Pharmacological activation of peroxisome proliferator-activated receptor γ (PPAR-γ) protects against hypoxia-associated fetal growth restriction

Abstract

The hypoxia of high-altitude (HA) residence increases the risk of intrauterine growth restriction (IUGR) and preeclampsia 3-fold, augmenting perinatal morbidity and mortality and the risk for childhood and adult disease. Currently, no effective therapies exist to prevent these vascular disorders of pregnancy. The peroxisome proliferator-activated receptor γ (PPAR-γ) is an important regulator of uteroplacental vascular development and function and has been implicated in the pathogenesis of IUGR and preeclampsia. Here, we used a model of HA pregnancy in mice to determine whether hypoxia-induced fetal growth restriction reduces placental PPAR-γ protein expression and placental vascularization and, if so, to evaluate the effectiveness of the selective PPAR-γ agonist pioglitazone (PIO) for preventing hypoxia-induced IUGR. Hypoxia resulted in asymmetric IUGR, placental insufficiency, and reduced placental PPAR-γ expression; PIO prevented approximately half of the fetal growth restriction and attenuated placental insufficiency. PIO did not affect fetal growth under normoxia. Although PIO was beneficial for fetal growth, PIO treatment reduced placental vascular density of the labrynthine zone in normoxic and hypoxic (Hx) conditions, and mean vascular area was reduced in the Hx group. Our results suggest that pharmacological PPAR-γ activation is a potential strategy for preventing IUGR in pregnancies complicated by hypoxia, although further studies are needed to identify its likely metabolic or vascular mechanisms.—Lane, S. L., Dodson, R. B., Doyle, A. S., Park, H., Rathi, H., Matarrazo, C. J., Moore, L. G., Lorca, R. A., Wolfson, G. H., Julian, C. G. Pharmacological activation of peroxisome proliferator-activated receptor γ (PPAR-γ) protects against hypoxia-associated fetal growth restriction.

Intrauterine growth restriction (IUGR) increases the risk of stillbirth 4-fold and neonatal death 8–20-fold (1). The adverse impact of IUGR extends across the lifespan, increasing the likelihood of disease in childhood, as well as obesity, metabolic syndrome, and cardiovascular disease in later life (2). IUGR and preeclampsia, a hypertensive disorder of pregnancy that shares key pathophysiologic features with IUGR, often coexist. Together, these vascular disorders of pregnancy account for nearly 30% of premature deliveries (3, 4) with an annual cost in excess of $2.1 billion for short-term care in the United States alone (5). Despite its public health significance, the underlying pathophysiology of IUGR remains unclear, and no effective prevention or treatment strategies exist. Among the numerous determinants of IUGR is the chronic hypoxia of high-altitude (HA; ≥2500 m) residence. Specifically, HA reduces birth weight and increases the incidence of IUGR 3-fold, caused primarily by a leftward shift of the birth-weight distribution rather than shortened gestation (6). For this reason, investigations of HA pregnancy are uniquely poised to identify the physiological and molecular mechanisms by which hypoxia influences fetal growth.

Normal fetal development depends on the adequate delivery of oxygen and nutrient-rich blood to the uteroplacental and fetoplacental circulations (7). To meet increasing metabolic demands of the fetus across gestation, pregnancy prompts extensive adaptations of the maternal cardiovascular system, lowering uteroplacental vascular resistance and permitting a >10-fold rise in blood flow throughout gestation (8). Placental factors also contribute to a reduction in uteroplacental vascular resistance. Specifically, during early human gestation, extravillous trophoblasts (invasive placental cells derived from the cytotrophoblast) migrate through the decidua to initiate extensive vascular remodeling of the uterine spiral arteries; this process is critical for establishing placental perfusion. Placental hypoperfusion, a common pathophysiologic feature of IUGR, has largely been attributed to impaired trophoblast differentiation, incomplete spiral and myometrial artery remodeling, reduced fetoplacental neovascularization (9), minimal branching of the villi capillaries (10), and reduced uterine artery blood flow (11). We propose that the peroxisome proliferator-activated receptor γ (PPAR-γ) plays a central role for the effect of hypoxia to reduce fetal growth.

PPAR-γ is a ligand-activated transcription factor of the nuclear hormone receptor superfamily that governs gene expression by forming a heterodimeric complex with the retinoid X receptor at the regulatory region in target genes. Primary endogenous PPAR-γ ligands include essential fatty acids (e.g., linoleic acid), prostaglandins, and oxidized fatty acids (e.g., 13-hydroxyoctadecadienoic acid) (12). Although predominantly recognized for its role in metabolic homeostasis, PPAR-γ is vital for establishing the uteroplacental villous circulation (13), modulating vasoreactivity by suppressing the synthesis of the potent vasoconstrictor endothelin-1 (ET-1) (14) and increasing NO production (15), and regulating placental vascularization (16). Consistent with these functions, PPAR-γ is abundantly expressed in the labyrinthine zone of the rodent placenta (17) and porcine placental vascular endothelial cells (18), as well as human trophoblasts (13) and vascular smooth muscle cells (19).

Abundant evidence implicates PPAR-γ in the development of IUGR and preeclampsia. In humans, for instance, placental PPAR-γ mRNA expression is 2 times lower in IUGR cases compared with controls and is directly related to fetal weight (20). In recent work, we found that maternal HA (3600–4100 m) residence during pregnancy down-regulates the expression of key PPAR-γ pathway genes in maternal peripheral blood cells and that this effect occurred in association with reduced birth weight and head circumference (21). Here, we used an experimental murine model to: 1) determine whether hypobaric hypoxia reduces placental PPAR-γ protein expression and diminishes placental vascularization in association with lower fetal weight; and 2) determine whether dietary supplementation with a selective synthetic PPAR-γ agonist [pioglitazone (PIO)] during the latter portion of pregnancy protects against hypoxia-associated IUGR and normalizes placental vascularization. Emphasizing the translational potential of such studies, renewed calls to evaluate the utility of PPAR-γ agonists for the treatment of vascular disorders of pregnancy have been issued (22). We considered that the present studies may therefore accelerate the development of interventions using existing (or next-generation) pharmacological compounds for the treatment or prevention of IUGR.

MATERIALS AND METHODS

Experimental animals, treatments, and exposures

All murine experiments were approved by the Institutional Animal Care and Use Committee of the University of Colorado Denver (Denver, CO, USA) and were consistent with the Guide for the Care and Use of Laboratory Animals (National Institutes of Health, Bethesda, MD, USA).

Nulliparous C57/BL6 females and males (The Jackson Laboratory, Bar Harbor, ME, USA) were paired overnight under standard conditions (12-h light/dark cycles at 23°C and 60% humidity) in the animal care facility at the University of Colorado Denver Anschutz Medical Campus (Aurora, CO, USA). Confirmation of mating was defined by the presence of a copulatory plug and considered as gestational day (GD) 0.5. Pregnant dams were randomly assigned to receive either a control diet (Control) (AIN-76A; TestDiet, St. Louis, MO, USA) or a PPAR-γ agonist diet (PIO) (AIN-76A with 0.05% actos containing 0.0125% PIO HCl; TestDiet) under normoxic (Norm) or hypoxic (Hx) conditions, generating the following 4 study groups for comparison: Control-Norm, PIO-Norm, Control-Hx, and PIO-Hx. Activation of PPAR-γ nuclear receptors occurs via natural PPAR-γ ligands or pharmacologic compounds including thiazolidinediones (e.g., PIO, rosiglitazone, troglitazone). PIO was selected for this study, given that the compound is a high-affinity, selective agonist for PPAR-γ and is widely used for the treatment of type II diabetes (23).

For all animals, exposures and experimental diets began on GD 14.5. For the Norm and Hx groups, dams were placed into a normobaric chamber at a simulated altitude of 0 feet (P_B ∼760 mmHg) or a hypobaric chamber at a simulated altitude of ∼15,000 feet (P_B ∼460 mmHg), respectively. The PIO dose chosen was based on preparatory work in which the effect of 3 different PIO concentrations (0.05, 0.025, and 0.0125%) on fetal viability were compared. Norm dams receiving the 0.05 or 0.025% PIO diet had significantly reduced fetal numbers, whereas the 0.0125% PIO dose did not affect fetal number. All groups were fed ad libitum and had free access to water at all times. On GD 18.5, the dams were weighed and subsequently euthanized by CO₂ inhalation followed by cervical dislocation. Immediately postmortem, the gravid uterus and uterine vasculature were carefully removed and placed in a dissecting dish containing cold PBS (Thermo Fisher Scientific, Waltham, MA, USA) with protease and phosphatase inhibitors (MilliporeSigma, Burlington, MA, USA). The number of fetuses was noted, fetuses and their placentas were weighed separately without membranes and umbilical cords, and fetal biometry [crown-rump length (CRL) and biparietal diameter] was recorded.

Biological specimen collection

Within 10 min of euthanasia, whole placentas were washed in cold PBS and cut into 2 equal portions. One half of the placenta was fixed in 4% paraformaldehyde, embedded in paraffin, and sectioned to visualize the entire thickness of the placenta from the decidua to the chorionic plate in a single section. These sections were used to assess the effect of PIO and hypoxia on placental vascularization. The remainder of the placenta was snap frozen in liquid N₂ and stored at −80°C for Western blotting for semiquantitative measures of PPAR-γ protein expression. Given evidence that PIO increases the phosphorylation of AMPK in several tissues, including skeletal muscle (24), placental phosphorylated (p)-AMPK protein levels were also evaluated to determine whether PIO treatment affected the organ of interest.

Placental PPAR-γ and p-AMPK protein expression by Western blot

Placental tissue was mechanically homogenized in Meso Scale lysis buffer [150 mM NaCl, 20 mM Tris (pH 7.5), 1 mM EDTA, 1 mM EGTA, and 1% Triton X-100; Meso Scale Diagnostics, Rockville, MD, USA] containing a mixture of water-soluble protease inhibitors with broad specificity for the inhibition of serine, cysteine, aspartic, and metalloproteases (Halt Inhibitor Cocktail; Thermo Fisher Scientific) using a sonicator. Protein concentration was determined using the Pierce BCA Protein Assay Kit (Pierce, Rockford, IL, USA). PPAR-γ protein expression was analyzed in placental tissue homogenates using commercial antibodies (2535S; Cell Signaling Technology, Danvers, MA, USA) with β-actin (13E5; Cell Signaling Technology) as the loading control. Briefly, protein (15 μg/lane) was denatured, separated by SDS-PAGE on a 4–20% Mini-Protean TGX Precast Gel (Bio-Rad, Hercules, CA, USA), then transferred to the nitrocellulose membrane for 90 min at 30 V at 4°C. All block steps and incubations were performed in Li-Cor Odyssey TBS Blocking Buffer (Li-Cor Biosciences, Lincoln, NE, USA). Membranes were exposed to the primary antibody overnight at 4°C and to Li-Cor fluorescent secondary antibodies for 1 h at room temperature. Membranes were imaged and analyzed using Image Studio software (Li-Cor Biosciences, Lincoln, NE, USA) on a Li-Cor Odyssey CLx. Data are presented as fluorescent signals normalized to β-actin.

Assessment of placental vascularity

Formalin-fixed, paraffin-embedded placental tissue sections (5 μm) were used for hemotoxylin and eosin (H&E) and CD34 (vascular endothelial marker) staining and processed for histological staining and immunohistochemistry to assess placental vascularity. The CD34-stained slides were quantified for vascularity using the Matlab Image Processing Toolbox (MathWorks, Natick, MA, USA) as previously described in Cookson et al. (25). Fields of view were captured across the placenta at ×20 magnification. Briefly, image analysis measured the area of CD34 using brightfield colorimetric thresholding to identify vessels stained as positive for endothelial marker. The vascular-mask threshold identified individual vessel density and area. Vessel density was defined by the average number of vessels per unit area of tissue per image. Vessel area of an image was defined as the total area of vessels stained in an image in square microns per field of view.

Data analysis

Continuous variables were tested for normality using the Kolmogorov-Smirnov test. Two-way ANOVA was used to determine the effects of PIO treatment and Hx exposure on maternal weight, fetal biometry, fetal weight, and placental characteristics. All models included PIO treatment and exposure as main effects and an interaction term for PIO × exposure. Sidak’s test was used to identify between-group differences. Individual dam and fetal data are reported as an average within each group (e.g., Control-Norm). All analyses were conducted using GraphPad Prism (GraphPad Software, La Jolla, CA, USA). Data are expressed as the means ± the sd, and sample sizes for each comparison are provided in the table and figure legends. Statistical differences were considered significant when the 2-tailed value of P < 0.05 unless specified otherwise, and trends were reported when 0.05 < P < 0.10.

RESULTS

In this section, results are structured to compare the effect of Hx on our primary outcomes of interest for each treatment group (Control and PIO) and the effect of PIO treatment within each exposure group (Norm and Hx). Significant interaction effects (Hx × PIO) indicate that the effect of hypoxia on a given variable differs between Control- and PIO-treated animals.

Effect of Hx and PIO on maternal and fetal characteristics

Prior to the initiation of exposures and experimental diets (GD 8.5), dam weight was equivalent among all groups (Table 1). At the time of tissue harvest (GD 18.5), Control-Hx dams weighed 27% less than Control-Norm dams, whereas PIO-Hx and PIO-Norm dams were of similar weight (Table 1). PIO-Hx dams weighed 16% more than Control-Hx, indicating that the effect of Hx to reduce dam weight was blunted with PIO treatment (interaction effect: Hypoxia × PIO, P < 0.0001).

TABLE 1.

Effect of hypoxia and PIO on dam and fetal characteristics

Parameter	Control		PIO		P
	Norm	Hx	Norm	Hx	Hypoxia	PIO	Hypoxia × PIO
Dam weight, GD 8.5, g	23.7 ± 0.8	23.4 ± 0.5	24.0 ± 0.7	23.5 ± 0.7	NS	NS	NS
Dam weight, GD 18.5, g	42.9 ± 3.9a,b,c	31.3 ± 0.7b,c,d	37.0 ± 1.8a,d	37.4 ± 2.6a,d	<0.0001	NS	<0.0001
Offspring, #	9.0 ± 0.0c	8.0 ± 0.8	8.2 ± 1.3	6.5 ± 2.2d	<0.05	NS	NS
CRL, cm	1.7 ± 0.2a,b,c	1.5 ± 0.1b,c,d	1.9 ± 0.1a,c,d	1.8 ± 0.1a,b,d	<0.0001	<0.0001	<0.05
BPD, cm	0.55 ± 0.08	0.55 ± 0.06	0.51 ± 0.05c	0.56 ± 0.09b	<0.05	NS	<0.05

Data were analyzed by 2-way ANOVA with Sidak’s multiple comparisons to detect sources of difference. Sample sizes: dams (Control-Norm, n = 6; Control-Hx, n = 5; PIO-Norm, n = 6; and PIO-Hx, n = 10), and fetal characteristics (Control-Norm, n = 54; Control-Hx, n = 26; PIO-Norm, n = 43; and PIO-Hx, n = 63). BPD, biparietal diameter; NS, not significant. ^a–dValues are significantly different from Control-Norm (a), Control-Hx (b), PIO-Norm (c), and PIO-Hx (d).

Although Hx reduced fetal number overall, there was no statistically significant effect in Control or PIO animals considered separately (Table 1). Similarly, the number of resorptions visible at the time of harvest was unaffected by Hx or PIO (unpublished data).

In Controls, Hx reduced fetal weight 21% (P < 0.001) compared with a 9% reduction in dams treated with PIO (P < 0.05; Fig. 1A). As a result, PIO-Hx fetuses weighed 16% more than Control-Hx (P < 0.001). In contrast, fetal weight was similar between the Control-Norm and PIO-Norm groups. Hx was associated with reduced fetal CRL in Controls, whereas this was not the case in fetuses of dams treated with PIO. Specifically, compared with Control-Norm, fetal CRL was greater in PIO-Norm and PIO-Hx (Table 1). Hx had no effect on fetal biparietal diameter in Controls, whereas values were greater for PIO-Hx compared with PIO-Norm.

Figure 1.

Effect of Hx and PIO on fetal and placental weight. Fetal weight, placental weight, and placental-to-fetal weight ratios were compared between treatment groups (PIO and Control) and exposure groups (Norm and Hx) using 2-way ANOVA and Sidak’s multiple comparisons tests to identify sources of difference. Data are shown as individual values for fetuses. A) PIO protected against the effect of Hx to reduce fetal weight. In Controls, Hx reduced fetal weight 20.6% vs. a 9.3% reduction in the PIO group. Notably, whereas fetuses of dams treated with PIO weighed 16% more than Controls under Hx conditions, there was no difference in fetal weight between treatement groups under Norm conditions. B) Overall, placental weight was greater under Hx compared with Norm (main effect, Hx); considering PIO and Controls group separately, however, placental weight was unaffected by Hx. PIO reduced placental weight overall (main effect, PIO) and under Hx conditions. C) In Controls and dams treated with PIO, Hx increased placental insufficiency, as assessed by the ratio of placental-to-fetal weight. Notably, the magnitude by which Hx increased placental insufficiency was reduced with PIO treatment (interaction effect: Hx × PIO). Significance values for the main effects of Hx, PIO, and the interaction of Hx × PIO are shown within each plot. Sample sizes: Control-Norm, n = 54; Control-Hx, n = 26; PIO-Norm, n = 43; PIO-Hx, n = 63. Solid lines with asterisks indicate the following significant post hoc comparisons: *P < 0.05, ***P < 0.001, ****P < 0.0001.

Effect of Hx and PIO on placental weight and placental insufficiency

Placental weight was unaffected by Hx in Control or PIO dams (Fig. 1B). Under Hx conditions, placental weight was lower in PIO vs. Controls (P < 0.05, Fig. 1B), whereas there was no difference between and Control and PIO groups under Norm conditions. Hx increased placental insufficiency, as assessed by the ratio of placental-to-fetal weight, in Control vs. PIO dams (Fig. 1C). Moreover, PIO dampened the effect of Hx to increase placental insufficiency (interaction effect of Hx × PIO, P < 0.01; Fig. 1C).

Effect of Hx on placental PPAR-γ protein expression

In Controls, placental PPAR-γ protein expression was 46% lower under Hx conditions, as assessed by Western blot (P < 0.05; Fig. 2A). Placental PPAR-γ protein levels were not compared between PIO and Controls, given that PIO selectively stimulates the PPAR-γ nuclear receptor to modulate the transcription of downstream genes rather than affecting PPAR-γ itself. Given that thiazolidinediones, including PIO, activate AMPK, we assessed the placental p-AMPK protein in each group to determine whether PIO was acting on the placenta, the organ of interest. In Controls, placental p-AMPK protein levels trended lower under Hx conditions (Fig. 2B). Conversely, placental p-AMPK protein levels were elevated in PIO-Norm vs. Control-Norm (P < 0.05) and trended higher in Hx conditions as well (Fig. 2B), suggesting that PIO but not hypoxia activated AMPK.

Figure 2.

The effect of Hx on placental PPAR-γ expression and the effect of hypoxia and PIO on placental AMPK activation. Ratios of PPAR-γ protein expression to β-actin expression were compared between Control-Norm and Control-Hx using Student’s t test. Ratios of p-AMPK to β-actin expression were compared between treatment groups (PIO and Control) and exposure groups (Norm and Hx) using 1-way ANOVA and Sidak’s multiple comparisons test to identify sources of difference. A) In untreated animals, hypoxia reduced PPAR-γ protein expression in the placenta by half. B) In untreated animals, placental p-AMPK was 81% lower with hypoxia. In Norm animals, PIO treatment tripled placental p-AMPK. In Hx animals, p-AMPK trended higher with PIO treatment but remained lower than levels in Norm, untreated placenta. Sample sizes: Control-Norm, n = 5; Control-Hx, n = 6; PIO-Norm, n = 6; PIO-Hx, n = 6. * P < 0.05.

Effect of Hx and PIO treatment on placental vascularization

Hx had no effect on placental vascularization, as assessed by vessel density or mean vascular area per field of view, in Controls or PIO (Fig. 3A, B). In contrast with our expectations, however, placental vascular density and mean vascular area were lower in PIO-Hx compared with Control-Hx (Fig. 3A, B). Mean placental vessel size was unaffected by PIO treatment (P = NS, all comparisons). Representative histology images of H&E and CD34 staining of placental sections for each group are shown in Fig. 3C, with the labyrinthine and junctional zones distinguished. Qualitatively, placental vascularization appears markedly sparse and diffuse in PIO-Norm and PIO-Hx vs. Control-Norm and Control-Hx (Fig. 3C).

Figure 3.

The effect of hypoxia and PIO treatment on placental vascularization. Placental vascular density and mean vascular area were compared between groups [Control-Norm (n = 4), Control-Hx (n = 11), PIO-Norm (n = 4), and PIO-Hx (n = 14)] using 2-way ANOVA and Sidak’s multiple comparisons tests to identify sources of difference. A) Neither vessel density (the number of vessels per square micrometer) nor the mean vascular area per field of view was affected by hypoxia; this was true for Controls and PIO-treated animals. Placentas of PIO-treated animals showed reduced vessel density under Norm and Hx conditions as well as lower mean vascular area in hypoxia. Solid lines with asterisks indicate the following significant posthoc comparisons: *P < 0.05, **P < 0.01. B) Representative histology images of H&E and CD34 staining of placental sections for each exposure and treatment group, with the labrynthine (LZ) distinguished; qualitatively, placental vascularization of PIO-treated animals appears markedly sparse and diffuse compared with Controls under Norm and Hx conditions. Magnification values, ×5.

DISCUSSION

The PPAR-γ pathway plays a vital role for several processes required for successful pregnancy and has been implicated in the development of IUGR and preeclampsia. In this first report to explore the role of PPAR-γ in hypoxia-induced IUGR, we found that Hx exposure during late murine pregnancy reduces placental PPAR-γ protein expression in association with diminished fetal growth and evidence of placental insufficiency. Our findings further indicate that dietary supplementation with PIO, a selective PPAR-γ agonist, protects fetal growth under Hx conditions. Given evidence that PPAR-γ has potent angiogenic properties in the placenta, we hypothesized that pharmacological activation of PPAR-γ would increase placental vascularization. Contrary to our expectations, placental vascularization, particularly within the labyrinthine zone, decreased with PIO treatment under Norm and Hx conditions. Based on these findings, we propose that the mechanism by which PPAR-γ activation in late pregnancy acts to protect against hypoxia-induced fetal growth restriction is not the result of proangiogenic effects in the placenta, but rather may be caused by the metabolic properties of PPAR-γ or its effects on vascular function.

Our Hx model induced a 21% reduction of fetal weight, a magnitude of growth restriction similar to the 21–26% reduction seen with a near-term, 5-d normobaric hypoxia (10% O₂) in mice (26). Other reports, however, suggest that 12% O₂ exposure from GD 14.5 to 18.5 (27) or 13% O₂ for the entirety of gestation (28) induces a more modest restriction of fetal growth (7 and 12%, respectively). Such variability may be caused, in part, by the fact that the dams in our study spent early pregnancy at moderate altitude (Denver, Colorado; 1609 m). Consistent with others’ work in rodents (26), we found that hypoxia increased the ratio of placental-to-fetal weight, indicating greater placental insufficiency under Hx conditions.

In our model, hypoxia had no effect on placental vascularization. In murine models, the effect of late-gestation hypoxia on placental vascularity has been reported to decrease or increase (27, 29). In humans, greater placental vascularization has been reported at HAs (30). For instance, the intermediate and terminal villi of placentas obtained from women of Kirghiz ethnicity residing at HAs (2200–2800 m) tended to have a higher volume fraction of fetal capillaries and increased capillary length density compared with their lowland (760 m) counterparts (30). However, placentas obtained from uncomplicated pregnancies at HAs in Colorado (3100 m) have decreased capillary volume and unchanged capillary length density, total length, and diameter compared with sea level (31). Variability between studies with respect to placental vascularization may be caused, in part, by differences in population ancestry and the altitudes being compared.

Our finding that hypoxia suppressed placental PPAR-γ protein expression agrees with reports of lower PPAR-γ mRNA and protein expression in isolated murine and human cells cultured in low oxygen environments as well as evidence that PPAR-γ is regulated, in part, by the hypoxia-inducible factor 1α (HIF1α). Specifically, PPAR-γ mRNA and protein expression are attenuated in murine trophoblasts incubated under 2% O₂ (32). Hypoxia also suppresses PPAR-γ mRNA and protein expression in human pulmonary vascular smooth muscle cells. Further, we have shown that HA residence (3600 m) down-regulates PPAR-γ–pathway gene expression in maternal peripheral blood mononuclear cells during pregnancy (21). Finally, a recent study identified a highly conserved Hx response element located 1042 base pairs upstream of the PPAR-γ transcriptional start site (33), indicating that PPAR-γ is a direct HIF1α target. However, the relationship between HIF1α and PPAR-γ appears to be complex and tissue specific. Activation of HIF1α inhibits PPAR-γ gene transcription during adipocyte differentiation (34), whereas in cardiomyocytes, HIF1α activation increases PPAR-γ expression (33). In murine trophoblast stem cells, however, hypoxia has been reported to repress PPAR-γ expression in an HIF1α-independent fashion (32). Further studies are needed to determine whether the effect of impaired maternal oxygenation on reducing placental PPAR-γ expression that was observed in our study occurred via HIF1α-independent mechanisms.

Abundant human and experimental evidence supports the link between reduced placental PPAR-γ protein expression and a hypoxia-induced reduction of fetal weight. In particular, human placental PPAR-γ gene and protein expression is lower in IUGR cases vs. controls and is directly related to fetal weight (20). Further, although IUGR and preeclampsia are distinct entities, given that the 2 conditions share pathophysiological features and often occur in tandem, it is notable that a dominant loss-of-function mutation in the ligand-binding domain of PPAR-γ is associated with the development of severe preeclampsia (35), and reduced PPAR-γ ligand levels can be detected in the maternal circulation prior to preeclampsia diagnosis (36). Strengthening these associations, PPAR-γ agonist (troglitazone) treatment abolished hypertensive responses to angiotensin II infusion during pregnancy and normalized fetal growth in a genetic mouse model [regulator of G-protein signaling 5 (Rgs5) ^+/−] known to be susceptible to hypertension and IUGR (37). Conversely, antagonizing PPAR-γ via dietary supplementation with GW9662 or intraperitoneal delivery of T0070907 decreased fetal growth in rats (38, 39). Given recent calls for evaluating the utility of PPAR-γ agonists for the treatment of vascular disorders of pregnancy, including IUGR and preeclampsia, our observation that pharmacological PPAR-γ activation partially restores fetal growth under Hx conditions is compelling.

We considered that pharmacological activation of PPAR-γ may protect fetal growth by increasing placental vascularization. Our rationale was 2-fold. First, in humans, intermediate villi branching is minimal in IUGR placentas with absent or reversed end-diastolic velocity, and the terminal capillaries are thin and elongated (10, 40). Further supporting the importance of sparse placental vascularization for IUGR, fetoplacental neovascularization is impaired in a transgenic murine model of preeclampsia accompanied by fetal growth restriction (9). Second, evidence suggests that pharmacological activation of PPAR-γ increases placental vascularization. Specifically, in human placental explants, rosiglitazone increases placental villous vascularization, increasing capillary length and vessel size (16). In isolated porcine placental vascular endothelial cells, rosiglitazone not only promotes cellular proliferation and migratory capabilities but also increases the mRNA expression of VEGF and its receptor (18). Our study indicates that PPAR-γ activation in late murine pregnancy reduces vascularization of the labyrinthine zone, irrespective of ambient oxygen tension. One possible explanation for our observation is the variable effect of PPAR-γ activation on vascularization across gestation. In experimental animal models, for example, daily administration of rosiglitazone from GD 4.5 to 16.5 reduced placental vascularization and the expression of proangiogenic factors (18). For this reason, our study was designed to initiate the pharmacological intervention during latter pregnancy at a time that is roughly equivalent to the second and third trimester of human pregnancy, when IUGR most frequently becomes apparent. In further support of our design, although placental growth is limited during the last half of gestation, increasing fetal metabolic demands during this time frame are met, in part, by enhanced placental function and dramatic branching growth of the fetoplacental capillary bed (41). Taken together, however, our results do not support the hypothesis that the effect of PIO for preserving fetal growth is mediated by increased placental vascularization.

An alternative mechanism by which pharmacological activation of PPAR-γ may protect fetal growth relates to its role in the regulation of vascular function and growth. Normal pregnancy initiates extensive maternal cardiovascular responses that collectively serve to sustain sufficient uteroplacental perfusion to support fetal development. Uterine artery diameter and blood flow, for instance, rise profoundly across gestation, caused, in part, by changes in the production or activity of vasodilators [e.g., NO and large‐conductance Ca²⁺-activated K⁺ potassium (BK_Ca) channels] and vasoconstrictors (e.g., ET-1) (8). HA reduces the pregnancy-associated rise in uteroplacental blood flow, an effect that is directly associated with slowed fetal growth (42, 43). In experimental animals, hypoxia also reduces uterine vasodilator responses, exaggerates myogenic tone, and reduces NO-induced uterine artery vasodilation (44, 45). Suggesting a vasoprotective role for PPAR-γ, activation of the PPAR-γ pathway suppresses ET-1 synthesis in human vascular endothelial cells (14). PPAR-γ ligands also increase NO production via PPAR-γ–dependent mechanisms (15). Further indicating vascular sites of action, PPAR-γ regulates myogenic tone of the mesenteric artery via Rgs5-mediated control of PKC and BK_Ca channel activity (46). In a rodent model of preeclampsia (reduced uterine perfusion pressure), administration of rosiglitazone from GD 16 to 18 improved mesenteric artery vasorelaxation (47). Conversely, antagonizing PPAR-γ during pregnancy impairs vasodilation of radial uterine arteries and induces a constellation of preeclampsia-like symptoms including hypertension and endothelial dysfunction (38, 39). Thus, further study is warranted to determine the effect of PPAR-γ activation on the vasoreactivity of uterine vessels responsible for uteroplacental perfusion (Fig. 4).

Figure 4.

Schematic of current research findings and alternative hypotheses. The present study demonstrates that hypoxia (Hx) reduces fetal weight and increases placental insufficiency without affecting placental weight or vascularization. Whether such effects are mediated through HIF1α or PPAR-γ are unclear; however, PIO partially protects against hypoxia-induced fetal growth restriction and placental insufficiency without affecting placental vascularization. Alternative mechanisms by which pharmacological activation of PPAR-γ may preserve fetal growth include the pleiotrophic actions of PPAR-γ to: 1) regulate myogenic tone via Rgs5-mediated control of PKC and BK_Ca channel activity; 2) increase vascular growth via increasing the expression of VEGF and its receptor (VEGFR2); or 3) decrease ET-1 synthesis. Black text indicates current observations and gray text represents alternative hypotheses.

Emphasizing the translational potential of our research, PPAR-γ agonists are widely used for the treatment of type II diabetes and polycystic ovarian syndrome (48, 49) and are the focus of clinical trials to reduce the risk of stroke and myocardial infarction (50). To maximize the therapeutic potential of PPAR-γ agonists for vascular disorders of pregnancy, continued efforts are needed to: 1) clarify the physiological role of PPAR-γ activation during normal pregnancy; 2) identify the molecular pathways that govern PPAR-γ activation and its downstream transcriptional targets in reproductive tissues; and 3) determine whether the effects of pharmacological PPAR-γ agonists are caused by the activation of PPAR-γ itself or off-target effects. Since our results suggest a lack of improved placental vascularization with late-gestation PPAR-γ activation, ongoing studies will evaluate whether PPAR-γ activation improves vasoregulation and fetal growth. Other potential avenues for future research include experimental animal and human studies to determine whether PPAR-γ–mediated regulation of glucose homeostasis (51), oxidative stress (52), and nutrient transporters (53) also influence fetal growth at HA.

Glossary

BK_Ca

large-conductance Ca²⁺-activated K⁺ potassium

CRL

crown-rump length

ET-1

endothelin-1

GD

gestational day

H&E

hemotoxylin and eosin

HA

high altitude

HIF1α

hypoxia-inducible factor 1α

Hx

hypoxic

IUGR

intrauterine growth restriction

Norm

normoxic

PIO

pioglitazone

PPAR-γ

peroxisome proliferator-activated receptor γ

Rgs5

regulator of G-protein signaling 5

AUTHOR CONTRIBUTIONS

S. L. Lane, R. B. Dodson, and C. G. Julian analyzed data; S. L. Lane, A. S. Doyle, H. Park, H. Rathi, C. J. Matarrazo, G. H. Wolfson, and C. G. Julian performed research; S. L. Lane, L. G. Moore, R. A. Lorca, and C. G. Julian wrote or contributed to manuscript; and C. G. Julian designed research.