The Latest in Animal Models of Pulmonary Hypertension and Right Ventricular Failure

Abstract

Pulmonary hypertension (PH) describes heterogeneous population of patients with a mean pulmonary arterial pressure greater than 20 mmHg. Rarely, PH presents as a primary disorder but is more commonly part of a complex phenotype associated with co-morbidities. Regardless of cause, PH reduces life expectancy and impacts quality of life. The current clinical classification divides PH into one of 5 diagnostic groups to assign treatment. There are currently no pharmacological cures for any form of PH. Animal models are essential to help decipher the molecular mechanisms underlying the disease, to assign genotype-phenotype relationships to help identify new therapeutic targets, and for clinical translation to assess the mechanism of action and putative efficacy of new therapies. However, limitations inherent of all animal models of disease limit the ability of any single model to fully recapitulate complex human disease. Within the PH community, we are often critical of animal models due to the perceived low success upon clinical translation of new drugs. In this review, we describe the characteristics, advantages and disadvantages of existing animal models developed to gain insight into the molecular and pathological mechanisms and test new therapeutics, focusing on adult forms of PH from groups 1 to 3. We also discuss areas of improvement for animal models with approaches combining several hits in order to better reflect the clinical situation and elevate their translational value.

Pulmonary hypertension (PH) encompasses a heterogenous group of disorders collectively defined by a mean pulmonary artery (PA) pressure above 20 mmHg at rest¹. PH is classified into five groups by the World Health Organization (WHO) according to etiology. Group 1 refers to pulmonary arterial hypertension (PAH), which is characterized by a progressive and severe obliteration of PAs due to excessive vasoconstriction and vascular cell proliferation. The steady rise in pulmonary vascular resistances (PVR) increases right ventricular (RV) afterload pushing the RV to undergo adaptive remodeling events. Almost inevitably, these coping mechanisms become insufficient, and patients rapidly succumb to RV failure. PAH can be heritable, idiopathic, induced by drugs or associated with various conditions such as scleroderma and HIV infection. Group 2 (due to left-sided heart disease) and group 3 (secondary to chronic lung disease and/or alveolar hypoxia) are the most prevalent forms of PH². Group 4 PH is related to obstruction of the pulmonary vasculature from thromboembolism. Lastly, group 5 consists of several forms of PH with multifactorial or unclear mechanisms. Regardless of the underlying etiology, PH is characterized by varying degrees of PA remodeling with restricted blood flow in the lung. Despite significant progress in basic research and improvements of patient management, PH still remains a life-threatening-disease with a poor prognosis.

Animal models are essential to help decipher the molecular mechanisms underlying the disease to identify new actionable therapeutic targets, and for clinical translation to assess the impact of putative therapies. However, because animals are not humans and PH is a complex and multifactorial disease, no single animal model faithfully reproduces the full clinical spectrum of PH, or indeed each PH subgroup. The limitation of PH animal models is further indirectly stressed by the fact that many drugs that were promising during preclinical evaluation have low success during subsequent clinical testing; a translational gap often referred to as the “Valley of Death”. In the present review, we detail the characteristics, advantages and drawbacks of different animal models developed to gain insight into the mechanisms and mimic aspects of the disease, focusing on adult forms of PH from groups 1 to 3 (Table 1).

Table 1.

Animal models of group 1 to 3 pulmonary hypertension

EXPERIMENTAL MODELS	ANIMAL SPECIES	PH Group	PULMONARY VASCULAR REMODLEING	RVSP	RV DYSFUNCTION	COMMENTS	REFERENCES
MCT (single injection 60mg.kg−1)	RAT	1	+++	++++	++++	Simple implementation Fast, Low cost Causes pronounced vascular remodeling Allow to interrogate the adaptive and maladaptive changes of the RV Myocarditis Reversible at lowest concentrations	5–7
MCT + Hypoxia	RAT	1	++++	++++	++++	Myocarditis	19
MCT + Pneumonectomy	RAT	1	++++	++++	++++	Requires technical skills Causes pronounced vascular remodeling and RV dysfunction	20
MCT + aortocaval shunt	RAT	1	++++	++++	++++	Requires technical skills Causes pronounced vascular remodeling and RV dysfunction	21, 22
Sugen-Hypoxia	RAT	1 (3)	+++	++++	+++	Relatively inexpensive – Lack of harmonization among laboratories Causes pronounced vascular remodeling with development of plexiform-like lesions Strain variability	25–28
Sugen-Hypoxia	MOUSE	1 (3)	++	+++	+	Partially reversible	33, 34
Inducible overexpression dominant-negative BMPR2 targeted to SMCs (Tg+/Tagln-rtTA; Tg+/tetO-Bmpr2Δex4)	MOUSE	1		++	+		42
Inducible overexpression dominant-negative BMPR2 (Tg+/ROSA26-rtTA; Tg+/ tetO-Bmpr2R899X)	MOUSE	1	+	++	++	Variable penetrance PH reversed by rhACE treatment Lipid deposition in CMs	44, 45
Inducible overexpression mutated BMPR2R899X targeted to SMCs (Tg+/Tagln-rtTA; Tg+/tetO-Bmpr2R899X)	MOUSE	1	+++	++			43
Bmpr2+/−	MOUSE	1	+/−	+/−	+/−	More susceptible to develop PH under stress conditions	47–51
Bmpr2+/R899X	MOUSE	1	+	+	−		46
Bmpr2+/R899X; Smad1+/−	MOUSE	1		++	+		46
Bmpr2flox/flox; Tg+/ALK1-Cre	MOUSE	1	++	+		Incomplete penetrance	53
Bmpr2+/Δ71Ex1	RAT	1	+	+	+	Incomplete penetrance	54
Bmpr2+/Δ527Ex1 or Bmpr2+/Δ16Ex1	RAT	1	−	−	−	Increased responsiveness to 5-LO-mediated inflammation	55
Alk1+/−	MOUSE		+	+	+		57
Eng+/−	MOUSE		+	+			58
Cav-1−/−	MOUSE			++	+++		59
Kcnk3Δ94ex1/Δ94ex1	RAT		+	+		Incomplete penetrance Low expressivity	63
Ucp2−/−	MOUSE		+	+/−	+		93
Sirt3−/−	MOUSE		++	++	+		94
Ucp2−/−; Sirt3−/−	MOUSE		++++	+++	++	Development of plexiform-like lesions Insulin resistance	95
Irp1−/−	MOUSE			+			99
Nfu1G206C/G206C	RAT		++	+			100
Iron-deficient diet	RAT		++	++			105
Overexpression of IL-6 by airway epithelial cells (Tg+/Scgb1a1-Il6)	MOUSE		+	+		Increased responsiveness to chronic hypoxia	68
Overexpression of TNFα by airway/alveolar epithelial cells (Tg+/Sftpc-Tnfa)	MOUSE			++	+	Causes airspace enlargement	70
Tet2flox/flox; Tg+/Vav1-Cre	MOUSE		+	+	+		75
Hif2aG536W/G536W	MOUSE		+	++	+		78
Egln1flox/flox;Tg+/Cdh5-Cre	MOUSE		++	+++	++	Robust PH prevented by loss of Hif2a function in ECs	79, 81
Egln1flox/flox;Tg+/Tie2-Cre	MOUSE		+++	+++	++	Robust PH prevented by loss of Hif2a function in ECs Development of plexiform-like lesions	80
VhlR200W/R200W	MOUSE		++	++		Partial rescue of the phenotype by reduced gene dosage of Hif2a	82
Ppargflox/flox;Tg+/Tagln-Cre	MOUSE		+	+	+		84
Overexpression of Fra2 (Tg+/H2k-Fra2)	MOUSE	1 (3)	+++	+			88–90
Overexpression of S100A4/Mts1 (Tg+/Hmgcr-s100A4)	MICE		+++	+		Development of plexiform-like lesions Specific to females	114, 115
Dexfenfluramine	MOUSE		+	+	−	Prevented by loss of Tph1 function	116
overexpression SERT targeted to SMCs (Tg+/Tagln-SERT)	MOUSE		++	+	+		117
cyclophosphamide	MOUSE RAT RABBIT	PVOD	++	+ (mice) ++ (rat)	++	Low penetrance in mice	121
Mitomycin-C	RAT	PVOD	++	+	++	Female more sensitive	122
SIV or SHIV-nef infection	MACAQUE	1	+++	++		High maintenance costs Variability among animals	127–131
Overexpression of a gag−pol−deleted HIV-1 provirus regulated by the viral promoter	RAT	1	+	+/−		Increase responsiveness to PH induced by chronic hypoxia exposure or chronic administration of cocaine	132–134
Chronic Hypoxia	MOUSE RAT	3	+		+	Simple implementation Low cost Helpful to determine whether genetically modified animals are protected or more susceptible to the disease	139, 140
Cigarette smoke exposure	MOUSE RAT HAMSTER GUINEA-PIG	3	++	++		Vascular remodeling often precedes airspace enlargement	147–152
Bleomycin	MOUSE RAT	3	+			Simple implementation Low cost Limited vascular remodeling	154–157
Bleomycin + MCT	RAT	3	+++		+++	Simple implementation Low cost Marked pulmonary vascular remodeling	156, 157
Intra-tracheal adenovirus−mediated active TGF-β1 gene transfer	RAT	3	+++		+++		159, 160
Fawn-Hooded Rat	RAT	3 (1)	++		+	Spontaneous development of PH at sea level. Increased severity upon exposure to mild hypoxia PH associated with defects in lung development	174–178
Transverse Aortic Constriction	MOUSE RAT PIG	2	++	++	+/−	Robust model of load stress Doesn’t model human disease well Surgical expertise required	182–185
ApoE −/−	MOUSE	2	+	+	−		189
AKR	MOUSE	2	+	+	+	Easy to implement Mild phenotype	191, 192
Sugen + HFD	ZSF1 RAT	2	++	++	+	Robust phenotype Doesn’t model human disease well (Sugen)	195
Supracoronary Banding + HFD or Olanzapine	RAT	2	++	+++	++	Robust model Surgical expertise required	196

1. Animal models used in the investigation of PAH

Monocrotaline

Although tested in large animals such as minipigs and dogs^{3, 4}, injection of monocrotaline (MCT) in rat is the more classical and widely used in vivo PAH model for over half a century. Primarily carried out in rats due to the inability of mice to convert MCT to its active pneumotoxic metabolite (monocrotaline pyrrole) by hepatic cells, a single subcutaneous or intraperitoneal injection of MCT (60mg/kg in most cases) is known to produce robust pulmonary vascular remodeling^{5, 6}. This in turn causes marked elevation of the mean PA pressure, RV dysfunction and death within 4–6 weeks⁷. Aside from the its direct pulmonary effects, MCT also causes significant toxicity in multiple organs, which can add confounding effects. The response to MCT is dose sensitive, with lower doses (30–40 mg/kg) causing acute muscularization of distal PAs, increased PVR and decreased cardiac output (CO) at 4 weeks after MCT administration with almost complete recovery at 8- and 12-weeks post MCT injection^{8, 9}. Mechanistically, MCT initially induces PA endothelial cell dysfunction, interstitial oedema, vasoconstriction, perivascular inflammatory infiltrate and cytokine secretion followed by progressive medial hypertrophy and adventitial fibrosis¹⁰. As observed in PAH patients¹¹, these histological and structural alterations are accompanied by a glycolytic switch in vascular cells with enhanced expression of pyruvate dehydrogenase kinase¹², contributing to their enhanced capacity to survive and proliferate^{11, 13}. Reduced BMPR2 expression¹⁴, increased growth factor signaling (PDGF, TGF-β, etc.)^{15, 16} as well as sustained activation of the DNA damage response^{17, 18}, to name a few, are also key molecular traits shared by the remodeling vasculature of MCT rats and PAH patients. However, the typical manifestation of severe PAH, the so-called plexiform lesions, are not found in MCT rats (Figure 1). To create a model presenting this characteristic, monocrotaline has been combined with either chronic hypoxia exposure¹⁹ or left pneumonectomy²⁰. In these two-hit models, a neointimal pattern of remodeling and severe RV hypertrophy occur providing a model that closely mimics the advanced stage of the disease. Similarly, increased pulmonary blood flow generated by abdominal aortocaval shunt coupled with MCT administration was implemented in rat^{21, 22}. At similar PA pressures, MCT rats with shunt exhibited pronounced RV failure, increased morbidity and mortality but with a similar neo-intimal hyperplastic response when compared with non-shunted MCT-treated animals^{21, 22}. Although these animal models produce more severe hemodynamic changes and histological alterations (Figure 1), they require surgical skills that significantly limit their use. It should be stressed that, besides its low cost and technical simplicity, interest of the MCT model lies in its ability to sequentially reproduce the adaptive and maladaptive responses (fibrotic, metabolic, inflammatory and hemodynamic) of the RV^{7, 23} driven by progressive and severe pulmonary vascular remodelling. Despite some caveats, the MCT rat remains a valuable model to explore the therapeutic benefit of anti-remodeling approaches¹⁰ (Figure 1). In this regard, it is often qualified as “easy to reverse”; a consideration perhaps biased by its popular use and a tendency to preferentially publish positive results.

Figure 1. Schematic progression of pulmonary arterial hypertension (PAH) with animal models commonly used to address different aspects of the pathophysiology and test candidate treatments or preventive interventions.

CM: cardiomyocyte; EndoMT: endothelial-to-mesenchymal transition; PA: pulmonary artery; LV; left ventricle; RV: right ventricle.

Sugen/Hypoxia (Su/Hx)

Although exposure to chronic hypoxia is long recognized to produce PH in rodents, the model doesn’t match the severity seen in PAH patients (see corresponding section below). As a result, the search of a more robust animal model has become a requirement. Taking benefit from findings showing that apoptosis of pulmonary endothelial cells and pruning of the pulmonary artery tree occurs in adult Sprague Dawley (SD) rats chronically treated with the VEGF receptor antagonist Sugen (Su) 5416²⁴, the same group demonstrated that combination of Su (weekly injection) and chronic hypoxia for 3 weeks markedly worsens pulmonary hypertension as witnessed by hemodynamic measurements and the appearance of plexiform-like lesions²⁵. Similar obliterative lesions were noted in SD rats receiving a single injection of Su (20mg/kg) before exposure to hypoxia for 3 weeks followed by a return to normoxia for an additional 10 to 11 weeks²⁶. Temporal examination of hemodynamic and histological changes revealed that RVSP gradually increases to reach a plateau (100 mmHg) at 3–5 weeks after SU exposure, a deterioration correlating with progression of PA vascular remodeling²⁷. Not surprisingly, strain-dependent phenotypic variabilities were documented. This is exemplified by Fisher rats that display an increased susceptibility to develop RV failure and death as compared to SD rats despite comparable hemodynamic afterload and pulmonary vascular remodeling²⁸. In addition, in support of the prevailing model of PAH development; i.e. initial apoptosis of PA endothelial cells (PAECs) followed by extensive proliferation of adjacent PA smooth muscle cells (PASMCs) and apoptosis-resistant PAECs²⁹, the development of the rat Sugen-hypoxic model represents a significant advance providing a complementary alternative to the MCT rat model. Indeed, due to its ability to cause severe and irreversible PH with the appearance of angio-obliterative lesions, the Su/Hx rat model (while certainly not without criticism) has become one of the most used preclinical models of PAH with therapeutic intervention initiated at either 3 weeks (immediately after the end of hypoxia exposure) or 5–8 weeks (after 2–5 weeks of normoxic recovery) post Su injection^{26, 30–32} (Figure 1). More recently, the rat Su/Hx model was adapted to mice³³ in order to take benefit of existing genetically modified mice and thus gain mechanistic insights into the pathogenesis of PAH. Mice subcutaneously injected with Su (20mg/kg once weekly) over a 3-week period of hypoxia were found to develop higher RV pressure (~50mmHg), RV hypertrophy and vascular remodeling compared with hypoxic control mice³³. Contrary to rats, all these aforementioned parameters were found to be decreased over 10 weeks of normoxic follow-up^{33, 34}. Thus, it is clear that rats are not simply huge mice, and that each species possesses advantages and disadvantages. The smaller size of mice makes them more cost-effective as a lower amount of drug is needed. On the other hand, assessment of RV function by echocardiography and closed-chest catheterization is more technically challenging. Finally, although largely assumed to be representative of group 1 PH, exposure to Su with or without chronic hypoxia, was reported by some groups to cause an airspace enlargement reminiscent of emphysema^{24, 35 36}. This suggests that this model may also be helpful for studying group 3 PH. Be that as it may, the rat Su/Hx model should be given priority when the primary purpose is to evaluate the potential of novel new anti-remodeling agents (Figure 1), with mice more useful for assessing genotype-phenotype consequences in transgenics.

BMPR2 transgenic/knockout animals

The first evidence of genetic contributions to PAH was discovered following linkage studies in which heterozygous germline mutations in bone morphogenetic protein receptor 2 (BMPR2) were detected^37–39. Although BMPR2 loss-of-function mutations are the underlying cause in approximatively 75% of patients with hereditary PAH and 10–40% of idiopathic PAH patients, only 20–30% of carriers develop the disease⁴⁰. This low penetrance indicates that additional environmental or genetic factors are required for its clinical manifestation. Based on these findings, multiple genetically engineered mouse lines have been created to address the impact of BMPR2 signaling deficiency on the development of PH. To circumvent the embryonic lethality of Bmpr2^−/− mice⁴¹, a transgenic mouse expressing an inducible and smooth muscle specific dominant-negative form of BMPR2 was generated⁴². Examination of these mice revealed that postnatal activation of the transgene induces a modest increase in PA medial thickness and a mild elevation of RVSP by 8 weeks of age, a phenotype slightly aggravated under hypoxia⁴². Similarly, a significant proportion of mice expressing postnatally a cytoplasmic tail mutation (R899X) of the Bmpr2 gene restricted or not to smooth muscle cells develop PH, as evidenced by an elevated RVSP and marked pulmonary vascular structural changes^43–45. Notably, mice with universal expression of the Bmpr2R899X transgene demonstrated an impaired RV hypertrophic response associated with an accumulation of lipid within cardiomyocytes (CMs), a feature not seen in mice harboring the SMC-restricted conditional, inducible Bmpr2R899X transgene⁴⁵.

Using a knock-in approach that fully preserves the native expression and splicing pattern, Long and colleagues found that Bmpr2^+/R899X mice develop age-related PH with substantial worsening after loss of one copy of SMAD1, an important downstream mediator of BMPR2 signaling⁴⁶. In parallel, different studies were undertaken using Bmpr2^+/− mice. Although contradictory results have been found regarding whether or not heterozygous mice spontaneously develop mild PH under basal conditions^47–51, most of these works demonstrated that Bmpr2^+/− mice are more prone to develop signs of PH under stressed conditions such as prolonged hypoxia^{47, 48}, 5-lipoxygenase- or lipopolysaccharide-mediated inflammation^{49, 50}, IL-1⁵² and chronic serotonin infusion⁵¹. In addition, Hong and coworkers found that a subset of mice with endothelial cells-specific heterozygous or homozygous Bmpr2 inactivation display sustained elevation of RV pressure associated with perivascular accumulation of inflammatory cells and conspicuous occlusion of distal PAs⁵³.

By taking benefit of major advances in rat genome editing, including the CRISPR/Cas9 system, and advantages of rat as a model organism more sensitive to PH development than mice, several Bmpr2-mutated rats were recently generated and characterized⁵⁴. In a first study, zinc finger nuclease technology was employed to target the first exon of the Bmpr2 gene leading to deletion of 71 bp associated with the loss of the translation initial codon. Heterozygous Bmpr2(Δ71) mutant rats showed age-dependent spontaneous mild PH with low penetrance (~20%). Interestingly, interleukin-6 overexpression was identified as a discriminating factor between rats that spontaneously develop PH to those that did not⁵⁴. In a second study, the same methodology was used to create two strains of monoallelic mutant rats with frameshift mutations within exon 1 (Bmpr2^+/Δ527Ex1 and Bmpr2^+/Δ16Ex1) The follow up of these rats over one year revealed no sign of PH⁵⁵. These mutant strains were then assessed for increased responsiveness to several putative PAH triggers, namely, 5-lipoxygenase exposure, MCT, Sugen, chronic hypoxia and Su/Hx. Among these, only the combined presence of Bmpr2 haploinsufficiency and overexpression of 5-lipoxygenase was shown to produce additive/synergistic effects on PAH development⁵⁵. Collectively, these findings somewhat disappointing in view of the incomplete penetrance and variable expressivity of rodent Bmpr2 mutant phenotypes support a “two-hit” model in which two independent insults (BMPR2 deficiency and stressful experience) are assumed to be necessary for the development of PAH.

Rodent models related to other genes involved in heritable PAH

Since 2000, mutations in other genes related to BMPR2 signaling, including mutations in Activin receptor-like kinase 1 (ALK1), Endoglin (ENG), Caveolin-1 (CAV-1), and SMAD8 have been identified in PAH patients⁵⁶. To examine the mechanisms linking mutations of these genes to the development of PAH, genetically modified mice have been generated. Most but not all Alk1 heterozygous mice were shown to develop signs of PH by 9 weeks of age, including a significant elevation of RVSP associated with RV hypertrophy and remodeling of peripheral arteries⁵⁷. A similar phenotype was observed in Eng^{+/− 58} and Cav1^−/− mice⁵⁹. More recently, screening of patients with familial or idiopathic PAH by whole exome-sequencing identified mutations in the potassium (K⁺) channel subfamily K member 3 (KCNK3, also called TASK1)⁶⁰. Decreased expression and activity of K⁺ channels in PASMCs is known to cause membrane depolarization, enhanced calcium (Ca²⁺) influx and increased cytosolic Ca²⁺ concentration which in turn, stimulates vasoconstriction and proliferation⁶¹. However, neither male nor female Kcnk3^−/− mice manifested apparent signs of PH⁶² and only 45% (3/8) of homozygous mutant rats (Kcnk3^{Δ94ex1/Δ94ex1}) exhibited RVSP above 40mmHg at 1 year⁶³.

Animal models related to inflammation

Pulmonary inflammation is recognized as a key pathogenic driver of PH. This is underscored by the recruitment of inflammatory cells, such as macrophages, T and B-lymphocytes and mast cells around and within the remodeled vessels and the presence of high circulating levels of pro-inflammatory mediators in PAH patients that correlate with clinical outcomes^64–66. Among these, the pro-inflammatory cytokine IL-6 has gained considerable attention with several studies demonstrating augmented serum concentration in severe PAH and up-regulation by smooth muscle cells as a result of deficient BMPR2 signaling^{64, 67}. To test the hypothesis that increased expression of IL-6 actively participates in the remodeling process, transgenic mice that overexpress IL-6 in the airway epithelium were generated⁶⁸. Detailed phenotypic characterization of these mice revealed a subtle elevation of the RVSP under normoxic conditions compared to littermate controls (~30 vs 20mmHg). However, transgenic mice exposed to 3-week hypoxia developed significantly worse PH (RVSP ~65mmHg) compared to hypoxic control mice (RVSP ~25mmHg) with extensive pulmonary vessel occlusion⁶⁸. In support of the pathological role of IL-6 in PAH, subsequent studies shown that IL-6 directly stimulates survival of pulmonary vascular cells and that genetic or pharmacological inhibition of IL-6 signaling attenuates PH in multiple preclinical models⁶⁹. Taken together, these findings indicate that IL-6 overexpression is necessary but not sufficient to cause PAH. Aside from IL-6, overexpression of tumor necrosis factor (TNF)-α by lung epithelial cells was documented to produce a marked elevation of RV systolic pressures in mice reared at moderate altitude (Denver, CO, 5280 feet). Although histological and morphometric analysis of pulmonary vessels were not conducted, a noticeable airspace enlargement (probably due to impaired alveolarization ± emphysema) was noted in transgenic mice⁷⁰, potentially making them a more appropriate model for group 3 PH. The complex interplay between lung and vascular inflammation, perhaps highlighted by the overlap of patients with Group 1 PAH and Group 3 PH associated with lung disease is exemplified by work demonstrating that the TNF-related apoptosis-inducing ligand (TRAIL) can regulate neutrophil apoptosis in lung injury models to resolve inflammation ⁷¹ but while delivery of recombinant TRAIL reduces lung injury and fibrosis⁷², deficiency of TRAIL is protective to development of PAH^{73, 74}.

Importance of inflammation in PAH onset and progression is further highlighted by a recent study examining PAH patients harboring rare and predicted deleterious variants in the Tet methylcytosine dioxygenase 2 (TET2) gene⁷⁵, encoding for an enzyme known to promote DNA demethylation and repress cytokine gene expression in inflammatory cells⁷⁶. The authors demonstrated that TET2 variant carriers exhibit increased overall inflammation compared to age/sex-matched healthy and PAH non-mutated patients. Similarly, they found that conditional inactivation of Tet2 in hematopoietic cells resulted in persistent lung inflammation associated with increased muscularization of distal PAs and mild elevation of the RVSP in 7–10-month-old mice; a phenotype reversed by administration of IL-1β neutralizing antibodies⁷⁵.

Animal models related to abnormal activity of transcription factors

As the name implies, Hypoxia-inducible factors alpha (HIFα) family members are pivotal transcription factors that orchestrate adaptive responses under low oxygen conditions. Under normoxia, the HIFα protein levels are low due to constant degradation via the collaborative action of prolyl hydroxylases and the von Hippel-Lindau (VHL)-containing E3 ubiquitin ligase complex. Not surprisingly, both HIF1α and HIF2α isoforms have been extensively studies in PH⁷⁷. Increased activation of the different isoforms was reported to regulate a large array of genes that promote a metabolic switch to glycolysis, cell survival and proliferation as well as endothelial to mesenchymal transition⁷⁷. Mice bearing a missense gain-of-function mutation in the Hif2a gene were shown to develop PH with a significant increase in RVSP in heterozygotes (46mmHg) and an even more robust increase in homozygotes (65mmHg)⁷⁸. In direct connection with this, consistent in vivo results were obtained by different groups regarding the implication of prolyl hydroxylases and their negative regulation of HIF abundance. Published data demonstrated that loss of prolyl hydroxylase domain protein 2 (PHD2, encoded by the Egln1 gene) targeted to endothelial cells induces severe PH in mice (RVSP reaching 60 to 75 mmHg in 3.5-month-old animals)^79–81. Occlusive pulmonary vascular remodeling, severe RV hypertrophy and failure and progressive mortality were seen in mutant mice. Consistent with prior findings showing that PHD2 deficiency induces stabilization of HIFα transcription factors, both HIF-1α and HIF-2α were found elevated in lungs of Egln1 conditional knockout mice. To determine whether HIF-1α or HIF-2α derepression causes the phenotype, endothelial specific Egln1 inactivated mice were bred with Hif1a^flox/flox or Hif2a^flox/flox mice. Using this elegant approach, the authors demonstrated that activation of HIF-2α (not HIF-1α) is responsible for the obvious alterations seen in Egln1 conditional knockout mice^{79, 80}. Similarly, mice homozygous for the R200W mutation in the tumor suppressor gene Vhl developed pulmonary hypertension independently of polycythemia similar to Chuvash patients. Lungs from homozygous carriers displayed pulmonary vascular remodeling, hemorrhage, edema, and macrophage infiltration, aggravated by fibrosis in older mice. As depicted in Egln1 conditional knockout mice, loss of 1 allele of Hif2α partially rescued the lung phenotype of Vhl^R200W/R200W mice⁸², providing genetic evidence that dysregulated HIFα signaling contributes to aberrant pulmonary vascular remodeling in PH. Besides HIFα signaling, a large body of evidence implicate peroxisome proliferator-activated receptor gamma (PPARγ) as a master regulator of vascular homeostasis in the lung, with diminished expression in remodeled PAs of PH patients⁸³. Activation of PPARγ secondary to BMPR2 signaling was shown to exert anti-proliferative effects on PASMCs⁸⁴. Additional protective effects of PPARγ also include anti-inflammatory and anti-vasoconstrictive effects. Accordingly, mice deficient for Pparg in smooth muscle cells presented a slight elevation of the RVSP as compared to controls (29 vs 21mmHg)⁸⁴. Although transcription factors are usually regarded as difficult to target, several preclinical studies have shown that PPARγ agonists (rosiglitazone or pioglitazone) significantly prevent or reverse the pathological hallmarks of PAH⁸⁵, adding further evidence that small-molecule modulation of transcription factor activity has substantial potential for PAH.

Systemic sclerosis (SSc) is a multisystem autoimmune disease characterized by chronic inflammation, proliferative vasculopathy and fibrosis of the skin and various organs. The autoimmune process can affect the lung resulting in scarring and vascular injury contributing to the development of PAH with significant negative impact on survival⁸⁶. Although several animal models of SSc have been generated to dissect the cellular and molecular mechanisms of SSc, rare are those reproducing the pulmonary manifestations of the disease, pushing researchers to borrow animal models from idiopathic pulmonary fibrosis (IPF) to get valuable information about the pathogenesis of the human disease. This is illustrated by a recent randomized trial showing that Nindenadib (approved for the treatment of IPF) is effective in slowing progression of SSc-associated pulmonary fibrosis ⁸⁷. Overexpression in mice of Fos-related protein 2 (FRA2), a subunit of the transcription factor AP-1, was shown to replicate many features of SSc with predominant alterations in the lung^{88, 89}. Comprehensive characterization of the pulmonary phenotype of these mice at 12–17 weeks of age demonstrated that perivascular inflammation associated with concentric laminar lesions are the first apparent pathological alterations followed by prominent interstitial fibrosis^{88, 89}. Accordingly, mutant mice exhibited a modest but significant elevation of RVSP (~30mmHg)⁹⁰. Although enhanced PDGF-BB-mediated PDGFRβ signaling was proposed as a molecular link of adverse lung remodeling in Fra-2 transgenic mice⁸⁹, how FRA2 mechanistically contributes to lung scaring and pulmonary vascular remodeling remains largely unknown. Nevertheless, this animal model may provide invaluable clues in better understanding SSc-related PAH and a promising entry point to the development of much-needed new therapies.

Animal models related to mitochondria dysfunction

Metabolic dysfunctions are well-established critical components of PAH^{91, 92}. Like cancer cells, vascular cells in PAH exhibit multiple metabolic alterations including a shift from oxidative phosphorylation to glycolysis⁹². This metabolic rewiring confers growth and survival advantages, in part, through hyperpolarization of the mitochondrial membrane and epigenetic regulation of gene expression programs⁹². Guided by accumulating evidence pinpointing occurrence of metabolic dysfunction in PAH patients, signs of PH were researched in mice deficient for some factors critically implicated in the regulation of mitochondrial function. Mice lacking mitochondrial uncoupling protein 2 (UCP2) or Sirtuin 3 (SIRT3, a mitochondrial deacetylase) were reported to spontaneously develop pulmonary vascular remodeling and mild PH^{93, 94}. Interestingly, loss-of-function UCP2 and SIRT3 polymorphisms were found often both in the same patient in a homozygous or heterozygous manner and their presence correlated positively with the degree of PAH upon referral and outcomes (death, transplantation). Accordingly, Sirt3/Ucp2 double knockout mice displayed an increased severity of PH associated with RV dysfunction and dilatation, along with insulin resistance. Histological examination of lungs from double-mutant mice revealed the presence of multifocal inflammatory plexiform-like lesions occurring before the elevation of the RVSP (Figure 2), making these mutant mice one of the rare mouse models recapitulating many features of human PAH and thus a significant tool in the armamentarium of the PAH researcher⁹⁵.

Figure 2. Plexogenic arteriopathy and severe PAH in mice lacking both Sirt3 and Ucp2.

(A) Representative histology (Hematoxylin-Eosin staining) of a lung from a Sirt3^−/−;Ucp2^−/− mouse shows numerous plexogenic lesions (back arrows). (B) Representative photomicrograph of confocal fluorescence immunohistochemistry of small pulmonary arteries from wild-type WT and Sirt3^−/−;Ucp2^−/− mice shows smooth muscle cells (green) and endothelial cells (magenta) in plexogenic lesions. (C) Representative tracings of RV and PA pressures from WT and Sirt3^−/−;Ucp2^−/− mice by close-chest right heart catheterization through the jugular vein shows that Sirt3^−/−;Ucp2^−/− mice have a significant increase in RV and PA pressure. SMA: smooth muscle actin, vWF: von Willebrand Factor (marking endothelial cells), DAPI (marking nuclei). (Images courtesy of Dr Gopinath Sutendra.)

The metabolic theory of PH is also supported by clinical reports describing the development of a metabolic syndrome associated with the occurrence of PAH in individuals carrying mutations in genes involved in iron-sulfur cluster biogenesis^{96, 97}. Mainly assembled by the mitochondria, iron-sulfur clusters are essential cofactors for many proteins involved in immune response and various core cellular functions such as mitochondrial respiration, DNA replication and repair, and gene expression⁹⁸. To examine the functional relevance of mitochondrial iron homeostasis in the onset of PH, different animal models have been established. Deletion, mutation, or knockdown of multiple iron metabolism-related genes such as Iron regulatory protein 1⁹⁹, NFU1¹⁰⁰, BolA family member 3¹⁰¹ and Frataxin¹⁰² were documented to promote hemodynamic manifestations of PH in rodent by decreasing oxidative phosphorylation. Along with defects in mitochondrial iron metabolism, iron deficiency was found prevalent in PAH patients and correlated with worse clinical conditions^{103, 104}. Importance of iron homeostasis in the cardio-pulmonary system was further supported by experimental animal studies showing that SD rats fed an iron-deficient diet for 4 weeks display an abnormal muscularization of PAs associated with up-regulation of HIFα and acquisition of a glycolytic state as well as increased PA pressure and RV hypertrophy¹⁰⁵.

Animal models related to drug-induced PH and PVOD

According to the current classification, PAH can be associated with drug use¹. Indeed, several types of medications such as anorexigens or anti-cancer agents as well as abuse consumption of illicit drugs were confirmed to be risk factors for PAH. One of the best-documented causes is the appetite-suppressants fenfluramine and aminorex. This association was derived from epidemiological studies conducted in the 1990s showing that the risk of developing PAH was 10 to 30 times higher in people who took certain appetite suppressants than in the general population¹⁰⁶, eventually leading to their withdrawal from the world market. Since fenfluramide and aminorex are serotonin transporter substrates and indirect serotonergic agonists, many experimental studies have concentrated on further linking serotonin signaling to PAH development. In PAH patients, endothelial-derived serotonin was shown to activate serotonin receptors and enter the cell via the serotonin transporter (SERT) to induce contraction and proliferation of adjacent PASMCs^107–109. The molecular mechanisms implicated in this response are complex and involve the modulation of voltage-gated K⁺/voltage-operated Ca²⁺ channels^{110, 111} and the serotonylation (covalent linkage of serotonin to proteins) of several intracellular targets^{112, 113}. Not surprisingly, serotonin was reported to stimulate expression and release of various factors contributing to the pro-proliferative phenotype of PAH-PASMCs, including Ca^2+-dependent S100A4 (Mts1) protein¹⁰⁹, for which overexpression promotes the development of plexiform-like lesions with minimal elevation of the RVSP in female mice selectively^{114, 115}. The instrumental role of serotonin signaling in the development and maintenance of PAH is strongly suggested by published data showing that pulmonary vascular remodeling and elevated RVSP were seen in mice chronically administered with dexfenfluramine¹¹⁶ or displaying increased internalization of serotonin secondary to SERT overexpression^116–118.

In addition to appetite suppressants, tyrosine kinases inhibitors (TKI) such as Dasatinib and Xalkori have been associated with PAH¹¹⁹. Results gained from experimental studies demonstrated that PAECs exposed to these anti-cancer-agents present cell damage, as reflected by initiation of a global endoplasmic reticulum stress response, mitochondrial dysfunction and induction of apoptosis. However, the insults appear insufficient to cause PH, as rats chronically treated with these TKIs do not experience any major changes in pulmonary hemodynamic parameters and vascular remodeling^{9, 120}. However, when given prior to PH inducers (MCT or hypoxia), these anti-cancer agents were shown to potentiate the development of the disease by exacerbating hemodynamical and structural changes^{9, 120}, providing strong support to the two-hit theory of the pathophysiology of TKI-PAH.

Pulmonary veno-occlusive disease (PVOD), an uncommon form of PAH characterized by preferential remodeling of pulmonary veins, has long suffered from the lack of an animal model. Following repetitive case reports mentioning the occurrence of PVOD in patients receiving chemotherapy, the ability of alkylating agents to cause PVOD was tested in different animal species. In a first study, cyclophosphamide was reported to induce venular remodeling and PH in a dose-dependent manner preferentially in female rats and rabbits, whereas mice were less responsive¹²¹. In a second study, the same group demonstrated that a single injection of mitomycin-C produced a slight elevation of mPAP accompanied by a remodeling of pulmonary venous and capillary compartments¹²². Together, these findings highlight a probable cause-effect relationship between alkylating class chemotherapeutic drugs and PVOD. Of note, although mutations in the EIF2AK4 (also called GCN2) gene were identified as the major genetic cause of PVOD¹²³, Eif2ak4^{Δ152Ex1/ Δ152Ex1} knockout mutant rats failed to present elevation of RVSP¹²⁴.

Human immunodeficiency virus (HIV)-associated PAH (group 1.4.2)

PAH is overrepresented in individuals diagnosed as HIV-positive and again more prevalent in HIV-infected intravenous drug-abusing users^{125, 126}. Nonhuman primates infected with simian immunodeficiency virus (SIV) or SHIV-nef (a chimeric viral construct containing the HIV nef gene in a SIV backbone) have been used to mimic the salient features of HIV-1 infection in humans and acquire scientific knowledge on the occurrence of pulmonary vasculopathy. Consistent with that seen in human, pulmonary vascular lesions characterized by intimal and medial hyperplasia, adventitial fibrosis and recruitment of inflammatory cells were commonly encountered in SIV or SHIV-infected animals^127–130. Longitudinal evaluation of cardiopulmonary function in rhesus macaques before and after SIV infection demonstrated that around 50% of animals had elevated mean PA pressure greater than 25mmHg 6-month post infection, among these 50% exhibiting worsening at 1 year¹³¹. All animals presented evidence of pulmonary inflammation. Histological and metabolic analyses revealed an increased collagen deposition and glucose uptake in RV from SIV-infected macaques with hemodynamic evidence of PAH. However, no RV hypertrophy was observed and RV function was preserved¹³¹. Due to complex infrastructure to house nonhuman primates and high maintenance costs, efforts have focused on the development of small animals for investigating the pathogenesis of HIV. Signs of PH were researched in HIV-1 transgenic rats exposed or not to chronic stress. Different results were obtained with some showing a significant elevation of the RVSP in HIV-1 transgenic rats¹³², whereas other found no difference¹³³. However, expression of the HIV-1 transgene was consistently shown to increase responsiveness to PH induced by either chronic hypoxia exposure¹³³ or chronic administration of cocaine¹³⁴. Although findings gained from this model can be difficult to extrapolate to human, it offers significant advantages as a noninfectious small-animal model of HIV-PAH.

2. Animal models of pulmonary hypertension due to hypoxia and/or lung diseases (Group 3 PH)

Hypoxic PH often complicates the course of patients with advanced lung diseases including but not limited to: sleep apnea, chronic obstructive pulmonary disease (COPD), idiopathic pulmonary fibrosis (IPF), and developmental lung disorders¹³⁵.

Animal models related to chronic hypoxia

Although often employed to draw conclusions in group 1 PH, exposure to animals to chronic hypoxia is largely considered as a model recapitulating the pulmonary vascular changes observed during exposure to prolonged high altitude. It is well established that exposure to rodents to chronic alveolar hypoxia (usually 3 weeks) primary causes constriction of resistant vessels in the pulmonary vascular bed. When hypoxia is sustained, a moderate increase in the muscularization of proximal and distal pulmonary vessels occurs contributing to elevation of pulmonary vascular resistance. Several non-exclusive mechanisms for the de novo muscularization of small PAs have been proposed including distal expansion of pre-existing vascular smooth muscle cells and endothelial to mesenchymal transdifferentiation^136–138. The magnitude of the hypoxic response varies across species with further intra-species differences relating to the strain, age and gender¹³⁹. For instance, for the same hypoxic challenge, SD rats are more prone to develop pulmonary vascular remodeling than C57BL/6 mice¹⁴⁰. Furthermore, younger animals are more susceptible to develop severe PH. Indeed, newborn calves placed under hypoxic conditions develop a more severe PH accompanied by a pronounced medial wall and adventitial thickening of PAs, as compared to rodent kept under chronic hypoxia¹⁴¹. The major shortcomings of this model are the moderate severity of PH, as determined in part, by the limited extent of pathological remodeling of the pulmonary vasculature and inability to replicate the maladaptive events leading to RV failure as well as the reversibility of structural and functional alterations in the pulmonary vasculature upon return to hypoxia¹⁴² (Figure 1). Despite these limitations, the chronic hypoxia model has and continues to provide us valuable insights into the vasoconstrictor component of PH. Past results obtained with this model are the groundwork for currently available PAH-based therapies Moreover, it offers the advantage to be relatively inexpensive, easy to implement, highly reproducible and remains an essential tool to appreciate the susceptibility of genetically modified mice to PH development.

Exposure for rodents to chronic intermittent hypoxia (CIH) is also used to mimic the cyclic pattern of hypoxemia observed in patients experiencing obstructive sleep apnea (OSA)^{143, 144}; a condition complicated by mild or moderate PH if untreated¹⁴⁵. Indeed, exposure to CIH for several weeks were shown to induce signs of PH, as illustrated by a thickening of small pulmonary vessels, a mild elevation in RV systolic pressure and a slight increase in RV mass^{144, 146}.

Mild-to-moderate PH is frequently encountered in COPD, an irreversible lung condition due to chronic and persistent inflammatory response and leading to chronic bronchitis and emphysema. Because the disease most commonly results from long-standing inhalation of irritants, in particular cigarette smoke, cigarette smoke exposure has become one of the most utilized animal models. Findings from mice, rats, hamsters or guinea pigs revealed that chronic exposure to cigarette smoke, usually over a 4–6-month period, is accompanied by an airspace enlargement, thickening of the PA wall and elevated RVSP^147–152. Analysis of the time course of lung injury in rodent models repeatedly demonstrated that pulmonary vascular remodeling and PH may precede the development of emphysema^{148, 149}, providing evidence that cigarette smoke has a direct impact on the pulmonary vasculature and that PH in COPD is not necessarily driven by loss of capillaries in emphysema and chronic alveolar hypoxia.

Idiopathic pulmonary fibrosis (IPF)

Different animal models have been generated displaying parenchymal lung fibrosis with varying degrees of pulmonary vascular remodeling and PH. Although decried due to its tendency to spontaneously resolve in young mice or after a single challenge, administration of the chemotherapeutic agent bleomycin (BLM) in rodents is the best-characterized animal model of pulmonary fibrosis, and as such, is considered as the greatest model for preclinical testing¹⁵³. In this relatively easy to handle animal model with many variations in terms of dosage and route of delivery, muscularization of pulmonary vessels accompanied with or without the presence of mild PH have been repeatedly described^154–157. Since elevation of RVSP in BLM-challenged rodents is most often modest and occurs concomitantly with the appearance of fibrotic lesions, a two-hit rat model by sequential exposure to BLM and MCT was recently developed to more adequately recapitulate group 3 PH^{156, 157}. As revealed by histological analyses, hydroxyproline assay, lung function testing and right heart catheterization, the combination of these insults results in pronounced fibrosis associated with altered mechanical properties of the respiratory system and a robust PH phenotype^{156, 157}, making this model more appropriate to assess the impact of drugs on the development of PH in pre-existing pulmonary fibrosis. Besides these drug-induced animal models, intratracheal delivery of an adenoviral gene vector encoding biologically active transforming growth factor (TGF)-β1 has been shown to produce prolonged severe fibrosis in rat lung leading to marked pulmonary vascular remodeling and PH^{158, 159}. Using this animal model, Bellaye and coworkers found that Macitentan (a dual endothelin receptor antagonist approved for the treatment of PAH) given two weeks after instillation of AdTGF-β1 for two weeks significantly prevents progression of lung fibrosis and PH¹⁶⁰.

Lung immaturity

In the early 1990s, the British epidemiologist David Barker proposed that adverse influences early in development could elicit permanent structural changes thereby predisposing individuals to an increased risk of disease in adulthood ¹⁶¹. This concept, commonly referred to as “the Fetal Origins of Adult Disease” or “Fetal Programming of Adult Disease” expanded greatly during the past decades. It is now clearly established that extremely preterm-born infants and survivors of bronchopulmonary dysplasia (BPD, an important complication of prematurity) have a higher risk to develop COPD later in life^162–164. Although less documented, the long-term effects of in utero and early life factors on pulmonary vascular dysfunction in adulthood have been highlighted by several longitudinal cohort studies. Jayet and coworkers found that preeclampsia, a complex placental dysfunction disorder and prenatal risk factor for chronic lung disease in very low birth weight infants, is associated with a persistent dysfunction in the pulmonary circulation of the offspring at age 13–14 years with a roughly 30% higher PA pressure than in control subjects¹⁶⁵. In a direct link with this concept, premature birth was associated with an increased risk of adult PH in a Swedish population-based study¹⁶⁶. Similarly, in a prospectively followed US cohort, young adults born preterm demonstrated early vascular disease characterized by elevated mean PA pressure (with 45% > 19mmHg) and RV dysfunction^{167, 168}. These studies provide compelling evidence that interruption of normal pulmonary vascular development due to prematurity or early lung injury causes short-term and long-term cardiopulmonary sequelae and increases the risk of developing PH later in life. The predisposition of perinatal insults to pulmonary vascular dysfunction in adulthood is also exemplified by a series of studies conducted in rodents. Treatment of newborn rats with Sugen (Su) is known to result in reduced pulmonary angiogenesis, increased PA wall thickness, and decreased alveolarization, mimicking anomalies encountered in premature newborns who develop BPD^{169, 170}. Despite catch-up growth, these anomalies persist into adulthood since diminished PA density, increased PA wall thickness, airspace enlargement and RV hypertrophy were also detected in adult rats¹⁷⁰. The durable impairment of postnatal hyperoxia on the adult pulmonary vascular bed and RV function was also underscored by Goss and coworkers ¹⁷¹. In this study, newborn rats were exposed to hyperoxia for 2 weeks to mimic some aspects of BPD. At the end of the 2-week exposure period, animals were then allowed to age out 1 year in standard animal housing. Aged rats that had experienced hyperoxia during their first days of life demonstrated a decreased pulmonary vascular surface area and a significant increase in RV systolic pressure and hypertrophy ¹⁷¹. Additionally, brief perinatal hypoxia was documented in rats to increase severity after re-exposure to hypoxia at 2 weeks of age ¹⁷² and hypoxia exposure of adult rats treated with SU as newborns led to a more severe RV hypertrophy compared with those exposed only as adults, indicating that early lung injuries are accompanied with long-term pulmonary sequelae predisposing to the development of PH. In a similar fashion, postnatal treatment of newborn rats with glucocorticoids decreased alveolarization and PA density, leading to long-term alterations including increased PA pressure ¹⁷³.

The Fawn-Hooded rat (FHR) strain, characterized by a chromosome 1 abnormality and increased serotonin blood levels due to a platelet storage pool deficiency, is reported to spontaneously develops pulmonary hypertension at sea level, a phenotype amplified with exposure to mild hypoxia^{174, 175}. Mechanistic studies conducted in adult rats have revealed that the chromosome 1 abnormality is responsible for mitochondrial defects. Indeed, as observed in PAH-PASMCs, PASMCs from FHR rats exhibit fragmented and hyperpolarized mitochondria along with diminished expression of superoxide dismutase 2 (SOD2) leading to normoxic activation of hypoxia-inducible factor 1 alpha (HIF-1α) and subsequent downregulation of potassium voltage-gated channel subfamily A member 5 (Kv1.5)^{176, 177}, all contributing to their abnormal pro-proliferative and apoptosis-resistant phenotype. As such, this genetic strain is classically used as a model of pulmonary hypertension in adult rats. Interestingly, when compared to the Sprague Dawley rat strain, lung architecture of FHR pups exhibits a marked airspace enlargement secondary to impaired alveolarization¹⁷⁸. Supplementation in oxygen during the combined prenatal and early postnatal period significantly improved alveolarization and reduced the development of PH as reflected by measurements of right ventricular hypertrophy and wall thickness of small PAs, reinforcing the notion that pulmonary hypertension detected in adult life originates at least in part from developmental defects¹⁷⁸.

3. Animal Models of Group 2 Pulmonary Hypertension

Group 2 PH is thought to occur as a result of increased back pressure due to left atrial, left ventricular, or valvular (mitral or aortic) pathology. While many models that have been proposed to study the pathogenesis of valvular disease¹⁷⁹, systolic heart failure¹⁸⁰, and diastolic heart failure¹⁸¹, very few reports have directly assessed the degree of pulmonary vascular remodeling in these models. While no models perfectly recapitulate the heterogeneity of Group 2 PH in humans, a number of models have been proposed that provide mechanistic insight into how comorbidities known to associate with Group 2 PH may cause or exacerbate pulmonary vascular remodeling.

Pressure overload

Elevated left sided filling pressures are a key hemodynamic finding in the diagnosis of Group 2. A robust model for causing pressure overload in animals is the transverse aortic banding model^{182, 183}. In addition to the development significant fibrosis and upregulation of fibrotic gene programs in the lungs of mice with heart failure, pressure overload results in a large leukocytic infiltrate with concomitant upregulation of pro-inflammatory cytokines, injury, muscularization of small vessels, and increase in endothelin-1 and nitric oxide mediated signals. While a shortcoming of this model is the lack of clinical correlation with an acute load imposed by aortic banding, it remains a robust model for studying the load-dependent changes in pulmonary vascular remodeling. Notably, this model has been replicated in larger animal models that more closely approximate human pulmonary vascular pathology^{184, 185}.

Obesity and Metabolic Syndrome

Given the known risk factors of diabetes and obesity in exacerbating Group 2 PH^{186, 187}, a number of models have utilized mice prone to the development of metabolic syndrome to identify mechanisms by which obesity and diabetes contribute to Group 2PH. Hansmann et al utilized the ApoE^−/− knockout mouse, which is prone to development of atherosclerotic heart failure, to show that the development of insulin resistance was associated with histologic and hemodynamic features of pulmonary hypertension that could be reversed with rosiglitazone, an activator of PPARγ^{188, 189}. The AKR mouse, a mouse strain susceptible to development of hypertension and metabolic syndrome¹⁹⁰, has been used to study the mechanisms underlying obesity-related pulmonary vascular remodeling in Group 2 PH^{191, 192}. Strengths of these models are that they are relatively simple models to implement for the studying the effect of metabolic syndrome upon pulmonary vascular remodeling, but they each have the notable limitation of producing only a mild pulmonary vascular phenotype.

Multi-comorbidity Interactions

A number of studies have now shown that a combination of comorbid conditions induced by hemodynamic stressors, metabolic stressors, vascular injury, and/or inflammatory stressors are more robust models of Group 2 PH. Inflammation has been postulated to play an important role in the pathogenesis of Group 2 PH¹⁹³. Lawrie et al expanded on previous work by Hansmann et al to show that pulmonary vascular remodeling in ApoE^−/− mice was at least partially dependent on inflammation mediated by IL-1¹⁹⁴. Treatment with an IL-1 receptor antagonist prevented high-cholesterol diet-mediated pulmonary vascular remodeling. Lai et al utilized a combination of vascular injury induced by Su and a metabolic stress induced by high fat diet in obese, ZSF1 (leptin receptor deficient) rats to cause pulmonary vascular remodeling with concomitant cardiac hypertrophy¹⁹⁵. They further demonstrated that chronic treatment with nitrite and metformin prevented pulmonary vascular remodeling via activation of SIRT3 and AMP-activated protein kinase (AMPK). Finally, a recent model by Ranchoux et al utilized a combination of hemodynamic stress induced by supracoronary banding and metabolic stress induced by either high fat diet or olanzapine treatment (an anti-psychotic drug with known side effects of metabolic syndrome) to demonstrate an important role for IL-6 mediated STAT3 activation in the pathogenesis of Group 2 PH¹⁹⁶. These models have the distinct advantage of likely recapitulating the mechanisms by which multiple comorbidities promote pulmonary vascular remodeling in Group 2 PH, but future translational studies are needed to determine whether the identified pathways are indeed actionable for the prevention or reversal of Group 2 PH.

4. Animal Models of RV Failure

Although the pulmonary vasculature is the locus of the initial insult, survival of PAH patients is intimately related to RV function¹⁹⁷. Generally considered as a feature to combat, RV hypertrophy is classically divided into 2 patterns on the basis of morphological and functional changes: adaptive remodeling (compensated state) characterized by more concentric remodeling, minimal dilatation and fibrosis, normal ejection fraction, and CO, and maladaptive remodeling (decompensated state) associated with increased infiltration of inflammatory cells, accumulation of extracellular matrix due to fibroblast activation and proliferation, cardiomyocyte loss and capillary rarefaction resulting in altered ejection fraction, low CO, and diminish exercise capacity^{198, 199}. While there is broad consensus on the fact that hypertrophy can be either beneficial or detrimental, the molecular mechanisms and pathways governing the process from adaptive to maladaptive hypertrophy remain largely unknown, and as a consequence, there are no clinically established treatments for RV failure. Along with the MCT model mentioned above, the constriction of the pulmonary artery with a clip or a ligature represents a suitable model to study RV remodeling and dysfunction in the setting of PAH. This surgical procedure, called pulmonary artery banding (PAB), is commonly used in various species, including rodents and larger animals such as sheep, to generate RV afterload^{200, 201}. In this model, different degrees of constriction can be applied leading to either long-term compensatory hypertrophy (mild constriction) or RV failure (tight constriction)^{202, 203} (Figure 1). There are a number of benefits to using this model, first of all to replicate the adaptive and maladaptive continuum of the RV stress response and thus capture the cellular and molecular events implicated in the stress coping response, or inversely, disruption of the latter. To this end, a comprehensive hemodynamic characterization of PAB or MCT-challenged animals determined by right heart catheterization (mPAP, CO, RVEDP…) followed by their subsequent categorization into compensated or decompensated state are prerequisite^{7, 204, 205}. In addition, the PAB model serves as a complementary tool to conventional PAH models in determining whether a drug that shows anti-remodeling effects on the pulmonary vasculature has a beneficial, direct damaging or no effect on the remodeled RV^{22, 206}. This is all the more important given the increasing interest in “anti-remodeling“ approaches and that, paradoxically, the main purpose of such approaches is to stop proliferation and induce death of unwanted PA cells, while preserving survival and function of cardiomyocytes is desired. In direct connection with this, it can be hypothesized that the beneficial effects of the drug on the pulmonary vasculature masks potential deleterious effects on the stressed RV.

Conclusion and concluding remarks

Given the complex multifactorial nature of PH, no single animal model can per se completely mimic the human condition. Ideally, a model of PH would be more prevalent in female but more severe in male; progressive and irreversible in nature (culminating in RV failure and death); should mimic the pathophysiology of the disease (i.e. similar genetic, epigenetic, metabolic and inflammatory profiles leading to profound vascular lesions). Finally respond to standard of care should be limited as seen in patients. Such model can probably be generated by combining the existing ones. So far, multiple animal models have been developed, each of which shed light on different processes that appear to play a role (Table 1). Animal models have proven to be a great value in deciphering the molecular physiopathology of PH as well as key information on therapeutic strategies that cannot be obtained by using only alternative methods. Rather than an alternative, ex-vivo studies, such as human precision cut lung slice, can be considered as helpful complementary approaches allowing the researcher to bridge the gap between animal models and human patients, thus reducing the number of animals used to a minimum. It must keep in mind that PH animal models provide only clues, not final answers, for human disease. It is crucial to select the most appropriate animal model based on aspects they can reproduce. For instance, if the primarily goal is to assess the effects of a molecule on pulmonary vascular remodeling, the MCT and Su/Hx models should be privileged. Similarly, the MCT and PAB models should be prioritized to study the molecular mechanisms responsible for RV failure and evaluate the cardioprotective effects of new approaches. In parallel, improvement of animal models of PH by combining different risk factors or hits is needed. In that respect, the comparison of “one” versus multi-hit models that mix genetics with environmental interactions, various mechanisms and pathways, comorbidities and other diseases that lead to PH is encouraged to better capture the multifactorial nature of PH. Recent studies showing that (i) Sirt3;Ucp2 double knockout mice developed several features of the human disease that are typically absent in other PAH mouse models⁹⁵ and (ii) phenotypically silent BMPR2 mutations predispose rats to inflammation-induced PAH⁵⁵ echoes this view. Lastly, due to inherent limitations of currently available animal models, it is recommended that, where possible, the effects of an intervention be tested in multiple animal models, thereby maximizing the chance of translational success.

Nonstandard Abbreviations and Acronyms

BPD

bronchopulmonary dysplasia

BMPR2

bone morphogenetic protein receptor 2

COPD

chronic obstructive pulmonary disease

FHR

Fawn-Hooded rat

HIV

Human immunodeficiency virus

IPF

idiopathic pulmonary fibrosis

MCT

monocrotaline

mPAP

mean pulmonary artery pressure

PA

pulmonary artery

PAB

pulmonary artery banding

PAEC

pulmonary artery endothelial cell

PAH

pulmonary arterial hypertension

PASMC

pulmonary artery smooth muscle cell

PH

pulmonary hypertension

PVOD

Pulmonary veno-occlusive disease

PVR

pulmonary vascular resistance

RV

right ventricle/ventricular

RVSP

right ventricular systolic pressure

SD

sprague-dawley

SSc

Systemic sclerosis

Su/Hx

Sugen/Hypoxia