| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1,1,2
,2
* State Key Laboratory of Cancer Biology, Department of Medical Genetics and Developmental Biology;
Department of Ophthalmology, Xijing Hospital; and
Department of Hematology, Tangdu Hospital, Fourth Military Medical University, Xi’an, China
2Correspondence: H.H., Department of Medical Genetics and Developmental Biology, Fourth Military Medical University, Chang-Le Xi St. #17, Xi’an 710032, China. E-mail: huahan{at}fmmu.edu.cn; Y.S.W., Department of Ophthalmology, Xijing Hospital, Fourth Military Medical University, Xi’an 710032, China. E-mail: wangys{at}fmmu.edu.cn.com
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(RBP-J) in endothelial cells strikingly induced spontaneous angiogenesis in multiple tissues, including retina and cornea, as well as in internal organs, such as liver and lung. In a choroidal neovascularization model, which mimics the angiogenic process in tumor growth and age-related macular degeneration, RBP-J deficiency induced a more intensive angiogenic response to injury. This could be transmitted by bone marrow, indicating that RBP-J could modulate bone marrow-derived endothelial progenitor cells in adult angiogenesis. In addition, in the absence of RBP-J, proliferation of endothelial cells increased significantly, leading to accumulative vessel outgrowth. These findings suggest that in adults RBP-J-mediated Notch signaling may play an essential role in the maintenance of vascular homeostasis by repressing endothelial cell proliferation.—Dou, G.-R., Wang, Y.-C., Hu, X.-B., Hou, L.-H., Wang, C.-M., Xu, J.-F., Wang, Y.-S., Liang, Y.-M., Yao, L.-B., Yang, A.-G., Han, H. RBP-J, the transcription factor downstream of Notch receptors, is essential for the maintenance of vascular homeostasis in adult mice.
Key Words: angiogenesis • Notch signaling • endothelial cells • choroidal neovascularization
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. However, in some inherited abnormalities or in response to pathological stimuli such as trauma or malignant growth, mature ECs regain the ability of proliferating and remodeling, leading to neovascularization in adults. Notably, it is widely accepted at present that vasculogenesis also contributes to postnatal neovascularization through endothelial progenitor cells (EPCs; ref. 2
). In many diseases such as ocular neovascularization and tumors, deteriorating destabilization of vascular homeostasis acts as an etiological clue or contributes to disease progression. Therefore, it is critical to understand the potential mechanisms controlling vascular homeostasis and angiogenesis/vasculogenesis in adults.
A number of signaling pathways are involved in the control of vessel growth and stability (3)
. Notch signaling, an evolutionarily conserved pathway governing cell fate specification, plays a pivotal role in embryonic vascular development (4)
. Notch ligands and receptors such as Delta-like (Dll) 4, Jagged1, Jagged2, Notch1, Notch2, and Notch4 are expressed by vascular endothelium and/or supporting cells (5
6
7)
. Genetic inactivation of these ligands and receptors, or of other key components of the Notch pathway such as Hey1 (Hesr1) and Hey2 (Hesr2), leads to embryonic lethality in mice due to failures in vessel formation (9
10
11
12
13
14
15
16
17)
. Cellular analyses have shown that in the absence of Notch signaling, vessel remodeling, the key step in angiogenesis, turns out to be defective, with embryonic vasculogenesis appearing intact. Notch signaling acts cell autonomously in ECs (18)
. Recently, by using a zebra fish model and Dll4 haploinsufficient mice, several groups (19
, 20)
have demonstrated that Notch signaling regulates angiogenesis during embryonic and early postnatal stages by restricting endothelial tip cell specification and behaviors. In addition, Notch signaling is essential for the establishment of arteriovenous identity. Endothelial expression of constitutively active Notch4 in adult endothelium causes artery-vein shunting and subsequent arteriovenous malformations (8)
. In summary, Notch signaling participates in embryonic angiogenesis by establishing an arteriovenous identity and regulating endothelial tip cell specification and proliferation, leading to vascular remodeling (21)
.
Although the Notch pathway is crucially implicated in embryonic vascularization, its function and mechanisms in the adult vascular system remain largely uninvestigated. The gene responsible for cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL), a degenerative vascular disease characterized by recurrent subcortical ischemia and vascular dementia, has been mapped to Notch3, suggesting a noticeable role of Notch signaling in adult vascular stability (22)
. Similarly, in Allagilles syndrome, haploinsufficiency of Jagged1 leads to inherited vascular anomalies (23)
. Preferential expression of Notch pathway components in tumor ECs implicates their significance in pathological angiogenesis in adults (24)
. This has been further supported by the reported results that blockade of Dll4 increases EC proliferation and tumor vascular density but inhibits tumor growth due to poor function of new vessels and studies taking Notch signaling as a new target of cancer therapy have been inspired ever since (25)
. Furthermore, in vitro evidence has proven that Notch signaling may be involved in the transition of ECs from the state of proliferation to quiescence, a state recognized as differentiation-associated cell-cycle arrest, as well as vessel remodeling (20
, 26)
. Also by using in vitro cultured ECs, Taylor et al. (27)
have shown that Notch signaling is active in EC contact inhibition by providing negative signals on proliferation. Despite the above-mentioned results, the role of Notch signaling in maintaining the vasculature of adults is still unclear.
Recombination signal-binding protein J (RBP-J), also named C-promoter binding factor1 (CBF1) in mammals and suppressor of hairless [Su(H)] in Drosophila melanogaster, is a DNA-binding protein recognizing a consensus sequence C(T)GTGGGAA and is one member of the CSL family of transcription factors. RBP-J primarily acts as a transcription repressor of promoters possessing RBP-J-binding sites by recruiting other corepressors. The most important function of RBP-J, however, is to mediate signals from Notch receptors (28)
. On ligand binding, Notch signaling is initiated by 
-secretase-mediated proteolytic cleavage and liberation of the Notch intracellular domain (NICD). NICD subsequently translocates into the nucleus where the interaction between NICD and RBP-J occurs, through which NICD activates the transcription of downstream genes. Interestingly, knockout of RBP-J resulted in embryonic failure with severe vascular phenotypes resembling those of the Notch1 Notch4 double mutants (29)
. Vascular defects in endothelial-specific knockout of RBP-J embryos were more profound than those in embryos lacking Dll4, suggesting that additional Notch ligands may play important roles during early vascular development (30
, 31)
.
In the present study, the function of RBP-J in adult angiogenesis was investigated using Cre-loxP-mediated conditional knockout mice (32)
. The results have demonstrated that, when RBP-J was disrupted in ECs of adult mice, spontaneous angiogenesis occurred in multiple organs. In addition, loss of RBP-J in adults markedly enhanced neovascular outgrowth from the existing vessel network in vivo, as well as injury-induced angiogenesis. At the cellular level, RBP-J-deficient ECs exhibited hyperproliferating phenotypes. Finally, deletion of RBP-J in adult ECs led not only to the down-regulation of a series of downstream genes such as Hes1, Hesr1, and VEGFR1 but also to the up-regulation of VEGFR2 and p21WAF1. To summarize, our results suggested that in adults RBP-J-mediated Notch signaling may be a key element in the maintenance of vascular homeostasis.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. ROSA26R mice were obtained from Jackson Laboratory (Bar Harbor, ME, USA). All mice were maintained in specific pathogen-free conditions. Mice used in the choroidal neovascularization (CNV) model were handled in accordance with guidelines from the Association for Research in Vision and Ophthalmology. All animal experiments were approved by the Animal Experiment Administration Committee of the Fourth Military Medical University.
Histology
Tissues were fixed in Bouin’s fixative or 4% paraformaldehyde, embedded in paraffin, sectioned at 6 µm, and stained with hematoxylin and eosin (H&E) by standard methods. Masson’s triple staining was also performed according to standard procedures. To examine LacZ expression, mouse ears and flat-mounted retinas were fixed in 0.25% glutaraldehyde/PBS for 5 min, rinsed three times, and stained at 37°C in the X-gal buffer (1.3 mg/ml potassium ferrocyanide, 0.2% Triton X-100, 1 mmol/L MgCl2, and 1 mg/ml X-gal in PBS) for 4 h.
For immunofluorescence, corneas and retinas were fixed overnight with 4% paraformaldehyde, flat-mounted, and immunostained with fluorescein isothiocyanate (FITC) -conjugated anti-CD31 (Chemicon International, Temecula, CA, USA) or anti- proliferating cell nuclear antigen (PCNA; Sigma). In some cases, rhodamine-labeled ricinus communis agglutinin l (Vector Laboratories, Burlingame, CA, USA) was used for staining. Images were taken using a fluorescence microscope (BX51, Olympus, Tokyo, Japan) with a CCD camera (DP70, Olympus) or a confocal microscope (FV1000, Olympus). In some experiments, pictures were imported into Image Pro Plus 5.1 software, and pixels for each color were analyzed to quantitatively represent positively stained cells.
Immunohistochemistry was performed by standard procedures with rat anti-mouse CD31 (Chemicon International, 1:500 dilution) or monoclonal anti-mouse VEGFR2 antibody (Chemicon International, 1:200 dilution) as primary antibodies. Secondary antibodies included goat anti-mouse IgG (Boster BioTec, Wuhan, China), biotinylated rabbit anti-rat IgG, and horseradish peroxidase-conjugated streptavidin (Zhongshanxinqiao, Beijing, China).
For scanning electron microscopy (SEM), tissues were washed in 0.1 M phosphate buffer, fixed with 2.5% glutaraldehyde for 24 h, and washed again in 0.1 M phosphate buffer. Tissues were then rinsed with distilled water and dehydrated with ethanol. Samples were coated with gold particles and examined under a scanning electron microscope (S-3400N, Hitachi, Japan).
Fluorescence-activated cell sorter
Single-cell suspension was prepared and resuspended in PBS containing 2% fetal calf serum (FCS) and 0.02% NaN3. Cells were stained for 30 min on ice with antibodies (Pharmingen, San Diego, CA, USA) that included the following: anti-CD31-FITC (390), biotinylated anti-VEGFR2 (Avas12a1), anti-CD133-FITC (13A4), anti-CD106-FITC (429MVCAM.A), and streptavidin-APC. Stained cells were washed and analyzed using a fluorescence-activated cell sorter (FACSCalibur; BD Immunocytometry System, San Jose, CA, USA). Dead cells were excluded by propidium iodide gating. Data were analyzed using the CellQuest software.
Bone marrow transplantation
Bone marrow (BM) was flushed from femurs and tibiae, and single-cell suspension was obtained by passing through an 18-gauge needle and filtered with a nylon filter, followed by erythrolysis in a buffer containing 0.14 M NH4Cl. Recipient mice were sublethally irradiated and transfused with 5 x 106 BM cells through a caudal vein. Recipients were analyzed 2 months later.
Experimental CNV assay
Pupils of anesthetized mice were dilated using a compound tropicaide ophthalmic solution (Double Crane Pharmaceutical, Shandong, China). Animals were placed on a Mayo stand, and fundi were visualized using a three-mirror contact lens and sodium hyaluronate. A multifrequency Nd:YAG laser beam (Iris Medical Instruments, Mountain View, CA, USA), with wavelength of 532 nm and power of 140 mW, was delivered to retina for 0.1 s through a slit lamp biomicroscope (Haag-Streit; Mason, OH, USA). Six laser burns were performed in the 2-, 4-, 6-, 8-, 10-, and 12-o’clock positions of the posterior pole around the optic nerve of both eyes. Lesions with subretinal cavitation bubbles or slight hemorrhages in choroids were permitted to further experiments.
The photocoagulation-induced CNV (33)
was evaluated by fundus fluorescence angiography (FFA) 1 and 2 wk after the initial laser treatment. Mice were anesthetized and were injected intraperitoneally with 0.1 ml of 2.5% sodium fluorescein (Wuzhou Pharmaceutical, Guangxi, China). Recording of FFA started 3 min after the injection, using a digital imaging system (Heidelberg Engineering, Heidelberg, Germany). Images were captured using a 35 mm Kowa hand-held-fundus camera (Genesis, Tokyo, Japan) at 3, 5, and 8 min after the injection. The presence of an increasing hyperfluorescent lesion in size at the site of laser irradiation was defined as leakage, indicating the incidence of CNV. The percentage of CNV in six laser photocoagulation spots was calculated for each eye.
Histological analysis of CNV lesions was carried out as described above, with serial sections of CNV lesions (
80 sections at 6 µm for each eye). Images were photographed, and data were imported into the Image Pro Plus 5.1 software to analyze the depth and area of CNV lesions.
Matrigel neovascularization assay
Mice were injected subcutaneously with 0.2 ml of heparinized Matrigel (BD Bioscience, San Jose, CA, USA). The animals were sacrificed 5 days later, the gel plugs were carefully recovered by en bloc resection, and each one was divided into half for hemoglobin quantification using the Drabkin method and histology analysis. Some of the gel plugs were fixed and sectioned for H&E staining, as well as for immunohistochemical staining for VEGFR2, as described above.
EC culture
Mice were sacrificed by cervical dislocation, and the aorta was dissected and cut coronally into segments under sterile conditions. The segments were washed and digested in collagenase IV/PBS for 1 h at 37°C. Dispersed cells were washed again and cultured in 24-well plates with Dulbecco’s modified Eagle’s medium supplemented with 20% FCS and endothelial cell growth supplements (BD Bioscience). Cells were characterized by FACS after being stained with anti-CD31, anti-VEGFR2, and anti-CD106. Cells between passages 2 and 5 were used in experiments. For growth curve depiction, cells (5x104) were seeded in 48-well plates and were counted every 2 days.
Reverse transcriptase-polymerase chain reaction (RT-PCR) and quantitative analysis
Liver sinus endothelial cells (LSECs) were sorted from mouse liver (400 mg) using anti-LSEC magnetic beads (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany) following the recommended protocol. Generally, from 1.5 x 108 cells, 4 x 106 LSEC were enriched with purity >85%, as confirmed by FACS with anti-VEGFR2 and anti-CD31 antibodies. Sorted cells were immediately dissolved in the Trizol reagent (Invitrogen, Carlsbad, CA, USA), and total cellular RNA was prepared according to the manufacturer’s instructions. cDNA was prepared from the total RNA using a reverse-transcription kit from Toyobo (Toyobo Co., Osaka, Japan). cDNA from equal amount of total RNA was diluted 4-, 16-, and 64-fold and was used to amplify target genes by PCR. Real-time PCR was performed using a kit (SYBR Premix EX Taq, TaKaRa) and the ABI PRISM 7300 Real-Time PCR system, with GAPDH as a reference control. Primers used in RT-PCR were as follows: Hes1: 5'-TACCGATCACTAAGTAGCCCTAA and 5'-ATTCTTGCCCTTCGCCTC; Hesr1: 5'-GAGACCGAATCAATAACAGT and 5'-AAAGTCGCCAGTAAGTAAGTCAG; VEGFR1 (flt1): 5'-GTGTGAACGGCTGCCCTATG and 5'-ATCACCATCAGAGGCCCTCC; VEGFR2 (KDR/flk1): 5'-CATTATCCTCGTCGGCACTGC and 5'-TGCTCGCTGTGTGTTGCTCC; and β-actin: 5'-CTGGAGAAGAGCTATGAGCTGG and 5'-CAACGTCACACTTCATGATGG. Primers used for real time RT-PCR included the following: GAPDH: 5'-TCGACAGTCAGCCGCATCTTCTT and 5'-GCGCCCAATACGACCAAATCC; VEGFR2 (KDR/Flk1): 5'-GGGATGGTCCTTGCATCAGAA and 5'-ACTGGTAGCCACTGGTCTGGTTG; Hesr1 (Hey1): 5'-TGAGATCTTGCAGATGACTGTGGA and 5'-CAACTTCGGCCAGGCATTC; p21WAF1: 5'-CTGTCTTGAACTCTGGTGTCTGA and 5'-CCAATCTGCGCTTGGAGTGA; and VEGFR1 (Flt1): 5'-TAATGACGATGGCAACAGGGTAGA and 5'-TGTGCACGACCTAAGCACACAG.
Statistics
Mann-Whitney’s U test was used to compare the leakage score in FFA between RBPf/f-MxCre and the control. Pearson 
2 test was used for the CNV production rate. Student’s t test was used for the rest statistical analyses. Values of P < 0.05 were considered statistically significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A) as well as in skin vessels (Supplemental Fig. 1) by X-gal staining.
|
Two months after the induction, corneas of RBPf/f-MxCre mice became opaque. Gross observation by a slit lamp exhibited surprising neovascularization in the cornea and iris of RBPf/f-MxCre mice (Fig. 1B
). The formation of retinal vasculature in normal mouse is completed soon after birth and is largely quiescent in adults. We explored the retinal vasculature of RBPf/f-MxCre mice by staining flat-mounted retina with rhodamine-agglutinin. The result showed that compared with RBP+/f-MxCre mice there were a large number of microvessel sprouts growing from the superficial plexus in the retina of RBPf/f-MxCre mice. These neovasculatures grew perpendicularly from the superficial plexus of the retina, forming ring-like structures in the flat-mounted retina under the microscope (Fig. 1C
). Quantification of these ring-like structures (Fig. 1D
), and filopodia (Fig. 1E
) representing the leading edge of angiogenesis, indicated that there was a significant increase of spontaneous angiogenesis in the retinal vasculature in the absence of RBP-J. Therefore, RBP-J may play a role in maintaining vascular homeostasis in adults.
Vascularization of avascular cornea in RBP-J knockout mice
The avascular statue of normal cornea is essential for optical clarity and optimal vision, which is maintained by soluble VEGFR1 (34)
. In the RBP-J knockout mouse, remarkable vascularization of cornea was observed (Figs. 1B
and 2A)
. This was further demonstrated by whole-mount staining of cornea with anti-CD31, an endothelial marker. The result showed that compared with avascular cornea of RBP+/f-MxCre mice the cornea of RBPf/f-MxCre mice was full of CD31+ ECs (Fig. 2
B), with some of them forming cord-like structures (bottom). H&E staining, immunohistochemistry of VEGFR2, as well as SEM, of corneal sections all demonstrated the formation of vessels in corneal stroma (Fig. 2C
). Interestingly, some erythrocytes were leaking out of vessel lumens as seen under SEM (Fig. 2C
, bottom), suggesting that the neovasculature formed in the absence of RBP-J was immature. Therefore, in addition to soluble VEGFR1, the signaling mediated by RBP-J might provide another critical pathway in maintaining corneal avascularity.
|
Increased neovascularization in internal organs of RBP-J knockout mice
We further investigated neovascularization in other organs of RBP-J knockout mice. The mesenchymal structure in the lung of RBP-J knockout mice appeared abnormal, with a remarkable increase of CD31+ signal on immunohistochemical staining (Fig. 3
A, top). Masson’s triple staining exhibited significant accumulation of extracellular matrix that might be derived from ECs (Fig. 3A
, middle). SEM confirmed the increased lung mesenchyme (Fig. 3A
, bottom), within which a significant increase of capillaries could be seen in RBPf/f-MxCre mice (Fig. 3B
). Similarly, in the liver of RBP-J knockout mice, LSECs increased significantly as shown by immunohistochemical staining with anti-VEGFR2 (Fig. 3C
) and anti-CD31 (data not shown) but not by an isotype control antibody (data not shown). These findings indicated that deficiency of RBP-J might result in extensively spontaneous angiogenesis in adult tissues.
|
Disruption of RBP-J induces vascular outgrowth
To further demonstrate that lack of RBP-J induces angiogenesis directly in vivo, we performed a Matrigel plug experiment. Matrigel containing heparin was injected subcutaneously into RBPf/f-MxCre and RBP-J+/f-MxCre mice. Five days later, Matrigel plugs embedded in RBP+/f-MxCre mice were transparent without visible vascularization. However, gel plugs from RBPf/f-MxCre mice were heavily vascularized with severe "tissue" hemorrhage (Fig. 4
A). This was confirmed by measuring the hemoglobin in the gel extracts, which showed a significantly higher amount of hemoglobin in gels from RBPf/f-MxCre mice (Fig. 4B
). Histological analysis indicated that in gel plugs from RBPf/f-MxCre mice there were a large number of VEGFR2+ ECs and vessel lumens filled with blood cells, while few vessels were seen in gel plugs from RBP+/f-MxCre mice (Fig. 4C, D
). These results demonstrated that disruption of RBP-J induced neovascular outgrowth but the neovasculature induced by RBP-J deficiency might be less mature.
|
Enhanced CNV in RBP-J knockout mice
CNV represents the end stage of age-related macular degeneration (AMD), which accounts for most of elderly human blindness in advanced countries. We used the CNV model to determine whether RBP-J plays a role in modulating angiogenic response to injuries. CNV was induced by photocoagulation of retinal pigment epithelium (RPE) and the underlying Bruch’s membrane at multiple spots with a red diode laser. Seven and fourteen days after photocoagulation, neovasculatures penetrated through RPE and the subretinal space to form neovascular complexes, which were characterized by fluorescein leakage detected with FFA. In RBPf/f-MxCre and RBP+/f-MxCre mice, newly formed vascular complexes after the Bruch’s membrane damage extended from the choroid to subretinal space and caused hyperfluorescence on FFA, similar to previous studies (33)
. However, CNV lesions in RBPf/f-MxCre mice displayed a significant increase in fluorescence leakage as compared with those of control mice on both day 7 and 14 after photocoagulation, leading to higher frequency of visible neovascular complexes at the irradiated spots (Fig. 5
A, B). Moreover, we sectioned serially CNV lesions on day 14 and estimated the size and depth of CNV lesions after H&E staining. The results revealed that the average CNV area and CNV thickness were significant higher in RBPf/f-MxCre mice than those in RBP+/f-MxCre mice (Fig. 5C, D
). Immunohistochemistry of EC markers (CD31 and VEGFR2) confirmed neovasculatures in the lesions (Supplemental Fig. 2). Interestingly, on day 14 after photocoagulation, compared with in RBP+/f-MxCre mice, we found that the CNV lesions in RBPf/f-MxCre mice had not been encapsulated by pericytes and RPE cells (Fig. 5E
), consistent with the idea that Notch signal may play a role in arteriogenesis (35)
. In summary, these data indicated that RBP-J-mediated signaling inhibited angiogenic response to tissue injuries.
|
Enhanced CNV in RBP-J knockout mice can be transferred by BM transplantation
In recent years, BM-derived EPCs have been recognized as participants in adult pathological angiogenesis, such as tumor and myocardial infarction, as well as in the incidence and progression of laser-induced CNV (36
, 37)
. We thus investigated the effects of RBP-J on EPCs by BM transplantation followed by laser-induced CNV in recipient mice. As shown in Fig. 6
A, incidence of CNV was significantly increased in mice receiving BM from RBPf/f-MxCre mice. By a semiquantitative analysis of fluorescein leakage from CNV lesions on days 7 and 14 after laser photocoagulation, we found that in mice receiving BM from RBPf/f-MxCre mice CNV lesions were significantly promoted (Fig. 6B
). Histological analysis of CNV lesions showed that the CNV area and thickness developed remarkably faster and were more severe than the control (Fig. 6C, D
; Supplemental Fig. 3).
|
We further detected EPCs and circulating ECs (CECs) in the peripheral blood of BM-transfused mice by FACS after staining with anti-VEGFR2 and anti-CD31 (CECs) or anti-VEGFR2 and anti-CD133 (EPCs). The results showed that, while the number of EPCs did not change significantly between RBPf/f-MxCre and RBP+/f-MxCre BM-transfused mice, the number of CECs increased greatly in the peripheral blood of mice accepting RBP-J knockout BM (Fig. 6E
; Supplemental Fig. 4). These data suggested that the RBP-J might also affect the differentiation of EPCs.
Interruption of RBP-J induces EC proliferation
RBP-J mediates Notch signaling, which has been demonstrated to induce growth arrest in ECs (38
, 39)
. Therefore, we investigated in vivo the proliferation of ECs in RBP-J knockout mice by immunofluorescent staining of PCNA, a marker of proliferating cells. Because it is difficult to count cells after this staining, we analyzed the pictures using the Image Pro Plus 5.1 software by which ECs and PCNA-positive cells were represented by red areas (rhodamine-agglutinin) and green areas (FITC), respectively. The result showed that, in the retinal vasculature of RBP-J knockout mice, the expression of PCNA in ECs was increased remarkably. These proliferating ECs existed both at growing filopodia and vessel trunks of the superficial retinal vasculature (Fig. 7
A, B). Subsequently, we cultured aortic ECs from RBPf/f-MxCre and RBP-J+/f-MxCre mice. These results showed that ECs from RBPf/f-MxCre mice had a significantly higher potential of proliferation (Fig. 7C
). Labeling of the in vitro cultured ECs with 5-carboxyfluorescein diacetate succinimidyl ester (CFSE) and examining cell proliferation also showed that, consistent with the in vivo results, disruption of RBP-J induced ECs proliferation (Fig. 7D
). These findings indicated that RBP-J-mediated signaling induced growth inhibition in ECs.
|
Up-regulation of VEGFR2 in ECs of RBP-J knockout mice
We further assessed possible molecular mechanisms underlying the growth arrest in RBP-J-deficient ECs, by comparing the expression of critical downstream molecules in LSECs magnetically sorted from RBPf/f-MxCre and RBP+/f-MxCre mice. RBP-J directly regulates Hes-family members (40
, 41)
. We found that deficiency of RBP-J led to down-regulation of Hes1 and Hesr1, consistent with published results (42
; Fig. 7E, F
). Moreover, knockout of RBP-J up-regulated VEGFR2 (Flk1/KDR) expression but down-regulated VEGFR1 (Flt1) mRNA. p21WAF1, a cyclin-dependent kinase (Cdk) inhibitor downstream of Notch signaling, was up-regulated in RBP-J knockout ECs. These results, in combination with published data, suggested that Notch-RBP-J signaling might maintain vascular homeostasis by inhibiting EC proliferation through turning down the VEGFR signaling and influencing the cell cycle.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. However, because loss-of-function mutations of major molecules of the Notch pathway unavoidably result in embryonic lethality, the exact role of Notch signaling in adult vasculature remains obscure, even though a number of in vitro studies have suggested critical roles of Notch signaling in angiogenesis. To solve this issue, we used a conditional loss-of-function mutation of RBP-J, the essential transcription factor mediating signaling of all four Notch receptors in mammals. We have demonstrated that disruption of RBP-J in adult mice results in spontaneous angiogenesis by three pieces of evidence. First, disruption of RBP-J leads to spontaneous angiogenesis in retina and internal organs such as lung and liver. Second, RBP-J knockout results in vascularization of avascular cornea. Third, knockout of RBP-J induces vessel outgrowth into artificial matrices embedded in mouse tissues. Previous studies (18
, 29)
have reported that the deletion of RBP-J recapitulates the phenotypes of individual receptor knockouts in the embryonic stage. Until now, Notch-independent functions of RBP-J have not been described in the mammalian system and RBP-J deletion essentially generates a specific blockade of the canonical Notch signaling pathway (21)
. Therefore, our results suggest that in normal adults, the Notch-RBP-J pathway provides a repressive signal for angiogenesis, which contributes to homeostasis of normal vasculature.
Destabilization of vascular homeostasis is critical in initiation and progression of many diseases (23)
. CNV is a common end-stage process leading to severe visual loss in numerous ocular diseases (44)
. In our model, disruption of the Notch signaling resulted in a more severe neovascularization at the site of laser-induced retinal injury. We hypothesized that the intensified CNV response in RBP-J knockout mice may be attributed to at least three mutually unexclusive reasons. One is the increased angiogenesis, as shown by the fact that disruption of Notch signaling induces spontaneous angiogenesis. A second reason might be the increased vasculogenesis. Many research groups have reported that BM-derived EPCs take part in CNV, at least in experimental models. Our results showed that mice receiving BM from RBP-J knockout mice exhibited stronger CNV responses. FACS analyses showed that, despite the fact that the number of EPCs did not change in the recipient mice, CECs indeed increased remarkably in mice receiving BM from RBP-J knockout mice. These results suggest that Notch-RBP-J signaling might be involved in the regulation of EPC-mediated vasculogenesis in adults. Thirdly, immaturity of neovasculature may also contribute to a stronger CNV response in RBP-J knockout mice. It should be noted that in the Matrigel plug assay, even though ECs created prominent lumens filled with blood cells, severe hemorrhages were displayed in the plugs, suggesting that with disrupted RBP-J new lumens would be defective and fail to create a vascular network in proper density and framework. Similar defective lumens were also noticed in vessels formed in corneal stroma in RBP-J knockout mice. Last but not least, CNV is the most severe complication of AMD, which mostly leads to blindness in elderly patients. A recent study (45)
has implicated that Notch signaling is involved in age-related disorders. As our results have demonstrated that loss of RBP-J-mediated signaling promotes CNV, it might be interesting to examine dynamics of Notch molecule expression in ECs of different ages so as to figure out whether Notch signaling is involved in human AMD.
How does Notch signaling maintain vascular homeostasis? In mature vasculature, most ECs remain quiescent except for the menstruous ovary. Initiation of angiogenesis involves entry of ECs into cell cycle, proliferation and migration, and re-exits from the cell cycle to form mature vessel lumens. Notch signaling influences cell fate decision, apoptosis and cell proliferation by enabling cell-cell communication between contacting cells. In RBP-J knockout mice, we did not detect alteration in apoptosis compared with control mice (data not shown), suggesting that Notch signaling might not target cell apoptosis in our system. However, by using PCNA staining, we did reveal elevated EC proliferation in RBPf/f-MxCre mice compared with RBP+/f-MxCre mice, supporting previous studies (46
, 47
, 49)
using -secretase inhibitors, selective antibodies, or soluble antagonists based on ligand-receptor interaction. To confirm these observations, we cultured aortic ECs from RBPf/f-MxCre and RBP+/f-MxCre mice in vitro. RBP-J knockout ECs have remarkable advantage in proliferation, demonstrating that the RBP-J-mediated Notch signaling pathway represses EC proliferation in normal adults. This contact-mediated growth inhibition might be important not only for the maintenance of vascular homeostasis but also for ECs to exit from the cell cycle in the formation of vessel lumens, as we noticed that when cultured for 7 days in vitro, RBP-J-deficient ECs overlaid and stacked to form sheeted structures (data not shown).
Reciprocal interaction between Notch signaling and VEGF-VEGFR signaling has been implicated in ECs (38
, 48
, 49)
. Previous in vitro studies (27
, 50)
have shown that Notch signaling represses EC proliferation through Hes family members and VEGFR2. Indeed, semiquantitative RT-PCR as well as quantitative real-time PCR showed that in LSECs from RBPf/f-MxCre mice, expression of Hes1 and Hesr1 was significantly reduced. VEGFR2 was up-regulated, consistent with the proliferation phenotype and previous reports. This suggests that in adulthood, Notch-RBP-J signaling maintains vascular homeostasis by repressing EC proliferation possibly through down-regulation of VEGFR2 expression. Of note is that the expression of VEGFR1 is significantly down-regulated in ECs of RBP-J deficient mice. According to studies reported recently by Ambati et al. (34)
, soluble VEGFR1 (Flt1) is responsible for maintaining the corneal avascularity by antagonizing proangiogenic functions in cornea. A recent in vitro study (51)
indicated the regulative role of Notch signaling on VEGFR1. Our results suggest that Notch-RBP-J signaling may be involved in maintaining the avascularity of cornea by up-regulating the expression of soluble VEGFR1. However, on the other hand, our observation raises another open question, namely how Notch-RBP-J signaling regulates two VEGF receptors in an opposite manner.
Down-regulation of p21WAF1 in ECs has been reported when Notch is activated (39)
. Consistently, our quantitative analysis has demonstrated that when RBP-J is deleted, the expression of p21WAF1 is up-regulated. Although up-regulation of p21WAF1 on Notch activation has been implicated in differentiation-associated growth arrest in keratinocytes (52
, 53)
, our results suggest that at least in ECs, up-regulation (in this study by RBP-J deletion) of p21WAF1 expression may not be sufficient for growth arrest and subsequent differentiation. This difference may be due to the cell type-specific context, and detailed molecular analyses are required to elucidate the control mechanism for the growth and differentiation of ECs.
A series of recent studies (54)
implicated Notch signaling as a new therapeutic target for pathological angiogenesis by the application of agonists or antagonists. Here, our findings demonstrate the essential role of Notch-RBP-J signaling in the maintenance of vascular homeostasis in adults, which will be instructive for developing and administration of such therapies.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication October 9, 2007. Accepted for publication November 29, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
