Elastin and collagen fibers are the major load-bearing extracellular matrix (ECM) constituents of the vascular wall. Arteries function differently than veins in the circulatory system; however as a result from several treatment options, veins are subjected to sudden elevated arterial pressure. It is thus important to recognize the fundamental structure and function differences between a vein and an artery. Our research compared the relationship between biaxial mechanical function and ECM structure of porcine thoracic aorta and inferior vena cava. Our study suggests that aorta contains slightly more elastin than collagen due to the cyclical extensibility, but vena cava contains almost four times more collagen than elastin to maintain integrity. Furthermore, multiphoton imaging of vena cava showed longitudinally oriented elastin and circumferentially oriented collagen that is recruited at supraphysiologic stress, but low levels of strain. However in aorta, elastin is distributed uniformly, and the primarily circumferentially oriented collagen is recruited at higher levels of strain than vena cava. These structural observations support the functional finding that vena cava is highly anisotropic with the longitude being more compliant and the circumference stiffening substantially at low levels of strain. Overall, our research demonstrates that fiber distributions and recruitment should be considered in addition to relative collagen and elastin contents. Also, the importance of accounting for the structural and functional differences between arteries and veins should be taken into account when considering disease treatment options.
Vein grafts are commonly used for coronary artery bypass surgery, lower extremity ischemia, or arteriovenous fistulas for dialysis treatment, in addition to vena cava grafts and stents specifically being relevant to leiomyosarcoma, complications from liver transplants, and aortocaval fistulas from ruptured atherosclerotic abdominal aortic aneurysms [1–7]. Vein graft failure is a common finding (30–50%) in patients following coronary artery bypass graft surgery . Improved understanding of vein mechanics can aid in the development of better vein grafts with lower incidence of failure. Thrombosis or the development of a blood clot, neointimal hyperplasia characterized by thickening of the innermost layer of a vein or artery, and atherosclerosis are the three progressive stages that lead to stenosis or occlusion of the graft . In vein segments grafted into arteries, exposure to high pulsatile pressure load leads to significant wall remodeling within a few days of grafting. Specifically, an inflammatory response, neointima development, collagen and elastin turnover, cellular proliferation, and apoptosis occur [8–10]. Previous research has shown that the abrupt increase in mechanical wall stresses and flow fields in the vein graft activates various cellular and intracellular signaling molecules, which stimulate vascular smooth muscle cells in the media to be synthetic and migratory, accelerating the progression of neointimal hyperplasia [11,12].
It is assumed that at least a part of biomechanical changes developing in the grafted vein wall is a physiological adaptation mechanism that normalizes the extremely high circumferential stress . Understanding the mechanisms of vein graft failure is predicated upon elucidating the structural and functional mismatch between arteries and veins. From the biomechanical point of view, the proliferative structural remodeling of a vein graft occurs in order to maintain the wall stresses at its biologically preset level . Following implantation, the vein graft, which has been subjected to an internal pressure of ∼10 mmHg, is immediately subjected to arterial pressure (∼100 mmHg) and increases in flow as well as wall stresses . Interestingly, experimental evidence supports the idea that certain morphological and biochemical alterations of grafted veins are reversible [14–17]. The relationship between biological remodeling and mechanical function is obviously causal.
It is important to understand that veins and arteries are very different in both function and structure. Both arteries and veins have three layers: tunica intima, media, and adventitia . The thinnest layer is the innermost layer, the tunica intima, which is made up of a single layer of endothelial cells, and its main purpose is to provide hemocompatibility with the blood. In arteries, the thickest layer is the tunica media, i.e., the middle layer of the vessel wall. The tunica media layer contains smooth muscle cells that are embedded in a matrix of elastic lamella, collagen fibers, as well as aqueous ground substance matrix containing proteoglycans. Arteries tend to have a thicker tunica media layer in order to accommodate the pulsatile blood flow. However, veins have a relatively thin tunica media layer . The outermost layer in the vessels is the tunica adventitia or tunica externa, which provides structural stability to the vessels. It consists primarily of a dense network of type I collagen fibers.
Arterial wall mechanics in response to pulsatile blood flow from the heart plays a significant role in normal physiological functions and cardiovascular diseases, and has been studied broadly . Early research by Roach and Burton suggested the contributions of compliant elastin fibers at lower levels of stress and stiff collagen fibers at higher levels of stress result in the nonlinear mechanics of the arterial wall . Arterial structure–function has since been further explored from the macro- to nanoscale with the advancement of experimental approaches including mechanical characterization and optical imaging [20–24], and in computational modeling [25–27]. Alterations in structure and function to the arterial wall due to various pathological conditions are also being extensively studied, with the goal of early diagnosis, prevention, and the development of treatment options [28–30]. Here we refer to several previous publications for a more thorough review [18,31,32].
In contrast to many studies of aorta mechanics, only limited prior study concerning biaxial mechanical properties of vena cava has been performed despite 60–80% of blood volume being localized in veins and vein mechanics being relevant to many clinical interventions [6,33–35]. Uniaxial tests have been used to determine the mechanical properties of veins, but these results are inadequate for fully characterizing the anisotropic behavior of vena cava [36–38]. Moreover, the important biomechanical roles of ECM constituents, such as elastin and collagen, in controlling arterial and venous function are under-recognized. The overall goal of this study was to study and elucidate the differences in ECM structure and biaxial mechanical properties for aorta and inferior vena cava. ECM structure was studied with histology, biochemical assays, and multiphoton microscopy, which was paired with mechanical function characterization using planar biaxial tensile testing. Findings from this study will help us understand the fundamental structure–function relationship between ECM fiber distribution and recruitment with the complex nonlinear and anisotropic vascular behavior.
Materials and Methods
Porcine (12–24 months, 160–200 pounds) descending thoracic aortae and inferior vena cavae were obtained from a local abattoir and cleaned of loose connective/fatty tissue. Duplicate square samples (∼20 × 20 mm) were cut from three of each blood vessel type, for a total of six aorta samples and six vena cava samples from six pigs. Samples were cut with one edge parallel to the longitudinal direction and the other edge parallel to the circumferential direction of the blood vessel. Aortae were cut at the first intercostal arterial side branch, and vena cavae were cut between the renal vein side branch and common iliac bifurcation. Samples were mechanically tested within 24 h and kept frozen at −80 °C until multiphoton imaging.
Histology studies were performed to assess the overall structural composition of blood vessels. Samples were fixed in 10% formalin buffer (Fisher Scientific, Waltham, MA) and then imbedded in paraffin. Sections of approximately 6 μm in thickness were obtained and stained with Movat's pentachrome which stains collagen yellow, elastin black, smooth muscle cells red, and glycosaminoglycans blue.
After samples were cut for biaxial tensile testing, adjacent tissue from each blood vessel was used for biochemical assays. Samples were analyzed using a Sircol collagen assay kit following manufacture instructions (Biocolor, Ltd., Carrickfergus, Northern Ireland). The Sircol collagen kit uses a quantitative dye-binding method and absorbance was measured using a SpectraMax M5 plate reader (Molecular Devices, Sunnyvale, CA) at a 540 nm wavelength. Collagen content was expressed as microgram of collagen per milligram of wet tissue weight. Elastin content was measured using Fastin elastin assay kit (Biocolor, Ltd.) [39–41]. Following manufacture's protocols, our assays measured soluble tropoelastins and insoluble elastin that was solubilized into α-elastin polypeptides. The optical density was measured at 513 nm using the microplate reader. Elastin content was expressed as microgram of elastin per milligram of wet tissue weight.
Biaxial Tensile Test.
where F is the applied load, λ is the stretch, Lo is the initial length, and t is the thickness of the tissue. The subscripts 1 and 2 correspond to the longitudinal and circumferential directions, respectively.
To understand the contribution of collagen and elastin fibers to the structure and function of aorta and vena cava wall, mechanical loading was paired with multiphoton imaging of tissue. Vena cava and aorta samples were imaged with an LSM 710 NLO microscope system (Carl Zeiss, Jena, Germany) equipped with a Chameleon Vision–S tunable femtosecond IR pulse laser (Coherent, Santa Clara, CA) with an excitation wavelength of 800 nm. Each sample was imaged with a field-of-view of 425 μm at five locations spread over an area of 1 cm2 to obtain the average structural properties of the sample. Samples were imaged from both the adventitial and medial surfaces to generate second-harmonic generation from adventitial and medial collagen (417/80 nm) as well as two-photon-excited fluorescence from medial elastin (525/45 nm). Image z-stacks were acquired with 2 μm spacing, and maximum intensity projections were used for analysis. All samples were imaged with the circumferential direction of the tissue aligned horizontally. Thus, fibers oriented at 0 deg and ±90 deg are in the circumferential (C) and longitudinal (L) anatomic directions of the blood vessel, respectively. Vena cava samples were imaged at 0%, 5%, and 10% equibiaxial strain, then 10%C–20%L and 10%C–30%L due to the higher degree of anisotropy. The images were then used in the analysis of fiber straightening and recruitment in response to mechanical loading.
which is equal to 1 for a straight line . With increasing strain, adventitial collagen fibers are recruited as they go from wavy to straight and contribute to the nonlinear stiffening of the vascular wall.
Blood vessel thickness, collagen and elastin assay results, stress–stretch data, and straightness parameter are summarized and plotted with mean ± standard error of the mean. Two-tailed independent t-tests were used for analysis with p < 0.05 considered as statistically significant.
Movat's histological stain revealed some interesting differences between aorta and vena cava (Figs. 1(a)–1(c)). In aorta, collagen (yellow) is most visible in the adventitia, but also exists throughout the media where elastin (black) is most prominent. In vena cava, collagen is visible throughout the thickness; however, elastin appears to be discontinuous. Smooth muscles cells (red) are also visible in the aorta media, but are less apparent in the vena cava. Glycosaminoglycans (blue) can be seen in parts of the aorta media, especially toward the inner wall. The adventitia–media boundary is also more obvious in aorta than it is in vena cava. Additionally, the average aorta thickness (1.32 ± 0.09 mm) was significantly (p < 0.05) thicker than vena cava (0.66 ± 0.07 mm). Collagen and elastin assay results show that aorta has significantly (p < 0.05) less collagen (35.78 ± 5.41 μg/mg) than elastin (54.28 ± 7.57 μg/mg), but vena cava has significantly (p < 0.05) more collagen (83.96 ± 7.92 μg/mg) than elastin (22.39 ± 2.02 μg/mg) (Fig. 1(d)).
The average equibiaxial stress–stretch curves demonstrate marked differences in the mechanical properties between aorta and vena cava (Fig. 2(a)). Both aorta and vena cava show a stiffer circumference than longitude; however, much greater anisotropy was observed in vena cava. At 200 kPa, the average vena cava circumference stretch is 1.06 ± 0.01, whereas the average longitude stretch is 1.30 ± 0.06. In contrast, average aorta circumference and longitude stretch to the range of 1.19 ± 0.01 to 1.22 ± 0.01 at 200 kPa. Tangent modulus was obtained by taking the derivative of a sixth-order polynomial fit to the stress–stretch data (Fig. 2(b)). For aorta, the tangent modulus ranges from 0.21 ± 0.04 to 1.59 ± 0.40 MPa in the circumferential direction and 0.54 ± 0.08 to 1.11 ± 0.14 MPa in the longitudinal direction, with the circumferential direction being slightly stiffer when stretches are higher than 1.10. For vena cava, however, the tangent modulus in the circumferential direction increases dramatically upon loading from 0.30 ± 0.75 to 6.35 ± 4.05 MPa. In the longitudinal direction, the vena cava tangent modulus is in a similar range as aorta when stretches are less than 1.10, and then increases to over twice that of aorta up to 2.27 ± 0.91 MPa at higher levels of stretch.
A detailed exploration of the structural differences between aorta and vena cava was made possible by using a multiphoton microscope. Figure 3 shows the organization of elastin (green) and collagen (blue) in the medial and adventitial layers of the two vascular types. Overall, the ECM fibers in aorta appear to be more uniformly distributed than vena cava. Compared to aorta, the elastin fibers in vena cava show an obvious longitudinally orientation, whereas medial and adventitial collagen fibers are predominantly circumferentially oriented. It is also noticed that the vena cava adventitial collagen seems to have a tighter crimp than aorta. Looking specifically at multiphoton images of vena cava at different levels of strain in Fig. 4, adventitial collagen clearly begins straightening with 5% strain and is fully recruited at 10% strain. Additionally, medial collagen can be seen crimped at 0%, but straighter at 5% and straight at 10% strain. Elastin fibers are longitudinally oriented and show minimal visual changes.
Straightness parameter (Ps) was used to quantify the engagement of adventitial collagen in aorta and vena cava (Fig. 5), which did not show significant difference at unstretched (zero equibiaxial strain) state between aorta (0.85 ± 0.10) and vena cava (0.87 ± 0.01) (Fig. 2). However, the adventitial collagen of vena cava becomes significantly (p < 0.05) straighter with the first 5% increase in strain (0.96 ± 0.003), and the fibers at 10% equibiaxial strain are essentially straight (Ps = 0.98 ± 0.002). Aorta adventitial collagen showed a delayed engagement at about 20% of strain and was not fully straightened until about 30–40% strain (Ps = 0.97 ± 0.02 to 0.99 ± 0.01).
Figure 6 compares the distributions of ECM fibers in aorta and vena cava with mechanical loading. In Table 1, fiber distribution was further quantified by determining the ratio of circumferentially (0 ± 20 deg) to longitudinally (90 ± 20 deg) oriented fibers . A ratio closer to 1 indicates uniformly distributed fibers, and a ratio greater than 1 indicates more circumferentially distributed fibers. The distribution of elastin fibers shows minimal changes with increasing strain. Both medial and adventitial collagen fibers show preferential distributions in the circumferential direction within the 10% strain range. At 0% strain, aorta has relatively more circumferentially oriented collagen than vena cava. At 5% strain, aorta and vena cava have a similar proportion of collagen in the circumferential direction, but at 10% strain, vena cava actually has proportionally more circumferentially aligned collagen. Overall, both the medial and adventitial collagens in vena cava become more circumferentially aligned with increasing strain.
To study fiber reorientation with deformation, vena cava was subjected to 10%C–20%L and 10%C–30%L nonequibiaxial strain, where C and L represents the circumferential and longitudinal directions, respectively (Fig. 7). Minimum changes were observed in elastin fibers except that the longitudinal elastin peaks become higher. However, the circumferentially oriented medial and adventitial collagen fibers show wider distributions, suggesting that the circumferentially orientated fibers begin to shift toward the longitudinal direction in response to the large longitudinal deformation.
Our study provided new understandings on the striking differences in ECM fiber distribution and recruitment between aorta and vena cava, and their close correlation with the biaxial stress–stretch mechanical behavior of blood vessels from a large animal model representative of human physiology. Aorta and vena cava have very different functions in the circulatory system . Whereas the arterial wall is subjected to a pulsatile transmural pressure of approximately 76–114 mmHg , mechanical loads borne by veins are quite different. Blood pressure in veins is not pulsatile, but steady and much lower than arterial pressure. Veins are capacitance vessels with greater compliance at low pressure but very stiff at supraphysiologic values (>30 mmHg transmural pressure) .
Collagen and elastin assay results show vena cava is primarily collagen, but aorta has more elastin (Fig. 1(d)). Adventitia generally comprises 60–75% of vena cava and 10% of aorta, with the media of vena cava not being as clearly defined as in aorta because there is no internal/external elastic lamina . Our finding of a ∼4:1 (collagen:elastin) ratio in vena cava and 1:1.5 in thoracic aorta is in agreement with previous studies [13,45]. The collagen to elastin ratio was reported to be ∼5:1 in saphenous vein and ∼1.5:1 in canine abdominal aorta , and thoracic aorta has been found to have more elastin than abdominal aorta . Elastin fibers in large elastic arteries such as aorta are essential for its cyclical extensibility. However, vena cava contains much more collagen than elastin mainly to maintain integrity without the need to sustain cyclic loading.
There is a marked difference in elastin alignment between aorta and vena cava. Elastin has remarkably long half-life and can elongate up to 150% of its original length [13,31]. In terms of a structure–function relationship, we are reminded that aorta experiences repeated multiaxial cyclic extension from pulsatile blood flow. Therefore, it makes sense that aorta would have a rather uniformly oriented elastin. The ECM structural changes with mechanical loading in aorta have been studied in detail by Chow et al. . Elastin in vena cava, however, is longitudinally oriented (Figs. 3 and 4). The scattered appearance of elastin in the histology images in Figs. 1(b) and 1(c) is due to looking at the cross section of longitudinally oriented elastin. Longitudinal elastin has also been found in porcine jugular veins . One possible reason for longitudinally oriented elastin in veins maybe to handle the extension due to breathing , since the torso elongates during inhalation and imposes a longitudinal stretch on the vena cava.
The large anisotropy seen in the vena cava shows strong correlation between ECM fiber structure and mechanical function of blood vessels. The unique distribution of elastin in vena cava results in a compliant longitudinal stress–strain behavior (Fig. 2(a)). With the absence of elastin in the circumferential direction and tighter crimp in the adventitial collagen of vena cava (Fig. 3), the stress–strain curve shows a rapid stiffening behavior at lower levels of stretch. In contrast, aorta shows much less anisotropy than vena cava (Fig. 2(a)). The tangent modulus is in the range of 0.21–1.59 MPa for aorta and 0.18–6.35 MPa for vena cava (Fig. 2(b)), which supports the known role of veins as capacitance vessels that are compliant at low physiologic pressure, but very stiff beyond the physiologic range . It is important to note that such findings were based on tissue-level experiments and may not be revealed from local measurements. For example, Akhtar et al. used nanoindentation to compare the mechanical properties across the thickness of ferret aorta and vena cava . For aorta, the elastic modulus linearly decreased from ∼30 MPa at the adventitia to ∼8 MPa at the intimal surface. For vena cava, the modulus was ∼35 MPa at the adventitia and intima, but ∼20 MPa in the medial layer. The distribution of ECM components plays an important role in local property measurements. In addition to difference between the percentages that adventitia comprises aorta versus vena cava, it is also worth noting that medial collagen is generally considered to be mainly type III and the adventitia mainly the stiffer type I collagen.
The structural characteristics such as fiber distribution and recruitment need to be considered in addition to the relative collagen and elastin contents. The wavier collagen fibers seen in the aorta at low strain (e.g., 5–10%) is correlated with the higher in vivo strains (about 20%) for aorta [31,47,48]. Additionally, aorta has more elastin to contribute at lower levels of strain, especially in the circumferential direction, and thus a delayed recruitment for collagen at higher levels of strain that compliments the observed nonlinear stiffening . Collagen undergoes continuous turnover during a lifetime. There is also evidence that faster loading rates in aorta can lead to a shorter half-life (∼22 days) compared to a less demanding mechanical environment [47,49,50] such as vena cava. In addition, collagen in vena cava is deposited in the vessel wall in the absence of pulsatile motion. Thus, the adventitial collagen in vena cava straightens at low levels of strain (Figs. 4 and 5) and reveals a narrower circumferential distribution due to fiber straightening (Fig. 6), which correlates well with the very stiff circumference of the vena cava (Fig. 2(a)).
The present study focuses on porcine blood vessels and may not be completely generalizable to humans. Also, the sample size of this study was adequate to demonstrate the observed differences in structure–function relationship between aorta and vena cava; however, future studies with larger sample size would be beneficial. The variability in the stress–stretch behavior in Fig. 2, calculated as the standard deviation as percent of average deformation, ranges from 12% to 27%. Planar biaxial testing of human descending thoracic aorta variability was reported to range from ∼10% to 15% , and porcine ascending thoracic aorta variability was up to 33% . The variability in the stress–stretch behavior of vena cava in this study ranges from 32% to 42%, and the closest comparison is uniaxial ovine vena cava with variability of up to ∼23% within an individual . Finally, the choice of a sixth-order polynomial was used to obtain the tangent modulus . However, we acknowledge that the use of a fifth-order polynomial has been used to successfully determine the tangent modulus .
This study establishes several new fundamental understandings on the differences between aorta and vena cava at the structural and functional levels. In summary, aorta contains less collagen than elastin, but vena cava contains almost four times more collagen than elastin. In vena cava, the elastin is longitudinally oriented and the collagen becomes more circumferentially oriented and straightens at low levels of strain. These structural observations support the functional finding that vena cava is more anisotropic than aorta because the circumference stiffens substantially at low levels of strain. Importantly, these findings should be put in context with the fact that physiologic pressure in vena cava is much lower than aorta. Overall, this study improves the understanding of the contribution of ECM constituents to the mechanics of blood vessels. Specifically, it points to the need to consider ECM structure in addition to the relative collagen and elastin amounts. When comparing the mechanics of vena cava to aorta, it is important to remember that ECM fiber distribution, recruitment, and content each play an important role in vascular function. Finally, this study demonstrates the importance of accounting for the structural and functional difference between arteries and veins when considering disease treatment options.
The authors would like to acknowledge the Multiphoton Microscope Core Facility at Boston University School of Medicine for training and access to equipment. This work was supported, in part, by a grant (CMMI 1463390) from National Science Foundation, and a grant (R01HL0950825) and a predoctoral training grant (2T32HL007969) from National Institutes of Health.