Lysophosphatidic Acid Signaling Protects Pulmonary Vasculature from Hypoxia-induced Remodeling

Abstract

Objective

Lysophosphatidic acid (LPA) is a bioactive lipid molecule produced by the plasma lysophospholipase D (lysoPLD) enzyme autotaxin that is present at ≥ 100 nM in plasma. Local administration of LPA promotes systemic arterial remodeling in rodents. To determine if LPA contributes to remodeling of the pulmonary vasculature, we examined responses in mice with alterations in LPA signaling and metabolism.

Methods and Results

Enpp2^+/− mice, heterozygous for the autotaxin-encoding gene, that have reduced expression of autotaxin/lysoPLD and ~half normal plasma LPA, were hyper-responsive to hypoxia-induced vasoconstriction and remodeling, as evidenced by the development of higher right ventricular (RV) systolic pressure, greater decline in peak flow velocity across the pulmonary valve, and a higher percentage of muscularized arterioles. Mice lacking LPA₁ and LPA₂, two LPA receptors abundantly expressed in the vasculature, also had enhanced hypoxia-induced pulmonary remodeling. With age, Lpar1^−/−2^−/− mice spontaneously developed elevated RV systolic pressure and RV hypertrophy that was not observed in Lpar1^−/− mice or Lpar2^−/− mice. Expression of endothelin-1, a potent vasoconstrictor, was elevated in lungs of Lpar1^−/−2^−/− mice, and expression of ET_B receptor, which promotes vasodilation and clears endothelin, was reduced in Enpp2^+/− and Lpar1^−/−2^−/− mice.

Conclusions

Our findings indicate that LPA may negatively regulate pulmonary vascular pressure through LPA₁ and LPA₂ receptors, and that in the absence of LPA signaling, upregulation in the endothelin system favors remodeling.

Introduction

Arterial smooth muscle cells (SMC) regulate vessel contraction and relaxation and, following systemic vascular injury, can undergo phenotypic modulation to a proliferative/synthetic state that likely contributes to the development of atherosclerosis and restenosis ¹. In the highly-compliant pulmonary circulation, large changes in blood flow are accommodated with little alteration of pressure by tightly coordinated vasoactive mediators. Dysregulation in the balance between pulmonary vasodilators and vasoconstrictors can trigger pulmonary vascular remodeling, characterized by an increase in pulmonary SMC proliferation, hypertrophy, and the number of SMC-wrapped small arterioles. As a consequence, pulmonary hypertension (PH) with elevated pulmonary vascular resistance and sustained high pulmonary arterial pressure develops ^2–4. Endothelial dysfunction, including an increase in endothelial permeability and apoptosis, is associated with and likely promotes PH ^{2, 4, 5}. With time, the increase in pulmonary vascular resistance and pressure results in right ventricular (RV) hypertrophy, dilatation, and dysfunction ^{4, 6}. Cardiopulmonary disorders associated with elevated pulmonary artery pressures are one of the leading cause of RV hypertrophy and failure ⁷. Unfortunately, at present, very little is known about the mediators that initiate pulmonary vascular changes underlying PH, making prevention and effective treatment of the disease challenging ⁸.

Lysophosphatidic acid (LPA) is a bioactive lipid molecule present in plasma concentrations (≥ 100 nM) predicted to activate its signaling receptors ⁹, ²⁶. LPA has effects on cultured SMC and endothelial cells that make it an attractive candidate regulator of the pulmonary vasculature in vivo ¹⁰. LPA is a potent autocrine regulator of phenotypic modulation of cultured vascular SMC. Indeed, LPA has been proposed to be the major mediator in serum responsible for dedifferentiation of cultured vascular SMC. Moreover, local administration of LPA in carotid arteries of rodents triggers a remodeling response, which requires SMC dedifferentiation and proliferation ^11–16. LPA may also have important effects on endothelial cell function and endothelial barrier stability ^17–19. In isolated cell systems, LPA has been reported to affect endothelial nitric oxide synthase (eNOS) activity and, consequently, nitric oxide levels as well as endothelin levels, both important regulators of pulmonary vascular tone ^20–22. Thus, LPA has actions on both vascular smooth muscle cells and endothelial cells that trigger a dysregulation in their function that is similar to what is observed in a variety of pathological conditions, including pulmonary hypertension. However, until recently, tools to probe the role of the autotaxin-LPA signaling nexus in vascular pathophysiology were not available.

Recent progress in identifying enzymes and receptors responsible for LPA metabolism and signaling has enabled the development of mouse models and pharmacological tools to investigate the involvement of LPA in cardiovascular regulation and function. LPA signaling responses are mediated by selective G protein-coupled receptors (GPCRs). Eight GPCRs for LPA (LPA_1–8) have been identified ^{17, 19,20, 21}. LPA₁ and LPA₂ are abundantly expressed in vascular smooth muscle cells and endothelial cells ^{22, 23}. In vitro studies in cell lines reveal that LPA receptors share certain function redundancy ^{17, 18, 24}. Although a complete range of sub-type specific receptor antagonists is not presently available, use of LPA receptor selective probes and studies of mice with genetic deletion of LPA receptors have provided valuable information on their pathophysiologic roles in diverse settings, including neuropathic pain, lung inflammation, reproduction, and vascular injury responses ²⁵.

LPA is produced in the blood and other tissues through hydrolysis of lysophosphatidylcholine, catalyzed by the secreted enzyme autotaxin/lysophospholipase D (lysoPLD) encoded by the ectonucleotide pyrophosphatase/phosphodiesterase 2 (ENPP2) gene ²⁶. Enpp2^−/− mice die embryonically and display defects on vessel maturation. Enpp2^+/− mice are viable but have approximately half normal circulating autotaxin/lysoPLD and LPA levels ²⁷.

Substantial evidence identifies a role for LPA in regulating phenotypic modulation of cultured vascular SMC that can be recapitulated in vivo by local administration of LPA to systemic arteries ^{11, 16}. Use of mice with genetic deficiencies and small molecule inhibitors of LPA receptors support a role for the autotaxin-LPA signaling nexus in systemic arterial remodeling in rodent models. In this study we used mice with alterations in LPA signaling and metabolism and previously characterized models of PAH ²⁸ to identify an unexpected role for LPA in regulation of the pulmonary vasculature and hypoxia-induced remodeling.

Methods

All procedures conformed to the recommendations of Guide for the Care and Use of Laboratory Animals (Department of Health, Education, and Welfare publication number NIH 78-23, 1996) and were approved by the Institutional Animal Care and Use Committee. Generation and characterization of Enpp2^+/− and Lpar1^−/−2^−/− mice were as previously described ^{27, 29–31}. The Enpp2^+/− mice were backcrossed >10 generations to the FVB background, whereas the Lpar1^−/−, Lpar2^−/−, and Lpar1^−/−2^−/− mice were backcrossed >10 generations to the BalbC background. Mice were housed in cages with HEPA filtered air in rooms on 14-hour light cycles and fed Purina 5058 rodent chow ad libitum. Details of the methods are included in the Supplemental Materials

Statistics

All results were expressed as mean ± SD. In vitro studies were repeated a minimum of 3 times, and results were analyzed by Student t test or ANOVA. Statistical significance within strains was determined using ANOVA with multiple pair-wise comparisons. Statistical analysis was performed using Sigma-STAT software version 3.5 (Systat Software Inc). Changes with a probability value of less than 0.05 for a type I error were considered significant ³².

Results

Reduction in autotaxin/lysoPLD levels promotes hypoxia-induced pulmonary remodeling in mice

LPA promotes phenotypic modulation of cultured vascular SMC and systemic arterial remodeling in experimental models ^{11, 16}. To determine if LPA contributes to remodeling of the pulmonary vasculature, we examined responses in mice engineered to have reduction in LPA synthesis. Heterozygous Enpp2^+/− mice have 50% of normal plasma LPA levels, but no obvious vascular developmental abnormalities ²⁷. We previously reported lower autotaxin protein levels in the lungs of Enpp2^+/− mice ²⁹. In keeping with those observations, Enpp2 mRNA expression in lungs of Enpp2^+/− mice was 20 ± 10 % that in lungs from WT littermate controls (P<0.05). No difference in lung expression of LPA1-5 receptor mRNA was observed in Enpp2^+/− mice (data not shown).

We used a hypoxia model to stimulate pulmonary vasoconstriction and promote vascular remodeling ²⁸. Exposure of WT mice to hypoxia (FiO₂=0.1) increases pulmonary arterial pressure, which can be monitored by measuring right ventricular systolic pressure (RVSP) (Figure 1A). Following exposure to hypoxia for 3 weeks, a significant increase in RVSP occurred in WT mice (30.3 ± 2.3 mmHg; n = 6) compared to RVSP in age-matched normoxic littermates (25.7 ± 2.1 mmHg; n = 6; P<0.05). Under normoxic conditions, Enpp2^+/− mice displayed a similar RVSP (23.8 ±1.8 mmHg; n =5) to their WT littermates (Figure 1A). However, following a 3-week exposure to hypoxia, Enpp2^+/− mice developed higher RVSP (36.2 ± 5.2 mmHg; n=7; P<0.05) than WT littermate mice (Figure 1A). Peak flow velocity across the pulmonary valve decreased by 8.0 ± 5.7% following hypoxia in WT mice. The decline in peak flow velocity was approximately twice as great (16.2 ± 8.4%) in Enpp2^+/− littermates after exposure to hypoxia (Table 1).

Figure 1. Response of mice with reduced ATX levels (Enpp2^+/−) to hypoxia.

A. Right ventricular systolic pressure (RVSP) of WT (n=6) and Enpp2^+/− (n=5) mice housed in normoxic conditions and WT (n=6) and Enpp2^+/− (n=7) mice with 3 weeks of hypoxia exposure. Individual values (dots) and medium with 25 and 75 confidence intervals (box plots) are presented. Data were analyzed by 2-way ANOVA. *P<0.05. B. Percent muscularization of distal pulmonary arterioles in normoxic or hypoxic WT and Enpp2^+/− mice. Lung sections were immunostained with α-smooth muscle actin and scored as described in Materials and Methods. Non = non muscularized vessels, partial = partially muscularized vessels, and full = fully muscularized vessels. Data are presented as averages of 4 mice/group and analyzed by 2-way ANOVA. The percentages of non-muscularized vessels were used for statistical analysis. *P<0.05.

Table 1.

Peak flow velocity across the pulmonary valve before and after 3 weeks of exposure to normoxia or hypoxia.

Genotype	Condition	N	Baseline	After 3 weeks
Enpp2+/+	Normoxia	4	879 ± 44	993 ± 70
Enpp2+/+	Hypoxia	6	899 ± 76	824 ± 55
Enpp2+/−	Normoxia	5	928 ± 66	923 ± 41
Enpp2+/−	Hypoxia	7	961 ± 93	802 ± 95

^*Results are presented as mean ± SD in mm/sec.

To evaluate the role of pulmonary vascular remodeling in the enhanced RVSP and peak flow response of the Enpp2^+/− mice to hypoxia, we examined muscularization of small pulmonary arterioles. Lung sections stained with α-smooth muscle actin, a marker for smooth muscle cells, were scored for non-, partially, and fully muscularized small pulmonary arterioles. The overall pulmonary histology of Enpp2^+/− mice was not notably different from WT controls; in particular, WT and Enpp2^+/− littermates maintained in normoxic conditions did not differ in the percentage of muscularized arterioles. Following a 3-week exposure to hypoxia, the percentage of muscularized small arterioles increased by ~1.75 fold in WT mice. Relative to the WT controls, lungs from Enpp2^+/− mice exposed to hypoxia had a higher percentage of fully muscularized and a lower percentage of non-muscularized distal small arterioles than did WT controls (Figure 1B). Thus, by three parameters, Enpp2^+/− mice showed an exaggerated pulmonary vascular response to hypoxia consistent with an augmentation in hypoxia-induced pulmonary hypertension.

Genetic inactivation of LPA₁ and LPA₂ promotes hypoxia-induced pulmonary remodeling in mice

The exaggerated response of the pulmonary vasculature of Enpp2^+/− mice to hypoxia suggests that LPA signaling may be protective in this setting. We therefore predicted that inactivation of the relevant LPA receptors would also result in exaggerated remodeling responses to hypoxia. LPA₁ and LPA₂ receptors are reported to be abundantly expressed in cultured endothelial and vascular SMC ^{22, 23}. We observed relatively high expression in murine blood vessels, making these receptors candidate to mediate the pulmonary vascular remodeling responses. WT and Lpar1^−/−2^−/− mice (8 – 14 weeks old) were subjected to hypoxia for 3 weeks. Exposure of Enpp2+/−, which are on the FVB background, to hypoxia did not result in mortality. However, the Lpar1^−/−2^−/− mice are on the BalbC background, and ~10% of the WT mice and ~55% of the Lpar1^−/−2^−/− mice died during the 3-week exposure to hypoxia. The surviving Lpar1^−/−2^−/− mice displayed evidence of hemodynamic compromise (HR <300 bpm) and, therefore, we were unable to obtain accurate RVSP measurements following exposure to hypoxia. Because sustained elevation in RVSP produces RV hypertrophy, we analyzed heart cross-sections to detect RV hypertrophy as an indirect marker for RVSP. In comparison to WT mice, Lpar1^−/−2^−/− mice developed more pronounced RV hypertrophy after hypoxia (Figure 2A and 2B). Similar to Enpp2^+/− mice, the Lpar1^−/−2^−/− mice had an approximately 2-fold greater reduction in peak flow velocity across the pulmonary valve after hypoxia than their WT controls (Table 2). Analysis of the lungs revealed significantly more muscularized and less non-muscularized pulmonary arterioles in Lpar1^−/−2^−/− mice than in age-matched WT controls (Figure 2C). Likewise, Lpar1^−/−2^−/− mice had thicker pulmonary arteriolar walls, suggesting positive arteriolar remodeling (Figure 2D). Together, these results indicate that mice lacking LPA signaling through LPA₁ and LPA₂ receptors were hyper-responsive to hypoxia-induced pulmonary remodeling and RV hypertrophy, providing further support for a protective role of LPA signaling in the pulmonary vasculature and indicating that these effects are mediated by LPA₁ and LPA₂.

Figure 2. Response of LPA₁- and LPA₂- double deficient mice to hypoxia.

A. Representative cross-section images of hearts from normoxic and hypoxic WT and Lpar1^−/−2^−/− mice. Hearts were sectioned at the widest point transversely, and stained with H&E staining. B. Quantification of RV free wall thickness to cross-section diameter ratio in normoxic and hypoxic WT and Lpar1^−/−2^−/− mice (n=5 per group). C. Percent muscularization of distal pulmonary arterioles in normoxic and hypoxic WT and Lpar1^−/−2^−/− mice. Lung sections were immunostained with α-smooth muscle actin and scored as described in Materials and Methods. Non = non muscularized vessels, partial= partially muscularized vessels, and full= fully muscularized vessels. Data are presented as mean ± S.D from 4 mice/group and were analyzed by 2-way ANOVA. *P<0.05. D. Pulmonary arteriolar wall thickness in WT and Lpar1^−/−2^−/− mice under normoxic and hypoxic conditions. Data are presented as mean ± S.D from four mice/group. *P<0.05.

Table 2.

Peak flow velocity across the pulmonary valve before and after 3 weeks of exposure to hypoxia.

Genotype	N	Baseline	After 3 weeks
WT	6	633.99 ± 90.22	552 ± 154
Lpar1−/−2−/−	5	607.66 ± 90.78	466 ± 149

^*Results are presented as mean ± SD in mm/sec.

Mice lacking LPA₁ and LPA₂ develop pulmonary hypertension with age

Under normoxic conditions, the pulmonary arterioles in Lpar1^−/−2^−/− mice were more muscularized and thicker than those in WT control mice (Figure 2 C and D). Changes in elastin content often accompany vascular remodeling. Elastin gene expression (Eln) in lungs from Lpar1^−/−2^−/− mice was 1.9 ± 0.37 fold higher than in WT controls, and quantification of elastin staining on lung sections confirmed an increase in perivascular elastin content in Lpar1^−/−2^−/− mice (Figure 3A–C). Together, these observations indicate alterations in the pulmonary vasculature occur in Lpar1^−/−2^−/− mice even in the absence of hypoxia. To determine if these alterations promoted the subsequent development of pulmonary arterial hypertension, we measured RVSP in young and old Lpar1^−/−2^−/− mice. Although no significant difference in RVSP was observed in WT and Lpar1^−/−2^−/− mice at age 8 – 14 weeks (Figure 4A), Lpar1^−/−2^−/− mice developed an age-dependent increase in RVSP (58.5 ± 4.5 versus 28.6 ± 8.0 mmHg in WT mice; P<0.05) that did not occur in age-matched Lpar1^−/− or Lpar2^−/− mice (Figure 4B). In agreement with the age-dependent elevation in RVSP observed in Lpar1^−/−2^−/− mice, peak flow velocity across the pulmonary valve was 11% lower in Lpar1^−/−2^−/− (n = 18) than in WT (n= 22; P<0.05) mice at 2 – 3 months of age and declined further in Lpar1^−/−2^−/− mice with age, such that by 6 – 9 months of age, the peak flow velocity was 18% lower in the Lpar1^−/−2^−/− mice (n = 14) than WT controls (n=11; P<0.001) (Table 3). In keeping with the elevated RVSP, histological sections of hearts revealed thicker RV free walls and enlarged RV chambers in older Lpar1^−/−2^−/− mice (Figure 4C), which was not observed in Lpar1^−/− or Lpar2^−/− mice (Figure 4D). No difference in the number of cardiomyocyte nuclei in the RV was observed in heart cross-sections. The average cross-section area of individual cardiomyocytes in the Lpar1^−/−2^−/− RV tended to be larger than in WT mice, but did not reach statistical significance (P=0.057) (Supplementary Figure 1). We previously reported that systemic blood pressure was normal in Lpar1^−/−2^−/− mice ¹. No difference in LV cardiomyocytes size, LV wall thickness, or LV ejection fraction was observed in Lpar1^−/−2^−/− mice, suggesting the alterations were RV-specific (Supplementary Figure I and data not shown). No histological evidence of cardiac fibrosis or inflammation was observed in the Lpar1^−/−2^−/− cardiac tissue. Taken together, the physiological and histological changes indicate that Lpar1^−/−2^−/− mice develop pulmonary hypertension and possibly RV hypertrophy spontaneously with age.

Figure 3. Perivascular elastin deposition in WT and Lpar1^−/− 2^−/−lungs.

A. Representative images of lung sections of 12 weeks old WT and Lpar1^−/−2^−/− mice. Sections were stained with elastin staining. Vessels are indicated by arrowheads. B. Quantitative RT-PCR analysis of elastin expression level in lungs of 12 week old WT (n=3) and Lpar1^−/−2^−/− (n=3) mice. The expression level of WT is set as 1 and data are presented as mean ± S.D. C. Quantification of perivascular elastin content in 12-week old Lpar1^−/−2^−/− (n=4) mice and age-matched WT (n=4) mice.

Figure 4. Elevated right ventricular systolic pressures (RVSP) and RV hypertrophy in olderLpar1^−/−2^−/− mice.

A. RVSP in 8–14 week old WT (n=6) and Lpar1^−/−2^−/− (n=4) mice. B. RVSP in aged (> 45 weeks old) WT (N=6), Lpar1^−/− (n=5), Lpar2^−/− (N=3), and Lpar1^−/−2^−/− (n=4) mice. Individual values (dots) and medium with 25 and 75 confidence intervals (box plots) are presented. Data were analyzed by 2-way ANOVA. NS: not statistically significant. C. Representative images of cross-sections of hearts of age-matched WT and Lpar1^−/−2^−/− mice (> 45 weeks old). Hearts were sectioned at the widest point transversely and stained with H&E staining. D. Quantification of RV free wall thickness to cross-section diameter ratio of aged (> 45 weeks old) WT (N = 6), Lpar1^−/− (n=4), Lpar2^−/− (n=2), and Lpar1^−/−2^−/− (n=6) mice. Data were analyzed by one-way ANOVA.

Table 3.

Peak flow velocity across the pulmonary valve of different age groups of Lpar1^−/−2^−/− mice.

Genotype	Age (months)	n	PV velocity
WT	2 – 3	22	737 ± 76
WT	6 – 9	11	712 ± 100
Lpar1−/−2−/−	2 – 3	18	667 ± 152
Lpar1−/−2−/−	6 – 9	14	584 ± 48

^*Results are presented as mean ± SD in mm/sec.

We excluded hypoxia as a cause for pulmonary arterial hypertension in Lpar1^−/−2^−/− mice. Arterial oxygen saturation in WT mice was 96 ± 1% and 97 ± 1% at 1 – 2 months and 6 months of age, respectively. Corresponding values in the Lpa1^−/−2^−/− mice were 95 ± 2%, 97 ± 0.3% and 96 ± 2% at 1 – 2 months, 3 – 4 months and >6 months of age, respectively. Additionally, hemoglobin (17.6 ± 1.3 g/dl) and hematocrit values (53.3 ± 4.4 %) in Lpar1^−/−2^−/− mice did not differ from age-matched WT (17.4 ±1.6 g/dl and 55.3 ± 7.9 %), indicating that the Lpar1^−/−2^−/− mice did not develop secondary polycythemia, which would be expected to occur in the setting of prolonged hypoxia (Supplementary Figure II). Because alterations in endothelial permeability have been reported to contribute to the development of pulmonary hypertension, we examined pulmonary vascular permeability in Lpar1^−/−2^−/− mice. Following intravascular administration, Evans Blue dye accumulation in the lung tissue of WT and Lpar1^−/−2^−/− mice was similar after normalization to wet lung weight (Supplementary Figure IIIA). In accordance with Evans Blue data, no histological evidence of edema was seen in the lung tissue of Lpar1^−/−2^−/− mice, and wet lung weight/leg bone length was the same in Lpar1^−/−2^−/− and WT mice. Examination of pulmonary valves did not reveal any notable abnormalities of Lpar1^−/−2^−/− mice, excluding pulmonary stenosis as a cause for the elevated RVSP (Supplementary Figure IIIB). Plastic casting and microCT analysis of the pulmonary vasculature also did not reveal obvious obstruction in the structure of the main pulmonary artery (data not shown).

In light of the established role of LPA as a stimulus for arterial vascular remodeling and in provoking acute increases in systemic blood pressure in WT and Lpar1^−/−2^−/− mice ^{1, 16}, the heightened pulmonary vascular remodeling in mice with attenuated LPA signaling suggested that perhaps LPA has different effects in the systemic and pulmonary vasculature. Therefore, we investigated acute LPA responses in an isolated pulmonary-cell and organ models. Analogous to observations in aortic SMCs ¹, LPA stimulated migration of cultured human pulmonary arterial SMCs by 11.6 fold and reduced by 34% expression of SMC-specific myosin heavy chain, a SMC differentiation marker. Infusion of LPA (up to 10 µM) into an isolated buffer-perfused and ventilated mouse lung system did not alter pulmonary artery pressure, under conditions where thrombin (1 U/ml) increased pulmonary artery pressure on average by 130% (from a mean of 5.4 ± 1.2 to 7.1 ± 2 mm Hg). Together, these results suggest that the phenotype observed in the Enpp2^+/− and Lpar1^−/−2^−/− mice may be due to an indirect effect of loss of LPA signaling on another mediator(s).

Alterations in endothelin signaling in Enpp2^+/− and Lpar1^−/−2^−/− mice

To understand the basis for the enhanced response in the pulmonary vasculature of mice with attenuated LPA production, we examined expression of mediators implicated in regulating pulmonary vasoconstriction and vasodilation in the Enpp2^+/− mice. Vegfr2, Pde5a, and Nos3 expression were normal in Enpp2^+/− lungs. Exposure to hypoxia produced a 30 – 40% increase in Vegfr2 and Pde5a expression and a decline of 50% in Nos3 expression in WT mice. Similar changes occurred in Enpp2^+/− mice. Exposure to hypoxia increased expression of Edn1 (encoding for endothelin-1, the most potent vasoconstrictor in the pulmonary circulation) in WT and Enpp2^+/− by 50%. Hypoxia had no effect on expression of Ednra (ET_A receptor), but reduced lung expression of Ednrb (ET_B receptor) by 61 ± 11% in WT mice (P<0.05). Interestingly, Ednrb expression was 30 ± 1% lower in Enpp2^+/− than WT mice at baseline (P<0.05) and declined further with hypoxia (P<0.05). On the other hand, the expression levels of Edn1 and Ednra were similar in WT and Enpp2^+/− mice at baseline or with hypoxia (Supplementary Figure IV). ET_B is a vasodilatory and clearing receptor for endothelin. Reduced levels of ET_B have been associated with acceleration of pulmonary hypertension ³³. Thus, an alteration in ET_B expression is a possible explanation for the enhanced remodeling response of Enpp2^+/− mice. Hypoxia treatment did not have significant effect on the expression levels of Lpar 1–5 in Enpp2^+/− mice (data not shown).

Next, we examined expression profiles of mediators known to influence development and hypoxia responses in LPA₁ and LPA₂ deficient mice. In keeping with the observations in Enpp2^+/− mice, no differences in Vegfr2, Pde5a, and Nos3 expression were observed between WT and Lpar1^−/−2^−/− mice exposed to normoxic conditions, and both responded with a similar increase in Vegfr2 and Pde5a expression and decline in Nos3 expression following hypoxia (data not shown). However, the expression levels of Edn1, which encodes for endothelin-1 (ET-1) and Ednra, which encodes the vasoconstrictive ET_A receptor, were significantly higher in Lpar1^−/−2^−/− mice than in WT controls (Figure 5A). As in the Enpp2^+/− mice, Ednrb expression was lower in Lpar1^−/−2^−/− mice (Figure 5A). The elevation in expression of Edn1 and Ednra and the reduced expression of Ednrb were evident as early as postnatal day 3 and persisted at days 7, 21 and adulthood. In agreement with mRNA levels, ET-1 peptide content in the lung tissue was also higher in Lpar1^−/−2^−/− mice at postnatal day 7 (Figure 5B), and protein levels of ET_B receptor were lower and ET_A were higher (Figure 5C and D). Given the pronounced vasoconstrictive effects of the endothelin pathway, the changes observed in Lpar1^−/−2^−/− mice could account for their propensity to develop pulmonary vascular remodeling and hypertension.

Figure 5. Expression of ET-1 and its receptors in lungs of Lpar1^−/−2^−/− mice.

A. Quantitative RT-PCR analysis of expression of Edn1(encoding ET-1), Ednra (ET_A) and Ednrb (ET_B) in lungs of age-matched WT (slashed bars) and Lpar1^−/−2^−/− (open bars) mice. All results were graphed from three experiments and presented as mean ± S.D. The expression level of WT mice is set as 1. *P<0.05. B. ET-1 protein content in lungs of age-matched WT (n=5; slashed bars) and Lpar1^−/−2^−/− (n=10; open bars) mice. C. ET_B and D. ET_A protein levels quantitated by immunoblot analysis and normalized to β actin expression (n = 3 per group). Results are presented as mean ± S.D.

Discussion

By studying mice with loss of function mutations in genes required for LPA production and signaling, we have identified an unanticipated role for this mediator in protection from hypoxia-induced pulmonary hypertension. We found that heterozygous inactivation of the LPA-producing enzyme autotaxin promotes the development of hypoxia-induced pulmonary arterial hypertension. Mice lacking LPA₁ and LPA₂ receptors were also more susceptible to hypoxia-induced pulmonary remodeling and spontaneously develop pulmonary arterial hypertension, characterized by elevated RVSP, increased pulmonary arteriolar muscularization and wall thickness, as well as RV hypertrophy with age. Our findings implicate both LPA₁ and LPA₂ as the main contributors to LPA’s effect in the pulmonary vasculature. Because neither Lpar1^−/− nor Lpar2^−/− mice display the dramatic development of pulmonary hypertension observed in Lpar1^−/−2^−/− mice, we conclude that these receptors act redundantly to promote the relevant signaling pathways.

Our findings raise the possibility that LPA has distinct effects in the pulmonary and systemic vasculature. LPA has been widely reported to stimulate cultured SMC migration and proliferation ^12–14; direct LPA administration stimulates neointimal formation in rodent carotid arteries ¹⁶; and systemic LPA administration elevates blood pressure acutely in rodents ¹. Based on its ability to promote migration and dedifferentiation of pulmonary vascular cells, LPA might be predicted to promote or exacerbate pulmonary hypertension. LPA₁ and LPA₂ receptors regulate migratory responses in several cell types, including SMC ^{1, 34}. It is possible that, in the setting of attenuated LPA signaling, abnormal SMC migration results in a developmental abnormality of vessel formation that promotes pulmonary hypertension. An alternative, but not exclusive possibility, is that LPA signaling plays a fundamental role in maintaining pulmonary vascular tone, and mice with inherited deficiencies in LPA signaling develop compensatory mechanisms to balance the loss of LPA, such as upregulation of endothelin signaling, that contribute to pathology. The ability of LPA to affect acutely systemic blood pressure in certain species was reported over 30 years ago. However, to our knowledge, this is the first report to document the consequences of long-term deficiency of LPA signaling on vascular tone in an experimental model. Of interest is the observation that acute LPA administration alters systemic blood pressure but does not affect pulmonary resistance, whereas genetic deficiency in LPA signaling alters the pulmonary vasculature without effecting systemic blood pressure. It is possible that in the absence of LPA signaling the endothelin pathway is upregulated to maintain systemic pressure at the expense of the pulmonary vasculature. LPA and endothelin share downstream signaling mediators, such as PLC and PLD, and other work has identified unexpected consequences of genetic deletion of signaling pathways on other systems. Similar effects, for example mediated by alterations in coupling of receptors to GPCRs, could account for upregulation of endothelin signaling in the absence of LPA receptors.

The development of pulmonary hypertension may require several events, including a trigger (such as hypoxia), an imbalance in the ratio of vasodilators to vasoconstrictors, and a permissive genotype ³⁵. The spontaneous development of pulmonary hypertension in the Lpar1^−/−2^−/− mice may have been facilitated by permissive genetic modifiers on the BalbC background. BalbC mice are more prone than C3H and C57BL/6 mice to developing pulmonary vascular muscularization following smoke exposure ³⁶. Enhanced hypoxic pulmonary vasoconstriction in BalbC mice compared to C57BL/6 mice has been observed ³⁷. The BalbC background also modifies the phenotype of TGFβ1-deficient mice ³⁸, and TGFβ signaling has been implicated in the development of pulmonary hypertension. Differences of baseline cardiovascular phenotypes in mice on various genetic backgrounds are reported and may also explain the differences in peak flow velocity of normoxic Enpp2 and Lpar1^−/−2^−/− mice ³⁹.

The development of pulmonary hypertension in the Enpp2^+/− and the Lpar1^−/−2^−/− mice was likely promoted by an imbalance in the ratio of vasodilators (nitric oxide and prostacyclins) to vasoconstrictors (ET-1 and thromboxane A₂). No difference in expression of enzymes responsible for production of NO (eNOS) or termination of its signaling (PDE-5) was observed, nor were there changes in urinary prostacyclin or thromboxane metabolites to account for the development of pulmonary hypertension (data not shown). However, in both Enpp2^+/− and the Lpar1^−/−2^−/− mice, lower expression of Ednrb could promote pulmonary vasoconstriction. The elevation in ET-1 levels, which occurred in Lpar1^−/−2^−/− mice by postnatal day 7, may also contribute to the propensity of these animals to develop pulmonary hypertension.

Our findings suggest that there may be significant crosstalk between LPA and endothelin signaling systems that occurs directly or functionally. LPA and ET-1 stimulate many of the same vasoconstrictive and phenotypic changes in cultured SMC ^{40, 41}. In rat aortic endothelial cells, LPA stimulates preproET-1 mRNA levels ⁴² and, in non-vascular SMC, some of the effects of ET-1 may be related to production of LPA and signaling through LPA₁ ⁴³. Both LPA and endothelin are shown to increase blood pressure acutely ^{1, 44}. Thus, in the settings of long-term deficiency of LPA signaling, such as the cases of Enpp2^+/− and the Lpar1^−/−2^−/− mice, compensatory changes in endothelin signaling pathways may occur that promote pathology. Precedent for this exists; mice that are deficient in cyclooxygenase 2 have an upregulation in ET_A receptors and develop exacerbated hypoxia-induced pulmonary hypertension ⁴⁵. Experiments are underway to elucidate the mechanism(s) of downregulation of Ednrb and hence enhanced endothelin signaling in the deficiency of LPA.

In conclusion, by studying mouse models with reduced LPA levels, as well as impaired LPA signaling, our data identify important roles for LPA₁ and LPA₂ receptors in pulmonary development and in the maintenance of normal pulmonary vascular pressure. These results focus attention on the possibility that components of the LPA signaling system could be targets for pharmacological treatment of pulmonary hypertension.