CCM1-ICAP-1 complex controls β1 integrin-dependent endothelial contractility and fibronectin remodeling

Abstract

The endothelial CCM complex regulates blood vessel stability and permeability. Loss-of-function mutations in CCM genes are responsible for human cerebral cavernous malformations (CCMs), which are characterized by clusters of hemorrhagic dilated capillaries composed of endothelium lacking mural cells and altered sub-endothelial extracellular matrix (ECM). Association of the CCM1/2 complex with ICAP-1, an inhibitor of β1 integrin, prompted us to investigate whether the CCM complex interferes with integrin signaling. We demonstrate that CCM1/2 loss resulted in ICAP-1 destabilization, which increased β1 integrin activation and led to increased RhoA-dependent contractility. The resulting abnormal distribution of forces led to aberrant ECM remodeling around lesions of CCM1- and CCM2-deficient mice. ICAP-1-deficient vessels displayed similar defects. We demonstrate that a positive feedback loop between the aberrant ECM and internal cellular tension led to decreased endothelial barrier function. Our data support that up-regulation of β1 integrin activation participates in the progression of CCM lesions by destabilizing intercellular junctions through increased cell contractility and aberrant ECM remodeling.

Figure 1.

CCM1 and ICAP-1 proteins are destabilized upon CCM1 and CCM2 loss. (A) ICAP-1, CCM1, and CCM2 protein content in total lysates of KD HUVECs was analyzed by Western blot. (B) Quantification of the three proteins normalized to actin in silenced HUVEC; Error bars are ± SEM (n = 3). ICAP-1 and CCM1 proteins were strongly reduced in the three conditions. (C) Q-PCR measurements show that knock-down of one protein had not effect on the expression of the two other partners. Error bars are means ± SEM (n = 3). (D) Western blot of CCM2 +/+ and −/− embryo lysates showing the loss of CCM1 and ICAP-1 proteins upon CCM2 depletion. Remaining CCM2 comes from maternal material. (E) Western blot of ICAP-1 andCCM1 in liver lysates from ICAP-1+/+ and ICAP-1−/− mice. One fourth of CCM1 protein remains in ICAP-1−/− mice (see Fig. S1 C for densitometric analysis). Black lines indicate that intervening lanes have been spliced out for presentation purposes. (F) Overexpressed ICAP-1 protein in CHO is stabilized by the coexpression of CCM1 but not of the N192-Y195 non-interacting CCM1 mutant. Cycloheximide was added at t = 0 to block protein synthesis with or without the proteasomal inhibitor MG132. Results are representative of three independent experiments. (G) Quantification of ICAP-1 protein level over a time-course after addition of cycloheximide when expressed alone or in different combinations with CCM1 and/or CCM2. Results are representative of two independent experiments.

Figure 2.

CCM1/2 proteins control RhoA-dependent actin cytoskeleton organization by regulating β1 integrin activation. (A) Flow cytofluorometry using 9EG7 shows an increase in β1 integrin activation upon silencing of CCM1, CCM2, or ICAP-1. Error bars are means ± SEM (n = 4). (B) HUVECs depleted in ICAP-1, CCM1, or CCM2 spread for 1 h on low density of FN displayed more and larger β1 integrin containing focal adhesions (stained with 9EG7 antibody) localized all over their ventral face. Bar, 5 µm. (C) Quantification of the percentage of cells displaying central plaques. Error bars are means ± SEM (n = 3). (D and E) Quantification of RhoA activation upon ICAP-1, CCM1, or CCM2 depletion alone (D) or with additional β1 integrin silencing (E) by RhoGTP pull-downs. Error bars are means ± SEM (n = 3). RhoG-LISA measurements show that the level of RhoAGTP returned to that of CT upon ICAP-1 or CCM1/2 depletion in absence of β1 integrin (see Fig. S3 B). (F) Transversal actin bundles were observed in elongated HUVECs upon depletion of ICAP-1, CCM1, or CCM2 spread on FN for 4 h. Additional siRNA depletion of β1 but not of β3 integrin (Fig. S3 A) abolished their formation. (G) As sparse cells, confluent HUVECs displayed transversal actin stress fibers and their junctional VE-cadherin and β-catenin stainings appeared thinner and discontinuous (see Fig. S4). Transversal actin fibers were abolished upon additional depletion in β1 but not in β3 integrin. (H) Quantification of the percentage of sparse cells with transversal actin bundles in the absence or presence of additional β1 or β3 integrin depletion. Error bars are means ± SEM (n = 3). *, P < 0.05; **, P < 0.005; ***, P < 0.0005.

Figure 3.

CCM1/2 proteins control β1 integrin–dependent distribution of traction forces and FN remodeling. (A) pMLC and zyxin delocalized from the peripheral focal adhesions to transversal actin stress fibers upon depletion in ICAP-1, CCM1, or CCM2 after 4 h of spreading. Bar, 5 µm. These images are presented again in Fig. S3 C, where it is shown that pMLC and zyxin relocalized to the cell periphery upon additional silencing of β1 integrin but not of β3 integrin. (B) Representative traction forces maps obtained by TFM. Forces were delocalized from the cell periphery to all over the ventral face upon CCM depletion. (C) Quantification of the percentage of cells displaying central traction forces. Between 14 and 23 cells of each were analyzed. Error bars are means ± SEM (n = 2). Similar fold-increase of the percentage of cells bearing central forces was measured for CCM1-depleted HUVECs (n = 1; not depicted). Bar, 10 µm. Remodeling of glass-coated FN was studied on sparse cells after 24 h (D) and on confluent cells after 48 h (E). Bars, 5 µm. FN staining in the yellow box in E is magnified to highlight the linear pattern of parallel FN fibers produced by ICAP-1/CCM-depleted HUVECs. Bars, 2 µm. These data are representative of more than three independent experiments.

Figure 4.

Abnormal FN deposition around mouse CCM1 or CCM2 lesions and ICAP-1–deficient blood vessels. (A) Sections of iCCM2 were labeled with total FN, cFN, and PECAM. CCM lesions are annotated by asterisks and normal blood vessels by arrows. cFN was abundant around CCM lesions, but barely detectable around peri-lesion blood vessels. Bar, 50 µm. (B and C) Numerous FN fibers appeared around the lesions (B, arrow), whereas FN signal was weak and diffuse around unaffected peri-lesion vessels (B, asterisk) or CT blood vessels (D). Strong collagen III staining was detected around CCM lesions but not around unaffected peri-lesion vessels or CT blood vessels. n = 3 mice for iCCM1, n = 6 mice for iCCM2. Bars, 10 µm. (E) Projections of confocal z-stacks of whole-mount of ear blood vessels of ICAP-1 +/+ and −/− mice immunostained for PECAM and FN. Blood had been washed out by intracardiac perfusion of PBS. The plot of intensity of the FN signal over the surface delineated by the yellow square is presented in the middle. Increased FN deposition as a mixture of soluble and thin fibrils around ICAP-1−/− blood vessels differed from the discrete FN fibers around ICAP-1+/+ blood vessels. n = 3 from two litters for each genotype. Bars, 10 µm.

Figure 5.

Defective ECM organization increases cellular tension and decreases barrier function of naive HUVECs. (A) ICAP-1-, CCM1-, or CCM2-depleted HUVECs were seeded at confluency on FN-coated slides. Cells were removed 48 h later by a triton/ammonium hydroxide treatment. Naive HUVECs were seeded on these decellularized remodeled matrices. (B and C) FN and ColIV staining of decellularized matrices remodeled in the absence (B) or presence of FUD (C). Note the linearity of FN fibers and ColIV in ICAP-1- and CCM1/2-depleted conditions. Bars, 1 µm. (D) Quantification of the percentage of naive HUVECs with central β1 integrin focal adhesions after 1 h of spreading. (E) Quantification of the percentage of naive HUVECs with transversal stress fibers after 4 h of spreading. (F) β-catenin and actin staining of confluent naive HUVECs on the remodeled matrices after 24 h. Bars, 5 µm. (G) Quantification of β-catenin staining thickness showing thinner junctions of CCM KD matrices. (H) Time-course measurement of the electrical impedance of a monolayer of naive HUVECs on these matrices. The data shown are from a single representative experiment out of three repeats. (I) Normalized impedance to that on CT matrix showing a significant decrease of monolayer resistance barrier on CCM KD matrices due to structural changes in FN remodeling. Error bars are ± SEM (n = 3). *, P < 0.05; **, P < 0.005; ***, P < 0.0005 using GLMM with Tukey’s test.

Figure 6.

Ultrastructural defects of the basal lamina and interstitial ECM around ICAP-1–deficient blood vessels are associated with increased permeability to small molecules. (A) Transmission electron micrographs of retinal capillaries showing dilation of ICAP-1−/− capillary compared with ICAP+/+. Bar, 1 µm. Enlargements of the yellow square show that basal lamina (BL) around ICAP-1−/− capillary present zones of multilayering due to dispersion of the dense matter and zones of accumulation of matter in spots (arrows). EC, endothelial cell; P, pericyte. Bar, 0.5 µm. (B) Arrows indicate zones of tearing of the basal lamina around an ICAP-1−/− venule. SMC, smooth muscle cell. Bar, 1 µm. (C) Fenestrated choroidal capillaries showing multilayering of the basal lamina (BL) juxtaposed to EC and enlargement of the interstitial ECM, which appears scattered around ICAP-1−/− capillary. Arrowhead points to detachment of the EC from its BL. EC–EC junctions are present (arrow). Bar, 1 µm. n = 4 for each genotype from three litters. (D) Gallery of ear pictures over time under a fluorescence microscope after injection of Alexa Fluor 546/10-kD dextran in the blood circulation. See the progressive increase in fluorescence of the ear tissue in ICAP-1−/− mouse compared with wild-type littermate (yellow rectangle). Bar, 100 µm. (E) Diffusion rate of the 10-kD dextran was calculated between 90 and 300 s after injection. Error bars are means ± SEM (n = 5 for each genotype from five litters for +/+ and two litters for −/−). *, P < 0.05 one-tailed paired t test.

Figure 7.

ICAP-1–deficient mice present dilated, more branched and tortuous blood vessels. (A) Mesenteric veins of ICAP-1−/− mice were dilated and more branched than in their wild-type littermates. Bar, 5 mm. (B) Quantification of the number of final branches connecting the intestine per mesenteric vein. Points correspond to the number of branches of an individual vein and horizontal bars are the means per genotype (ICAP-1+/+, n = 5 mice from five litters; ICAP-1−/−, n = 4 mice from three litters). *, P < 0.05. (C) Projections of confocal z-stacks of ear whole-mount staining of blood vessels with PECAM. Bar, 100 µm. (D) Histogram of the distribution of blood vessel diameter categories (ICAP-1+/+, n = 4 mice from three litters; ICAP-1−/−, n = 5 mice from two litters). (E) Projections of confocal z-stacks of ear whole-mount staining of blood vessels with PECAM, smooth muscle actin (SMA), and desmin. Bar, 10 µm.

Figure 8.

β1 activation upon loss in CCM1 or 2 drives a loop between endothelial cell contractility and ECM with deleterious effect on cell–cell junctions. In a quiescent vessel, ICAP-1 maintains low β1 integrin activation. Endothelial cells are well joined and VE-cadherin adherens junctions are stabilized by a cytoplasmic adaptor complex that recruits junctional actomyosin cytoskeleton. Upon CCM1 or CCM2 depletion, ICAP-1 protein is destabilized and lost. β1 integrin is activated and activates in turn RhoA/ROCK-dependent actin stress fiber formation. Increased β1 integrin activation and cell contractility result in aberrant remodeling of ECM in linear and parallel fibers onto which cells spread and flatten. A self-sustaining mechanical loop is initiated that increases intra- and external tensions destabilizing cell–cell junctions.