Deficiency of Cardiac Natriuretic Peptide Signaling Promotes Peripartum Cardiomyopathy-Like Remodeling in the Mouse Heart

Introduction

Editorial, see p 589

During pregnancy, delivery, and the postpartum period, maternal homeostasis is balanced by several endogenous hormones, including progesterone and estrogen, which are essential for maintaining pregnancy, and prolactin (PRL) and oxytocin from the anterior and posterior pituitary, respectively, which regulate lactation and contraction of the uterus after delivery. During gestation, circulatory blood volume is increased to maintain blood flow towards the infant and the developed uterus, and the maternal heart undergoes transient and reversible cardiac hypertrophy to compensate for the increased cardiac output.¹ In some cases, pregnancy and delivery are associated with pathologic cardiac events such as peripartum cardiomyopathy (PPCM), a high-mortality disease that causes acute heart failure in puerperants.²

Cardiac hypertrophy is classified into 2 types—physiologic and pathologic—according to hemodynamic stress, cardiac geometrics, and functional phenotypes.³ In physiologic cardiac hypertrophy induced by exercise, the PI3-kinase–Akt pathway is activated; the expression of fetal cardiac genes such as Nppa, Nppb, Myh7, and Acta1 is not upregulated; and cardiac function is preserved.⁴ In contrast, in pathologic cardiac hypertrophy, the expression of fetal cardiac genes is increased, and cardiac function is decreased.⁴ The molecular pathways underlying pathologic cardiac hypertrophy induced by pressure overload have been investigated extensively,^3,5 particularly the involvement of calcineurin and nuclear factor of activated T cells (NFAT).⁶ Calcineurin, a Ca²⁺-activated protein phosphatase, dephosphorylates NFAT in the cytoplasm and induces its nuclear translocation.⁷ The transient receptor potential cation channels (eg, TRPC3, TRPC4, and TRPC6) are involved in the calcineurin–NFAT signaling pathway.⁸ In addition, ERK1/2 (extracellular signaling-regulated kinase), CaMKII (Ca²⁺/calmodulin-dependent protein kinase II), and p38 mitogen-activated protein kinase signaling contribute to pathologic cardiac hypertrophy.^3,5

Two cardiac natriuretic peptides, atrial natriuretic peptide (ANP) and brain natriuretic peptide (BNP), are produced in and secreted from the heart. They control circulatory homeostasis through a common receptor, NPR1, encoded by the Npr1 gene.⁹ ANP and BNP induce diuresis, natriuresis, and dilation of blood vessels.¹⁰ Soon after delivery, sodium and water preserved during pregnancy are excreted. ANP is thought to be involved in this process, because the plasma ANP level in women increases significantly immediately after delivery.¹¹ Elevated plasma ANP and BNP levels in patients with pregnancy-induced hypertension and preeclampsia¹² imply that ANP and BNP are involved in these pregnancy-related diseases. The precise physiologic and pathophysiologic roles of ANP and BNP during the perinatal period are not yet fully understood.

In this study, we sought to clarify the physiologic and pathophysiologic roles of the endogenous ANP/BNP–NPR1 system, focusing particularly on the maternal heart during the perinatal period, by comparing the phenotypes of wild-type and Npr1-knockout mice.

Methods

A full description of the methods is presented in the online-only Data Supplement. The data, methods, and study materials will be made available to other researchers on request for the purposes of reproducing the results or replicating the procedures.

Mice

All animal studies were approved by the National Cerebral and Cardiovascular Center Ethics Committee for Animal Experiments and were conducted in accordance with the guidelines of the Physiological Society of Japan. Npr1^−/− mice were generated at the Howard Hughes Medical Institute (University of Texas Southwestern Medical Center, Dallas).¹³ Mice with floxed Npr1 (Npr1^fl/fl) were generated as described previously.¹⁴ Angiotensin II type 1a receptor–deficient (Agtr1a^−/−) mice were generated at the University of Tsukuba (Japan).¹⁵

Mice with floxed Agtr1a (Agtr1a^fl/fl) were generated as follows. To conditionally knock out the Agtr1a gene, TT2 embryonic stem cells derived from an F1 hybrid of C57BL/6 and CBA mice¹⁶ were transfected with an Agtr1a-targeting vector, selected in the presence of G418, and screened for homologous recombination by polymerase chain reaction analysis and Southern blotting. To generate heterozygous Agtr1a^fl/+ mice, chimeric mice with high embryonic stem cell contributions were crossed with cytomegalovirus–Cre mice (C57BL/6 strain background; RIKEN BioResource Center, Tsukuba, Japan), which express Cre recombinase under the control of the cytomegalovirus promoter; the resultant Agtr1a^fl/+ mice were then intercrossed to obtain Agtr1a^fl/fl mice. To establish tissue-specific Agtr1a-knockout mice, Agtr1a^fl/fl mice were crossed with cytomegalovirus–Flp mice (C57BL/6 background; RIKEN BioResource Center), which express Flp recombinase under the control of the cytomegalovirus early enhancer and chicken β-actin promoter, to remove the Pr–Neo–pA cassette. The resultant allele was designated as Agtr1a^fl/fl (CDB accession number CDB0831K; http://www2.clst.riken.jp/arg/mutant%20mice%20list.html). The genotypes of wild-type and Agtr1a^fl/fl mice were confirmed by polymerase chain reaction using primer sets F1–R1 and F2–R2: F1, 5′-TGTGAGAGTACAGGCTGCTT-3′; R1, 5′-GAGCGTATGCCCACGTACT-3′; F2, 5′-GCAGAAATGCGTTCAACAGAC-3′; R2, 5′-AAACGAGACCGCTGCGATT-3′ (Figure I in the online-only Data Supplement).

Mice with floxed Nr3c2 (Nr3c2^fl/fl) were provided by Dr Günther Schütz (German Cancer Research Center, Heidelberg, Germany).¹⁷Nestin–Cre mice, AQP2–Cre mice, and Rosa26–LacZ reporter mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Transgenic mice expressing Cre recombinase under the control of the α-myosin heavy-chain promoter (αMHC–Cre mice) were provided by Dr Kinya Otsu (King’s College London British Heart Foundation Centre of Research Excellence, London, UK).¹⁸Tie2–Cre mice were provided by Dr Thomas N. Sato (Advanced Telecommunications Research Institute International, Kyoto, Japan).¹⁹

CLC-KB–Cre mice were generated as reported previously.²⁰ In brief, the 11-kb 5′ flanking region of the human CLC-KB gene was fused to Cre recombinase cDNA upstream of the SV40 polyA signal sequence. To confirm that the CLC-KB–Cre transgene was expressed in the distal tubules of the kidney, CLC-KB–Cre mice were crossed with Rosa26-LacZ reporter mice, and X-gal staining was performed on frozen kidney sections of double-transgenic mice (Figure II in the online-only Data Supplement).

To generate neuron-specific, cardiac myocyte–specific, endothelial cell–specific, and collecting duct– and distal tubule–specific Npr1-knockout mice, Npr1^fl/fl mice were crossed with Npr1^fl/fl; Nestin–Cre^+/−, Npr1^fl/fl; αMHC–Cre^+/−, Npr1^fl/fl; Tie2–Cre^+/−, and Npr1^fl/fl; AQP2–Cre^+/−; CLC-KB-Cre^+/− mice, respectively. Mice not harboring the Cre transgene (Npr1^fl/fl mice) served as the corresponding control mice. Tissue-specific Agtr1a-knockout and Nr3c2-knockout mice were generated in the same manner.

To examine the involvement of neuronal aldosterone–mineralocorticoid receptor (MR) signaling in the lactation-induced cardiac hypertrophy in Npr1^−/− mice, the neuronal Nr3c2 was deleted by crossing the Npr1^−/−; Nr3c2^fl/fl mice with Npr1^−/−; Nr3c2^fl/fl; Nestin–Cre^+/− mice. Mice not harboring the Cre transgene (Npr1^−/−; Nr3c2^fl/fl mice) served as the corresponding control mice.

All mice used in this study were on the C57BL/6 background. Mice were group-housed under a 12:12-hour light:dark cycle at 25°C and had unrestricted access to food and water. Female mice (age 8 weeks) were used for experiments.

Western Blot Analysis

The primary and secondary antibodies used for Western blotting are listed in Table I in the online-only Data Supplement. Uncropped immunoblots are shown in Figure III in the online-only Data Supplement.

Statistical Analysis

All data are presented as mean ± standard error of the mean. Animals were assigned randomly to experimental groups. Kaplan-Meier analysis, followed by a log-rank test, was used to compare survival among mice. Pairwise comparisons were performed by using 2-tailed unpaired Student ttests. Differences among 3 or more groups were analyzed by using 1-way or 2-way analysis of variance with the Tukey-Kramer post hoc test. For comparisons between tissue-specific knockout mice, 1-way analysis of variance followed by the Dunnett post hoc test was applied. In the microarray analysis, differential gene expression was determined by using ttests. Adjustment for multiple testing was performed according to the Benjamini-Hochberg procedure, which controlled the false discovery rate at the 0.05 level.²¹ For all comparisons, a P value of less than 0.05 was considered to indicate statistical significance. The detailed results of 2-way analysis of variance are summarized in Table II in the online-only Data Supplement.

Results

PPCM-Like Cardiac Remodeling in Postpartum Npr1-Knockout Mice

The experimental protocol is shown in Figure 1A. The survival rate of Npr1-knockout (Npr1^−/−) mice was significantly lower than that of wild-type (Npr1^+/+) mice over consecutive pregnancy–lactation cycles (Figure 1B; P = 0.0008 versus Npr1^+/+ mice). Surprisingly, of the Npr1^−/− dams that died during consecutive pregnancy–lactation cycles, 75% died during the lactation period (Figure IV in the online-only Data Supplement). After the fifth consecutive pregnancy–lactation cycle, the hearts of Npr1^−/− mice were markedly larger than those of Npr1^+/+ mice and were accompanied by increased lung weight, interstitial fibrosis, and upregulation of mRNA expression of genes related to cardiac hypertrophy (Figure 1C and 1D; Figure V in the online-only Data Supplement). However, the angiogenesis was not impaired in the hearts of Npr1^−/− mice after the fifth consecutive pregnancy–lactation cycle (Figure V in the online-only Data Supplement). Consistent with a previous report describing hypertrophied hearts in Npr1^−/− mice,²² the ratio of heart weight to tibial length (HW/TL) in the nulliparous state was slightly, but significantly, higher in Npr1^−/− mice than in Npr1^+/+ mice (Figure 1E and 1F). However, HW/TL after the first pregnancy–lactation cycle was increased significantly in postpartum Npr1^−/− mice (Figure 1E and 1F), and HW/TL after the second consecutive pregnancy–lactation cycle (2PP) was increased significantly not only in postpartum Npr1^−/− mice but also in Npr1^+/+ mice (Figure 1E and 1F). The cardiac hypertrophy in 2PP Npr1^+/+ mice was completely restored to the nulliparous level by 8 weeks after weaning (Figure 1F). By contrast, postpartum cardiac hypertrophy in 2PP Npr1^−/− mice had not completely dissipated by 8 weeks after weaning (Figure 1F). The ratio of lung weight to tibial length, indicative of lung congestion attributable to heart failure, was almost identical between nulliparous Npr1^+/+ and Npr1^−/− mice; whereas it remained consistent throughout pregnancy–lactation cycles in Npr1^+/+ mice, it significantly increased in Npr1^−/− mice in the first pregnancy–lactation cycle and 2PP (Figure 1G). In addition, regression of ratio of lung weight to tibial length by 8 weeks after weaning tended to be impaired in 2PP Npr1^−/− mice (P = 0.064 versus nulliparous Npr1^−/− mice; Figure 1G). In postpartum Npr1^+/+ mice, cardiac fibrosis was not increased (Figure 1H and 1I), but cardiomyocytes were enlarged significantly (Figure 1J and 1K). Both fibrotic area (Figure 1H and 1I) and cardiomyocyte size (Figure 1J and 1K) increased significantly over consecutive pregnancy–lactation cycles in Npr1^−/− mice.

• Download figure
• Download PowerPoint

Cardiac function and geometry of nulliparous and 2PP mice were assessed by echocardiography (Figure 2A and 2B; Table III in the online-only Data Supplement) and the pressure–volume conductance catheter technique (Figure 2C and 2D; Table III in the online-only Data Supplement). In echocardiographic analysis, significant thickening of the interventricular septum and left ventricular posterior wall, enlargement of the left ventricular end-diastolic dimension, decrease in fractional shortening, and increase in left ventricular mass occurred in 2PP Npr1^−/− mice. Although these echocardiographic changes also were recognized in 2PP Npr1^+/+ mice, they were moderate, and differences were not statistically significant except for left ventricular mass. In hemodynamic analysis, end-systolic volume was significantly greater in 2PP Npr1^+/+ mice than in nulliparous Npr1^+/+ mice. Both end-diastolic and end-systolic volumes of 2PP Npr1^−/− mice were significantly larger than those of nulliparous Npr1^−/− mice. The ejection fraction was significantly reduced in both 2PP Npr1^+/+ and 2PP Npr1^−/− mice. Cardiac output and end-diastolic pressure were not significantly altered in 2PP Npr1^+/+ or Npr1^−/− mice in comparison with the corresponding nulliparous state. Collectively, these data show that Npr1^+/+ and Npr1^−/− mice that experience consecutive pregnancies and deliveries undergo cardiac remodeling.

• Download figure
• Download PowerPoint

To explore the mechanisms underlying postpartum cardiac remodeling, particularly the hypertrophic change, we monitored the expression of signaling molecules potentially related to cardiac hypertrophy in the hearts of postpartum Npr1^+/+ and Npr1^−/− mice. Interestingly, postpartum levels of RCAN1 (regulator of calcineurin 1) and subsequent nuclear translocation of NFATc3 were greater in the hearts of 2PP Npr1^−/− mice than in 2PP Npr1^+/+ mice (Figure 3A and 3B). Although phosphorylation of CaMKII in the hearts of nulliparous mice was comparable to that in the hearts of 2PP mice, phosphorylation of ERK1/2 in the heart was significantly greater in 2PP mice than in nulliparous mice (Figure 3A and 3B). This effect was further increased in Npr1^−/− mice (Figure 3A and 3B). In contrast, the phosphorylation levels of Akt and p38 mitogen-activated protein kinase did not differ significantly between nulliparous and 2PP mice in Npr1^+/+ or Npr1^−/− mice (data not shown). The mRNA expression levels of Trpc3 and Trpc6 in the hearts of 2PP Npr1^−/− mice were comparable with those in 2PP Npr1^+/+ mice (Figure 3C).

• Download figure
• Download PowerPoint

Lactation, But Not Pregnancy, Induces Cardiac Hypertrophy in Mice

To determine which process (ie, pregnancy, delivery, or lactation) is responsible for cardiac hypertrophy in Npr1^+/+ and Npr1^−/− mice, we examined maternal phenotypes during all 3 processes. The experimental protocol is shown in Figure 4A. Consistent with a previous report,²²Npr1^−/− mice showed higher blood pressure in the nulliparous state than did Npr1^+/+ mice (Figure 4B). Although a previous study reported that mice deficient in pro-ANP convertase develop pregnancy-induced hypertension,²³Npr1^−/− mice did not exhibit that phenotype (Figure 4B). In both Npr1^+/+ and Npr1^−/− mice in their first pregnancy–lactation cycle, maternal body weight was highest during late gestation (Figure 4C). Plasma ANP peaked bimodally immediately after delivery and after 2 weeks of lactation in Npr1^+/+ mice, whereas that in Npr1^−/− mice peaked at 2 weeks after delivery (Figure 4C). In contrast, HW/TL did not increase in late gestation or within 3 days after first parturition in either Npr1^+/+ or Npr1^−/− mice (Figure 4D). Furthermore, expression of RCAN1 and phosphorylation of ERK1/2 were not increased during the late phase of pregnancy (E18.5) in either Npr1^+/+ or Npr1^−/− mice (Figure VI in the online-only Data Supplement). These findings imply that cardiac hypertrophy in peripartum mice is not induced by the volume overload associated with pregnancy.

• Download figure
• Download PowerPoint

In contrast, HW/TL and left ventricular end-diastolic dimension were significantly increased within 2 weeks of lactation in Npr1^+/+ and Npr1^−/− mice after first parturition (Figure 4E; Table IV in the online-only Data Supplement). The mRNA expressions of genes associated with cardiac hypertrophy (Nppa, Nppb, and Acta1) were significantly upregulated during lactation in both Npr1^+/+ and Npr1^−/− mice after first parturition (Figure 4F). In comparison, the mRNA expression of genes related to fibrosis (Col3a1, Fn1, and Tgfb1) was significantly upregulated in lactating Npr1^−/− mice only (Figure 4F). However, systolic blood pressure after 2 weeks of lactation did not differ from that immediately postpartum in either Npr1^+/+ or Npr1^−/− mice (Figure VII in the online-only Data Supplement).

The cardiac hypertrophy that occurred in Npr1^+/+ or Npr1^−/− mice after the first pregnancy–lactation cycle was absent when litters were removed immediately after birth (Figure 4E; 2 weeks without lactation). By removing the litters, both the hypertrophied cardiomyocytes and increased mRNA expression related to cardiac hypertrophy were diminished in Npr1^−/− mice after first parturition (Figure VIII in the online-only Data Supplement). However, preventing lactation had no effect on the cardiac function in either Npr1^+/+ or Npr1^−/− mice after first parturition (Table IV in the online-only Data Supplement). These results indicate that lactation, not pregnancy, induces cardiac hypertrophy in mice, thereby illustrating the importance of the ANP/BNP–NPR1 system, which potentially suppresses hypertrophic cardiac remodeling during lactation.

Effects of PRL on the Maternal Heart

Previous work showed that cathepsin D–cleaved 16 kDa PRL induces postpartum cardiomyopathy in mice.²⁴ Hence, we next examined the involvement of mature PRL (23 kDa) and cleaved PRL (16 kDa) in postpartum cardiac hypertrophy. Four-week treatment with an anti-PRL agent (bromocriptine, a dopamine D2 receptor agonist) during lactation suppressed cardiac hypertrophy in 2PP Npr1^−/− but not in 2PP Npr1^+/+ mice (Figure 5A). However, the effect of bromocriptine on cardiac fibrosis, the cross-sectional area of cardiomyocytes, and mRNA expression was minimal (Figure IX in the online-only Data Supplement). Although 2 weeks of administration of mature PRL to Npr1^+/+ mice significantly increased the plasma concentration of PRL (Figure 5B; Figure X in the online-only Data Supplement), mature or cleaved PRL did not influence body weight (Figure 5C) and failed to induce hypertension, fibrosis, or cardiac hypertrophy in nulliparous mice, in mice after first parturition but without lactation, or in mice after first parturition with lactation (Figure 5C and 5D; Figure XI in the online-only Data Supplement). These results imply that factors other than PRL that are suppressed by bromocriptine promote postpartum cardiac hypertrophy in Npr1^−/− mice.

• Download figure
• Download PowerPoint

Plasma Aldosterone Level Is Elevated in Npr1^−/− Mice During the Lactation Period

To identify the factors that induce cardiac hypertrophy in Npr1^−/− mice during the lactation period, we considered the involvement of the renin–angiotensin–aldosterone system, which counteracts the natriuretic peptide system.²⁵ In Npr1^+/+ mice, plasma renin activity tended to decrease during lactation (Figure XII in the online-only Data Supplement). Because the angiotensin II type 1a receptor (Agtr1a) is expressed throughout the body (eg, in brain, heart, blood vessels, and renal tubules),²⁶ we examined the tissue-specific roles of Agtr1a in lactation-induced cardiac hypertrophy by using mice with tissue-specific deletion of Agtr1a in neurons, cardiomyocytes, endothelial cells, or collecting ducts and distal tubular cells of the kidney. However, we observed no significant suppression of lactation-induced cardiac hypertrophy in any of the tissue-specific Agtr1a-knockout mice (Figure 6A and 6B). In addition, cardiac hypertrophy during the lactation period was not inhibited by systemic deletion of Agtr1a (Figure 6C). These results suggest that the tissue angiotensin II–Agtr1a system is unlikely to affect cardiac hypertrophy during the lactation period.

• Download figure
• Download PowerPoint

Two weeks of lactation significantly increased the plasma aldosterone level but not cortisol in the Npr1^−/− mice after first parturition (Figure 6D; Figure XIII in the online-only Data Supplement). This increase in plasma aldosterone was supported by the absence of a lactation-dependent increase in serum potassium (Figure XIV in the online-only Data Supplement). Removal of the litter immediately after birth significantly suppressed the increase in plasma aldosterone level in Npr1^−/− mice after first parturition (Figure 6D). In addition, bromocriptine administration during the lactation period significantly decreased the plasma aldosterone and PRL levels in both Npr1^+/+ and Npr1^−/− mice after first parturition (Figure 6D; Figure XV in the online-only Data Supplement). Expression levels of Cyp11b1 and Cyp11b2 mRNA in the adrenal gland were not influenced by lactation in Npr1^+/+ or Npr1^−/− mice (Figure XVI in the online-only Data Supplement).

Lactation-Dependent Excessive Cardiac Hypertrophy in Npr1^−/− Mice Might Be Induced Through Activation of Neuronal Aldosterone Signaling

To clarify the involvement of aldosterone–MR signaling in lactation-dependent cardiac remodeling, we performed pharmacologic analysis using eplerenone, a selective aldosterone blocker, and genetic analysis using mice lacking MR (encoded by Nr3c2[nuclear receptor subfamily 3, group C, member 2]). For the genetic experiments, we used a tissue-specific knockout because null mutation of Nr3c2 results in death attributable to dehydration at approximately day 10.²⁷ Although eplerenone treatment during lactation did not affect systolic blood pressure (Figure XVII in the online-only Data Supplement), it significantly suppressed the increase of HW/TL after first parturition in Npr1^−/− mice but not Npr1^+/+ mice (Figure 6E). Eplerenone treatment tended to decrease the mRNA expression of Nppa and Acta1 in both Npr1^+/+ and Npr1^−/− mice (Figure XVIII in the online-only Data Supplement). Furthermore, because MR is expressed widely throughout the body (eg, in brain, heart, blood vessels, and renal collecting ducts and distal tubules),^28,29 we evaluated tissue-specific (neurons, cardiomyocytes, endothelial cells, or collecting ducts and distal tubular cells of the kidney) Nr3c2-knockout mice to determine the tissues in which MR makes important contributions to lactation-induced excessive cardiac hypertrophy. In neuron-specific Nr3c2 knockout mice (Nr3c2^fl/fl; Nestin–Cre mice), the hypothalamic Nr3c2 mRNA expression was reduced by half and the lactation-induced physiologic cardiac hypertrophy was significantly suppressed (Figure 6F; Figures XIX and XXA in the online-only Data Supplement). Furthermore, neuron-specific deletion of Nr3c2 almost completely abolished lactation-dependent excessive cardiac hypertrophy in Npr1^−/− mice (Figure 6G; Figure XXB in the online-only Data Supplement). Deleting the neuronal MR signaling tended to decrease the mRNA expressions of Nppa, Nppb, and Acta1 in lactating Npr1^−/− mice (Figure XXI in the online-only Data Supplement). However, the intracerebroventricular infusion of aldosterone to nulliparous Npr1^+/+ mice failed to induce cardiac hypertrophy (Figure XXII in the online-only Data Supplement). These data suggest that cardiac remodeling in Npr1^−/− mice depends on the excessive activation of aldosterone–MR signaling in neural cells, which the endogenous ANP/BNP–NPR1 system counteracts in Npr1^+/+ mice.

Lactation-Dependent Cardiac Hypertrophy in Npr1^−/− Mice Is Mediated in Part by an Interleukin-6–Induced Inflammatory Response

Aldosterone-induced activation of brain MR modulates cardiovascular inflammation, oxidative stress, and sympathetic nerve activity.^30–32 Therefore, we next performed microarray analysis using heart tissue–derived RNA and cDNA from nulliparous and 2-week lactating Npr1^+/+ and Npr1^−/− mice after first parturition. Greater than 2-fold altered expression levels occurred in 3246 and 2336 probes in the hearts of Npr1^+/+ and Npr1^−/− mice, respectively (Tables V and VI in the online-only Data Supplement). In addition, significant (>1.5-fold) changes in gene expression occurred in 27 probes in the hearts of lactating Npr1^−/− mice compared with lactating Npr1^+/+ mice (Figure XXIII and Table VII in the online-only Data Supplement). Evaluation of upstream regulators through Ingenuity Pathway Analysis revealed that cardiac inflammatory cytokines might contribute to lactation-induced cardiac hypertrophy in Npr1^−/− mice (Table VIII in the online-only Data Supplement). Because interleukin (IL)-6 and IL-1β play important roles in the development of cardiac hypertrophy,³³ we examined the mRNA levels of IL-6 and IL-1β in the hearts of nulliparous and lactating Npr1^+/+ and Npr1^−/− mice after first parturition. The experimental protocol is shown in Figure 7A. Compared with that in nulliparous mice, the Il6 mRNA level in the heart tended to be increased in lactating Npr1^+/+ mice but was significantly upregulated in lactating Npr1^−/− mice (Figure 7B). The mRNA level of Il1β tended to increase during lactation in both Npr1^+/+ and Npr1^−/− mice after first parturition (Figure 7B). We also examined the protein expression and phosphorylation levels of signal transducer and activator of transcription 3 (STAT3), a downstream target of IL-6, through Western blotting. Lactation markedly increased phosphorylated STAT3 (p-STAT3α) in the hearts of Npr1^−/− mice, but not Npr1^+/+ mice, after first parturition (Figure 7C). Although lactation had no effect on the plasma levels of IL-6 in either Npr1^+/+ or Npr1^−/− mice, the number of CD68-positive cells in the heart was greater in lactating Npr1^−/− mice compared with lactating Npr1^+/+ mice (Figures XXIV and XXV in the online-only Data Supplement). Eplerenone treatment seemed to decrease Il6 expression in the hearts of Npr1^−/− mice after first parturition (Figure 7D). Furthermore, weekly intraperitoneal injection of anti–IL-6 receptor antibody (MR16-1) tended to suppress lactation-induced cardiac hypertrophy in Npr1^−/− but not in Npr1^+/+ mice after first parturition (Figure 7E). However, pharmacologic modification of sympathetic or parasympathetic nerve activity through the administration of either metoprolol (β1 adrenergic receptor antagonist) or nicotine (α7-nicotinic acetylcholine receptor agonist) during the lactation period did not suppress cardiac hypertrophy in Npr1^−/− mice (Figure 7F). Moreover, the administration of tempol (radical scavenger) to lactating Npr1^−/− mice for 2 weeks did not suppress lactation-dependent cardiac hypertrophy (Figure 7F).

• Download figure
• Download PowerPoint

ANP/BNP–NPR1 System in Cardiomyocytes and Endothelial Cells Protects the Maternal Heart From Inflammation-Mediated Hypertrophic Remodeling During the Lactation Period

Because NPR1 is widely expressed in various cell types and tissues (eg, brain, cardiomyocytes, endothelial cells, distal segments of the renal tubule),³⁴ we compared the HW/TL of tissue-specific Npr1-knockout mice to identify the tissues in which the ANP/BNP–NPR1 system inhibits lactation-dependent cardiac hypertrophy. Mice with cardiomyocyte- or endothelium-specific deletion of Npr1 exhibited increased lactation-dependent cardiac hypertrophy (Figure 7G). These results imply that the endogenous ANP/BNP–NPR1 system in cardiomyocytes and endothelial cells protects the maternal heart from inflammation-mediated hypertrophic remodeling during lactation.

Discussion

We found that consecutive pregnancy–delivery–lactation cycles resulted in reversible cardiac hypertrophy in wild-type mice, which was accompanied by an increase in the transcription of fetal cardiac genes, a reduction in cardiac systolic function, and an increase in activated ERK1/2 phosphorylation but not Akt phosphorylation. These changes are not seen in exercise-induced cardiac hypertrophy, which is accompanied by unchanged fetal cardiac gene mRNA expressions, preserved cardiac function, and an activated PI3-kinase–Akt pathway.⁴ Compared with that in their wild-type counterparts, lactation in Npr1^−/− mice lowered the survival rate and caused more severe cardiac hypertrophy and dysfunction, which are phenomena similar to PPCM. Lactation-induced cardiac hypertrophy in Npr1^−/− mice was blocked by administration of bromocriptine but was not dependent on PRL. We suspect that neural aldosterone–MR signaling–induced IL-6–mediated cardiac inflammation is involved in the lactation-induced cardiac hypertrophy in Npr1^−/− mice. In the hearts of lactating Npr1^−/− mice, the calcineurin–NFAT pathway and STAT3 phosphorylation were activated. In addition, in this study, we revealed that the ANP/BNP–NPR1 system in cardiomyocytes and endothelial cells functions to suppress lactation-dependent inflammation in the heart (Figure 8). Collectively, these findings suggest that lactation is a latent stress on the hearts of mothers.

• Download figure
• Download PowerPoint

The increase in the plasma ANP level immediately after delivery that we observed in this study was consistent with previous reports in humans¹¹ and rodents.³⁵ During pregnancy, plasma volume increases by ≈30% in humans and 80% in mice.³⁵ A previous study has demonstrated that pregnancy-related physiologic cardiac hypertrophy occurs during the late phase of pregnancy.³⁶ Increased ANP in the early postpartum period is thought to contribute to the excretion of excess sodium and water, which accumulates in the maternal body during gestation. In mice, plasma ANP level remained high at day 15 of lactation.³⁵ Other reports show that the plasma volume of maternal mice increases during the midlactation period (days 6 through 15), as well as during gestation.³⁷ Consistent with those previous reports, we observed that plasma ANP level was high on day 14 of the lactation period, nearly equivalent to that immediately after delivery. Cardiac hypertrophy in wild-type mice occurred predominantly during the lactation period rather than during gestation. In addition, it is well known that production of ANP and BNP is increased in the hypertrophied heart.³⁸ Taken together, these findings indicate that the endogenous ANP/BNP–NPR1 system plays an important role in the regulation of plasma volume and the adaptive hypertrophic response in the heart, not only immediately after delivery but also during the lactation period.

Preserved cardiac output and end-diastolic pressure in both 2PP wild-type mice and 2PP Npr1^−/− mice indicated that left ventricular dilation fully compensated for the postpartum reduction in the ejection fraction. However, a significant postpartum increase in ratio of lung weight to tibial length was observed in 2PP Npr1^−/− mice but not in 2PP wild-type mice, and the majority of postpartum Npr1^−/− mice died between 1 and 3 weeks into the lactation period. The plasma ANP level, which increases in heart failure,³⁹ reached a peak at 2 weeks into the lactation period in both wild-type and Npr1^−/− mice. Collectively, these findings suggest that cardiac decompensation could have existed at around 2 weeks into the lactating period in Npr1^−/− mice.

In Npr1^−/− mice, the severe cardiac hypertrophy and dysfunction caused by lactation was similar to PPCM. Because >40% of patients with PPCM have hypertension, pregnancy-induced hypertension is a major risk factor of PPCM.⁴⁰ Mice lacking corin, a transmembrane cardiac serine protease that activates ANP before section, develop a pregnancy-induced hypertension–like phenotype.²³ This effect was not observed in Npr1^−/− mice. The reason for the discrepancy in blood pressure alteration during gestation between corin-deficient mice and Npr1^−/− mice remains unclear; it might be caused in part by a difference in the blood pressure response to salt between corin-knockout (salt-sensitive)⁴¹ and Npr1^−/− (salt-resistant) mice.¹³ Collectively, these observations suggest that Npr1^−/− mice have the potential to serve as a model for PPCM unaccompanied by pregnancy-induced hypertension.

Cardiac hypertrophy is classified as physiologic when cardiac function remains normal and as pathologic when cardiac dysfunction progresses.⁵ Physiologic cardiac hypertrophy occurs in athletes, as well as during pregnancy, and is unaccompanied by the expression of fetal cardiac genes (eg, Nppa, Nppb, Myh7, and Acta1) or fibrosis. The PI3-kinase–Akt pathway is an important mediator of physiologic cardiac hypertrophy.^4,42 By contrast, pathologic cardiac hypertrophy induced by prolonged hemodynamic stress (eg, hypertension) is accompanied by fibrosis, cardiac dysfunction, and fetal cardiac gene expression.⁴² The calcineurin–NFAT pathway, CaMKII pathway, and proinflammatory cytokines are key factors involved in pathologic cardiac hypertrophy.^6,43 In the current study, lactation-induced cardiac hypertrophy in wild-type mice was accompanied by an increase in the phosphorylation of ERK1/2 and an increase in fetal cardiac gene mRNA. These results suggest that lactation-induced cardiac hypertrophy is distinct from exercise- or pregnancy-induced cardiac hypertrophy. Lactation-dependent activation of ERK1/2 signaling was observed regardless of the genotype. Therefore, it is likely that elevated fetal cardiac gene expression in the heart in lactating wild-type mice is at least partially mediated by ERK1/2. The mechanisms underlying the lactation-induced ERK1/2 phosphorylation and increased expression of fetal cardiac gene mRNA in wild-type mice heart should be examined in future studies. In contrast, enhanced RCAN1 protein expression and increased translocation of NFATc3 into nuclei, which indicates activation of the calcineurin–NFAT pathway,^44,45 were detected only in postpartum Npr1^−/− mice. The hearts of postpartum Npr1^−/− mice exhibited reduced cardiac function and remarkable fibrosis. These results suggest that the calcineurin–NFAT pathway plays an important role in the pathologic cardiac remodeling of the maternal heart in Npr1^−/− mice.

Because both ANP and BNP inhibit the secretion of aldosterone,¹⁰ a lack of these hormones leads to an increase in the plasma aldosterone level. Consistent with this, the plasma aldosterone level in Npr1^−/− mice was significantly elevated during the lactation period. Conversely, the mRNA levels of Cyp11b1 and Cyp11b2 in the adrenal glands were not upregulated, implying that the activity of aldosterone synthase is elevated in lactating Npr1^−/− mice. Because the concentration of aldosterone in the brain is affected by the circulating aldosterone level,⁴⁶ brain MR in the Npr1^−/− mice may be activated during lactation. Central aldosterone–MR signaling activates sympathetic nerve activity³¹ and increases oxidative stress.³⁰ However, lactation-induced excessive cardiac hypertrophy in Npr1^−/− dams was not diminished by either metoprolol or tempol, suggesting that sympathetic activity and oxidative stress make minimal contributions to lactation-induced excessive cardiac hypertrophy in Npr1^−/− dams. Treatment with intracerebroventricular infusion of aldosterone for 2 weeks significantly elevated systolic blood pressure but did not induce cardiac hypertrophy in nulliparous wild-type mice. This result implies that the changes in maternal environment that occur during lactation are crucial for neuronal MR activation to promote lactation-induced cardiac remodeling.

The significant increase in HW/TL in nulliparous cardiomyocyte-specific Nr3c2-knockout (Nr3c2^fl/fl; αMHC–Cre) mice compared with Nr3c2^fl/fl mice agrees with previous findings.⁴⁷ Although deletion of MR in cardiomyocytes typically has a cardioprotective effect after cardiac insult,^47,48 the lactation-induced cardiac hypertrophy of Nr3c2^fl/fl; αMHC–Cre mice did not demonstrate a similar beneficial response. This result implies that aldosterone–MR signaling in cardiomyocytes is not involved in lactation-induced cardiac hypertrophy. In contrast, HW/TL was significantly decreased in nulliparous neuron-specific Nr3c2 knockout (Nr3c2^fl/fl; Nestin–Cre) mice. This result might be attributable to the relatively long tibiae in Nr3c2^fl/fl; Nestin–Cre mice, because the heart weight to body weight ratio was almost identical between nulliparous Nr3c2^fl/fl and Nr3c2^fl/fl; Nestin–Cre mice. Even though the lactation-associated increase in body weight remained intact in Nr3c2^fl/fl; Nestin–Cre mice (data not shown), the heart weight to body weight ratio was significantly decreased only in lactating Nr3c2^fl/fl; Nestin–Cre mice. This result implies that the neuronal MR activation plays an important role in lactation-dependent cardiac hypertrophy. The reason why breeding Npr1^−/− mice with Nr3c2^fl/fl; Nestin–Cre mice all but cancelled the influence of the neuronal Nr3c2 deletion on nulliparous HW/TL remains unknown.

The neuronal blockade of Nr3c2 nearly eliminated lactation-induced cardiac hypertrophy in Npr1^−/− mice, whereas eplerenone treatment was only marginally beneficial during lactation. The permeability of the blood–brain barrier to eplerenone might influence the drug’s effect on lactation-induced cardiac hypertrophy in Npr1^−/− mice.

Central aldosterone–MR signaling activation increases the levels of circulating proinflammatory cytokines, including tumor necrosis factor–α, IL-1β, and IL-6, in rats with heart failure,⁴⁹ implying a role for a brain–heart interaction in the pathogenesis of heart failure. Continuous activation of the IL-6–glycoprotein-130 signaling cascade induces cardiac hypertrophy.⁵⁰ The cardiac levels of IL-6 mRNA and phosphorylated STAT3 protein were increased to a greater extent in lactating Npr1^−/− mice than in wild-type mice. Furthermore, MR16-1 administration tended to attenuate cardiac hypertrophy in Npr1^−/− mice. ERK1/2 and STAT3 are located downstream in the IL-6–glycoprotein-130 signaling pathway in cardiomyocyte hypertrophy.⁵¹ Moreover, ERK1/2 and the calcineurin–NFAT pathway interact in cardiac myocytes.⁵² Taken together, these findings suggest that lactation-induced cardiac hypertrophy in Npr1^−/− dams is attributable to the increased level of cardiac IL-6, which is mediated through central aldosterone–MR signaling. However, MR is distributed widely throughout the brain, including the hippocampus, hypothalamus, and circumventricular organs.⁵³ Therefore, additional studies are warranted to elucidate the precise mechanisms underlying the cardiac inflammatory remodeling mediated through lactation-induced neuronal aldosterone–MR signaling in Npr1^−/− mice.

Female mice with cardiomyocyte-specific deletion of stat3 develop PPCM.²⁴ In contrast, the lactation-induced cardiac remodeling in Npr1^−/− mice was accompanied by an upregulation in the Il6 mRNA level and elevated phosphorylation of STAT3. Weekly intraperitoneal injection of anti–IL-6 receptor antibody tended to suppress the lactation-induced cardiac hypertrophy in Npr1^−/− mice. Because continuous (for as long as 2 weeks) glycoprotein-130 (a subunit of common receptor for IL-6)–STAT3 activation in the infarcted murine heart promotes cardiac inflammation and leads to an adverse outcome,⁵⁴ our results imply that the lactation-induced cardiac remodeling in Npr1^−/− mice is mediated through the IL-6–glycoprotein-130–STAT3 pathway. However, the precise mechanism through which STAT3 in cardiomyocytes influences postpartum cardiovascular disorders remains unknown. Further investigation is needed to reconcile the apparent discrepancies between these 2 mouse models.

Mice with cardiomyocyte- or endothelium-specific deletion of Npr1 exhibited increased lactation-induced cardiac hypertrophy. These results might be attributable to increased susceptibility to IL-6 in these mice. In fact, whereas cardiac Il6 mRNA transcripts tended to be increased in wild-type mice, hearts of lactating wild-type mice showed no STAT3 phosphorylation. These results suggest that the endogenous ANP/BNP–NPR1 system inhibits IL-6 signaling in the heart. In addition, the lack of enhanced lactation-induced cardiac hypertrophy in mice with neuron-specific deletion of Npr1 indicates that the neuronal ANP/BNP–NPR1 system does not contribute directly to the development of lactation-induced cardiac hypertrophy. Taken together, the excessive cardiac hypertrophy during the lactation period in Npr1^−/− mice might reflect contributions from both the lack of inhibitory effect of aldosterone secretion and the enhanced susceptibility to inflammatory cytokines in cardiomyocytes and endothelial cells. Additional research is needed to elucidate the relationship between the ANP/BNP–NPR1 system and IL-6 signaling in the heart.

We evaluated the involvement of PRL in the onset of lactation-induced cardiac hypertrophy in Npr1^−/− mice. We found that excessive postpartum cardiac hypertrophy in Npr1^−/− mice was significantly suppressed by the removal of litters or by anti-PRL treatment using bromocriptine, a dopamine D2 receptor agonist. In contrast, 2 weeks of administration of either 23 kDa or 16 kDa PRL failed to induce hypertrophic changes in the hearts of Npr1^−/− mice, calling into question the hypertrophic effect of PRL on cardiomyocytes. Dopamine D2 receptor signaling inhibits secretion of aldosterone and PRL.⁵⁵ Indeed, 2 weeks of bromocriptine treatment during lactation significantly decreased both the plasma aldosterone and PRL level in Npr1^−/− mice. Therefore, the antihypertrophic effect of bromocriptine on the hearts of Npr1^−/− mice during lactation might be caused by suppression of aldosterone secretion, but may be independent of the inhibition of PRL secretion.

Our findings reveal that the endogenous ANP/BNP–NPR1 system plays a protective role on the lactating dam heart by suppressing excessive increases in plasma aldosterone and cardiac IL-6 expression. Considering the reduction in plasma aldosterone level and the subsequent suppression of the IL-6 pathway, the administration of ANP or BNP, an antialdosterone agent, or an anti–IL-6 agent might exert therapeutic effects against postpartum cardiac disorders. Our findings emphasize the importance of carefully monitoring mothers’ hearts during the postpartum period, especially during lactation.

Acknowledgments

We thank Dr Kyoko Shioya for support with animal experiments; Dr Tamaki Mabuchi, Dr Xin-Mei Yu, Atsuko Ikenaga, Mika Kitazume, Mitsuko Nakatani, and Manami Sone for technical assistance; Dr Günther Schütz, Dr Kinya Otsu, and Dr Thomas N. Sato for providing the Nr3c2-floxed mice (Nr3c2^fl/fl), αMHC–Cre mice, and Tie2–Cre mice, respectively; and Dr Naoki Mochizuki for advice.

Footnotes