Development and Initial Characterisation of a Localised Elastin Degradation Ex Vivo Porcine Aortic Aneurysm Model

1. Introduction

Aneurysms of the thoracic and abdominal aorta constitute a clinical disorder defined by dilation beyond 50% of their normal diameter [1]. They affect approximately 4.8% of the world’s population [2] and cause over 150,000 fatalities each year [3]. Endovascular aneurysm repair using a stent graft is the preferred mode of treatment for both thoracic and abdominal aortic aneurysms (TAA, AAA, respectively) due to lower short-term morbidity and mortality compared to open surgical repair [4,5]. However, current endovascular devices have high rates of device-related complications often necessitating re-intervention. The most common complications include endoleaks (15-30% within 30 days), endograft migration (10% within 1 year), and collapse [5]. There is a need to optimise endovascular aneurysm repair devices to maintain functionality in the challenging mechanical environment present in aortic aneurysms (AA) [6,7,8,9-9].

An implanted stent graft must match both the dilation and recoil of the vessel wall during the cardiac cycle and withstand the haemodynamic forces imposed by flowing blood to maintain a proper seal and fixation [5]. Not only are the geometry of the aorta and the pressure-flow characteristics of the blood important, but the compliance of the vessel wall must also be taken into account [10]. The healthy arterial wall exhibits non-linear hyperelastic behaviour characterised by an increasing stiffness with increasing deformation resulting in a J-shaped stress-strain curve [11]. This behaviour is attributed to the initial recruitment of elastin fibres in conjunction with the progressive strain-dependent recruitment of stiff collagen fibres [6,12]. Destruction and degeneration of mural elastin fibres is widely believed to cause the increases in stiffness seen in AA through the resultant load being taken up by the stiffer collagen fibres [7,13,14]. As such, the mechanical environment present in AA differs from that found in healthy aortae because of both the locally dilated vessel geometry and decreased mural compliance. Optimisation of endovascular devices using in vitro benchtop models, which recapitulate these changes in geometry and compliance, offers a means to address, and ultimately reduce, the incidence of endovascular device complications secondary to mechanical failures.

In vitro models represent an invaluable tool in facilitating the refinement and optimisation of device designs prior to costly preclinical in vivo trials [15,16]. The ex vivo porcine aorta is the most utilised model based on its widespread availability and structural similarity to healthy human aortae [17,18,19,20]. However, it does not replicate the dilation nor the stiffening seen in AA [21]. While synthetic models can accurately replicate the geometric features of aneurysms, the materials used poorly replicate the compliance and rupture properties of arterial tissues [10,22,23,24,25]. Currently available tissue engineered in vitro models are more suited to studies of aneurysm pathophysiology and have not yet been manufactured on a scale suitable for endovascular device testing [26,27,28]. At present, there are no suitable in vitro models which accurately reproduce a physiologically relevant environment for endovascular device testing.

Enzymatic treatments represent a promising avenue to induce microstructural changes which mimic the increases in diameter and stiffness seen in AA. In a seminal study by Roach et al., human iliac arteries treated with crude trypsin for elastin degradation demonstrated a qualitatively greater diameter and significantly greater stiffness under low hydrostatic pressures [6]. Following on from this, Dobrin et al. reported that elastin-depleted human iliac arteries exhibit decreased distensibility and only moderate dilation under physiological pressures; therefore, demonstrating the feasibility of this approach as a mechanical model for aneurysms [29]. In elastin-depleted strips of porcine aorta, Kratzberg et al. reported that a near aneurysmal increase in unloaded circumferential length can be achieved through application of a creep-loading protocol [30]. Furthermore, Chow et al. reported that a mild elastase treatment produces gradual changes in the biaxial stress-strain behaviour of porcine aortic strips characterised by increasing stiffness with increasing elastase incubation time [31]. The viability of this approach is further supported by the successful use of elastase treatments to produce models of aneurysms in small vessels in an ex vivo setting [32,33] and in vivo AA models in small and large animals [34,35]. Researchers have been studying the effects of elastase on a wide range of in vitro and in vivo arterial tissues since the 1950s [6]. Despite this, a scaled-up ex vivo model of an aneurysm, suitable for endovascular device benchtop testing, has yet to be developed.

We hypothesized that the localised application of elastase to porcine aortic segments would constitute a feasible methodology to recapitulate the critical features of the mechanical environment present within AA, specifically: increased diameter and stiffness under physiological pressures. To test this hypothesis, the effect of elastase treatment was first studied on individual rings of porcine aorta, to demonstrate proof-of-principle, before application to whole porcine aortae. Rings were incubated in elastase across a range of times with changes in elastin structure and content assessed qualitatively and quantitatively through histology. The circumference and thickness of treated rings were measured using ImageJ analysis software and mechanically characterised via ring tensile testing. From this, an incubation time was selected, and the methodology scaled up to achieve localised degradation in whole porcine aortae using a specially designed test rig. The effect of spatially confined degradation on pressure-diameter behaviour was measured via inflation extension testing. Altogether this work aimed to show the viability of a localised degradation approach to produce porcine aortae with aneurysmal geometry and compliance; therefore, making it a viable preclinical ex vivo model for endovascular device testing.

2. Materials and Methods

2.1. Tissue Preparation and Chemical Degradation

Thoracic aortae from 6-month-old healthy white pigs were obtained from a local abattoir. Loose connective tissue was debrided, and samples were cryopreserved within 3 hours of sacrifice as described previously [36]. Briefly, debrided samples were placed in 50 mL falcon tubes containing tissue freezing medium consisting of RPMI medium (21875035, BioSciences), sucrose (S0389, Sigma), and dimethylsulfoxide (PIER20688, VWR International) before being cooled at a controlled rate of 1°C/min to -80°C in a Mr. Frosty container (C1562, Sigma). Samples were thawed at 37°C and rinsed with phosphate buffered saline (PBS) to remove excess cryoprotectant prior to experiments. Elastase treatment solution consisted of 1 U/mL elastase (E1250, Sigma Aldrich); 0.35 mg/mL of trypsin inhibitor (10109886001, Sigma); and Dulbecco’s modified eagle medium (DMEM) (61965026, BioSciences). Controls were treated with DMEM alone.

Individual ring samples were enzymatically digested to inform the degradation protocols for whole aorta models. Experimental groups are provided in Table 1. For the ring analysis, 2 mm wide rings were cut adjacently from descending porcine aortae and placed in PBS in a 12-well plate (N=5 pigs, n=5 rings per pig). Once all rings were cut, the PBS was replaced with 2 mL of either elastase or control solution so as to cover the entire ring and well plates were placed on a shaker at 99 RPM in an oven at 37°C for the designated incubation time. After incubation, samples were rinsed with PBS and cryopreserved as described above. All rings underwent two cryopreservation cycles.

For whole aorta models, a customised apparatus was designed, printed (PLA), and used to localise treatment to a 1 cm intraluminal section of a 6 cm long aorta (Figure 1). The 3D printed apparatus consisted of a bath, grips and proximal and distal grip holder components as shown in Figure 1a. Treatment was localised to a 1 cm segment by creating a seal between externally applied zip ties and intra-luminally placed cylindrical hollow grips (Figure 1b,c). Parafilm was applied to both intra- and extra-luminal surfaces to minimise tissue damage associated with zip-tying and to limit enzyme perfusion through the vessel wall. Once secured, aortae were cannulated to tubing and placed in the degradation rig as shown in Figure 1d. Table 1 includes the two whole aorta treatment groups. Elastase and control solutions were prepared as described previously and perfused through the intra-luminal tubing to ensure there were no air bubbles. The outer bath was filled with pre-heated PBS at 37°C. The whole degradation rig was then placed on a shaker at 99 RPM in an oven at 37°C and the aorta rotated 120° every 1.5 hours throughout the 9-hour incubation period. Following incubation, aortae were removed, rinsed in PBS and cryopreserved. All whole aortae underwent two cryopreservation cycles.

2.2. Mechanical Testing

Individual ring models were assessed via ring tensile testing, whereas full aortae were assessed via inflation-extension testing.

2.2.1. Ring Tensile Testing

A materials testing device (ZwickRoell Ltd.) with a 200 N load cell was fitted with horizontally oriented pins and used to conduct tensile tests on rings of porcine aorta. Prior to testing, rings were imaged against graph paper to allow measurement of wall thickness and circumference optically using ImageJ (version 1.53r) analysis software [37]. The thickness of each ring was measured perpendicular to tangent lines constructed at the inner mural border at four equidistant locations on each sample. The circumferential length was measured as the circumference of an ellipse fit to the outer mural border for each sample. The pins used were 1 mm in diameter and all tests were performed within a PBS bath kept at a temperature of 37°C. To begin testing, rings were loaded to hang freely from the top pin without contacting the lower pin and the force was set to zero. The test protocol involved preloading rings to 0.05 N at 10 mm/min and five preconditioning cycles to 10% strain at a rate of 60% the initial gauge length (L₀) per minute. Data was collected from the end of the fifth preconditioning cycle, when rings were stretched from 0% strain to failure at a fixed strain rate of 40% L₀/min. After failure, force-displacement and gauge length data were exported and the rings were fixed for histological analysis.

Force-displacement data was converted to Cauchy and Green-Lagrange definitions of stress and strain, respectively, using equations published by Macrae et al. [38]. First, the circumferential stretch was calculated as [38]:

$${\lambda }_1=\frac{l+\pi r_w}{l_0+\pi r_w}$$

(1)

where $\lambda$ is the calculated stretch factor, $r_w$ is the pin radius, and $l$ and $l_0$ represent the distance between the center of each pin in the deformed and undeformed configurations, respectively; subscript 1 denotes the circumferential axis of loading. Green strain $E$ was calculated as [38]:

$$E_{11}=\frac{1}{2}\left({\lambda }_1^2-1\right)$$

(2)

Cauchy stress was calculated on the assumptions of plane stress and incompressibility [38]:

$${\mathsf{\sigma }}_1=\frac{F_1{\lambda }_1}{2tw}$$

(3)

where $F$ is the force, $t$ the thickness and $w$ the width. Cauchy stress and Green strain data were plotted as stress-strain curves, from which the elastic moduli were calculated over initial and final regions as described by Campbell et al. [39]. Initial and final moduli were calculated as the slope of linear regressions plotted from 0-0.05 Green strain and 0.25-0.3 MPa Cauchy stress respectively.

2.2.2. Inflation-Extension Testing

Inflation-extension testing assessed the quasi-static pressure-diameter behaviour of whole aortic samples at 30 mmHg pressure steps from 0 mmHg to 150 mmHg using another custom designed rig, shown in Figure 2a. Sample ends were secured to cylindrical grips using zip-ties allowing insertion into the testing rig. The outer bath was filled with PBS at 37°C and a physiologic axial strain of 1.35 [40] was imposed using an attached micrometer. Samples were flushed with PBS to remove any air bubbles prior to testing. Intraluminal hydraulic pressurisation was achieved by threading a condom through the vessel and inflating this on account of their watertight nature, negligible stiffness, and greater unstretched diameter with respect to the aortic specimens [41]. Aortic side branches were closed with sutures. Pressures were maintained at each step using closed loop feedback system between an intraluminally placed FOP-M260 pressure probe (FISO); a signal conditioner; a laptop with PID Pump Data Log software (Harvard Apparatus) installed; and a PHD Ultra 703007 Syringe Pump (Harvard Apparatus) equipped with a 50mL syringe reservoir of PBS. Desired pressures were set on the laptop using the PID Pump Data Log software and maintained using this closed-loop feedback. Samples were given two minutes to equilibrate to each pressure step prior to imaging. Aortae were marked at three positions along the length of the vessel: 25% (proximal), 50% (middle/treated), and 75% (distal) using a permanent marker as shown in Figure 2b,c and vessel diameters were measured at each location using ImageJ analysis software. Following inflation samples were rinsed in PBS, sectioned into 2 mm wide rings at the marked sites and fixed for histological analysis. The degree of dilation exhibited by the middle region at each pressure step was calculated separately by normalising the diameter measurement with respect to the initial middle diameter at 0 mmHg for absolute dilation and to the average diameter of the adjacent proximal and distal regions at that pressure step for relative dilation. Lastly, the physiological compliance was calculated as the change in absolute dilation over the 90-120 mmHg pressure range.

2.3. Histological Processing and Histomorphometric Quantification

After mechanical testing, all samples were fixed in 10% formalin solution (HT501128-4L, Sigma) at 4°C for 48 hours, stepwise dehydrated using ethanol and xylene, and embedded in paraffin wax. Wax-embedded samples were sectioned along the transverse axis using a microtome (RM-2125RT, Leica) into 7 μm thick slices. Slides were stained with Verhoeff’s elastic stain and imaged using an Aperio CS2 microscope installed with ImageScope software (V12.3.). QuPath (V0.2.3) image analysis software was used to quantify elastin content using a colour deconvolution approach comprising three distinct steps [42]. First, the RGB values defining stain colour (referred to as the stain vector) were determined from a subset of images. Next, a thresholder was defined to distinguish the tissue from the background in images. Lastly, using the stain vector, a separate thresholder was applied to identify the percentage surface area fraction of the tissue area stained for elastin (Equation (4)).

$$\mathrm{\%}\mathrm{A}\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{f}\mathrm{r}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}=\frac{areastainedforelastin}{tissuearea}*100$$

(4)

Protocols for stain vector determination and tissue and stain thresholding were standardised for all measurements. For elastin quantification of individual rings, each ring was sectioned into four slices, each of which was assessed at three ROIs, and as such, the individual values for each ring represent the average of 12 individual measurements. Each ROI encompassed the full wall thickness.

2.4. Statistical Analysis

Quantitative data are presented as mean and standard deviation. P < 0.05 was determined as the threshold of significance. Outliers were identified and removed using the ROUT method with a set maximum false discovery rate of 1% (Q = 1%). In the case of the presence of an outlier it is stated with the individual result. The Shapiro-Wilk normality test determined use of either a one-way ANOVA (parametric) or Kruskall-Wallis test (non-parametric) and post hoc comparisons were made with either Tukey’s or Dunnett’s correction, respectively. For one-way ANOVA, if the Brown-Forysthe test determined unequal standard deviations then Welch’s correction was applied. Due to the repeated nature of measures for dilation of whole aorta models, a two-way repeated measures ANOVA was performed for analyses of the absolute and relative dilation of the whole aorta models. For this analysis, sphericity was assumed and post-hoc comparisons were made using Šídák’s multiple comparisons test. All statistical analyses were performed with GraphPad Prism software (V 9.3.1).

3. Results

3.1. Impact of Elastase Treatment on Aortic Elastin Structure and Content

Elastase treatment produced a time dependent degradation of elastin in porcine aortic rings. Verhoeff’s elastin staining revealed partially degraded samples which were characterised by radially distinct regions: elastin-depleted inner and outer edges with an elastin-conserved medial region (Figure 3b). This clearly indicates that degradation occurred from enzyme-contacting surfaces moving radially inwards. The innermost medial elastin fibres were mainly conserved following 9 hours of incubation, fragmented at 18 hours, and completely absent following 27 hours (Figure 3b,d). Elastin quantification confirmed the observed decreases in elastin content with increasing elastase incubation time (Figure 3e).

3.2. Impact of Elastase Treatment on Aortic Ring Morphology

Enzymatic treatment with elastase produced progressive changes in aortic ring appearance and morphology with increasing incubation time as demonstrated in Figure 4a-e. Rings transformed from a well-defined tissue in control groups (Figure 4a,e) to semi-translucent, non-rigid, gel-like tissues seen in the 18- and 27-hour elastase treated groups (Figure 4c,d). The average ring thickness and circumference significantly increased following 18- and 27-hour elastase incubations (*** P<0.0001), which reduced ring elastin content to less than 1% (Figure 3e). These changes contributed to increases in the average initial gauge length with elastase treatment from 17.3 ± 2.35 and 19.1 ± 1.25 mm in 9- and 27- hour control groups, respectively, to 35.6 ± 3.43 mm in the 27-hour treated group (*** P=0.0009, ** P=0.008) (Figure 4h). Increases in sample thickness and initial gauge length exhibited by samples exposed to longer enzyme incubation times contributed to increases in the force and displacement induced per unit stress and strain calculated for these samples in subsequent analyses.

3.3. Impact of Elastase Incubation on the Tensile Properties of Aortic Rings

Control rings at 9- and 27-hours demonstrated the J-shaped stress-strain behaviour characteristic of arterial tissues (Figure 5a,b). The 9-hour elastase treatment group demonstrated a decrease in resistance to initial deformation evidenced by a flattening of the initial portion of the curve followed by more rapid stiffening (Figure 5c). Rings treated for 18- and 27-hours (Figure 5d,e) further demonstrated a decrease in initial stiffness and increase in the final slope of the curve where stiffening occurred at lower strains in these groups. The strain at which stiffening occurred varied widely in the 18-hour elastase treated group (Figure 5d). The 27-hour elastase treated group, which contained virtually no elastin, as seen in Figure 3c, demonstrated very stiff behaviour almost immediately upon sample loading. Of note, the small flat region at the beginning of these curves is consistent with the 10% strain range over which samples were preconditioned prior to testing as highlighted in Supplementary Figure 1. Overall, elastase incubation induced progressive decreases in the initial resistance to deformation followed by progressively earlier and more rapid stiffening.

Quantitative analyses of stress-strain behaviour in the initial and final regions of the stress-strain curves in Figure 5 are shown in Figure 6. First, the initial tangent moduli confirmed observations of progressive decreases in resistance to initial deformation upon loading with decreasing elastin content up to the 18-hour incubation time. Notably, the 27-hour control group demonstrated increased initial moduli compared to the 9-hour control group, further highlighting that the decreases seen in the elastase treated groups are due to elastin depletion. The final tangent moduli saw significant increases with increasing elastase incubation time from 0.71 ± 0.114 and 0.88 ± 0.133 kPa in the 9- and 27-hour controls, respectively, to 2.77 ± 0.812 kPa in the 27-hour treatment group. Elastase treatment induced time-dependent decreases in initial stiffness and concomitant increases in final stiffness in rings of porcine aorta.

Treatment with elastase resulted in a notable reduction in the strain level achieved at the maximum stress applied of at 300 kPa, particularly in the 18- and 27-hour elastase treated samples (Figure 7a). However, analysis of the pin-to-pin distance (proxy for circumferential length) at the maximum stress revealed that elastase treatment caused a significant increase in circumferential length: from 31.7 ± 3.69 (9-hour control) and 32.6 ± 1.31 mm (27-hour control) to 41.9 ± 3.49 mm in the 9-hour elastase treatment group, which was also present for the longer incubation times (Figure 7b). As such, while near complete elastin degradation caused stiffening to occur at lower strains, elastase-treated rings demonstrated significantly increased initial sample gauge lengths (Figure 4h) resulting in these stiffer samples actually reaching higher final lengths at 300 kPa compared to the control samples. These findings highlight that progressive elastase treatment can be used to produce porcine aortic tissues with increased arterial diameter and progressive increases in stiffness across a given stress range (0.25-0.3 MPa in this instance). Thus, there exists a threshold level of elastin degradation at which porcine aortic tissues demonstrate a greater arterial diameter and increased mural stiffness similar to those seen in AA under physiological stresses.

3.4. Ex Vivo Porcine Model with Localised Elastin Degradation

Spatially controlled elastin degradation of a 1 cm portion of 6 cm long aortic segments was produced by localised exposure to elastase. Representative Verhoeff’s elastin-stained sections of rings cut at proximal, middle (treated) and distal positions along the vessel length for one aorta from each group are shown in Figure 8. These demonstrate spatially controlled elastin degradation which was limited to the treated segment of the aortae (Figure 8d–f). All other samples in the control and treatment groups can be seen in Supplementary Figure 2 and Figure 3. The mid-sections (treated) of the elastase-treated aortae (Figure 8e) exhibit less elastin degradation compared to the aortic rings (Figure 3d) exposed to the same 9-hour treatment protocol.

3.5. Inflation-Extension of the Localised Elastin-Degradation ex Vivo Aneurysm Model

The spatially controlled elastin degradation produced a localised increase in aortic dilation, but no significant change in stiffness under physiological pressures (90-120 mmHg) (Figure 9). The elastase-treated region of the aortae demonstrated a small but significant (* P=0.047) increase in absolute dilation compared to control-treated aortae, Figure 9a. Furthermore, the treated regions of elastase-treated aortae demonstrated a significantly (* P = 0.017) increased relative dilation (diameter normalised with respect to the untreated proximal and distal regions) compared to that of the control aortae, Figure 9b. However, this increased dilation did not yield any significant change in physiological arterial compliance (the change in absolute dilation across 90-120 mmHg), Figure 9c.

This increased dilation is consistent with the mechanical data from the ring tests in which elastase treatment produced increases in arterial diameter under the same applied tension. However, the lack of change in stiffness indicates that the increase in aortic deformation produced at this level of elastin degradation was below the threshold at which strain-induced stiffening occurs. Overall, elastase treatment was still able to induce localised changes in pressure-diameter behaviour which are consistent with those seen in the ring analysis.

4. Discussion

Elastase treatment induced increases in the circumferential length and tensile modulus in rings of porcine aortae when compared to untreated controls, similar to that seen in AA. Using insight from these elastase-treated ring samples, we established a local elastin degradation protocol for whole aortae which caused localised dilation under physiological pressures, similar to that seen in AA. These findings highlight a promising avenue to use localised elastin degradation to produce an ex vivo aneurysm model.

Elastase-treated aortic rings demonstrated significant increases in circumferential length both in the unloaded state (Figure 4h) and under 300 kPa Cauchy stress (Figure 7b), results which are consistent with findings from other groups) [12,29,30,31,43]. In healthy tissues, resistance to deformation is provided by the initial recruitment of pliable elastin fibres followed by the progressive recruitment of stiff collagen fibres with increasing strain [6,44]. This strain-dependent recruitment is permitted by the wavy ‘crimped’ architecture of groups of collagen fibres allowing them to be ‘pulled taut’ and thus recruited at specific levels of tissue deformation [12,45]. Using a multiphoton microscopy approach, Chow et al. found that elastase-induced increases in the loaded length of porcine aortic tissues were attributed to increases in collagen fibre uncrimping with respect to untreated controls under the same conditions of stress similar to that seen in mechanical loading in untreated tissues [12,45]. As such, increases in loaded length result from the increased deformation required to recruit sufficient collagen fibres to make up for decreases in elastin mediated resistance [12,43]. Additionally, the increases in unloaded circumference seen in the highly depleted elastin aortic rings indicates that collagen crimp in unloaded arterial tissues is maintained by a compressive force imposed by the elastic fibres in the resting state [6,30,43]. As such, elastase-mediated increases in unloaded length secondary to a loss in elastin-mediated compression act similar to an ‘initial stretch’ on the collagenous architecture in the unloaded state compared to untreated tissues. This stretching nature of elastase-induced increases in unloaded circumferential length is recognised by groups such as Chow et al. in porcine aortae and Fonck et al. in rabbit carotids. [31,43]. Overall, increases in aortic circumferential length due to elastase treatment in both loaded and unloaded configurations are attributed to the increased stretch, and thus decreased crimp, exhibited by collagen fibres under the same load in elastin depleted tissues. This work builds upon existing literature to highlight the role of elastic fibres in mediating vessel circumference and compliance in loaded and unloaded states through their role in preserving collagen crimp.

Elastase treatment induced increased compliance under low stress, and increased stiffness under high stress in rings of porcine aorta. Chow et al. found that the tensile properties of porcine aortae incubated in elastase solution progressed through distinct sequential stages of ‘initial softening’ characterised by a flattening of the initial region of the curve; ‘extensible but stiff’ behaviour characterised by an increase in the slope of the final region; and ‘collagen scaffold’-like behaviour characterised by immediate stiffening upon loading with increasing elastase incubation time similar to that seen in our study [31]. Furthermore, these findings are in line with those of Fonck et al. who report that elastase-treated rabbit carotids were more pliable under low circumferential stretch and less pliable under large circumferential stretch [43]. Decreases in initial stiffness are attributed to reductions in the number and integrity of elastic fibres present to provide resistance to initial deformation with increasing elastase incubation time [6,31].

However, while elastase treatment induced significant increases in loaded and unloaded circumferential length, measured increases in circumferential length fall short of the 50% dilation exhibited in human aneurysms [1]. Kratzberg et al. reported that application of a creep loading protocol produced a further plastic deformation of 23.6 ± 10.5% and 22.5 ± 7.1% in partially and completely elastin-depleted strips of porcine aorta, respectively [30]. Perhaps application of a similar creep loading approach would increase dilation to similar levels in the ex vivo porcine aortae to yield the 50% dilation pathognomonic of human aneurysms.

Elastin degradation represents a feasible method to produce porcine aortic tissues with similar compliance to human AA. Significant heterogeneity with regards to testing and analysis methodologies used to quantify the tensile properties of human AA in the literature limits direct comparison of measured values [46]. Individual studies indicate that human TAA and AAA demonstrate respective increases in circumferential stiffness of 44% and 60% compared to healthy human aortae under the same degree of stress [8,47]. Notably, a comparative analysis conducted by de Beaufort et al. found that healthy porcine aortae demonstrated similar compliance to aortae from young humans under 100 mmHg of pressure, but only 41% of the stiffness found in aortae from humans over 60 years of age [21] – in which aneurysms most frequently occur [48]. This indicates that human AAA and TAA may have 350 and 390% greater circumferential stiffness, respectively, compared to healthy porcine aortae. Elastase treatments conducted as part of this study produced dramatic increases in the circumferential stiffness of rings of porcine aorta with rings in the 27-hour treatment group demonstrating a 380% increased circumferential stiffness compared to time-matched controls over a Cauchy stress range of 250-300 kPa. A follow-up study including mechanical assessments of aneurysmal human and elastin-depleted porcine aortae would allow optimisation of elastase incubation time to produce porcine aortic models with similar stiffness to that seen in human aneurysms.

The model produced and characterised as part of this work demonstrates several key advantages. First, the proposed model meets an established need for an ex vivo large animal model representative of the geometric and compliance features of aneurysms suitable for endovascular device testing. It represents a comprehensive reproduction of the mechanical features of aneurysm and its use could facilitate better testing and optimisation of endovascular devices prior to in vivo testing. Second, the produced model can be directly assembled from widely available low-cost parts allowing for rapid production and use. Lastly, this approach can produce models of aneurysms demonstrating distinct geometric or anatomical features, e.g., aortic root or arch aneurysms, facilitating the improvement of endovascular treatments in these areas [5]. Overall, our proposed model demonstrates clear advantages: it replicates the mechanical features of AA better than existing in vitro models; requires minimal skill and resource requirements; and has the potential for customisability to specific aneurysm pathologies.

However, the produced model requires further optimisation prior to its progression in the testing of endovascular devices. The degree of elastin degradation produced in whole aortae as part of this work resulted in an approximate 5% increase in dilation and no change in compliance under physiological pressures compared to controls. Human aneurysms demonstrate a 50% increased diameter [1] and a 44% and 60% increased circumferential stiffness at thoracic or abdominal locations, respectively, compared to non-aneurysmal aortae [8,47]. While our findings are consistent with the level of degradation seen, a larger degree of elastin degradation is needed to produce a more physiologically relevant model. Furthermore, the 1 cm length utilized as part of this study is not representative of the lengths of aneurysmal dilation in human aneurysms [49]. Reproduction of our localised degradation over a longer time period and larger aortic segment, would produce a model more robust for use in endovascular device testing.

This work is subject to several methodological limitations. First, degradation produced by enzyme soaking techniques was mediated by enzyme diffusion into the tissues resulting in the production of tissues with inhomogeneous elastin degradation and this limited the direct translation of protocols from aortic rings to the whole aorta model. This method was selected for use in this study despite these limitations due to its successful use in relevant background literature to produce the desired changes in arterial geometry and compliance [12,30,31,43]. Future studies could assess the role of hydrostatic pressurisation, or axial strains to reduce the time, and improve the homogeneity of elastin degradations [50,51]. The formation of a translucent gelatinous precipitate on elastin-depleted rings impacted the accuracy with which unloaded sample dimensions could be measured optically. This finding has been reported by other groups and was minimised in this study by imaging each ring against a high contrast graph paper background and with the light source at different positions [31,52]. Histomorphometric quantifications of constituent composition are sensitive to variations in selected image analysis parameters and component staining lending a subjective component to obtained values [53]. To maximise measurement objectivity, all samples in this study were prepared and analysed in the same manner [53]. Use of this method in this study offered several advantages including semi-quantitative assessment of the time-degradation relationship, the variability of produced degradations, and is consistent with work performed by other groups in this field [7,8,14,54]. Use of pins with a smaller radius than the vessel thickness may have contributed to stress concentrations at the pins predisposing to tissue failure at these points [55]. As such, the failure characteristics of analysed tissues were not assessed as part of this study. Lastly, this study utilised measures of average strain such as crosshead displacement and video extensometry in ring tensile and inflation extension testing respectively, rather than more resource-intensive measures of local strain such as those employed by Shazly et al. and Lane et al. respectively [32,56]. This work chose to measure average strain as this was appropriate to the characterisation of a segmentally homogenous approximate model of overall aneurysmal pressure-dilation behaviour [38]. Furthermore, cross-head displacement and video extensometry are validated accurate measures of average strain and are utilised widely by other groups [7,38,41].

Endovascular repair represents the preferred method of surgical repair for AA, due to reduced perioperative morbidity rates and more rapid recovery compared to open surgical techniques. The development of better in vitro models represents a useful tool to allow refinement of device designs to optimise these factors prior to costly in vivo trials. Of these, ex vivo animal models allow for the most comprehensive representation of aortic mechanics replicating both the non-linear stress-strain and material failure behaviours such as endothelial injury, aortic dissection, and rupture, all of which are not possible with synthetic models [10,57]. However, as of yet, no model representative of the dilation and stiffening seen in AA exists on a scale suitable for ex vivo endovascular device testing. Our ex vivo model using whole porcine aortae with localised elastin degradation represents a promising avenue to address this need. Our model could facilitate new device designs leading to further reductions in the morbidity and mortality of AA through improvements in endovascular device applications and performance.

5. Conclusions

This work assessed the feasibility of a localised elastin degradation approach in porcine aortae to produce an ex vivo model of AA suitable for testing of endovascular devices. Elastase treatment of aortic rings produced time-dependent decreases in elastin content and resulted in significant changes in morphological and tensile properties. Elastase-treated rings demonstrated increases in loaded circumferential length and mural stiffness similar to those seen in human aneurysms. Following from this, elastase treatment was successfully localised to segments of whole porcine aortae. Elastase treatment produced significant changes in local pressure-diameter behaviour compared to untreated aortae which was consistent with the level of degradation produced. Overall, a localised elastin degradation of porcine aortic segments represents a feasible method to produce aortae which better exhibit the local dilation and stiffening seen in human aneurysms than what is currently seen in existing models. Use of our physiologically relevant ex vivo model of AA in the design and optimisation of endovascular devices has the potential to alter the current benchtop test beds for such devices and ultimately contribute to reductions in the morbidity and mortality of AA.