QUICK SEARCH:   [advanced] Author: Keyword(s):
Home     Help     Feedback     Subscriptions     Archive     Search     Table of Contents

First published online March 1, 2004
doi: 10.1242/10.1242/dev.01036

This Article Summary Full Text Full Text (PDF) Supplemental Data Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Related articles in Development Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Whitehead, K. J. Articles by Li, D. Y. Search for Related Content PubMed PubMed Citation Articles by Whitehead, K. J. Articles by Li, D. Y.

Ccm1 is required for arterial morphogenesis: implications for the etiology of human cavernous malformations

Kevin J. Whitehead^1,2,*, Nicholas W. Plummer^3,*, Jennifer A. Adams¹, Douglas A. Marchuk^3, and Dean Y. Li^1,2,

¹ Program in Human Molecular Biology and Genetics, University of Utah, Building 533 Room 4220, 15 N 2030 East, Salt Lake City, Utah 84112, USA
² Department of Medicine, University of Utah, 4C104 SOM, 30 N 1900 East, Salt Lake City, Utah 84132, USA
³ Department of Molecular Genetics and Microbiology, Duke University Medical Center, 268 CARL Building, Box 3175, Durham, North Carolina 27710, USA

View larger version (40K): [in a new window]   Fig. 1. Generation of mice lacking Ccm1. (A) Targeting strategy to generate the Ccm1tm1Dmar allele. Mice heterozygous for the allele will be referred to as Ccm1+/-, and mice homozygous for the allele will be referred to as Ccm1-/-. Our construct was designed to replace most of the sixth and the entire seventh coding exon of mouse Ccm1 with an internal ribosomal entry site (IRES) and the E. coli lacZ gene. Although this construct successfully ablated Ccm1 gene expression, we were unable to detect any ß-galactosidase protein or enzyme activity in Ccm1+/- or Ccm1-/- embryos. (B) Southern blotting of KpnI-digested genomic DNA detects a 3.1 kb band from the recombinant allele of a heterozygous animal. This same probe detects a 4.3 kb band from the parent strains (C57BL/6J, labeled B6, and 129X1Sv/J, labeled 129), from the wild-type allele of a heterozygous animal and from a wild-type littermate. Long-range PCR and sequencing of the resulting product also confirmed homologous recombination and conserved sequence at the 3' and 5' ends. (C) Genotyping of mice was performed using allele specific PCR primers. The wild-type primers amplify a 466 bp product, and the mutant primers amplify a 310 bp product. (D) In situ hybridization for Ccm1 detects ubiquitous expression at E8.5 in a wild-type embryo. No transcript is detected in Ccm1-/- embryos using this probe, which spans exons 3 through 7. Scale bars: 200 µm.

View larger version (142K): [in a new window]   Fig. 2. Dilated cranial vessels in Ccm1-/- mice phenocopy human cavernous malformations. (A-D) Immunohistochemical staining for the endothelial antigen Pecam on cross-sections from cranial portions of the embryo. (A,B) Immunostaining reveals the plexus of head vessels at E8.5. Significant enlargement of cranial vessels is observed early in Ccm1-/- embryos (B). (C,D) Examination of embryos at E9.5 shows further enlargement of cranial vessels in Ccm1-/- embryos, which now occupy much of the region normally encompassed by mesenchymal tissue. By this stage, Ccm1-/- mice are readily distinguished from their phenotypically normal littermates and no further embryonic enlargement is observed. (E,F) Comparison of a human cavernous malformation (E) with a similar magnification of a Ccm1-/- mouse embryo (F) shows cavernous vascular channels surrounded by a thin layer of Pecam-stained endothelium (stained brown). Human lesions contain vessels of widely varying sizes, some larger than those shown. Scale bars: 100 µm.

View larger version (109K): [in a new window]   Fig. 3. Vascular dilatation in mice lacking Ccm1. (A-D) Immunohistochemical stains for the endothelial antigen Pecam on cross-sections taken from caudal regions of the embryo (as indicated in diagrams to the left). (A,B) The paired dorsal aortae and midline vitelline arteries are apparent at E8.5. Significant enlargement is observed in the dorsal aortae of Ccm1-/- embryos. (C,D) A comparison of wild-type (C) and Ccm1-/- (D) embryos at E9.0. There is marked enlargement and midline fusion of the dorsal aortae in the Ccm1-/- embryo. The enlarged vessel occupies almost the entire volume of the embryo and distorts the closed neural tube. (E-G) Endothelial proliferation at E8.5 in vivo as determined by immunofluorescent double-labeling with antibodies for Pecam and phosphorylated histone 3, a marker of mitosis. Sections were counterstained with DAPI to define cell nuclei. (E,F) Sections taken from wild-type and homozygous mutant embryos, respectively. Two double-positive (mitotic) endothelial nuclei are demonstrated in the Ccm1-/- embryo (arrows in F). (G) A significantly increased endothelial cell proliferative rate is observed for the dilated aortae of Ccm1-/- embryos, distal to the heart (light gray portion of embryo diagram to left). Proliferative rates from more rostral sections of aorta and branchial arch arteries were similar between Ccm1+/+ and Ccm1-/- embryos. Data bars represent the mean values from three separate embryo pairings, and a total of 137 Ccm1+/+ and 192 Ccm1-/- aortic cross-sections. Error bars represent s.e.m. Scale bars: 100 µm.

View larger version (109K): [in a new window]   Fig. 4. Proximal narrowing limits flow into dilated distal vessels of Ccm1-/- mice. (A-D) Confocal immunofluorescent detection of the endothelial antigen Pecam. (A,B) Composite whole-mount images of E8.5 Ccm1+/+ and Ccm1-/- embryos. The amnion and yolk sac have been removed, allowing the embryos to be extended (some kinking of the midpoint of the Ccm1+/+ embryo resulted). Note the enlarged diameter of the dorsal aorta of the caudal Ccm1-/- embryo (arrow in B). The intersomitic arteries that extend dorsally from the aorta are dilated and more prominent in Ccm1-/- embryos. (C,D) Higher magnification views of the first branchial arch artery and proximal aorta (see boxes in A and B). The wild-type embryo has a uniform, broad dorsal aorta (double arrow in C). By contrast, the Ccm1-/- embryo shows narrowing of the branchial arch artery and adjacent dorsal aorta (single arrows in D). Distally, the aorta restores to a more normal diameter (double arrow in D). (E-H) Injection of India ink into the ventricle of embryonic hearts. (E,F) At E8.5, ink fills the first branchial arch artery and dorsal aorta of a wild-type embryo, eventually entering the head veins. Ink fails to enter the dorsal aorta of a Ccm1-/- embryo, and instead flows in a retrograde manner through the sinus venosus and into the common cardinal vein (arrow in F). Yolk sacs were removed following injection to improve visualization. (G,H) At E9.5, injected ink flows primarily through the second and third branchial arch arteries to fill the dorsal aorta (arrowheads in G). Ink also fills the venous system of the embryo. By contrast, injection of a Ccm1-/- embryo fails to opacify the dorsal aorta. A small amount of ink is observed in the first and second branchial arch arteries (arrowheads in H) with a minimal amount entering the adjacent aorta. Scale bars: 200 µm.

View larger version (139K): [in a new window]   Fig. 5. Branchial arch arteries and adjacent dorsal aortae become narrowed in mice lacking Ccm1. (A-F) Immunohistochemical stains for the endothelial antigen Pecam on sagittal sections of embryos. Yellow arrows indicate one of the paired dorsal aortae; blue arrowheads highlight one of the paired branchial arch arteries. (A,B) No differences are observed between Ccm1+/+ and Ccm1-/- embryos at E8.0. A patent lumen is present in the bilateral dorsal aortae. A cord of endothelial tissue extends bilaterally from the aortic sac, through the branchial arch towards the dorsal aortae. (C,D) At E8.5, the Ccm1+/+ embryo has formed a widely patent first branchial arch artery, whereas the branchial arch arteries of the Ccm1-/- embryo fail to enlarge. (E,F) Further enlargement of these vessels is present at E9.0 in a wild-type embryo. The Ccm1-/- embryo, however, is left with only a vestige of the branchial arch arteries and dorsal aortae. Scale bars: 100 µm.

View larger version (134K): [in a new window]   Fig. 6. Neural patterning is not perturbed in Ccm1-/- mice. (A-H) In situ hybridization of Ccm1+/+ and Ccm1-/- embryos at E8.5. Arrowheads identify the isthmus between midbrain and hindbrain. (A,B) The expression of the homeobox gene Six3 is restricted to the forebrain. (C,D) The expression of the transcription factor Otx2 remains restricted to the forebrain and midbrain in both Ccm1+/+ and Ccm1-/- mice. (E,F) Transcript for Fgf8 is detected from the isthmus, bounded by the midbrain and hindbrain (arrowheads in all panels). (G,H) The homeobox gene Gbx2 is expressed throughout the hindbrain, up to the junction with the midbrain (or isthmus) in both wild-type and Ccm1 homozygous mutant embryos. (I,J) The expression of the leucine zipper transcription factor MafB is restricted to rhombomeres 5 and 6 of the hindbrain in both Ccm1+/+ and Ccm1-/- mice. (K,L) The secreted molecule shh is expressed in the notochord and ventral floor plate of the neural tube. (M,N) The homeobox gene Pax7 is expressed along the dorsal neural tube of both genotypes. (O,P) The homeobox gene Nkx2-2 is expressed in the ventral floor plate of the forebrain at E8.5. (Q,R) The axon guidance molecule Netrin1 is expressed in the ventral floor plate, as well as in adjacent somites, in Ccm1-/- and Ccm1+/+ mice. All markers studied showed similar expression between Ccm1+/+ and Ccm1-/- mice. Scale bars: 200 µm.

View larger version (91K): [in a new window]   Fig. 7. Impaired arterial identity in mice lacking Ccm1. (A,B) Immunohistochemical staining for -smooth muscle actin shows that the arterial endothelium in Ccm1-/- embryos fails to recruit vascular smooth muscle cells. Sections at the level of the atrium of the heart at E9.5 show actin-positive cells along the medial aspect of the wild-type dorsal aorta (arrows in A) without staining of the adjacent cardinal vein. No actin staining is present in the enlarged dorsal aorta of the Ccm1-/- embryo (B), although cardiac myocytes stain positive. da, dorsal aorta; v, vein; g, gut; pp, pericardial-peritoneal canal; hrt, heart. (C-F) Staining with X-Gal to examine the expression of a tau-lacZ transgene driven by the mouse Efnb2 promoter (Wang et al., 1998). Embryos shown are all heterozygous for the transgene at the Efnb2 locus. (C,D) Efnb2 is expressed in the somites, the nephrogenic cords (arrowheads) and the hindbrain of both genotypes, with arterial expression observed in the dorsal aorta and vitelline artery (arrows, C) of the wild-type embryo. No arterial expression is observed in the Ccm1-/- embryo. (E,F) Efnb2 transgene expression is detected in the yolk sac arteries of a Ccm1+/+ embryo at E9.0, with no expression observed in the Ccm1-/- yolk sac. (G-J) In situ hybridization for Efnb2 in E8.5 yolk sacs. (G,H) Efnb2 transcript is detected in the arterial endothelial network of the caudal pole of the yolk sac (asterisks) of a Ccm1+/+ embryo, with no endothelial stain observed from the same region of a Ccm1-/- yolk sac. (I,J) Higher magnification view of yolk sacs (corresponding to boxes in G,H). Efnb2 stain is observed in individual endothelial cells lining the yolk sac arterial network, with no stain observed in yolk sacs lacking Ccm1. Scale bars: 100 µm.

View larger version (113K): [in a new window]   Fig. 8. Ccm1 lies genetically upstream of Notch signaling and downstream of Vegf. (A-F) Whole-mount in situ hybridization of Ccm1+/+ and Ccm1-/- tissues at E8.5, before the mutant phenotype can be distinguished grossly. (A-D) Disruption of arterial Notch gene expression is demonstrated using probes for Dll4 and Notch4. (A,B) Hybridization with a probe for the Notch ligand Dll4 shows marked downregulation of Dll4 transcript throughout the dorsal aorta of Ccm1-/- embryos, compared with normal controls. (C,D) A decrease in Notch4 transcript is evident in the branchial arch artery and proximal aorta of a Ccm1-/- embryo. The Notch4 signal intensity was similar, although weak, for the caudal aorta of both the wild type and the mutant. (E,F) In situ hybridization with a probe for Vegfa shows a similar intensity of ubiquitous signal from both Ccm1+/+ and Ccm1-/- embryos. (G) Quantification of transcript levels by real-time quantitative PCR at E8.8. Comparison of three pairs of Ccm1+/+ and Ccm1-/- embryo cDNA samples confirmed downregulation of Efnb2 expression (despite intact extravascular domains of expression shown in Fig. 7A-D). Marked downregulation of Dll4 with modest downregulation of Notch4 was also observed, in agreement with the in situ hybridization data. Transcript levels for Vegfa were similar between genotypes.

View larger version (106K): [in a new window]   Fig. 9. Downregulation of NOTCH4 in arteries associated with CCM from patients with loss-of-function mutations in CCM1. (A,B) Immunohistochemistry using antibodies against NOTCH4 on human tissues. (A) A section of normal brain taken from an autopsy specimen showing NOTCH4 signal from endothelium and vascular smooth muscle cells of two arteries (red arrowheads), without significant expression from an adjacent vein (blue arrow). (B) A section from a surgically excised specimen from a patient with a previously characterized mutation of CCM1 (Sahoo et al., 1999). An artery (red arrowhead) from brain tissue adjacent to a cavernous malformation shows little NOTCH4 protein in the endothelium or vascular smooth muscle cells. There is no NOTCH4 staining in an adjacent venule (blue arrow). Similar findings were present from two other members of the same family. Scale bars: 100 µm.