Abstract
None of the animal models have been able to reproduce all aspects of CTEPH because of the rapid resolution of the thrombi in the pulmonary vasculature. The aim of this study was to develop an easily reproducible large-animal model of chronic pulmonary hypertension (PH) related to the development of a postobstructive and overflow vasculopathy. Chronic PH was induced in 5 piglets by ligation of the left pulmonary artery (PA) through a midline sternotomy followed by weekly transcatheter embolization of the right lower-lobe arteries. Sham-operated piglets (n = 5) served as controls. Hemodynamics, RV function, lung morphometry, and endothelin-1 (ET-1) pathway gene expression (ET-1 and its receptors ETA and ETB) were assessed after 5 weeks in the obstructed (left lung and right lower lobe) and unobstructed (right upper lobe) territories. All animals developed chronic PH within 5 weeks. Compared to controls, chronic-PH animals had higher mean PA pressure (28.5 ± 1.7 vs. 11.6 ± 1.8 mmHg, P = 0.0001) and total pulmonary resistance (784 ± 160 vs. 378 ± 51 dyn s−1 cm−5, P = 0.05). Echocardiography showed RV enlargement, RV wall thickening (56 ± 5 vs. 30 ± 4 mm, P = 0.0003), decreased tricuspid annular plane systolic excursion (11.3 ± 0.9 vs. 14.4 ± 0.4 mm, P = 0.01), and paradoxical septal motion. In obstructed territories, morphometry demonstrated increases in the number of bronchial arteries per bronchus (8.7 ± 0.9 vs. 2 ± 0.17, P < 0.0001) and in distal PA media thickness (60% ± 2.8% vs. 29% ± 0.9%, P < 0.0001), consistent with postobstructive vasculopathy. Distal PA media thickness was increased in unobstructed territories (70% ± 2.4% vs. 29% ± 0.9%, P < 0.0001). ET-1 was overexpressed in unobstructed territories, compared to controls and obstructed territories. In conclusion, the large-animal model described here is reproducible and led to the development of PH in a relatively short time frame.
INTRODUCTION
Chronic thromboembolic pulmonary hypertension (CTEPH) is a subtype of pulmonary hypertension (group 4 in the Dana Point classification 1 ) that can be effectively treated with pulmonary endarterectomy 2 to prevent the development of right ventricle (RV) failure. CTEPH is due to pulmonary arterial bed obstruction by persistent organized clots related to one or more episodes of acute pulmonary embolism. The absence of clot lysis and the pathophysiology of CTEPH remain unexplained. Although the increase in pulmonary vascular resistance is believed to result from a combination of postobstructive vasculopathy in the obstructed territory and overflow vasculopathy in the unobstructed territory,3,4 the mechanisms underlying lesion development in the obstructed and unobstructed peripheral vascular beds remain unknown. Histopathological studies have shown distal pulmonary artery (PA) remodeling with media hypertrophy and the development of plexiform lesions in obstructed or unobstructed territories.4,5 Moser et al. 5 were the first to describe two compartments in the pulmonary vascular bed, an obstructed compartment downstream of the occluded vessels and subjected to chronic ischemia responsible for bronchial circulation hypertrophy and an unobstructed compartment subjected to increased flow and shear stress. Since then, no major advances have been achieved in elucidating the pathobiology of CTEPH, despite dramatic progress in the field of pulmonary hypertension. 6 The absence of an animal model replicating all the features of CTEPH in the lungs and heart may explain the dearth of experimental studies on this disease. The aim of this study was to develop a reproducible large-animal model of chronic pulmonary hypertension (PH) characterized by postobstructive and overflow vasculopathy in the obstructed and unobstructed pulmonary vascular territories, respectively.
METHODS
We studied 10 Large White piglets weighing 21 ± 3 kg. The study complied with the Principles of Laboratory Animal Care developed by the National Society for Medical Research and was approved by our local ethics committee on animal experiments.
Experimental design
The piglets were randomly allocated to two groups of 5 animals each. The pulmonary vasculopathy induced in the chronic-PH group was compared to that in sham-operated controls. All animals were studied 5 weeks after the first procedure. We performed hemodynamic and echocardiographic studies as well as a lung morphometry analysis and investigations of the endothelin-1 (ET-1) pathway.
Surgical procedures
All procedures were performed under general anesthesia, as described elsewhere.7,8 Chronic PH was induced by ligation of the left PA through a midline sternotomy 9 followed by weekly embolization of enbucrilate tissue adhesive (Histoacryl; B. Braun, Melsungen, Germany) into the right lower-lobe arteries for 5 weeks. 10 The amount of Histoacryl injected did not exceed 1 mL/week and depended on hemodynamic and respiratory tolerance. Usually we stopped the embolization when the mean systemic pressure dropped below 60 mmHg, oxygen saturation was <90%, and/or cardiac output (CO) was <2 L/minute. Sham-group piglets underwent left PA dissection without ligation through a median sternotomy, followed by weekly saline solution injections for 5 weeks. Right heart catheterization and pulmonary embolization were conducted through an 8-Fr introducer (Radiofocus Introducer II; Terumo, Somerset, NJ) inserted into the jugular vein and carried out with a 5-Fr vertebral catheter (Terumo) positioned at the origin of the right lower-lobe artery under angiographic monitoring.
Echocardiography
Transthoracic echocardiography was performed before ligation of the left PA and then before each embolization into the right lower lobe. In the sham group, all assessments were performed once a week for 5 weeks. Echocardiographic images (M-mode, 2-dimensional, Doppler) were obtained with a Vivid E9 machine (General Electric Medical Systems, Milwaukee, WI). Images were obtained at end-exhalation for cardiac dimension measurement and were analyzed on a comprehensive workstation (EchoPAC Dimension'06; General Electric Medical Systems). Tricuspid annular plane systolic excursion (TAPSE), defined as the maximum systolic displacement of the lateral portion of the tricuspid annular plane, was measured on M-mode images to evaluate RV longitudinal systolic function. The pulmonary annulus was measured at maximal pulmonary valve opening during systole. RV wall thickness was measured on the 4-chamber apical view in front of the maximal end-diastolic dimension of the mid-RV diameter. Systolic PA pressure was estimated from the tricuspid regurgitation.
Thoracic computed tomography (CT)
An enhanced thoracic CT scan was performed in the chronic-PH piglets to assess pulmonary vascular obstruction, cardiac remodeling, and the systemic vessel supply to the lung.
Hemodynamic measurements
Pulmonary hemodynamic variables, aortic blood flow, pressures, and blood gases were measured before and after each experiment. Right heart catheterization was performed with the Seldinger technique, with an 8-Fr sheath inserted via the jugular vein. The right heart catheter was a 7-Fr, 2-lumen, thermodilution pressure–measuring tipped catheter (Edwards Lifesciences, Irvine, CA). CO was measured with the thermodilution technique, and CO data corresponded to the average of 3 measurements. Total pulmonary resistance (TPR) was calculated as (mPAP/CO) α 80, where mPAP is mean pulmonary arterial pressure; TPR is expressed in dyn s−1 cm−5.
Lung harvesting
Biopsies weighing 300–500 mg were taken from the left upper lobe and right upper and lower lobes, snap-frozen in liquid nitrogen, and stored at −70αC or fixed in 4% paraformaldehyde solution.
Light microscopy and morphometry
Fixed lung sections were processed with standard histological techniques and embedded in paraffin. First, we sought to identify 30–40 PAs measuring <200, 200–400, 400–600, and 600–800 αm in diameter in the lung biopsies from each piglet. Medial thickness was calculated as (ED – ID/ED) α 100, where ED is the external diameter and ID the internal diameter. We also assessed the systemic vasculature to the lung by counting the submucosal bronchial arteries supplying each bronchus. All morphometric studies were performed in a blinded manner.
Real-time quantification by polymerase chain reaction assay (RT-PCR)
RT-PCR assays were conducted as previously described 11 on lung tissue from the left lungs of chronic-PH animals to study the postobstructive vasculopathy, on tissue from the right upper lungs of chronic-PH animals to study the overflow vasculopathy, and on tissue from the left lungs of sham-group animals. The primers for ET-1 and its receptors ETA and ETB were the specific porcine primers described by Rondelet et al. 11
Statistical analysis
All results are reported as mean ± SEM. The Kolmogorov-Smirnov test was used on each continuous variable to determine whether the values were normally distributed. One-way analysis of covariance was performed, followed by Fisher's test for between-group comparisons. Two-way analysis of variance for repeated measurements, followed by a post hoc test, was used for between-group comparisons of hemodynamic and echocardiographic data collected during the experiments. All statistical analyses were performed with Statview IV (Abacus Concept, Berkeley, CA).
RESULTS
Echocardiography
Echocardiography (Fig. 1) showed progressive RV remodeling starting as early as 2 weeks after the first embolization. RV thickness increased gradually over the 5-week observation period. At the end of the experiments, the chronic-PH piglets had enlarged RVs, with increases in tricuspid annulus diameter (31.8 ± 2.8 vs. 24 ± 1.8 mm, P = 0.0003), systolic PA pressure (41.5 ± 1.3 vs. 18.8 ± 1.8 mmHg, P < 0.0001), and RV wall thickness (56 ± 5 vs. 30 ± 4 mm, P = 0.0003); decreased TAPSE (11.3 ± 0.9 vs. 14.4 ± 0.4 mm, P = 0.01); and paradoxical septal motion.
Echocardiography results comparing CTEPH (chronic thromboembolic pulmonary hypertension) and sham groups. Piglets in the CTEPH group had right ventricle (RV) remodeling with significant increases in thickness, tricuspid annulus diameter, and systolic pulmonary artery pressure and a significant decrease in tricuspid annular plane systolic excursion (TAPSE). Asterisk indicates P < 0.05 for the CTEPH group versus the sham group.
Hemodynamic study
Compared to the sham-group animals, the chronic-PH animals had higher values for mPAP (28.5 ± 1.7 vs. 11.6 ± 1.8 mmHg, P = 0.0001) and TPR (784 ± 160 vs. 378 ± 51 dyn s−1 cm−5, P = 0.05). Starting at the first embolization, mPAP and TPR remained significantly higher in the chronic-PH group than in the sham group. CO did not differ significantly between the two groups (2.8 ± 0. vs. 2.2 ± 0.7 L/min, P = NS [not significant]). See Figure 2.
Hemodynamic measurements showing higher mean pulmonary artery pressure (mPAP) and total pulmonary resistance (TPR) values in the CTEPH (chronic thromboembolic pulmonary hypertension) group, compared to the sham group, from the first embolization to the fifth week. After 5 weeks of embolization, mPAP was >25 mmHg at rest. Asterisk indicates P < 0.05 for the CTEPH group versus the sham group.
Enhanced thoracic CT
Thoracic CT in the chronic-PH piglets showed RV enlargement (Fig. 3A), proximal vascular obstruction by unresolved material (Fig. 3B), and hypertrophy of the bronchial circulation (Fig. 3C).
Enhanced thoracic computed tomography showing right ventricle remodeling and enlargement (A), unresolved proximal obstruction of the right lower-lobe arteries (B), and bronchial artery hypertrophy (C).
Gross anatomy of the harvested heart-lung block
Bronchial circulation hypertrophy was noted in the left lung and, to a lesser extent, in the right lower lobe in the chronic-PH piglets. Moreover, unresolved material composed of Histoacryl and fibrin was found inside the PA downstream from the right upper lobe. This material adhered to the arterial wall and molded the arterial tree (Fig. 4). The right upper lungs seemed identical in appearance to the lungs of the sham-group piglets.
Gross anatomy of the harvested lungs, showing the different vascular lung territories and the unresolved intravascular material. RLL: right lower lobe.
The RV was enlarged and thickened in the CTEPH group. RV free-wall thickness was increased in the chronic-PH group compared to that in the sham group (56 ± 2.5 vs. 31.5 ± 2.85 mm, P = 0.0004).
Light microscopy and morphometry
Postobstructive vasculopathy with media hypertrophy (Fig. 5) was found in obstructed territories (left lung and right lower lobe), while overflow vasculopathy was observed in the unobstructed territories (right upper lobe), regardless of external artery diameter. There was no vascular remodeling in the lungs of the sham-group animals.
Results of the morphometric study showing pulmonary vasculopathy with media hypertrophy in obstructed and unobstructed territories, compared to the sham group. MT: media thickness; RUL: right upper lobe; Isch.: ischemia; RLL: right lower lobe.
The submucosal bronchial arteries were larger and more numerous in the left lung and right lower lobe (8.6 ± 5.1 and 5.5 ± 4 arteries per bronchus, respectively) than in the right upper lobe (2.2 ± 0.8 arteries per bronchus) and the sham group (2 ± 0.9 arteries per bronchus; Fig. 6). In the chronic-PH animals, the bronchial arteries were larger, more numerous, and thicker (Fig. 7).
Number of submucosal bronchial arteries per bronchus, showing hypertrophy of the systemic vessels supplying the ischemic lung (left lung and right lower lobe). Isch.: ischemia; RLL: right lower lobe; RUL: right upper lobe.
Fixed left lung section stained with hematoxylin and eosin stain (α20): enlarged and hypertrophic submucosal bronchial artery in a chronic thromboembolic pulmonary hypertension–group piglet.
Real-time quantification by PCR
Endothelin-1 (ET-1) and ETB messenger RNAs (mRNAs) in lung tissue were overexpressed in the presence of an overflow vasculopathy in the right upper lobe of the chronic-PH animals, compared to levels in the sham-group animals (Fig. 8). In contrast, the expression of ET-1 and ETB mRNAs was not increased in the presence of a postobstructive vasculopathy in the left lungs of chronic-PH animals (obstructed territory), compared to that in sham-group animals (P = 0.02 and 0.03, respectively). The level of ETA mRNA expression in the lung was higher in both the overflow-vasculopathy and postobstructive-vasculopathy territories of the chronic-PH animals, compared to that in the sham-group animals (P = 0.04).
Quantification of messenger RNA (mRNA), using real-time-polymerase chain reaction, for endothelin-1 (ET-1) and its receptors ETA and ETB in lung parenchyma. ET-1, ETA, and ETB were overexpressed in the right upper lobe subjected to high blood flow. ETA was overexpressed in the left lung subjected to chronic ischemia. Asterisk indicates P < 0.05 for the high-flow group versus the sham group; pound sign indicates P < 0.05 for the ischemia group versus the sham group.
DISCUSSION
We report the first large-animal model of chronic PH that is reproducible in a short time frame. Indeed, all animals had increased PA pressure and pulmonary resistance within 5 weeks after the start of the experiments. This model reproduces some of the morphological, hemodynamic, echocardiographic, and biological modifications seen in human CTEPH, with the presence of a postobstructive vasculopathy associated with an overflow vasculopathy in different territories of the lung. The development of chronic PH is also associated with significant RV remodeling.
CTEPH is due to chronic pulmonary vascular obstruction by organized clots, with subsequent right heart failure. The aim of our study was to induce PH (i.e., mPAP > 25 mmHg at rest 12 ) with RV remodeling and to mimic some of the specific features of CTEPH, most notably the combination of pulmonary vascular ischemia in obstructed territories and increased pulmonary blood flow in unobstructed territories. We had previously developed two animal models that separately replicate each of these compartments in piglets: left PA ligation for 5 weeks simulates PA obstruction, 9 and aortopulmonary shunting for 5 weeks induces high-flow vasculopathy. 7 The main drawbacks of these previous animal models were the absence of interactions between the two pulmonary vascular compartments and the absence of RV remodeling and dysfunction. In this study, we succeeded in replicating high-flow-induced vasculopathy with PA remodeling in the unobstructed right upper lobe 7 and postobstructive vasculopathy combining PA remodeling and bronchial artery hypertrophy in the chronically obstructed right lower lobe and left lung. 9 More importantly, the combination of the postobstructive and overflow vasculopathies resulted in PH with right heart remodeling, as shown by hemodynamic, echocardiographic, and morphological studies.
The ET-1 pathway is known to be involved in many causes of PH. 6 In the high-flow model, we previously documented overexpression of ET-1, ETA, and ETB after 5 weeks of shunting. 11 Similarly, in the present study, ET-1, ETA, and ETB expression was increased in the right upper lobe subjected to chronic overflow conditions due to CO redistribution. The ET-1 pathway was also involved in postobstructive vasculopathy. It has been shown that the increased ET-1 responses in PAs downstream of chronic PA ligation are attributable to an increase in the ratio of ETA to ETB. 13 These results are also consistent with the ETA mRNA levels found in the ischemic lung (left lung and right lower lobe) in the present study.
Since the 1990s, several attempts to develop animal models of CTEPH14–18 have failed because of the high fibrinolysis potential of pulmonary endothelial cells and the adaptive capabilities of the pulmonary circulation. Thus, 3 hours after acute pulmonary embolism in dogs, 14 only 30% of the initial injected thrombus volume remained inside the PA. Adding tranexamic acid 15 or plasminogen activator inhibitor-1 16 to delay thrombus resorption between injections failed to solve this problem. We consequently used nonabsorbable glue that solidifies immediately in contact with blood and adheres to the arterial wall. At this point, Histoacryl glue may exacerbate the intravascular inflammatory response, and this should be considered when using our model for pathophysiological studies. The result was proximal obstruction by unresolved material in the right lower-lobe artery, which gradually increased the mPAP and TPR. Enhanced CT showed persistent obstruction by the intravascular material after 5 weeks. This model produced both unobstructed territories (right upper lobe) and obstructed territories of different ages (right lower lobe and left lung), as seen in clinical practice.
The second issue that had to be dealt with was the adaptive capability of the pulmonary circulation, which required obstruction of more than half the pulmonary vasculature to increase pulmonary vascular resistance. Adaptation of the pulmonary vasculature explains the inability of other inert and nonabsorbable materials to simulate CTEPH when injected into the PAs. Thus, in a model of chronic PA injections, pulmonary pressures and resistances returned to normal within 1 week after each injection. 17 With microspheres, 60 days of repeated embolization were required to increase PA pressure. 18 After 60 days, signs of pulmonary hypertension started to develop, but there was no bronchial artery hypertrophy, postobstructive or high-flow vasculopathy, or proximal vascular obstruction. 18 In our model, left PA ligation was followed by progressive obstruction of the remaining pulmonary arterial tree to increase the extent of the obstructed territory, thereby inducing sustained PH (mPAP > 25 mmHg at rest). PA ligation led to dilation of most of the pulmonary arterial bed, thus overwhelming the adaptive potential of the pulmonary circulation and leading to PH within a few weeks. The progressive nature of the obstruction of weekly embolization allowed RV adaptation, thus preventing death by acute RV failure.
In our model, proximal PA obstruction occurred, as demonstrated by the development of postobstructive vasculopathy in the right lower lobe and left lung. However, vascular obstruction could be produced more distally by pushing the catheter farther into the PA. Thus, several types of vascular obstruction (lobar, segmental, and subsegmental) could be induced to replicate the different situations met in clinical practice.
In our model, pulmonary hypertension resulted in RV remodeling with increased RV free-wall thickness, paradoxical septal motion, and decreased TAPSE. However, although TAPSE decreased with successive embolizations, CO remained the same as in the sham group. Although our model replicated all the echocardiographic signs of RV dysfunction, we found no clinical evidence of RV failure. Further embolizations may be needed to study the mechanisms underlying CTEPH-induced RV failure.
Our study has several limitations. We did not determine whether pulmonary resistance continued to increase in the absence of additional embolizations, as observed in humans during the honeymoon period. 19 Further studies are needed to monitor pulmonary resistance at a distance from the last pulmonary embolization. The pulmonary vasculature of our piglets did not exhibit primary endothelial cell dysfunction, which is the mechanism believed to explain the absence of spontaneous clot lysis and the organization of clots seen in human CTEPH. 20 However, despite these limitations, our CTEPH model replicates all the major aspects of the human disease and should therefore prove useful for investigating the response of distal high-flow pulmonary vasculopathy and RV dysfunction to pharmacological agents or surgical reversal of PH by left PA reperfusion.
In conclusion, we describe the first large-animal model of chronic PH replicating the vascular abnormalities seen in both obstructed and unobstructed territories in human CTEPH, as well as RV remodeling, in a short time frame. The challenges raised by the adaptive properties of the pulmonary circulation and the marked fibrinolytic potential of pulmonary endothelial cells were overcome by ligating the left PA and then performing injections of a nonresorbable quick-set glue into the PAs over a 5-week period. Additional embolizations or a longer observation time may replicate the advanced stages of CTEPH with RV failure. This new model should prove valuable for studies investigating relationships between obstructed and unobstructed territories, right heart failure, and the effects of targeted treatments.