Current Hypertension Reviews, 2011, 7, 217-227 217 -RAMP2 System in Cardiovascular Development and Homeostasis

Takayuki Shindo*, Takayuki Sakurai, Akiko Kamiyoshi and Yuka Ichikawa-Shindo

Department of Organ Regeneration, Shinshu University Graduate School of Medicine, Japan

Abstract: Adrenomedullin (AM), originally identified as a vasodilating peptide, is now recognized to be a pleiotropic vasoactive molecule involved in both the pathogenesis of cardiovascular diseases and circulatory homeostasis. To elucidate the in vivo roles of AM, we have established and analyzed genetically engineered AM mice and its receptor components. Heterozygotes of AM knockout mice (AM+/-) showed blood pressure elevation, severe cardiac hypertrophy, and fibrosis by pressure overload or angiotensin II infusion. On the other hand, vascular-specific AM-overexpressing mice were resistant to neointimal formation by arterial injury. Therefore, endogenous AM exerts protective effects against stress-induced cardiac hypertrophy, fibrosis, and arteriosclerosis, as well as a vasodilating effect. Homozygotes of AM knockout mice (AM-/-) were lethal at mid-gestation with abnormalities in vascular development. This finding first clarified the indispensable roles of AM in vascular development. We also showed that AM possesses novel angiogenic properties not only during development, but also in adults. Based on these observations, there is hope that AM can be used therapeutically. However, AM has a short half-life in the blood stream and its application in chronic disease has limitations. We generated knockout mice of receptor activity- modifying 2 (RAMP2), a small membrane protein that associates with the AM receptor. We found that the important vascular phenotypes of AM knockout mice were reproduced in RAMP2 knockout mice. This shows that RAMP2 is the key determinant of the vascular functions of AM. RAMP2 could be an attractive therapeutic target in cardiovascular diseases. Keywords: Angiogenesis, arteriosclerosis, cardiac hypertrophy, knockout mice, vascular development.

INTRODUCTION roles in the maintenance of metabolic and cardiovascular homeostasis. In contrast, an imbalance between the factors Adrenomedullin (AM) is a 52-amino acid vasodilating may cause the onset of metabolic syndrome and cardio- peptide first identified from human pheochromocytoma in vascular diseases. Plasma AM also increases with obesity 1993 [1]. Since then, AM has been shown to be synthesized and shows a positive correlation with body mass index [15]. by a variety of tissues and cell types [2]. Based on The expression of AM was also detected in the adipose its structural homology and similar vasodilatory effects, tissue and is enhanced with obesity [16]. Furthermore, AM has been classified as a family peptide of calcitonin elevated expression of AM was reported in coronary plaques -related peptide (CGRP). Apart from its vasodilatory obtained from patients with unstable angina [17]. Therefore, effect, AM also exerts diuretic [3] and cardiotonic [4] AM could also be involved in metabolic syndrome. effects, and is involved in the regulation of hormone release [5, 6], inflammation [7], oxidative stress [8, 9] and the Mice with manipulation of the AM gene have served as a proliferation, migration and differentiation of various cell useful tool to elucidate the physiological and patho- types [10-12]. Plasma AM is elevated in the presence of physiological roles of AM in vivo. AM is a member of the such pathological conditions as hypertension, renal failure, calcitonin superfamily and acts via the G protein-coupled and heart failure [13, 14]. Therefore, AM is thought to be seven transmembrane domain receptor, calcitonin-receptor- involved in the pathophysiology of cardiovascular diseases like receptor (CLR) [18]. The specificity of CLR for as well as the regulation of circulatory homeostasis. its ligands is regulated by a group of three receptor- activity-modifying , RAMP1, -2 and -3. This review Recently, various humoral factors have attracted much summarizes the roles of AM and its related molecules, attention among researchers in terms of being a common which have been clarified through the analysis of genetically molecular basis that links the onset of metabolic syndrome engineered mice. with that of atherosclerosis. These factors play important AM IN VASCULAR DEVELOPMENT

*Address correspondence to these authors at the Department of Organ Three independent groups including ours have reported Regeneration, Shinshu University Graduate School of Medicine, Asahi 3-1- the outcome of targeted AM gene disruption [8, 19, 20]. Our 1, Matsumoto, Nagano, 390-8621, Japan; Tel: +81-263-37-2578; strategy to knockout the AM gene was to destroy a 2.4kb Fax: +81-263-37-3437; E-mail: [email protected]

1573-4021/11 $58.00+.00 © 2011 Bentham Science Publishers 218 Current Hypertension Reviews, 2011, Vol. 7, No. 4 Shindo et al. fragment encompassing the 1.3-kb 5’ -flanking region, exons abnormally constricted. These embryos also exhibited 1-3, and part of exon 4 of the AM gene [19]. accumulation of pericardial effusion suggestive of cardiac failure. We concluded that AM-/- embryos are lethal due to Heterozygous knockout mice of AM (AM+/-) were fully disruption of the vasculature and circulatory collapse. viable, and AM levels in various organs of AM+/- decreased to half that of wild-type mice (WT). In contrast, mice At E12.5, before hemorrhagic changes were detectable homozygous for AM null mutation (AM-/-) were embryonic macroscopically, endothelial cells had partially detached lethal, and no embryos could survive beyond the mid-term of from the basement structure in vessels. The endothelial cells gestation. The mortality rate among AM-/- embryos at in AM-/- embryos appeared cuboid in shape, rather than flat embryonic day (E) 13.5 was 83%, and none survived to as in WT, and thus stood out from the wall of the lumen E14.5. The most apparent abnormality in AM-/- embryos at (Fig. 1B). The three-layer structure of the basal membrane E13.5-14.0 was severe hemorrhage, which was observable formed by the lamina rara interna, lamina densa and lamina under the skin and in the lung and liver. Hemorrhage was not rara externa was not clearly recognizable in AM-/- embryos. yet detectable at E12.5-13.0, though the embryos showed These changes may have contributed to the fragility of the abnormal vitelline vessels on the yolk sac (Fig. 1A); vascular wall and subsequent hemorrhage. histological examination showed only the presence of poorly Scanning electron microscopic observation of the developed vitelline vessels. The umbilical artery was corrosion casts of the AM-/- embryonic cervical artery A! +/+! -/-!

B! C! EC! Vascular ! +/+! +/+! Vascular ! +/+! lumen! lumen!

EC! SMC! -/-! -/-! -/-! EC!

EC!

SMC!

Fig. (1). (A) Appearance of E13.0 yolk sac. Well-developed vitelline vessels were detected on yolk sac of wild-type (left) but were poorly developed in AM-/- (right). (B) Transmission electron micrographs and schemata of vascular structure of cervical arteries from E12.5 wild- type (upper panels) and AM-/- (lower panels). EC indicates endothelial cell, SMC smooth muscle cell, and arrows basement membrane of endothelial cells. The endothelial cells in AM-/- embryos were cuboid in shape (rather than flat as in wild-type), stand out from the wall of the lumen, and partially detach from the basement membrane. Scale bars = 2m. (C) Scanning electron micrograph of corrosion casts of cervical arteries from wild-type (upper) and AM-/- (lower) embryos. Numerous holes are detected on the surfaces of the casts of AM-/- artery, which correspond to protruding or detaching endothelial cells. Scale bars = 100m. Cited from the reference [19] with some modifications. Cardiovascular Regulation by Adrenomedullin-RAMP2 Current Hypertension Reviews, 2011, Vol. 7, No. 4 219 showed numerous holes (Fig. 1C), which most likely that CGRP contributes to the regulation of cardiovascular correspond to protruding or detaching endothelial cells. function through inhibitory modulation of sympathetic These findings indicate the indispensable roles of AM in nervous activity. It is noteworthy that the study of knockout vascular morphogenesis during development. mice first clarified the distinctly different roles between AM and CGRP, although both of them had been thought to Calcitonin gene-related peptide (CGRP) is a 37-amino possess similar physiological functions. acid vasoactive peptide produced by tissue-specific alternative splicing of the primary transcript of the ANGIOGENIC FUNCTIONS OF AM IN ADULT calcitonin/CGRP gene [21]. While calcitonin, which controls We have shown that AM possesses novel angiogenic calcium homeostasis, is expressed almost exclusively in the properties not only during development, but also in adults. C-cells of the thyroid gland, CGRP is widely distributed. We used a hind-limb ischemia model to evaluate the Based on its structural homology and similar biological angiogenicity of AM [23]. Following femoral artery actions, AM has been classified as a family peptide of occlusion, autoamputation of ischemic toes was frequently CGRP. observed in control mice, whereas no autoamputation was Calcitonin derives from exons 1-4, whereas CGRP observed in mice treated with AM (50 ng/h using osmotic derives from exons 1-3 and 5-6 from the calcitonin/CGRP pumps) or vascular endothelial growth factor (VEGF) (100 gene. We also generated CGRP knockout mice using a ng i.m.) (Fig. 2A). In laser Doppler perfusion imaging on targeting DNA construct, which replaces exon 5 encoding a day 12 after surgery, control mice showed about 70% blood CGRP specific region [22]. Unlike AM-/-, homozygotes of flow recovery in the ischemic limb, whereas AM- and CGRP knockout mice (CGRP-/-) were alive and apparently VEGF-treated mice showed almost full recovery (ratio of normal. Adult CGRP-/- showed higher blood pressure and blood flow: control mice, 0.73±0.08; AM-treated mice, heart rate compared with wild-type mice. We have shown 0.98±0.05; VEGF-treated mice, 1.07±0.09) (Fig. 2B).

A! B! Day 0! Day 7! Day 12!

Control! Control!

VEGF! VEGF!

AM! AM!

C!

Control!

AM!

Fig. (2). (A) Photographs of legs of mice taken 28 days after unilateral hind-limb ischemia induced by resecting the left femoral artery. Necrosis of ischemic toes (arrow head in the upper panel) was frequently observed in control mice; no such autoamputation was observed in the mice treated with VEGF or AM. (B) Laser Doppler perfusion image showing recovery of blood perfusion in the ischemic hind-limb (arrows). The blood flow recovery is enhanced by either VEGF or AM administration. (C) Laser Doppler perfusion image around the tumor in control and AM-treated mice 10 days after transplantation of sarcoma 180 (S180) cells. Arrows indicate the tumor transplantation site. Cited from the reference [23]. 220 Current Hypertension Reviews, 2011, Vol. 7, No. 4 Shindo et al.

Treatment with AM or VEGF significantly increased the In AMTg and AM+/-, we evaluated hemodynamic number of capillaries in the ischemic limb. parameters such as blood pressure and heart rate, by intraarterial cannulation [19, 29]. Blood pressure was To investigate the role of endogenous AM in blood flow significantly lower in AMTg than in WT littermates (mean recovery, we also subjected AM+/- mice and their wild-type blood pressure: AMTg 109.3±4.7 vs WT 124.4±2.7 mm Hg). littermates to hind-limb ischemia. On day 7 after surgery, No significant change in heart rate accompanying the AM+/- mice showed about 30% less blood flow recovery reduction in blood pressure was observed. To determine the than wild-type mice (blood flow ratios: wild-type, 1.0±0.10; extent to which the reduced blood pressure seen in AMTg AM+/-, 0.7±0.07). These findings indicate that both was due to increased nitric oxide (NO) release, we studied exogenous and endogenous AM exerts novel angiogenic G the effects of N -monomethyl-L-arginine (L-NMMA), a effects that influence recovery of blood flow in ischemic NO synthase inhibitor, on blood pressure. The pressor tissue. response elicited by intravenous injection of L-NMMA was High levels of AM expression have been detected in significantly higher in AMTg (AMTg 21.1±3.3 vs WT various types of cancer cells suggesting AM is involved in 10.7±1.3%); in fact, it offset the difference in blood pressure tumor growth [24-27]; in fact the presence AM is associated between the two groups. Plasma cGMP concentrations were with more aggressive tumor phenotypes in some cancer cell significantly higher in AMTg than in WT mice, which is lines [28]. We then examined the role of AM in tumor indicative of steady-state activation of the NO-cGMP angiogenesis [23]. Ten days after transplantation of sarcoma pathway. 180 (S180) tumor cells into mice, laser Doppler perfusion We also measured blood pressures in AM+/- mice and imaging showed that the blood flow at the transplantation found them to be significantly higher than in WT (mean site was increased in both AM-treated and control mice; blood pressure: AM±128.4 +- 2.2 vs WT 118.7±2.4 mm Hg). however, the flow was much more pronounced in the former (Fig. 2C). In addition, tumor weight and capillary density We next examined the effects of acetylcholine (ACh), around and within the tumors were significantly greater in AM, and AM receptor antagonists AM22-52 and CGRP8-37 AM-treated animals. on the renal perfusion pressure (RPP) of kidneys isolated from AMTg, AM+/- and WT [30]. Baseline RPP was This means that caution should be exercised when significantly lower in AMTg than in AM+/-, and WT mice considering the possible therapeutic applications of AM, as (AM+/- 93.4±4.6, WT 85.8±4.2, AMTg 72.4±2.4 mm Hg). one would not want to promote unfavorable angiogenesis ACh and AM caused a dose-related reduction in RPP, but associated with tumor growth. On the other hand, it means the degree of vasodilatation was smaller in AMTg than that AM antagonists may effectively inhibit tumor growth in AM+/- and WT (% RPP 10-7mol/L ACh: AM+/- 48.1±3.9, through suppression of tumor angiogenesis. AM22-52 is a C- WT 57.5±5.6, AMTg 22.8±4.8%). NG-nitro-L-arginine terminal fragment of AM and a selective AM receptor methyl ester (L-NAME), a NO synthase inhibitor, caused antagonist. Tumor weight and capillary density were actually greater vasoconstriction in AMTg (% RPP 10-4 mol/L L- reduced in both AM22-52-treated and AM+/- mice NAME: AM+/- 33.1±3.3, WT 55.5±7.2, AMTg 152.6±21.2%). compared with control mice. Both AM antagonists increased RPP in AMTg to a greater -6 We also found that these angiogenic properties of AM extent compared with AM+/- and WT (% RPP 10 mol/L are mediated in part by its ability to enhance VEGF CGRP8-37: AM+/- 12.8±2.6, WT 19.4±3.6, AMTg 41.8±8.7%). expression and serine/threonine protein kinase Akt activity. In this report, we also evaluated the renal function and AM could be a useful therapeutic tool for relieving ischemia. histology 24 hours after 45-minute bilateral renal artery Conversely, inhibitors of AM could be useful for the clinical clamping. In mice with ischemic kidneys, serum levels of management of tumor growth. urea nitrogen and renal damage scores showed smaller values in AMTg and greater values in AM+/- mice. BLOOD PRESSURE ALTERATION BY AM GENE However, differences in serum urea nitrogen and renal MANIPULATION damage scores among the 3 groups of mice were not AM was originally identified as a strong vasodilating observed in mice pretreated with L-NAME. Therefore, AM peptide and acute AM administration causes blood pressure plays a role in the physiological regulation of vascular tone depression in animals and humans. Using genetically and AM protects tissues from ischemia/reperfusion injury engineered mice, we analyzed whether chronic alteration of through its NO releasing activity. the AM level can also influence blood pressure. AM IN ARTERIOSCLEROSIS We established transgenic mice (AMTg) overexpressing Elevated expression of AM was reported in coronary AM driven by preproendothelin-1 (PPET-1) promoter. We plaques obtained from patients with unstable angina [17], constructed a transgene containing a 9.2-kb fragment of the suggesting its implications in atherosclerosis. We evaluated murine PPET-1 gene 5’-flanking region that included a 131- the effect of AM on neointimal hyperplasia and athero- bp sequence of exon 1 and a 0.7-kb SV40-derived sequence sclerosis using AMTg [31]. We used a cuff-injury model, in with an intron and poly-A additional signal, between which which a polyethylene tube cuff was placed around the AM cDNA was inserted. In this way, we gained AMTg femoral artery for 4 weeks. The area of neointima was overexpressing AM in a vessel-selective manner [29]. significantly smaller in AMTg mice than in wild littermates Cardiovascular Regulation by Adrenomedullin-RAMP2 Current Hypertension Reviews, 2011, Vol. 7, No. 4 221

(intima/media area ratio: AMTg 0.45±0.14 vs WT 1.31±0.41) cardiac morbidity and mortality [32] , and can be induced by (Fig. 3A). The bromodeoxyuridine labeling index in the mechanical stress as well as by humoral factors such as subendothelial layers, indicative of cell replication, was also angiotensin II (Ang II) [33, 34]. Several studies have significantly smaller in AMTg, indicating that proliferation suggested that application of exogenous AM suppresses the of vascular smooth muscle cells was suppressed by AM development of cardiac hypertrophy and renal damage [35, overexpression. The vasculoprotective effect of AM was 36]. Moreover, mechanical stretch causes AM production in inhibited by chronic administration of L-NAME, which cardiac myocytes [37], while Ang II stimulation causes its suggests that this effect was at least partially mediated by production in cardiac fibroblasts [38], suggesting AM serves NO. as a local paracrine and/or autocrine modulator of cardiac remodeling. We also evaluated the effect of AM on atherosclerosis induced by hypercholesterolemia. We crossbred AMTg mice We examined the effect of endogenous AM on cardiac with ApoE knockout mice (ApoEKO) and fed the mice an hypertrophy and fibrosis using AM+/- mice [39, 40]. Mice atherogenic diet for 2 months. Atheromatous lesions were were subjected to aortic constriction or Ang II infusion. The significantly smaller in ApoEKO/AMTg than in ApoEKO survival rate after aortic constriction was lower among (percent lesion area: ApoEKO/AMTg 12.0±3.9 vs ApoEKO AM+/- mice (Fig. 4A). Although diastolic left ventricular 15.8±2.8%) (Fig. 3B), although the lipid profiles before and dimension (LVDd) was wider in AM+/- mice in the sham- after being fed an atherogenic diet were not affected by AM operated group, the difference was offset by the wall overexpression. thickening after aortic constriction. Ejection fraction (EF) was smaller in AM+/- mice after aortic constriction. Collectively, AM has protective effects on injured or cholesterol-loaded vessels from neointima formation and Cardiac hypertrophy induced by aortic constriction or atherosclerosis, suggesting that AM may have therapeutic Ang II infusion was exacerbated in AM+/- mice (Fig. 4B). In potential in the treatment of atherosclerosis and vascular addition, cardiac myocytes from AM+/- mice showed greater remodeling. cross sectional lengths and greater numbers of TUNEL- positive cells. Perivascular fibrosis was also more severe in AM IN CARDIAC HYPERTROPHY AM+/- mice (Fig. 4C) and greater numbers of PCNA- Cardiac hypertrophy, which is recognized in many positive cells were observed, indicating stress-induced cardiovascular diseases, is an independent risk factor of proliferation of fibroblasts was enhanced.

A B ApoEKO/! WT! AMTg! ApoEKO! AMTg! Sham!

14 d!

28 d!

Fig. (3). (A) Evaluation of neointima formation by cuff-injury, in which a polyethylene tube cuff is placed around a femoral artery. Photomicrographs show cross sections of the femoral arteries after sham operation and 14 days or 28 days after cuff injury. Sections are stained by elastica van Gieson staining. Neointima formation is reduced in vascular-specific AM overexpressing mice (AMTg) compared with wild-type mice (WT). (B) Hypercholesterolemia-induced atherosclerotic lesions. Photographs show en face atherosclerotic lesions in the aorta from ApoE kncockout mice (APOEKO) and ApoE knockout/AM-overexpressing mice (ApoEKO/AMTg). Atherosclerotic lesions are visualized by staining with Sudan IV. Cited from the reference [31]. 222 Current Hypertension Reviews, 2011, Vol. 7, No. 4 Shindo et al. B 28 days after constriction +/+ +/-

A

Survival rate after constriction (%) 100 14 days after Ang II infusion +/+ +/+ +/- 80 * 60 +/- C Constriction Ang II infusion 40 +/+

20

0 0 7 14 21 28 +/- Days after constriction

Fig. (4). (A) Survival rate after aortic constriction. Percentages of surviving heterozygous knockout mice of AM+/- and wild-type mice+/+ are plotted. The survival rate after aortic constriction was lower among AM+/- mice. (B) Transverse sections of the heart at the papillary muscle level after 28 days of aortic constriction or 14 days of Ang II infusion in heterozygous knockout mice of AM+/- and wild-type mice+/+. Cardiac hypertrophy was exacerbated in AM+/- mice. (C) Accumulation of perivascular collagen (blue-stained in Masson trichrome staining) in the heart of heterozygous knockout mice of AM+/- and wild-type mice+/+. Perivascular fibrosis is exacerbated in AM+/- mice. Cited from the reference [40].

After 28 days of aortic constriction, expression of in AM+/- mice. By contrast, the PKC inhibitor H7 angiotensinogen, ACE, TGF- and collagen type I was suppressed ERK activation in both wild-type and AM+/- upregulated in the heart, and the effect was more pronounced mice, and the inhibitory effect was more pronounced in in AM+/- mice. At the same time, expression of SERCA-2 AM+/- mice, which suggests PKC activity is enhanced by was downregulated to a greater degree in AM+/- mice. Ang II in AM+/- mice. Thus, AM appears to inhibit the development of cardiac hypertrophy, at least in part, by When Ang II was administered to the cardiac myocytes suppressing ERK activation via inhibition of PKC. from AM+/- mice, we observed enlargement of cardiac myocytes that was accompanied by increases in protein Taken together, these results suggest that endogenous synthesis and ANP gene expression. Ang II stimulation also AM serves as an autocrine and/or a paracrine factor exerting enhanced proliferation of fibroblasts isolated from AM+/- a number of crucial protective effects when the cardio- mice and enhanced collagen type I gene expression. vascular system comes under stress. We also analyzed the mechanism by which AM exerts its AM-RAMP2 SYSTEM anti-hypertrophic effects. The Ras-ERKs pathway plays a Based on the observation of phenotypes from genetically key role in the induction of early response like c-fos engineered animal models of AM, AM is thought to be a key during pressure-overload [41]. We found ERKs to be factor involved in various aspects of cardiovascular diseases, activated within 30 min after the onset of aortic constriction, and it is hoped it may be of therapeutic use. As with other and the level of activation was much greater in AM+/- mice. growth factors, the clinical applicability of AM has two Treatment with 1 M Ang II increased ERK phosphorylation serious limitations: AM is a peptide with a short half-life in in both wild-type and AM+/- mice, but the effect was greater the blood stream, and the cost of the recombinant protein Cardiovascular Regulation by Adrenomedullin-RAMP2 Current Hypertension Reviews, 2011, Vol. 7, No. 4 223 makes its use in the treatment of chronic diseases the signals of multiple ligands, though the precise impractical. This prompted us to focus on the AM receptor mechanism remains largely unknown. system. We hypothesized that not only the receptor-ligand AM-signaling is regulated by a unique control system specificity, but also the diversity of the biological activities [18, 42-44]. The AM receptor is a seven-transmembrane of AM reflects its novel regulation by RAMPs. We analyzed domain G-protein coupled receptor (GPCR) named the expression of AM and its related genes during mid- calcitonin-receptor like receptor (CLR). CLR associates with gestational development (E11.5-E14.5), the stage at which an accessory protein, receptor activity modifying protein AM-/- embryos typically die. We found that in wild-type (RAMP), which is comprised of about 160 amino acids and mice, AM, CLR, RAMP2 and RAMP3 all continued to be includes a single membrane-spanning domain. Three RAMP expressed at mid-gestation (Fig. 5A). Using in situ subtypes have been identified. By interacting with RAMP1, hybridization, we detected AM expression in the vascular CLR acquires a high affinity for CGRP, whereas by system and found that, among the RAMPs, only RAMP2 interacting with either RAMP2 or -3, CLR acquires a high was specifically expressed in the vasculature at that stage affinity for AM. This novel system enables CLR to transduce (Fig. 5B). We therefore speculated that it is RAMP2, which A

CLR AM RAMP2 RAMP3 1.6 1.6 1.6 1.6 1.2 1.2 1.2 1.2 0.8 0.8 0.8 0.8

expression 0.4 0.4 0.4 0.4 Relative gene 0 0 0 0 11.512.513.514.5(E) 11.512.513.514.5(E) 11.512.513.514.5(E) 11.512.513.514.5(E)

B in situ hybridization of RAMP2 (E12.5)

sense antisense

224 Current Hypertension Reviews, 2011, Vol. 7, No. 4 Shindo et al.

Fig. (5). contd…. G Vitelline artery Wild RAMP2-/- E Vascular lumen Vascular lumen E

H Aorta Wild RAMP2-/-

Vascular lumen Vascular lumen

Fig. (5). (A) Real-time PCR analysis of gene expression during embryonic day (E)11.5-14.5. The expression levels of each gene are shown as ratios relative to those at E11.5. AM, CLR and RAMPs were still being expressed during mid-gestation. (B) In situ hybridization of RAMP2 in mouse embryo. Section of artery from wild-type embryo collected at mid-gestation (E12.5) was used. Intense RAMP2 expression was detected in the vascular ECs. Scale bars = 20m. (C) Appearance of yolk sac and vitelline arteries of E13.5 wild-type (left) and RAMP2-/- (right) embryos. (D) Whole mount immunofluorescence staining of CD31 using yolk sacs from E13.5 wild-type (left) and RAMP2-/- (right) embryos. Vitelline arteries were well-developed in wild-type mice but poorly developed in RAMP2-/- embryos. (E) Front views of wild-type (left) and RAMP2-/- (right) embryos at mid-gestation. Some RAMP2-/- embryos showed severe systemic edema. (F) Hemorrhagic changes observed in the liver section of RAMP2-/- (right) embryos at mid-gestation. Scale bars = 50m. (G) Transmission electron micrographs of vitelline arteries from E12.5 RAMP2-/- (right) and wild-type (left) embryos. RAMP2-/- showed the detachment of ECs (shown as E) from basement membrane (arrows). Scale bars = 2μm. (H) Transmission electron micrographs of aortas from E12.5 RAMP2-/- (right) and wild-type (left) embryos. RAMP2-/- showed thinner (compare the two headed arrows) and rougher smooth muscle cell layer. Scale bars = 25μm. Cited from the reference [45]. determines the function of AM during vascular development, expression of AM was upregulated by more than 5-fold in and proceeded to generate RAMP2-specific knockout mice RAMP2-/- mice, presumably as a compensatory response. to directly assess the functions of the AM-RAMP2 system in At E13.5, well-developed vitelline arteries were detected in vivo. on the yolk sacs of wild-type embryos, whereas RAMP2-/- RAMP2 KNOCKOUT MICE embryos had only poorly developed vitelline arteries (Fig. 5C). Histological examination revealed that the vitelline The targeting vector for RAMP2 knockout mice was arteries from RAMP2-/- embryos were smaller than those constructed to insert loxP sites encompassing exons 2 to 4 of from wild-type embryos and appeared disorganized (Fig. RAMP2 and the neomycin-resistance gene [45]. After 5D). The fact that these phenotypes resemble those of AM-/- obtaining heterozygotic floxed RAMP2 mice, we crossbred embryos [19] shows that deletion of RAMP2 is sufficient to them with Cre-deleter mice to delete exon 2 to 4 of the reproduce the major phenotypes of the vascular abnormality RAMP2 gene. seen in AM-/- mice. No RAMP2 homozygous knockout (RAMP2-/-) As for the embryos, the most apparent finding in newborns were obtained, and analysis of the embryos from RAMP2-/- mice was severe systemic edema (Fig. 5E) and timed RAMP2+/- intercrosses showed that the RAMP2-/- some had bleeding that was observable under the skin and genotype was lethal at mid-gestation. The mortality rate within organs (Fig. 5F). These phenotypes were also among RAMP2-/- embryos was 13% at E13.5, 92% at E14.5, observed in AM-/- embryos [19], though RAMP2-/- mice and 100% at E15.5. The most lethal stage (E13.5-14.5) was showed much more severe systemic edema than AM-/- mice. nearly identical to that of the AM-/- genotype. Expression of RAMP3 did not differ in RAMP2-/- and wild-type mice, At E12.5, electron microscopic observation of the which means there is no functional redundancy between endothelial cells of the vitelline arteries in RAMP2-/- RAMP2 and RAMP3 during development. Moreover, the embryos revealed deformity and detachment from the Cardiovascular Regulation by Adrenomedullin-RAMP2 Current Hypertension Reviews, 2011, Vol. 7, No. 4 225

A B Capillary area Relative area (ratio) of tight-junction 20 RAMP2 O/E RAMP2 O/E ** Control ** ** 1.2 Control ** ** 16 ** ** ** 1.0

12 0.8 ** ** ** ** 0.6 8 ** ** ** ** ** * 0.4 4 0.2

0 0 H2O2 (-) H2O2 (+) 1 3 5 7 9 11 13 15 17 19 day

Fig. (6). (A) Evaluation of capillary formation on Matrigel by EAhy926 endothelial cells, which overexpress RAMP2 (RAMP2O/E). Cells are cultured in 24-well culture plates coated with Matrigel in medium containing 10-7 M AM. Capillary area is shown relative to the cell surface area on day 1. RAMP2O/E cells exhibit much greater angiogenesis than control cells. (B) Comparison of the tight-junction formation between EAhy926 endothelial cells, which overexpress RAMP2 (RAMP2O/E) and control cells. Tight junctions were better preserved after H2O2-treatment in RAMP2O/E than control endothelial cells. Cited from the reference [45]. basement membrane (Fig. 5G). In addition, there was of human umbilical vein endothelial cells (HUVEC) and line abnormal thinning of the arterial walls in RAMP2-/- A 549/8 lung carcinoma cells. The resultant RAMP2O/E embryos (Fig. 5H). We also found that expression of cells are known to form capillary-like tubes on Matrigel [46, molecules involved in cell adhesion was altered in arteries 47]. We observed that RAMP2O/E cells showed much from RAMP2-/- mice. In particular, expression of VE- greater capillary formation than control cells in Matrigel cadherin, claudin 5 and type IV collagen was diminished. assays (Fig. 6A), clearly demonstrating that upregulation of These molecules are all important for the composition of the AM-RAMP2 system enhances angiogenesis. tight junctions, adherence junctions and the basement We then assessed endothelial barrier function by assaying membrane of vascular endothelial cells, and abnormalities vascular permeability in in vitro. Cells were seeded onto involving them lead to paracellular leakage from the vascular semipermeable membranes in permeability chambers, after lumen, which likely explains the severe edema seen in which the passage of FITC-dextran through confluent RAMP2-/- mice. endothelial cell monolayers was monitored. Monolayers of We found that neovascularization was diminished and RAMP2O/E cells were significantly less permeable than vascular permeability was increased not only during those of control cells, suggesting upregulation of the AM- development, but also in adult RAMP2+/- mice, which RAMP2 system also enhances vascular barrier function and showed reduced expression of RAMP2. We also found that reduces permeability. We hypothesized that the reduced RAMP2+/- mice have higher blood pressure than their wild- permeability reflected the firmer structure of the tight type littermates, which confirms that RAMP2 continues to junctions formed by RAMP2O/E cells. We treated be a crucial determinant of vascular function of AM in the RAMP2O/E and control cells with H2O2 (0.5 mM), which adult. Interestingly, we also found that the edema developed leads to formation of intercellular gaps and reduced tight by RAMP2+/- mice in various disease models was more junction formation between endothelial cells. After the severe than in wild-type mice, suggesting the AM-RAMP2 treatment, the tight junction structure was better preserved in system could be an attractive therapeutic target for treating RAMP2O/E than control cells (Fig. 6B). the edema often associated with vascular regenerative We also found that expression of eNOS, VEGF, and therapies, brain trauma and infarction. claudin 5 was upregulated in the RAMP2O/E cells and that ANGIOGENESIS REGULATED BY AM-RAMP2 treatment with the PI3K inhibitor LY294002 or a PKA SYSTEM inhibitor blocked those effects. Thus signaling via a PI3K and PKA-dependent pathway appears to play a key role in To analyze the mechanisms underlying the angiogenesis AM-RAMP2 mediated angiogenesis and vascular stability. and vascular stability mediated by the RAMP2, we next generated an endothelial cell line that stably overexpressed By contrast, endothelial cell lines overexpressing RAMP2 (RAMP2O/E). This was accomplished utilizing RAMP3 (RAMP3O/E) did not show either enhanced EAhy926 endothelial cells, an immortal, clonally pure, angiogenesis or improved vascular stability, although human endothelial cell line obtained through hybridization RAMP3 has been shown to work with CLR to function as 226 Current Hypertension Reviews, 2011, Vol. 7, No. 4 Shindo et al. another AM receptor [48]. Furthermore, our finding that be achieved by targeting RAMP2 than by targeting CLR, RAMP3 is expressed at wild-type levels in RAMP2-/- mice which can also function as a receptor for other ligands confirms that RAMP3 cannot compensate for the absence of including CGRP. In that context, our findings provide a clear RAMP2 during vascular development. Consistent with the basis for the development of drugs to modulate RAMP2 and, distinctly different physiological roles played by RAMP2 thereby, the vascular effects of AM. and -3, RAMP3-/- mice live apparently normally until old age [49]. In addition, whereas RAMP2 and CLR are CONFLICT OF INTERESTS downregulated in an endotoxemia model, RAMP3 is Declared none. markedly upregulated [50], and it has been suggested that RAMP3 may be involved in post-endocytic receptor ACKNOWLEDGMENT trafficking, as it presents a PDZ type I domain [51]. Declared none. Table 1. summarizes the phenotypical differences among REFERENCES knockout mice of AM and its receptor components. [1] Kitamura K, Kangawa K, Kawamoto M, et al. Adrenomedullin: a novel hypotensive peptide isolated from human pheochromocytoma. Table 1. Phenotypical Differences among Knockout Mice of Biochem Biophys Res Commun 1993; 192: 553-560. AM and Its Receptor Components [2] Ichiki Y, Kitamura K, Kangawa K, Kawamoto M, Matsuo H, Eto T. Distribution and characterization of immunoreactive adrenomedullin in human tissue and plasma. FEBS Lett 1994; 338: Knockout Mice Major Phenotypes 6-10. [3] Jougasaki M, Wei CM, Aarhus LL, Heublein DM, Sandberg SM, Burnett JC, Jr. Renal localization and actions of adrenomedullin: a AM-/- Embryonic lethal (E13.5-14.5) natriuretic peptide. Am J Physiol 1995; 268: F657-663. Abnormalities of cardiovascular development [4] Nishikimi T, Matsuoka H. Cardiac adrenomedullin: its role in cardiac hypertrophy and heart failure. Curr Med Chem Cardiovasc RAMP1-/- Alive Hematol Agents 2005; 3: 231-242. Hypertension, Dysregulated cytokine production [5] Samson WK, Murphy T, Schell DA. A novel vasoactive peptide, adrenomedullin, inhibits pituitary adrenocorticotropin release. RAMP2-/- Embryonic lethal (E14.5-15.5) Endocrinology 1995;136:2349-2352. Abnormalities of cardiovascular development [6] Petrie MC, Hillier C, Morton JJ, McMurray JJ. Adrenomedullin selectively inhibits angiotensin II-induced aldosterone secretion in RAMP3-/- Alive humans. J Hypertens 2000; 18: 61-64. Decreased body weight in aged mice [7] Isumi Y, Kubo A, Katafuchi T, Kangawa K, Minamino N. Adrenomedullin suppresses -1beta-induced tumor CLR-/- Embryonic lethal (E13.5-14.5) necrosis factor-alpha production in Swiss 3T3 cells. FEBS Lett Abnormalities of cardiovascular development 1999; 463: 110-114. [8] Shimosawa T, Shibagaki Y, Ishibashi K, et al. Adrenomedullin, an endogenous peptide, counteracts cardiovascular damage. Circulation 2002; 105: 106-111. CONCLUSION [9] Shimosawa T, Ogihara T, Matsui H, Asano T, Ando K, Fujita T. Deficiency of adrenomedullin induces resistance by These observations of a genetically engineered model increasing oxidative stress. Hypertension 2003; 41: 1080-1085. help us to understand the novel functions of AM in in vivo. [10] Kano H, Kohno M, Yasunari K, et al. Adrenomedullin as a novel antiproliferative factor of vascular smooth muscle cells. J AM could not be categorized as a simple vasodilating Hypertens 1996;14: 209-213. peptide since it also has protective effects on organ and [11] Miyashita K, Itoh H, Sawada N, et al. Adrenomedullin promotes vasculature at postnatal stages as well as an important role in proliferation and migration of cultured endothelial cells. Hypertens morphogenesis during development. Because of its wide Res 2003; 26 Suppl: S93-98. [12] Iwasaki H, Eguchi S, Shichiri M, Marumo F, Hirata Y. range of bioactivity, AM attracts much attention for clinical Adrenomedullin as a novel growth-promoting factor for cultured applications. However, AM is a peptide with a short half-life vascular smooth muscle cells: role of tyrosine kinase-mediated in the blood stream and its use in the treatment of chronic mitogen-activated protein kinase activation. Endocrinology 1998; diseases has limitations. 139: 3432-3441. [13] Kobayashi K, Kitamura K, Etoh T, et al. Increased plasma Signal transduction via GPCRs and the regulation of their adrenomedullin levels in chronic congestive heart failure. Am Heart J 1996; 131: 994-998. function has long attracted the interest of many researchers. [14] Ishimitsu T, Nishikimi T, Saito Y, et al. Plasma levels of Indeed, about 40% of the drugs in clinical use today target adrenomedullin, a newly identified hypotensive peptide, in patients GPCRs. As we showed, it is noteworthy that we are able to with hypertension and renal failure. J Clin Invest 1994; 94: 2158-2161. modulate the vascular function of AM by modulating [15] Kato J, Kitamura K, Uemura T, Kuwasako K, Kita T, Kangawa K, RAMP2. Thus, RAMP2 could be a therapeutic target via Eto T. Plasma levels of adrenomedullin and atrial and brain which to manipulate the vascular functions of AM. Because natriuretic peptides in the general population: their relations to age RAMP2 is a low molecular weight protein, structural and pulse pressure. Hypertens Res 2002;25:887-892. [16] Nambu T, Arai H, Komatsu Y, et al. Expression of the analysis and the synthesis of specific agonists or antagonists adrenomedullin gene in adipose tissue. Regul Pept 2005; 132: are much more realistic compared with AM’s receptor CLR, 17-22. which belongs to seven-transmembrane domain GPCRs. [17] Hojo Y, Ikeda U, Mizuno O, Katsuki TA, Shimada K. Adrenomedullin expression in coronary circulation after stent Moreover, because RAMP2 determines the vascular functions implantation as a prognostic factor for restenosis. Int J Cardiovasc of AM, it would be expected that greater specificity would Intervent 2003; 5: 190-194. Cardiovascular Regulation by Adrenomedullin-RAMP2 Current Hypertension Reviews, 2011, Vol. 7, No. 4 227

[18] McLatchie LM, Fraser NJ, Main MJ, et al. RAMPs regulate the [34] Baker KM, Aceto JF. Angiotensin II stimulation of protein transport and ligand specificity of the calcitonin-receptor-like synthesis and cell growth in chick heart cells. Am J Physiol 1990; receptor. Nature 1998; 393: 333-339. 259: H610-618. [19] Shindo T, Kurihara Y, Nishimatsu H, et al. Vascular abnormalities [35] Dobrzynski E, Wang C, Chao J, Chao L. Adrenomedullin gene and elevated blood pressure in mice lacking adrenomedullin gene. delivery attenuates hypertension, cardiac remodeling, and renal Circulation 2001; 104: 1964-1971. injury in deoxycorticosterone acetate-salt hypertensive rats. [20] Caron KM, Smithies O. Extreme hydrops fetalis and cardiovascular Hypertension 2000; 36: 995-1001. abnormalities in mice lacking a functional Adrenomedullin gene. [36] Nishikimi T, Mori Y, Kobayashi N, et al. Renoprotective effect of Proc Natl Acad Sci USA 2001; 98: 615-619. chronic adrenomedullin infusion in Dahl salt-sensitive rats. [21] Rosenfeld MG, Mermod JJ, Amara SG, et al. Production of a novel Hypertension 2002; 39: 1077-1082. neuropeptide encoded by the calcitonin gene via tissue-specific [37] Tsuruda T, Kato J, Kitamura K, et al. Enhanced adrenomedullin RNA processing. Nature 1983; 304: 129-135. production by mechanical stretching in cultured rat cardio- [22] Oh-hashi Y, Shindo T, Kurihara Y, et al. Elevated sympathetic myocytes. Hypertension 2000; 35: 1210-1214. nervous activity in mice deficient in alphaCGRP. Circ Res 2001; [38] Tsuruda T, Kato J, Kitamura K, et al. An autocrine or a paracrine 89: 983-990. role of adrenomedullin in modulating cardiac fibroblast growth. [23] Iimuro S, Shindo T, Moriyama N, et al. Angiogenic effects of Cardiovasc Res 1999; 43: 958-967. adrenomedullin in ischemia and tumor growth. Circ Res 2004; 95: [39] Niu P, Shindo T, Iwata H, et al. Accelerated cardiac hypertrophy 415-423. and renal damage induced by angiotensin II in adrenomedullin [24] Martinez A, Vos M, Guedez L, et al. The effects of adrenomedullin knockout mice. Hypertens Res 2003; 26: 731-736. overexpression in breast tumor cells. J Natl Cancer Inst 2002; 94: [40] Niu P, Shindo T, Iwata H, et al. Protective effects of endogenous 1226-1237. adrenomedullin on cardiac hypertrophy, fibrosis, and renal damage. [25] Li Z, Takeuchi S, Otani T, Maruo T. Implications of adrenomedullin Circulation 2004; 109: 1789-1794. expression in the invasion of squamous cell carcinoma of the [41] Aikawa R, Komuro I, Nagai R, Yazaki Y. Rho plays an important uterine cervix. Int J Clin Oncol 2001; 6: 263-270. role in angiotensin II-induced hypertrophic responses in cardiac [26] Forneris M, Gottardo L, Albertin G, Malendowicz LK, Nussdorfer myocytes. Mol Cell Biochem 2000; 212: 177-182. GG. Expression and function of adrenomedullin and its receptors in [42] Kuwasako K, Cao YN, Nagoshi Y, Kitamura K, Eto T. Conn's adenoma cells. Int J Mol Med 2001; 8: 675-679. Adrenomedullin receptors: pharmacological features and possible [27] Rocchi P, Boudouresque F, Zamora AJ, et al. Expression of pathophysiological roles. Peptides 2004; 25: 2003-2012. adrenomedullin and peptide amidation activity in human prostate [43] Parameswaran N, Spielman WS. RAMPs: The past, present and cancer and in human prostate cancer cell lines. Cancer Res 2001; future. Trends Biochem Sci 2006; 31: 631-638. 61: 1196-1206. [44] Morfis M, Christopoulos A, Sexton PM. RAMPs: 5 years on, [28] Hata K, Takebayashi Y, Akiba S, et al. Expression of the where to now? Trends Pharmacol Sci 2003; 24: 596-601. adrenomedullin gene in epithelial ovarian cancer. Mol Hum Reprod [45] Ichikawa-Shindo Y, Sakurai T, Kamiyoshi A, et al. The GPCR 2000; 6: 867-872. modulator protein RAMP2 is essential for angiogenesis and [29] Shindo T, Kurihara H, Maemura K, et al. Hypotension and vascular integrity. J Clin Invest 2008; 118: 29-39. resistance to lipopolysaccharide-induced shock in transgenic mice [46] Bauer J, Margolis M, Schreiner C, et al. In vitro model of overexpressing adrenomedullin in their vasculature. Circulation angiogenesis using a human endothelium-derived permanent cell 2000; 101: 2309-2316. line: contributions of induced gene expression, G-proteins, and [30] Nishimatsu H, Hirata Y, Shindo T, et al. Role of endogenous integrins. J Cell Physiol 1992; 153: 437-449. adrenomedullin in the regulation of vascular tone and ischemic [47] Edgell CJ, Curiel DT, Hu PC, Marr HS. Efficient gene transfer to renal injury: studies on transgenic/knockout mice of adrenomedullin human endothelial cells using DNA complexed to adenovirus gene. Circ Res 2002; 90: 657-663. particles. Biotechniques 1998; 25: 264-268, 270-262. [31] Imai Y, Shindo T, Maemura K, et al. Resistance to neointimal [48] Sexton PM, Albiston A, Morfis M, Tilakaratne N. Receptor activity hyperplasia and fatty streak formation in mice with adrenomedullin modifying proteins. Cell Signal 2001; 13: 73-83. overexpression. Arterioscler Thromb Vasc Biol 2002; 22: 1310- [49] Dackor R, Fritz-Six K, Smithies O, Caron K. Receptor activity- 1315. modifying proteins 2 and 3 have distinct physiological functions [32] Levy D, Garrison RJ, Savage DD, Kannel WB, Castelli WP. from embryogenesis to old age. J Biol Chem 2007; 282: 18094-18099. Prognostic implications of echocardiographically determined left [50] Ono Y, Okano I, Kojima M, Okada K, Kangawa K. Decreased ventricular mass in the Framingham Heart Study. N Engl J Med gene expression of adrenomedullin receptor in mouse lungs during 1990; 322: 1561-1566. sepsis. Biochem Biophys Res Commun 2000; 271: 197-202. [33] Komuro I, Katoh Y, Kaida T, et al. Mechanical loading stimulates [51] Bomberger JM, Parameswaran N, Hall CS, Aiyar N, Spielman WS. cell hypertrophy and specific gene expression in cultured rat Novel function for receptor activity-modifying proteins (RAMPs) cardiac myocytes. Possible role of protein kinase C activation. J in post-endocytic receptor trafficking. J Biol Chem 2005; 280: Biol Chem 1991; 266: 1265-1268. 9297-9307.

Received: July 10, 2011 Revised: September 07, 2011 Accepted: September 12, 2011