Captopril treatment during development alleviates mechanically induced aortic remodeling in newborn elastin knockout mice

Abstract

Deposition of elastin and collagen in the aorta correlate with increases in blood pressure and flow during development, suggesting that the aorta adjusts its mechanical properties in response to hemodynamic stresses. Elastin knockout (Eln^−/−) mice have high blood pressure, pathological remodeling of the aorta, and die soon after birth. We hypothesized that decreasing blood pressure in Eln^−/− mice during development may reduce hemodynamic stresses and alleviate pathological remodeling of the aorta. We treated Eln^+/+ and Eln^−/− mice with the anti-hypertensive medication captopril throughout embryonic development and then evaluated left ventricular (LV) pressure and aortic remodeling at birth. We found that captopril treatment decreased Eln^−/− LV pressure to values near Eln^+/+ mice and alleviated the wall thickening and changes in mechanical behavior observed in untreated Eln^−/− aorta. The changes in thickness and mechanical behavior in captopril-treated Eln^−/− aorta were not due to alterations in measured elastin or collagen amounts, but may have been caused by alterations in smooth muscle cell (SMC) properties. We used a constitutive model to understand how changes in stress contributions of each wall component could explain the observed changes in composite mechanical behavior. Our modeling results show that alterations in the collagen natural configuration and SMC properties in the absence of elastin may explain untreated Eln^−/− aortic behavior and that partial rescue of the SMC properties may account for captopril-treated Eln^−/− aortic behavior.

1. Introduction

During aortic development, hemodynamic forces, such as blood pressure and flow, increase concurrently with extracellular matrix (ECM) deposition (Wagenseil and Mecham 2009). ECM proteins, including elastin and collagen, provide elasticity and strength for proper cardiovascular function. When the balance of ECM proteins is upset, such as in genetic mutations that alter the amount or organization of ECM components, mechanical behavior of the aortic wall is changed, leading to altered mechanical stimuli on the smooth muscle cells (SMCs) within the wall. Pathological adaptation to these changes can lead to disease (Humphrey et al. 2014).

Supravalvular aortic stenosis (SVAS) is a genetic disease associated with elastin mutations and functional elastin haploinsufficiency (Li et al. 1997). The cardiovascular manifestation is a narrowed aortic lumen that increases the risk for sudden cardiac death (Wessel et al. 1994). Elastin knockout (Eln^−/−) mice represent a severe model of SVAS that do not survive beyond birth (Li et al. 1998). Eln^−/− mice have hypertension and the aorta is characterized by smooth muscle cell (SMC) overproliferation, luminal narrowing, and increased stiffness (Li et al. 1998; Wagenseil et al. 2009), which are all similar to the human disease phenotype. Previous studies on Eln^−/− mice at embryonic day (E) 18 and postnatal (P) day 1, demonstrate that increased aortic stiffness occurs first (at E18) (Wagenseil et al. 2010), followed by increased blood pressure, SMC overproliferation, and additional wall remodeling (at P1) (Wagenseil et al. 2009), suggesting that increased aortic stiffness due to the absence of elastin may be the initiating signal in downstream pathological events (Cheng and Wagenseil 2012).

As there are not yet any pharmaceutical approaches to increase the amounts of crosslinked elastin within the arterial wall in developing mice (Knutsen et al. 2018), we reasoned that pharmaceutically inhibiting the resulting increase in blood pressure may alleviate the additional wall remodeling observed in Eln^−/− aorta and limit the pathological effects of elastin insufficiency. We chose captopril, an angiotensin converting enzyme (ACE) inhibitor that has been shown to lower blood pressure in maturing mice (Huang et al. 2013) as our anti-hypertensive agent. We treated pregnant dams with or without captopril and then evaluated blood pressure and aortic remodeling, gene expression, and mechanical behavior in the newborn pups. We used a constitutive model to understand how the absence of elastin and resulting changes in collagen and SMC behavior could explain the resulting composite mechanical behavior of the aortic wall. Our results show that captopril-treatment during development alleviates mechanically induced aortic remodeling in newborn Eln^−/− mice. We suggest future research on blood pressure regulation as a treatment option to manipulate developmental aortic remodeling.

2. Methods

2.1. Animals and anti-hypertensive treatment

Eln^+/− mice (Li et al. 1998) were bred to produce one day old Eln^−/− and Eln^+/+ pups. Pregnant dams were treated with captopril, an anti-hypertensive that works through ACE inhibition, in drinking water (0.075 g/L) (Huang et al. 2013) or left untreated. All pups were sacrificed within 24 hours after birth by thoracotomy under 2.5% isoflurane or CO₂ inhalation. Ascending aortic segments from the root to the brachiocephalic artery were used for all aortic studies. Animals were used for blood pressure (N = 7 – 9 per group), protein quantification (N = 3 per pooled group of 5 samples), mRNA expression (n= 3 per pooled group of 8 – 10 samples for Agtr1a and Agtr2, n = 5 – 6 per group for all other genes), histology and image analysis (N = 6 – 8 per group), aortic dimensions (N = 8 – 15 per group), and mechanical testing (N = 5 – 8 per group). Sex of the animals was not determined. All animal protocols were approved by the Institutional Animal Care and Use Committee at Washington University.

2.2. Left ventricular blood pressure measurement

Newborn pups were anesthetized with 1.5% isoflurane. Body temperature was maintained with a heated stage and radiant lighting. Left ventricular (LV) pressure was measured using a fluid-filled 30G needle under ultrasound guidance (Vevo 2100, Visualsonics), as previously described (Le et al. 2012). The systolic LV pressure was calculated from the mean LV pressure, assuming that the LV diastolic pressure was zero.

2.3. Protein quantification

Five aortic segments per group were pooled and used to measure desmosine (an elastin-specific crosslink), hydroxyproline (an amino acid abundant in collagen), and most amino acids (total protein) as described in Kim et al. (2017). Aortic segments were stored at −80°C and then processed as a batch. Samples were hydrolyzed in 6 N HCl, dried, resuspended in ultrapure water, and stored at −20°C until use. Desmosine was measured through a competitive ELISA (Cheng et al. 2013). Hydroxyproline was measured through a Chloramine T reaction (Jamall et al. 1981). Total protein was measured through a ninhydrin assay (Starcher 2001). Desmosine and hydroxyproline amounts were normalized to total protein for each group.

2.4. mRNA expression

The expression of Agtr1a and Agtr2, the main angiotensin II (Ang II) receptor genes in the adult mouse aorta (Burson et al. 1994; Habashi et al. 2011), were measured to determine changes in expression of Ang II signaling genes resulting from captopril treatment. Additionally, the expression of Mki67 was measured as a marker of cell proliferation (Scholzen and Gerdes 2000) and Acta2, Cnn1, Myh11, Smtn, and Tagln were measured as markers of SMC phenotype (Rensen et al. 2007). Aortic segments were flash-frozen in liquid nitrogen and stored at −80° until use. For Agtr1a and Agtr2, 8 – 10 samples were pooled and three pooled samples were used for each group. Total RNA was extracted using a Bullet Blender (Next Advance) tissue homogenizer and the RNeasy Plus Mini Kit (Qiagen). For all other genes, a TissueLyser LT (Qiagen) was used for tissue homogenization and TRIzol Reagent (Invitrogen) for total RNA extraction, which provided a higher yield and did not require pooling of samples. For these genes, 5 – 6 aortic samples were used per group. Reverse transcription was performed using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems) and 180 ng of total RNA. qPCR was performed with TaqMan Fast Advanced Master Mix (Applied Biosystems) on a QuantStudio 12k machine and analyzed with accompanying software (Applied Biosystems). All experiments were run in triplicate using 2 μl (Agtr1a and Agtr2) or 1.2 μl (all other genes) of cDNA in a total volume of 20 μl. Mean C_t of each replicate was normalized to the average of β−2 microglobulin (β2M). Primers were Taqman Gene Expression Assays (Applied Biosystems). Fold change was calculated via the 2^−ΔΔCT method.

2.5. Histology and image analysis

Aortic segments were dissected, fixed in 10% formalin for 24 hours, and then transferred to 70% ethanol. They were sequentially dehydrated, embedded in paraffin, and cut into 5 μm sections. The sections were stained with hematoxylin and eosin (H&E) and imaged for general cellular morphology.

Color H&E images were converted to grayscale and manually thresholded to highlight cell nuclei. Three regions of the wall that showed a clear cross-section were outlined for each image and cell nuclei were counted using Fiji software (NIH). Cell density was determined by dividing the total nuclei count by the total outlined area for each aortic section. Cell shape was determined using the “Analyze Particles” macro in Fiji for the entire thresholded aortic section. Roundness (4*area/(π*major_axis^{^}2) of the fitted ellipse for each cell nucleus (varying from 257–1029 nuclei per image) was calculated. Roundness is the inverse of the aspect ratio and values vary from 0 (line) to 1 (circle). Roundness measurements were combined to generate a total histogram for each group.

2.6. Aortic dimensions and circumferential residual strain

An image was taken with a digital camera connected to a dissecting microscope to measure the length of the ascending aorta in vivo. After mechanical testing each aortic segment was placed in physiologic saline solution, cut into 150- to 250-μm thick rings, and imaged to obtain the unloaded diameter and thickness. A radial cut was then made in each ring to determine the residual strain as measured by the opening angle (Chuong and Fung 1986).

2.7. Mechanical testing and data analysis

Aortic segments were dissected and stored at 4°C in physiologic saline solution for up to 3 days before testing (Amin et al. 2011). Passive inflation tests were performed using a Myograph 110P (Danish Myotechnology), as previously described (Kim et al. 2017). The aorta was mounted on custom-designed stainless steel cannulae in a 37°C physiologic saline solution bath with 7–0 silk suture. The unloaded length of the aorta from suture to suture was measured with a caliper. The aorta was extended in the longitudinal direction to the approximate in vivo stretch ratio (λ_z = 1.1) (Kim et al. 2015) and preconditioning cycles were performed until the aorta exhibited repeated behavior for multiple pressurization cycles. The aorta was then pressurized from 0 to 60 mmHg in 5 mmHg step increments, holding for 8 seconds at each step. Three of these cycles were performed while recording outer diameter and lumen pressure at 2 Hz. Average data for each pressure step were determined by averaging all points within ± 1 mmHg of the target pressure. Due to collapse of the aorta at 0 mmHg, data between 5- and 60 mmHg were used for analysis.

Data from the last inflation cycle were used for further analyses. The mean circumferential stretch (λ_θ) at each pressure step was calculated by,

$${\lambda }_{\theta }=\frac{1}{2}\left(\frac{r_i}{R_i}+\frac{r_o}{R_o}\right),$$ Eqn. 1

where r is the loaded radius, R is the unloaded radius and subscripts i and o indicate inner and outer dimensions, respectively. The deformed inner radius was calculated assuming incompressibility (Faury et al. 1999). The mean circumferential stress (σ_θ) at each pressure step was calculated by,

$${\sigma }_{\theta }=\frac{{\text{ Pr }}_i}{r_o-r_i},$$ Eqn. 2

where P is the applied pressure. The stress-stretch curves were fit to,

$${\sigma }_{\theta }=b_1+b_2\text{exp}\left(\frac{{\lambda }_{\theta }{}^{b_3}}{b_4}\right),$$ Eqn. 3

where b_i are constants determined by regression in Matlab (Mathworks) (Staiculescu et al. 2017). The tangent physiologic modulus (E_phys) was calculated by,

$$E_{phys}=\frac{\partial {\sigma }_{\theta }}{\partial {\lambda }_{\theta }}{|}_{P=P_{phys}},$$ Eqn. 4

where P_phys = average LV pressure for each group. Any aortic segments that had leaks or did not show repeated behavior for multiple pressurization cycles were excluded from the mechanical testing data, but were used for unloaded dimension and residual strain measurements.

2.8. Constitutive modeling

Constitutive modeling followed previous approaches (Aars 1971; Gleason et al. 2004; Wagenseil 2011). The aortic wall was modeled as a constrained mixture where the stretch ratios for the entire mixture were calculated by Eqn. 1 above for the circumferential direction and by Eqn. 5 below for the axial direction,

$${\lambda }_z=\frac{l}{L},$$ Eqn. 5

where l is the in vivo length and L is the unloaded length. Radial stretch was calculated by incompressibility. It was assumed that each component has a natural configuration in the unloaded state (λ_i0) so that the actual stretch ratio of each component in the mixture can be calculated by,

$${\lambda }_i^n={\lambda }_{i0}^n*{\lambda }_i,$$ Eqn. 6

where i = r, θ, or z for the radial, circumferential, and axial directions, respectively, and n represents elastin (e), collagen (c), or SMCs (m) (Fig. 1).

Figure 1.

Depiction of the natural configurations (λⁿ_θ0 , λⁿ_z0) for each component (n = Elastin, Collagen, or SMCs) in the aortic wall. Relationships for the component natural configurations to the component stretch ratios (λⁿ_θ , λⁿ_z) and the stretch ratios for the composite mixture (λ_θ , λ_z) are shown, as defined in Eqn. 6.

The total stress in each direction $\left({\sigma }_i^{total}\right)$ is the sum of each constituent’s stress $\left({\sigma }_i^n\right)$ multiplied by its volume fraction (ϕⁿ),

$${\sigma }_i^{total}=\sum _n{\varphi }^n{\sigma }_i^n,$$ Eqn. 7

Constituent stresses were calculated in the absence of shear and assuming incompressibility. Elastin was considered an isotropic, neo-Hookean solid (Gleason et al. 2004),

$${\sigma }_{\theta }^e\left({\lambda }_{\theta }^e,{\lambda }_z^e\right)=2{\lambda }_{\theta }^e{b}_1\left(1-\frac{1}{{\lambda }_{\theta }^4{\lambda }_z^{e^2}}\right),{\sigma }_z^e\left({\lambda }_{\theta }^e,{\lambda }_z^e\right)=2{\lambda }_z^e{b}_1\left(1-\frac{1}{{\lambda }_{\theta }^2{\lambda }_z^{c^4}}\right),$$ Eqn. 8a,b

where b₁ is a material constant. Collagen was modeled as a Fung-type, transversely isotropic material with z as the preferred direction (Gleason et al. 2004),

$${\sigma }_{\theta }^c\left({\lambda }_{\theta }^c,{\lambda }_z^c\right)=2{\lambda }_{\theta }^c{b}_2{b}_3\left(1-\frac{1}{{\lambda }_{\theta }^{c^4}{\lambda }_z^{c^2}}\right)\text{exp}\left(Q^c\left({\lambda }_{\theta }^c,{\lambda }_z^c\right)\right),$$ Eqn. 9a

$${\sigma }_z^c\left({\lambda }_{\theta }^c,{\lambda }_z^c\right)=2{\lambda }_{\theta }^c{b}_z\left[b_3\left(1-\frac{1}{{\lambda }_{\theta }^{c^2}{\lambda }_z^{c^4}}\right)+2b_4\left({\lambda }_z^{c^2}-1\right)\right]\text{exp}\left(Q^c\left({\lambda }_{\theta }^c,{\lambda }_z^c\right)\right),$$ Eqn. 9b

$$\text{with}{Q}^c\left({\lambda }_{\theta }^c,{\lambda }_z^c\right)=b_3\left({\lambda }_{\theta }^{c^2}+{\lambda }_z^{c^2}+\frac{1}{{\lambda }_{\theta }^{c^2}{\lambda }_z^{c^2}}-3\right)+b_4{\left({\lambda }_z^{c^2}-1\right)}^2,$$ Eqn. 9c

where b₂ – b₄ are material constants. Passive SMCs were modeled as a transversely isotropic material with a linear term and an exponential term in the θ direction (Gleason et al. 2004),

$${\sigma }_{\theta ,pas}^m\left({\lambda }_{\theta }^m,{\lambda }_z^m\right)=2{\lambda }_{\theta }^{m^2}\left[b_5\left(1-\frac{1}{{\lambda }_{\theta }^{m^4}{\lambda }_z^{m^2}}\right)+2b_6{b}_7\left({\lambda }_{\theta }^{m^2}-1\right)\right]\text{exp}\left(Q^m\left({\lambda }_{\theta }^m\right)\right),$$ Eqn. 10a

$$\text{with}{Q}^m\left({\lambda }_{\theta }^m\right)=b_7{\left({\lambda }_{\theta }^{m^2}-1\right)}^2,$$ Eqn. 10b

$${\sigma }_{z,pas}^m\left({\lambda }_{\theta }^m,{\lambda }_z^m\right)=2{\lambda }_z^{m^2}{b}_5\left(1-\frac{1}{{\lambda }_{\theta }^{m^2}{\lambda }_z^{m^4}}\right),$$ Eqn. 10c

where b₅ – b₇ are material constants.

Volume fractions were taken from Wagenseil (2011) for a 3-day old mouse aorta, assuming that 75% of the wall is made up of water. Natural configurations were chosen with elastin under slight compression and SMCs in the unloaded state based on observations from the mechanical behavior in Fig. 7 and differences in the timing of elastin deposition and growth in the newborn compared to the adult aorta (Davis 1995; Kelleher et al. 2004). The natural configurations for collagen were calculated from equilibrium of the unloaded mixture. Given the assigned mass fractions and natural configurations, material constants were prescribed based on previous work (Gleason et al. 2004; Wagenseil 2011) and scaled to produce reasonable contributions for each component to approximate the circumferential mechanical behavior of the newborn Eln^+/+ untreated aorta. For Eln^−/− aorta, the elastin volume fraction is zero and water takes up a larger portion of the wall. The collagen natural configurations were then recalculated from equilibrium. Eln^−/− aorta was modeled with and without changes to the SMC material constants. The constants used for constitutive modeling are shown in Table 1.

Figure 7.

Circumferential mechanical behavior of the aorta from newborn Eln^−/− and Eln^+/+ mice treated with (Cap) and without (Unt) captopril during development. Unt Eln^−/− aorta has reduced outer diameter compared to Unt Eln^+/+ at pressures between 5 – 25 mmHg ($) (A). The reduced outer diameter in Eln^−/− aorta is increased at 5 and 10 mmHg after captopril treatment (#). Circumferential stretch-stress behavior shows the rightward shift (arrow) in Unt Eln^−/− aorta that appears to be rescued with captopril treatment (B). $ = P < 0.05 between Unt Eln^+/+ and Eln^−/−. # = P < 0.05 between Unt Eln^−/− and Cap Eln^−/−. Data are plotted as mean ± sem and lines are interpolated between measured data points. Tangent physiologic modulus was calculated at the mean LV pressure for each group (C). Each data point represents one aorta and lines are mean ± sd. N = 7 for Unt Eln^−/−, N = 5 for Cap Eln^−/−, N = 8 for Unt Eln^+/+, and N = 8 for Cap Eln^+/+.

Table 1.

Prescribed parameters for constitutive modeling. Natural configurations (Fig. 1) for elastin and SMCs were assumed based on observations from the mechanical behavior and differences in the timing of elastin deposition and growth in the newborn compared to the adult aorta. Collagen values were calculated by equilibrium in the unloaded state. Collagen natural configurations change for the Eln^−/− case when elastin is not present. Volume fractions were taken from Wagenseil (2011) for a 3-day old mouse aorta, assuming that 75% of the wall is made up of water. Elastin volume fraction is zero for the Eln^−/− case and water takes up a larger portion of the wall. Material properties were scaled from previous work (Gleason et al. 2004; Wagenseil 2011) to approximate the behavior of Eln^+/+ newborn aorta in the circumferential direction and to provide reasonable contributions for each wall component (Fig. 8A). For the first Eln^−/− case, no material properties were changed (Fig. 8B). For the second Eln^−/− case, SMC material constants were reduced, as suggested by experimental data (Espinosa et al. 2013; Karnik et al. 2003) (Fig. 8C). Changes compared to the baseline case are highlighted in bold text. e = elastin, c = collagen, m = SMCs. Constants defined in text and equations 6 – 10.

Group, Model Figure	Natural configurations	Volume fractions	Material properties
Eln+/+ Baseline, Fig. 8A	λeθ0 = 0.95, λez0 = 0.95	ϕe = .058	b1 (e) = 5 kPa
	λcθ0 = 1.31, λcz0= 1.30	ϕc = .019	b2 (c) = 5 kPa
	λmθ0 = 1.00, λmz0 = 1.00	ϕm = .173	b3 (c) = 0.5
			b4 (c) = 0.001
			b5 (m) = 8 kPa
			b6 (m) = 8 kPa
			b7 (m) = 0.05
Elr−/− No elastin, Figs. 8B and 8D	λeθ0 = N/A, λez0 = N/A	ϕe = 0	b1 (e) = N/A
	λmθ0 = 1.00, λmz0= 1.00	ϕc = .019	b2 (c) = 5 kPa
	λcθ0 = 1.00, λcz0 = 1.00	ϕm = .173	b3 (c) = 0.5
			b4 (c) = 0.001
			b5 (m) = 8 kPa
			b6 (m) = 8 kPa
			b7 (m) = 0.05
Elr−/− No elastin and reduced SMC linear constants, Figs. 8C and 8D	λeθ0 = N/A, λez0 = N/A	ϕe = 0	b1 (e) = N/A
	λmθ0 = 1.00, λmz0= 1.00	ϕc = .019	b2 (c) = 5 kPa
	λcθ0 = 1.00, λcz0 = 1.00	ϕm = .173	b3 (c) = 0.5
			b4 (c) = 0.001
			b5 (m) = 4 kPa
			b6 (m) = 4 kPa
			b7 (m) = 0.05

2.9. Statistical analysis

Data are presented as mean ± sd for all figures except the mechanical testing data. Mechanical testing data in Figs. 7A and B are presented as mean ± sem for clarity. Significant differences were determined by analysis of variance (ANOVA), followed by Fisher’s LSD for multiple comparisons in Graphpad Prism software. For the cell roundness measurements, visual examination of the histograms showed non-normal distributions, so the Mann-Whitney U test was used to determine mean rank differences between groups in Matlab software. All P values for comparisons between genotype and treatment for P < 0.1 are shown in the figures. P < 0.05 was considered significant.

3. Results

3.1. Captopril treatment during development decreases LV blood pressure in newborn mice

Captopril treatment decreases LV pressure from 41 to 36 mmHg in Eln^−/− mice and from 31 to 27 mmHg in Eln^+/+ mice (Fig. 2A). Heart rate is 12% higher in captopril treated Eln^−/− mice compared to the untreated group (Fig. 2B), but there is no significant relationship between heart rate and blood pressure (R² = .027, P = 0.33, not shown). The pressure measurements confirm that Eln^−/− mice have high LV pressure at birth and that captopril treatment during development reduces the LV pressure to the same level as untreated Eln^+/+ mice.

Figure 2.

LV blood pressure (A) and heart rate (B) in newborn Eln^−/− and Eln^+/+ mice treated with (Cap) and without (Unt) captopril during development. Each data point represents one animal. Lines are mean ± sd.

3.2. Captopril treatment decreases wall thickness and opening angle in Eln^−/− aorta

Captopril treatment during development does not affect the ascending aortic length (Fig. 3A) or unloaded outer diameter (Fig. 3B) of newborn mouse aorta. Eln^−/− aorta is 40% longer than Eln^+/+, with no significant difference in unloaded outer diameter for the untreated groups. After captopril treatment, Eln^−/− aorta has a 9% smaller outer diameter than Eln^+/+. Eln^−/− aorta has increased wall thickness (Fig. 3C) and decreased opening angle (Fig. 2D) compared to Eln^+/+ aorta. Captopril treatment reduces the unloaded wall thickness and opening angle in Eln^−/− aorta only (Fig. 3C and D).

Figure 3.

Aortic dimensions and residual strain in newborn Eln^−/− and Eln^+/+ mice treated with (Cap) and without (Unt) captopril during development. In vivo length (A), unloaded outer diameter (B), unloaded thickness (C), and opening angle (D) as a measure of residual strain are shown. Each data point represents one animal. Lines are mean ± sd.

3.3. Alterations in ECM protein amounts or Ang II receptor gene expression do not account for remodeling of Eln^−/− aortic wall in response to captopril treatment

Amounts of desmosine and hydroxyproline are not affected by captopril treatment (Fig. 4), indicating that changes in thickness and opening angle in Eln^−/− aorta are not due to altered deposition of crosslinked elastin or total collagen. As expected, there is no crosslinked elastin present in Eln^−/− aorta.

Figure 4.

Aortic desmosine (A) and hydroxyproline (B) were measured and normalized to total protein to quantify crosslinked elastin and total collagen amounts, respectively, in newborn Eln^−/− and Eln^+/+ mice treated with (Cap) and without (Unt) captopril during development. Each data point represents one group of 5 pooled aortic samples. Lines are mean ± sd.

Because captopril is an ACE inhibitor, treatment likely stimulates changes in Ang II signaling that may lead to wall remodeling independently of changes in blood pressure. Ang II activity is challenging to measure in newborn mouse aorta, so as a first pass, we measured expression of Ang II receptor genes. We found no significant changes in expression of Agtr1a or Agtr2 in Eln^−/− aorta in response to captopril treatment (Fig. 5). There is increased expression of Agtr1a and Agtr2 in Eln^+/+ compared to Eln^−/− in the untreated groups, suggesting reduced expression of Ang II signaling genes in Eln^−/− aorta at baseline. Expression levels of Agtr2 are reduced in Eln^+/+ aorta in response to captopril treatment, but are unchanged in Eln^−/− aorta (Fig. 5B). Together, these results suggest that alterations in expression of Ang II signaling genes are not the primary cause of aortic wall remodeling in Eln^−/− mice with captopril treatment.

Figure 5.

Aortic mRNA levels of Agtr1a (A) and Agtr2 (B) were measured to quantify Ang II signaling in newborn Eln^−/− and Eln^+/+ mice treated with (Cap) and without (Unt) captopril during development. Values were normalized to B2M and then to Unt Eln^+/+ controls. Each data point represents one group of 8 – 10 pooled aortic samples. Lines are mean ± sd.

3.4. SMCs appear more organized in Eln^−/− aorta with captopril treatment

H&E staining shows an increased number of cells in Eln^−/− aorta due to the thicker wall, but there are no significant differences in cell density between groups (Fig. 6). SMC nuclei appear round in cross-sections for untreated Eln^−/− aorta, indicating that they are not oriented in the circumferential direction, as they are in Eln^+/+ aorta. After captopril treatment, the SMCs appear more elongated and circumferentially oriented in Eln^−/− aorta. Quantification of nuclear shape confirms that untreated Eln^−/− cells are skewed toward higher roundness values than untreated Eln^+/+ cells and that the Eln^−/− distribution shifts toward lower roundness values after captopril treatment (Fig. 6F). Interestingly, captopril treatment has the opposite effect on the shape of Eln^+/+ nuclei, shifting them to higher roundness values. The changes in SMC organization may lead to the observed reduction in wall thickness and opening angle in Eln^−/− aorta in response to captopril treatment and may indicate alterations in SMC phenotype.

Figure 6.

H&E stained aortic sections of newborn Eln^−/− and Eln^+/+ mice treated with (Cap) and without (Unt) captopril during development (A – D). Note the circular shape of the SMC nuclei in the Unt Eln^−/− aorta (white arrow) that is less apparent in the Cap Eln^−/− aorta. Scale bar = 10 μm. Cell density (E) was quantified by counting cell nuclei in defined areas of the H&E images. Each data point represents one aortic image. Cell shape was quantified by determining the roundness of the cell nuclei in the entire aortic cross-section (F). Histograms include all cells counted in each group. Unt Eln^−/− nuclei are significantly more round than Unt Eln^+/+ and Cap Eln^−/−. Cap Eln^+/+ nuclei are significantly more round than Unt Eln^+/+ nuclei. * = P < .05.

3.5. Aortic mechanical behavior is altered in Eln^−/− aorta after captopril treatment

We performed mechanical testing to determine how aortic wall remodeling in response to captopril treatment alters structural and material behavior. For untreated and captopril treated groups, Eln^−/− aorta becomes more stiff, with little change in outer diameter with increased pressure, at lower diameters than Eln^+/+ (Fig. 7A). For the untreated group only, Eln^−/− aorta is smaller than Eln^+/+ at pressures between 5 – 25 mmHg. After captopril treatment, the Eln^−/− diameter at low pressures is shifted up near Eln^+/+ levels. The diameter of Eln^−/− aorta is significantly different between untreated and captopril treated groups at 5 and 10 mmHg. Pressure-outer diameter data were converted into circumferential stretch-stress data using the measured unloaded dimensions. The untreated Eln^−/− curve is shifted to the right of the other groups, indicating higher stretch with lower stress (Fig. 7B). After captopril treatment, the Eln^−/− curve is similar to the untreated Eln^+/+ curve. The tangent physiologic modulus was calculated at the average LV pressure for each group. Captopril treatment increases the modulus of Eln^−/− aorta to near Eln^+/+ levels (Fig. 7C).

3.6. Constitutive modeling suggests altered SMC material properties in Eln^−/− aorta after captopril treatment

We used constitutive modeling to understand how component contributions may be altered in Eln^−/− aorta in response to captopril treatment. The contributions of each component to the composite material behavior of the aortic wall for the baseline model are shown in Fig. 8A with the experimental stretch-stress curve for untreated Eln^+/+ aorta. The prescribed material parameters (Table 1) provide a reasonable approximation of the experimental data and contributions for each wall component. The collagen natural configurations in the Eln^+/+ case were calculated to maintain equilibrium in the unloaded state and are 1.3 in Fig. 8A. To model Eln^−/− aorta, we assumed that the elastin mass fraction was zero and water takes up a larger fraction of the wall since we did not measure any increases in hydroxyproline amounts (Fig. 4A) or cell density (Fig. 6E). We maintained the material constants for all components and the natural configurations for SMCs. Since elastin is no longer present, the collagen natural configurations have to be recalculated to maintain equilibrium and are equal to 1.0 (Table 1). Fig. 8B shows the resulting total and component stress contributions from the model with no elastin and the experimental data from untreated Eln^−/− aorta. The model with no elastin shifts the total stress-stretch curves to the right due to the change in collagen natural configuration, but does not quite match the experimental data.

Figure 8.

Constitutive modeling of the aortic mechanical behavior from newborn Eln^−/− and Eln^+/+ mice treated with (Cap) and without (Unt) captopril during development. Model parameters are summarized in Table 1. Panels A, B, and C show total model contributions, as well as the contributions of each wall component. Panel A = Baseline model compared to Unt Eln^+/+ experimental data. Panel B = Model with no elastin compared to Unt Eln^−/− experimental data. Panel C = Model with no elastin and reduced SMC material constants compared to Unt Eln^−/− experimental data. Panel D = Total model contributions with constants in Panels B and C compared to Unt Eln^−/− and Cap Eln^−/− experimental data.

There is evidence that Eln^−/− SMCs have reduced actin filaments (Karnik et al. 2003) and a decreased passive modulus (Espinosa et al. 2013) compared to Eln^+/+ SMCs. Hence, we reasoned that Eln^−/− aorta may require altered SMC material constants in our constitutive model. Fig. 8C shows model stresses and component contributions for the same case as Fig. 8B with no elastin, but includes 50% reductions in the linear SMC material constants (Table 1). The model approximates the untreated Eln^−/− experimental data in Fig. 8C.

After captopril treatment in Eln^−/− aorta, we have not changed elastin or collagen amounts (Fig. 4) or cell density (Fig. 6E), but we may have altered SMC organization and/or phenotype, as evidenced by the decreased roundness values in the cell orientation histogram (Fig. 6F). Fig. 8D shows the model predictions for the total aortic wall behavior without elastin, but with and without the 50% reduction in the linear SMC material constants. The model without the changes in SMC behavior approximates the experimental data for the captopril treated Eln^−/− aorta. Our modeling results indicate how complementary changes in collagen natural configurations and SMC material properties can produce mechanical behavior similar to untreated and captopril treated Eln^−/− aorta.

3.7. SMC gene expression is not significantly different between groups

The constitutive modeling suggests that captopril treatment may rescue SMC material properties in Eln^−/− aorta. To investigate SMC phenotype changes, we measured gene expression levels of a proliferation marker (Mki67) and SMC-specific genes (Acta2, Cnn1, Myh11, Smtn, and Tagln). There were no significant differences between groups (Fig. 9), indicating no changes in SMC phenotype as linked to expression of these marker genes.

Figure 9.

Aortic mRNA levels of Mki67 (A), Acta2 (B), Cnn1 (C), Myh11 (D), Smtn (E), and Tagln (F) were measured to quantify proliferation and SMC phenotype in newborn Eln^−/− and Eln^+/+ mice treated with (Cap) and without (Unt) captopril during development. Values were normalized to B2M and then to Unt Eln^+/+ controls. Each data point represents one aortic sample. Lines are mean ± sd.

4. Discussion

Developing vessels remodel in response to changes in mechanical stimuli. This was observed by Thoma in 1893 for vasculature in the chick embryo (Clark 1918) and has been supported by studies in the aorta of numerous species including chickens (Espinosa et al. 2018), rats (Gerrity and Cliff 1975), rabbits (Leung et al. 1977), and sheep (Langille 1993). In normal development, remodeling allows the aortic wall to adapt to mechanical demands for optimal cardiovascular function. In abnormal development, the remodeling may exacerbate disease pathology (Humphrey et al. 2014). The remodeling may also be harnessed to manipulate desired outcomes in tissue engineering or disease treatment. For example, culturing tissue engineered arteries under pulsatile pressure increases ECM deposition and burst strength compared to those cultured under static conditions (Niklason et al. 1999). Additionally, the use of anti-hypertensive medication influences the severity of vascular stiffness in Williams Syndrome (Kozel et al. 2014). Williams Syndrome is caused by a contiguous gene deletion that includes the elastin gene and has the same vascular phenotype as SVAS (Pober et al. 2008). In the current study, we show that treating Eln^−/− mice with captopril during embryonic development lowers blood pressure at birth and induces aortic remodeling that ameliorates the pathologic phenotype associated with the lack of elastin.

We found that newborn, untreated Eln^−/− mice have 31% higher LV pressure than Eln^+/+, consistent with previous data (Wagenseil et al. 2009). Treating mice with the ACE-inhibitor captopril during embryonic development lowers blood pressure in newborn Eln^−/− mice to near Eln^+/+ levels. Blood pressure was measured in 5 of the treated Eln^+/− dams to confirm that captopril was having the desired effect. The systolic pressure of the treated dams was 28% lower than comparable measurements from our laboratory on adult male Eln^+/− mice (Stoka et al, 2018). In previous studies (Huang et al. 2013), we found a 10 – 15% drop in blood pressure with captopril treatment in young mice, which is similar to the 12% drop in newborn mice in the current study. The resulting decrease in circumferential wall stress may alleviate the mechanically-mediated increase in wall thickness associated with high blood pressure (Wolinsky 1970), as evidenced by the reduced wall thickness in captopril-treated Eln^−/− aorta. Captopril-treated Eln^−/− aorta is still thicker than Eln^+/+ aorta, however, as aortic remodeling is also driven by biochemical signaling in the absence of elastin (Karnik et al. 2003). The decrease in residual strain as measured by opening angle is consistent with reduced hypertensive remodeling in captopril-treated Eln^−/− aorta (Liu and Fung 1989). Aortic length does not change with captopril treatment in Eln^−/− aorta, indicating that mechanically-stimulated remodeling in the axial direction is not affected by our chosen anti-hypertensive treatment. Increased wall thickness, decreased opening angle, and increased length in untreated Eln^−/− aorta compared to Eln^+/+ are similar to previous results in newborn mice (Kim et al. 2017; Wagenseil et al. 2009).

Prior studies investigating remodeling in the developing aorta found alterations in deposition of ECM proteins in response to blood pressure changes (Espinosa et al. 2018; Gerrity and Cliff 1975; Leung et al. 1977). We cannot alter elastin deposition in Eln^−/− aorta, however, collagen deposition is unchanged in response to reduced elastin levels and changes in blood pressure in the current study and in previous studies on Eln^+/− (Cheng et al. 2013; Faury et al. 2003) and Eln^−/− (Kim et al. 2017) mice. Expression of other ECM genes, including non-fibrillar collagens and elastic fiber associated molecules, are altered in Eln^+/− (Knutsen et al. 2018) and Eln^−/− (Staiculescu et al. 2018) aorta indicating that the ECM remodeling response is more complicated than previously appreciated.

The response of SMCs to mechanically-induced arterial remodeling includes hypertrophy (Ben Driss et al. 1997) and proliferation (Davis et al. 2005). SMCs overproliferate and are disorganized in Eln^−/− aorta (Li et al. 1998; Wagenseil et al. 2009). The SMC response is likely due to hypertension in Eln^−/− mice (Wagenseil et al. 2009), as well as the loss of biochemical elastin signaling (Karnik et al. 2003), and the absence of the physical barrier presented by elastic laminae (Cheng and Wagenseil 2012). We confirmed that cell nuclei are disorganized (rounder) in untreated Eln^−/− aorta than Eln^+/+ and become more organized in Eln^−/− aorta after captopril treatment. Despite previous studies showing increased SMC proliferation in late embryonic development (Li et al. 1998) and an increased amount of DNA in the aorta at birth (Wagenseil et al. 2009), Mki67 expression was not significantly different in untreated Eln^−/− aorta compared to Eln^+/+. It is possible that overproliferation of SMCs is downregulated in Eln^−/− aorta within 24 hours of birth. Previous work with cultured cells showed reduced expression of SMC marker genes in Eln^−/− SMCs compared to Eln^+/+ (Espinosa et al. 2013; Karnik et al. 2003), however we found no significant differences in SMC marker gene expression between groups. This may be due to differences between aortic tissue and cultured cells or to the high variability in our SMC gene expression data. The variability may be due to differences in developmental age of the pups at sacrifice (within 24 hours of birth) or individual responses to captopril treatment. Variability was reduced in the analysis of Ang II receptor genes, likely due to pooling of multiple samples, but at the expense of a four-fold increase in required animals. Additional work is needed to further investigate changes in SMC phenotype in response to captopril treatment.

Captopril reduces blood pressure through ACE inhibition and downstream effects on the renin-angiotensin pathway. Ang II is a potent modulator of SMC phenotype and ECM synthesis, hence aortic remodeling in the captopril treated group could be caused by direct effects of Ang II signaling. As measurements of Ang II levels are challenging to obtain in newborn aorta, we quantified mRNA expression of the two main Ang II receptors as an indication of renin-angiotensin signaling activity. We found that both receptors are expressed at lower levels in Eln^−/− aorta compared to Eln^+/+, suggesting that renin-angiotensin signaling is altered due to the absence of elastin. Ang II receptor expression is not changed by captopril treatment in Eln^−/− mice, suggesting that alterations in Ang II signaling do not explain the remodeling observed in captopril treated Eln^−/− aorta. However, receptor gene expression is not a direct measure of Ang II signaling. Previous studies found elevated plasma renin activity (Faury et al. 2003) and augmented Ang II induced vasoconstriction of resistance vessels (Osei-Owusu et al. 2014) in Eln^+/− mice suggesting hyperactivation of the renin-angiotensin system that may be independent of receptor expression. Future studies with additional anti-hypertensive treatments that act through alternative mechanisms are needed to support our assertion that captopril ameliorates Eln^−/− aortic remodeling through a reduction in blood pressure.

ECM and SMC amounts, organization, and behavior determine the structural and material properties of the aortic wall. We found that untreated Eln^−/− aorta has altered structural and material behavior compared to Eln^+/+ aorta, consistent with previous reports (Kim et al. 2017; Wagenseil et al. 2009). Captopril treatment has no effect on Eln^+/+ aorta, but shifts the structural and material behavior of Eln^−/− aorta toward Eln^+/+ values. The physiologic modulus values are comparable to previous studies on newborn mouse aorta (Kim et al. 2015; Staiculescu et al. 2017) and are significantly affected by captopril treatment in Eln^−/− aorta only. To determine mechanisms that could lead to the observed shifts in mechanical behavior of the composite aortic wall, we employed constitutive modeling. We considered the aortic wall as a constrained mixture of elastin, collagen, and SMCs where each component has a natural configuration (Humphrey and Rajagopal 2002) and we prescribed model constants for illustrative behavior. Our baseline model represents untreated Eln^+/+ aorta. Elastin has little mechanical contribution in newborn mouse aorta, as the elastic laminae are not yet completely formed (Davis 1995). Despite minimal contributions to the stretch-stress behavior, elastin plays a key mechanical role. We assumed that elastin is slightly compressed in the natural configuration, causing collagen to be stretched to maintain unloaded equilibrium. When elastin is not present in Eln^−/− aorta, collagen is no longer stretched in the natural configuration and the collagen contribution to the stretch-stress behavior decreases significantly, mimicking the rightward shift observed experimentally in the untreated Eln^−/− stretch-stress curves.

Previous work on adult arteries assumes that elastin is under tension in the natural configuration, which places collagen under compression (Ferruzzi et al. 2011; Fonck et al. 2007; Rachev and Shazly 2019). This set of natural configurations reproduces data on elastase-treated adult arteries where the stress-stretch curves are shifted left toward higher tangent modulus values after elastin degradation, which is the opposite of the untreated Eln^−/− mechanical behavior. In adult arteries, it is thought that elastin is under tension in the natural configuration because elastin is deposited during development, does not turnover, and must be stretched as the animal grows (Wagenseil and Mecham 2009). However, elastin production is at a high level in the newborn mouse aorta (Kelleher et al. 2004) and elastin may be compressed at this stage if elastin is being deposited into the elastic laminae at a faster rate than the arterial wall growth.

We assumed that passive SMCs play a large role in the newborn mouse aorta because the ECM is still being deposited (Kelleher et al. 2004). Due to the large SMC contribution, the model stress-stretch curve is not altered enough after the removal of elastin and new natural configurations for collagen to match the experimental data for Eln^−/− aorta. We have no experimental evidence for alterations to the collagen material properties, but there is prior evidence for reductions to the passive SMC material properties in Eln^−/− aorta. In culture, Eln^−/− SMCs have a reduced Young’s modulus as measured by atomic force microscopy (Espinosa et al. 2013), reduced expression of contractile proteins, and limited numbers of actin stress fibers (Karnik et al. 2003). Based on these observations, we reduced the linear SMC material constants by 50% in the model. With reduced SMC material constants, no elastin, and new collagen natural configurations, the model approximates the untreated Eln^−/− experimental data. With the original SMC material constants, no elastin, and new collagen natural configurations, the model approximates the captopril treated Eln^−/− experimental data, indicating that changes to the SMC properties alone could explain our results for captopril treated Eln^−/− aorta.

Our experimental results show that captopril treatment ameliorates Eln^−/− aortic remodeling and suggest that reducing blood pressure may be a viable treatment option for reducing or delaying pathological aortic remodeling due to elastin insufficiency. Our constitutive model suggests that the observed changes in aortic structural and material properties may be mediated by mechanically-stimulated alterations in SMC phenotype. Anti-hypertensive treatment in utero is not advised in humans due to possible birth defects, however, many children with elastin insufficiency are already prescribed anti-hypertensive medications (Bouchireb et al. 2010). It is possible that anti-hypertensive treatment in childhood would be sufficient to ameliorate some vascular phenotypes in humans with elastin insufficiency. There is evidence that elastin expression continues throughout life in adult human aorta (Fritze et al. 2012), even though lastin expression is turned off in adult mouse aorta (Kelleher et al. 2004), indicating that there may be a longer treatment window in humans than mice for elastin insufficiency. As there are currently no pharmaceutical strategies to increase elastin deposition in developing aorta, pharmaceutical strategies that manipulate aortic remodeling to improve structural and material properties through alternative mechanisms are an encouraging avenue for treatment of genetic elastinopathies.

Limitations and future directions

Sex of the animals was not determined, so sex-dependent variation for the effectiveness of captopril treatment in reducing blood pressure and the effects on the aortic mechanical behavior are unknown. Our conclusions are based on results with a single anti-hypertensive drug. Our results must be tested through additional studies with anti-hypertensives that work through different mechanisms to confirm that the changes are due to decreased blood pressure and not through captopril-specific pathways such as modulation of renin-angiotensin signaling. Captopril treatment increased heart rate and aortic diameter in Eln^−/− mice. These changes would affect the Wormersley number and the pulsatile nature of the blood flow which may have effects on endothelial cell signaling. Endothelial cell changes should be addressed in future research. Our mechanical test data and model are based on the assumption of elastic material behavior and ignore viscoelastic effects. Previous work has shown differences in viscoelastic behavior between Eln^+/+ and Eln^−/− aorta (Kim et al. 2017), which may also be affected by captopril treatment. Our model has several limitations including the prescribed material parameters and natural configurations that must be experimentally determined in future work.