Physiol. Genomics 31: 139-157, 2007.
First published July 17, 2007; doi:10.1152/physiolgenomics.00007.2006
1094-8341/07 $8.00
Received 11 January 2006;
accepted in final form 5 July 2007.
Physiological Genomics 31:139-157 (2007)
1094-8341/07 $8.00 © 2007 American Physiological Society
Maternal vitamin A alters gene profiles and structural maturation of the rat ductus arteriosus
Utako Yokoyama1,*,
Yoji Sato2,*,
Toru Akaike1,
Seiichi Ishida3,
Junichi Sawada4,
Taku Nagao2,
Hong Quan1,
Meihua Jin1,
Mari Iwamoto5,
Shumpei Yokota5,
Yoshihiro Ishikawa1,6 and
Susumu Minamisawa1,7
1 Department of Physiology, Yokohama City University, Yokohama
2 Division of Cellular and Gene Therapy Products, National Institute of Health Sciences, Tokyo
3 Division of Pharmacology, National Institute of Health Sciences, Tokyo
4 Division of Biochemistry and Immunochemistry, National Institute of Health Sciences, Tokyo
5 Department of Pediatrics, Yokohama City University, Yokohama, Japan
6 Department of Cell Biology and Molecular Medicine and Medicine (Cardiology), New Jersey Medical School, University of Medicine and Dentistry of New Jersey, Newark, New Jersey
7 Department of Life Science and Medical Bio-Science, Waseda University, Tokyo, Japan
 |
ABSTRACT
|
|---|
Retinoic acid (RA), a metabolite of vitamin A, has been proposed to regulate vascular remodeling and reactivity of the ductus arteriosus (DA). Using rat Affymetrix GeneChips, we found that a considerable number of genes in DA varied their expression levels in accordance with developmental mode: namely, preterm-, term-, and postnatal-dominant clusters. Among a total of 8,740 probe sets, maternal vitamin A administration (MVA) changed the expression levels of 91 genes (116 probe sets) >2.5-fold. About half of preterm- and term-dominant genes responded to MVA, whereas only 5% of postnatal-dominant genes responded to MVA, indicating that fetal-dominant genes were susceptible to RA signals. The expression levels of 51 genes in MVA-treated DA at preterm were similar to the expression levels in nontreated DA at term, indicating that the global gene profile at preterm resembled that of the control animal at term. We observed neointima formation in MVA-treated DA at preterm in accordance with upregulation of fibronectin and hyaluronic acid, whereas it was rarely observed in nontreated DA at preterm. Five fetal cardiac myofibrillar genes were also upregulated in MVA-treated in vivo DA, whereas they were developmentally downregulated in nontreated DA. The present study indicates that MVA-mediated alteration in gene profile was associated with early structural maturation of DA, although MVA-mediated maturation may differ from normal vascular remodeling of DA.
retinoic acid; vascular remodeling; development; myofibrillar differentiation
|
INTRODUCTION
|
|---|
PATENCY OF THE DUCTUS ARTERIOSUS (DA), a fetal arterial connection between the pulmonary artery and the descending aorta, is essential in gestation. The morphology and function of the DA change dramatically during development. At late gestation the deposition of extracellular matrix in the subendothelium is increased, and smooth muscle cells of the media migrate into this region, resulting in intimal thickening (2, 12). After birth, the DA closes immediately through the contraction of its smooth muscle by an increase in oxygen tension, and a dramatic decline in circulating prostaglandins (8). The DA later undergoes permanent closure through structural remodeling and apoptosis. The DA, however, remains patent after birth, especially in premature infants, which is responsible for significant morbidity and mortality. Therefore, it is important to understand the precise mechanisms by which the DA exhibits developmental mode-specific remodeling. We hypothesized that this specification of the DA is largely dependent on the expression of a distinct subsets of genes involved in developmental vascular remodeling.
Retinoic acid (RA), a metabolite of vitamin A, is one of the most important regulatory factors of gene transcription for cardiovascular development (9, 30). Maternal vitamin A administration (MVA) exerts its effect through oxidized metabolites of retinol. Colbert et al. (10) demonstrated that a strong RA response signal was predominantly detected in the developing DA and that the signal was colocalized with the expression of the adult-specific smooth muscle myosin heavy chain (MHC) isoform, SM2. They have suggested that RA plays an important role in differentiation and maturation of the DA. Accordingly, Momma's group has demonstrated that MVA enhances the susceptibility of contraction of the rat DA in response to oxygen exposure (35) or indomethacin (20) even in the premature fetus at day 19 of gestation, suggesting that MVA accelerates vascular maturation of the DA in premature fetuses, at least in terms of vascular contraction. On the basis of these animal experiments, vitamin A therapy is considered to be a potent, novel strategy for DA closure in premature infants (29). However, the first clinical trial using vitamin A therapy has failed in premature infants with patent DA (25). To determine whether vitamin A therapy is really worthwhile for premature infants with patent DA, the precise mechanism of MVA-mediated maturation of DA needs to be clarified. Since the developmental mode-specific remodeling of DA is largely dependent on precise regulation of gene expression, RA may alter the transcriptional profiling from immature to mature state of the DA. Accordingly, to understand the precise network of distinct subsets of genes that are regulated by MVA, we have performed comprehensive analyses of gene expression patterns in DA in the presence or absence of MVA.
|
MATERIALS AND METHODS
|
|---|
Vitamin A Administration
The experiments were approved by the Ethical Committee of Animal Experiments of Yokohama City University School of Medicine. Vitamin A (retinyl palmitate, 33 mg, which is equivalent to 18 mg of retinol) (Eisai, Tokyo, Japan) was diluted with polyoxyethylene castor oil and was injected intramuscularly daily into pregnant Wistar rats from the 17th day of gestation at doses of 1 mg (3,000 IU)/kg body wt as previously described (36). The dose of vitamin A was determined in accordance with the previous studies (20, 35). We confirmed a significant increase in serum retinol palmitate concentration in vitamin A-injected maternal rats (at least >300 µg/dl) by HPLC.
Tissue Collection and Preparation
Pooled tissues from the DA in the presence or absence of MVA were obtained from Wistar rat fetuses on the 19th day of gestation (preterm, n 
120) and the 21st day of gestation (term, n 120), and from neonates on the day of birth (n 120). Since the timing of tissue collection after birth is critical, we collected the postnatal DA samples between 3 and 6 h after birth. The total number of MVA was two times in the 19-day fetuses and four times in the 21-day fetuses and neonates. Therefore, it should be noted that the conditions of MVA treatment were not exactly the same among MVA-treated animals. After excision, tissues were immediately frozen in liquid nitrogen and stored at –80°C.
Total RNA Preparation and DNA Microarray Analysis
Total RNA was isolated from pooled tissues with TRIzol (Invitrogen). After treatment with DNase I (Promega, Tokyo, Japan), a second clean-up step was performed on the isolated RNA, using an RNeasy Mini total RNA Preparation Kit (Qiagen, Tokyo, Japan). Total RNA was converted to biotin-labeled cRNA that was hybridized to rat genome U34A GeneChip DNA microarray (Affymetrix, Santa Clara, CA) for 16–24 h at 45°C. The hybridization signals on the microarray were scanned and analyzed by GeneArray Scanner and Microarray Suite Software (Affymetrix), respectively. The hybridization experiments were performed in duplicate. At the first step, sequences without expression in DA at each developmental mode, which were indicated as "absent" by "absolute analysis" of Microarray Suite Software, were eliminated from the data set. Then, if the difference in the signal intensities of a given sequence tag was equal to the cut-off (=2.5-fold) or more at any developmental mode or between any two developmental modes, and if the "comparison analysis" of the Microarray Suite Software indicated "increased" or "decreased" at any developmental mode with 2.5-fold difference or between any two with such a difference, that sequence tag was employed for further analyses. The signal intensities of duplicate genes were averaged. For cluster analysis related to developmental changes in gene profiles, the signal intensities were standardized by subtracting the mean across groups of interest and dividing by the groups' standard deviation, so that the mean and the standard deviation for each sequence tag was 0 and 1, respectively. Expression pattern clusters were defined by subjecting the standardized signal intensities to the hierarchical tree-clustering algorithm, using the SYSTAT program (SYSTAT Software). For the hierarchical tree analysis, Euclidean distance and Ward's method were employed as the distance measure and the linkage rule, respectively. For cluster analysis related to MVA, the selected genes were defined by subjecting the difference in the signal intensities to "increase (2.5)" and "decrease (
0.4)" by MVA at each developmental mode. Then, the selected genes were patterned into four groups in accordance with the developmental differences of signal intensity affected by MVA. All the microarray data in the present study were deposited to the GEO repository (http://www.ncbi.nlm.nih.gov/projects/geo/; accession number GSE3420).
Primary Culture of Rat DA Smooth Muscle Cells
Vascular smooth muscle cells (SMCs) in primary culture were obtained from DA of Wistar rat fetuses at the 21st day of gestation. Minced tissues were transferred to 800 µl of collagenase-dispase enzyme mixture as described previously (36). DA SMCs were harvested in DMEM containing 0.1% fetal calf serum in ambient air in the presence or absence of all-trans retinoic acid (atRA) at a concentration of 10 nM for 4 days. DA SMCs were collected with TRIzol (Invitrogen).
Tissue Staining and Immunohistochemistry
Paraffin-embedded blocks containing DA tissues were prepared as previously demonstrated (37). Hyaluronic acid (HA) staining was done as described previously (37). For immunostaining of fibronectin, a rabbit anti-mouse antibody for fibronectin (Santa Cruz) was incubated at 4°C overnight.
SMC Migration Assay
DA SMCs were harvested in DMEM containing 0.1% fetal calf serum in the presence or absence of atRA at a concentration of 10 nM for 3 days. DA SMC migration assay was performed as previously demonstrated with some modifications (37).
Quantitative RT-PCR Analyses
To verify the microarray hybridization results, we performed quantitative RT-PCR analyses from the same RNA samples used for microarray hybridization as described previously (19, 33). The specifics for PCR primers used in RT-PCR analyses is provided in Supplemental Table S1.1
|
RESULTS
|
|---|
Development-dependent Gene Profiles in DA
Since the morphology and function of DA dramatically change during development, we investigated the developmental changes in the gene profiles in DA. Among a total of 8,740 probe sets, a total of 163 genes (219 probe sets) showed the significant difference (2.5-fold) between any two developmental modes. The Euclidean distance during hierarchical clustering provided the rationale to classify genes into three clusters (Fig. 1 and Supplemental Fig. S1): the first cluster contained 61 genes (86 probe sets) showing a preterm-dominant expression (preterm-dominant cluster), the second cluster contained 67 genes (80 probe sets) showing a term-dominant expression (term-dominant cluster), and the third cluster contained 38 genes (53 probe sets) showing a postnatal dominant expression (postnatal-dominant cluster) (Tables 1–3). Three probe sets encoding common genes (AF077354_g_at and AF077354_at for heat shock protein 4; rc_AA925762_at, rc_AA859896_at and rc_AA955167_s_at for myristoylated alanine-rich protein kinase C substrate; rc_AA818604_s_at, Z75029_s_at and L16764_s_at for heat shock protein 1) belonged to two different clusters, resulting in the inconsistency in numbers between a total of 163 genes and summation of numbers of each cluster (n = 166).
View larger version (32K):
[in this window]
[in a new window]
|
Fig. 1. Hierarchial clustering of 166 genes that were changed significantly during development in DA. Color bar indicates the relative changes in mean normalized intensity of expression of these genes. The Euclidean distance during hierarchical clustering gave rationale to classify genes into 3 clusters shown at right. Three line graphs showed the changes in the standardized expression levels at each developmental mode in DA. Data are expressed as means ± SD of standardized signal intensities. DA, ductus arteriosus.
|
|
Of 61 genes in the preterm-dominant cluster (Table 1), most of genes have not been characterized in DA. The preterm-dominant cluster includes five genes encoding proteins related to cytoskeletal organization and the extracellular matrix such as the MHC polypeptide 7 (Myh7) and cardiac 
-actin (Actc1). This cluster also contained many transport-related genes such as albumin and -fetoprotein.
Of 67 genes in the term-dominant cluster (Table 2), several genes such as fibronectin; prostaglandin E receptor 4 (subtype EP4) (Ptger4); cGMP-specific phosphodiesterase 5A (Pde5a); and ATP-binding cassette, subfamily C, member 9 (Abcc9) have physiological roles in DA. However, most genes have an unknown function in DA. Since the vascular remodeling of DA becomes prominent in the late stage of gestation (term), genes in the term-dominant cluster may be potential targets for regulatory factors of vascular remodeling and maturation in DA.
Of 38 genes in the postnatal dominant cluster (Table 3),most of genes have not been characterized in DA. There were many immediate-early genes such as Jun B, early growth response 1 (Egr1), zinc finger protein 36, dual specificity phosphatase 1, and B-cell translocation gene 2 (Btg2) in this cluster. Genes in this cluster may represent the response to the transition from fetal to neonatal circulation.
Changes in Gene Profiles by MVA in DA
Among a total of 8,740 probe sets, a total of 91 genes (116 probe sets) showed significant difference between DA in the presence or absence of MVA at least at one developmental mode. The effects of MVA on transcriptional profiles in the DA were dependent on its developmental mode. Based on the developmental mode-specific effects of MVA on gene expression, we classified the 116 probe sets into four groups: downregulation, preterm upregulation, perinatal upregulation, and postnatal upregulation (Fig. 2). Cyclin D2 was assigned to two groups in a probe set-dependent manner (D16308_at and rc_AA899106_ at), resulting in the inconsistency in numbers between a total of 91 genes and summation of numbers of each group (n = 92).
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 2. Clustering of 91 genes that showed the significant difference between DA in the presence or absence of MVA. Four line graphs showed the changes in the standardized expression levels at each developmental mode in DA. Data are expressed as means ± SD of standardized signal intensities. MVA, maternal vitamin A administration.
|
|
In cluster A (downregulation), in the absence of MVA, the averaged signal intensities of genes in this cluster were developmentally decreased from preterm to term. Of 27 genes, 22 genes belonged to the preterm-dominant cluster. MVA decreased the gene expression levels, especially at preterm (Table 4). Accordingly, the expression levels of MVA-treated genes at preterm became similar to expression levels of nontreated DA at term, suggesting that MVA promoted the downregulation of these genes. Many transport-related genes such as albumin and -fetoprotein belonged to this group.
In cluster B (upregulation at preterm), in the absence of MVA, the averaged signal intensities of genes in this cluster were developmentally increased from preterm to term. Of 24 genes, 19 genes belonged to the perinatal dominant cluster. MVA upregulated the gene expression at preterm (Table 5). Accordingly, the expression levels of MVA-treated genes at preterm became similar to expression levels of nontreated DA at term, indicating that MVA promoted the upregulated expression of genes. In both cluster A and B, MVA promoted the developmental changes in gene profiles, resulting in a global profile at preterm that resembled that of the control animal at term.
In cluster B, the physiological roles of three genes have been identified in DA: fibronectin 1 (3); Pde5a (32); and ATP-binding cassette, subfamily C (CFTR/MRP), member 9 (Abcc9) (21). Abcc9 is also known as sulfonylurea receptor 2 (SUR2), a component of ATP-sensitive potassium (KATP) channel. We will discuss the role of these genes in vascular remodeling and/or contraction later in Specific Genes of Interest. The physiological roles of other genes have not been clarified in DA.
In cluster C (perinatal upregulation), in the absence of MVA, the averaged signal intensities of genes in this cluster were developmentally decreased. MVA maintained the expression levels of five genes at term at similar or even higher levels of those at preterm (Table 6). Among them, four genes were related to cytoskeleton organization and belonged to the preterm-dominant cluster.
In cluster D (postnatal upregulation), in the absence of MVA, the averaged signal intensities of genes in this cluster were decreased at birth. MVA maintained the expression levels of these genes at birth at similar or even higher levels of those at preterm and term (Table 7). Of 37 genes, three and 14 genes belonged to the preterm- and term-dominant clusters, respectively. Vascular endothelial growth factor (VEGF), which plays an important role in vascular remodeling of DA (6), belonged to this cluster. The expression levels of integrin 1 and ß3 gene were significantly increased by MVA at birth. In addition, several genes involved in cell differentiation and apoptosis were found in cluster D. The physiological role of most genes, however, has not been identified in DA.
Interestingly, the expression of half of genes in the preterm-dominant (30/61) and term-dominant clusters (33/67) was altered in response to MVA, whereas that of only 5% of genes in postnatal dominant cluster (2/38) was changed by MVA, indicating that fetus-dominant genes were susceptible to retinoid signal.
Specific Genes of Interest
Extracellular matrix.
Since extracellular matrix has profound effects on vascular remodeling of DA (24), RA-mediated changes in extracellular matrix are of great interest. We found that MVA significantly increased fibronectin transcripts at preterm and integrin 1 and ß3 transcripts after birth (Fig. 3, A–C). The expression levels of other genes encoding extracellular matrix proteins including collagens, heparan sulfate proteoglycans, chondroitin sulfate proteoglycans, and laminins were not altered by MVA (Fig. 3, D–G). In addition to fibronectin, HA plays an important role in intimal cushion formation in DA (3, 37). No information regarding HA synthase (HAS) was available in the present study, since there was no probe for a gene encoding HAS in the rat genome U34A GeneChip. Therefore, by quantitative RT-PCR analysis, we investigated the expression of HAS2, a predominant isoform in vascular SMCs. MVA increased the expression of HAS2 mRNA by 1.8-fold at preterm (Fig. 3H).
Since previous studies have demonstrated that fibronectin and HA play an important role in intimal cushion formation (2, 17, 37), we hypothesized that upregulation of fibronectin and HA by MVA may accelerate the formation of intimal cushion at preterm. We found that intimal cushion formation was already visible in MVA-treated DA at preterm as usually seen in DA at term, whereas it was rarely found in control DA at preterm (Fig. 4A). Accordingly, the intima-media ratio was higher in MVA-treated DA than in control DA at preterm (Fig. 4B), even though the media of the smooth muscle layer was thicker in MVA-treated DA than in control DA at preterm (Fig. 4C). The thickness of intimal cushion was not different between MVA-treated DA than in control DA at term. We also found strong HA and fibronectin staining in MVA-treated DA at preterm at a similar level with that in control DA at term (Fig. 4A). Furthermore, we investigated the effects of atRA on DA SMC migration using a modified Boyden chamber method, because SMC migration into the subendothelial layer is an important cellular process for intimal cushion formation in DA. In the presence of atRA, SMC migration was slightly increased at basal level although the value did not reach statistical significance (Fig. 5A). PDGF, a potent enhancer of cell migration, significantly increased DA SMC migration. atRA administration further increased PDGF-induced migration (Fig. 5A). The expression of fibronectin was significantly increased in atRA-treated DA SMCs when compared with that in nontreated DA SMCs (Fig. 5B). The expression of HAS2 was also increased in atRA-treated DA SMCs, although the value did not reach statistical significance (Fig. 5B).
View larger version (56K):
[in this window]
[in a new window]
|
Fig. 4. MVA promoted intimal cushion formation at preterm. A: effect of MVA on intimal cushion formation and HA and fibronectin production in rat DA. Intimal cushion formation was already visible in MVA-treated DA at preterm as usually seen in DA at term, whereas it was rarely found in control DA at preterm. Intimal cushion formation was similar between MVA-treated and control DA at term. HA production was visualized by HABP stain (middle). The stain of HABP and fibronectin was increased in preterm DA in the presence of MVA. Scale bars, 100 µm. B: intima-media ratio was higher in MVA-treated DA than in control DA at preterm. C: media of DA was thicker MVA-treated DA than in control DA at preterm (n = 5–7). Data are expressed as means ± SE. Arrowheads indicate the region of intimal thickening. **P < 0.01 and ***P < 0.001. HABP, HA binding protein (HABP).
|
|
View larger version (9K):
[in this window]
[in a new window]
|
Fig. 5. Effects of atRA on DA SMC migration. A: cell migration in response to atRA exposure was measured by modified Boyden chamber method. The PDGF-BB-induced SMC migration was significantly increased by atRA (n = 16). Data are expressed as means ± SE. B: expression of fibronectin-1 and HAS2 mRNAs was measured by quantitative RT-PCR analysis (n = 6). Data are expressed as means ± SE. atRA, all-trans retinoic acid; PDGF-BB, platelet-derived growth factor-BB; ns, not significant; SMC, smooth muscle cell.
|
|
Cytoskeleton.
Since the composition of the genes related to cytoskeletal organization is known to be an excellent marker of muscle differentiation, we evaluated the effects of MVA on the expression of myofibril-related genes in DA in detail. As described above, the expression levels of fetal cardiac type of five myofibril-related genes remained the same or showed elevated expression levels compared with those at preterm by MVA (Fig. 6, A–E). Accordingly, MVA tended to prevent the genes encoding cardiac myofibrillar proteins from developmentally dependent downregulation in matured DA, resulting in retarded smooth muscle differentiation in DA. The expression of the gene encoding similar to Myh11 protein was upregulated at preterm by MVA (Fig. 6F). By contrast, the expression levels of the smooth muscle MHC genes such as SM22, SM1, and SM2 were not altered by MVA (Fig. 6, G–I). Although most of data were consistent between quantitative RT-PCR and microarray analyses, it should be noted that there was an inconsistency between quantitative RT-PCR and microarray data regarding the expression of SM22 mRNA (Supplemental Fig. S2). The microarray analysis showed that the expression of SM22 mRNA was further upregulated after birth, whereas quantitative RT-PCR analysis demonstrated that it was decreased from the 21-day of gestation to a postnatal period.
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 6. The pattern of expression changes in myofibril-related genes by MVA. The expression levels of Myh7 (A), Actc1 (B), Myl7 (C), Tnnt2 (D), Tnni3 (E), and Myh11 (F) mRNAs were upregulated by MVA. MVA did not change these expression levels of SM22 (G), SM1 (H), and SM2 (I) mRNAs. Data are expressed as means of duplicate experiments in expression value.  , Nontreated DA; •, MVA-treated DA; *2.5-fold difference between nontreated and MVA-treated DA. Myh7, myosin, heavy polypeptide 7, cardiac muscle, beta; Actc1, actin alpha cardiac 1; Myl7, myosin, light polypeptide 7, regulatory; Tnnt2, troponin T2, cardiac; Tnni3, troponin I type 3 (cardiac); Myh11, myosin, heavy polypeptide 11, smooth muscle; SM22, transgelin; SM1, smooth muscle myosin heavy chain 1; SM2, smooth muscle myosin heavy chain 2.
|
|
Because Colbert et al. (10) demonstrated that endogenous RA signaling, which colocalized with advanced expression of SM2, promoted SM2 induction during development and a number of studies have demonstrated that RA induces differentiation in vascular SMCs (15, 18), the above finding was somehow unexpected. Therefore, we investigated the effect of atRA on the expression of myofibrillar genes such as Myh7, Actc1, Tnni3, SM22, SM1, and SM2 using cultured DA SMCs. We found that atRA upregulated SM1, SM2, and SM22 mRNAs and downregulated Myh7 mRNA in cultured DA SMCs (Fig. 7). Thus, the result obtained from cultured DA SMCs was apparently different from the DNA microarray data obtained from in vivo MVA-treated DA.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 7. The changes in the expression of myofibril-related genes in cultured DA SMCs in the presence of atRA (10 nM) (n = 6–9). Data are expressed as means ± SE. Abbreviations are defined in Fig. 6 legend.
|
|
Phosphodiesterase.
The PDE gene superfamily consists of at least nine genes: Pde1 to Pde9 (14). The present microarray analyses revealed that Pde1, Pde2, Pde3, Pde4b, and Pde5a were present in DA at least at one developmental mode and that Pde5a was the only isoform whose expression was significantly altered by MVA (Fig. 8). Both quantitative RT-PCR and microarray analyses demonstrated MVA upregulated Pde5a mRNA at preterm, although they showed an inconsistency in the postnatal expression (Supplemental Fig. S3).
View larger version (10K):
[in this window]
[in a new window]
|
Fig. 8. The pattern of expression changes in genes encoding PDE by MVA. MVA increased PDE5A transcripts at preterm. Data are expressed as means of duplicate experiments in expression value. , Nontreated DA; •, MVA-treated DA; *2.5-fold difference between nontreated and MVA-treated DA. PDE, phosphodiesterase.
|
|
Ion channels and pumps.
Nakanishi et al. (21) have demonstrated that KATP channels play an important role in oxygen-induced vasoconstriction of DA. KATP channels are hetero-octameric proteins composed of inwardly rectifying potassium channels and SUR subunits. We found that MVA increased SUR2 (Abcc9) transcripts at preterm (Fig. 9A). Another SUR isoform, SUR1, was called "absent" in DA. Among inwardly rectifying potassium channels, Kir6.1 and Kir1.1 were called "present" in DA. The expression of both genes was not affected by MVA (Fig. 9, B and C). Other potassium channels, such as Kv1.5 and Kv2.1, were called "absent" in DA. We also found that two channel-related genes encoding potassium channel tetramerization domain containing 12 (predicted) and voltage-dependent anion channel 1 belonged to cluster B and D, respectively (Fig. 9, D and E). Sarcoplasmic reticulum Ca2+ ATPase (ATPase, Ca2+ transporting, cardiac muscle, slow twitch 2) mRNA was postnatally upregulated by MVA (Fig. 9F). Phospholamban, an endogenous inhibitory protein of sarcoplasmic reticulum Ca2+ ATPase, was called "absent" in DA.
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 9. Expression profile of genes encoding ion channels and pumps in the absence or presence of MVA. Data are expressed as means of duplicate experiments in expression value. , Nontreated DA; •, MVA-treated DA; *2.5-fold difference between nontreated and MVA-treated DA. Acc9, ATP-binding cassette, subfamily C (CFTR/MRP), member 9; Kir6.1, potassium inwardly rectifying channel, subfamily J, member 8; Kir1.1, potassium inwardly rectifying channel, subfamily J, member 1.
|
|
Other genes of which a functional role has been well characterized in DA.
Clyman et al. (6) has demonstrated that VEGF plays an important role in the formation of neointimal mounds and vasa vasorum in-growth during permanent ductus closure. MVA postnatally increased the expression of VEGF-A mRNA (Fig. 10A). Endothelin-1 has been known to induce contraction of DA, resulting in postnatal DA closure (11, 31). MVA did not alter the expression of endothelin-1 and endothelin converting enzyme (Fig. 10, B and C). Endothelin receptor A and B were called "absent" in DA. Since prostaglandin E receptor subtype 4 (EP4) and cyclooxygenase-1 and -2 (COX-1, -2) play a critical role in the patency of DA (16, 22, 27), the effect of MVA on the expression of these genes is of great interest. The expression levels of EP4, COX-1, and COX-2 mRNAs were not significantly altered by MVA (Fig. 10, D–F).
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 10. Expression profiles of DA-related genes in the absence or presence of MVA. MVA postnatally increased the expression of VEGF-A mRNA (A). Data are expressed as means of duplicate experiments in expression value. , Nontreated DA; •, MVA-treated DA; *2.5-fold difference between nontreated and MVA-treated DA. VEGF-A, vascular endothelial growth factor-A; ET-1, endothelin-1; ECE, endothelin converting enzyme; EP4, prostaglandin E receptor subtype 4; COX-1, cyclooxygenase-1; COX-2, cyclooxygenase-2.
|
|
Confirmation of Accuracy of Microarray Hybridization Results by Quantitative Real-Time RT-PCR Analyses
To confirm the accuracy of microarray hybridization results, we used quantitative RT-PCR analyses to examine the expression of 13 mRNAs: SM1, SM2, SM22, Myh7, Actc1, Tnnt2, Tnni3, Pde5a, Abcc9, Kcnj8, EP4, endothelin-1, and fibronectin 1. Although there were incongruent data between microarray and quantitative RT-PCR analyses in individual genes, we found significant correlation between microarray hybridization and quantitative RT-PCR results (Supplemental Fig. S4), indicating that the present results described above were verified.
|
DISCUSSION
|
|---|
The present microarray analysis revealed that the gene expression profiles in the DA dramatically changed during development. The changes in the expression of the distinct subsets of genes are responsible at least in part for the changes in morphology and function of the DA during development. Genes in the preterm-dominant cluster and in the term-dominant cluster may be associated with immaturity and maturity of the DA, respectively. Genes in the postnatal-dominant cluster may be associated with the response to the transition from fetal to neonatal circulation. In accordance with the developmental changes in gene profiles, one of the important findings of the present study is that the effects of MVA on gene expression profiles are dependent on DA development and conditions such as the transition from fetal to neonatal circulation. About half of the genes in the preterm-dominant and term-dominant clusters responded to MVA, whereas only 5% of genes in the postnatal-dominant cluster responded to MVA. These results indicated that fetus-dominant genes were susceptible to retinoid signal. On the other hand, it should be noted that a considerable number of genes were postnatally upregulated by MVA (see cluster D), even though these genes were exposed to RA stimuli from the 17th day of gestation. Therefore, the results suggest that in addition to RA signals, other postnatal factors such as oxygen must play a role in the regulation of MVA-related gene expression after birth. In this regard, Demary et. al. (13) have demonstrated that RA receptor activity is regulated by redox state. Costa et al. (11) also demonstrated that oxygen changed gene expression profiles in rat DA. Not only oxygen, but also declined circulating PGE2 may contribute to MVA-related postnatal upregulation of genes. Further study is required to identify the factor(s) that synergistically affect the gene profiles in MVA-treated DA after birth.
We found that MVA promoted the developmental changes in gene profiles, especially at preterm, resulting in a global gene profile at preterm largely resembling that of the control animal at term. These results suggested that MVA could promote an earlier maturation of the functions and structure of DA. In this sense, we found that MVA promoted intimal cushion formation in premature DA. Accordingly, genes in cluster Aand B are primary candidates to be involved in MVA-promoted intimal cushion formation. The present study showed that MVA increased fibronectin 1 and HAS2 transcripts at preterm, which is consistent with the previous studies demonstrating that RA increased the expression of fibronectin (28) and HAS2 (26) mRNAs. Previous studies have demonstrated that fibronectin and hyaluronan play a critical role in intimal cushion formation in DA (2, 3, 17, 37). We also found that atRA significantly increased PDGF-mediated DA SMC migration. Accordingly, the increases in fibronectin and HAS2 transcripts could contribute to MVA-promoted SMC migration and thus intimal cushion formation of the DA at preterm. Furthermore, among clusters A and B, several genes such as moesin (1), angiotensin II receptor type 2 (34), lysyl oxidase (23), insulin growth factor 2 (38), and coagulation factor 2 (F2: thrombin) (4) have been known to be associated with formation of neointimal thickening in other vascular smooth muscles. Therefore, it is likely that MVA also promotes maturation of DA at preterm in terms of vascular remodeling. Further investigation is warranted to explore the physiological role of these genes in vascular remodeling in DA.
In addition, previous studies have demonstrated that MVA-treated rat DA at preterm responded to vasoconstrictive stimuli such as oxygen to the same degree as control DA at term, whereas nontreated DA at preterm did not (20, 35), indicating that MVA accelerates functional maturation of DA at preterm. Among 51 genes in clusters A and B, Pde5 (32) and KATP (21) have been assigned a role in oxygen-induced vasoconstriction of DA. Therefore, the present study supported that Pde5a and SUR2 (Abcc9) are involved in the MVA-promoted sensitivity of premature DA to oxygen-induced contraction. In addition to Pde5a, we also found that MVA increased the expression of Pde4b gene in DA (Fig. 8), although it did not reach statistical significance. Since Pde4 modulates generation of reactive oxygen species through activation of cAMP-PKA pathway (14), Pde4b may also modulate the relaxant action of PGE2. Furthermore, the expression of potassium channel tetramerization domain containing 12 (predicted) mRNA was upregulated by MVA at preterm (Fig. 9D). The function of this gene has remained unknown. Since potassium channels are involved in oxygen-induced contraction of DA, it is intriguing to investigate whether potassium channel tetramerization domain containing 12 plays a role in oxygen-induced contraction of DA.
MVA postnatally increased the expression of VEGF-A and integrin 1 and ß3 mRNAs that also have been known to play a role in vascular remodeling in DA (5–7). Clyman et al. (6) has demonstrated that VEGF promotes the formation of neointimal mounds and vasa vasorum in-growth during permanent ductus closure. Integrin 1 and ß3 have been known to play a role in DA SMC adhesion and migration (5, 7). Therefore, MVA may contribute to permanent closure of DA after birth via increases in VEGF-A and integrin 1 and ß3 mRNAs. In addition, several genes involved in cell differentiation and apoptosis in cluster D may also contribute to promotion of postnatal vascular remodeling and thus permanent closure of DA.
Although we show that MVA accelerates maturation of DA, an unexpected, but important finding of the present study is that MVA prevented the downregulation of several cardiac myofibrillar genes in DA during development. We found that MVA did not significantly increase the transcripts of differentiated myofibrillar proteins such as SM1, SM2, and SM22 in DA. On the contrary, the expression levels of fetal cardiac type of myofibril-related genes remained the same or showed higher expression levels of those at preterm by MVA. Therefore, MVA may have the effect, at least in part, of diminishing the distinct profiles of SMC phenotype in DA. These results suggested that MVA retarded the smooth muscle specification of DA in term of myofibrillar genes. This finding was different from other previous studies (10, 15, 18) and our in vitro study using cultured DA SMCs. We assume that the difference is due to the experimental conditions. Because the transcripts of differentiated smooth myofibrillar proteins such as SM2 have already been highly upregulated in in vivo DA, the additional RA signal via MVA has little effect on it. On the other hand, cultured in vitro SMCs could respond to RA signals, because they may exhibit poorly differentiated character. To our knowledge, no study has demonstrated the effect of RA signal on the expression of fetal cardiac myofibrillar genes in an in vivo artery. Further study will be apparently required to understand the distinct responses of myofibrillar genes to RA signal in different experimental conditions of DA SMCs.
In conclusion, by using comprehensive microarray analyses, we found 91 RA-related genes in the DA during developmental vascular remodeling. In addition to the expected genes, microarray analysis uncovered many genes whose roles were not previously recognized in DA. MVA promoted the developmental changes in gene profiles, especially at preterm, which could be associated with earlier functional and structural maturation of DA. These data imply the beneficial effect of vitamin A therapy for premature infants with patent DA. On the other hand, vitamin A therapy may retard myofibrillar differentiation of DA SMCs. Moreover, the response to vitamin A was quite different before and after birth. These should be taken into account when early vitamin A therapy would be adopted in premature infants. Our data provide a basis for understanding the molecular mechanisms underlying the MVA/RA-mediated differentiation and remodeling of DA.
|
GRANTS
|
|---|
This work was supported in part by grants from National Institute of Biomedical Innovation (MF-16, 05-25), Yokohama Foundation for Advanced Medical Science (to U. Yokoyama, T. Akaike, S. Ishida, and Y. Ishikawa), and the Ministry of Education, Science, Sports and Culture of Japan (to S. Minamisawa, Y. Ishikawa, Y. Sato, and S. Ishida), the Miyata Cardiology Research Promotion Funds (to S. Minamisawa), and the Foundation for Growth Science (to S. Minamisawa), the Mother and Child Health Foundation (to S. Minamisawa), and the grant for 2005 Strategic Research Project of Yokohama City University (to S. Minamisawa).
|
ACKNOWLEDGMENTS
|
|---|
We are grateful to Drs. Sylvia M. Evans and Hemal H. Patel at University of California, San Diego, for critical reading and English editing of the manuscript. We thank Takayo Musuda, Remi Haruno, and Kayoko Fujishita for excellent technical assistance.
|
FOOTNOTES
|
|---|
Address for reprint requests and other correspondence: S. Minamisawa, Dept. of Physiology, Yokohama City Univ., 3-9 Fukuura, Kanazawa-ku, Yokohama 236-0004, Japan (e-mail: sminamis{at}med.yokohama-cu.ac.jp)
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).
* U. Yokoyama and Y. Sato contributed equally to this work. 
1 The online version of this article contains supplemental material.
|
REFERENCES
|
|---|
- Blindt R, Zeiffer U, Krott N, Filzmaier K, Voss M, Hanrath P, vom Dahl J, Bosserhoff AK. Upregulation of the cytoskeletal-associated protein Moesin in the neointima of coronary arteries after balloon angioplasty: a new marker of smooth muscle cell migration? Cardiovasc Res 54: 630–639, 2002.[Abstract/Free Full Text]
- Boudreau N, Rabinovitch M. Developmentally regulated changes in extracellular matrix in endothelial and smooth muscle cells in the ductus arteriosus may be related to intimal proliferation. Lab Invest 64: 187–199, 1991.[ISI][Medline]
- Boudreau N, Turley E, Rabinovitch M. Fibronectin, hyaluronan, and a hyaluronan binding protein contribute to increased ductus arteriosus smooth muscle cell migration. Dev Biol 143: 235–247, 1991.[CrossRef][ISI][Medline]
- Cao H, Dronadula N, Rizvi F, Li Q, Srivastava K, Gerthoffer WT, Rao GN. Novel role for STAT-5B in the regulation of Hsp27-FGF-2 axis facilitating thrombin-induced vascular smooth muscle cell growth and motility. Circ Res 98: 913–922, 2006.[Abstract/Free Full Text]
- Clyman RI, Mauray F, Kramer RH. Beta 1 and beta 3 integrins have different roles in the adhesion and migration of vascular smooth muscle cells on extracellular matrix. Exp Cell Res 200: 272–284, 1992.[CrossRef][ISI][Medline]
- Clyman RI, Seidner SR, Kajino H, Roman C, Koch CJ, Ferrara N, Waleh N, Mauray F, Chen YQ, Perkett EA, Quinn T. VEGF regulates remodeling during permanent anatomic closure of the ductus arteriosus. Am J Physiol Regul Integr Comp Physiol 282: R199–R206, 2002.[Abstract/Free Full Text]
- Clyman RI, Tannenbaum J, Chen YQ, Cooper D, Yurchenco PD, Kramer RH, Waleh NS. Ductus arteriosus smooth muscle cell migration on collagen: dependence on laminin and its receptors. J Cell Sci 107: 1007–1018, 1994.[Abstract]
- Coceani F, Olley PM. The control of cardiovascular shunts in the fetal and perinatal period. Can J Physiol Pharmacol 66: 1129–1134, 1988.[ISI][Medline]
- Colbert MC. Retinoids and cardiovascular developmental defects. Cardiovasc Toxicol 2: 25–39, 2002.[CrossRef][Medline]
- Colbert MC, Kirby ML, Robbins J. Endogenous retinoic acid signaling colocalizes with advanced expression of the adult smooth muscle myosin heavy chain isoform during development of the ductus arteriosus. Circ Res 78: 790–798, 1996.[Abstract/Free Full Text]
- Costa M, Barogi S, Socci ND, Angeloni D, Maffei M, Baragatti B, Chiellini C, Grasso E, Coceani F. Gene expression in ductus arteriosus and aorta: comparison of birth and oxygen effects. Physiol Genomics 25: 250–262, 2006.[Abstract/Free Full Text]
- De Reeder EG, Girard N, Poelmann RE, Van Munsteren JC, Patterson DF, Gittenberger-De Groot AC. Hyaluronic acid accumulation and endothelial cell detachment in intimal thickening of the vessel wall. The normal and genetically defective ductus arteriosus. Am J Pathol 132: 574–585, 1988.[Abstract]
- Demary K, Wong L, Liou JS, Faller DV, Spanjaard RA. Redox control of retinoic acid receptor activity: a novel mechanism for retinoic acid resistance in melanoma cells. Endocrinology 142: 2600–2605, 2001.[Abstract/Free Full Text]
- Dousa TP. Cyclic-3',5'-nucleotide phosphodiesterase isozymes in cell biology and pathophysiology of the kidney. Kidney Int 55: 29–62, 1999.[CrossRef][ISI][Medline]
- Herdeg C, Oberhoff M, Baumbach A, Schroeder S, Leitritz M, Blattner A, Siegel-Axel DI, Meisner C, Karsch KR. Effects of local all-trans-retinoic acid delivery on experimental atherosclerosis in the rabbit carotid artery. Cardiovasc Res 57: 544–553, 2003.[Abstract/Free Full Text]
- Loftin CD, Trivedi DB, Tiano HF, Clark JA, Lee CA, Epstein JA, Morham SG, Breyer MD, Nguyen M, Hawkins BM, Goulet JL, Smithies O, Koller BH, Langenbach R. Failure of ductus arteriosus closure and remodeling in neonatal mice deficient in cyclooxygenase-1 and cyclooxygenase-2. Proc Natl Acad Sci USA 98: 1059–1064, 2001.[Abstract/Free Full Text]
- Mason CA, Bigras JL, O'Blenes SB, Zhou B, McIntyre B, Nakamura N, Kaneda Y, Rabinovitch M. Gene transfer in utero biologically engineers a patent ductus arteriosus in lambs by arresting fibronectin-dependent neointimal formation. Nat Med 5: 176–182, 1999.[CrossRef][ISI][Medline]
- Miano JM, Berk BC. Retinoids: versatile biological response modifiers of vascular smooth muscle phenotype. Circ Res 87: 355–362, 2000.[Free Full Text]
- Minamisawa S, Wang Y, Chen J, Ishikawa Y, Chien KR, Matsuoka R. Atrial chamber-specific expression of sarcolipin is regulated during development and hypertrophic remodeling. J Biol Chem 278: 9570–9575, 2003.[Abstract/Free Full Text]
- Momma K, Toyono M, Miyagawa-Tomita S. Accelerated maturation of fetal ductus arteriosus by maternally administered vitamin A in rats. Pediatr Res 43: 629–632, 1998.[ISI][Medline]
- Nakanishi T, Gu H, Hagiwara N, Momma K. Mechanisms of oxygen-induced contraction of ductus arteriosus isolated from the fetal rabbit. Circ Res 72: 1218–1228, 1993.[Abstract/Free Full Text]
- Nguyen M, Camenisch T, Snouwaert JN, Hicks E, Coffman TM, Anderson PA, Malouf NN, Koller BH. The prostaglandin receptor EP4 triggers remodelling of the cardiovascular system at birth. Nature 390: 78–81, 1997.[CrossRef][Medline]
- Nuthakki VK, Fleser PS, Malinzak LE, Seymour ML, Callahan RE, Bendick PJ, Zelenock GB, Shanley CJ. Lysyl oxidase expression in a rat model of arterial balloon injury. J Vasc Surg 40: 123–129, 2004.[CrossRef][ISI][Medline]
- Rabinovitch M. Cell-extracellular matrix interactions in the ductus arteriosus and perinatal pulmonary circulation. Semin Perinatol 20: 531–541, 1996.[CrossRef][ISI][Medline]
- Ravishankar C, Nafday S, Green RS, Kamenir S, Lorber R, Stacewicz-Sapuntzakis M, Bridges ND, Holzman IR, Gelb BD. A trial of vitamin A therapy to facilitate ductal closure in premature infants. J Pediatr 143: 644–648, 2003.[CrossRef][ISI][Medline]
- Saavalainen K, Pasonen-Seppanen S, Dunlop TW, Tammi R, Tammi MI, Carlberg C. The human hyaluronan synthase 2 gene is a primary retinoic acid and epidermal growth factor responding gene. J Biol Chem 280: 14636–14644, 2005.[Abstract/Free Full Text]
- Segi E, Sugimoto Y, Yamasaki A, Aze Y, Oida H, Nishimura T, Murata T, Matsuoka T, Ushikubi F, Hirose M, Tanaka T, Yoshida N, Narumiya S, Ichikawa A. Patent ductus arteriosus and neonatal death in prostaglandin receptor EP4-deficient mice. Biochem Biophys Res Commun 246: 7–12, 1998.[CrossRef][ISI][Medline]
- Shanker G, Sawhney R. Retinoic acid: identification of specific receptors through which it may mediate transcriptional regulation of fibronectin gene in bovine lens epithelial cells. Cell Biol Int 20: 613–619, 1996.[CrossRef][ISI][Medline]
- Shaul PW. Maternal vitamin A administration and the fetal ductus arteriosus. Pediatr Res 49: 744–746, 2001.[ISI][Medline]
- Smith SM, Dickman ED, Power SC, Lancman J. Retinoids and their receptors in vertebrate embryogenesis. J Nutr 128: 467S–470S, 1998.[ISI][Medline]
- Taniguchi T, Azuma H, Okada Y, Naiki H, Hollenberg MD, Muramatsu I. Endothelin-1-endothelin receptor type A mediates closure of rat ductus arteriosus at birth. J Physiol 537: 579–585, 2001.[Abstract/Free Full Text]
- Thebaud B, Michelakis E, Wu XC, Harry G, Hashimoto K, Archer SL. Sildenafil reverses O2 constriction of the rabbit ductus arteriosus by inhibiting type 5 phosphodiesterase and activating BK(Ca) channels. Pediatr Res 52: 19–24, 2002.[ISI][Medline]
- Uemura N, Ohkusa T, Hamano K, Nakagome M, Hori H, Shimizu M, Matsuzaki M, Mochizuki S, Minamisawa S, Ishikawa Y. Down-regulation of sarcolipin mRNA expression in chronic atrial fibrillation. Eur J Clin Invest 34: 723–730, 2004.[CrossRef][ISI][Medline]
- Wilson DP, Saward L, Zahradka P, Cheung PK. Angiotensin II receptor antagonists prevent neointimal proliferation in a porcine coronary artery organ culture model. Cardiovasc Res 42: 761–772, 1999.[Abstract/Free Full Text]
- Wu GR, Jing S, Momma K, Nakanishi T. The effect of vitamin A on contraction of the ductus arteriosus in fetal rat. Pediatr Res 49: 747–754, 2001.[ISI][Medline]
- Yokoyama U, Minamisawa S, Adachi-Akahane S, Akaike T, Naguro I, Funakoshi K, Iwamoto M, Nakagome M, Uemura N, Hori H, Yokota S, Ishikawa Y. Multiple transcripts of Ca2+ channel alpha1-subunits and a novel spliced variant of the 1C-subunit in rat ductus arteriosus. Am J Physiol Heart Circ Physiol 290: H1660–H1670, 2006.[Abstract/Free Full Text]
- Yokoyama U, Minamisawa S, Quan H, Ghatak S, Akaike T, Segi-Nishida E, Iwasaki S, Iwamoto M, Misra S, Tamura K, Hori H, Yokota S, Toole BP, Sugimoto Y, Ishikawa Y. Chronic activation of the prostaglandin receptor EP4 promotes hyaluronan-mediated neointimal formation in the ductus arteriosus. J Clin Invest 116: 3026–3034, 2006.[CrossRef][ISI][Medline]
- Zaina S, Nilsson J. Insulin-like growth factor II and its receptors in atherosclerosis and in conditions predisposing to atherosclerosis. Curr Opin Lipidol 14: 483–489, 2003.[CrossRef][ISI][Medline]
<
TOP
ABSTRACT
INTRODUCTION
