Research ArticleAngiogenesisVascular biology Open Access | 10.1172/jci.insight.160372

Xi Yang,1 Shi-Ting Wu,1 Rui Gao,1 Rui Wang,1 Yixuan Wang,1 Zhenkun Dong,2 Lu Wang,1 Chunxiao Qi,1 Xiaohong Wang,1 M. Lienhard Schmitz,3 Renjing Liu,4 Zhiming Han,5,6 Lu Wang,2 and Xiangjian Zheng1

1Department of Pharmacology and Tianjin Key Laboratory of Inflammation Biology, School of Basic Medical Sciences, and Center for Cardiovascular Diseases, Tianjin Medical University, China.

2State Key Laboratory of Experimental Hematology, Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Tianjin, China.

3Institute of Biochemistry, Justus Liebig University, Member of the German Center for Lung Research, Giessen, Germany.

4Vascular Epigenetics Laboratory, Victor Chang Cardiac Research Institute, and St. Vincent’s Clinical School, University of New South Wales, Sydney, Australia.

5State Key Laboratory of Stem Cell and Reproductive Biology, Institute of Zoology, Chinese Academy of Sciences, and

6Beijing Institute for Stem Cell and Regenerative Medicine, Beijing, China.

Address correspondence to: Xiangjia Zheng, No.22 Qi Xiang Tai Road, Tianjin 300070, China. Phone: 86.22.833.6835; Email: xzheng@tmu.edu.cn.

Authorship note: XY and STW contributed equally to this work.

Find articles by Yang, X. in: JCI | PubMed | Google Scholar |

1Department of Pharmacology and Tianjin Key Laboratory of Inflammation Biology, School of Basic Medical Sciences, and Center for Cardiovascular Diseases, Tianjin Medical University, China.

2State Key Laboratory of Experimental Hematology, Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Tianjin, China.

3Institute of Biochemistry, Justus Liebig University, Member of the German Center for Lung Research, Giessen, Germany.

4Vascular Epigenetics Laboratory, Victor Chang Cardiac Research Institute, and St. Vincent’s Clinical School, University of New South Wales, Sydney, Australia.

5State Key Laboratory of Stem Cell and Reproductive Biology, Institute of Zoology, Chinese Academy of Sciences, and

6Beijing Institute for Stem Cell and Regenerative Medicine, Beijing, China.

Address correspondence to: Xiangjia Zheng, No.22 Qi Xiang Tai Road, Tianjin 300070, China. Phone: 86.22.833.6835; Email: xzheng@tmu.edu.cn.

Authorship note: XY and STW contributed equally to this work.

Find articles by Wu, S. in: JCI | PubMed | Google Scholar

1Department of Pharmacology and Tianjin Key Laboratory of Inflammation Biology, School of Basic Medical Sciences, and Center for Cardiovascular Diseases, Tianjin Medical University, China.

2State Key Laboratory of Experimental Hematology, Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Tianjin, China.

3Institute of Biochemistry, Justus Liebig University, Member of the German Center for Lung Research, Giessen, Germany.

4Vascular Epigenetics Laboratory, Victor Chang Cardiac Research Institute, and St. Vincent’s Clinical School, University of New South Wales, Sydney, Australia.

5State Key Laboratory of Stem Cell and Reproductive Biology, Institute of Zoology, Chinese Academy of Sciences, and

6Beijing Institute for Stem Cell and Regenerative Medicine, Beijing, China.

Address correspondence to: Xiangjia Zheng, No.22 Qi Xiang Tai Road, Tianjin 300070, China. Phone: 86.22.833.6835; Email: xzheng@tmu.edu.cn.

Authorship note: XY and STW contributed equally to this work.

Find articles by Gao, R. in: JCI | PubMed | Google Scholar

1Department of Pharmacology and Tianjin Key Laboratory of Inflammation Biology, School of Basic Medical Sciences, and Center for Cardiovascular Diseases, Tianjin Medical University, China.

2State Key Laboratory of Experimental Hematology, Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Tianjin, China.

3Institute of Biochemistry, Justus Liebig University, Member of the German Center for Lung Research, Giessen, Germany.

4Vascular Epigenetics Laboratory, Victor Chang Cardiac Research Institute, and St. Vincent’s Clinical School, University of New South Wales, Sydney, Australia.

5State Key Laboratory of Stem Cell and Reproductive Biology, Institute of Zoology, Chinese Academy of Sciences, and

6Beijing Institute for Stem Cell and Regenerative Medicine, Beijing, China.

Address correspondence to: Xiangjia Zheng, No.22 Qi Xiang Tai Road, Tianjin 300070, China. Phone: 86.22.833.6835; Email: xzheng@tmu.edu.cn.

Authorship note: XY and STW contributed equally to this work.

Find articles by Wang, R. in: JCI | PubMed | Google Scholar |

1Department of Pharmacology and Tianjin Key Laboratory of Inflammation Biology, School of Basic Medical Sciences, and Center for Cardiovascular Diseases, Tianjin Medical University, China.

2State Key Laboratory of Experimental Hematology, Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Tianjin, China.

3Institute of Biochemistry, Justus Liebig University, Member of the German Center for Lung Research, Giessen, Germany.

4Vascular Epigenetics Laboratory, Victor Chang Cardiac Research Institute, and St. Vincent’s Clinical School, University of New South Wales, Sydney, Australia.

5State Key Laboratory of Stem Cell and Reproductive Biology, Institute of Zoology, Chinese Academy of Sciences, and

6Beijing Institute for Stem Cell and Regenerative Medicine, Beijing, China.

Address correspondence to: Xiangjia Zheng, No.22 Qi Xiang Tai Road, Tianjin 300070, China. Phone: 86.22.833.6835; Email: xzheng@tmu.edu.cn.

Authorship note: XY and STW contributed equally to this work.

Find articles by Wang, Y. in: JCI | PubMed | Google Scholar

1Department of Pharmacology and Tianjin Key Laboratory of Inflammation Biology, School of Basic Medical Sciences, and Center for Cardiovascular Diseases, Tianjin Medical University, China.

2State Key Laboratory of Experimental Hematology, Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Tianjin, China.

3Institute of Biochemistry, Justus Liebig University, Member of the German Center for Lung Research, Giessen, Germany.

4Vascular Epigenetics Laboratory, Victor Chang Cardiac Research Institute, and St. Vincent’s Clinical School, University of New South Wales, Sydney, Australia.

5State Key Laboratory of Stem Cell and Reproductive Biology, Institute of Zoology, Chinese Academy of Sciences, and

6Beijing Institute for Stem Cell and Regenerative Medicine, Beijing, China.

Address correspondence to: Xiangjia Zheng, No.22 Qi Xiang Tai Road, Tianjin 300070, China. Phone: 86.22.833.6835; Email: xzheng@tmu.edu.cn.

Authorship note: XY and STW contributed equally to this work.

Find articles by Dong, Z. in: JCI | PubMed | Google Scholar

1Department of Pharmacology and Tianjin Key Laboratory of Inflammation Biology, School of Basic Medical Sciences, and Center for Cardiovascular Diseases, Tianjin Medical University, China.

2State Key Laboratory of Experimental Hematology, Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Tianjin, China.

3Institute of Biochemistry, Justus Liebig University, Member of the German Center for Lung Research, Giessen, Germany.

4Vascular Epigenetics Laboratory, Victor Chang Cardiac Research Institute, and St. Vincent’s Clinical School, University of New South Wales, Sydney, Australia.

5State Key Laboratory of Stem Cell and Reproductive Biology, Institute of Zoology, Chinese Academy of Sciences, and

6Beijing Institute for Stem Cell and Regenerative Medicine, Beijing, China.

Address correspondence to: Xiangjia Zheng, No.22 Qi Xiang Tai Road, Tianjin 300070, China. Phone: 86.22.833.6835; Email: xzheng@tmu.edu.cn.

Authorship note: XY and STW contributed equally to this work.

Find articles by Wang, L. in: JCI | PubMed | Google Scholar

1Department of Pharmacology and Tianjin Key Laboratory of Inflammation Biology, School of Basic Medical Sciences, and Center for Cardiovascular Diseases, Tianjin Medical University, China.

2State Key Laboratory of Experimental Hematology, Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Tianjin, China.

3Institute of Biochemistry, Justus Liebig University, Member of the German Center for Lung Research, Giessen, Germany.

4Vascular Epigenetics Laboratory, Victor Chang Cardiac Research Institute, and St. Vincent’s Clinical School, University of New South Wales, Sydney, Australia.

5State Key Laboratory of Stem Cell and Reproductive Biology, Institute of Zoology, Chinese Academy of Sciences, and

6Beijing Institute for Stem Cell and Regenerative Medicine, Beijing, China.

Address correspondence to: Xiangjia Zheng, No.22 Qi Xiang Tai Road, Tianjin 300070, China. Phone: 86.22.833.6835; Email: xzheng@tmu.edu.cn.

Authorship note: XY and STW contributed equally to this work.

Find articles by Qi, C. in: JCI | PubMed | Google Scholar

1Department of Pharmacology and Tianjin Key Laboratory of Inflammation Biology, School of Basic Medical Sciences, and Center for Cardiovascular Diseases, Tianjin Medical University, China.

2State Key Laboratory of Experimental Hematology, Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Tianjin, China.

3Institute of Biochemistry, Justus Liebig University, Member of the German Center for Lung Research, Giessen, Germany.

4Vascular Epigenetics Laboratory, Victor Chang Cardiac Research Institute, and St. Vincent’s Clinical School, University of New South Wales, Sydney, Australia.

5State Key Laboratory of Stem Cell and Reproductive Biology, Institute of Zoology, Chinese Academy of Sciences, and

6Beijing Institute for Stem Cell and Regenerative Medicine, Beijing, China.

Address correspondence to: Xiangjia Zheng, No.22 Qi Xiang Tai Road, Tianjin 300070, China. Phone: 86.22.833.6835; Email: xzheng@tmu.edu.cn.

Authorship note: XY and STW contributed equally to this work.

Find articles by Wang, X. in: JCI | PubMed | Google Scholar |

1Department of Pharmacology and Tianjin Key Laboratory of Inflammation Biology, School of Basic Medical Sciences, and Center for Cardiovascular Diseases, Tianjin Medical University, China.

2State Key Laboratory of Experimental Hematology, Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Tianjin, China.

3Institute of Biochemistry, Justus Liebig University, Member of the German Center for Lung Research, Giessen, Germany.

4Vascular Epigenetics Laboratory, Victor Chang Cardiac Research Institute, and St. Vincent’s Clinical School, University of New South Wales, Sydney, Australia.

5State Key Laboratory of Stem Cell and Reproductive Biology, Institute of Zoology, Chinese Academy of Sciences, and

6Beijing Institute for Stem Cell and Regenerative Medicine, Beijing, China.

Address correspondence to: Xiangjia Zheng, No.22 Qi Xiang Tai Road, Tianjin 300070, China. Phone: 86.22.833.6835; Email: xzheng@tmu.edu.cn.

Authorship note: XY and STW contributed equally to this work.

Find articles by Schmitz, M. in: JCI | PubMed | Google Scholar |

1Department of Pharmacology and Tianjin Key Laboratory of Inflammation Biology, School of Basic Medical Sciences, and Center for Cardiovascular Diseases, Tianjin Medical University, China.

2State Key Laboratory of Experimental Hematology, Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Tianjin, China.

3Institute of Biochemistry, Justus Liebig University, Member of the German Center for Lung Research, Giessen, Germany.

4Vascular Epigenetics Laboratory, Victor Chang Cardiac Research Institute, and St. Vincent’s Clinical School, University of New South Wales, Sydney, Australia.

5State Key Laboratory of Stem Cell and Reproductive Biology, Institute of Zoology, Chinese Academy of Sciences, and

6Beijing Institute for Stem Cell and Regenerative Medicine, Beijing, China.

Address correspondence to: Xiangjia Zheng, No.22 Qi Xiang Tai Road, Tianjin 300070, China. Phone: 86.22.833.6835; Email: xzheng@tmu.edu.cn.

Authorship note: XY and STW contributed equally to this work.

Find articles by Liu, R. in: JCI | PubMed | Google Scholar

1Department of Pharmacology and Tianjin Key Laboratory of Inflammation Biology, School of Basic Medical Sciences, and Center for Cardiovascular Diseases, Tianjin Medical University, China.

2State Key Laboratory of Experimental Hematology, Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Tianjin, China.

3Institute of Biochemistry, Justus Liebig University, Member of the German Center for Lung Research, Giessen, Germany.

4Vascular Epigenetics Laboratory, Victor Chang Cardiac Research Institute, and St. Vincent’s Clinical School, University of New South Wales, Sydney, Australia.

5State Key Laboratory of Stem Cell and Reproductive Biology, Institute of Zoology, Chinese Academy of Sciences, and

6Beijing Institute for Stem Cell and Regenerative Medicine, Beijing, China.

Address correspondence to: Xiangjia Zheng, No.22 Qi Xiang Tai Road, Tianjin 300070, China. Phone: 86.22.833.6835; Email: xzheng@tmu.edu.cn.

Authorship note: XY and STW contributed equally to this work.

Find articles by Han, Z. in: JCI | PubMed | Google Scholar

1Department of Pharmacology and Tianjin Key Laboratory of Inflammation Biology, School of Basic Medical Sciences, and Center for Cardiovascular Diseases, Tianjin Medical University, China.

2State Key Laboratory of Experimental Hematology, Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Tianjin, China.

3Institute of Biochemistry, Justus Liebig University, Member of the German Center for Lung Research, Giessen, Germany.

4Vascular Epigenetics Laboratory, Victor Chang Cardiac Research Institute, and St. Vincent’s Clinical School, University of New South Wales, Sydney, Australia.

5State Key Laboratory of Stem Cell and Reproductive Biology, Institute of Zoology, Chinese Academy of Sciences, and

6Beijing Institute for Stem Cell and Regenerative Medicine, Beijing, China.

Address correspondence to: Xiangjia Zheng, No.22 Qi Xiang Tai Road, Tianjin 300070, China. Phone: 86.22.833.6835; Email: xzheng@tmu.edu.cn.

Authorship note: XY and STW contributed equally to this work.

Find articles by Wang, L. in: JCI | PubMed | Google Scholar

1Department of Pharmacology and Tianjin Key Laboratory of Inflammation Biology, School of Basic Medical Sciences, and Center for Cardiovascular Diseases, Tianjin Medical University, China.

2State Key Laboratory of Experimental Hematology, Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Tianjin, China.

3Institute of Biochemistry, Justus Liebig University, Member of the German Center for Lung Research, Giessen, Germany.

4Vascular Epigenetics Laboratory, Victor Chang Cardiac Research Institute, and St. Vincent’s Clinical School, University of New South Wales, Sydney, Australia.

5State Key Laboratory of Stem Cell and Reproductive Biology, Institute of Zoology, Chinese Academy of Sciences, and

6Beijing Institute for Stem Cell and Regenerative Medicine, Beijing, China.

Address correspondence to: Xiangjia Zheng, No.22 Qi Xiang Tai Road, Tianjin 300070, China. Phone: 86.22.833.6835; Email: xzheng@tmu.edu.cn.

Authorship note: XY and STW contributed equally to this work.

Find articles by Zheng, X. in: JCI | PubMed | Google Scholar |

Authorship note: XY and STW contributed equally to this work.

Published January 24, 2023 - More info

Published in Volume 8, Issue 5 on March 8, 2023
JCI Insight. 2023;8(5):e160372. https://doi.org/10.1172/jci.insight.160372.
© 2023 Yang et al. This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Published January 24, 2023 - Version history
Received: March 29, 2022; Accepted: January 20, 2023 View PDF Abstract

Loss-of-function mutations in cerebral cavernous malformation (CCM) genes and gain-of-function mutation in the MAP3K3 gene encoding MEKK3 cause CCM. Deficiency of CCM proteins leads to the activation of MEKK3-KLF2/4 signaling, but it is not clear how this occurs. Here, we demonstrate that deletion of the CCM3 interacting kinases STK24/25 in endothelial cells causes defects in vascular patterning during development as well as CCM lesion formation during postnatal life. While permanent deletion of STK24/25 in endothelial cells caused developmental defects of the vascular system, inducible postnatal deletion of STK24/25 impaired angiogenesis in the retina and brain. More importantly, deletion of STK24/25 in neonatal mice led to the development of severe CCM lesions. At the molecular level, a hybrid protein consisting of the STK kinase domain and the MEKK3 interacting domain of CCM2 rescued the vascular phenotype caused by the loss of ccm gene function in zebrafish. Our study suggests that CCM2/3 proteins act as adapters to allow recruitment of STK24/25 to limit the constitutive MEKK3 activity, thus contributing to vessel stability. Loss of STK24/25 causes MEKK3 activation, leading to CCM lesion formation.

Graphical Abstract Introduction

Mutations in the cerebral cavernous malformation (CCM) genes, KRIT1, CCM2, and PDCD10, cause CCM disease. The CCM genes encode KRIT1 (also referred to as CCM1), CCM2, and PDCD10 (also referred to as CCM3) proteins that act as adapter proteins and can form a single signaling complex (1, 2). Biochemical studies revealed that CCM1 interacts with CCM2 and that CCM2 interacts with CCM3 (3–6). The interaction with CCM1 induces a conformation change in CCM2 and enhances its interaction affinity with CCM3 (4). Mutations that disrupt CCM complex formation causes human diseases (7, 8). The CCM complex also interacts with other proteins; CCM2 interacts with MEKK3 and Rac, whereas CCM3 can complex with the kinases STK24/25 and MST4, which belong to the germinal-center kinase III (GCKIII) subfamily of kinases (4, 5, 9). These kinases appear to initiate 2 downstream signaling pathways, namely the MEKK3-KLF2/4 and STK24/5-Ezrin/Radixin/Moesin (STK24/5-ERM) signaling pathways (4, 10, 11). Loss of CCM genes leads to elevated MEKK3-KLF2/4-ADAMTS4/5 signaling (10, 11). A causative role of this elevated signaling pathway leading to the development of cardiovascular defects and CCM lesions was shown by the beneficial effects of pharmacological MEKK3 inhibition or genetic reduction of MEKK3, KLF, or ADAMTS signaling (10–13). In support of this, the recent identification of a gain-of-function mutation in MAP3K3 in patients with CCM emphasized a causative role of this kinase in CCM pathogenesis (14). In addition, experiments in zebrafish and in cultured HUVECs, suggest that the interaction between STK and CCM3 employs a ERM-RHO signaling pathway to regulate cardiovascular development (4).

In this study, we generated mice with floxed alleles of Stk24 and Stk25 to delete both genes in endothelial cells. Defective STK24/25 expression in endothelial cells caused defects in vascular development and CCM lesion formation, akin to the phenotype observed with Ccm gene deletions. Biochemical experiments suggest that STK24/25 prevents CCM pathogenesis through restriction of constitutive MEKK3 activity.

Results

Deletion of STK24/25 in endothelial cells restricts lumen formation of BAA and the DA. Mass spectrometry and biochemical studies have identified the GCKIII subfamily kinases STK24, STK25, and MST4 as binding partners of CCM3 (4, 5, 15). To determine the role of endothelial STK24 and STK25 in vascular development, we crossed the Stk24fl/fl and Stk25fl/fl mice with the Tie2-Cre mice to delete the Stk24 and/or Stk25 genes in the endothelial lineage. Both the Tie2-Cre;Stk24fl/fl and Tie2-Cre;Stk25fl/fl mice were born at expected numbers and without an overt phenotype. Further genetic analyses showed that the simultaneous deletion of Stk24 and Stk25 in endothelial cells (the Tie2-Cre;Stk24fl/fl;Stk25fl/fl mice, hereafter referred to as the Stk24/25dECKO) led to embryonic lethality. Timed mating revealed that the Stk24/25dECKO embryos died before E11, while Tie2-Cre;Stk24fl/fl;Stk25fl/+ and Tie2-Cre;Stk24fl/+;Stk25fl/fl littermates were unaffected and appear grossly normal at E11.5 or E12 (Figure 1A, Supplemental Tables 1 and 2, and Supplemental Figure 1; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.160372DS1). Histologic analysis revealed that Stk24/25dECKO failed to form patent branchial arch arteries (BAA) and the dorsal aorta (DA) (Figure 1B and Supplemental Figure 2A). This restricted formation of BAA and DA prevent the formation of a functional circulation system. Injection of Indian ink and subsequent analysis of its distribution showed that the injected ink was confined to the heart in the Stk24/25dECKO embryos at E10 (Figure 1C and Supplemental Figure 2B). Whole-mount staining of endoglin confirmed the restricted DA (Figure 1C) and also showed the mispatterning of brain vasculature (Figure 1D and Supplemental Figure 3). The developmental patterning of BAA, DA, and brain vasculature in mice with deletion of only Stk24 (Tie2-Cre;Stk24fl/fl;Stk25fl/+) were normal (Supplemental Figures 1–3). These data indicate that double deletion of Stk24 and Stk25 is required to induce a vascular phenotype similar to that of Ccm gene deficiency (16, 17). These data suggests STK24 and STK25 complement each other and play a critical role in the CCM pathway to regulate the lumenization of the BAA and the DA as well as the patterning of the brain vasculature.

Figure 1

Deletion of Stk24 and Stk25 in endothelium results in vascular defects during embryonic development. (A) Stereomicroscopic images of developmental time course of littermate Stk24fl/fl;Stk25fl/fl and Stk24/25dECKO mice. Scale bars: 1 mm. (B) H&E staining and Pecam immunostaining of transverse sections of E10 Stk24fl/fl;Stk25fl/fl (n = 3) and Stk24/25dECKO (n = 4) embryos reveal the presence of normally lumenized dorsal aortas (DA) in the Stk24fl/fl;Stk25fl/fl embryos but not in Stk24/25dECKO embryos. White arrowheads indicate dorsal aortas. Scale bars: 100 μm. (C) Images of E9.5 embryo hearts of Stk24fl/fl;Stk25fl/fl (n = 9) and Stk24/25dECKO (n = 6) embryos with injection of Indian ink. Upper and middle panels show embryo overview and magnification of the boxed regions, showing injected ink flows primarily through the second and third BAA to fill the DA in the Stk24fl/fl;Stk25fl/fl embryos. In contrast, ink injected into the heart of Stk24/25dECKO embryos failed to opacify the DA. Accumulation of ink was observed in the heart due to the narrow BAA. Scale bars: 500 μm. Lower panel shows whole-mount immunostaining for the endothelial cell marker, endoglin, showing narrowed BAA and adjacent DA (red arrows) in Stk24/25dECKO embryos. Scale bars: 100μm. (D) Whole-mount immunostaining with endoglin showing the impaired vascular patterning (indicated by the white arrows) in the brain of E9.5 Stk24/25dECKO (n = 6) embryos in comparison with that of Stk24fl/fl;Stk25fl/fl (n = 3) littermate control embryos. Scale bars: 500 μm in overview panels and 100 μm in magnified panels. All the images presented are representatives of 3 or more independent experiments.

Induced postnatal deletion of Stk24/25 impairs angiogenesis and confers CCM lesion formation. Due to the lethality of the Stk24/25dECKO embryos during early embryonic development, we generated a genetic system that allow for the inducible ablation of these genes. To achieve this, the Stk24fl/fl;Stk25fl/fl mice were crossed with Cdh5-CreERT2 mice to generate the Cdh5-CreERT2;Stk24fl/fl;Stk25fl/fl mice (denoted as Stk24/25idECKO). Induction of Stk24/Stk25 deletion at P2 through intragastric injection of 4-hydroxytamoxifen (4-HT) led to decreased retinal vessel outgrowth (Figure 2, A–C) as well as defective remodeling of the retinal venous branches and peripheral vessel remodeling (Figure 2, D–G). The number of endothelial cells in the malformed areas in the Stk24/25idECKO mice were increased compared with littermate controls, but the number of proliferatively active cells (as determined by phosphorylated histone 3 [pH3] staining) were not increased (Figure 2, H–J). Tip cells of the retina vasculature of Stk24/25idECKO mice showed increased filopodia protrusions (Figure 2, K–M, and Supplemental Figure 4). Imaging of the brain vessels in vibratome sections revealed vasculature mispatterning and the presence of small cavernous vessels in the cerebrum and the cerebellum of the Stk24/25idECKO mice (Figure 2, N–Q, and Supplemental Figures 5 and 6).

Figure 2

Deletion of Stk24 and Stk25 in endothelium of newborn pups disrupt retina vascular extension and filopodia formation. (A and B) Whole-mount staining of retinal vasculature with IsoB4 in Stk24fl/fl;Stk25fl/fl (n = 5) and Stk24/25idECKO (n = 8) mice at P6. The white line highlights the total retinal area. Scale bars: 500 μm. (C) Quantitative analysis shows reduced retinal vascular outgrowth in Stk24/25idECKO mice compared with that of Stk24fl/fl;Stk25fl/fl mice. Each data point represents 1 mouse. (D–G) IsoB4 whole-mount stainings of P6 retinas show vascular remodeling close to arteries and veins (D and E) and peripheral vessel plexus (F and G) in Stk24fl/fl;Stk25fl/fl mice (n = 4) and Stk24/25idECKO mice (n = 4). A indicates arteries; V denotes veins. Scale bars: 100 μm. (H and I) Confocal images of IsoB4, pH3, and Erg costaining of P6 retina in Stk24fl/fl;Stk25fl/fl mice (n = 6) and Stk24/25idECKO mice (n = 6). The white circle indicates the pH3+ ECs. Scale bars: 100 μm. (J) Quantitative analysis shows increased Erg+ cells but decreased pH3+ ECs in Stk24/25idECKO mice compared with Stk24fl/fl;Stk25fl/fl mice. Each data point represents 1 mouse. (K and L) Confocal images of P6 retina lobe stained with IsoB4 in Stk24/25idECKO mice (n = 5) compared with Stk24fl/fl;Stk25fl/fl mice (n = 3). The red dots denote filopodia in the vascular front. Scale bars: 100 μm. (M) Quantitative analysis showing increased filopodia numbers in Stk24/25idECKO mice compared with Stk24fl/fl;Stk25fl/fl mice. (N–Q) Confocal images and magnification of IsoB4 staining in the cerebral cortical vasculature (N and O) and cerebellum (P and Q) of P8 mouse pups (n = 4 for both Stk24/25idECKO and Stk24fl/fl;Stk25fl/fl). The white arrows indicate the malformation vessel. Scale bars: 100 μm. All the images presented are representatives of 3 or more independent experiments. The quantitative data are presented as mean ± SD, and significance was determined using unpaired t test. ***P < 0.001, **P < 0.01.

We next determined whether the loss of Stk24/25 could also cause CCM as previous reported in Ccm gene deficient mice (18, 19). We again induced Stk24/25 gene deletion in Stk24/25idECKO mice by intragastric injection of 4-HT at P2 (Figure 3A). Robust CCM lesions were detected in both the cerebrum and the cerebellum of the Stk24/25idECKO mice starting from 5 days after 4-HT induction as determined by μCT (Figure 3A and Supplemental Figure 7) and histology (Figure 3, B and C). Malformed cavernous vessels were only detected in the brain and retina and were not found in other peripheral organs such as the lung, liver, or testes (Supplemental Figure 8). Administration of 4-HT to Stk24/25idECKO mice at P2 only allowed survival of the pups up to P10 (Figure 3D). The few Stk24/25idECKO mice that survived up to P10 displayed severe hemorrhage in the cerebellum (Figure 3A). Our data show that a loss of Stk24/25 in endothelial cells caused a more severe phenotype than in previously reported models of inducible endothelial deletion of Ccm1/2/3 by Cdh5-CreERT2 (10, 18, 20).

Figure 3

Development of cerebral cavernous malformations in the Stk24/25idECKO mice. (A) Schematic of 4-hydroxytamoxifen (HT) injection and sample collection. Pups were intragastrically injected with 4-HT at P2, and the brain tissues were harvested at specific time points (P7, P8, and P9). Stereomicroscopic images and μCT images of CCM lesions in control and the Stk24/25idECKO mice at P7, P8, and P9. Scale bars: 2 mm. Quantitative analysis of lesion volume in the cerebrum (n = 4) and cerebellum (n = 5) in the Stk24/25idECKO mice is shown on the right. Data in the quantitative plots are presented as mean ± SD. (B) H&E staining of brain sections in Stk24fl/flStk25fl/fl mice (n = 3) at P8 and Stk24/25idECKO mice (n = 3) at P7, P8, and P9. CCM lesions are shown as red masses. Scale bars: 100 μm. (C) Immunostainings of Pecam show cavernomas and reduced vascular number in Stk24/25idECKO mice (n = 4) compared with Stk24fl/fl;Stk25fl/fl mice (n = 4). The white arrowheads indicate malformed vessels. Scale bars: 50 μm. (D) The survival curve of Stk24/25idECKO (n = 20) and Stk24fl/fl;Stk25fl/fl (n = 21) mice after 4-HT induction at P2. Statistical analysis was performed using the Mantel-Cox test. ****P < 0.0001. Representative images from at least 3 or more independent experiments are shown.

The induction window of CCM lesion formation is limited to the first week of postnatal life when using established models of endothelial deletion of Ccm1/2/3 genes with Cdh5-CreERT2. Since the Stk24/25idECKO mice showed such a severe CCM lesion burden, we investigated whether a delayed deletion of Stk24/25 expression would still lead to the formation of the CCM lesions. When 4-HT was administered at P5, about 40% the Stk24/25idECKO mice survive up to P23 (Figure 4A) and robust CCM lesions were detected in brains at P12 and P21 (Figure 4B and Supplemental Figure 9A). In comparison with those mice induced at P2, the CCM lesion burdens in the P5 induced Stk24/25idECKO mutants were less severe in both the cerebrum and cerebellum (Figure 4B). When induced at P7, CCM lesions also developed in the brain and retina of Stk24/25idECKO mice at P21 and P24 (Figure 4, C and D). However, Stk24/25 gene deletion at P10 and P15 resulted in no detectable CCM lesion when imaged up to P30 (Figure 4, E and F, and Supplemental Figure 9, B and C). These results indicate that the induction window of Cdh5-CreERT2–driven Stk24/25 deletion was also limited to the first weeks after birth, similar to that of the Cdh5-CreERT2–driven Ccm1/2/3 mutant mice, with more robust lesion formation with Stk24/25 loss.

Figure 4

Limited induction time window of CCM formation in the Stk24- and Stk25-deficient mice. (A) Schematic of 4-HT injection and sample collection. Pups were intragastrically injected with 4-HT at P5, and the brain tissues were harvested at P12 and P21. The survival curves of Stk24/25idECKO (n = 14) and Stk24fl/fl;Stk25fl/fl mice (n = 16) after 4-HT induction at P5 are shown below. Statistical analysis was performed using the Mantel-Cox test. ****P < 0.0001. (B) Stereomicroscopic images and μCT images of CCM lesions in Stk24/25idECKO mice at P12 and P21 with 4-HT induction at P5. Scale bars: 2 mm. Quantitative analysis of lesion volume in cerebrum (n = 3) and cerebellum (n = 3) at P12 and P21 is shown on the right. Data are presented as mean ± SD. (C) Stereomicroscopic images and μCT imaging of CCM lesions in Stk24/25idECKO (n = 3) mice after tamoxifen induction at P7. Scale bars: 2 mm. (D) H&E staining of brain sections and whole-mount images showing CCM in brain and retina of Stk24fl/fl;Stk25fl/fl (n = 3) and Stk24/25idECKO mice (n = 3) at different time points. Scale bars: 200 μm. (E) Stereomicroscopic images and μCT images showing diminished CCM lesion formation in the Stk24/25idECKO mice (n = 3) at P30 after tamoxifen induction starting at P10. Scale bars: 2 mm. (F) Stereomicroscopic images and μCT imaging showing near absence of CCM lesion formation in the Stk24/25idECKO mice (n = 3) at P29 after tamoxifen induction starting at P15. Scale bars: 2 mm. Representative images from 3 or more independent experiments are shown.

STK24/25 function upstream of MEKK3-KLF2/4 signaling. MEKK3-KLF2/4 and bone morphogenetic protein (BMP) signaling has been reported to function downstream of CCM signaling for CCM lesion formation (10, 20, 21). As a first step, we analyzed the relative mRNA expression levels of genes encoding various components of these signaling pathways by quantitative PCR (qPCR). We treated P2 control and Stk24/25idECKO mice with 4-HT and isolated P6 endothelial cells from these mice. The Stk24/25idECKO endothelial cells demonstrated significantly enhanced MEKK3 signaling, as shown by the increased expression of downstream effectors, Klf2/4 and Adamts1, over littermate controls (Figure 5A). Among the genes implicated in BMP and endothelial-mesenchymal transition (EndoMT) signaling, the expression of Snai2 and Sca1 were also significantly increased, while Bmp4 expression was decreased (Figure 5B). The upregulation of Klf4 expression was also evident at the protein level, shown by the increased immunostaining in endothelial cells (Figure 5, C and D).

Figure 5

Stk24 and Stk25 deletion induces CCMs via MEKK3-KLF2/4 signaling activation. (A and B) Relative mRNA expression level of CCM-related genes in endothelial cells isolated from control and Stk24/25idECKO mice at P6 after induction at P2. n = 5 for each group — except Bmp2, Bmp4, and Bmp6, for which n = 3 was used. (C) Pecam and Klf4 immunofluorescence staining in endothelial cells of Stk24fl/fl;Stk25fl/fl (n = 3) and Stk24/25idECKO mouse brains (n = 3) at P6 after induction at P2. Scale bars: 100 μm. (D) Quantitative analysis showing increased Klf4+ EC in Stk24/25idECKO mice compared with Stk24fl/fl;Stk25fl/fl mice. The quantitative data (mean ± SD) from 3 independent experiments are reported, and significance was determined using unpaired t test. **P < 0.01. (E) Schematic representation of the interaction between MEKK3 and STK25-CCM2 hybrid protein consists of N terminal kinase domain of STK and C terminal MEKK3 interacting domain of CCM2. (F) Immunoprecipitation experiment shows that STK25/CCM2 hybrid protein interaction with MEKK3 was comparable with that of CCM2. (G) Representative images of in situ staining of cmlc2 and fluorescence imaging of the hearts of Tg (cmlc2:EGFP) zebrafish embryos in which myocardial cells express EGFP. The ccm2 morpholino induced dilated heart, while coinjection with mRNA expressing STK25-CCM2 rescued the dilated heart phenotype compared with injection of mRNA only expressing STK25(1-302) or STK25K49R-CCM2 hybrid protein. Scale bars: 100 μm. (H) Quantification of cmlc2+ cardiac area and cmlc2:EGFP area of zebrafish embryos with ccm2 morpholino and different cRNA. Data are presented as mean ± SD, and significance was determined using unpaired t test (A, B, and D) or 1-way ANOVA (H). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

CCM3 interacts with STK24/25, and CCM2 interacts with MEKK3, while the interaction between CCM2 and CCM3 is essential to prevent CCM lesion formation (2, 4, 5, 9). However, it is unknown whether a close proximity between STK24/25 and MEKK3 is required for the suppression of MEKK3 activation. To test this possibility, we generated a hybrid protein consisting of the N-terminal kinase domain of STK25 and the C-terminal domain of CCM2, which mediates the interaction between CCM2 and MEKK3, as schematically shown in