Significance
Sturge–Weber Syndrome (SWS) is a neurocutaneous disorder that involves atypical blood vessel overgrowth in the skin, brain, and eye. It is associated with facial capillary malformations (CM) (aka port wine birthmark), leptomeningeal CM in the brain visible with MRI, and glaucoma. The neurological sequalae involve seizures, cerebral atrophy and calcification, and intellectual disorders. Currently, there are no molecularly targeted therapies for nonsyndromic CM or SWS. Our study shows the involvement of MAPK pathway and the pro-angiogenic molecule ANGPT2 in endothelial permeability and suggests a path to target in GNAQ p.R183Q-driven CM.
Keywords: capillary malformation, GNAQ, endothelial permeability, angiopoietin-2, MAPK signaling
Abstract
Nonsyndromic capillary malformations (CM) are seen predominantly in skin. In Sturge–Weber Syndrome (SWS), CMs occur in the skin, leptomeninges of the brain, and choroid of the eye. >90% of CM are caused by a somatic mutation—GNAQ p.R183Q, the gene encoding the G-protein subunit Gαq. Longitudinal MRI of the brain in one SWS patient suggests developing vascular permeability. We modeled this in a transendothelial electrical resistance assay and found endothelial cells with GNAQ p.R183Q (EC-R183Q) exhibited increased permeability compared to EC wild-type. Increased vascular permeability was confirmed in a Gnaq p.R183Q mouse model. Knockdown of elevated angiopoietin-2 (ANGPT2) in EC-R183Q partially restored the EC barrier, as did a MEK1,2 inhibitor, implicating MAPK/ERK signaling. The combination of ANGPT2 knockdown and trametinib further restored the EC barrier in an additive manner. indicating the two operate in separate pathways. In summary, we found that EC-R183Q exhibits increased permeability, reflecting the compromised endothelial barrier in CMs.
Capillary malformation (CM) is a slow-flow vascular malformation consisting of abnormally enlarged blood vessels with uneven mural cell coverage (1–4). In most cases, CM are nonsyndromic and confined to skin. In the rare neurocutaneous disorder Sturge–Weber Syndrome (SWS), facial CMs coincide with CM in the leptomeninges and the choroid of eye (5). The majority of SWS patients are at high risk for seizures, stroke-like episodes, as well as cognitive impairment (6). Only symptomatic treatments are available and to date there are no targeted molecular therapies for nonsyndromic or syndromic CM.
In 2013, a somatic mutation in GNAQ (p.R183Q) was found in 92% of nonsyndromic skin CMs and 88% of SWS skin and brain specimens (7). The mutant allelic frequency in CM affected tissues ranged from 1 to 18.1% (7), underscoring the mosaicism of the somatic mutation. To date, multiple studies have confirmed the presence of the GNAQ p.R183Q mutation (8–10). A somatic missense mutation (p.R183C) in GNA11, a GNAQ homolog, was found in patients with diffuse CM on extremities and milder neurological symptoms, suggesting a possible genotype–phenotype correlation (11, 12). Our laboratory showed the GNAQ p.R183Q is enriched in endothelial cells (ECs) in skin and brain CMs (13, 14).
GNAQ encodes Gαq, an alpha subunit of heterotrimeric G proteins that link G-protein-coupled receptors (GPCR) to primarily activate phospholipase C (PLC)β3. GPCR activation causes Gαq to dissociate from the βγ subunits (15) and increase the catalytic rate constant of PLCβ3 (16). Activated PLCβ3 hydrolyzes phosphatidylinositol 4,5 biphosphate (PIP2) to generate diacylglycerol (DAG) and inositol 1,4,5 triphosphate (IP3). PIP2 has long been established as a major contributor to the intracellular signaling stimulated by PLCβ (17). DAG activates protein kinase C (PKC), which plays a role in various cellular processes including activation of MAPK signaling (16). IP3 triggers the release of calcium ions from intracellular stores in the endoplasmic reticulum, which also leads to signaling and cellular responses. Gαq has intrinsic guanosine-5′-triphosphate hydrolase (GTPase) activity. Loss of arginine (R) at position 183, located in the conserved GTP-binding domain, is predicted to reduce GTPase activity and thereby increase GTP bound Gαq, which results in increased activity and increased signaling (18).
The GNAQ p.Q209L mutation in uveal melanoma, an aggressive cancer in the eye, causes a stronger activation of the Gαq GTPase activity (19). A protein–protein interaction network analysis suggests that p.R183Q and p.Q209L may have distinct downstream effectors leading to distinct cellular responses (20, 21). Nonetheless, the p.Q209L variant provides a template for understanding the consequences of these activating mutations (22). In contrast, the pathogenic downstream signaling pathways activated by GNAQ p.R183Q are not fully delineated. This gap in knowledge needs to be closed to advance our understanding of CM in both nonsyndromic and syndromic SWS cases.
Previous studies demonstrated that human ECs expressing the GNAQ p.R183Q mutation show activation of PLCβ3 and PKC, leading to increased angiopoietin-2 (ANGPT2), a potent angiogenic factor that promotes vascular permeability and sprouting. Furthermore, the p.R183Q ECs formed enlarged CM-like vessels in a murine xenograft model. Suppression of ANGPT2 using short hairpin RNA (shRNA) prevented the enlarged vessel phenotype, indicating its role in CM (18). Increased ANGPT2 has been reported in other cellular models of GNAQ p.R183Q ECs as well (1, 18). Analysis of brain histopathology sections has shown that CMs are associated with increased extravascular fibrin surrounding the enlarged and irregular blood vessels (4). To study the apparent increase in vascular permeability, we analyzed the ability of p.R183Q-expressing ECs to form electrically resistant endothelial barriers in vitro. Consistently, the EC-R183Q formed significantly weaker endothelial barriers, i.e., showed increased permeability, compared to EC-wild-type (EC-WT). Furthermore, we modeled the somatic mosaicism in CM by titrating EC-R183Q into EC-WT cultures and learned that a small percentage of mutant ECs have an outsized effect on the endothelial barrier. We used shRNA and small interfering RNA (siRNA) to examine roles of ANGPT2 and MEK1 (aka MAP2K1) in endothelial barrier formation and tested drugs against these molecules as orthogonal approaches. Taken together, our study pinpoints MEK1 and ANGPT2 as potential targets to reverse the cellular consequences of the mutant Gαq in CMs.
Results
Weakened Endothelial Barrier Formation in CM and in p.R183Q Endothelial Cells.
In 1977, it was hypothesized that cerebral vessels in SWS have abnormal vessel permeability (23). Here and in our previous study, we showed by Martius Scarlet Blue staining that CMs in the brain have increased extravascular fibrin, suggesting permeability or leakiness (4). We extend these findings to skin and brain CMs; both have increased extravascular fibrin compared to their non-CM controls (Fig. 1A). Brain MRI studies with intravenous gadolinium contrast administration can provide indirect evidence of vascular permeability abnormalities in patients with SWS (24). Longitudinal MRI studies on a SWS patient at 1 and 15 mo of age allows visualization of abnormally prominent and tortuous small vessels, coupled with abnormal diffuse contrast enhancement of the adjacent leptomeninges overlying the cortex seen at 15 mo of age, which strongly suggests altered capillary permeability where the enhancement indirectly reflects leakage of gadolinium across the compromised blood–brain barrier, highlighting the developing functional abnormality of these vessels (Fig. 1B). Consistent with this, we found increased vascular permeability in a mouse model of Gnaq-driven CM (Fig. 1C). Adult transgenic Tg-Cdh5CreERT2+/−;R26GT-Gnaq-GFP/+ mutant mice (25), versus control mice (Tg-Cdh5CreERT2+/− and/or R26GT-Gnaq-GFP/+) were given Evans Blue dye by tail vein injection one week after tamoxifen injections (Fig. 2D). Twenty minutes later, mice were euthanized and dissected according to the approved procedure by the Boston Children’s Institutional Animal Care and Use Committee. Lungs, intestines, and brain were imaged. The highly visible blue coloration of the lungs and intestines in the Tg-Cdh5CreERT2+/−;R26GT-Gnaq-GFP/+ mutant mice indicates increased vascular permeability (Fig. 1E); however, we did not observe Evans blue leakage in brain (SI Appendix, Fig. S1).
Fig. 1.
Vascular permeability in capillary malformations (CM). Martius Scarlet Blue staining of (A) CM brain and skin (Top), and non-CM brain and skin (Bottom). Fibrin (red), collagen (blue), erythrocytes (yellow), and nuclei (blue/black). N = 4. (Scale bar, 200 μm.) (B) Brain MRI studies of SWS patient at 1 mo and 15 mo of age revealing progressive atrophy of the Right cerebral hemisphere with associated progressive enlargement and tortuosity of subarachnoid small vessels (yellow arrows) and associated progression of abnormal leptomeningeal enhancement along the cortical surface (red arrowheads). (C) Breeding strategy of mutant Tg-Cdh5CreERT2+/−;R26GT-Gnaq-GFP/+ and control Tg-Cdh5CreERT2+/− and R26GT-Gnaq-GFP/+ mice. (D) Timeline of tamoxifen injections for Cre-mediated recombination prior to Evans Blue tail vein injections. Schematic created in BioRender. (E) Increased Evans blue extravasation in lung and small intestine in Tg-Cdh5CreERT2+/− (control mice, Top) and R26GT-Gnaq-GFP/+ Tg-Cdh5CreERT2+/− (mutant mice, Bottom). (Scale bar, 50 mm.)
Fig. 2.
GNAQ p.R183Q ECs form reduced barrier compared to wild-type ECs. (A) Schematic of CRISPR editing to generate the p.R183Q mutation in ECs. (B) Western blot for Angiopoietin2 (ANGPT2, 69 kDa), pERK1/2 (42 to 44 kDa), ERK1/2 (42 to 44 kDa), and Gαq (40 kDa), in EC-WT versus EC-R183Q grown in 2% FBS- (Left) or 0% FBS-containing media (Right) media for 24 h. Tubulin (50 kDa) served as a loading control. Full blots and their quantification are in SI Appendix, Fig. S3. (C) Schematic of Gαq dissociation from βγ subunits, blocked by the Gαq YM254890 inhibitor (26). (D) Absolute endothelial resistance (Ω) in EC-WT (black open circles), EC-R183Q (red open circles), EC-WT treated with YM254890 (black closed circles), EC-R183Q treated with YM254890 (red closed circles) measured over 26 h. (E) Quantification of percent change in resistance at 20 h ± YM354890 (100 nM). P-values were calculated by Brown–Forsythe and Welch ANOVA tests followed by Dunnett T3 multicomparison test. (D and E) n = 4 wells/condition were measured. Three independent experiments were performed. A representative experiment shown.
To assess the vascular permeability and its causes, human ECs were CRISPR-edited to contain a single copy of the p.R183Q allele at the endogenous GNAQ locus. EC-WT were mock edited in parallel (Fig. 2A and SI Appendix, Fig. S2). We confirmed increased ANGPT2 (18) in the EC-R183Q versus EC-WT grown in either reduced serum (2% FBS) (P = 0.0448) or serum-free (0% FBS) (P = 0.058). pERK1/2, ERK1/2, and Gαq are shown for comparison. (Fig. 2B and SI Appendix, Fig. S3). We measured permeability in the EC-R183Q and EC-WT using the TEER assay to quantify the EC–EC barrier formation. EC-R183Q displayed a sustained reduction (−36.54 ± 0.75%; P < 0.0001) in TEER compared to EC-WT (Fig. 2 D and E). The results were similar when the TEER was conducted in 2% FBS (−34.64 ± 0.55%) versus 0% (−44.43 ± 0.54%) FBS (SI Appendix, Fig. S4 B and C). Using a second human EC model (18), we confirmed that the mutant ECs formed a reduced barrier (−15.47 ± 0.63%, P < 0.0001) compared to the EC-WT (SI Appendix, Fig. S5). To verify the reduced barrier was a consequence of Gαq activity, we conducted the TEER experiment ± the Gαq inhibitor, YM254890 (100 nM) (26–28) (Fig. 2C). YM254890 restored resistance (Ω) in EC-R183Q to EC-WT levels over the 26-h period of TEER measurements (Fig. 1D). YM254890 had no effect on the TEER in EC-WT (2.69 ± 0.44%), possibly because the EC–EC contacts were fully established (Fig. 2E). YM254890 fully restored TEER in EC-R183Q (P = 0.1458) to level seen in EC-WT at 20 h (Fig. 2E). This experiment establishes a role for the R183Q mutant Gαq in compromised endothelial barrier formation.
Modeling the Somatic Mosaicism in CM.
EC populations isolated from CM specimens present mutant allelic frequencies up to 21%, reflecting the mosaic nature of the GNAQ p.R183Q mutation in CM (14). To assess the impact of mosaicism on endothelial permeability, we titrated EC-R183Q into EC-WT cultures, ranging from 0 to 100%. Increasing percentages of EC-R183Q reduced TEER over 24 h (Fig. 3A). Quantification at 20 h showed that 10% EC-R183Q reduced TEER significantly (−14.78 ± 0.37; P < 0.0001), suggesting that a small percentage of mutant ECs can have a detrimental effect on barrier formation (Fig. 3B). Indeed, a trend toward increased permeability was seen at 5% EC-R183Q. We confirmed the mutant allelic frequency in the titration experiment by droplet digital PCR (SI Appendix, Fig. S6). We also confirmed reduced TEER in ECs isolated from two different CM specimens (SI Appendix, Fig. S7). Overall, increased permeability was observed in three cellular models of endothelial GNAQ p.R183Q, including in CM patient ECs. We next analyzed pERK (P = 0.060) and ANGPT2 (P = 0.0016) levels in cell lysates and found both increased in cultures with 50% EC-R183Q (Fig. 3 C–E and SI Appendix, Fig. S8). We also determined ANGPT2 gene expression levels (Fig. 3F) and ANGPT2 protein levels measured by ELISA (Fig. 3G). Both were significantly increased in cultures with 25% EC-R183Q. These results indicate that reduction in the endothelial barrier can occur before measurable increases in ANGPT2 or pERK (Fig. 3H).
Fig. 3.
Modeling p.R183Q mosaicism. (A) Titration experiments were performed by adding EC-R183Q from 0 to 100% into EC-WT cultures. Absolute endothelial resistance (Ω) was measured over 26 h. (B) Quantification of percent change in resistance at 20 h. The P-values were calculated by Brown–Forsythe and Welch ANOVA test followed by Dunnett T3 multicomparison test. n = 4 wells/condition were measured. Three independent experiments were performed. A representative experiment shown. (C) WB for Gαq (pERK1/2, total ERK1/2 and ANGPT2 in EC-R183Q/EC-WT cultures. Tubulin served as loading control. (D and E) Quantification of pERK1/2 and ANGPT2, normalized to tubulin, in C. The P-values were calculated by One-Way ANOVA test followed by Dunnett’s multiple comparison test. Three independent experiments were performed. (F) ANGPT2 gene expression levels by qPCR at the end of TEER assay. The P-values were calculated by One-Way ANOVA test followed by Tukey’s multiple comparison test. n = 3. (G) Secreted ANGPT2 levels were measured by ELISA at the end of TEER experiment. ANGPT2 levels are normalized to 0% EC-R183Q levels. The P-values were calculated by One-Way ANOVA test followed by Tukey’s multiple comparison test. n = 4. (H) Schematic showing increasing EC-R183Q coincides with increases in ANGPT2 and pERK and reduced TEER.
MAPK Pathway Is Involved in Gαq-R183Q Endothelial Permeability.
PKC is one of the major effectors downstream of Gαq, and it is known to be involved in endothelial tight junctions (29). Therefore, we tested a pan-PKC inhibitor, AEB071 (30), in the TEER assay (Fig. 4A) and found significant rescue of TEER (23.9 ± 0.8%, P < 0.0001) in EC-R183Q at 20 h compared to untreated EC-R183Q (Fig. 4B). One of the downstream targets of PKC is MAPK signaling, and thus, we used an inhibitor of MEK, trametinib, which is currently used to treat fast-flow vascular malformations (31–33) as well as metastatic melanoma (34, 35). We measured the effects of increasing concentrations of trametinib (0.1, 0.5, and 1 µM) in the TEER assay (Fig. 4C). Significant yet partial changes were detected from 45.48 ± 3.37 to 40.98 ± 2.12% at 20 h in EC-R183Q (Fig. 4D). Trametinib also reduced ANGPT2 levels in a dose-dependent manner, with 0.1 μM trametinib causing significantly reduced ANGPT2 protein (P = 0.0038) compared to untreated EC-R183Q. Similarly, 0.5 μM (P = 0.0002) and 1 μM (P = 0.0016) trametinib significantly reduced ANGPT2 levels (SI Appendix, Fig. S9 A and B). All three concentrations of trametinib reduced pERK significantly (P < 0.001), as expected (SI Appendix, Fig. S9 A and C). The elevated level of Gαq in EC-R183Q was not significantly changed by increasing doses of trametinib (SI Appendix, Fig. S9). Overall, these results suggest the MAPK pathway in EC-R183Q is involved in endothelial barrier formation and may contribute to elevated ANGPT2 levels in EC-R183Q. To confirm effects seen with trametinib, we tested two additional MAPK inhibitors: U0126, which selectively inhibits the activation of MEK1/2 (36) and RDEA119, a MEK inhibitor also known as Refametinib (37). U0126 (2.5 µM) rescued the deficit in the endothelial resistance in the TEER assay by 32.4 ± 3.29%, (P = 0.021) in EC-R183Q at 20 h (SI Appendix, Fig. S10 A and B). RDEA119 (2.5 µM) also rescued the endothelial resistance in the TEER assay by 29.2 ± 5.49%, (P = 0.0434) at 20 h (SI Appendix, Fig. S10 C and D).
Fig. 4.
PKC and MAPK pathways are involved in EC-R183Q barrier function. (A) Absolute endothelial resistance (Ω) in EC-WT (black open circles), EC-R183Q (red open circles), EC-WT treated with AEB071 (black circles), and EC-R183Q treated with AEB071 (red circles) measured over 26 h. (B) Quantification of percent change in resistance at 20 h EC-WT and EC-R183Q ± AEB071 (10 µM). The P-values were calculated by Brown–Forsythe and Welch ANOVA tests followed by Dunnett T3 multicomparison test. n = 4 wells/condition were measured. Three independent experiments were performed. A representative experiment shown. (C) Absolute endothelial resistance (Ω) in EC-WT (open symbols) and EC-R183Q (closed symbols) ± increasing concentrations of trametinib (0 to 1 µM) measured over 24 h. (D) Quantification of percent change in resistance at 20 h EC-WT and EC-R183Q ± trametinib. The P-values were calculated by Brown–Forsythe and Welch ANOVA tests followed by Dunnett T3 multicomparison test. n = 4 wells/condition were measured. Three independent experiments were performed. A representative experiment shown. (E) Schematic of MAP2K1 siRNA knock down experiment (F) MAP2K1 gene expression measured by qPCR 48 h after siRNA transfection. (G) Absolute endothelial resistance (Ω) in EC-WT siControl (gray open squares), EC-WT siMAP2K1 (red open circles or blue open triangles), EC-R183Q siControl (gray closed squares), EC-R183Q siMAP2K1 (red closed circles or blue closed triangles) measured over 24 h. (H) Quantification of percent change in resistance at 20 h EC-WT and EC-R183Q ± siControl and two different siMAP2K1. The P-values were calculated by Brown–Forsythe and Welch ANOVA tests followed by Dunnett T3 multicomparison test. n = 4 wells/condition were measured. Three independent experiments were performed. A representative experiment shown. (I) Summary of PKC and MAPK inhibitors that restored endothelial barrier function in EC-R183Q.
We next explored the consequences of MEK inhibition in EC-R183Q by knocking down MAP2K1 gene expression. Two siMAP2K1 (Fig. 4E) showed successful knockdowns of 72% and 69% (Fig. 4F). Cells were treated with siMAP2K1 or siControl for 48 h prior to measuring TEER (Fig. 4G). Similar to trametinib, partial yet significant changes were observed: the two siMAP2K1 rescued the deficit in TEER by 21.9 ± 3.64%, (P = 0.0069), and 23.47 ± 5.76%, (P = 0.0041), respectively (Fig. 4H). siMAP2K1 had no significant effect on TEER in EC-WT. These results suggest that PKC and MAPK pathways are involved in reduced endothelial barrier in the EC-R183Q (Fig. 4I).
ANGPT2 in Gαq-R183Q Endothelial Permeability.
It has been previously reported that the addition of ANGPT2 to primary murine brain ECs leads to decreased TEER (38). Since ANGPT2 is significantly increased in EC-R183Q, we tested its role in the reduced barrier formation observed in these cells. We first used siRNA to knockdown ANGPT2 (Fig. 5A), with two successful knockdowns of 72% and 62% (Fig. 5B). siANGPT2 or siControl was applied to EC-WT and EC-R183Q for 48 h prior to measuring TEER (Fig. 5A). The two siANGPT2 partially yet significantly rescued the deficit in TEER in the EC-R183Q by 50 ± 7.4%, (P < 0.0001) and 33 ± 1.66% (P = 0.0002), respectively (Fig. 5 C and D). This finding was confirmed with shRNA lentiviral knockdown of ANGPT2 (Fig. 5E); knockdown efficiency was determined by qPCR and found to be 54% (Fig. 5 E and F). Similar to siANGPT2, shANGPT2 partially yet significantly rescued the TEER deficit by 28.75 ± 2.86%, (P < 0.0001) in EC-R183Q at 20 h (Fig. 5 G and H). We next tested a function-blocking antibody against human ANGPT2 (REGEN910) (39, 40) (Fig. 5I). REGEN910 selectively binds to ANGPT2 and blocks its binding to the TIE2 receptor. We found REGEN910 had no effect on TEER in EC-WT or EC-R183Q. Quantification at 20 h showed no significant difference in EC-R183Q treated with control IgG compared to EC-R183Q treated with REGEN910 (7.49 ± 3.03% P = 0.3604) (Fig. 5 J and K and SI Appendix, Fig. S11). The activity of REGEN910 was verified (SI Appendix, Fig. S12) using phorbol 12-myristate 13-acetate stimulated normal human ECs. Overall, these results show that knockdown of endogenous ANGPT2 resulted in increased TEER whereas addition of a function blocking anti-ANGPT2 had no effect. This suggests intracellular ANGPT2—not accessible to the RGEN910—is responsible for the diminished endothelial barrier formation. Immunofluorescence staining showed strong intracellular ANGPT2 in the EC-R183Q, colocalized with TIE2, consistent with this possibility (Fig. 5L and SI Appendix, Fig. S13).
Fig. 5.
ANGPT2 contributes to the reduced endothelial barrier in EC-R183Q. (A) Schematic of siRNA knockdown experiment (B) ANGPT2 mRNA levels measured by qPCR at 48 h after siRNA transfection. (C) Absolute endothelial resistance (Ω) in EC-WT siControl (gray open squares), EC-WT siANGPT2 (red open circles or blue open triangles), EC-R183Q siControl (gray closed squares), EC-R183Q siANGPT2 (red closed circles or blue closed triangles) measured over 26 h. (D) Quantification of percent change in resistance at 20 h EC-WT and EC-R183Q ± siControl and two different siANGPT2. The P-values were calculated by Brown–Forsythe and Welch ANOVA tests followed by Dunnett T3 multicomparison test. n = 8 wells/condition were measured. Three independent experiments were performed. A representative experiment shown. (E) Schematic of shRNA knockdown experiment. (F) ANGPT2 gene expression levels by qPCR after shRNA transfection. (G) Absolute endothelial resistance (Ω) in EC-WT shControl (black open circles), EC-R183Q shControl (red open circles), EC-WT shANGPT2 (black closed circles), and EC-R183Q shANGPT2 (red closed circles) measured over 24 h. (H) Quantification of percent change in resistance at 20 h EC-WT and EC-R183Q ± siControl and shANGPT2. The P-values were calculated by Brown–Forsythe and Welch ANOVA test followed by Dunnett T3 multicomparison test. n = 4 wells/condition were measured. Three independent experiments were performed. A representative experiment shown. (I) Schematic of anti-human ANGPT2 (REGEN910) (39, 40) addition to EC-R183Q. Schematic created in Biorender. (J) Absolute endothelial resistance (Ω) in EC-WT (black open circles), EC-R183Q (red open circles), EC-WT treated with REGEN910 (black closed circles), and EC-R183Q treated with REGEN910 (red closed circles) measured over 26 h. (K) Quantification of percent change in resistance at 20 h EC-WT and EC-R183Q ± REGEN910 (24 nM). The P-value were calculated by Brown–Forsythe and Welch ANOVA test followed by Dunnett T3 multicomparison test. n = 4 wells/condition were measured. Three independent experiments were performed. A representative experiment shown. A representative experiment shown. (L) Immunostaining for ANGPT2 (yellow) TIE2 (magenta) and DAPI (cyan) in EC-WT (Left) and EC-R183Q (Right). Cells were treated overnight without serum. High magnification images (red box) are on the Right. (Scale bar, 15 μm.)
Combination of ANGPT2 Knockdown and Trametinib Effectively Rescues Reduced Permeability Phenotype in Gαq-R183Q Endothelial Cell.
Since MEK inhibitors (trametinib, U0126, RDEA119) and silencing MEK partially increased TEER in EC-R183Q (Fig. 4 and SI Appendix, Fig. S9) and knockdown of ANGPT2 had a similar partial effect (Fig. 5 and SI Appendix, Figs. S11 and S12), we hypothesized that combining the two would lead to a more complete rescue of endothelial barrier formation. To test this, EC-WT and EC-R183Q transduced with shANGPT2 were treated ± 1 μM trametinib. shControl for both EC-WT and EC-R183Q were treated ±1 μM trametinib in parallel. Levels of cell-associated ANGPT2 protein were evaluated by Western blot. We confirmed a significant (P = 0.0265) reduction of ANGPT2 protein levels with the knockdown (Fig. 6 A and B and SI Appendix, Fig. S14). We next found shControl-EC-R183Q cells treated with trametinib have noticeably reduced ANGPT2 (Fig. 6 A and B and SI Appendix, Fig. S13 and S14). These results are consistent with SI Appendix, Fig. S9 A and B where we observed trametinib significantly reduced ANGPT2 in naive EC-R183Q. These differences could potentially be from the manipulation of EC-R183Q with sh-lentiviral system. Importantly, shANGPT2 had no effect on pERK levels (Fig. 6 A and C and SI Appendix, Fig. S14), which places MAPK upstream of ANGPT2. As expected, trametinib decreased pERK levels (Fig. 6C). We next tested the shANGPT2-EC-R183Q treated ± trametinib in TEER assay (Fig. 6D). EC-WT-shControl ± trametinib were used as control. shANGPT2-EC-R183Q treated with trametinib showed significantly restored TEER at 20 h, with 67.45 ± 3.39%, (P < 0.0001) of the deficit erased, and TEER levels within 10.5 ± 1.55% of EC-WT (Fig. 6E). In contrast, shANGPT2 alone or trametinib alone rescued about 1/3 of the deficit in EC-R183Q (Fig. 6E). Overall, these results suggest that the combination of ANGPT2 knockdown and trametinib restores the barrier function in an additive manner.
Fig. 6.
Combined ANGPT2 and MAPK inhibition restores EC-R183Q endothelial barrier. (A) pERK1/2 (42 to 22 kDa), total ERK1/2 (42 to 22 kDa) and ANGPT2 (68 kDa) in EC-WT and EC-R183Q treated with shControl, shANGPT2, shControl treated with trametinib, and shANGPT2 treated with trametinib were detected by WB. Tubulin served as loading control. 1 µM trametinib was used. WB quantification of (B) ANGPT2 and (C) pERK1/2 normalized to tubulin. The P-values were calculated by Two-Way ANOVA test followed by Šídák’s multiple comparison test. n = 3. (D) Absolute endothelial resistance (Ω) in EC-WT (open symbols) and EC-R183Q (closed symbols) ± treated with shControl (gray squares), shANGPT2 (red circles), shControl treated with trametinib (blue triangles), shANGPT2 treated with trametinib (purple diamonds) measured over 24 h. (E) Quantification of percent change in resistance at 20 h EC-WT and EC-R183Q ± treated with shControl (gray squares), shANGPT2 (red circles), shControl treated with trametinib (blue triangles), shANGPT2 treated with trametinib (purple diamonds). The P-value were calculated by Brown–Forsythe and Welch ANOVA tests followed by Dunnett T3 multicomparison test. n = 8 wells/condition were measured. Three independent experiments were performed. A representative experiment shown.
Using telomerase-immortalized microvascular Gαq-EC-R183Q cells, Zecchin and colleagues showed that there is active calcium influx from outside the cells in the mutant ECs that was sensitive to the CRAC channel inhibitor CM4620 (41). We tested CM4620 at concentrations 0.125 to 1 μM in the TEER assay and found no effect on the endothelial barrier in EC-R183Q or EC-WT (SI Appendix, Fig. S15), suggesting that calcium flux into the cytoplasm does not affect endothelial barrier formation. We also tested tacrolimus (200 nM), an inhibitor of calcineurin pathway, and found no change in the TEER assay measurements (SI Appendix, Fig. S16). These two findings suggest that calcium and the calcineurin pathway play minimal roles in the reduced endothelial–endothelial barrier in EC-R183Q. Moreover, we tested the mTOR inhibitors everolimus and rapamycin in our TEER assay and found no effects on the EC-R183Q barrier function (SI Appendix, Fig. S17).
Discussion
In this study, we show that GNAQ p.R183Q expressing ECs (EC-R183Q) form an endothelial monolayer with increased permeability, indicated by reduced TEER, compared to EC-WT. Furthermore, we show that a small percentage of mutant EC in the endothelial monolayer increased permeability in an outsized manner, which may be relevant to the mosaic nature of CM vessels. Genetic knockdown and pharmacologic inhibition experiments show that ANGPT2 and MEK1 contribute to reduced endothelial barrier formation in the EC-R183Q (Fig. 7). We posit that the weakened endothelial barriers formed in the presence of EC-R183Q correlate with the extravascular fibrin seen in histological sections of CM and the apparent permeability seen by contrast MRI in a child with SWS.
Fig. 7.
Summary of the Gαq-R183Q pathway and role of downstream effectors in endothelial barrier function. Inhibitors with no effect on Gαq-R183Q endothelial barrier function shown in gray. Schematic created in BioRender.
Permeability abnormalities of the endothelium can be qualitatively and quantitatively assessed through brain MRI studies with intravenous gadolinium contrast administration. The characterization of abnormal leptomeningeal contrast enhancement in the affected regions of the brain serves as a strong qualitative indicator of altered capillary permeability, reflecting a compromised blood–brain barrier (24). In future studies, the degree of abnormal permeability could be further evaluated using advanced sequences where capillary permeability measurements (Ktrans) can be obtained using dynamic contrast-enhanced magnetic resonance perfusion (42–44). This technique can measure the accumulation of gadolinium-based contrast agent in the extravascular-extracellular space, which can be seen in different conditions affecting the brain, including tumors, vascular malformations, infections, and different inflammatory processes.
In addition to human studies, permeability in affected blood vessels has been described in one of the first animal models of GNAQ p.R183Q capillary malformation (45). In this mouse model, pretreatment with a subseizure dose of kainate for 1.5 h, followed by perfusion of Evans blue dye, showed severe leakage and irregular expression of claudin-5 in the leptomeningeal vessels in the mutant mice. This is consistent with the lack of junctional markers (zona occludin1 and claudin-5) we previously observed in CM brain and skin blood vessels (46). Here, we show additional evidence for vascular permeability in a second mouse model wherein endothelial expression of the Gnaq p.R183Q is induced in adult mice with tamoxifen (25); however, Evans blue leakage was not seen in the brain. Whether pretreatment with kainate would reveal a phenotype in the brain could be tested in future studies. Vascular permeability has also been shown in other vascular malformations such as cerebral cavernous malformation (CCMs) (47, 48) and arteriovenous malformation (49–52). Particularly in CCMs, knockdown of CCM3 in brain ECs was shown to reduce TEER in vitro and tight junction stability in vivo (53, 54).
Our experiments show that ANGPT2 and MAP2K1 contribute to the increased permeability in EC-R183Q monolayers. Angiopoietins are crucial angiogenic factors that play fundamental roles in vascular development and angiogenic processes. Angiopoietin-1 binds to the receptor tyrosine kinase TIE2 to promote vessel formation and stabilization, whereas ANGPT2 in most settings antagonizes TIE2 signaling, resulting in vessel destabilization, remodeling, and angiogenesis. Upregulation of ANGPT2 occurs in many types of advanced tumors (55–60) and vascular anomalies (18, 61–64). ANGPT2 has also been proposed as a serum biomarker for lymphatic anomalies (65) and has been associated with blood–brain barrier leakiness in early Alzheimer’s disease (66). CCMs exhibit brain hyperpermeability, cerebral hemorrhage, seizures, and stroke (48, 67). CCM3 loss of function results in elevated ANGPT2 in ECs. A neutralizing antibody against ANGPT2 was shown to normalize defects caused by CCM3 deficiency (63). In another study associated with cerebrovascular leakage and brain edema, ANGPT2 was shown to mediate blood–brain barrier permeability via endothelial paracellular and transcellular routes (38). Its role in EC-R183Q barrier formation suggests targeting ANGPT2 in CM would promote vascular integrity and prevent leakage.
Trametinib selectively blocks phosphorylation and activation of ERK1/2, thereby reducing downstream signaling needed for cell proliferation, survival, and angiogenesis. It is approved for certain cancers and is currently in phase 2 clinical trial for arteriovenous malformation. In KRAS-mutant EC models, trametinib reduced arteriovenous malformation size and normalized vessel morphology (31). Sirolimus, a mammalian target of rapamycin (mTOR) inhibitor, has been tested in SWS patients for improvements in cognitive functions and topically for improvement in pulse dye laser treatment of skin CMs. Sirolimus was well tolerated by adult SWS patients in a phase 2/3 clinical trial (68). Sirolimus combined with pulsed dye laser for patients with port wine birthmark showed reduced skin pigmentation and lower number of vessels in histologic analysis (69). In our in vitro TEER model, we tested two mTOR inhibitors, everolimus and sirolimus (aka rapamycin), for effects on EC-R183Q barrier formation. Neither mTOR inhibitor influenced the endothelial barrier formation. Sirolimus may affect other properties of the GNAQ-R183Q endothelium, which could be explored in future studies.
There are several limitations in our study. For most of the in vitro experiments, we used telomerase-expressing ECs. The cells express endothelial markers and proliferate well in culture; however, the telomerase ECs may not fully represent the growth and functional properties of ECs from CMs in the skin and brain. A second limitation is that the knockdown of ANGPT2 in EC-R183Q was not 100% such that remaining ANGPT2 might have affected some results. Third, experiments to examine potential paracrine signaling from neighboring cells, for example, by incorporating mural cells into the experimental design, would be informative.
In conclusion, our study demonstrates that Gαq-R183Q expressing ECs form reduced endothelial barrier in the TEER assay. Signs of increased vascular permeability are evident in human studies of CM and SWS and in an EC-specific Gnaq p.R183Q mouse model, which strengthens the relevance of our findings using GNAQ p.R183Q ECs in cell culture. Histopathology shows that there is extravascular fibrin in both brain and skin CMs and longitudinal MRI of the brain of a child with SWS indicates capillary permeability in the leptomeninges develops during infancy. Additional studies are needed as this is a n = 1 case. A murine developmental model of brain CM also shows increased permeability in brain vessels of mice with inducible expression of GNAQ p.R183Q (45) which further demonstrates endothelial permeability as a pathologic feature of CM. Our finding of increased permeability in human ECs expressing the mutant Gαq provides a platform for drug testing that we hope will address clinical needs of these patients.
Methods
Additional and detailed information is available in SI Appendix.
Droplet digital PCR ddPCR was used to measure the mutant allelic frequency in tissue sections and in cultured cells (4, 46). Briefly, genomic DNA was isolated from cell pellets using DNeasy Blood & Tissue Kit (Qiagen, 69506). ddPCR with probes to detect the GNAQ R183Q mutation is listed in Major Resource Table. 15 to 30 ng DNA per sample was mixed with ddPCR SuperMix (BioRad, 186-3010), primers for a 74 bp amplicon in GNAQ Exon4, FAM-labeled quenched mutant probes, and HEX-labeled quenched wild-type probes (See Major Resource Table), divided into droplets (QX200 AutoDG) and amplified for 40 PCR cycles using 60 °C as annealing temperature. The droplet fluorescence was read in a QX200 droplet reader and analyzed in QuantaSoft software (BioRad).
Cell culture The CRISPR edited cells designated EC-R183Q and EC-WT were expanded on 0.1 µg/cm2 fibronectin (Millipore Sigma-Aldrich, FC010-10MG) coated plates at 15,000 to 20,000 cells/cm2 in endothelial cell growth medium-2 (EGM2, Lonza, CC3162), which contains endothelial cell basal medium-2 (EBM2), SingleQuot supplements (all except hydrocortisone and GA-1000), 10% heat-inactivated fetal bovine serum (FBS, Cytiva), and 1× glutamine-penicillin-streptomycin (GPS, Thermo Fisher). This growth media will be referred to as EGM2. For experiments, the cells were grown in EBM2 without growth factors and with either 2% FBS or 0% FBS for 16 to 24 h to promote quiescence and throughout the remainder of the experiment as indicated.
Trans Endothelial Electrical Resistance (TEER) Assay First, 30,000 cells/well were seeded on a fibronectin-coated (0.1 mg/cm2) 96-well array with 10 interdigitated electrodes per well (Applied BioPhysics, 96W10idfPET ECIS array) in EGM2. Once cells adhered and formed a monolayer overnight (16 h), medium was replaced with reduced serum (2% FBS) EBM2 supplemented with thrombin (0.5 U/mL) (Sigma-Aldrich, 605190) to disrupt the nascent EC–EC contacts and follow subsequent barrier formation measured by electrical resistance (Ω). TEER was recorded after thrombin addition over 24 h, with and without inhibitors. TEER was measured at multiple frequencies (62.5 to 64,000 Hz) in real-time using an ECIS™ Zθ system (Applied BioPhysics). All ECIS measurements are reported in ohms (Ω) at 4,000 Hz, which is the optimal frequency for ECs. The percentage change in resistance at 5 h and 20 h was quantified and normalized to appropriate control cells. Four technical replicates for each condition were measured. Each experiment was repeated at least 3 independent times. A representative of the three is shown for each TEER experiment.
Supplementary Material
Appendix 01 (PDF)
Acknowledgments
We would like to thank the Vascular Biology Program Microscopy Imaging Core, and the Histology Core at Boston Children’s Hospital. We would like to thank Dr. William T. Pu, Department of Cardiology, Boston Children’s Hospital for providing the inducible Cdh5-Cre Tg(Cdh5-cre/ERT2)1Rha+/− mouse line. We would like to thank Drs. Raj Gupta and Gavin Schnitz for donating the teloHAECs used for CRISPR editing and subsequent experiments. We would also like to thank the Biopolymers Facility at Harvard Medical School for assistance with generating and characterizing the mutant teloHAECs cell lines. Sources Of Funding: Research reported in this manuscript was supported by the National Heart, Lung, and Blood Institute, part of the NIH, under Award Number 5R01HL127030 (J.B. and A.K.G). S.N. was supported by F32HL172637 and is currently supported by K99 Career Development Award—K99HL177326. M.V. was supported by F32HD107878. We also extend gratitude and thanks to the Taylor/McDonald Family. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.
Author contributions
S.N. and J.B. designed research; S.N., M.B., J.W.-S., M.P.V., A.H., L.M., and Y.S.C. performed research; S.N., M.P.V., P.S., C.A., and A.P. contributed new reagents/analytic tools; S.N., M.B., and J.B. analyzed data; and S.N. and J.B. wrote the paper.
Competing interests
The authors declare no competing interest.
Footnotes
This article is a PNAS Direct Submission.
Preprint server: Deposited to BioRxiv.https://www.biorxiv.org/content/10.1101/2025.03.31.646063v1.
Contributor Information
Sana Nasim, Email: sana.nasim@childrens.harvard.edu.
Joyce Bischoff, Email: joyce.bischoff@childrens.harvard.edu.
Data, Materials, and Software Availability
All study data are included in the article and/or SI Appendix.
Supporting Information
References
- 1.Galeffi F., et al. , A novel somatic mutation in GNAQ in a capillary malformation provides insight into molecular pathogenesis. Angiogenesis 25, 493–502 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Liu L., et al. , Histological analysis of different types of port-wine stains to guide clinical decision making: A retrospective study. Indian J. Dermatol. Venereol. Leprol. 89, 204–212 (2023). [DOI] [PubMed] [Google Scholar]
- 3.Tan W., et al. , Pathological alterations involve the entire skin physiological milieu in infantile and early-childhood port-wine stain. Br. J. Dermatol. 177, 293–296 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Nasim S., et al. , MRC1 and LYVE1 expressing macrophages in vascular beds of GNAQ p. R183Q driven capillary malformations in Sturge Weber syndrome. Acta Neuropathol. Commun. 12, 47 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Hammill A. M., Boscolo E., Capillary malformations. J. Clin. Invest. 134, e172842 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Yeom S., Comi A. M., Updates on Sturge-Weber syndrome. Stroke 53, 3769–3779 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Shirley M. D., et al. , Sturge-weber syndrome and port-wine stains caused by somatic mutation in GNAQ. N. Engl. J. Med. 368, 1971–1979 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Frigerio A., et al. , Genetic variants associated with port-wine stains. PLoS One 10, e0133158 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Lee K. T., et al. , Phenotypic association of presence of a somatic GNAQ mutation with port-wine stain distribution in capillary malformation. Head Neck 41, 4143–4150 (2019). [DOI] [PubMed] [Google Scholar]
- 10.Nakashima M., et al. , The somatic GNAQ mutation c.548G>A (p.R183Q) is consistently found in Sturge-Weber syndrome. J. Hum. Genet. 59, 691–693 (2014). [DOI] [PubMed] [Google Scholar]
- 11.Couto J. A., et al. , A somatic GNA11 mutation is associated with extremity capillary malformation and overgrowth. Angiogenesis 20, 303–306 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Dompmartin A., et al. , GNA11-mutated Sturge-Weber syndrome has distinct neurological and dermatological features. Eur. J. Neurol. 29, 3061–3070 (2022). [DOI] [PubMed] [Google Scholar]
- 13.Couto J. A., et al. , Endothelial cells from capillary malformations are enriched for somatic GNAQ mutations. Plast. Reconstr. Surg. 137, 77e–82e (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Huang L., et al. , Somatic GNAQ mutation is enriched in brain endothelial cells in Sturge-Weber syndrome. Pediatr. Neurol. 67, 59–63 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Falzone M. E., MacKinnon R., Gbetagamma activates PIP2 hydrolysis by recruiting and orienting PLCbeta on the membrane surface. Proc. Natl. Acad. Sci. U.S.A. 120, e2301121120 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Falzone M. E., MacKinnon R., The mechanism of Galpha(q) regulation of PLCbeta3-catalyzed PIP2 hydrolysis. Proc. Natl. Acad. Sci. U.S.A. 120, e2315011120 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Harraz O. F., Hill-Eubanks D., Nelson M. T., PIP(2): A critical regulator of vascular ion channels hiding in plain sight. Proc. Natl. Acad. Sci. U.S.A. 117, 20378–20389 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Huang L., et al. , Endothelial GNAQ p.R183Q increases ANGPT2 (Angiopoietin-2) and drives formation of enlarged blood vessels. Arterioscler. Thromb. Vasc. Biol. 42, e27–e43 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Van Raamsdonk C. D., et al. , Frequent somatic mutations of GNAQ in uveal melanoma and blue naevi. Nature 457, 599–602 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Martins L., et al. , Computational analysis for GNAQ mutations: New insights on the molecular etiology of Sturge-Weber syndrome. J. Mol. Graph. Model. 76, 429–440 (2017). [DOI] [PubMed] [Google Scholar]
- 21.Feng X., et al. , Hippo-independent activation of YAP by the GNAQ uveal melanoma oncogene through a trio-regulated rho GTPase signaling circuitry. Cancer Cell 25, 831–845 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Schrenk S., et al. , MEK inhibition reduced vascular tumor growth and coagulopathy in a mouse model with hyperactive GNAQ. Nat. Commun. 14, 1929 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Norman M. G., Schoene W. C., The ultrastructure of sturge-weber disease. Acta Neuropathol. 37, 199–205 (1977). [DOI] [PubMed] [Google Scholar]
- 24.Freeze W. M., et al. , CSF enhancement on post-contrast fluid-attenuated inversion recovery images; A systematic review. NeuroImage: Clinical 28, 102456 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Smits P., et al. , An endothelial specific mouse model for the capillary malformation mutation Gnaq p. R183Q. Angiogenesis 28, 33 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Onken M. D., et al. , Targeting primary and metastatic uveal melanoma with a G protein inhibitor. J. Biol. Chem. 296, 100403 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Kamato D., et al. , Ga(q) proteins: Molecular pharmacology and therapeutic potential. Cell. Mol. Life Sci. 74, 1379–1390 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Zhang H., Nielsen A. L., Stromgaard K., Recent achievements in developing selective G(q) inhibitors. Med. Res. Rev. 40, 135–157 (2020). [DOI] [PubMed] [Google Scholar]
- 29.Cong X., Kong W., Endothelial tight junctions and their regulatory signaling pathways in vascular homeostasis and disease. Cell. Signal. 66, 109485 (2020). [DOI] [PubMed] [Google Scholar]
- 30.Wu X., Li J., Zhu M., Fletcher J. A., Hodi F. S., Protein kinase C inhibitor AEB071 targets ocular melanoma harboring GNAQ mutations via effects on the PKC/Erk1/2 and PKC/NF-kappaB pathways. Mol. Cancer Ther. 11, 1905–1914 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Nguyen H. L., Boon L. M., Vikkula M., Trametinib as a promising therapeutic option in alleviating vascular defects in an endothelial KRAS-induced mouse model. Hum. Mol. Genet. 32, 276–289 (2023). [DOI] [PubMed] [Google Scholar]
- 32.Lekwuttikarn R., Lim Y. H., Admani S., Choate K. A., Teng J. M. C., Genotype-guided medical treatment of an arteriovenous malformation in a child. JAMA Dermatol. 155, 256–257 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Pastura P., et al. , NRAS(Q61R) mutation drives elevated angiopoietin-2 expression in human endothelial cells and a genetic mouse model. Pediatr. Blood Cancer 71, e31032 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Lugowska I., Kosela-Paterczyk H., Kozak K., Rutkowski P., Trametinib: A MEK inhibitor for management of metastatic melanoma. Onco Targets Ther. 8, 2251–2259 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Gouda M. A., Subbiah V., Expanding the benefit: Dabrafenib/Trametinib as tissue-agnostic therapy for BRAF V600E-positive adult and pediatric solid tumors. Am. Soc. Clin. Oncol. Educ. Book 43, e404770 (2023). [DOI] [PubMed] [Google Scholar]
- 36.You Y., et al. , U0126: Not only a MAPK kinase inhibitor. Front. Pharmacol. 13, 927083 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Iverson C., et al. , RDEA119/BAY 869766: A potent, selective, allosteric inhibitor of MEK1/2 for the treatment of cancer. Cancer Res. 69, 6839–6847 (2009). [DOI] [PubMed] [Google Scholar]
- 38.Gurnik S., et al. , Angiopoietin-2-induced blood-brain barrier compromise and increased stroke size are rescued by VE-PTP-dependent restoration of Tie2 signaling. Acta Neuropathol. 131, 753–773 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Daly C., et al. , Angiopoietin-2 functions as a Tie2 agonist in tumor models, where it limits the effects of VEGF inhibition. Cancer Res. 73, 108–118 (2013). [DOI] [PubMed] [Google Scholar]
- 40.Papadopoulos K. P., et al. , A phase I first-in-human study of nesvacumab (REGN910), a fully human anti-angiopoietin-2 (Ang2) monoclonal antibody, in patients with advanced solid tumors. Clin. Cancer Res. 22, 1348–1355 (2016). [DOI] [PubMed] [Google Scholar]
- 41.Zecchin D., et al. , GNAQ/GNA11 mosaicism causes aberrant calcium signaling susceptible to targeted therapeutics. J. Invest. Dermatol. 144, 811–819 e814 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Chu J. P., et al. , Pilot study on evaluation of any correlation between MR perfusion (Ktrans) and diffusion (apparent diffusion coefficient) parameters in brain tumors at 3 Tesla. Cancer Imaging 12, 1–6 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Chaganti J., et al. , Novel magnetic resonance KTRANS measurement of blood-brain barrier permeability correlated with covert HE. Hepatol. Commun. 7, e0079 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Ying Y., et al. , Heterogeneous blood-brain barrier dysfunction in cerebral small vessel diseases. Alzheimers Dement. 20, 4527–4539 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Chase Solomon M. M., Singh Pratibha, Nemeth Christina, Comi Anne, R183Q GNAQ Sturge-Weber syndrome leptomeningeal and cerebrovascular developmental mouse model. J. Vascular Anom. 5, e099 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Nasim S., et al. , Similarities and differences between brain and skin GNAQ p. R183Q driven capillary malformations. Angiogenesis 27, 931–941 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Phillips C. M., Johnson A. M., Stamatovic S. M., Keep R. F., Andjelkovic A. V., 20 kDa isoform of connexin-43 augments spatial reorganization of the brain endothelial junctional complex and lesion leakage in cerebral cavernous malformation type-3. Neurobiol. Dis. 186, 106277 (2023). [DOI] [PubMed] [Google Scholar]
- 48.Girard R., et al. , Vascular permeability and iron deposition biomarkers in longitudinal follow-up of cerebral cavernous malformations. J. Neurosurg. 127, 102–110 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Queisser A., Seront E., Boon L. M., Vikkula M., Genetic basis and therapies for vascular anomalies. Circ. Res. 129, 155–173 (2021). [DOI] [PubMed] [Google Scholar]
- 50.Nakisli S., Lagares A., Nielsen C. M., Cuervo H., Pericytes and vascular smooth muscle cells in central nervous system arteriovenous malformations. Front. Physiol. 14, 1210563 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.Nakisli S., Fanelli K., LaComb J., Arnold L. J., Nielsen C. M., CNS resident macrophages exhibit region-specific states and immunogenic responses during Rbpj-deficient brain arteriovenous malformation. Sci. Rep. 15, 3932 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Pedrosa L., et al. , Profiling tight junction protein expression in brain vascular malformations. Int. J. Mol. Sci. 26, 4558 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53.Johnson A. M., et al. , Connexin 43 gap junctions contribute to brain endothelial barrier hyperpermeability in familial cerebral cavernous malformations type III by modulating tight junction structure. FASEB J. 32, 2615–2629 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar] [Retracted]
- 54.Stamatovic S. M., Sladojevic N., Keep R. F., Andjelkovic A. V., PDCD10 (CCM3) regulates brain endothelial barrier integrity in cerebral cavernous malformation type 3: Role of CCM3-ERK1/2-cortactin cross-talk. Acta Neuropathol. 130, 731–750 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55.Etoh T., et al. , Angiopoietin-2 is related to tumor angiogenesis in gastric carcinoma: Possible in vivo regulation via induction of proteases. Cancer Res. 61, 2145–2153 (2001). [PubMed] [Google Scholar]
- 56.Ochiumi T., et al. , Clinical significance of angiopoietin-2 expression at the deepest invasive tumor site of advanced colorectal carcinoma. Int. J. Oncol. 24, 539–547 (2004). [PubMed] [Google Scholar]
- 57.Park H. R., et al. , Angiopoietin-2-dependent spatial vascular destabilization promotes T-cell exclusion and limits immunotherapy in melanoma. Cancer Res. 83, 1968–1983 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 58.Abdul Pari A. A., et al. , Tumor cell-derived angiopoietin-2 promotes metastasis in melanoma. Cancer Res. 80, 2586–2598 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 59.Wu X., et al. , Angiopoietin-2 as a biomarker and target for immune checkpoint therapy. Cancer Immunol. Res. 5, 17–28 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 60.Huang H., et al. , Specifically targeting angiopoietin-2 inhibits angiogenesis, Tie2-expressing monocyte infiltration, and tumor growth. Clin. Cancer Res. 17, 1001–1011 (2011). [DOI] [PubMed] [Google Scholar]
- 61.Engel E. R., Cras T. D., Ricci K. W., How we use angiopoietin-2 in the diagnosis and management of vascular anomalies. Pediatr. Blood Cancer 71, e30921 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 62.Vates G. E., Hashimoto T., Young W. L., Lawton M. T., Angiogenesis in the brain during development: The effects of vascular endothelial growth factor and angiopoietin-2 in an animal model. J. Neurosurg. 103, 136–145 (2005). [DOI] [PubMed] [Google Scholar]
- 63.Jenny Zhou H., et al. , Endothelial exocytosis of angiopoietin-2 resulting from CCM3 deficiency contributes to cerebral cavernous malformation. Nat. Med. 22, 1033–1042 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 64.Crist A. M., et al. , Angiopoietin-2 inhibition rescues arteriovenous malformation in a Smad4 hereditary hemorrhagic telangiectasia mouse model. Circulation 139, 2049–2063 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 65.Le Cras T. D., et al. , Angiopoietins as serum biomarkers for lymphatic anomalies. Angiogenesis 20, 163–173 (2017). [DOI] [PubMed] [Google Scholar]
- 66.Van Hulle C., et al. , Elevated CSF angiopoietin-2 correlates with blood-brain barrier leakiness and markers of neuronal injury in early Alzheimer’s disease. Transl. Psychiatry 14, 3 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 67.Mikati A. G., et al. , Vascular permeability in cerebral cavernous malformations. J. Cereb. Blood Flow Metab. 35, 1632–1639 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 68.Sebold A. J., et al. , Sirolimus Treatment in Sturge-Weber Syndrome. Pediatr. Neurol. 115, 29–40 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 69.Marques L., et al. , Topical rapamycin combined with pulsed dye laser in the treatment of capillary vascular malformations in Sturge-Weber syndrome: Phase II, randomized, double-blind, intraindividual placebo-controlled clinical trial. J. Am. Acad. Dermatol. 72, 151–158 e151 (2015). [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Appendix 01 (PDF)
Data Availability Statement
All study data are included in the article and/or SI Appendix.







