy Imaging of Unfixed, Unstained Vessels

The other five specimens were used for multiphoton microscopy (MPM) studies; i.e., to conduct simultaneous mechanical loading-imaging. MPM imaging was performed using a combined second harmonic generation (SHG) and two-photon excitation fluorescence (TPEF) modalities at room temperature (20°C) (3). All images were obtained for an average excitation power of 40 mW with 830-nm excitation wavelength, and using band-pass emission filters at the SHG (405/40 nm) and the TPEF (520/40 nm) wavelengths. The fluorescent microsphere (excitation wavelength of 540 nm and emission wavelength of 584 nm), which was used to track the scan area of a specimen, was also two-photon excited by 830-nm excitation, and only a small portion of their strong signal passed through the 520/40-nm band-pass filter and appeared in TPEF images. The field of view was 120 × 120 μm², and each acquired image was integrated over two frames to improve the signal-to-noise ratio.

The coronary adventitia consists of layered inner as well as thick and loose outer adventitia (3). The latter is difficult to completely peel off of the vessel. Although the penetration depth of MPM is 100 μm beneath the tissue surface, the images become too blurred to distinguish individual fibers and their deformation beyond ∼50-μm depth of the specimens (2). Hence, we focused on the layered inner adventitia (with average thickness of 65 ± 5 μm) and its microscopic deformation under mechanical load. The external adventitia, with fewer elastin and looser collagen fibers, prevents the overdistension of the vessel and connects the vessel with the surrounding tissue (3). We thus scanned the inner adventitia from the top surface of the inverted segment. Since the adventitia is thin (thickness to diameter ratio of 0.07 ± 0.01), the strain is nearly uniform through the wall, and the inversion process induces very little strain or stress in the adventitia segment. The specimen was scanned from the surface toward the inside through adventitia wall (z-direction). The step size of Z-stack was typically set at 0.5 μm, and the scanning depth was limited to 40 μm in thickness of adventitia (totally 80 slices). The scan time for an entire z-stack was <5 min. The total scan time for a set of loading conditions was <2 h.

Simultaneous Biaxial Mechanical Loading-Imaging Protocol

A segment of the adventitia was mounted on a custom-made chamber device, consisting of an organ bath chamber, an elastic balloon-tip catheter (Catheter Research, Indianapolis, IN) and two micromanipulators (World Precision Instrument, Sarasota, FL) (3). Before the mechanical test, a ∼1-mm-long ring was cut from the segment to measure the no-load state and ZSS. The segment was then immersed in a physiological solution for the duration of the study and axially stretched by one of the micromanipulators for a given scan. The balloon-tip catheter was inserted into the lumen of vessel and inflated by a syringe pump, which was connected to a second micromanipulator. We considered circumferential stretch ratio λ_θ (λ_θ = c/C, where c refers to the midwall circumference of the loaded vessel, and C refers to the corresponding midwall circumference at ZSS, as calculated by the mean of outer and inner circumferences, respectively) as the control variable instead of pressure, since the pressure inside the balloon may be different from the luminal pressure of the adventitia, given the elastic balloon may take up some of the load. The specimen was distended to the desired λ_θ by inflating the balloon. A charge-coupled device (CCD) camera with a macro lens (f = 100 mm) was used to record the diameter and length of specimen, and the corresponding circumferential distension and axial extension were determined by λ_θ and λ_z, respectively. A number of 10-μm-diameter fluorescent microspheres were randomly dispersed on the outer surface of the specimen to track the scan area under various mechanical loads. Specifically, areas with microspheres at edges of the field of view were selected (white box in Fig. 1A), and images with lower magnification (covering an area of 240 × 240 μm²) were captured first to ensure the scan area was focused on the middle of microspheres (Fig. 1, A and B). The TPEF and SHG images with higher magnification (covering an area of 120 × 120 μm²) were then collected simultaneously (Fig. 1, C and D) (details are given in Ref. 2). As the load was varied, the specimen was axially stretched or inflated slowly to allow simultaneous adjustment of the x- and y-position of the microscope stage and the z-position of the objective to continuously follow the same scan area throughout the loading protocol.

Fig. 1.

Simultaneous biaxial mechanical loading-imaging. Fluorescent microspheres were used to track the scan area and deformation of individual fibers. Two-photon excitation fluorescence (TPEF) image with lower magnification (240 × 240 μm²) was first captured, and an area with fluorescent microspheres at edges (white arrows) was tracked at elevated loads λ_z = 1.0 (A) to λ_z = 1.3 (B), where λ_z is axial stretch ratios. C and D: TPEF images with 120 × 120 μm² magnification were collected at corresponding loads, and a representative sublayer was selected based on remarkable structural features (white circles). Elastin fibers were carefully selected and numbered to ensure the same fibers were followed at elevated loads. X denotes axial direction, and Y denotes circumferential direction in all images.

The specimen was inflated to three λ_θ values: 1) λ_θ ≈ 1.0; i.e., no-distension load state (0 mmHg); 2) λ_θ = 1.5 (∼90 mmHg) for pig coronary arteries (21, 22, 29, 42); and 3) λ_θ = 1.8, corresponds to pressure-overload (>160 mmHg) (21, 22, 29, 42). At each distension load, the following six axial stretch loads were considered to investigate biaxial loading on fibers: λ_z = 1.0, 1.1, 1.2, 1.3, 1.4, and 1.5. At each load, in situ MPM images of both collagen and elastin fibers of the adventitia specimen were acquired. Immediately after measurements, 2- to 3-mm-long rings were cut from the segment and photographed as the no-load state. The ring was then cut open radially to release residual stress and strain as the ZSS. The specimen slice was placed in a glass slide, immersed in a water-drop (no cover slip), and then scanned by MPM to obtain morphometric data of collagen and elastin fibers at ZSS.

Statistical Analysis of SHG and TPEF Images

Morphological data of fibers at ZSS.

The elastin and collagen fibers formed multiple concentric densely packed sheets in the inner adventitia, while collagen fibers tended to distribute randomly at the outer adventitia with few elastin fibers (3). The SHG/TPEF images of ∼40-μm-depth wall were measured to obtain accurate ZSS morphological data of the inner adventitia, which occupied 30–40% of the total adventitial thickness. There usually exists about five sublayers in this depth, beyond which the SHG/TPEF images become indistinct. Specifically, a representative image of each sublayer was selected and then statistically analyzed.

The selected images were subsequently processed, and three geometric parameters of fibers: orientation angle, waviness, and width were measured (3). The orientation angle θ was determined by the angle between the center line of the fiber and circumferential direction of specimen θ (θ is 0° or 180° in circumferential direction and 90° in axial direction: 0° ≤ θ ≤ 180°). The waviness was determined by the ratio of total arc length to end-to-end straight length of fiber, while the width of single fibril bundle was measured directly in images. The width was measured to quantify the layered heterogeneity of inner adventitia. At least 20 individual fibers were drawn, and 3 measurements were made in each selected image. Each geometrical parameter was averaged over three measurements, and all measurements were made by two individuals. The statistical distributions of the parameters at the ZSS were analyzed based on five samples. Specifically, the population of a geometric parameter was calculated in each sublayer, and the populations of all sublayers (a total of 25 sublayers for 5 specimens) were grouped together to determine the distribution pattern of the parameter (Fig. 2, A and B). The parameter-layer relation was provided to elucidate transmural features of fibers, especially for fiber width. The mean width was obtained by averaging over five specimens at each sublayer (Fig. 2C).

Fig. 2.

Morphometric data of elastin and collagen fibers at ZSS (zero-stress state). A: the orientation distribution of collagen and elastin fibers of inner adventitia. B: the bell-shaped distribution of the collagen fibers waviness. C: layer-to-layer heterogeneity of fiber width. Data point presents averaged fiber width over 5 specimens, and error bar indicates standard error (SE, which is used in all figures).

In situ deformation of individual fibers.

The scan area of each specimen was marked by fluorescent microspheres, and a representative sublayer was selected based on fiber features in MPM images (white circles in Fig. 1, C and D), to ensure that the same fibers were tracked during biaxial deformation (fibers 1–4 in Fig. 1, C and D). Similar to ZSS, at least 20 distinct fibers were selected in the tracked sublayer where fibers seldom move out of plane at higher loads. As shown in Fig. 1, C and D, the deformation of individual fibers can be continuously tracked in MPM images. The mean of geometrical parameter (orientation angle and waviness) of fibers in a selected sublayer of each specimen was considered as an independent statistical sample, and statistical deformation-loading relations of fibers were obtained. The results were expressed as means ± SD (standard deviation), and the significance of the difference between the parameter under various mechanical loading was evaluated by a one-way ANOVA test. The results were considered statistically different when P < 0.05.

RESULTS

Only samples of high-quality images were considered for measurements (e.g., typically 5 from 7 samples were high quality where fibers were clearly visible) and the interoperator measurement difference was found to be <10%. The measurements below summarize the statistical data obtained over multiple optical sections and fibers.

Morphometry of Coronary Adventitia at ZSS

A layered structure of adventitia was observed at the ZSS, where orientation and width of fibers show significant transmural variation. In each sublayer, the orientation angles of collagen fibers followed a normal distribution, of which the mean (referred to as the principal direction) varied from layer to layer, while the main orientation angles of elastin fibers were parallel with collagen fibers but with secondary principle directions (Figs. 1 and 5). The orientation angles of fibers varied randomly in subsequent layers, and hence the orientation angles of fibers in sublayers were grouped together to present overall fiber arrangement (Fig. 2A). The overall orientation of collagen fibers followed a bimodal distribution with two peaks, where the first lower peak was nearly in the circumferential direction with mean 22.7 ± 12.6° and the second higher peak was nearly in the axial direction with mean 110.4 ± 28.0°. The overall orientation distribution of elastin fibers, having multiple directions in a sublayer, also followed a bimodal distribution similar to collagen, with the first peak occurred at 20.7 ± 13.1° and the second peak at 115.2 ± 31.7°.

Fig. 5.

TPEF images of elastin fibers (collected simultaneously with SHG images shown in Fig. 4) under different biaxial mechanical loading. A–C: elastin at λ_θ = 1.0 and λ_z = 1.0, 1.3, and 1.5, respectively. D–F: elastin at λ_θ = 1.5 and λ_z = 1.0, 1.3, and 1.5, respectively. G–I: elastin at λ_θ = 1.8 and λ_z = 1.0, 1.3, and 1.5, respectively. X denotes axial direction, and Y denotes circumferential direction in all images.

There was no significant transmural change for the fiber waviness. The waviness of the collagen fibril bundle of inner coronary adventitia followed a bell-shaped distribution with mean of 1.30 ± 0.11, while the waviness of 70% of the population of rod-like elastin fibers was <1.02, and the waviness of 20% of the population was between 1.05 and 1.1 (Fig. 2B). The dimension of collagen fibers increased significantly toward the exterior of the adventitia, whereas relatively thinner elastin fibers were found throughout the layers (Fig. 2C).

Macroscopic Mechanical Responses of Coronary Adventitia Under Biaxial Loads

At lower axial strain, the toe part of stress component of the vessel wall ended at a higher λ_θ, as shown in Fig. 3, A and B, due to initial collagen fiber orientation. Specifically, the critical λ_θ, beyond which circumferential and axial stresses increased suddenly, was about λ_θ ≈ 1.7 at no-extension state λ_z = 1.0. This was reduced to λ_θ ≈ 1.6 at elevated axial load λ_z = 1.3, and further reduced to λ_θ ≈ 1.5 at λ_z = 1.5. The axial stress increased significantly with increase of the axial load. There is a statistically significant difference between the first (λ_z = 1.0) and third groups of axial stresses (λ_z = 1.5) (P < 0.05), and between the second (λ_z = 1.3) and third groups (P < 0.05). There was no statistically significant difference between the first and second group (P = 0.129), as shown in Fig. 3B.

Fig. 3.

The macroscopic stresses of coronary adventitia. A and B: circumferential Cauchy stresses varying with circumferential stretch ratio λ_θ at λ_z = 1.0, λ_z = 1.3, and λ_z = 1.5. There are not statistically significant differences between three groups of circumferential stresses (P = 0.540, P = 0.848, and P = 0.431, respectively). B: axial Cauchy stresses varying with λ_θ at λ_z = 1.0, λ_z = 1.3, and λ_z = 1.5. There is a statistically significant difference between the first (λ_z = 1.0) and third groups (λ_z = 1.5) (P < 0.05), between the second (λ_z = 1.3) and third groups (P < 0.05), while there is not a statistically significant difference between the first and second group (P = 0.129). The ratio of axial to circumferential stress of coronary adventitia and intact vessel at different axial loads λ_z = 1.3 and λ_z = 1.5 with circumferential loads λ_θ = 1.3 (C) and λ_θ = 1.5 (D) are shown. *Significant differences (P < 0.05). **Data of intact vessel were cited from Ref. 21.

As shown in Fig. 3, C and D, the ratio of axial to circumferential stresses of the adventitia reflects a much more significant axial response compared with that of intact coronary arteries observed in a previous study (21) (where the species were the same and the size of animals was similar with the present study; and the conditions and mechanical protocol were identical to those used in this study). Before collagen recruitment (λ_θ = 1.3), the axial stress of the adventitia was 235% of the circumferential stress at λ_z = 1.3 (under equi-biaxial stretch), while the axial stress was about five times of the circumferential stress at λ_z = 1.5. The ratios were much larger than that of intact vessel, suggesting that, at a low strain level (λ_θ = 1.3), the adventitia, being stiffer in the axial direction, is in line with initial arrangement of elastin fibers at the ZSS. At λ_θ = 1.5, the axial stress was still higher for the adventitia (especially at equi-biaxial stretch ratio λ_z = 1.5), as a direct result of the axially oriented collagen fibers recruited to take up the load, while it was much lower than the circumferential stress of the intact vessel (Fig. 3D).

Microstructural Deformation of Coronary Adventitia Under Biaxial Loads

Collagen fibers tended to uniformly align toward the principal direction (Fig. 4), while elastin reoriented more randomly (Fig. 5) in a sublayer at elevated loads. At lower λ_θ = 1.0 (Fig. 4, A–C, and Fig. 5, A–C), both collagen and elastin fibers were stretched towards the axial direction with an increased λ_z, and corresponding stresses are very small (Fig. 3, A and B), suggesting that undulated collagen does not resist pressure under low loads. When λ_z = 1.5, a considerable number of collagen fibers became straightened to take up loads (Fig. 4C), which resulted in a relatively large axial stress of the adventitia (when λ_θ = 1.0 and λ_z = 1.5 in Fig. 3B). At the higher λ_θ = 1.5, fibers were first shifted to circumferential direction and then gradually reoriented toward the axial direction with an increase of the axial stretch. At λ_z = 1.0, collagen fibers were circumferentially stretched but still undulated so the macroscopic stresses of the adventitia were very low [Fig. 3, A and B (when λ_z = 1.0 and λ_z = 1.5)], at this load. At physiological loads (λ_θ = 1.5 and λ_z = 1.3, Fig. 4E), a few collagen fibers were completely straightened to take up load, suggesting that macroscopic stresses were not very high at this load (Fig. 3, A and B). When λ_z approached 1.5, more collagen fibers were recruited, and the axial stress increased significantly to more than twice the circumferential stress, as shown in Fig. 3, C and D. The elastin fibers with multiple orientation angles in each sublayer gradually reoriented toward the loading direction to form a network in Fig. 5, D–F. Moreover, most collagen fibers were significantly stretched in the circumferential direction and became straightened at λ_θ = 1.8 (Figs. 4, G–I and 6F), and elastin fibers further stretched to a network-like structure (Fig. 5, G–I). Correspondingly, the macroscopic stresses increased sharply and the axial stress became comparable to the circumferential stress at this load (Fig. 3, A and B).

Fig. 4.

Second harmonic generation (SHG) images of collagen fibers under different biaxial mechanical loading. A–C: collagen at λ_θ = 1.0 and λ_z = 1.0, 1.3, and 1.5, respectively. D–F: collagen at λ_θ = 1.5 and λ_z = 1.0, 1.3, and 1.5, respectively. G–I: collagen at λ_θ = 1.8 and λ_z = 1.0, 1.3, and 1.5, respectively. X denotes axial direction, and Y denotes circumferential direction in all images.

Fig. 6.

The deformation-loading relations of elastin and collagen fibers. A, C, and E: fiber reorientation, presented by normalized orientation angles, at different biaxial loading. B, D, and F: change of collagen waviness at different biaxial loading. Each data point presents average parameter over a specimen. A and B: λ_θ = 1.0. C and D: λ_θ = 1.5. E and F: λ_θ = 1.8.

Figure 6 shows quantitative data of in situ biaxial deformation of fibers. Fiber reorientations of five specimens varying with an λ_z, at each λ_θ (P < 0.05 between λ_z = 1.0 and 1.5), were plotted in Fig. 6, A, C, and E. The fiber reorientation is presented by normalized orientation angle in reference to initial orientation angle of each sample at ZSS. The correlations between normalized orientation angle and λ_z are summarized in Table 1. The slopes of three correlations of collagen were nearly the same (α ≈ 0.41), suggesting that changes of collagen orientation are similar at the three distension loads. It was also found that elastin shifted toward the axial direction more gradually (α ≈ 0.34) at each distention load. The change of elastin orientation angle, however, was more random compared with that of collagen fiber, especially at higher loads. At λ_θ = 1.0 and 1.5, the change of elastin orientation was significant under elevated λ_z (P < 0.05), but became insignificant (P > 0.05) at higher load λ_θ = 1.8. The changes of the waviness of collagen fibers were plotted in Fig. 6, B, D, and F. The waviness of collagen fibers decreased rapidly with an increase of λ_z at λ_θ = 1.0 (P < 0.05), as shown in Fig. 6B, but decreased relatively slowly at λ_θ = 1.5 (P < 0.05, Fig. 6D). At λ_θ = 1.8, most of collagen fibers became completely straightened, as shown in Fig. 6F.

Table 1.

The relations between axial stretch ratio λ_z and fiber geometric parameters (the orientation angle and waviness) are assumed as linear function: y = αx + β, determined by least squares method

		Collagen			Elastin
Geometric Parameter	Circumferential Stretch Ratio λθ	α	β	R2	α	β	R2
Normalized orientation angle	1.0	0.41	0.61	0.81	0.36	0.65	0.73
	1.5	0.41	0.46	0.65	0.31	0.63	0.47
	1.8	0.42	0.27	0.50	0.34	0.49	0.27
Waviness	1.0	−0.38	1.59	0.88
	1.5	−0.21	1.38	0.66
	1.8	−0.02	1.03	0.02*

R² is the correlation coefficient of fit.

^*At circumferential stretch ratio λ_θ = 1.8, most collagen fibers become straightened, and fiber waviness remains as ≈1.0.

Table 2 summarizes the geometrical parameters of collagen and elastin fibers at ZSS and at biaxial loads (predicted by the correlations between parameters and stretch). It also provides the corresponding ratio of axial stress to circumferential of adventitia, which confirms that the macroscopic mechanical properties of vascular tissue stem from microstructure.

Table 2.

The mean of fiber geometrical parameters measured at ZSS and predicted by geometric parameter-stretch relations (Table 1) at biaxial loads, and the corresponding macroscopic stress ratio of coronary adventitia

	ZSS	Physiological Loads: λθ = 1.5 and λz = 1.3	Biaxial Load: λθ = 1.5 and λz = 1.5
Orientation angle, °
Collagen	110 ± 28*	109 ± 28	118 ± 30
Elastin	115 ± 31*	118 ± 32	125 ± 34
Waviness
Collagen	1.30 ± 0.1	1.11 ± 0.1	1.07 ± 0.1
Elastin	1.02	1.0	1.0
Stress ratio, %
Adventitia	Undefined	120	215

Values are means ± SD.

ZSS, zero-stress state; λ_θ, circumferential stretch ratio; λ_z, axial stretch ratio.

Stress ratio denotes the ratio of axial stress to circumferential stress of the adventitia.

^*The second peak of bimodal distribution of fiber orientation was used.

DISCUSSION

Although there are many pseudostructural and structural constitutive models developed (4, 26, 28, 30, 41, 43, 46) for biological tissue in the past decades, these models cannot accurately predict stress and strain on individual fibers or cells. To accurately predict the mechanical microenvironment of blood vessel, it is necessary to use realistic microstructural data and the biaxial loading-deformation relations of fibers. This study provides quantitative data on in situ deformation of both elastin and collagen fibers of blood vessel adventitia and reveals that the longitudinally stiffer adventitia is a direct result of initial fiber alignment. The overall mechanical behavior of the tissue highly depends on biaxial deformation of fibers, and the measured relations between biaxial loading and fiber structure demonstrated the structure-function relation of coronary adventitia. Future studies should focus on identification of the elastic material parameters of single fibers based on measured microstructural data and structural constitutive models. These material parameters are expected to have physical meaning as opposed to empirical curve fit parameters in phenomenological models.

Simultaneous Mechanical Loading-Imaging on Fresh Vessels Under Physiological Loads

Simultaneous mechanical loading-imaging of soft tissues has advanced in the past decade. Some previous loading-imaging studies (16, 19, 24, 39) splayed a vessel open and scanned a planar tissue, which introduces artifacts to the microstructure, as well as nonphysiological stresses and strains to the specimens. Recently, simultaneous mechanical loading-imaging on intact fresh, unfixed, and unstained vessels were conducted to investigate the microstructure and their response to external mechanical loading (1, 3, 25). Arkill et al. (1) and Keyes et al. (25) investigated the reorientation of collagen and elastin fibers under pressurized conditions and provided the corresponding statistical data. These studies did not track the scan area between various load states, however, and did not measure in situ deformation of individual fibers.

A previous study overcame the above limitations by using fluorescent microspheres as markers to track the scan area as well as the deformation of individual fibers under uniaxial distension load (3). The combination of simultaneous loading-imaging and fluorescent markers enables the measurement of in situ deformation of individual fibers to accurately reflect the loading-deformation relation of collagen and elastin fibers. The present work extended the previous methodology to conduct simultaneous loading-imaging under biaxial loading conditions to reveal the in situ deformation of individual fibers at physiological loads (λ_θ = 1.5 and λ_z = 1.3), pressure-overload (λ_θ = 1.8) or increased axial stretch (λ_z = 1.5), to provide insight into the local mechanical behavior and tissue mechanics. Furthermore, the overall behavior of the adventitia was related to the deformation of microstructure to understand the anisotropic mechanical properties of coronary adventitia in relation to arrangement of elastin and collagen fibers.

Anisotropic Properties of Vessels Stem From Microstructure

Our laboratory previously reported that most of the collagen fibers in sublayers aligned approximately with the axial direction, and elastin fibers oriented more randomly at no-distension state with physiological stretch ratio (λ_θ = 1.0 and λ_z = 1.3) (3). The measured distributions of fiber orientations (at ZSS) are consistent with previous observations, where collagen and elastin fibers were more undulated than those at λ_θ = 1.0 and λ_z = 1.3 state (3). Moreover, the diameters of fibers were found to be slightly larger than those at λ_θ = 1.0 and λ_z = 1.3 state, suggesting that the physiological stretch ratio λ_z = 1.3 slightly stretches the fibers. We also studied the biaxial incremental elastic moduli of coronary arteries (31) and found that the intact coronary vessel is a heterogeneous and anisotropic material, where the adventitia is stiffer in the longitudinal direction, and the media is stiffer in the circumferential direction. In the present work, we obtained the morphometry of coronary adventitia fibers at the ZSS to show that the large longitudinal modulus of the adventitia is a direct result of the initial fiber (collagen and elastin) alignment (Fig. 2A). Similar results have also been observed in other biological tissues, such as in the inner media of the carotid artery (39).

In Situ Deformation of Collagen and Elastin Fibers of Coronary Adventitia

The initial geometry of a single fiber strongly affects the extent of its deformation. Hence, the nearly uniform distribution (either orientation angle or waviness) of collagen in each sublayer makes their deformation homogeneous at higher loads (Fig. 4), while elastin fiber deformed heterogeneously with multiple orientations (Fig. 5). Under biaxial loads, fibers gradually reoriented toward the loading directions, and the fibers aligning in the loading direction were stretched much more than those aligned off from this direction, but reoriented relatively less. Therefore, elastin fibers reoriented toward the axial direction with an increase of axial load (Fig. 5, A–C). When subjected to circumferential stretch (Fig. 5, D–F), however, they gradually unfolded to a network as shown in Fig. 5, D–I, due to the elastin fiber alignment in sublayers, where the minor orientation of elastin was approximately orthometric to the major orientation. This observation has previously been made for the mouse aorta, where the elastin fibers, with preferential axial alignment, showed a spread of orientation with two almost orthometric peaks after pressurization (25). Consequently, elastin fibers resemble a network, while collagen fibers align uniformly in a sublayer. It was confirmed that the orientation of fibers dictates the macroscopic anisotropy of vascular tissues in the physiological state. Specifically, the noncollagenous matrix material, predominated by elastin fibers, accounts for the isotropic mechanical response, while collagen fiber contributes mainly to anisotropic deformation (4, 18, 28, 46).

The relation between λ_z and fiber reorientation showed that changes of collagen orientation at three λ_θ values were similar, implying that the fibers are sensitive to both circumferential and axial mechanical loads. Normalized orientation angle of elastin decreased similarly with lower slopes (Table 1), but became highly heterogeneous at higher distensions (Fig. 6, C and E, with lower R², Table 1). This is a result of elastin gradually forming a netlike structure in adventitia sublayers. The change of collagen waviness with an increase of axial strain was different than that of fiber orientation. At the no-distension state (λ_θ = 1.0), the mean waviness of collagen decreased rapidly as axial stretch increases (Fig. 6B) and then decreased relatively slowly with an increase of axial stretch while collagen became straightened gradually to take up the load at the physiological distension (λ_θ = 1.5, Fig. 6D). At λ_θ = 1.8, more collagen fibers became completely straightened even at low axial stretch (Fig. 4, G and H) and predominated the mechanical function of adventitia (3, 44). The fibers aligning in the longitudinal direction remain undulated at high distension with low axial stretch. These observations reveal microscopic mechanical responses of vascular tissue and are essential to understand tissue mechanical properties and response at macroscopic level.

Limitations of Study

There are several potential limitations in the present study. First, the present study focused only on in situ deformation of fibrous components of arterial adventitia. The adventitia consists of dense fibers, ground substance, and a few fibroblasts (10, 33). Ground substance is an amorphous gellike structure that mainly contains glycosaminoglycans, proteoglycans, and glycoporteins (9, 14), while fibroblasts synthesize and secrete extracellular matrix and play an important role in healing wounds. They both have negligible effects on passive mechanical behaviors of tissues (6, 8, 34, 38). To understand the interaction between smooth muscle tone and fibers, a comprehensive microstructural database of not only the fibers but also of smooth muscle cells is needed. Our laboratory has recently published a study on the morphometry and deformation of smooth muscle cells (5). Future studies are needed to quantify the geometry of collagen and elastin fibers in media, as well as their interaction with cells, and then integrate all the microstructural data in a structural model for an entire blood vessel.

Second, fibers may be longer than the field of view (longer than 120 μm), but the present study limited the field of view to 120 × 120 μm². The material area became larger as the tissue was stretched, but the field of view remained the same size. To address this, we focused on those fibers that were in the middle of the field of view at various magnifications. Moreover, the collagen and elastin fibrils are cross-linked by aldehyde formations from lysine or hydroxylysine side chains (11), whose molecular dimension is in the order of nanometer (35). Hence, the dimensions of the cross-links are far lower than the current magnitude for fiber bundles (micrometer), and the related information cannot be obtained with the current spatial resolution of MPM.

Third, the inversion of the adventitia may introduce some residual strain in the circumferential direction. This effect is likely to be small, given the small thickness of the adventitia. Furthermore, the effect on the deformation pattern of fibers is likely negligible, given that most fibers in the inner adventitia aligned toward the axial direction, which are unaffected by small changes in circumferential deformation. In addition, the loading was exerted on adventitia specimens by inflating a custom-made elastic balloon (3), which was used to avoid microvibration and possible movement induced by potential fluid leakage through the adventitia with direct pressurization. Since the distension loading was determined by the ratio of midwall circumference of loaded vessels to that of ZSS (i.e., λ_θ), it was not exactly the same for each specimen (average distension loads were 0.91 ± 0.09, 1.5 ± 0.1, and 1.83 ± 0.08, respectively). This may have led to some dispersion in data of Fig. 6, as well as the reduced R² in Table 1. Moreover, the present study focuses on ex vivo vessels under physiological loading due to the lack of availability of in vivo MPM technology.

Finally, the present findings correspond to the elastic response of the fibers and vessels after preconditioning. The viscoelastic property of the coronary arteries and fibers is not important during the cardiac cycles, since the time constant of a cardiac cycle (<1 s) is much shorter than the time constants of viscoelasticity (>> 1 s) (36) of coronary arteries. Furthermore, although preconditioning may alter the internal structure of vessels to ensure that the mechanical testing is reproducible, it is a standard in mechanical testing of soft tissues to ensure a steady-state condition. Hence, the obtained results are in accordance with the standards of mechanical testing and represent the preconditioned arrangement of fibers.

Conclusions

This study provided quantitative data on in situ deformation of both elastin and collagen fibers of blood vessel adventitia and revealed that the longitudinally stiffer adventitia is a direct result of initial fiber alignment. The overall mechanical behavior of the tissue highly depends on biaxial deformation of fibers, and the measured relations between biaxial loading and fiber structure demonstrated the structure-function relation of coronary adventitia. These findings can be incorporated in microstructural models to accurately predict the biaxial mechanical behaviors of vascular tissues.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grant 1 R01 HL-117990.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: H.C. and Y. Lanir conception and design of research; H.C., M.N.S., Y. Liu, and J.-X.C. performed experiments; H.C., M.N.S., Y. Liu, X.Z., J.-X.C., and Y. Lanir analyzed data; H.C., M.N.S., Y. Liu, X.Z., J.-X.C., Y. Lanir, and G.S.K. interpreted results of experiments; H.C. prepared figures; H.C. drafted manuscript; H.C., M.N.S., Y. Liu, X.Z., and G.S.K. edited and revised manuscript; H.C., M.N.S., Y. Liu, X.Z., J.-X.C., Y. Lanir, and G.S.K. approved final version of manuscript.