Sex, Gender, and Sex Hormones in Pulmonary Hypertension and Right Ventricular Failure

Abstract

Pulmonary hypertension (PH) encompasses a syndrome of diseases that are characterized by elevated pulmonary artery pressure and pulmonary vascular remodeling and that frequently lead to right ventricular (RV) failure and death. Several types of PH exhibit sexually dimorphic features in disease penetrance, presentation, and progression. Most sexually dimorphic features in PH have been described in pulmonary arterial hypertension (PAH), a devastating and progressive pulmonary vasculopathy with a 3-year survival rate <60%. While patient registries show that women are more susceptible to development of PAH, female PAH patients display better RV function and increased survival compared to their male counterparts, a phenomenon referred to as the “estrogen paradox” or “estrogen puzzle” of PAH. Recent advances in the field have demonstrated that multiple sex hormones, receptors, and metabolites play a role in the estrogen puzzle and that the effects of hormone signaling may be time and compartment specific. While the underlying physiological mechanisms are complex, unraveling the estrogen puzzle may reveal novel therapeutic strategies to treat and reverse the effects of PAH/PH. In this article, we (i) review PH classification and pathophysiology; (ii) discuss sex/gender differences observed in patients and animal models; (iii) review sex hormone synthesis and metabolism; (iv) review in detail the scientific literature of sex hormone signaling in PAH/PH, particularly estrogen-, testosterone-, progesterone-, and dehydroepiandrosterone (DHEA)-mediated effects in the pulmonary vasculature and RV; (v) discuss hormone-independent variables contributing to sexually dimorphic disease presentation; and (vi) identify knowledge gaps and pathways forward.

Introduction

Several cardiopulmonary diseases are characterized by sex and gender differences and have been the focus of comprehensive research efforts (145). However, few of these diseases have seen as much progress in understanding the biological basis of these differences as pulmonary hypertension (PH), a pulmonary vasculopathy resulting in elevated pulmonary artery (PA) pressures (376). PH is not a single disease but rather a syndrome that encompasses a heterogeneous group of acute and chronic diseases of different origins and etiologies that share the common feature of mean pulmonary artery pressure (mPAP) higher than 20 to 25 mmHg (377). The current PH classification guidelines differentiate five major groups that differ in their etiologies and phenotypes (Figure 1) (377). If left untreated, PH of any etiology can lead to right ventricular (RV) failure and death. The majority of sex and gender differences in PH have been described in pulmonary arterial hypertension (PAH; Group 1 PH), a disease characterized by progressive pulmonary vascular remodeling resulting in severely increased pulmonary vascular resistance (PVR) and a high likelihood of RV failure and death (326, 429, 430). Sexually dimorphic features have also been described in other types of PH but are typically not as prevalent or pronounced as in PAH.

Figure 1. Current classification of pulmonary hypertension (PH) and subtypes with evidence for sexually dimorphic features.

PH classification from 6^th World Symposium (Nice, 2018) according to Simonneau et al. (351). In addition to the data presented here, one study in a large cohort of veterans with all types of PH (predominantly Group 2 and 3 PH; n = 15,464 patients) demonstrated that women with PH exhibit higher pulmonary vascular resistance and pulmonary artery pulse pressure, yet lower RAP as well as 18% greater survival compared to men with PH. *These analyses predominantly included patients with idiopathic PAH and also patients with heritable PAH and drug- and toxin-associated PAH (no subgroup analyses performed). ^#Attenuated hypoxia-induced PH in women not consistently found across studies. BMPR2, gene encoding bone morphogenic protein receptor 2; CYP1B1, gene encoding cytochrome P450 1B1; CYP19A1, gene encoding aromatase; ESR1, gene encoding estrogen receptor α; HFpEF, heart failure with preserved ejection fraction; HIV, human immunodeficiency virus; HT, hormone therapy; LVEF, left ventricular ejection fraction; PCH, pulmonary capillary hemangiomatosis; PVOD, pulmonary veno-occlusive disease; PVR, pulmonary vascular resistance; RV, right ventricle; SNP, single-nucleotide polymorphism; SSc, systemic sclerosis.

Sexual dimorphism in PAH exists in disease prevalence, severity of hemodynamic alterations, RV adaptation, treatment responses, and, importantly, survival. In particular, women are more prone to develop PAH, but exhibit a more favorable hemodynamic profile, better RV function, a better response to treatment with endothelin receptor antagonists (ERAs), and better survival. Men, on the other hand, are less prone to develop PAH and are more likely to respond to treatment with phosphodiesterase type 5 inhibitors but are more likely to die from this disease. More favorable hemodynamic profiles and higher survival rates have also been described in women with non-PAH types of PH; however, data from these cohorts is less abundant than for PAH.

This article comprehensively reviews the rapidly expanding biological and epidemiological knowledge regarding sex and gender differences in PAH and PH. We review the role of sex hormones, their metabolites and their receptors, and the role of nonhormonal factors in the pulmonary vasculature and RV in health and disease. We discuss cell culture systems, animal studies, and studies in humans. Knowledge gaps will be identified, and pathways forward will be proposed.

We use the term “PH” when discussing PH in general and the term “PAH” when specifically referring to this disease. By convention, elevated PA pressure in animal models is referred to as “PH,” while “PAH” is reserved for the human condition. According to the definitions published by the Institute of Medicine and embraced by the APS Journals (279, 468), we use the term “sex” when biologic concepts are described, but use the term “gender” when cultural or behavioral influences may play a role (e.g., in human studies).

A list of commonly used abbreviations is provided in the Abbreviations and Acronyms Section.

PH Classification and Definitions

PH has traditionally been defined as an mPAP > 25 mmHg with a classification scheme divided into five groups based on the predominant underlying pathology and clinical phenotype (Figure 1). These groups encompass Group 1 (PAH), Group 2 (PH due to left heart disease), Group 3 (PH due to lung disease and/or hypoxia), Group (4 PH due to PA obstructions such as chronic thromboembolic pulmonary hypertension [CTEPH]), and Group 5 (PH due to unclear or multifactorial mechanisms) (377). A detailed discussion of all five PH groups is beyond the scope of this article; the most up-to-date classification from the Proceedings of the 6^th World Symposium on Pulmonary Hypertension in Nice is presented by Simonneau et al. (377).

Most recently, the hemodynamic definition of PH was changed to an mPAP cut-off of >20 mmHg (377), two standard deviations above the upper limit of normal for the pulmonary circulation, although this remains controversial. Regardless of the threshold used for mPAP, the various PH phenotypes can also be classified based on the localization of the pathology in the pulmonary vascular compartment. Precapillary PH is characterized by (i) an elevated mPAP, (ii) a pulmonary arterial wedge pressure (PAWP) ≤ 15 mmHg, and (iii) a PVR ≥ 3 Wood units. Precapillary PH occurs in Groups 1, 2, 3, and in some cases of Group 5 PH (377). Postcapillary PH, on the other hand, is characterized by an elevation of both mPAP (to >20 or 25 mmHg) and PAWP (≥ 15 mmHg). This may occur in isolation (without an elevation in PVR to >3 Wood units) or combined with precapillary PH such that mPAP, PAWP, and PVR are increased. Both isolated and combined postcapillary PH occur in Group 2 PH and in some forms of Group 5 PH.

Pathophysiology of PAH and PH

The pathophysiology of PAH and PH has been reviewed in detail elsewhere (166, 326). An overview is provided in Figure 2. Briefly, PH occurs as a consequence of lesions in the arterial, capillary, or venous compartment of the pulmonary vasculature. In certain subtypes and associated conditions (e.g., pulmonary veno-occlusive disease (PVOD), drug- and toxin- and connective tissue disease-associated PAH, and CTEPH), a spectrum of lesions may occur that span more than one compartment. PH can also occur in a fairly normal pulmonary vasculature as a consequence of venous congestion due to left heart disease or increased pulmonary blood flow in the setting of hypervolemia or hyperdynamic states.

Figure 2. Pathophysiology of PAH.

(A) Arterial cross section illustrating PAH pathology in the pulmonary arteries. Proliferation of endothelial cells (ECs), smooth muscle cells (SMCs), and fibroblasts leads to vascular remodeling with eventual occlusion of diseased vessels. Neoangiogenesis driven by apoptosis-resistant proliferative ECs, SMCs, and other resident PA and recruited cells promotes formation of plexiform vascular lesions, which are the hallmark of PAH. Plexiform lesions may be seen within pulmonary vessels as well as extending into the adventitial tissue (not shown). Infiltration of PH vascular lesions by immune cells and bone marrow-derived cells drives a pro-inflammatory and pro-proliferative state in the tissue. (B) Transverse section of the heart. High pulmonary vascular resistance in PAH produces increased afterload on the RV, resulting in adaptation and RV failure in PAH. RV hypertrophy may be adaptive and compensatory to overcome PVR and maintain cardiac output (not shown). On the other hand, RV hypertrophy may be maladaptive, marked by vessel rarefaction, metabolic dysfunction, inflammation, cell death, fibrosis, and increased RV dilatation. Maladaptive RV remodeling is associated with RV ischemia and decreased RV ejection fraction and cardiac output, resulting in RV failure.

A tremendous amount of progress has been made in our understanding of PAH pathobiology (166, 326). Over the past several decades, discoveries that endothelial dysfunction and vascular remodeling occur in PAH from dysregulation of nitric oxide (NO), endothelin-1 (ET-1), and prostacyclin pathways informed drug development and led to the approval of numerous pulmonary vasodilators (169). We now understand PAH to be an even more complex and systemic disease. Numerous cells in and around the vascular compartment, such as endothelial cells (ECs), smooth muscle cells (SMCs), adventitial fibroblasts, and inflammatory cells contribute to disease pathogenesis and are influenced by the immune and hematopoietic systems as well as abnormalities in cellular energetics and metabolism. The current paradigm is that PAH occurs as a consequence of a single or repetitive pulmonary vascular injury mediated by increased pulmonary blood flow, shear stress, inflammatory processes, excessive vasoconstrictor stimuli, and/or EC damage. While not all individuals with a pulmonary vascular injury develop PAH, disease development is more likely to occur in the setting of genetic predisposition (e.g., mutations in bone morphogenetic protein receptor 2 [BMPR2]), previous vascular injury (e.g., premature birth and environmental exposures), and/or coexposures (e.g., hormonal and metabolic abnormalities, and substance abuse in HIV infection) (166, 291). Epigenetic changes may further modify the disease course; these multiple potential “hits” to the pulmonary vasculature are being targeted for potential intervention. Deep phenotyping efforts are also underway to understand common molecular mechanisms that may underpin and influence the severity of various forms of PH across all five PH groups and provide targets for precision-based medicine (152).

Right Ventricular (RV) Adaptation in PH

RV failure is an important cause of morbidity and mortality in PAH as well as Group 2 and 3 PH from highly prevalent chronic heart and lung diseases. An estimated 70 million individuals in the United States may have right heart dysfunction (171, 239, 274, 282, 292, 334, 434), yet there are no well-established biologic or clinical determinants of RV structure and function and no approved treatments for right heart failure. Unlike the left ventricle (LV), the thin-walled, compliant RV has difficulty accommodating increases in resistance such that even incremental increases or fluctuations in afterload over time may lead to RV sequelae (448, 449). There is, however, great variability in the clinical trajectory of patients, and they often present at later stages of disease, when RV dysfunction has already occurred. While RV failure is the proximate cause of death in PAH, mechanisms of RV adaptation (and maladaptation) have garnered much interest but remain understudied (212).

Current knowledge of the pathophysiology of RV failure has been discussed in detail elsewhere (346, 447, 449) and is beyond the scope of this article. A brief overview is presented here and in Figure 2. Initially, as RV afterload increases during PH development, the RV employs compensatory mechanisms that include structural changes, neurohormonal activation, and increased contractility (346, 449). At the cellular level, these changes are accompanied by increased angiogenesis, changes in mitochondrial function and substrate utilization, increased production of reactive oxygen species, changes in myosin isoform expression, and changes in sarcomere organization and structure (346). It is thought that these changes allow for a state of adaptive (or compensated) RV hypertrophy, characterized by a cardiac output that is still sufficient to meet the metabolic demands of the body (448, 449). However, with ongoing increases in RV afterload, the RV’s compensatory mechanisms will eventually be exhausted and cause a transition to a maladaptive (or decompensated) form of RV hypertrophy (448,449). Consequently, RV failure with decreased cardiac output and decreased oxygen delivery occurs. At a cellular and molecular level, maladaptive RV hypertrophy purportedly is characterized by ischemia, impaired or insufficient angiogenesis, inflammation, oxidative stress, metabolic dysfunction, and impaired calcium handling, all associated with myocardial fibrosis and cell death (34, 447, 449). The individual contribution of each of these processes may vary from patient to patient and exhibit marked temporal and spatial variations (212).

A brief overview of PAH/PH epidemiology and subtypes, with a focus on those subgroups with a known gender bias, as well as a review of gender differences in RV adaptation across all forms of pulmonary vascular disease follows.

Gender Bias in PAH Epidemiology

The earliest modern description of idiopathic PAH by Dresdale et al. (80) in 1951 included three young women. The first prospective multicenter registry from the National Institutes of Health (NIH), which included patients with idiopathic, heritable PAH and PAH associated with anorexigen use, reported a mean age of 36 ± 15 years and a ratio of women:men of 1.7:1. Before the advent of targeted PAH therapy, 1-, 3-, and 5-year survival for this cohort was 68%, 48%, and 34%, respectively, with an estimated median survival of 2.8 years (69). This early description of then “primary pulmonary hypertension,” a rare disease affecting young women of child-bearing age, has evolved in recent years.

The prevalence of Group 1 PAH is estimated between 15 cases/million (5.9 cases/million for idiopathic PAH) with an incidence of 1.1 to 3.7 cases/million/year (96, 168, 231, 318). Multiple registries have captured survival in both the pre- and posttreatment era (28, 69, 96, 98, 167, 177, 185, 231, 307, 412, 487). Short-term survival has improved over time and is approximately 90% at 1 year and 75% at 3 years. Longer term survival remains poor, however, with registries survival rates between 21% and 75% at 5 years.

While a gender bias similar to that reported in the NIH registry has been noted in recent registries throughout the world (60, 96, 98, 158, 168, 177, 224, 231, 268, 318, 412, 450), others have described a more marked predominance among women. In modern registries including various Group 1 etiologies, as many as 70% of participants are women, and the average age of all participants is older (5^th decade of life) (168, 185, 411, 412). A large European registry, which enrolled patients from 2007 to 2011 (The Comparative Prospective Registry of Newly Initiated Therapies for Pulmonary Hypertension [COMPERA]), demonstrated a ratio of 1.8 women:1 men that was most pronounced among younger patients (158). The largest US-based registry, the Registry to Evaluate Early And Long-term pulmonary arterial hypertension disease management (REVEAL), began enrolling patients in 2006 and reported among idiopathic PAH patients an average age of 53 ± 15 years, 80% of whom were women (18, 262). Whether these observations signal a true change in disease biology or a significant survival bias among women because of a predominantly prevalent (vs incident) study population is not known. In China, where targeted PAH therapies have only recently become available, the earliest registry of incident (i.e., untreated) patients included 71% women (a ratio of 2.4 women:1 men), similar to the US NIH registry (177, 338). In a follow-up Chinese study from a treated/prevalent cohort, 76% of participants were women, and there were 3.1 times as many women enrolled as men (487). Table 1 summarizes the gender biases of modern population-based registries for Group 1 PH.

Table 1.

Gender ratios in major PH registries across the world.

Registry	References	Enrollment period	P(A)H types included	Patient age (years; mean unless indicated otherwise)	Number of patients included	Ratio, Women:men
Latvian	(380)	2007–2016	IPAH, APAH, drug-induced PAH, CTEPH	PAH: 65 (median) CTEPH: 67 (median)	PAH: 130 CTEPH: 44	PAH: 2.7:1 CTEPH: 1.6:1
Swedish (SPHAR)	(329)	2008–2014	IPAH, HPAH, APAH, CTEPH	PAH: 67 (median) CTEPH: 70 (median)	PAH: 457 CTEPH: 183	PAH: 1.8:1 CTEPH: 1:1
European (COMPERA)	(307)	2007–2013	IPAH, HPAH, APAH	68 (median)	1283	1.8:1
International CTEPH Registry	(319)	2007–2009	CTEPH	63 (median)	679	1:1
UK/Ireland	(231)	2001–2009	IPAH, HPAH, anorexigen-induced PAH	50	493	1.4:1
Spanish	(96)	2007–2008	IPAH, APAH, TOS-PAH, PVOD, CTEPH	PAH: 45 CTEPH: 61	PAH: 866 CTEPH: 162	PAH: 2.4:1 CTEPH: 1.5:1
REVEAL	(28)	2006–2007	IPAH, HPAH, APAH, anorexigen-induced PAH	53	2525	4.1:1 IPAH 3.8:1 APAH
Chinese	(177)	1999–2004	IPAH, HPAH	36	72	2.4:1
French	(170)	2002–2003	IPAH, HPAH, APAH, anorexigen-induced PAH	50	674	1.9:1
Scottish	(318)	1986–2001	IPAH, APAH	51	374	2.3:1
NIH	(338)	1981–1985	“Primary PH” (IPAH and HPAH)	36	187	1.7:1

APAH, associated pulmonary arterial hypertension; CTEPH, chronic thromboembolic pulmonary hypertension; HPAH, hereditary pulmonary arterial hypertension; IPAH, idiopathic pulmonary arterial hypertension; NIH, National Institutes of Health; PVOD, pulmonary veno-occlusive disease; TOS-PAH, toxic oil syndrome-induced pulmonary arterial hypertension.

While female gender has long been established as the major clinical risk factor for PAH, in 2010, both the French (167) and US REVEAL (28) registries published similar findings and found male gender to double the risk of death in PAH. This risk was shown to be independent of established measures of disease such as the six-minute walk distance (6MWD) and cardiac index (CI) (28, 167, 170). Follow-up data from REVEAL continued to demonstrate significant differences in outcome based on gender irrespective of whether the PAH diagnosis was incident or prevalent, such that 5-year survival estimates for newly diagnosed (incident) men were 53% ± 4% versus 63% ± 2% for women and 57% ± 2% versus 68% ± 1% for previously diagnosed (prevalent) men versus women (98). Interestingly, this survival benefit occurs despite more profound vascular remodeling and more plexiform lesions in PAH women (390), a constellation suggestive of better adaptation to vascular remodeling in women (reviewed in more detail below). Table 2 provides a summary of gender differences in survival, hemodynamic alterations, and treatment responses in PAH.

Table 2.

Major findings from studies investigating gender differences, sex hormones, or sex hormone pathway components in human PAH.

Study finding	References
General findings
↑ Susceptibility to PAH development in women	See Table 1
↑ Pulmonary vascular remodeling in female PAH patients	(390)
↑ Survival in female PAH patients	(28, 167, 170, 175, 307, 365)
↓ Hemodynamic alterations in female PAH patients (↓ RAP, ↓ mPAP, ↑ CI)	(365, 445)
↑ RVEF in female PAH patients	(175, 192,405)
↑ Improvement in RVEF in females after initiation of PAH treatment responsible for survival advantage in females	(175)
↑ Response to treatment with endothelin receptor antagonists or intravenous prostacyclins in female PAH patients	(108, 118)
↑ Response to treatment with tadalafil in male PAH patients	(258)
Menopause is risk factor for scleroderma-associated PAH (SSc-PAH); HT attenuates SSc-PAH	(30, 363)
Findings specific to sex hormone signaling
Altered estrogen metabolism ↑ penetrance of hereditary PAH; ↓ urine 2-hydroxyestradiol/16α-hydroxyestrone ratios in patients with hereditary PAH vs. unaffected BMPR2 mutation carriers	(12)
↑ ESR1 mRNA expression in PAH patients	(331, 391)
SNPs in ESR1 and CYP19A1 associated with ↑ risk for development of portopulmonary hypertension; SNPs in CYP19A1 associated with ↑ E2 plasma levels	(340)
E2 plasma levels correlate negatively with 6MWD and functional class in male and female PAH patients	(19, 441)
↑ E2 and E2/testosterone ratio and ↓ testosterone and γ progesterone associated with ↑ risk of PAH in males; ↑ E2 associated with ↑ mortality in male patients	(471)
Genetic variations in E2 metabolism and androgen signaling associated with RV morphology in a gender-specific manner (MESA-RV, healthy cohort)	(443)
Aromatase inhibition safe and ↑ 6MWD in postmenopausal and male PAH patients (small proof-of-concept study)	(193)
Lower DHEA-S levels in men and postmenopausal women with PAH compared to matched controls; lower DHEA-S associated with more severe PAH, RV dysfunction	(19, 441)
Lower DHEA-S and metabolites associated with poor survival	(19)
Open-label DHEA treatment in small (n = 8) study of COPD-PH improved 6MWD and hemodynamics	(90)

The upper section lists general findings; the lower section lists findings related to the study of specific sex hormones. BMPR2, bone morphogenetic protein receptor 2 gene; CI, cardiac index; CYP19A1, aromatase gene; ESR1, estrogen receptor α gene; HT, hormone therapy; mPAP, mean pulmonary artery pressure; RAP, right atrial pressure; RVEF, right ventricular ejection fraction; SNP, single-nucleotide polymorphism; 6MWD, six-minute walking distance.

Age may be an important modifier of the relationship between gender and outcomes in PAH (28, 96, 167, 171, 307, 445), which suggests that temporal changes in the hormonal milieu may impact disease risk and severity throughout the lifespan. Among older patients, gender-based differences in PAH prevalence appear to be diminished (158, 231, 445). In a large study (n = 1211) of patient-level pooled data from 11 clinical trials in PAH, women with idiopathic PAH and connective tissue disease (CTD)-associated PAH had more favorable hemodynamic indices (lower right atrial pressure [RAP], lower PVR, and higher CI) as compared to men with idiopathic and CTD-associated PAH (445). Younger men had higher mPAP than younger women, but this difference was attenuated after age 45 years. In both men and women with idiopathic PAH, hemodynamic burden, including mPAP, tended to decrease to similar levels with age, such that a gender difference in mPAP was only seen in patients <45 years old. Similar observations have been made in the COMPERA registry, which demonstrated a strong gender bias toward women among younger patients with PAH that dissipated after age 65 (158), and in the REVEAL registry, men had higher RAP and mPAP at diagnosis (as well as worse survival especially in those older than 60 years of age) (28, 365). These observations have not been consistent across all registries, and further work is needed to refine the sex-age interaction in pulmonary vascular disease (231,487).

Race/ethnicity may also modify the relationship between sex and PAH. In the NIH registry, the gender ratio was even more skewed toward women (4.3:1) among African Americans (69). This observation was also made in the United Kingdom and Ireland, where 85% of nonwhite patients were women (as compared to 70% women in white patients) and in the United States, where the ratio of women:men was 5.4:1 in African Americans from the REVEAL registry (18, 231).

Gender Bias in PAH Subtypes

Mutations in BMPR2, a gene encoding a member of the transforming growth factor (TGF)-β family, are present in 70% to 80% of families with PAH and roughly 25% to 30% of patients with idiopathic PAH (73, 291, 388). These mutations are transmitted in an autosomal dominant fashion with incomplete penetrance. Female mutation carriers are more than twice as likely to be affected with PAH as carrier men (31); in a large cohort of individuals with BMPR2 mutations, roughly 70% of the population were women (97). Cross talk between BMPR2 pathways and estrogen signaling has been a major area of study as reviewed below. Rarer mutations in activin-like receptor kinase-1 (ALK-1), endoglin (ENG), mothers against decapentaplegic homologue (SMADs), caveolin 1 (CAV1), and the potassium channel subfamily K member 3 (KCNK3) genes have also been identified (15, 245, 302, 372). Recent studies identified that rare variants in ATP13A3, AQP1, and SOX17 and common genetic variations at loci in an enhancer near SOX17 and in HLA-DPA1/DPB1 are associated with PAH (129, 336). Biallelic mutations in EIF2AK4 have been linked to pulmonary capillary hemangiomatosis (PCH) and PVOD, very rare forms of PAH (31). Mutations in TBX4 are associated with childhood-onset PAH (198). The penetrance of PAH in these rare mutations is not known to vary by gender.

A number of systemic diseases are associated with the development of pulmonary vasculopathy, although the mechanisms by which PAH develops in these varied conditions are poorly understood. In some of these subgroups, female gender has been described as a risk factor for the development of PAH, including CTD-associated PAH and portopulmonary hypertension. Approximately 12% of systemic sclerosis (SSc) patients develop PAH, and it is a major cause of death (137, 295, 392). Additional CTDs such as systemic lupus erythematosus, mixed CTD, and rheumatoid arthritis are also associated with PAH. While the true prevalence of PAH in these conditions is unknown, PAH occurs less commonly than in SSc and is associated with better outcomes than when associated with SSc (60, 64). After idiopathic PAH, CTD-PAH patients are the second most represented subgroup in registry studies (18, 96, 168, 318). CTD-PAH patients tend to be older, have less hemodynamic impairment, and are more likely to have mixed phenotype PH from concurrent interstitial lung disease, pulmonary venous involvement, and left heart disease (18, 487). As in idiopathic PAH, female sex is arisk factor in CTD-PAH (18,487). CTD itself occurs more commonly in women than in men, and when associated with PAH patients are 3.8 to 10 times more likely to be women (18, 60,168,185,487). Women with SSc are eight times more likely than men with SSc to be affected by PAH (59). While scleroderma-associated PAH is more common in women, it is interesting to note that PAH in scleroderma patients frequently does not occur until after menopause (30, 363). In patients with systemic lupus erythematosus, women are 17 times more likely to be affected with PAH than men (60). Chung et al. demonstrated an almost fourfold increase in the risk of death (hazard ratio 3.9, 95% CI 1.1–13.9, p = 0.03) among men as compared to women with PAH associated with SSc (58), however.

The presence of portal hypertension without other clinical risk factors or associated conditions in a patient with PAH is designated as portopulmonary hypertension. This condition occurs in roughly 3% to 6% of patients with cirrhosis referred for liver transplantation (110). The French Registry reported that 40% of patients with portopulmonary hypertension were women (168). While the degree of cirrhosis does not influence the risk of portopulmonary hypertension, female gender and autoimmune hepatitis are independent risk factors for the development of PAH in these patients (63, 138, 195). Female gender as a risk factor for portopulmonary hypertension has been confirmed in several registry studies (168, 195, 205). Although the pathobiology of portopulmonary hypertension has not been entirely elucidated, abnormalities in sex hormone signaling have been implicated in disease development, as discussed below (303, 340). Survival is generally poorer in portopulmonary hypertension than in idiopathic PAH (205).

Additional PAH subtypes and their associated conditions do not appear to have a strong gender bias. These include drug- and toxin-induced PAH, human immunodeficiency virus (HIV) infection-associated PAH, congenital heart disease (CHD)-associated PAH, and schistosomiasis-induced PAH. This may be because hormonal factors do not play a pathobiologic role in all forms of PAH or because these conditions are less well studied. Certain drugs and toxins have been implicated in the development of PAH, some as “definite” causes of PAH and others as “possible” (377). The most classic example of drug-induced PAH are the anorexigens such as fenfluramine (1, 40, 79, 255, 339, 357, 450). Although tyrosine kinase inhibitors have been studied to treat PAH, the use of dasatinib for chronic myelogenous leukemia has been associated with the development of PAH (285, 333, 353). Methamphetamine was recently reclassified as a “definite” cause of PAH; in one study, patients with methamphetamine-associated PAH were less likely to be women, had more severe disease, and worse outcomes as compared to patients with idiopathic PAH (486). Treatment with interferon has also been identified as a possible risk factor for PAH (44, 74, 111, 178, 352, 358, 376).

The prevalence of pulmonary vascular disease in HIV-infected patients is approximately 0.5% (308, 379, 389), which has not decreased in frequency despite the advent of antiretroviral therapy (379). Disease characteristics are similar to idiopathic PAH patients, although gender does not appear to be a risk factor for the development of PAH in HIV (168, 321). Pulmonary vascular disease has been reported in 4% to 34% of adults with CHD (87, 93, 242). Modern PAH registries have reported that CHD-associated PAH makes up 11% to 24% of Group 1 PAH patients; this does not appear to vary by gender (18, 26, 96, 168, 231). In recent years, consensus guidelines have included a more detailed subclassification of pulmonary vascular disease associated with CHD, which distinguishes between precapillary/Group 1 PAH and Group 2 PH due to congenital/acquired cardiovascular conditions leading to postcapillary PH (376, 377). Chronic schistosomiasis is likely the most common cause of PAH worldwide given the widespread prevalence of schistosomiasis mansoni infection but is incompletely understood. While direct pulmonary vascular exposure to egg antigens does not appear to cause schistosomiasis-PAH, many of the mechanistic pathways implicated in idiopathic PAH (e.g., TGF-β and inflammatory cytokines) have been implicated in the development of Schistosoma-related pulmonary vascular injury (130-132). A gender bias in Schistosoma-associated PAH has not been described.

Gender Bias in RV Function in PAH

Women have better RV systolic function in both health and PH, including Group 1 (PAH), Group 2 (left heart disease), and Group 3 (chronic lung disease/hypoxia) (192, 196, 269, 323, 444). The Multi-Ethnic Study of Atherosclerosis (MESA)-RV is the largest population-based, cardiovascular disease-free cohort with available RV indices measured via cardiac magnetic resonance imaging (MRI), the gold standard for RV assessment. Female gender was associated with higher right ventricular ejection fraction (RVEF), lower RV mass, and smaller RV volumes after adjustment for multiple comorbid factors and body size in MESA-RV (194, 196). These gender-based differences in RVEF and volumes were replicated in The Framingham Heart Study Offspring cohort (106). Both the MESA and Framingham studies showed age to be an important modifier of the relationships between gender and measures of RV morphology (106, 194, 196).

Until recently, the RV had not been robustly studied in PAH. This is important, since changes in RV structure and function with PAH therapies are more strongly tied to survival in PAH than changes in PVR (434-436). Kawut et al. demonstrated that male gender is associated with lower RVEF measured by radionuclide angiography in a single-center cohort of subjects with PAH (192). This finding has since been corroborated by several other investigators (175, 405). Among a Dutch cohort of 101 patients with idiopathic PAH, heritable PAH, or anorexigen associated PAH, men and women had similar reductions in PVR, but RVEF declined in men and improved in women over time with PAH therapies (175). A significant proportion (39%) of the transplant-free survival difference seen between men and women was explained by treatment-related improvements in RVEF. This suggests that the survival bias conferred by female gender in PAH may be explained at least in part by gender or sex hormone-mediated effects on the RV. These observations have led to increased interest in the study of sexual dimorphism in RV function and failure in PAH. Studies of gender differences in RV function in PAH are listed in Table 2.

Gender Bias in Treatment Responses in PAH

In addition to the observational studies reviewed above that demonstrate gender-based differences in PAH prevalence, RV function, and survival, differential responses to PAH-specific treatments have also been described. In a patient-level pooled analysis from six randomized placebo-controlled trials of ERAs submitted to the US Food and Drug Administration, Gabler et al. noted that women exhibited a better response in 6MWD to treatment with ERAs (118). Similarly, women were more likely to respond to treatment with prostacyclin analogues (108). On the other hand, a post-hoc analysis of subjects enrolled in a trial of the phosphodiesterase type 5 inhibitor tadalafil demonstrated that men were more likely to improve their 6MWD and quality of life after starting tadalafil treatment as compared to women (258).

Gender Bias in Non-PAH PH

Studies in non-PAH PH are sparse and in general less robust than those in PAH. No clear signal exists indicating that female gender is a risk factor for disease development in non-PAH PH. Heart failure with preserved ejection fraction (HFpEF) is frequently associated with PH and is more common in postmenopausal women (257, 359), but female gender is not a risk factor for HFpEF-PH per se. On the contrary, some studies exist suggesting that women exhibit less hypoxia-induced PH and less chronic mountain sickness than men; however, such effects are not consistently found across studies (39, 225, 490). Two retrospective studies identified male gender as a risk factor for the development of high-altitude pulmonary edema (HAPE), a disease characterized by exaggerated and uneven hypoxic pulmonary vasoconstriction (HPV) (165, 386). Similar to PAH, there is evidence that women with non-PAH PH demonstrate superior RV function and higher survival rates. A review of the Veterans Affairs Clinical Assessment Reporting and Tracking (CART) Program database demonstrated that in a large cohort (n = 15,464 patients) of veterans with all types of PH (but predominantly Group 2 and 3 PH), women exhibited higher PVR and PA pulse pressure, yet lower RAP (442). This constellation of findings is indicative of better RV adaptation despite higher RV afterload. Interestingly, women veterans with PH had 18% greater survival compared to men with PH. When the cohort was limited to veterans with precapillary PH, women with PH were 29% more likely to survive. Women are also less likely to develop RV dysfunction in the setting of HFpEF (269) and exhibit better RV function in the setting of chronic lung disease (323). No gender bias in prevalence has been demonstrated in CTEPH (319). In a Japanese cohort of CTEPH patients, women exhibited better cardiac output at baseline but higher residual PVR after pulmonary thrombendarterectomy (371).

Proposed Mechanisms of Gender Differences in Human PAH/PH

Taken together, the data reviewed above suggest profound gender differences in PAH and other types of PH (Tables 1 and 2). Female gender is one of the strongest risk factors for PAH development, but also a robust protective factor once the disease has been acquired. On the other hand, with the possible exception of PH from HFpEF, female gender does not appear to be a risk factor for non-PAH types of PH. In both PAH as well as non-PAH PH, female gender is associated with better RV adaptation, indicative of persistent gender-based phenotypes across various types of pulmonary vascular disease.

These findings could be due to direct effects of sex hormones on cardiopulmonary function, genetically determined factors, environmental or epigenetic influences, and/or cultural phenomena. Several lines of evidence in humans suggest that biologically relevant effects of sex hormones indeed play a role in mediating gender differences in the pulmonary vasculature and RV. First, there is a high prevalence of exposure to hormone therapy (HT) in women with PAH (404). Second, genetic alterations in estrogen-metabolizing enzymes and estrogen receptors (ERs) have been found in various forms of PAH (12, 331, 340). Third, 17β-estradiol (E2) plasma levels correlate positively with RV function in healthy postmenopausal HT users yet negatively with 6MWD and functional class in PAH patients (19, 441, 444), and lower dehydroepiandrosterone-sulfate (DHEA-S) levels correlate with worse hemodynamics, RV function, 6MWD, and functional class in PAH patients (19, 441). Of note, lower DHEA-S levels are also associated with lower survival (19, 337). Fourth, at least one study described an absence of hemodynamic differences between men and women with PAH once they are older than 45 years, suggesting that the menopausal transition (and other hormone-related life cycle changes like waning testosterone and/or DHEA) may modify disease risk (445). In addition, as mentioned above, menopause represents a risk factor for the development of PAH in scleroderma patients, while HT may attenuate the risk of PAH in these patients (30, 363). Lastly, genetic variations in 17β-estradiol metabolism and androgen signaling are associated with RV morphology in a gender-specific manner in cohort of subjects without clinical cardiovascular disease (MESA-RV) (443). The contributions of sex hormones in human PAH/PH will be reviewed in detail within this article. The roles of genetics, epigenetics, environmental exposures, and cultural factors have been much less well studied and represent a significant knowledge gap in the field.

The “Estrogen Puzzle” in PAH

The marked discrepancy between increased susceptibility to PAH among women on the one hand and better disease outcomes in women on the other hand has been described as the “estrogen paradox” and has been the topic of many editorials, reviews, and discussions at scientific meetings. A second “estrogen paradox” has been identified in the area of basic and cellular investigation. In particular, this refers to the finding that estrogens are protective against disease development in several animal models of PAH, but detrimental (disease promoting) in others. Lastly, some investigators refer to a paradox in the observation that estrogens have been uniformly shown to be cardioprotective in the RV, whereas in the pulmonary vasculature they may exert disease-promoting effects. Together, these inconsistencies have led to increased interest in the study of sexual dimorphism in PAH and RV failure. However, since the term “paradox” implies an observation or finding that is logically unacceptable or self-contradictory, we prefer the terms “estrogen puzzle” or “estrogen conundrum.” This is based on the rationale that we believe biological explanations exist for the observed sex/gender differences in PAH but have not yet been fully identified. For example, many “paradoxical” effects can be explained with dose-, timing-, or compartment-specific effects of estrogen. We would also argue that to focus solely on estrogen as the hormone of interest is too narrow a scope. While estrogens clearly have been implicated as clinically important disease modifiers in PAH, one should note that estrogen-independent factors, such as other sex hormones, sex chromosomes, genetics, and epidemiological factors, likely play significant roles as well.

Sex Differences in Animal Models of PAH

Much of the knowledge about mechanisms of gender and sex differences in PAH and PH has been obtained from the study of animal models. While several animal models have been developed that display features of the pulmonary vascular remodeling and/or right heart hypertrophy/failure that are common to PAH, recapitulating the sexually dimorphic disease presentation and progression in these models has been challenging. In contrast to human PAH, female sex is protective in many classical models of PAH, such as hypoxia- or monocrotaline (MCT)-induced PH. However, several models (many of them transgenic) have been developed over the past two decades that demonstrate a female bias with regard to disease susceptibility or severity, and animal studies have allowed for a more nuanced understanding of the effects of sex and sex hormones on disease development. In addition, progress has been made in understanding sex differences in RV function and failure. Commonly used animal models of PH and RV failure have been reviewed extensively elsewhere (126, 212, 347, 394). Here, we briefly describe each model of PAH and the impact of sex on pulmonary and RV remodeling.

A synopsis is provided in Table 3. Contributions of individual sex hormones to PH development in these models and their interactions with specific pathways will be described later in this article.

Table 3.

Overview of sex differences observed in animal models of PH.

Species	Male bias	References	Female bias	References
Mouse	Hypoxia-induced PH	(478, 479)	SERT+ (with or without hypoxia)	(461, 463)
	eNOS−/−	(278)	S100A4/Mts1+	(72)
	VIP−/−	(349)	Dexfenfluramine-induced PH	(71)
	ApoE−/− + high-fat diet	(147)	Smooth muscle-specific STAT5++/− or STAT5−/− + chronic hypoxia	(479)
	miR-214−/− + Su5416/chronic hypoxiaa	(395, 479)	Cyp2c44−/− + chronic hypoxia	(182)
			Gonadectomized “four core genotypes” mouseb	(431)
Rat	Monocrotaline-induced PH	(421,432, 483)	Athymic rnu/rnu rats + Su5416 or chronic hypoxiac	(408)
	Hypoxia-induced PH/HPV	(92, 208, 327, 335)	4,4′-Methylenedianiline (DAPM)-induced PH	(46)
	Su5416/hypoxiad	(114, 213)
	Athymic rnu/rnu rats + semaxanib	(135)
Chicken	Hypoxia-induced PH	(42, 350)	n/a
Sheep	HPV	(459, 460)	n/a
Swine	Hypoxia-induced PH	(267)	n/a

ApoE, apolipoprotein E; athymic rnu/rnu rats, T-cell-deficient athymic rats; Cyp2c44e, cytochrome P450 2c44e; eNOS, endothelial nitric oxide synthase; HPV, hypoxic pulmonary vasoconstriction; miR-214, microRNA 214; n/a, no studies available; SERT+, serotonin transporter overexpression; STAT, signal transducer and activator of transcription; Su5416; sugen (VEGF receptor 2 inhibitor); S100A4/Mts1 +, S100A4/Mts1 overexpression; VIP, vasoactive intestinal peptide.

^aMore RV hypertrophy in males; no difference in RV systolic pressure or pulmonary artery remodeling.

^bNot a typical study comparing sexes, but Y-chromosome protective.

^cSex differences abolished after repletion with CD4⁺CD25^hlgh T regulatory cells.

^dNo published sex differences in RV systolic pressure, but better RV function, less RV remodeling, and higher survival in females, as well as disparate pulmonary artery remodeling in males and females.

Chronic Hypoxia-Induced Pulmonary Hypertension (HPH)

Hypoxia has classically been used in both rodent and non-rodent models to induce pulmonary vascular remodeling. Histologically, HPH (hypoxia-induced pulmonary hyper-tension) induces media hypertrophy in the pulmonary vasculature, but plexiform or vaso-occlusive lesions are not seen (394). In addition, damage to the pulmonary vasculature is reversible after reexposure to room air, and this model induces RV hypertrophy but not failure (212, 393). While hypoxia alone does not recapitulate Group 1 PAH pathology, HPH shares certain signaling pathways and disease mechanisms with human PAH and could be considered a model of mild or early PAH (393). Some investigators, on the other hand, suggest that HPH may be a better model for Group 3 PH (345). Contrary to human data, females of many HPH model species are more resistant to HPH, with smaller increases in right ventricular systolic pressure (RVSP), RV hypertrophy, and pulmonary vascular remodeling. This effect has been shown in vivo in rats, mice, swine, and chickens (42, 92, 208, 267, 327, 335, 350), as well as in isolated ovine lungs, which display reduced HPV (459, 460). Reduced HPV has also been demonstrated in isolated pulmonary arteries (PAs) from female rats (214). While the mechanisms of contraction during the acute phase of HPV are separate from the mechanisms governing contraction and remodeling during chronic HPH, the observation that sex impacts the contractile response of isolated pulmonary vessels illustrates the dramatic sexual dimorphism of tissues implicated in PH pathogenesis. The in vivo data in rats is particularly compelling, as ovariectomized rats become vulnerable to severe HPH, and supplemental E2 treatment in ovariectomized animals rescues this effect (92, 335). HPH is driven in part by hypoxia-induced erythrocytosis (which leads to increased blood viscosity and increases in PA pressure) (299, 437), and it has been demonstrated that the female resilience in HPH is at least in part due to lower hematocrit levels (208, 296). However, direct effects on the pulmonary vasculature (less vasoconstriction and remodeling) play a role as well. Compared with rats, mice are relatively resistant to HPH regardless of sex (126, 161, 406). Nevertheless, female HPH mice also display more favorable hemodynamics, less RV hypertrophy, and less PA remodeling (478, 479). While favorable hemodynamics and decreased RV hypertrophy are akin to humans, decreased PA remodeling is not. Female HPH mice also express higher levels of angiogenic factors such as VEGF-A in the RV as compared with hypoxic males (35).

Monocrotaline-Induced PH (MCT-PH)

Administration of the toxic pyrrolizidine alkaloid MCT in rats is another classical model of PAH. Circulating MCT is converted to its bioactive form dehydromonocrotaline by the cytochrome P450 (CYP) enzyme CYP3A family in the liver (189). The exact mechanism through which dehydromonocrotaline induces PH is unknown, but it likely acts primarily through damage to pulmonary artery endothelial cells (PAECs) (342). For reasons that are not entirely clear, this model does not work in mice; this may be due to unpredictable CYP3A subtype 4 activity and/or species-specific resistance to MCT-induced vascular injury (127). Muscularization of PAs, increased PVR, RV hypertrophy, and eventual RV failure and death are seen in this model. MCT-PH is accompanied by systemic inflammation and possibly myocarditis and hepatic veno-occlusive disease and has therefore been proposed as a model of inflammation-induced PAH, such as PAH associated with CTD (127). Similar to findings in hypoxia, female sex is protective in the rat model of MCT-PH. Specifically, female sex or exogenous estrogens ameliorate the phenotype of MCT-PH compared with males, while ovariectomy (OVX) exacerbates disease progress (3, 99, 298, 421, 432, 483). Sex differences in the MCT model may result at least in part from decreased CYP3A activity in the female liver, leading to reduced levels of dehydromonocrotaline (189). Recently, MCT administration has been combined with chronic hypoxia to develop a more severe phenotype of PH, characterized by thrombotic, neointimal, and plexiform-like lesions in the pulmonary vasculature (66, 289). Only data from male rats has been published using this combined injury model. Similarly, MCT in combination with pneumonectomy causes more severe PH and vascular remodeling but has only been published in males (466).

Sugen/Hypoxia-Induced PH (SuHx-PH)

A more recent model of PAH was published in 2001, which more closely resembles the human phenotype (409). Here, administration of the vascular endothelial growth factor (VEGF) receptor 2 antagonist Su5416 (sugen) to young rats, followed by hypoxia and subsequent reexposure to normoxia produces severe PH with RV failure and mortality. In this model, Su5416 administration leads to initial PAEC apoptosis, followed by exuberant proliferation of the remaining PAECs, resulting in pronounced and progressive PA remodeling and formation of vaso-occlusive lesions (409). Two recent studies demonstrated that female sugen/hypoxia-induced pulmonary hypertension (SuHx-PH) rats exhibit better RV function than their male counterparts (both at rest and after acute exercise), including improved stroke volume index, CI, RV compliance, and reduced RV hypertrophy (114,213). These findings were accompanied by more favorable antioxidant, pro-survival, and pro-angiogenic responses as well as less fibrosis and lower pro-inflammatory cytokine expression in female RVs. Interestingly, higher cardiac indices were also noted in healthy females versus males, mirroring the better RV function noted in healthy humans (167, 175, 192, 405). In both studies, no sex differences were found in RVSP increase or pulmonary vascular remodeling (114, 213). A study by Rafikova et al. (330), on the other hand, demonstrated sexually dimorphic pathology in the pulmonary vasculature, with female SuHx-PH rats displaying increased pulmonary vascular wall thickness compared to male SuHx-PH animals. However, the pulmonary vasculature from male SuHx-PH rats displayed increased fibrosis and inflammatory markers, and female SuHx-PH rats displayed less RV hypertrophy and increased survival (330). Taken together, these studies show that SuHx-PH exhibits sexually dimorphic features in rats, with better RV adaptation and survival in females, despite potentially more pronounced vascular remodeling. Importantly, this mirrors the human PAH phenotype and suggests that SuHx-PH is a suitable model to study sex differences. The lung vascular remodeling, fibrosis, and inflammation data by Rafikova et al. suggest that sex differences in the pulmonary vasculature can be nuanced and that a detailed examination of different compartments and pathways is necessary to capture the full spectrum of sex differences in the lung vasculature in experimental PH and human PAH. SuHx-PH has also been employed in mice, albeit with a less-consistent and less-severe phenotype (126, 212, 347, 394). Induction of SuHx-PH in mice requires maintained hypoxia as well as weekly injections of Su5416. In the only published male-female comparison in SuHx-PH mice, no sex differences were noted in RVSP or RV hypertrophy (61).

Mutant and Transgenic Rodent Strains

Genetically modified rodents have produced mixed and occasionally contradictory findings in regard to sex-based differences in PH. While some models display a female bias, others show the opposite. For example, overexpression of the calcium-binding protein S100A4/Mts1 (72) or the serotonin transporter SERT (462, 464) increases female penetrance of PH and disease severity, while ovariectomy attenuates these effects. Disease outcomes are similarly more severe in female mice globally lacking the epoxygenase CYP2c44 (182) or lacking the transcription factor Stat5 in SMCs (479). Conversely, female mice lacking genes for vasoactive intestinal peptide (VIP) (349), endothelial nitric oxide synthase (eNOS) (278), or apolipoprotein E (ApoE) (the latter in combination with a high-fat diet) (147) develop a much less-severe PH phenotype compared with male littermates. Furthermore, hypoxic female miR-214 knockout mice develop similar hemodynamic and PA remodeling alterations as their male counterparts but exhibit less RV hypertrophy (395). Two studies demonstrated that certain transgenic manipulations can abolish the male bias in experimental PH, suggesting that these pathways may be involved in making males more susceptible or females less susceptible. Specifically, smooth muscle-specific deletion of Stat5 (479) or the transcriptional repressor Bcl6 (478) abrogates female protection in female rats with HPH. In the case of Stat5 deletion, female protection is reversed to female susceptibility.

As mutations in BMPR2 underlie most cases of heritable PAH, and since decreased BMPR2 activation has been noted in idiopathic PAH (246), several strains of Bmpr2 mutant mice have been developed to examine the role of this signaling pathway in disease progression (reviewed in Ref. 345). Global knockout of the Bmpr2 gene is embryonically lethal in mice, while heterozygous mutant mice spontaneously develop a mild form of PH (29). Targeted deletion of Bmpr2 in either ECs (160) or SMCs (458) is sufficient to produce PH features such as increased RVSP and pulmonary vascular remodeling. A recent report from Hautefort et al. describes a similar phenotype in Bmpr2 mutant rats (150). A gender bias has not been reported in Bmpr2 mutants, with similar penetrance of PH seen in both male and female mice. However, Bmpr2 mutant mice are more vulnerable to 16-OHE1 administration than wild-type controls, leading to increased PH penetrance (52). This phenomenon is associated with aberrant estrogen signaling within cells of the pulmonary vasculature (101). Even though there appear to be no sexually dimorphic features in Bmpr2 mutant strains, altered estrogen signaling in Bmpr2 mutants may be relevant to unraveling the estrogen puzzle in human PAH.

While these transgenic mouse models allow focused hypothesis testing regarding the role of a particular gene/protein in PH and are useful in identifying mechanisms that may explain sex differences in PH, it is important to remember that many transgenic models display only a mild hemodynamic or vascular phenotype of PH, thus limiting their clinical relevance (24). Additionally, knockout models are primarily available in mice, which generally display a less-severe PH phenotype than rats, and robust models of PH such as MCT or SuHx either do not work in mice (in the case of MCT) or result in only a mild phenotype (in the case of SuHx) (61, 126). Despite these caveats, transgenic mouse models have been successfully used to investigate the estrogen puzzle and have advanced the field. Recent advances in genetic manipulation such as CRISPR/Cas9 are increasing the number of available transgenic rat strains (105); this generates an exciting opportunity to perform gain- and loss-of-function studies in more robust PH models.

Rodent Models of Immunity and PH

Two studies have been published employing T-cell-deficient athymic rnu/rnu rats. In the first study, administration of the VEGFR 1 and 2 antagonist semaxanib induced more severe PA remodeling, more RV hypertrophy, and more profound RV systolic dysfunction in males relative to females (135). Impairment of RV-PA coupling efficiency was observed only in males, and pulmonary artery smooth muscle cells (PASMCs) switched from a contractile state to a dedifferentiated state in males only. However, a more recent study demonstrated the opposite effect (408). Here, sugen administration or chronic hypoxia led to a more severe phenotype in female rats. In particular, female rats exhibited greater pulmonary inflammation; augmented RV fibrosis; lower plasma levels of prostacyclin; decreased lung expression of cyclooxygenase, prostacyclin synthase, programmed death ligand-1 (PDL-1), and heme oxygenase-1; and reduced PDL-1 levels in the RV. Treg immune reconstitution protected against PH development in both sexes and abrogated sex differences in Treg-deficient animals. While the reason for the contradictory findings between the two studies in athymic rats is unclear and must be resolved, the implication that immunity may underlie sex differences in PH is intriguing and parallels data showing sex- and sex hormone-mediated differences in immune function and dysfunction (227) and studies that identify immune dysregulation as a contributing factor to PAH in humans (328). More studies are required to investigate the interplay among sex, sex hormones, and immunity in PAH/PH. Ultimately, such studies would be expected to shed further light on the underlying mechanisms of sex differences in experimental PH and human PAH.

A critical role of immune cells in PH was recently demonstrated by Hu et al. in a study where humanized mice were engrafted with human hematopoietic CD34⁺ progenitor cells (resulting in circulating human leukocytes) and subsequently exposed to chronic hypoxia. In contrast to nonhumanized mice, humanized mice displayed significantly increased RVSP and PA muscularization (163), suggesting that species-specific immune responses may underlie the reduced acuity of murine PH models at least in part.

Drug-Induced PH

While gender does not play a clear role in human drug- and toxin-induced PAH (perhaps because of psychosocial or behavior influences), two models of drug-induced PAH predominantly affect female rodents. Specifically, administration of the anorexigen dexfenfluramine or the industrial compound 4,4′-diaminodiphenylmethane (DAPM) results in PH in female animals only (46, 71). The mechanism of both drugs appears to involve increased serotonergic signaling and altered estrogen metabolism to favor pro-proliferative metabolites. Interestingly, dexfenfluramine-induced PH involves upregulation of the estrogen-metabolizing enzyme CYP1B1, and CYP1B1 knockout animals are not susceptible to dexfenfluramine-induced PH (71).

Pulmonary Artery Banding (PAB)

Pulmonary artery banding (PAB) is a model of RV hypertrophy with or without RV failure that is independent of changes in the pulmonary vasculature (212). No investigations focusing on sex differences in this model have been published to date.

Summary of Studies of Sex as a Disease Modifier in Animal Models

Taken together, animal models have identified sex as an important disease modifier in experimental PH. As in humans, animal models of PAH demonstrate an effect of sex on disease penetrance and severity. As in humans, the results from animal models can be complex and occasionally contradictory. In the classical models of PAH (HPH and MCT-PH), female sex is protective. Female sex is also protective in several transgenic mouse models. On the other hand, a female bias with regard to disease susceptibility or severity has been noted in transgenic mice, Treg-deficient rats and models of drug exposure. SuHx-PH rats exhibit complex sex-specific features in the pulmonary vasculature but better RV function and survival in females. It is worth noting, however, that the pulmonary hypertensive or RV failure phenotype in some of these models is modest, thus limiting their relevance to human PAH. In particular, results from studies in mice may be of limited translational value because of the mild PH phenotype produced. New animal models, including transgenic, immunological, pharmacological, and surgical models that employ various “hits” to the pulmonary vasculature rather than one single insult (thus mimicking the pathogenesis of human PAH) are likely to provide novel information on the role of sex in PH (212). For example, SuHx-PH rats recapitulate many of the hallmarks of human PAH (126, 212, 347, 394). However, it is critical to look at animal models as tools to dissect specific components or mechanisms of disease development rather than use one single model to represent and study the entire spectrum of gender differences in human PAH (38, 324). In addition, important modifiers such as animal age, phase of the estrous cycle, and external influences (e.g., dietary phytoestrogens (149) and gender of the animal handler (387)) need to be considered. Only the study of several animal models in conjunction and the consideration of these modifiers will generate sufficient and relevant new data to solve the estrogen puzzle.

What are Possible Explanations for Discrepancies in Sex/Gender Differences and Sex Hormone Effects between Rodent Models and Human PAH?

While the animal models described here recapitulate many features of PAH, no animal model is a perfect analog for human physiology. Animal models of estrogen signaling are particularly difficult as the human menstrual cycle is a radically different physiological phenomenon than the much shorter estrous cycles of laboratory animal rodents (356). Hormone levels, cycle duration, and the aging endocrine profile are all species specific and may play a vital role in PAH/PH penetrance and progression. These differences may underlie seemingly contradictory findings, such as an association of high E2 levels with PH in humans (19, 441, 471), but reduced E2 levels in rodent models of PH (114, 482). It is possible that PAH/PH alters hormone metabolism and secretion differently between humans and model species and that the profile of estrogen metabolite production and distribution is different between species. The nature and physiological impact of increased E2 levels in human PAH patients is also unclear. Without a baseline endocrine profile that predates disease onset, it is impossible to tell if increased E2 levels are a result of PAH onset or if elevated E2 levels instead directly contribute to PAH development. Also, because hormone secretion is controlled via a negative feedback mechanism at the level of the hypothalamus and pituitary, it is possible that impaired ER signaling could lead to a surplus of circulating hormone in the circulation. In this scenario, elevated E2 may be serving as a biomarker of impaired downstream signaling, rather than as a disease mediator per se. More detailed investigations into hormone metabolism and species-specific endocrine profiles are warranted to develop new models of PAH that more closely align with human physiology.

What are the Underlying Mechanisms Mediating Sex and Gender Differences in Rodent Models and Human PAH?

The data reviewed above clearly demonstrate that profound gender and sex differences exist in human PAH and PH as well as in experimental PH. Substantial research over the past two decades has identified sex hormone-dependent and -independent mechanisms as mediators of these differences. The latter include genetic, epigenetic, environmental, and behavioral factors. In the remainder of this article, we review sex hormone-dependent and -independent factors involved in sex and gender differences in etiology, physiology, hemodynamics, treatment responses, and outcomes in PAH and PH. Given the critical role of RV function in PAH, we highlight sex and gender differences in RV function in dedicated s. Since the majority of the published literature involves data on estrogen signaling and metabolism, a large part of this article will focus on this area, but we also discuss the currently available knowledge about testosterone, progesterone, and DHEA(-S) as well as genetic, environmental, and behavioral modifiers. The conglomerate of the data reviewed demonstrates that there has been significant progress in the field and that we are getting closer to solving the “gender puzzle” in PAH and PH.

Overview of Sex Hormone Synthesis and Metabolism

Detailed reviews of steroid hormone production and metabolism are available in the literature (317, 417) and are beyond the scope of this article. However, since basic concepts of steroidogenesis and sex hormone signaling are key to understanding the sexually dimorphic pathogenesis and presentation of PAH/PH, clinically important key concepts will be reviewed here. The steroidogenic pathway is illustrated in Figure 3.

Figure 3. Sex hormone synthesis and estrogen metabolism.

Steroidogenic enzymes represented here are present in the lung, heart, and/or vascular tissue (148, 197, 314, 317, 480). Abundance of boxed enzymes is altered in PH/PAH (12, 326, 340, 457, 462-464). Compounds in red have been targeted in clinical trials of PAH therapies (193). 2-Hydroxyestradiol and 2-methoxyestradiol exert ER-independent antiproliferative, anti-inflammatory effects that appear to mitigate vascular remodeling in animal models of PAH (22, 82, 421). Conversely, the estrogen metabolites 4-hydroxyestradiol, 4-methoxyestradiol, and 16α-hydroxyestrone signal via estrogen receptors and promote a mitogenic, inflammatory, and antiapoptotic phenotype that exacerbates PH/PAH (54, 463). *Multiple CYP enzymes are capable of catalyzing estrogen hydroxylation; CYP1A1 and CYP1B1 appear to be the most relevant isoforms in PAH pathology. Factors including diet (240, 275), hypoxia (107), inflammation (11), genetics (12, 146), and drug exposure (71) may alter estrogen metabolism. Effects of enzymes and metabolites depicted here may be cell-, tissue-, and/or organ specific.

All steroid hormones are derived from cholesterol, a 27-carbon sterol consisting of four hydrocarbon rings attached to a hydrocarbon tail (top left in Figure 3). Sex hormone synthesis is characterized by the progressive cleavage of carbon atoms surrounding the hydrocarbon ring structure. Sex hormones can therefore be broken into three classes based on the number of carbon atoms present in their structure: 21-C, 19-C, and 18-C hormones (green, blue, and pink boxes, respectively, in Figure 3). Hormone synthesis and modification within each of these classes will be briefly discussed before examining the role of specific sex hormones in PAH/PH.

In the first step of steroidogenesis (Figure 3), CYP11A (side-chain cleavage enzyme) cleaves the hydrocarbon tail from cholesterol to yield pregnenolone, a 21-carbon compound that is the common precursor of all steroid hormones. CYP11A is located on the inner mitochondrial membrane, and transport of cholesterol across the outer mitochondrial membrane by steroidogenic acute regulatory (StAR) protein is the rate-limiting step in steroid hormone synthesis (396). The enzyme 3β-hydroxysteroid dehydrogenase (3β-HSD) converts pregnenolone to progesterone, a biologically active steroid critical to establish and maintain pregnancy. Progesterone is primarily secreted by the ovary and placenta, but also serves as a metabolic intermediary in the adrenal gland. The role of progesterone in PAH/PH is a fairly unexplored avenue of research that will be discussed later.

CYP17A catalyzes conversion of either pregnenolone or progesterone to the 19-carbon hormones dehydroepiandrosterone (DHEA) or androstenedione, respectively (blue box in Figure 3). Circulating DHEA and its sulfate ester DHEA-S are the most abundant steroid hormones in the human body (252). DHEA can be converted to the weak androgens androstenedione or androstenediol, both of which are immediate precursors of the potent androgen testosterone. The enzyme 5a-reductase converts testosterone to dihydrotestosterone (DHT), the most potent biological androgen.

18-Carbon steroid hormones (estrogens) are derived from androstenedione or testosterone. CYP19 (aromatase) catalyzes the conversion of androstenedione or testosterone into estrone (E1) or estradiol (E2), respectively (pink box in Figure 3). Estrone may be converted to estradiol through the action of 17β-hydroxysteroid dehydrogenase (17β-HSD). Estrogens may be further processed into bioactive metabolites via hydroxylation by multiple CYP enzymes including CYP1A1 and CYP1B1. The resulting hydroxyestrogens may then undergo methylation by the enzyme catechol-O-methyl transferase (COMT) to form methoxyestrogens (220, 427, 476). One hydroxylated metabolite, 16α-hydroxyestradiol, is commonly referred to as estriol (E3), an estrogenic compound that plays an important role during pregnancy (4, 310). The various estrogen metabolites exert radically different effects on cell signaling and function that may impact many proliferative and inflammatory processes, such as cancer, chronic inflammatory diseases, and PAH (86). The role of these metabolites in PAH is described in the following section. Sex steroids typically signal through specific receptors (e.g., progesterone-, androgen-, and estrogen-receptors), even though receptor-independent actions of certain metabolites have been described as well (described in detail later).

Circulating hormone levels vary widely among individuals and tend to decline with age (Table 4). Women of reproductive age also exhibit cyclical surges and withdrawal of estrogens and progesterone according to the phase of each menstrual cycle. Pregnancy and menopause are marked by their own distinct endocrine profiles. While serum hormone levels are constantly in flux, a high proportion of circulating sex hormones are bound to carrier proteins such as albumin or sex hormone-binding globulin (SHBG), which sequester hormones from receptors in surrounding tissues. This results in a relatively small percentage of circulating hormones being biologically active (143). Further complicating the issue, many nongonadal tissues contain the necessary enzymes (e.g., 3β-HSD, 17β-HSD, and aromatase) to synthesize both androgens and estrogens from DHEA. Peripheral production of androgens and/or estrogens is possible in many organs and occurs in adipocytes, ECs, SMCs, cardiac fibroblasts, and cardiomyocytes (148, 197, 314, 317, 465, 480). Peripheral aromatization to form E1 and E2 is particularly important as this is the main source of estrogens in men and post-menopausal women (183, 378). Local production of steroids by a target cell has been termed “intracrinology” (207, 360). This is biologically and clinically relevant as it may lead to significant local steroid signaling not reflected in measurements of circulating hormone levels.

Table 4.

Normal values for sex hormone levels in humans across the life span.

			Women of reproductive age
Hormone	Men <60	Men >60	Follicular phase	Preovulatory phase	Luteal phase	Pregnancy	Menopause
Progesterone	0.25–0.90 ng/ml	0.25–0.90 ng/ml	0.3–1.2 ng/ml	0.7–2.5 ng/ml	1–30 ng/ml	9–300 ng/ml	<0.2–1.1 ng/ml
	1–3 nM	1−3 nM	1–4 nM	2–9 nM	3–100 nM	25–1000 nM	0.6–3.5 nM
DHEA	1.5–10.4 ng/ml	0.3–3.5 ng/ml	3–10 ng/ml	3–10 ng/ml	3–10 ng/ml	1.5–17 ng/ml	0.4–4.6 ng/ml
	5–36 nM	1–12 nM	10–35 nM	10–35 nM	10–35 nM	5–60 nM	1.5–16 nM
Testosterone	2–15 ng/ml	1–10 ng/ml	0.2–0.8 ng/ml	0.2–0.8 ng/ml	0.2–0.8 ng/ml	1–1.4 ng/ml	0.2–0.8 ng/ml
	6–50 nM	3.5–33 nM	0.7–2.5 nM	0.7–2.5 nM	0.7–2.5 nM	3.5–5n M	0.7–2.5 nM
Estradiol	0.015–0.05 ng/ml	0.002–0.04 ng/ml	0.02–0.1 ng/ml	0.15–0.4 ng/ml	0.06–0.20 ng/ml	1−40 ng/ml	0.01–0.03 ng/ml
	0.05–0.2 nM	0.007–0.15 nM	0.07–0.4 nM	0.5–1.5 nM	0.2–0.8 nM	4–150 nM	0.04–0.12 nM

Values extracted from (45, 250, 304, 309, 426).

Gauging the effect of sex and sex hormones on physiology is complex due to conversion between hormone classes, changing levels of bioactivity, steroid metabolism into other bioactive compounds, and hormone synthesis in nonclassical steroidogenic tissues. In target cells, changes in hormone-receptor isotype abundance or localization (91, 294, 325, 355), variations in activity of enzymes responsible for sex hormone synthesis (47, 202, 253), interactions with genomic and nongenomic cofactors (265), additive or opposing effects of multiple hormones or receptors (142, 191), and epigenetic phenomena (219) can drastically alter the downstream effect of hormone-receptor binding. At a broader level, the expression and activity of sex steroid receptors, as well as hormone-metabolizing enzymes, may be affected by sex, age, diet, environmental exposures, fluctuations in endogenous sex hormone levels (e.g., menstrual cycle and menopause), and various disease states (reviewed in Ref. 297). It is therefore not surprising that many sex hormone-mediated effects are compartment-, time-, and concentration dependent. Carefully devised experimental systems are required to determine the impact of steroid hormones, particularly in a complicated syndrome like PH. The development of new animal models, new biological tools, and the dedication of many researchers has allowed recent advances in unraveling the role of steroid hormones in PAH/PH and moved the field closer to unraveling the “estrogen puzzle” in PAH.

Estrogen in PH

Estrogen Signaling and Metabolism

E1, E2, and E3 are the three main estrogens in the human body. Of these, E2 is the most potent estrogen and the primary mediator of estrogen signaling. An overview of E2 levels in humans across the life span is provided in Table 4. E1, E2, and E3, as well as many of their metabolites, signal through interaction with three ERs. Two of these receptors, ERα and ERβ, are members of the nuclear receptor superfamily (reviewed in Refs. 151, 271, 272, 297, 375). ERα and ERβ are encoded by independent genes located on separate chromosomes. In humans, ESR1 (estrogen receptor α gene) encodes ERα, a protein composed of 595 amino acid residues with a molecular weight of 66.2 kDa. ESR2 (estrogen receptor β gene) encodes ERβ, a slightly smaller protein composed of 530 amino acids and a molecular weight of 59.2 kDa (305). The two receptors share high homology for the DNA-binding domain, as well as the ligand-binding domain (95% and 60%, respectively) (206). Both receptors also contain two activation function domains (AF-1/AF-2) which alter transcription through interaction with nuclear coactivators/repressors (201, 454). ERα and ERβ share less than 30% homology of the AF domains, which contributes to the unique transcriptional profile of each receptor despite near-perfect homology of the DNA-binding domain (488).

ERα and ERβ are widely expressed throughout the body (reproductive, cardiovascular, respiratory, central nervous, endocrine, immune, and skeletal systems). In the pulmonary vasculature, ERα and ERβ are expressed in PAECs (141, 439, 440), PASMCs (141, 187), and fibroblasts (78), all of which are involved in vascular remodeling during PH. ERs are also expressed in other lung cells including epithelial cells (141, 176), alveolar cells (315), and alveolar macrophages (401, 439). In the heart, ERα and ERβ are expressed in ECs, cardiomyocytes, and fibroblasts (reviewed in Refs. 271, 272). ERα and ERβ are also expressed in progenitor cells (140) and immune cells (204, 397), where they modify developmental, regenerative, and immune signaling processes in health and disease. While there is overlap in functions of ERα and ERβ, significant differences exist between the two ERs in tissue expression and function.

Classical estrogen signaling occurs via these two receptors in what is termed the genomic pathway. Here, estrogen diffuses through the cell membrane and binds to cytoplasmic ERα or ERβ. This estrogen-ER complex subsequently dimerizes with another estrogen-ER complex (either as a homo- or heterodimer) and translocates to the nucleus. The estrogen-ER dimer then binds to an estrogen responsive element (ERE) in the DNA. In this context, estrogen acts as a classic transcription factor to alter gene expression. Coactivators and corepressors of gene transcription interact with the estrogen-ER complex and contribute to regulating gene expression. In a variation of this pathway, referred to as tethered estrogen signaling, the estrogen-ER complex indirectly regulates gene expression by binding to and modulating the activity of other transcription factors. In a third pathway, nonestrogen ligands such as epidermal growth factor phosphorylate and activate cytoplasmic ER in the absence of ER binding (67). Lastly, estrogens can bind to ERα and ERβ anchored to the cell membrane. Activation of these membrane-bound receptors leads to rapid, nongenomic cellular effects. In this signaling pathway, estrogen binds to a membrane-bound receptor, leading to activation of a second messenger such as MAPK (367). This pathway does not require altered gene transcription and can induce rapid and diverse cellular effects such as ion channel activation, or activation of eNOS or prostacyclin synthase (374). Nongenomic signaling occurs within seconds to minutes and is particularly important in the cardiovascular and respiratory systems (16, 49, 53, 157, 367, 370, 402, 481). While nongenomic estrogen signaling initiates rapid dynamic cellular responses, this pathway may also ultimately produce changes in gene transcription through activation and phosphorylation of downstream transcription factors and activators (33, 151, 297). A third ER, GPR30 (G-protein-coupled receptor 30), is a G-protein-coupled receptor that has been shown to bind estrogen and signal exclusively through the nongenomic pathway.

Estrogen signaling can be altered by changes to either the ligand or receptor. Each endogenous estrogen (E1, E2, and E3) has a unique ER affinity, and the downstream effects of estrogen signaling also depend on the type of ER (ERα, ERβ, and GPR30), their interactions with each other, and their tissue-specific expression patterns. Both circulating estrogen levels and ER expression patterns change based on sex, age, disease state, and fluctuations during the menstrual cycle (reviewed in Ref. 297). Changes in ER ubiquitination can increase proteosomal degradation of ERs, while posttranslational modifications including methylation, acetylation, or S-nitrosylation may drastically impact ER signaling. In the genomic pathway, nuclear coactivators and corepressors may alter transcription by interacting with the ER in the nucleus (264). One final modulator of ER signaling is the existence of several ER splice variants. These variants demonstrate a range of activity during in vitro studies from constitutive activation to dominant-negative regulation of ER signaling (156). A relative increase in the expression of ER splice variants (as a percentage of total ER isoforms) is associated with human disease pathology, particularly tumorigenesis (156,410). The role of ER splice variants has not been investigated in PH.

E1 or E2 may be further metabolized to form new bioactive compounds, usually by hydroxylation at carbon 2, 4, or 16. The 2-hydroxylation pathway transiently produces 2-hydroxyestradiol (2-OHE2) or 2-hydroxyestrone (2-OHE1) before conversion to methoxyestrogens by the enzyme COMT (220, 427, 476). 2-Methoxyestradiol (2-ME2) is a potent metabolite with no ER affinity, which displays antimitogenic and proapoptotic properties in both tumor (215, 322, 362) and vascular SMCs (82, 428, 484). All enzymes necessary for conversion of E2 to 2-ME2 (CYP1A1/2, CYP1B1, and COMT) are present in cardiovascular tissue, and 2-ME2 synthesis occurs in vascular ECs and SMCs (81, 82, 220, 417, 485). In contrast to the antimitogenic, nonestrogenic metabolites resulting from the 2-hydoxylation pathway, the 16-hydroxylation pathway produces 16α-hydroxyestradiol (E3) or 16α-hydroxyestrone (16α-OHE1). While E3 is a weak estrogen, 16α-OHE1 exerts comparable estrogenic effects to E2 (103) and promotes inflammatory, proliferative, and angiogenic cellular processes (417). 16α-OHE1 has lower affinity for SHBG than E2 and may covalently bind to ERs causing hyperestrogenic signaling (103, 403, 417). Several CYP450 isoforms are capable of catalyzing 16α-hydroxylation (17), and the enzyme CYP1B1 has been the primary focus of PAH research. CYP1B1 is expressed in cardiovascular tissue (314) and is upregulated in PASMCs in several animal models of PAH (71, 179). Because of its pro-mitogenic properties, 16α-OHE1 is currently explored as a therapeutic target in clinical trials for various cancers (e.g., NCT02525159). The role of CYP1B1 and 16α-OHE1 metabolism in PAH is currently under investigation and will be discussed in detail later in this article. A minor pathway of estrogen metabolism is the 4-hydroxylation pathway, leading to the formation of 4-hydroxyestradiol (4-OHE2), which exerts estrogenic and carcinogenic effects (54, 228, 417). It should be noted that the biological effects and properties of estrogen metabolites have primarily been explored in reproductive, endocrine, and malignant tissues and that several of the reported effects may be tissue-, time-, and/or context dependent and have not been fully explored in the cardiopulmonary system. The ratio of various hormones and metabolites may also be more important for net biological effects than the absolute levels of one single hormone (145).

Estrogen in the Pulmonary Vasculature

Cell Culture and Animal Studies

In the systemic vasculature, estrogen signaling promotes healthy vessel function and is protective in the face of disease or vascular injury (48, 136, 313,425,470). In particular, ERα facilitates EC recovery after vascular injury, blocks monocyte adhesion to ECs, and inhibits vasoconstrictor responses (32, 36, 244, 313, 381, 467, 475). Furthermore, loss-of-function mutations in ESR1 have been linked to endothelial dysfunction, coronary artery disease, myocardial infarction, and stroke (361, 368, 369, 398, 399). Many studies support a similar role of estrogen and its receptors in the pulmonary vasculature. Both ERα and ERβ activation in cultured PAECs increases eNOS activity (49, 157, 216, 247, 370), while ERβ-mediated signaling has been shown to increase prostacyclin synthesis (247). Increased activation and synthesis of these potent vasodilators suggests that estrogen signaling in PAECs promotes vasodilation. Indeed, endogenous (214) or exogenous (210) estrogen exposure attenuates both phenylephrine-induced vasoconstriction and HPV in isolated rat PA rings. These effects were later shown to be mediated by specific ER isoforms as the ERa-agonist propylpyrazole triol (PPT) attenuates phenylephrine-induced vasoconstriction, while the ERβ agonist DPN (diarylpropionitrile) ameliorates HPV (209). Inhibition of eNOS eliminates the vasodilatory effect of both ERα and ERβ, indicating that NO plays a central role in estrogen’s effect in the pulmonary vasculature (209). Several other groups have demonstrated that estrogens, and in particular E2, attenuate HPV (116, 210, 211, 214,473) (see Table 5 for details).

Table 5.

Major animal studies of E2 or estrogen metabolites in PH.

Model	Species	Major findings	References
Isolated lungs or isolated PAs	Rat, sheep	Female sex, high estrogen states (pregnancy, proestrus), exogenous E2 or selective ERα or ERβ agonist ↓ HPV and/or drug-induced PA vasocontraction	(116, 208-210, 214, 335, 473)
Hypoxia-induced PH (HPH)	Rat, chicken, swine, mouse	Females protected; OVX ↑ PH, E2 replacement in OVX ↓ PH	(42, 92, 267, 327, 335)
		E2 administration ↓ HPH in male rats	(208, 474)
		E2 ↓ ET-1, ERK1/2, Akt, Skp2	(92, 208, 474)
		E2 ↑ p27Ki1, LC3B	(208, 474)
		ERα and ERβ ↓ pro-proliferative signaling	(208)
		ER-mediated anti-proliferative E2 effects on PAECs	(208)
		↓ proliferation in hypoxic PASMCs from proestrus rats	(463)
		CYP1B1 ↑ in male and female mice; inhibition protective; knockout ↓ PH and PA remodeling in male mice only	(473)
		16α-OHE1 ↑ in HPH; treatment of HPH mice with 16α-OHE1 ↑ PH	(463)
		Hypoxia ↑ ERβ in rat lungs and cultured PAEC from rats and humans; ERβ ↓ HIF-activation and proliferative processes in PAECs and HPH lungs	(115)
		E2 regulates proliferative and inflammatory gene expression via ER during hypoxia; E2 ↓ gremlin expression and ↑ BMPR2 signaling in hypoxic lungs	(112)
		Aromatase inhibition ↓ PH in hypoxic female mice	(248)
Monocrotaline-induced PH (MCT-PH)	Rat	Females protected; OVX ↑ PH, 2ME2 replacement in OVX ↓ PH E2 metabolites (2OHE2, 2ME2, 2EE) protective E2 pro-angiogenic and anti-inflammatory	(421, 432, 483) (418, 420, 421, 423) (432)
		ERβ-mediated protection	(432)
		MCT-PH “estrogen-deficient” state (↓ lung aromatase, lung ERa, plasma E2, ↑ CYP1B1)	(483)
		E2 ↓ ET-1, ↑ NO, ↑ PGI2	(483)
		Activation of nongenomic ER (GPR30) in male or OVX female MCT rats: ↓ RVH, ↓ RVSP, ↑ exercise endurance	(5, 6)
		E2 treatment ↓ RVSP, ↓ RVH, ↓ pulmonary vascular remodeling after MCT injection of aged ApoE~−/− mice	(433)
		Phytoestrogen genistein ↓ MCT-PH by ↓ miR206 and ↑ pulmonary angiogenesis	(260, 366)
Sugen/hypoxia-induced PH	Rat, mouse	Only mild hemodynamic alterations in female rats CYP1B1 ↑ in male and female mice; CYP1B1 inhibition ↓ PH	(348) (463)
		E2 protective in OVX SuHx rats ↓ RVSP, ↓ PA muscularization, ↓ RVH, ↑ CI, ↑ VO2max ERα agonist replicates E2 effects	(114)
		E2 administration in males: ↓ RVH, ↓ apoptotic signaling, ↑ apelin	(114)
		E2 ↑ RV adaptation after acute exercise in SuHx-PH rats	(213)
		E2 treatment in SuHx-PH OVX mice: ↓ RV afterload, ↓ PA muscularization, ↑ PA compliance, ↑ RVEF, ↑ CO, maintains pulmonary hemodynamics	(232, 234, 235)
		E2 treatment preserves RV mitochondria number and function in OVX rats	(233)
		SuHx-PH ↑ CYP1A1, ↑ aromatase. Inhibition of aryl hydrocarbon receptor (AhR) reversed this effect	(70)
		Aromatase inhibition ↓ PH in female rats	(248)
Serotonin transporter overexpression (SERT+)	Mouse	Female mice develop ↑ PA pressure at normoxia and ↑ PH during hypoxia exposure; OVX protective; E2 detrimental	(462)
		E2 ↑ proliferation, Tph-1, 5-HT1B receptor, and SERT expression in human PASMCs	(462)
		CEBPβ, FOS, CYP1B1 ↑ in female hypoxic SERT+ mice; E2 ↑ these factors in human PASMCs	(464)
		SERT+ mice overexpress CYP1B1; CYP1B1 inhibition prevents spontaneous PH phenotype	(179)
S100A4/Mts1 overexpression	Mouse	Female MTS1+ mice more susceptible to PH development than males: ↑ RVSP, ↑ pulmonary vascular remodeling in females	(72)
		E2 treatment ↑ Mts1 and ↑ proliferation in hPASMCs in a RAGE-dependent manner	(72)
Dexfenfluramine (Dfen)-induced PH	Mouse	Only females develop PH; OVX protective CYP1B1 necessary for PAH development in Dfen-treated mice	(71) (71)
		Dfen and E2 treatments ↑ CYP1B1 and Tph1 expression in cultured PAH-PASMCs	(71)
BMPR2 mutation- induced PH	Mouse	16α-OHE1 ↑ disease penetrance and ↑ RV dysfunction 16α-OHE1 ↓ BMPR2 signaling in control mice but not in BMPR2 mutants	(101) (101)
		16α-OHE1 ↓ cytokine expression but ↑ alterations in genes related to platelet function, angiogenesis, Wnt pathway, and energy metabolism	(101)
		Lack of protective effect of 2-ME2	(101)
		Altered intracellular localization of ERα in BMPR2 mutant pulmonary microvascular endothelial cells (associated with insensitivity to activation by E2)	(101)
		Estrogen inhibition ↓ PH	(51)
		16α-OHE1 ↑ PH via upregulation of microRNA-29 (miR-29)	(52)

Studies organized by model system. 16α-OHE1, 16-alpha hydroxyestrone; 2EE, 2-ethoxyestradiol; 2ME2, 2-methoxyestradiol, Akt, RAC-alpha serine/threonine-protein kinase; ApoE, apolipoprotein E; Bmpr2, bone morphogenetic protein receptor 2; CEBPβ, CCAAT enhancer binding protein beta; CI, cardiac index; CO, cardiac output; CYP1B1, cytochrome P450 1 subfamily B member 1; ERK1, extracellular signal-regulated kinase 1; ET1, endothelin 1; FOS, Fos proto-oncogene, AP-1 transcription factor subunit; eNOS, endothelial nitric oxide synthase; HIF, hypoxia-inducible factor; HPH, hypoxia-induced pulmonary hypertension; HPV, hypoxic pulmonary vasoconstriction; LC3B, autophagy-related ubiquitin-like modifier LC3 B; MCT, monocrotaline; NO, nitric oxide; OVX, ovariectomy; P27^Kip1, cyclin-dependent kinase inhibitor 1B; PGI2, prostacyclin; RAGE, receptor for advanced glycation end products; RVEF, right ventricular ejection fraction, RVH, right ventricular hypertrophy, SERT⁺, serotonin transporter over-expression; SKP2, S-phase kinase-associated protein 2; SuHx, sugen-hypoxia; Tph1, tryptophan hydroxylase 1; VO2_max, maximal oxygen uptake.

The Estrogen Puzzle in Animal Models of PAH

Studies in classic (HPH and MCT) animal models of PH have shown that estrogen ameliorates the progression of PH in the pulmonary vasculature, a finding that is not surprising given the protective role of female sex in these animal models. The SuHx-PH rodent models have produced seemingly contradictory findings as both estrogen administration and estrogen antagonism have been shown to ameliorate PH outcomes. Still other research models have identified estrogen as a potent disease mediator in the pulmonary vasculature. In the following sections, we review what is known about estrogen’s effect in the lung during PH progression, followed by a summary of estrogen’s effects on the RV. The results of studies focusing on bioactive estrogen metabolites will be discussed separately. An overview of the major studies of E2 signaling in experimental PH is provided in Table 5.

Animal Studies Demonstrating Protective Estrogen Effects in the Pulmonary Vasculature

Protective effects of endogenous and exogenous estrogen exposure in female and male animals during chronic hypoxia exposure have been demonstrated by multiple research groups (92, 115, 208, 335, 473, 474). Specifically, E2 administration has been shown to oppose hemodynamic alterations and pulmonary vascular remodeling in HPH both by attenuating hypoxia-induced upregulation of pro-angiogenic/pro-proliferative factors such as ET-1 (92), erythropoietin (296), and ERK1/2 (208) and promoting antiproliferative factors including the cell cycle inhibitor p27^Kip1 (208, 474) and the autophagy marker LC3-B (208).

Several studies have focused on identifying the role of ER signaling and activation in the pulmonary vasculature during HPH. Lahm et al. found that E2 attenuates HPH in an ER-dependent manner, and that E2 exerts antiproliferative effects on hypoxic, but not normoxic, PAECs (208). In a study by Frump et al., a microarray analysis of HPH rat lungs treated with E2 or E2 plus the ER-antagonist ICI182,780 revealed that E2 regulates several genes that mediate proliferative and inflammatory processes during hypoxia and that these processes are dependent on ER (112). This method also revealed new ER targets in HPH including the bone morphogenic protein antagonist gremlin 1, which was upregulated in hypoxia but reduced by E2 treatment (112). Other studies sought to identify the relative importance of each ER subtype in protecting the pulmonary vasculature during HPH. These studies revealed that both ERα and ERβ play a role in E2-mediated protection during HPH as blockade of either receptor eliminates E2’s inhibitory effects on PA muscularization and ERK1/2 activation in hypoxic PAECs (208). However, multiple lines of evidence suggest that ERβ-mediated signaling may be more vital for E2-mediated protection in HPH. For example, chronic hypoxia upregulates ERβ but not ERα in both rat and human PAECs (115, 208). This occurs in a hypoxia-inducible factor 1α (HIF-1α)-dependent manner. ERβ activation, in turn, induces expression of the HIF inhibitor prolyl hydroxylase 2 (PHD2). Esr2 knockout mice display increased hypoxia-induced PA muscularization compared with wild-type or Esr1 knockout mice when treated with E2 during chronic hypoxia (115). ERβ and PHD2 therefore constitute a negative feedback loop that limits hypoxia-induced HIF-1α signaling and pulmonary vascular remodeling (115).

E2 supplementation is protective against pulmonary remodeling in MCT-PH. Farhat and colleagues demonstrated that E2 supplementation in male rats reduces PA remodeling 4 weeks after MCT administration (99). A study by Yuan et al. in ovariectomized female rats demonstrated that E2 administration reduced pulmonary vascular muscularization and reduced total pulmonary resistance after MCT injection (483). Protective effects of E2 in the pulmonary vasculature were associated with suppressed macrophage infiltration in the pulmonary vasculature, as well as increased lung NO and prostacyclin levels, and reduced ET-1 expression. The authors also demonstrated that MCT administration reduced plasma E2 levels and reduced aromatase expression in the lung tissue. In contrast, the E2-metabolizing enzymes CYP1A1 and CYP1B1 were upregulated in lung tissue. The authors suggested that MCT administration elicits an estrogen-deficient state. However, studies in PAH patients suggest a more complicated picture: while decreased E2 levels indeed have been reported in premenopausal women with PAH (482), another study demonstrated increased E2 levels in postmenopausal PAH patients (19), suggesting that E2 levels in PAH may be age dependent.

Two studies investigated the role of ERs in MCT-PH. A study by Umar et al. demonstrated that E2 administration attenuates MCT-induced PH and prevents disease progression to RV failure and death in male rats (432). In that study, E2 administration reduced PA muscularization, reduced lung inflammation and fibrosis, and induced pulmonary neoan-giogenesis. Most impressively, these results were shown in a rescue model, in which the rats received E2 supplementation after PH was established. E2’s protective effects were mediated by ERβ signaling, as coadministration of an ERβ antagonist removed E2-mediated protection, while treatment with the ERβ agonist DPN recapitulated E2’s effects (432). Another study, however, suggested that the protective effects of E2 in this model are not mediated solely by ERβ. Here, chronic treatment with the GPR30 agonist G1 after MCT exposure was also able to ameliorate the PH phenotype (5). Specifically, G1 treatment was associated with decreased MCT-induced pulmonary vascular remodeling, decreased pulmonary fibrosis, and higher levels of eNOS protein in the pulmonary vasculature compared with vehicle treated rats.

Several recent studies interrogated E2 effects on vascular remodeling and disease progression in the SuHx-PH model. SuHx-PH in rats is associated with reduced circulating E2 levels in female rats (114), similar to what has been described in MCT-PH (482). Ovariectomy-induced E2 depletion increased RVSP, whereas pulmonary vascular remodeling was not affected. E2 supplementation in ovariectomized female SuHx-PH rats, on the other hand, reduced RVSP and PA muscularization compared with ovariectomized SuHx-PH females and even intact SuHx-PH females. This suggests that E2 repletion in ovariectomized SuHx-PH females creates “super responders”, a phenomenon whose underlying mechanisms require further study. In the same study, male SuHx-PH rats were implanted with subcutaneous E2 pellets to elevate circulating E2 levels to those of female rats, but E2, while attenuating right ventricular hypertrophy (RVH) and rescuing CI to control levels, did not attenuate RVSP increases or prevent PA vascular remodeling (114). In a separate study, the same group employed the SuHx-PH rat model to examine the interaction between PH and exercise tolerance (213). In this experiment, ovariectomy worsened increases in RVSP (but not PA remodeling), and E2 supplementation in ovariectomized rats decreased PA pressures and reduced the abundance of fully muscularized vessels in the pulmonary vasculature. More critically, E2 supplementation was associated with a significant reduction in postexercise total pulmonary resistance (a surrogate for PVR (203)), indicating superior performance of the E2-supplemented pulmonary vasculature after strenuous exercise. Liu et al. examined the effect of E2 supplementation on pulmonary hemodynamics in ovariectomized females in the SuHx-PH mouse model and found that E2 decreased PA elastance (a marker of RV afterload) and increased PA global compliance and transpulmonary vascular efficiency (defined as the ratio of cardiac output to total hydraulic power over the cardiac cycle) (232, 234, 235). Structurally, the authors demonstrated that E2 supplementation rescued SuHx-induced increases in PA wall thickness and collagen content. One curiosity is that Liu et al. did not find medial hypertrophy in the distal pulmonary arterioles as a result of SuHx-PH, likely because mice show a less robust response to this treatment than rats (446). Taken together, the results from the SuHx-PH studies suggest that E2, while being able to attenuate PA pressure increases and remodeling, has somewhat less-consistent effects on the PA than in the traditional PAH models such as HPH and MCT-PH (with the exception of E2 repletion in ovariectomized females, which dramatically rescues the PH phenotype). This is in stark contrast to E2’s profound RV effects in this model, which will be discussed further below. One important caveat is that sugen administration has been shown to upregulate both aromatase and CYP1A1 in rat lung tissue, indicating that pulmonary estrogen metabolism may be changed in this model of PAH (70). How these changes correlate with estrogen signaling in human PAH is currently under investigation.

Animal Studies Identifying Estrogen as a Disease Mediator in the Pulmonary Vasculature

Not all studies have demonstrated a protective effect of estrogen in the pulmonary vasculature. In fact, estrogen has been shown to promote PH development in several transgenic mouse models of PH—SERT⁺ mice (462), BMPR2^−/− mice (51), and S100A4/Mts1 overexpressing mice (72). For example, female mice overexpressing the serotonin transporter (SERT⁺) develop severe PH when exposed to hypoxia, an effect not seen after ovariectomy. Additionally, exogenous E2 administration restores PH susceptibility after ovariectomy in SERT⁺ females and induces proliferation in cultured human PASMCs in a manner dependent on de novo serotonin synthesis and activation of the serotonin receptor 5HTB (1B) (462). A latter study by the same research group demonstrated that the PH phenotype of SERT⁺ mice could be ameliorated by treatment with the ERα inhibitor MPP, which eliminated the proliferative pulmonary vascular phenotype while increasing expression of BMPR2 in the lungs of SERT⁺ mice (469).

Heritable PAH in humans is associated with mutations in the TGFβ superfamily receptor BMPR2 (62), and BMPR2 knockout mice spontaneously develop a mild form of PH (160, 458). Recently, Chen et al. demonstrated that the PH phenotype and the occurrence of muscularized pulmonary arterioles were reduced by estrogen inhibition (with fulvestrant, anastrozole, or tamoxifen) in female BMPR2 mutants (51). Interestingly, when BMPR mutants were crossed with Esr1 or Esr2 knockout mice, loss of Esr1 reduced the rate of total pulmonary vessel occlusion while knockout of Esr2 completely eliminated vessel occlusion. This implies that ERβ signaling mediates the pathologic effects of estrogen in this model. The authors also linked estrogen signaling and Esr2 to the occurrence of metabolic defects such as oxidized lipid formation and insulin resistance as well as decreased abundance of the metabolic modulators peroxisome proliferator-activated receptor-γ and CD36.

In mice overexpressing the calcium-binding protein S100A4/Mts1, E2 administration further upregulated the expression of S100A4/Mts1 in PASMCs (72). This effect was associated with increased activation of S100A4’s receptor RAGE (receptor for advanced glycosylation end products) and a proliferative PASMC phenotype, leading to occlusive lesions in the pulmonary vasculature. There is convincing evidence that RAGE overexpression by PASMC contributes to PAH in humans, and RAGE inhibitors are effective in treating both MCT-PH and SuHx-PH in rats (270). Effects of endogenous and exogenous estrogens on mediating RAGE-induced SMC proliferation remain under investigation.

In another study of detrimental pulmonary estrogen signaling, Mair et al. showed that aromatase inhibition attenuated disease progression in both a mouse HPH model and a rat SuHx-PH model, but only in female animals (248). In both disease models, the aromatase inhibitor anastrozole decreased pulmonary vascular remodeling in a dose-dependent manner in females. Anastrozole also rescued PH-mediated decreases in BMPR2 protein levels in the lungs of female animals. Interestingly, this study showed that SuHx treatment increases the level of endogenous circulating estrogens in female rats, contrary to studies from other research groups (114). Mair et al. also demonstrated that E2 can increase proliferation and inhibit BMPR2 signaling in PASMCs from healthy controls (249).

In summary, animal studies examining estrogen signaling in the pulmonary vasculature during PH have produced conflicting and paradoxical results (Table 5). Exogenous E2 administration improves PH outcomes and limits pulmonary vascular remodeling in HPH (92, 115, 208, 296, 335, 474), MCT-PH (99, 432, 483), and SuHx-PH (114, 213, 232, 234). On the other hand, attenuation of estrogen signaling by aromatase inhibition or ER antagonists appears to be protective in the HPH, SuHx-PH, and BMPR2 mutation models of disease in female rodents (248). Transgenic mouse models of PH have also identified estrogen as a necessary cofactor and mediator of specific disease pathways that contribute to pulmonary vascular remodeling (51, 72, 462). Differences in model species (mouse vs rat), animal age, or estrogen source (endogenous vs. exogenous) between studies may contribute to these conflicting results. Dosing strategies for estrogen or its inhibitors need to be taken into consideration. This is of particular importance, as E2 clearly can exhibit dose-dependent effects and since ER inhibitors can also be partial agonists (109, 263). Tissue- or cell-type-specific effects of estrogen or its inhibitors can be pronounced and induce variability as well. Furthermore, the timing of estrogen administration or inhibition relative to the onset and progression of PH may be an important factor. Lastly, confounders such as diurnal variations, estrogen coexposure in animal feeds, and even the gender of the animal handler need to be considered (383). More research into estrogen signaling in the lung considering these modifiers is required to parse the differential and contradictory effects shown here.

Estrogen Metabolites in Pulmonary Hypertension

Several estrogen metabolites have been identified as disease modifiers in PAH and may explain some of the divergent and paradoxical effects of estrogen signaling in disease models. 2-ME2, a nonestrogenic metabolite of E2, has been shown to be protective in several models of PH. Tofovic et al. were the first to demonstrate that 2-ME2 administration attenuates MCT-PH in male rats (420). 2-ME2’s protective effects were further characterized in several studies in which this metabolite attenuated MCT-PH in both male and female rats (418, 421) as well as in a model of bleomycin-induced PH and fibrosis (using ovariectomized female rats (422)). The synthetic analogue of 2-ME2, 2-ethoxyestradiol, exerts similar antimitogenic effects in the pulmonary vasculature and attenuates MCT-PH (423). 2-ME2 exerts strong antimitotic effects in ECs and inhibits ET-1 and MAPK activity even more acutely than E2 (81, 84). It has also been shown to inhibit systemic vascular remodeling by downregulating Akt and ERK1/2 activation, while upregulating cyclooxygenase-2 and the cell cycle inhibitor p27 (22). As 2-ME2 can be synthesized from circulating E2 by ECs and SMCs (82,485), it is possible that the protective effects of E2 are mediated at least in part by conversion to 2-ME2. Indeed, E2 does not inhibit proliferation of SMCs collected from Comt knockout mice (484), which do not express the necessary enzyme for formation of methoxyestradiol metabolites (2-ME2 and/or 4-ME2), indicating that synthesis of 2-ME2 may be important for estrogen effects in some cell types. Furthermore, pathological conditions such as hypoxia and inflammation (both factors in PAH/PH etiology and progression) have been shown to decrease activity of CYP1A1 (one of the enzymes responsible for E2 conversion to the 2-ME2 precursor 2-OHE2). Since this would result in limiting the 2-hydroxylation pathway (107, 407) and 2-ME2 production, it is conceivable that decreased 2-ME2 production could contribute to hypoxia- or inflammation-mediated PH. However, at least in hypoxia, this conversion is not necessary for E2 to attenuate PH development (208). In addition, selective ER agonists recapitulate some or all of the effects of E2 administration in many PH models, indicating that 2-ME2 is not the sole mediator of E2’s salutary effects. Lastly, at least in one study, 2-ME2 administration exhibited only modest effects (101).

In contrast to the antiproliferative effects of 2-ME2, the estrogen metabolite 16α-OHE1 has been identified as a contributor to PAH development. An association among BMPR2 genetic status, 16α-OHE1 production, and PAH development was first identified by West et al. (457) and Austin et al. (12). The authors found that a single-nucleotide polymorphism (SNP) in CYP1B1 (one of the enzymes responsible for production of 16α-OHE1 as well as 2-OHE2; Figure 3) results in lower CYP1B1 activity and a lower urinary 2-OHE2/16α-OHE1 ratio in BMPR2 carriers that developed PAH as compared to nonaffected carriers. The data from animal models of PH supports the theory that increased 16α-OHE1 activity promotes PH. White et al. demonstrated that pulmonary CYP1B1 levels are increased by both hypoxia and SuHx-PH, that Cyplbl knockout mice are resistant to HPH, and that administration of the CYP1B1 inhibitor 2,3′,4,5′-tetramethoxystilbene (TMS) attenuates PH resulting from both hypoxia or SuHx (463). While it is unclear why White et al. found increased CYP1B1 expression as compared to the decreased activity noted by West and Austin, both groups implicated 16α-OHE1 in PAH development. In vitro experiments by White et al. confirmed that 16α-OHE1 provoked proliferation in human PASMCs, particularly PASMCs collected from PAH patients. Finally, the authors demonstrated that 16α-OHE1 injection could elicit a PH phenotype in mice (463). Other models of PH have also been shown to exhibit upregulated CYP1B1 expression in the lung, including SERT+ mice, anorexigen-induced PH female mice, and female MCT-PH rats (71, 464, 483), identifying altered CYP1B1 activity as a disease mediator in experimental PH. Interestingly, administering 16α-OHE1 to BMPR2 mutant mice doubles the penetrance of PH (101) and disrupts cellular metabolism through upregulation of miRNA-29 (52). Both CYP1B1 and 16α-OHE1 are attractive therapeutic targets due to a conserved expression pattern between humans and animal models, and the dramatic effect of inhibition in animal models.

Taken together, while there is a discrepancy between animal and human studies with regard to CYP1B1 activity/expression, there is robust evidence that 16α-OHE1 promotes a PAH phenotype in vivo and in vitro. 2-ME2, on the other hand, can promote adaptive processes. These studies have led to a paradigm of “bad” and “good” estrogen metabolites and altered estrogen metabolism in PAH.

Human Studies

Despite the epidemiologic observations reviewed above, until recently few clinical studies have explored the role of estrogens in PAH pathogenesis, risk, and outcomes. Higher levels of circulating E2 have been found in both men and postmenopausal women with PAH as compared to age- and body mass index-matched healthy controls (19, 441, 471). In a study comparing 23 men with WHO Group 1 PAH (including idiopathic PAH, heritable PAH, and CTD-associated PAH) and 67 sex-, age-, and weight-matched healthy controls, higher E2 levels were associated with the risk of PAH, such that a 1-unit increase in E2 increased the risk of PAH 50-fold, and also shorter 6MWD in PAH cases (441). These results were replicated in a larger prospective cohort of 95 men with idiopathic PAH from China; in this study, higher E2 levels were also independently associated with death in PAH patients (471). In 112 postmenopausal women ≥ age 55 with idiopathic PAH, heritable PAH, or CTD-associated PAH, PAH cases had higher circulating E2 levels as compared to matched controls, and higher E2 levels were associated with shorter 6MWD (each per unit increase was associated with a 32-m decrement in 6MWD, the minimally important difference for the 6MWD in PAH), higher RAP, and worse functional class (19, 259). In postmenopausal women with limited cutaneous SSc, the use of HT (most commonly combination estrogen/progesterone) after menopause decreased the risk of the development of echocardiographic PH, although there were no differences in single measurements of E2 levels between those who did and did not develop PH (30). A second study demonstrated that menopause was an independent risk factor for the development of PH in SSc (363). There have been no published studies that measure circulating estrogens in premenopausal women with PAH, perhaps because it is difficult to control for the impact of menstrual cycle variation on these levels.

As mentioned above, 2-OHE2 and 16α-OHE1 have different effects on inflammation and mitogenesis and have previously been implicated in oncogenesis and vascular metastatic invasion (83). Austin and colleagues discovered alterations in E2 metabolism in heritable PAH and demon-strated that a CYP1B1 polymorphism increased the risk of PAH penetrance in women with BMPR2 mutations, but not in carrier men (12). In addition, the presence of the CYP1B1 mutation in PAH patients was associated with lower urinary 2-OHE2/16α-OHE₁ ratios (12, 13, 463). In a small study, 10 men with heritable PAH had a higher proportion of the mitogenic metabolite (16α-OHE1) compared to the anti-inflammatory/antiproliferative metabolite (2-OHE2) as compared to healthy controls (101). This same group described direct ERα binding to the BMPR2 promoter, leading to reduced BMPR2 gene expression in females, and demonstrated that 16α-OHE1 promotes the development of heritable PAH via upregulation of microRNA-29, which plays a key role in cellular energetics and metabolism (13, 52). Interestingly, increased expression of 16α-OHE1 has been demonstrated in pulmonary arterioles from the explanted or autopsied lungs of PAH patients as compared to non-PAH lungs (463). These important connections among sex, E2 metabolite balance, altered BMPR2 expression, and effects on cellular metabolism may underpin the female predominance of heritable (and perhaps idiopathic) PAH. The role of these pathways in men with PAH and nonheritable disease has not been elucidated.

The Pulmonary Vascular Complications of Liver Disease Cohort enrolled patients with chronic liver disease and then performed a case-control study to determine genetic risk factors for portopulmonary hypertension in 15 candidate genes of interest (340). Polymorphisms in ESR1 and CYP19A1 (which encodes for aromatase) were associated with altered risk of developing portopulmonary hypertension. Moreover, biologic activity of the aromatase variants was suggested by a correlation between increased circulating E2 levels and genotype (340). Both ERα and aromatase are present in human lungs. Specifically, increased ERα expression has been demonstrated predominantly in PASMCs from explanted lungs of women with PAH as compared to control lungs (469). Peripheral aromatase activity (which accounts for most of the estrogen production in postmenopausal women and men) is present in SMCs of the small muscular PAs of explanted lungs from women with PAH. Complementary animal experiments implicate this enzyme in pulmonary vascular disease pathogenesis (248, 469).

A recently completed randomized, double-blind, and placebo-controlled trial demonstrated a significant reduction in circulating E2 levels with anastrozole treatment versus placebo, but no effect on echocardiographic RV measures at 12 weeks in 18 men and postmenopausal women with Group 1 PAH (193). Other hormone levels (specifically testosterone, progesterone, and DHEA-S) were not affected. Active treatment also resulted in an improvement in 6MWD (+26 m) compared to placebo (−12 m) (median % change from baseline was +8% versus −2%, respec-tively [p = 0.042]), and there were no adverse events. This small pilot study demonstrates the feasibility of hormonal manipulation as a treatment strategy in PAH. A longer, larger Phase II multicenter, randomized, double-blind, and placebo-controlled trial of anastrozole is ongoing (NCT03229499) in men and postmenopausal women with WHO Group 1 PAH. Tamoxifen, a selective ER modulator, is also being studied in a single-center, randomized, double-blind, and placebo-controlled Phase II trial in subjects with idiopathic, heritable, drug- or toxin-induced, or CTD-associated PAH (NCT03528902).

Estrogens and RV function

Cell Culture and Animal Studies

Previous studies in the LV demonstrated that estrogen signaling promotes the maintenance of myocardial metabolism and inhibits inflammation, fibrosis, and apoptosis in response to acute or chronic injuries such as pressure overload (104, 173, 216, 230, 452, 453). Of note, cardioprotective effects of the PDE5 inhibitor sildenafil in two animal models of left ventricular dysfunction are estrogen dependent (354). Healthy women exhibit superior RV systolic function compared to men, a relationship that correlates with E2 levels and that persists among patients with PAH (192, 196, 444). These observations indicate that E2 may exert direct RV-protective effects during PAH/PH, altering disease progression independent of its effects in the pulmonary vasculature. Indeed, several animal studies have confirmed this notion.

E2 supplementation of male rats during hypoxia reduced RVSP and RV hypertrophy, while increasing cardiac output (208). Cotreatment with the ERα antagonist MPP increased RV hypertrophy and reduced cardiac output to levels seen in untreated hypoxic animals. Nonselective (dual) ER blockade was required to attenuate other E2-mediated effects inthe RV such as ERK1/2 inhibition. These results indicate that estrogen signaling in the chronically hypoxic RV may involve multiple ERs, which initiate both unique and redundant downstream signaling. However, it should be noted that these effects on RV structure, function, and signaling may be secondary to the lower RVSP and PA remodeling noted with E2 treatment.

Several studies found that E2 mediates RV-protective effects in MCT-PH. Umar et al. treated male rats with E2, ERα agonist (PPT), or ERβ agonist (DPN) after MCT injection (432). E2 supplementation improved RVEF and decreased RV hypertrophy. E2 also promoted neoangiogenesis in the RV, thereby increasing vessel density. This observation is critical as vessel rarefaction and impaired angiogenesis in the RV is suggested to play a major role in RV failure during PH (reviewed in Ref. 113). Protective effects of E2 were recapitulated by DPN, suggesting that ERβ is involved in mediating its cardioprotective effects. Another study by the same group demonstrated similar effects of E2 administration after MCT injection of aged (12–14 months) female ApoE knockout mice, indicating that estrogen may be vital in attenuating PH in aged subjects with a disrupted reproductive cycle (433). Nadadur et al. demonstrated that E2 treatment reduces RV fibrosis after MCT administration (298). Effects of E2 in this study were largely recapitulated by the ERβ agonist DPN, similar to the effects noted by Umar et al. In recent studies from Alencar et al., activation of GPR30 with the GPR30 agonist G1 attenuated the effects of MCT in a rat model (5, 6), indicating that nongenomic estrogen signaling may also be critical to RV protection. As in the studies performed in HPH models, E2 or ER agonist also affected RVSP and PA remodeling in all these studies, making it difficult to dissect whether ER signaling exerts direct effects on the RV or whether RV structure and function improved as a result of lower RV afterload. However, Nadadur et al. demonstrated that E2 administration decreases fibrosis markers in cultured cardiac fibroblasts (298), indicating that this cell population is a direct target of E2.

More recent studies evaluated E2’s effects in angiopro-liferative PH. Using the SuHx-PH animal model, multiple groups have identified robust RV-protective effects of E2 (114, 213, 232-235). In one study, both endogenous (intact females) and exogenous (E2 repletion in ovariectomized rats) estrogens improve cardiac output and exercise capacity, and attenuate SuHx-induced increases in expression of pro-inflammatory and proapoptotic mediators as well as markers of mitochondrial dysfunction and oxidative stress. In addition, E2 increased abundance of the pro-angiogenic and pro-contractile peptide apelin (114). SuHx-PH also decreases expression of ERα in the RV, while E2 repletion increased ERα abundance. ERβ, GPR30, and aromatase, on the other hand, were not altered. E2 supplementation was also RV protective in male SuHx-PH rats, and administration of an ERα agonist replicated these effects. ERβ agonist treatment, on the other hand, was less efficacious, suggesting that ERα is primarily mediating E2’s RV-protective effects. A later study demonstrated that E2 abrogates decreases in RV function in SuHx-PH induced by an acute exercise challenge, an effect that was accompanied by increased RV antiapoptotic signaling, eNOS activation, and signs of improved autophagic flux. Similar to the results from Nadadur et al., Lahm et al. noted that E2 inhibited RV fibrosis (213). Concomitant inhibitory effects on RVSP and PA remodeling in both studies precluded determining whether E2’s effects on the RV were direct or indirect. This question was elegantly addressed by Liu et al. (233, 234). Using a SuHx-PH mouse model with E2 repletion in ovariectomized females, these authors demonstrated that E2-mediated cardioprotective effects in this model were both direct (by increasing RV contractile function) and indirect (by decreasing collagen accumulation and increasing compliance in the proximal PA). These authors also examined mitochondrial function in the RV and found that E2 supplementation improved both mitochondrial density and respiratory function compared with placebo-treated animals (233). Finally, this group demonstrated that E2 treatment preserves PA compliance after SuHx treatment, which reduces the pulsatile load on the RV, leading to improved RV function and improved ventricular-vascular coupling (232, 235).

In summary, data from multiple animal models clearly demonstrate that E2 improves RV function, structure, and biochemical processes in PH (Table 5). Interestingly, even in animal models of E2-mediated pulmonary vascular prolifer-ation, RV-protective effects such as reduced RV hypertrophy are often observed (419, 462). We are now beginning to understand the mechanisms of estrogenic signaling in the RV. All three ERs have been implicated in mediating RV protection (albeit with differences between model system), but their individual contributions and importance must be studied in more detail. In addition, PA-independent effects of estrogenic signaling (e.g., using a PAB model) need to be studied in more detail.

Human Studies

A key and unanswered question in human PAH is whether estrogens have a direct role in RV adaptation, which may serve to explain why more women than men develop PAH but have preserved RV function and better survival (Table 2). While such a role of estrogens is clearly suggested by the animal studies reviewed above, there are no human studies directly linking circulating estrogens to RV morphology in PAH, and the only observational data to date has been performed in an epidemiologic cohort without clinical cardiovascular disease (the MESA-RV Study). In postmenopausal women from MESA-RV using HT, higher E2 levels were associated with higher RVEF and lower RV end-systolic volume, but this relationship was not seen in non-HT users or men (444). The association seen in HT users only may be explained by greater E2 levels, a higher degree of variation in E2, altered or unmeasured estrogen metabolites, or protein/receptor interactions. Exogenous HT may lead to upregulated ER tissue expression and altered E2 sensitivity (76). High E2 states have been associated with heart neovascularization, and human ventricular myocardium contains functional ERs, which may result in adaptive remodeling and better RV systolic function (174, 272).

Cytochrome P450 enzymes are preferentially expressed in the RV (as compared to the LV) in humans (414). A follow-up study in MESA-RV demonstrated that genetic variation in CYP1B1 was independently associated with increased RVEF in postmenopausal women (443). This polymorphism is tightly linked to the variant that increased the risk of PAH in BMPR2 carriers (12) described above as well as variants tied to angioinvasion in cancer (75, 190, 312, 456). There were no associations noted in other candidate genes important in estrogen signaling and metabolism including ESR1 or ESR2 or CYP19A1. Urinary estrogen metabolites were also measured and were associated with RVEF but did not mediate the CYP1B1 SNP-RVEF relationship. Interestingly, the CYP1B1 polymorphism-RVEF association was strongest in black women, who have the highest female predominance in PAH. Activity and by-products of the cytochrome P450 subfamilies and E2 metabolite balance can be altered acutely during hypoxia, inflammation, with the onset of vascular disease, and with daily dietary changes (184, 275). This suggests that the impact of these metabolites on the cardiopulmonary unit is complex and may vary depending on an individual’s race/ethnicity, age, hormonal milieu (endogenous and exogenous), and disease course. These observations lend support to the hypothesis that sex hormones and their genotypes may have pleiotropic effects on the pulmonary circulation and RV and give rise to unique sex-based phenotypes in PAH. Given the profound effects of E2 on RV function, the two ongoing trials of E2 reduction as a treatment strategy in PAH (anastrozole, NCT03229499 and tamoxifen, NCT03528902) are incorporating echocardiographic measures of RV function as safety and efficacy end points in subjects with PAH.

In summary, data from animal and human studies confirm that estrogens are a clinically relevant modifier of RV function in PAH/PH (Tables 2 and 5). Elucidating the exact role of estrogenic signaling during RV adaptation to increased afterload may allow for the development of targeted therapies that improve cardiac adaptation in PAH/PH while avoiding potential off-target effects in the pulmonary vasculature.

Testosterone in PH

Testosterone Signaling and Metabolism

Testosterone and its metabolite DHT are ligands for the androgen receptor (AR). DHT is a far more potent androgen than testosterone (approximately 10 times) due to its relatively higher binding affinity and slower dissociation rate from the AR (134). Like other steroid hormones, androgens signal primarily through a genomic pathway to alter gene expression. Ligand binding initiates AR dimerization, nuclear localization, binding to androgen response elements in the DNA, and transcriptional modification of target genes. Nuclear cofactors (coactivators/corepressors), chromatinmodifying enzymes, and posttranslational modification of the AR all play a role in modifying androgen signaling in target cells (261). Like estrogens, androgens may also exert effects through a nongenomic signaling pathway acting through cytoplasmic or membrane-bound ARs. By doing so, testos-terone or DHT can rapidly initiate cell signaling pathways including rapid calcium influx (128), MAPK signaling (320), PI3K/AKT activation (121), cytoskeletal reorganization (311), or apoptosis (65, 413). Androgen signaling is vital for development and function of the male reproductive tract and development of secondary sex characteristics in males, but also plays a critical role in the cardiovascular system (85). Particularly relevant to PAH/PH, the AR is expressed in vascular SMCs, EC, lung tissue, and both atrial and ventricular cardiomyocytes (85, 237, 238, 276). Circulating testosterone levels fall in aged males (Table 4), and animal models indicate that this effect may be compounded by a reduced androgen sensitivity of the vasculature (94). Isolated coronary arteries from aged male rats display a muted response to testosterone in vitro (94), while AR expression in heart tissue is dependent on circulating testosterone levels (124). Careful study of androgen signaling in the cardiovascular systems of men and women in the context of aging, estrogen/androgen balance, and pulmonary vascular disease may provide data relevant to the sexually dimorphic progression of PAH.

As discussed in the hormone synthesis section, circulating testosterone may be converted to E2 by aromatase in peripheral tissues. It is therefore conceivable that any experimental effects of endogenous or exogenous testosterone administration could be mediated by estrogen signaling in target tissues after aromatization. To this point, the ratio of circulating E2/testosterone has been associated with cardiovascular disease risk in epidemiologic studies (68, 489). Careful monitoring of hormone levels and/or administration of nonaromatizable DHT is required in experimental systems to accurately identify the effect of androgens in PH.

Testosterone in the Pulmonary Vasculature

Cell Culture and Animal Studies

Studies in isolated human (343, 382) and rat (95) pulmonary vessels demonstrated that testosterone is a powerful vasodilator in this vascular bed. While testosterone elicits vasodilation in tissue collected from either sex, the vasodilatory response appears to be greater in male tissue (95, 343). Acute testosterone-induced vasodilation is nongenomically mediated by antagonistic effects on voltage-gated calcium channels and a subsequent reduction in calcium influx into SMCs (139, 181, 364). While the acute vasodilation response to testosterone involves vascular SMCs, androgens also act through genomic signaling pathways to promote NO synthesis in cultured systemic ECs (123, 280). Both classical AR signaling and activation of ER signaling after aromatization have been implicated in this process (reviewed in Ref. 241), and the pulmonary vasculature of male eNOS knockout mice exhibits increased muscularization compared to female knockouts (278). While testosterone-induced vasodilation would be expected to protect the male lung against incipient PAH/PH, effects of testosterone on other clinically relevant processes in PAH (e.g., proliferation, metabolism, and inflammation) are largely unknown. In particular, the role of androgens in pulmonary vascular remodeling during PAH has not been studied. In cultured systemic vascular SMCs and ECs, androgenic signaling promotes proliferation (117, 384). In addition, androgen signaling opposes EC dysfunction, oxidative stress, and inflammation (55). These data suggest that testosterone may exert biologically relevant effects in the pulmonary vasculature (Table 6) as well and provide a rationale to study these pathways in more detail.

Table 6.

Animal studies of progesterone and testosterone in PH.

Model	Species	Major findings	References
Isolated lungs or isolated PAs	Rat, sheep	Testosterone, progesterone ↑ PA vasorelaxation	(95, 116, 180, 226, 382)
Hypoxia-induced PH (HPH)	Rat	Testosterone ↑ RV hypertrophy in castrated male rats	(286)
Monocrotaline-induced PH (MCT-PH)	Rat	Progesterone protective in OVX females; ↓ RVSP, ↓ RV hypertrophy, ↓ PA remodeling, ↓ mortality	(416)
Pulmonary artery banding	Mouse	Testosterone ↓ RV function and ↑ RV remodeling	(154)

OVX, ovariectomy; PA, pulmonary artery; RV, right ventricle; RVSP, right ventricular systolic pressure.

Human Studies

Testosterone deficiency has been demonstrated in a number of chronic diseases (341, 424). Two studies have measured circulating testosterone levels in men with PAH with discordant results. In a cohort of 95 Chinese men with idiopathic PAH, testosterone deficiency was found in 54% of PAH patients as compared to matched healthy controls (471). In 23 men with idiopathic, heritable, or CTD-associated PAH, there were no differences between total and bioavailable testosterone levels in PAH cases as compared to matched controls, perhaps due to the smaller sample size or greater variation in PAH diagnoses and race/ethnicity in this study (441). In both studies, the findings with circulating E2 levels (higher levels associated with PAH and more severe disease) were more robust and drove associations between greater E2/testosterone ratios and the risk of PAH. Total testosterone levels were not associated with disease severity in either study nor survival in the Chinese cohort. Lower levels of total testosterone and bioavailable testosterone increased the odds of PAH threefold in a study of postmenopausal women with Group 1 PH, but there were no consistent associations observed between lower testosterone (or bioavailable testosterone) levels and markers of PAH severity (19). Taken together, testosterone levels may be lower in PAH than in health (as has been described for many chronic diseases), but currently available human studies do not suggest that lower circulating testosterone levels are associated with worse pulmonary vascular disease.

Testosterone in the RV

Cell Culture and Animal Studies

RV hypertrophy in response to increased PVR is a hallmark of PH. Androgens promote cardiomyocyte hypertrophy in vitro (256), in animal models (43), and in cases of anabolic steroid abuse (2). However, this effect has primarily been studied in the LV. Compounding this effect, cardiac hypertrophy significantly elevates expression of 5α-reductase, increasing conversion of testosterone to DHT and promoting a positive feedback loop of androgen signaling and hypertrophy (415). Two recent studies have examined the role of androgen signaling in cardiac hypertrophy resulting from pressure overload. Montalvo et al. showed that male mice exhibit more severe LV dilatation in response to transverse aortic constriction (TAC) compared with female or castrated animals, and that this effect was TGFβ dependent (284). Zwadlo et al. demonstrated that DHT appears to drive this phenotype as inhibition of 5α-reductase significantly reduced LV hypertrophy in male mice after TAC (491). While there are physiological differences between the left and right ventricle, animal studies focusing on the RV demonstrate a similar hypertrophic effect of androgens. Specifically, castrated male rats demonstrate RV hypertrophy when administered testosterone (286). Interestingly, the effects of testosterone on RV hypertrophy are additive to those of hypoxia, an effect not seen with other steroid hormones. In a PAB model of RV dysfunction, castrated male mice exhibit less RV hypertrophy, while testosterone replacement after castration leads to increased levels of hypertrophy (154). In parallel, testosterone promotes RV fibrosis, whereas testosterone deprivation appears to improve survival. This indicates that testosterone may promote a maladaptive type of RV hypertrophy.

Human Studies

Testosterone increases the myocardial inflammatory response and promotes cardiac remodeling (200, 300). Epidemiologic studies have shown that left heart failure is characterized by testosterone deficiency and is associated with worse cardiovascular outcomes in men (199, 251). The role of androgens in cardiovascular health remains controversial, however, because of mixed results with testosterone supplementation in human studies (41, 102). In the same study from MESA-RV, which demonstrated that higher E2 levels were linked to higher RVEF in postmenopausal women HT users, bioavailable and total testosterone levels were associated with greater RV mass and larger RV volumes (including RV stroke volume) in men only and were independent of LV measures (444). It is unknown whether these associations may be adaptive or maladaptive in pulmonary vascular disease as this study was cross-sectional in nature and performed in a cohort without clinical cardiovascular disease. There were no relationships noted between circulating testosterone or bioavailable testosterone levels and echocardiographic RV structure or function or natriuretic peptide levels in postmenopausal women with Group 1 PH; studies performed in men with PAH did not assess RV function (19).

In the follow-up genotype-RV phenotype study from MESA-RV, two polymorphisms in the AR gene were associated with RV end-diastolic volume and mass in men only and were dependent on circulating testosterone levels, indicating that these variants may have biologic relevance (443). ARs are present in human cardiomyocytes and stimulate hypertrophy with testosterone binding (256). Testosterone also directly regulates AR transcription during left ventricular hypertrophy in human hearts (415). AR interactions may lead to changes in RV morphology via both genomic and nongenomic effects that depend on the androgen, hormone concentration, cardiac receptor density, and sex of the individual.

Taken together, it appears that androgens may be protective in the pulmonary vasculature by promoting vasodilation, but detrimental to RV remodeling in the face of increased afterload (Tables 2 and 6). This hypothesis fits the clinical data in which men are less likely to develop PAH but display decreased survival rates compared with women. However, there is a paucity of mechanistic studies examining the role of androgen signaling in proliferative processes in the pulmonary vasculature. Studies from the systemic vasculature suggest that testosterone and androgenic signaling may promote PA wall cell homeostasis and also enhance proliferation, suggesting that pulmonary vascular effects of testosterone need to be studied in more detail. More mechanistic studies evaluating androgenic signaling in the RV are required as well.

Progesterone in PH

Progesterone Signaling and Metabolism

Progesterone is primarily synthesized by the ovarian corpus luteum during the menstrual cycle as well as by the placenta during pregnancy. Progesterone is one of the most critical hormones during pregnancy, and many physio-logical changes of pregnancy are progesterone mediated (reviewed in Ref. 153). Progesterone signaling is necessary for differentiation and maintenance of female reproductive tissues including the uterine and mammary epithelium (133). Like other steroid hormones, progesterone signals through binds to a steroid receptor (progesterone receptor, PR) and subsequently modulates gene transcription as well as through nonclassical pathways, which include genomic and nongenomic signaling cascades (120). Cofactors including transcriptional activators/repressors, chromatin-modifying enzymes, and posttranslational modification of the PR may alter progesterone signaling in target cells (133).

PR is expressed in ECs, including proliferative ECs and myofibroblasts found in plexiform lesions isolated from PAH patients (21,455). PR is also highly expressed in the systemic vasculature and in cardiac tissue, such as ECs and SMCs of the aorta, carotid and coronary arteries, and cardiomyocytes (172). Vascular expression of PR in the uterine vasculature varies according to the phase of the menstrual cycle and tends to decrease with age (217). It is possible that PR expression in other tissues demonstrates similar cyclicity. Interestingly, estrogen upregulates expression of the PR in cardiovascular tissue (186, 229), an effect that may be relevant to PAH disease progression.

In vitro culture of ECs revealed that PR activation suppresses cytokine production (122) and inhibits ET-1 synthesis (288). PR knockout mice demonstrate increased vascular medial hypertrophy and SMC proliferation after vascular injury, and isolated vascular SMCs from PR knockout mice are hyperproliferative in culture (188). These studies indicate that PR signaling might play a protective role by limiting the inflammatory, angiogenic, and proliferative phenotypes of PAH/PH; however, exogenous progesterone has also been shown to intensify vascular injury response in wild-type mice (188). Studies to clearly define the role of progesterone signaling in the pulmonary vasculature generally and in the context of PAH specifically are clearly needed.

Progesterone in the Pulmonary Vasculature and RV

Cell Culture and Animal Studies

Investigations of sex and gender differences in PAH/PH have focused on the role of estrogens, while remarkably few studies have examined the role of progesterone. This imbalance is surprising given that both hormones are much more abundant in women compared to men (Table 4).

Progesterone regulates proliferation of both ECs and SMCs (222, 223, 287, 438) and has been shown to be vasodilatory in pulmonary vessels isolated from both rats (95) and rabbits (226). While one study in humans found that oral progesterone supplementation opposed vasodilatory effects of estradiol (281), this study was not conducted in the pulmonary vasculature.

Tofovic et al. demonstrated that progesterone administration mitigates MCT-induced PH in rats (416). Specifically, progesterone supplementation in ovariectomized MCT-PH rats attenuated MCT-induced increases in RVSP, RV hypertrophy, PA remodeling, and mortality compared with untreated ovariectomized MCT-PH rats. Despite the encouraging results of this study, no data on the role of progesterone in animal models of PH has been published in the subsequent decade. Further studies that mechanistically evaluate pulmonary vascular effects of progesterone as well as the interaction between progesterone and estradiol in the context of PAH/PH could generate new knowledge to expand our understanding of sex steroid signaling and sex/gender differences in this field.

Little is also known about the role of progesterone during PH-induced RV remodeling. Progesterone promotes cardiac hypertrophy in other contexts, implying that it may drive RV adaptation as well. Progesterone can induce cardiac protein synthesis (125) as well as initiate hypertrophy in isolated rat cardiomyocytes (57). Additionally, progesterone is the dominant hormone of pregnancy, a condition that results in transient cardiac hypertrophy. Progesterone promotes cardiac hypertrophy in vitro, and PR activation inhibits apoptosis in cultured rat cardiomyocytes (293). One intriguing hypothesis is that progesterone may promote “physiological” (or adaptive) cardiac hypertrophy (56, 57, 283, 472) rather than the maladaptive cardiac hypertrophy associated with severe PH. However, this hypothesis has yet to be tested in the context of PAH/PH. On the other hand, pregnant women with PAH are at a particular high risk for increased morbidity and mortality (153, 273, 400). Whether this is directly or indirectly linked to the increased progesterone levels of pregnancy is unknown. A better understanding of the cardiopulmonary effects of progesterone in health and disease is critical for understanding the mechanisms of adaptive and maladaptive changes in both pregnant and nonpregnant women with PAH.

Human Studies

While progesterone has known effects on the respiratory system, its impact on the cardiovascular system is less well understood (20, 77). Two studies (described above) measured progesterone levels in PAH patients (19, 471). In postmenopausal women, there were no differences in circulating progesterone levels in subjects with PAH as compared to matched controls and no associations between progesterone levels and disease severity in PAH subjects including RV function assessed by echocardiography (19). In the study of men with idiopathic PAH from China, lower progesterone levels increased the risk of PAH and were associated with worse functional class, shorter 6MWD, and more severe hemodynamic impairment (471). There was no association between progesterone levels and mortality in this study, and RV structure and function were not assessed. There have been no studies of circulating progesterone levels in premenopausal women with PAH, which may be more informative than in postmenopausal women, and no human studies to date of the relationship between progesterone and RV performance in PAH or in health.

Taken together, the limited body of literature on progesterone in PAH/PH demonstrates that this hormone may be beneficial in the pulmonary vasculature and possibly even in the RV (Tables 2 and 6). However, more studies of progesterone in PH are needed to more clearly define its role.

DHEA in PH

DHEA Signaling and Metabolism

Pregnenolone is the prohormone to progesterone and DHEA. DHEA and DHEA-S are precursors in the biosynthesis of androgens and metabolized directly to androstenedione, testosterone, and subsequently estrogens. DHEA and DHEA-S are produced predominantly in the adrenal cortex and are the most abundant circulating endogenous steroids but wane with aging. The hormone has been shown to have direct genomic and nongenomic effects on vasculature as well as cardiomyocytes (254). DHEA binds directly to vascular endothelium to activate NO synthase and regulates ET-1 production, two key drivers in PAH pathobiology that are also major treatment targets (50, 236, 254, 301). When human systemic vein ECs are exposed to DHEA, inflammatory signaling is reduced (7, 451). DHEA has been shown to rescue cardiomyocyte hypertrophy induced by ET-1 (301) and to prevent myocardial fibrosis and contractile dysfunction through the restoration oxidative balance and downregulation of advanced glycation end products (AGEs) and its receptors, reducing tissue levels of collagen and fibronectin (9, 10).

DHEA in the Lung Vasculature

Cell Culture and Animal Studies

DHEA exposure at variable concentrations has been shown to induce phenotypic changes in human ECs in vitro (7, 23, 164, 221, 316, 451). Effects include enhanced eNOS expression, NO synthesis, and variable ET-1 secretion (50, 164). Human PAECs actively metabolize DHEA, and treatment of PAECs from PAH patients decreases activation of STAT3 (277, 316), an important mediator of pulmonary vascular remodeling. DHEA appears to be consistently beneficial in experimental PH. Although these studies have been done in predominantly male animals, DHEA has been used as both a prevention and rescue strategy following exposure to hypoxia, altitude, MCT, MCT-pneumonectomy, and SuHx (8, 37, 88, 144, 159, 316). These studies and the mechanisms by which DHEA is proposed to prevent or reverse experimental PH as well as the effects on cardiomyocytes are summarized in Table 7.

Table 7.

Experimental data supporting beneficial effects of DHEA in PH.

Model	DHEA effects	Proposed mechanism
Hypoxia	Prevent, rescue PH	BKCa channels (37, 144)
	Rescue PH	BKCa inhibits 5HT-, KCl-induced SMC growth (89)
	Rescue RV	Reduces cardiomyocyte proliferation (88)
Altitude	Prevent, rescue PH	Enhances sGC (306)
MCT	Prevent, rescue PH	Inhibits Src, STAT3, Pim1
PASMCs		Increases BMPR2, miR-204 (316)
MCT-PNX	Prevent, rescue	Inhibits RhoA/Rho kinase (159)
Sugen/hypoxia	Rescue RV > PH	Reduces RV capillary rarefaction, apoptosis, ROS via STAT3 (8)
Cardiomyocytes	Antichronotropic, antihypertrophic	Reduces T-type Ca channels (254) Inhibition of natriuretic peptide expression
	Antihypertrophic	Reduces ET-1 induced hypertrophy (301) Inhibits BNP

BK_Ca, Large conductance Ca²⁺-activated K channel; 5HT, serotonin; KCl, potassium chloride; SMC, smooth muscle cells; RV, right ventricle; GC, soluble guanylate cyclase; MCT, monocrotaline; STAT3, signal transducer and activator of transcription 3; BMPR2, bone morphogenetic receptor type II; PNX, pneumonectomy; ET-1, endothelin 1; BNP, B-type natriuretic peptide.

Human Studies

In addition to high levels of circulating E2 in PAH, studies of both men and postmenopausal women have demonstrated significantly lower (50%) levels of DHEA-S (which is more stable in blood samples than DHEA) in PAH subjects as compared to age- and body mass index-matched controls (19, 441). In men with PAH, lower levels of DHEA-S were associated with worse hemodynamics (higher RAP and higher PVR). There were more robust associations with disease severity in postmenopausal women (i.e., lower levels of DHEA-S were associated with worse 6MWD, functional class, hemodynamics, and worse RV function by echocar-diography) as well as an association with worse survival such that every unit decrease in DHEA-S was associated with a doubling in the risk of death (19). While DHEA is a prohormone of E2 and testosterone (and testosterone can be aromatized to E2), there were no strong correlations among hormone levels (DHEA-S, E2, and testosterone) and no evidence of effect modification among interrelated hormones, implying that DHEA may have a direct role in the development and progression of PAH and RV adaptation. A second study of unbiased metabolomic profiling demonstrated that DHEA-S and its metabolites were reduced in PAH patients compared to healthy controls and that lower circulating levels of DHEA-S were associated with mortality in PAH patients (337). In eight patients with PH related to chronic obstructive pulmonary disease, open-label treatment with 3 months of DHEA was associated with a significant increase in 6MWD and improvements in hemodynamics without adverse effects (90). These studies have led to the third randomized clinical trial of hormonal modulation as a treatment strategy in PAH; a single-center, double-blind, and placebo-controlled crossover trial of DHEA supplementation in subjects with Group 1 PH is ongoing (NCT03648385).

DHEA and the RV

Animal Studies

Several studies have demonstrated a role for DHEA in cardiomyocyte adaptation to injury (Table 7). In a PH/RV failure model, chronic DHEA treatment over 5 weeks in rats exposed to SuHx reduced RVSP and rescued CI and echocardiographic RV function (8). Treatment with DHEA inhibited capillary rarefaction, apoptosis, oxidative stress and NADPH levels, and fibrosis in the RV of these animals via reduced Rho kinase activity and inhibition of transcription factors implicated in maladaptive cardiac remodeling, STAT3 and NFATc3 (8). In rats exposed to chronic hypoxia followed by reoxygenation, DHEA increased RV myocyte density and proliferation, reduced mitochondrial fragmentation, and prevented RV dysfunction during the recovery phase of these experiments (88). These studies demonstrate that DHEA may have a direct and RV-specific effect independent of downstream hormones like E2.

Human Studies

Low circulating DHEA-S levels have been associated with an increased risk of death in heart disease, cardiac allograft vasculopathy, and heart failure severity (25, 100, 155, 290, 373) (Table 2). In the MESA-RV study, higher levels of DHEA were associated with lower RVEF (calculated from RV stroke volume/RV end-diastolic volume), higher RV stroke volume (calculated from RV end-diastolic—RV end-systolic volume), and larger RV end-diastolic volume in postmenopausal women who did not have any clinical cardiovascular disease. While a lower RVEF and larger RV stroke volume seem difficult to reconcile, higher RV end-diastolic volume with higher DHEA levels would result in larger RV stroke volume and numerically but not pathologically lower RVEF given the derivation of these parameters. In fact, virtually all participants had a normal RVEF (these are “disease-free” adults) in MESA and, as reviewed above, in postmenopausal women with PAH, lower DHEA-S levels were associated with worse RV dilatation and dysfunction by echocardiogram (444). The primary end point of the ongoing trial of DHEA supplementation in PAH is RV contractile function measured by cardiac MRI (NCT03648385).

In summary, the conglomerate of basic and clinical studies of DHEA in PH suggests beneficial effects on both pulmonary vascular and RV function (Tables 2 and 7).

Sex Hormone-Independent Effects

Sex steroid signaling clearly is a major driver of sex differences in susceptibility and disease progression in PH/PAH. However, the hormonal milieu is not the sole factor that impacts gender and sex-based differences in pulmonary vascular disease. Emerging research indicates that nonhormonal factors such as immune cell regulation, iron metabolism, and the Y chromosome itself may lead to sex-based differences in PH penetrance and progression.

A recent study demonstrated sexually dimorphic immune responses in experimental PH. In particular, the authors demonstrated that in regulatory T-cell (Tregs)-deficient rats exposed to sugen or hypoxia, females developed more severe PH than males (408). Interestingly, Treg repletion abolished this sex difference. In additional studies, the authors showed that protective vascular effects of Tregs were ER dependent, suggesting a cross talk between the immune system and sex steroid signaling. This observation suggests that females are reliant on normal Treg function to counteract detrimental effects of pulmonary vascular insults. Patient registry data supports this hypothesis as PAH is often associated with autoimmune disorders, and many of these disorders share similar or more skewed gender ratios as in idiopathic PAH (27). This indicates that altered immune responses may contribute to the female predominance in PAH. How sex-based differences in immunity impact PAH, as well as the role of sex steroid signaling in immune cell regulation, remains to be investigated.

Iron deficiency is a common comorbidity with PAH in humans (344, 385). Since iron is a required for degradation of HIFs (prominent drivers of PAH development), iron deficiency may contribute to the development and progression of PAH in some patients (332). Intravenous iron supplementation is currently the focus of a phase II clinical trial in PAH patients (162). Both globally (266) and in the United States (119, 218), iron deficiency is two to three times more prevalent in women than in age-matched men, suggesting that women may be disproportionately vulnerable to iron deficiency-related HIF activation in the pulmonary vasculature. However, this hypothesis has yet to be tested.

Finally, recent research indicates that the Y chromosome itself may play a protective role in PAH progression. Umar et al. used the Four Core Genotype (FGC) mouse model, in which the chromosome complement is independent of gonadal sex, to demonstrate that the presence of the Y chromosome protects mice from HPH development regardless of sex (431). An explanatory mechanism for this phenomenon was proposed by Yan et al. who demonstrated that the Y-chromosome-encoded transcription factor SRY (sex-determining region of the Y chromosome) promotes Bmpr2 expression in cultured male dermal fibroblasts (477). Altered BMP signaling plays a major role in vascular dysfunction in PH/PAH (reviewed in Ref. 243), and germ line mutations in BMPR2 are found in most cases of heritable PAH (14), making the mechanism discovered by Yan et al. conceptually sound. Additional studies evaluating potential SRY-regulated Bmpr2 expression in the lung vasculature and in vivo would be of tremendous value to the field.

How to Put it all Together: Common Themes, Knowledge Gaps, and Pathways Forward

The aggregate of studies reviewed in this article demonstrates that there has been significant progress in the study of sex/gender differences and sex hormone signaling in PAH/PH. In a relatively short time span, the field has moved from mechanistic cell culture and animal studies to human studies including clinical trials of hormonal modulation, suggesting that harnessing sex hormone signaling may provide a powerful new strategy to treat PAH/PH. The observation that sex hormones interact with several major disease modifiers such as BMPR2 signaling, metabolic function, and RV adaptation indicates that sex hormone signaling is indeed a major disease modifier. The major biological effects of the most abundant sex hormones as well as their major effects in animal studies and their association with PAH outcomes in human studies are depicted in Figure 4. Several “themes” have emerged: E2 has pleiotropic and compartment-specific effects, 16α-OHE1 promotes PAH development, and DHEA seems to be uniformly protective (Figure 4). E2’s RV-protective effects may explain why women with PAH have better RV function and live longer than their male counterparts. At the same time, E2 or its metabolites (in particular, 16α-OHE1) may promote pulmonary vascular remodeling and make certain women more prone to PAH development, especially in the context of additional predispositions, such as a BMPR2 mutation. The roles of testosterone and progesterone, on the other hand, have not been well studied. Since sex hormones exert diverse and pleiotropic effects, and since sex/gender differences are mediated by multiple factors (Figure 5), further research is needed to identify context- and compartment-specific signaling pathways and sex-based phenotypes. Nonhormonal factors, such as Y-chromosome-mediated effects, aging, and immunity, have recently been identified and may affect disease development and/or progression. Consideration of these (as well as potential unidentified) factors and nuances will ultimately solve the “estrogen puzzle” of PAH. An overview of the major current knowledge gaps in the field is provided in Table 8. Given the pleiotropic effects of many sex hormones (and in particular, E2), it remains to be determined whether inhibiting or enhancing one specific hormone will be meritorious. Selectively targeting one receptor or one metabolite may be a more precise approach with less off-target effects. In addition, sex-based treatment strategies may depend on factors such as receptor expression, sex hormone abundance, ratios between various sex hormones, age, comorbidities, or specific genetic or epigenetic landscapes.

Figure 4. Overview of major biological effects of the most abundant sex hormones (A) and their net effects in animal studies as well as reported associations with PAH risk and outcomes in human studies (B).

Note that effects of specific sex hormones on outcomes in human studies may be limited to men or pre- or postmenopausal women only. Human studies measured DHEA-S (dehydroepiandrosterone sulfate). *Data based on study only. #Inconsistent associations across studies. DHEA, dehydroepiandrosterone; PA, pulmonary artery; PAEC, pulmonary artery endothelial cell; PAH, pulmonary arterial hypertension (human studies); PASMCs, pulmonary artery smooth muscle cells; PH, pulmonary hypertension; RV, right ventricle.

Figure 5. Simplified overview of factors contributing to sex/gender differences and sexual dimorphism in PAH and PH.

Note that significant cross talk exists between the factors and mediators listed in this figure. BMPR2, gene encoding bone morphogenic protein receptor 2; CYP1B1, gene encoding cytochrome P450 1B1; CYP19A1, gene encoding aromatase; DHEA, dehydroepiandrosterone; ESR1, gene encoding estrogen receptor α; SNP, single-nucleotide polymorphism.

Table 8.

Knowledge gaps and pathways forward.

Are there gender differences in RV remodeling/adaptive or maladaptive signaling in human RV failure?

What is the role of sex steroid receptors in the human RV?

Study of the entire “hormonosome” and interactions with cellular metabolism in experimental and human studies

What are the roles of genetics, epigenetics, environmental exposures, and cultural factors in mediating gender differences in human PAH?

Need for carefully executed human studies in premenopausal women with PAH and other life cycle transitions (e.g., puberty, menopause, and transgender individuals)

Need to study X- and Y-chromosome-mediated effects

Improve translational potential in animal studies—e.g., sex balanced experiments, age and estrous as a modifiers of experimental conditions

Summary and Conclusion

This article has comprehensively reviewed gender differences in human PAH and sex differences in animal studies as well as the physiology of sex steroid signaling in health and PAH. The role of nonhormonal contributors to sex and gender differences in PAH/PH is less well described, but these factors may play a significant role as well. Sex, gender, and sex hormones clearly are major disease modifiers in experimental PH as well as human PAH. Manipulation of sex steroid signaling pathways may open up several new treatment strategies. In addition, several sex-based phenotypes exist, suggesting that therapeutic strategies may need to be tailored toward such specific phenotypes. A better understanding of sex hormone signaling and sex steroid-independent factors will lead to novel and targeted therapeutic approaches for PAH and PH patients of either sex.

Didactic Synopsis.

Major Teaching Points

• Pulmonary hypertension (PH) encompasses a heterogeneous group of diseases organized into five groups based on their predominant underlying pathology and clinical phenotype (Figure 1).

• The “estrogen puzzle” refers to two observations in PH research: (i) Many PH classes, particularly group 1 (PAH), are marked by sexually dimorphic disease presentation wherein women are at increased risk for disease development but display increased survival compared with men and (ii) animal models demonstrate contradictory effects of estrogen signaling in PH disease progression (protective as well as detrimental).

• Human and animal studies have shown varied effects of 17β estradiol (E2) in the pulmonary vasculature in PAH/PH, but consistently show that E2 promotes healthy RV function and adaptation. Disease-promoting effects of E2 in the pulmonary vasculature are mediated at least in part by pro-proliferative metabolites (e.g., 16α-hydroxyestrone).

• Few studies have examined the effect of progesterone or androgen signaling in PH though these hormones possibly play a role in disease susceptibility and progression.

• DHEA is protective in animal models of PH, and circulating DHEA levels in PAH patients correspond positively with PH endpoints and RV function in humans.

List of Abbreviations

ApoE

apolipoprotein E

AR

androgen receptor

BMPR2

bone morphogenetic protein receptor 2

COMT

catechol-O-methyl transferase

CYP

cytochrome P450

DHEA

dehydroepiandrosterone

DHEA-S

dehydroepiandrosterone-sulfate

DHT

dihydrotestosterone

DPN

diarylpropionitrile (ERβ agonist)

E1

estrone

E2

17β-estradiol

E3

estriol

ER

estrogen receptor

EC

endothelial cell

eNOS

endothelial nitric oxide synthase

ESR1

estrogen receptor α gene

ESR2

estrogen receptor β gene

ET-1

endothelin-1

GPR30

G-protein-coupled receptor 30

HIF-1α

hypoxia-inducible factor 1-alpha

HPAH

hereditary pulmonary arterial hypertension

HPH

hypoxia-induced pulmonary hypertension

HPV

hypoxic pulmonary vasoconstriction

HT

hormone therapy

IPAH

idiopathic pulmonary arterial hypertension

LV

left ventricle/left ventricular

MCT

monocrotaline

NO

nitric oxide

OVX

ovariectomy/ovariectomized

PA

pulmonary artery

PAB

pulmonary artery banding

PAEC

pulmonary artery endothelial cell

PAH

pulmonary arterial hypertension (Group 1 PH)

PASMC

pulmonary artery smooth muscle cell

PH

pulmonary hypertension (all groups)

PPT

propylpyrazole triol (ERα agonist)

PR

progesterone receptor

PVR

pulmonary vascular resistance

RV

right ventricle/right ventricular

RVEF

right ventricular ejection faction

RVH

right ventricular hypertrophy

RVSP

right ventricular systolic pressure

SERT+

serotonin transporter overexpression

SMC

smooth muscle cell

SNP

single-nucleotide polymorphism

SuHx-PH

PH induced by sugen combined with chronic hypoxia

16α-OHE1

16α-hydroxyestrone

2-OHE2

2-hydroxyestradiol

2-ME2

2-methoxyestradiol

4-OHE2

4-hydroxyestradiol