RASA1, a negative regulator of Ras-MAPK signaling, is essential for the development and maintenance of lymphatic vessel valves. However, whether RASA1 is required for the development and maintenance of lymphovenous valves (LVV) and venous valves (VV) is unknown. In this study, we show that induced disruption of Rasa1 in mouse embryos did not affect initial specification of LVV or central VV, but did affect their continued development. Similarly, a switch to expression of a catalytically inactive form of RASA1 resulted in impaired LVV and VV development. Blocked development of LVV was associated with accumulation of the basement membrane protein, collagen IV, in LVV-forming endothelial cells (EC), and could be partially or completely rescued by MAPK inhibitors and drugs that promote collagen IV folding. Disruption of Rasa1 in adult mice resulted in venous hypertension and impaired VV function that was associated with loss of EC from VV leaflets. In conclusion, RASA1 functions as a negative regulator of Ras signaling in EC that is necessary for EC export of collagen IV, thus permitting the development of LVV and the development and maintenance of VV.

The Ras signaling pathway is a ubiquitous intracellular signaling pathway that is triggered by different growth factor receptors (GFR) in numerous cell types (Buday and Downward, 2008; Wennerberg et al., 2005). Ras is a small GTP-binding protein, tethered to the inner leaflet of the cell membrane, which switches between inactive GDP-bound and active GTP-bound states. GFR activate Ras through recruitment to membranes of one or more Ras guanine nucleotide exchange factors (RasGEFs) that eject GDP from the Ras guanine nucleotide-binding pocket, thus permitting Ras to bind GTP (Bos et al., 2007). Activated GTP-bound Ras induces different downstream signaling cascades including the mitogen-activated protein kinase (MAPK) and phosphatidylinositol 3-kinase (PI3K) signaling cascades that couple GFR and Ras to cellular responses (Butler et al., 2011; Rubinfeld and Seger, 2005). Inactivation of Ras is mediated by Ras GTPase-activating proteins (RasGAPs) that increase the ability of Ras to hydrolyze bound GTP to GDP by several orders of magnitude (King et al., 2013).

There are ten different members of the RasGAP family that each comprise a catalytically active GAP domain together with one or more modular binding domains (King et al., 2013). Most members of the RasGAP family are expressed broadly in different tissues and cell types. Nonetheless, one RasGAP family member, p120 RasGAP (also known as RASA1), has emerged as an essential regulator of blood vessel (BV) and lymphatic vessel (LV) systems in both humans and mice. In humans, germline inactivating mutations of the RASA1 gene cause the autosomal dominant blood vascular disorder, capillary malformation-arteriovenous malformation (CM-AVM) (Eerola et al., 2003; Revencu et al., 2013, 2008). Recent evidence indicates that the development of BV lesions in CM-AVM is dependent upon the acquisition of somatic inactivating second hit mutations in the intact RASA1 allele in endothelial cells (EC) or their progenitors during development (Lapinski et al., 2018; Macmurdo et al., 2016). This is consistent with the variable phenotype and location of BV lesions in CM-AVM patients. In addition, LV abnormalities, including lymphedema, chylous ascites, chylothorax and LV hyperplasia, have also been observed in some CM-AVM patients (Burrows et al., 2013; de Wijn et al., 2012; Macmurdo et al., 2016; Revencu et al., 2013, 2008; Sevick-Muraca and King, 2014).

Non-conditional disruption of the Rasa1 gene in mice results in mid-gestation lethality [embryonic day (E) 10.5] as a consequence of abnormal vasculogenesis and developmental angiogenesis, in which primitive vascular plexuses are remodeled into hierarchical arterial-capillary-venous networks (Henkemeyer et al., 1995; Lubeck et al., 2014). In addition, disruption of Rasa1 in neonatal and adult mice results in impaired retinal angiogenesis and pathological angiogenesis, respectively (Chen et al., 2019; Lapinski et al., 2007, 2012). We recently demonstrated that, in the absence of RASA1, EC are unable to export the extracellular matrix (ECM) protein, collagen IV, which is the major constituent of vascular basement membranes (BM). This inability to export collagen IV accounts for the impaired developmental, neonatal and pathological angiogenic responses in the absence of RASA1.

Concerning the LV system, induced disruption of Rasa1 in adult mice results in LV leakage in the form of chylous ascites and chylothorax (Lapinski et al., 2012). This can be explained on the basis that RASA1 is required for the maintenance of intra-luminal valves in collecting LV that normally prevent backflow of lymph fluid (Lapinski et al., 2017). Upon Rasa1 disruption, lymphatic EC (LEC) are lost from LV valve leaflets at the rate of approximately one LEC per leaflet per week, until a threshold point is reached at which the valve is unable to prevent fluid backflow and is unable to function (Lapinski et al., 2017). RASA1 is also necessary for the development of LV valves in late gestation. Disruption of Rasa1 just before LV valvulogenesis at E15.5 does not impact upon initial LV valve specification characterized by increased expression of the PROX1 transcription factor in valve-forming LEC (Lapinski et al., 2017; Sabine et al., 2012). However, all PROX1hi valve-forming LEC undergo apoptosis shortly thereafter and valve development ceases (Lapinski et al., 2017).

Lymph fluid is returned to the blood circulation via LV that connect with the blood vasculature at the junction of the internal jugular vein (IJV), external jugular vein (EJV) and subclavian veins (SCV) with the superior vena cava (SVC) (Geng et al., 2016; Srinivasan and Oliver, 2011). Lymphovenous valves (LVV) guard the four points of connection of these LV with the blood vasculature (two points of connection on either side of the animal) and, thereby, prevent backflow of blood into LV (Geng et al., 2016; Srinivasan and Oliver, 2011). LVV first form at E12 at the point of contact of lymphatic sacs (LS) with the blood vasculature. LVV specification is marked by the upregulation of PROX1 by both LEC on the lymphatic side and LVV-forming EC on the venous side of the nascent valve structure (Geng et al., 2016; Srinivasan and Oliver, 2011). LVV development then proceeds through the successive stages of delamination, aggregation and maturation as LVV leaflets form and extend into the vein lumen; a process that is considered complete by E16.5 (Geng et al., 2016). Mature LVV leaflets comprise two different types of EC attached to a central ECM core (Geng et al., 2016). EC on the lymphatic and venous sides of leaflets continue to express high levels of PROX1. LEC on the lymphatic side of leaflets express high levels of vascular endothelial growth factor receptor 3 (VEGFR3; also known as Flt4), the receptor for VEGFC. In contrast, EC on the venous side of LVV leaflets express low levels of VEGFR3, and the rest of the venous EC are VEGFR3 negative. As RASA1 is necessary for the development of LV valves, it is important to determine whether it performs a similar function in the development of LVV and, if so, at which stage.

VV prevent backflow of blood and are required for unidirectional flow of blood in veins (Geng et al., 2017). Dysfunction of VV has serious consequences that include chronic venous insufficiency, venous hypertension and deep vein thrombosis (Meissner et al., 2007a,b). Development of VV in central veins, such as the IJV, EJV and SCV, is initiated at E14.5 and is completed by E16.5 (Geng et al., 2016; Munger et al., 2016). In contrast, VV in peripheral veins, such as the saphenous vein, develop postnatally (Bazigou et al., 2011; Munger et al., 2016). VV specification, like LV valve and LVV specification, is marked by increased expression of PROX1 in valve-forming EC (Geng et al., 2016, 2017). Mature VV are structurally similar to LV valves in that leaflets of both type of valve comprise a single type of EC attached to a central ECM core. VV and LV valve leaflet EC both continue to express high levels of PROX1 and share several other molecular characteristics, including high level expression of the FOXC2 transcription factor and α9 integrin (Geng et al., 2016, 2017). Given the similarity between VV and LV valves, investigation as to whether or not RASA1 is necessary for the development and maintenance of VV is also important.

Here, we used different mouse models of RASA1 deficiency to demonstrate a role for RASA1 in the development and maintenance of LVV and VV. Furthermore, we established that RASA1 is necessary for the export of collagen IV from valve-forming EC, which explains its role in valve development. Studies are of potential relevance to an understanding of the pathogenesis of CM-AVM and are of additional significance to other cardiovascular diseases.

RASA1 is not required for the specification of LVV and central VV

Specification of LVV occurs at E12 and is marked by increased expression of PROX1 in LEC and LVV-forming EC at the points that LS make contact with BV in the region of the junction of the IJV, EJV, SCV and SVC (Geng et al., 2016; Srinivasan and Oliver, 2011). Therefore, to examine whether RASA1 plays a role in LVV specification, we administered tamoxifen (TM) to pregnant Rasa1fl/fl mice carrying Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos at E10.5 and analyzed embryos at E12.5. The Cdh5 promoter drives expression of the ert2cre transgene within EC specifically (Wang et al., 2010). As assessed by staining of frontal tissue sections for PROX1, VEGFR3 and CD31, induced disruption of Rasa1 at E10.5 in Rasa1fl/fl Cdh5ert2cre embryos did not affect LVV specification (Fig. 1A).

Fig. 1.

RASA1 is not required for the specification of LVV and central VV. (A,B) Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos were administered TM at E10.5 or E12.5 and harvested at E12.5 or E14.5, respectively. Longitudinal sections through the neck region were stained with the indicated antibodies. LVV are indicated by arrows and VV by arrowheads. Note, normal specification of LVV at E12.5 and VV at E14.5 (n=2 Rasa1fl/fl and n=5 Rasa1fl/fl Cdh5ert2cre embryos at E12.5 and n=4 Rasa1fl/fl and n=5 Rasa1fl/fl Cdh5ert2cre embryos at E14.5). IJV, internal jugular vein; LS, lymphatic sac; SVC, superior vena cava.

Fig. 1.

RASA1 is not required for the specification of LVV and central VV. (A,B) Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos were administered TM at E10.5 or E12.5 and harvested at E12.5 or E14.5, respectively. Longitudinal sections through the neck region were stained with the indicated antibodies. LVV are indicated by arrows and VV by arrowheads. Note, normal specification of LVV at E12.5 and VV at E14.5 (n=2 Rasa1fl/fl and n=5 Rasa1fl/fl Cdh5ert2cre embryos at E12.5 and n=4 Rasa1fl/fl and n=5 Rasa1fl/fl Cdh5ert2cre embryos at E14.5). IJV, internal jugular vein; LS, lymphatic sac; SVC, superior vena cava.

We also examined whether RASA1 was required for the specification of central VV. Specification of the VV that guard the points of entry of the IJV, EJV and SCV into the SVC occurs at E14.5 (Geng et al., 2016; Munger et al., 2016). Similar to LVV and LV valve specification, VV specification is characterized by increased expression of PROX1 and VEGFR3 in VV-forming EC. Therefore, to examine whether RASA1 is required for VV specification, we administered TM to pregnant Rasa1fl/fl mice carrying Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos at E12.5 and analyzed embryos at E14.5 for PROX1 and VEGFR3 expression at the respective venous junctions (Fig. 1B). Loss of RASA1 in EC at E12.5 did not affect VV specification. Therefore, RASA1 is dispensable for the specification of VV as well as LVV.

RASA1 is required for the continued development of LVV and central VV

Development of LVV and central VV continues through E16.5 as valve leaflets are extended into the vessel lumen (Geng et al., 2016; Srinivasan and Oliver, 2011). To determine whether RASA1 is required for continued LVV and central VV development, we administered TM to Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos at E12.5 and analyzed embryos at E16.5. LVV and VV structure was analyzed by immunostaining of tissue sections and by scanning electron microscopy (SEM) (Fig. 2A,B). In Rasa1fl/fl control embryos, one pair of normal LVV was identified at the vast majority of examined LV-BV junctions (Fig. 2C). In contrast, in Rasa1fl/fl Cdh5ert2cre embryos, most LV-BV junctions contained no LVV and a minority contained only one LVV. None of the examined LV-BV junctions contained two LVV (Fig. 2C). Similarly, in Rasa1fl/fl control embryos, normal VV were found at all examined junctions of the IJV, EJV and SCV with the SVC (Fig. 2D,E). In contrast, in Rasa1fl/fl Cdh5ert2cre embryos, development of central VV was severely impaired. In the majority of cases, all three VV were absent, and in the remaining cases only a rudimentary EJV VV could be identified, whereas the IJV and SCV VV were also absent (Fig. 2D,E). Thus, although RASA1 is not required for initial specification of LVV and central VV, it is required for their continued development.

Fig. 2.

RASA1 is necessary for continued LVV and central VV development after initial specification. (A,B) Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos were administered TM at E12.5 and harvested at E16.5. (A) Longitudinal sections through the neck region (dorsal to ventral from top to bottom) were stained with the indicated antibodies. LVV and VV are indicated by arrows and arrowheads, respectively. (B) The structure of LVV and central VV was analyzed by SEM. LVV and VV are pseudo-colored in magenta and green, respectively. In A and B, note the absence of LVV and VV in Rasa1fl/fl Cdh5ert2cre embryos. (C) Percentage of LV-BV junctions from Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos in A and B that contain the indicated number of LVV. Rasa1fl/fl n=11, Rasa1fl/fl Cdh5ert2cre n=13. (D) VV structure categories. (E) Percentage of BV junctions for Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos in A and B that fall into the indicated VV categories in D. Rasa1fl/fl n=11, Rasa1fl/fl Cdh5ert2cre n=13. A, artery; EJV, external jugular vein; IJV, internal jugular vein; LS, lymphatic sac; SCV, subclavian vein; SVC, superior vena cava.

Fig. 2.

RASA1 is necessary for continued LVV and central VV development after initial specification. (A,B) Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos were administered TM at E12.5 and harvested at E16.5. (A) Longitudinal sections through the neck region (dorsal to ventral from top to bottom) were stained with the indicated antibodies. LVV and VV are indicated by arrows and arrowheads, respectively. (B) The structure of LVV and central VV was analyzed by SEM. LVV and VV are pseudo-colored in magenta and green, respectively. In A and B, note the absence of LVV and VV in Rasa1fl/fl Cdh5ert2cre embryos. (C) Percentage of LV-BV junctions from Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos in A and B that contain the indicated number of LVV. Rasa1fl/fl n=11, Rasa1fl/fl Cdh5ert2cre n=13. (D) VV structure categories. (E) Percentage of BV junctions for Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos in A and B that fall into the indicated VV categories in D. Rasa1fl/fl n=11, Rasa1fl/fl Cdh5ert2cre n=13. A, artery; EJV, external jugular vein; IJV, internal jugular vein; LS, lymphatic sac; SCV, subclavian vein; SVC, superior vena cava.

Impaired development of LVV and central VV in embryos that express RASA1 R780Q alone

Some studies have indicated that RASA1 can participate in intracellular signaling pathways independent of its ability to promote Ras hydrolysis of GTP (King et al., 2013). Therefore, we examined whether the function of RASA1 in LVV and central VV development was dependent upon its catalytic GAP activity. For this purpose, we used a Rasa1R780Q allele that we reported previously (Lapinski et al., 2017). R780 of RASA1 is the ‘arginine finger’ of the GAP domain that is essential for GAP activity (King et al., 2013; Lubeck et al., 2014). Mutation of this arginine to glutamine in R780Q thus abrogates GAP activity, although is predicted to leave all other putative RASA1 functions intact. As Rasa1R780Q/R780Q embryos die at E10.5 of gestation (Lubeck et al., 2014), we generated Rasa1fl/R780Q embryos with ubiquitin promoter-driven ert2cre (Ubert2cre) to examine a requirement of RASA1 GAP activity for LVV and central VV development. TM was administered to Rasa1fl/R780Q Ubert2cre embryos and control Rasa1fl/fl and Rasa1fl/R780Q embryos at E12.5 and LVV and VV development was analyzed at E16.5 as before (Fig. 3A). Compared with controls, Rasa1fl/R780Q Ubert2cre embryos showed impaired development of LVV and VV. The majority of LV-BV junctions in Rasa1fl/R780Q Ubert2cre embryos contained only one LVV (Fig. 3B). Furthermore, the majority of examined vessels from Rasa1fl/R780Q Ubert2cre embryos contained no VV at the junctions of the IJV and SCV with the SVC (Fig. 3C). Nonetheless, the impact of a switch to expression of RASA1 R780Q alone upon LVV and VV development was less than that observed following complete loss of RASA1 in EC (Fig. 2). This finding is consistent with the notion that RASA1 regulates LVV and VV development through both GAP domain-dependent and GAP domain-independent mechanisms.

Fig. 3.

Development of LVV and central VV in RASA1 R780Q embryos. (A) Rasa1fl/R780Q females were crossed with Rasa1fl/R780Q Ubert2cre males. TM was administered to pregnant females at E12.5 and embryos were harvested at E16.5. Longitudinal sections of the neck region (dorsal to ventral from top to bottom) were stained with the indicated antibodies. Shown are representative images from Rasa1fl/R780Q and Rasa1fl/R780Q Ubert2cre embryos. Arrows and arrowheads indicate LVV and VV, respectively. (B,C) Percentage of LV-BV junctions from Rasa1fl/R780Q Ubert2cre embryos and control Rasa1fl/fl and Rasa1fl/R780Q embryos from A that contain the indicated number of LVV (B) and categories of VV (C, see Fig. 2D). Rasa1fl/R780Q Ubert2cre n=6, Rasa1fl/fl n=4, Rasa1fl/R780Q n=4. EJV, external jugular vein; IJV, internal jugular vein; LS, lymphatic sac; SCV, subclavian vein; SVC, superior vena cava.

Fig. 3.

Development of LVV and central VV in RASA1 R780Q embryos. (A) Rasa1fl/R780Q females were crossed with Rasa1fl/R780Q Ubert2cre males. TM was administered to pregnant females at E12.5 and embryos were harvested at E16.5. Longitudinal sections of the neck region (dorsal to ventral from top to bottom) were stained with the indicated antibodies. Shown are representative images from Rasa1fl/R780Q and Rasa1fl/R780Q Ubert2cre embryos. Arrows and arrowheads indicate LVV and VV, respectively. (B,C) Percentage of LV-BV junctions from Rasa1fl/R780Q Ubert2cre embryos and control Rasa1fl/fl and Rasa1fl/R780Q embryos from A that contain the indicated number of LVV (B) and categories of VV (C, see Fig. 2D). Rasa1fl/R780Q Ubert2cre n=6, Rasa1fl/fl n=4, Rasa1fl/R780Q n=4. EJV, external jugular vein; IJV, internal jugular vein; LS, lymphatic sac; SCV, subclavian vein; SVC, superior vena cava.

MAPK inhibition partially rescues LVV development in the absence of RASA1

PROX1hi EC in developing LVV in Rasa1fl/fl Cdh5ert2cre embryos administered TM at E12.5 showed increased activation of MAPK at E14.5 (Fig. 4A,B). Similarly, PROX1hi EC in VV of Rasa1fl/fl Cdh5ert2cre embryos administered TM at E12.5 showed increased activation of MAPK at E15.5 (Fig. S1). Therefore, we next asked whether inhibition of MAPK could rescue development of LVV and central VV in TM-treated Rasa1fl/fl Cdh5ert2cre embryos. Thus, we administered the MAPK inhibitor, AZD6244, along with TM to Rasa1fl/fl Cdh5ert2cre and Rasa1fl/fl control embryos at E12.5 and on each subsequent day thereafter until embryo harvest at E16.5. AZD6244 had a significant rescue effect upon LVV development in TM-treated Rasa1fl/fl Cdh5ert2cre embryos (Fig. 4C,D). Although in the absence of AZD6244 treatment, the majority of LV-BV junctions contained no LVV, in the presence of AZD6244 the majority of LV-BV junctions contained at least one LVV, and in some cases two LVV were identified. AZD6244 had no influence upon LVV development in TM-treated Rasa1fl/fl embryos. These findings indicate that impaired LVV development results in part from dysregulated activation of the MAPK pathway. Surprisingly, despite the finding of significant impairment of central VV development in embryos that express RASA1 R780Q alone, AZD6244 was largely unable to rescue central VV development in TM-treated Rasa1fl/fl Cdh5ert2cre embryos (Fig. 4C,E). Therefore, impaired development of central VV in the absence of RASA1 may be more dependent upon dysregulated activation of distinct effector pathways downstream of activated Ras.

Fig. 4.

MAPK activation in LVV and effect of MAPK inhibition upon LVV and VV development in induced EC-specific RASA1-deficient embryos. (A) Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos were administered TM at E12.5 and embryos were harvested at E14.5. Longitudinal sections through the neck region were stained with the indicated antibodies. Representative sections show an increased number of phospho-ERK MAPK (pERK)+ EC among PROX1hi EC in LVV (arrows) of Rasa1fl/fl Cdh5ert2cre embryos compared with controls. (B) Mean+s.e.m. of the percentage pERK+ EC of PROX1hi EC in LVV of embryos of the indicated genotypes (Rasa1fl/fl n=3 LVV, Rasa1fl/fl Cdh5ert2cre n=5 LVV). **P<0.01, Student's 2-sample t-test. (C) Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos were administered TM and AZD6244 at E12.5 and AZD6244 every day thereafter until embryo harvest at E16.5. Longitudinal sections through the neck region (dorsal to ventral from top to bottom) were stained with the indicated antibodies. (D,E) Percentage of LV-BV junctions from Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos from C that contain the indicated number of LVV (D) and categories of VV (E, see Fig. 2D). Rasa1fl/fl n=4, Rasa1fl/fl Cdh5ert2cre n=12. Note that data from Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos treated with TM alone from Fig. 2C,E (red and black bars) is also plotted for comparison. EJV, external jugular vein; IJV, internal jugular vein; LS, lymphatic sac; SCV, subclavian vein; SVC, superior vena cava.

Fig. 4.

MAPK activation in LVV and effect of MAPK inhibition upon LVV and VV development in induced EC-specific RASA1-deficient embryos. (A) Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos were administered TM at E12.5 and embryos were harvested at E14.5. Longitudinal sections through the neck region were stained with the indicated antibodies. Representative sections show an increased number of phospho-ERK MAPK (pERK)+ EC among PROX1hi EC in LVV (arrows) of Rasa1fl/fl Cdh5ert2cre embryos compared with controls. (B) Mean+s.e.m. of the percentage pERK+ EC of PROX1hi EC in LVV of embryos of the indicated genotypes (Rasa1fl/fl n=3 LVV, Rasa1fl/fl Cdh5ert2cre n=5 LVV). **P<0.01, Student's 2-sample t-test. (C) Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos were administered TM and AZD6244 at E12.5 and AZD6244 every day thereafter until embryo harvest at E16.5. Longitudinal sections through the neck region (dorsal to ventral from top to bottom) were stained with the indicated antibodies. (D,E) Percentage of LV-BV junctions from Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos from C that contain the indicated number of LVV (D) and categories of VV (E, see Fig. 2D). Rasa1fl/fl n=4, Rasa1fl/fl Cdh5ert2cre n=12. Note that data from Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos treated with TM alone from Fig. 2C,E (red and black bars) is also plotted for comparison. EJV, external jugular vein; IJV, internal jugular vein; LS, lymphatic sac; SCV, subclavian vein; SVC, superior vena cava.

Loss of RASA1 results in accumulation of collagen IV in EC of LVV

We recently determined that global or EC-specific disruption of Rasa1 at E13.5 results in the apoptotic death of all dermal EC at E18.5 or E19.5, respectively (Chen et al., 2019). In the absence of RASA1, the vascular BM protein, collagen IV, is retained within the endoplasmic reticulum (ER) of EC and smooth muscle cells. As a result of the reduced amounts of collagen IV in BM, EC fail to attach properly to BM and undergo apoptotic death. Based upon these findings, we asked whether induced loss of RASA1 at E12.5 results in intracellular accumulation of collagen IV in EC of LVV. An inability of LVV EC to export collagen IV for deposition in the ECM core of developing LVV leaflets would provide an explanation for failed LVV development. In embryos administered TM at E12.5, developing LVV are present at E14.5 (Fig. 1B) but absent at E16.5 (Fig. 2). Therefore, we examined E15.5 embryos for any evidence of EC collagen IV accumulation in LVV EC. At E15.5, LVV in TM-treated (at E12.5) Rasa1fl/fl Cdh5ert2cre embryos were frequently abnormal and separation of the two EC layers was often observed (Fig. S2). We examined collagen IV accumulation in embryos that additionally carried a Prox1-eGFP transgene (Fig. 5) (Choi et al., 2011). In LVV EC of Rasa1fl/fl embryos, eGFP should be present throughout the cell cytoplasm and nucleus resulting in the appearance of a broadly distributed cellular green fluorescence signal. In contrast, in LVV EC of Rasa1fl/fl Cdh5ert2cre embryos, green fluorescence signals should be visibly absent in regions of the cell that contain large intracellular accumulations of collagen IV, i.e. in the ER. As shown in Fig. 5 and Fig. S3, in LVV of Rasa1fl/fl Cdh5ert2cre embryos, individual LVV EC could be identified that contained large accumulations of collagen IV that displaced the cytoplasmic eGFP signal. No such accumulations of collagen IV were identified in LVV EC of Rasa1fl/fl embryos (Fig. 5). Intracellular retention of collagen IV in EC of LVV of Rasa1fl/fl Cdh5ert2cre embryos was confirmed by co-staining for the cell surface marker CD31 (Pecam1) (Fig. S4). In contrast to collagen IV, the ECM protein fibronectin was distributed normally in LVV of Rasa1fl/fl Cdh5ert2cre embryos and there was no evidence of intracellular accumulation (Fig. S5).

Fig. 5.

Accumulation of collagen IV in EC of developing RASA1-deficient LVV.Rasa1fl/fl Prox1-eGFP and Rasa1fl/fl Cdh5ert2cre Prox1-eGFP embryos were administered TM at E12.5 and harvested at E15.5. Longitudinal sections through the neck region were stained with antibodies against collagen IV and the Hoechst nuclear stain. Bottom panels show magnification of boxed areas above. LS, lymphatic sac. Note intracellular accumulation of collagen IV that corresponds to regions of displacement of the eGFP signal in EC of Rasa1fl/fl Cdh5ert2cre LVV (arrows). Additional images of collagen IV accumulation in EC of LVV of Rasa1fl/fl Cdh5ert2cre embryos are provided in Fig. S3.

Fig. 5.

Accumulation of collagen IV in EC of developing RASA1-deficient LVV.Rasa1fl/fl Prox1-eGFP and Rasa1fl/fl Cdh5ert2cre Prox1-eGFP embryos were administered TM at E12.5 and harvested at E15.5. Longitudinal sections through the neck region were stained with antibodies against collagen IV and the Hoechst nuclear stain. Bottom panels show magnification of boxed areas above. LS, lymphatic sac. Note intracellular accumulation of collagen IV that corresponds to regions of displacement of the eGFP signal in EC of Rasa1fl/fl Cdh5ert2cre LVV (arrows). Additional images of collagen IV accumulation in EC of LVV of Rasa1fl/fl Cdh5ert2cre embryos are provided in Fig. S3.

Partial rescue of LVV development by a chemical chaperone that promotes folding of collagen IV

Collagen IV becomes trapped in the ER of RASA1-deficient EC as a consequence of impaired collagen IV folding. Accordingly, 4-phenylbutyric acid (4PBA), a chemical chaperone that promotes folding of collagen IV in the ER (Jeanne et al., 2015; Kuo et al., 2014), rescues EC export of collagen IV and EC apoptosis (Chen et al., 2019). Therefore, we examined whether 4PBA could rescue LVV and central VV development in the absence of RASA1. 4PBA was administered along with TM to Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos at E12.5 and on all subsequent days until embryo harvest at E16.5. 4PBA partially rescued LVV development in these experiments (Fig. 6A,B). In TM plus 4PBA-treated Rasa1fl/fl Cdh5ert2cre embryos, approximately half of examined LV-BV junctions contained one LVV and the remaining half contained two LVV (compared with no LVV in the majority of LV-BV junctions in Rasa1fl/fl Cdh5ert2cre embryos that received TM alone), albeit that rescued valves were ∼25% shorter than in control LVV (Fig. 6C). No influence of 4PBA upon LVV development was noted in TM-treated Rasa1fl/fl embryos (Fig. 6A,B). In addition, 4PBA partially rescued development of the EJV VV, although was unable to rescue development of the IJV and SCV VV (Fig. 6A,D). These findings provide support for the idea that impaired development of LVV and, in part, the EJV VV, is a consequence of an inability of EC to export collagen IV for deposition in the ECM core of the respective valve leaflets. Consistent with this, close inspection of 4PBA-rescued LVV revealed normal export of collagen IV by LVV EC and deposition in the LVV ECM core (Fig. 6E).

Fig. 6.

Partial rescue of LVV development in induced EC-specific RASA1-deficient embryos by the chemical chaperone 4PBA. (A-E) Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos were administered TM and 4PBA at E12.5 and 4PBA every day thereafter until embryo harvest at E16.5. Longitudinal sections of the neck region (dorsal to ventral from top to bottom) were stained with the indicated antibodies. In A, arrows and arrowheads indicate LVV and VV, respectively. (B,D) Percentage of LV-BV junctions from Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos from A that contain the indicated number of LVV (B) and categories of VV (D, see Fig. 2D). Rasa1fl/fl n=4, Rasa1fl/fl Cdh5ert2cre n=13. Note that data from Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos treated with TM alone from Fig. 2C,E (red and black bars) is also plotted for comparison. (C) Mean+s.e.m. of LVV length of embryos of the indicated genotypes treated with 4PBA (Rasa1fl/fl n=6 LVV, Rasa1fl/fl Cdh5ert2cre n=10 LVV). *P<0.05, Student's 2-sample t-test. (E) Normal deposition of collagen IV in LVV of Rasa1fl/fl Cdh5ert2cre embryos treated with 4PBA. Right panel shows magnification of boxed area on left. EJV, external jugular vein; IJV, internal jugular vein; LS, lymphatic sac; SCV, subclavian vein; SVC, superior vena cava.

Fig. 6.

Partial rescue of LVV development in induced EC-specific RASA1-deficient embryos by the chemical chaperone 4PBA. (A-E) Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos were administered TM and 4PBA at E12.5 and 4PBA every day thereafter until embryo harvest at E16.5. Longitudinal sections of the neck region (dorsal to ventral from top to bottom) were stained with the indicated antibodies. In A, arrows and arrowheads indicate LVV and VV, respectively. (B,D) Percentage of LV-BV junctions from Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos from A that contain the indicated number of LVV (B) and categories of VV (D, see Fig. 2D). Rasa1fl/fl n=4, Rasa1fl/fl Cdh5ert2cre n=13. Note that data from Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos treated with TM alone from Fig. 2C,E (red and black bars) is also plotted for comparison. (C) Mean+s.e.m. of LVV length of embryos of the indicated genotypes treated with 4PBA (Rasa1fl/fl n=6 LVV, Rasa1fl/fl Cdh5ert2cre n=10 LVV). *P<0.05, Student's 2-sample t-test. (E) Normal deposition of collagen IV in LVV of Rasa1fl/fl Cdh5ert2cre embryos treated with 4PBA. Right panel shows magnification of boxed area on left. EJV, external jugular vein; IJV, internal jugular vein; LS, lymphatic sac; SCV, subclavian vein; SVC, superior vena cava.

Rescue of LVV development by an inhibitor of collagen IV modifying enzymes

Folding of 2α1 and 1α2 collagen IV monomers into a triple helical protomer conformation in the ER is a highly regulated process that requires multiple post-translational modifications of monomers mediated by protein disulphide isomerases (PDIs), protein peptidyl isomerases (PPIs), proline-3 and proline-4 hydroxylases (P3H1-3 and P4HA1-3, respectively) and lysine hydroxylases and glycosylases (procollagen-lysine, 2-oxoglutarate 5-dioxygenases; PLODs 1-3) (Chioran et al., 2017; Ishikawa and Bächinger, 2013; Mao et al., 2015; Ricard-Blum, 2011). Under- or over-post-translational modification of collagen IV monomers can affect folding of monomers into the triple helical conformation. Proteomic analyses of induced RASA1-deficient embryonic blood vascular EC (BEC) established that, of the P3H, P4HA and PLOD enzymes that could be detected in these cells (seven out of a total of nine), all were increased in abundance compared with control RASA1-sufficient BEC (Chen et al., 2019). P3Hs, P4HAs and PLODs all belong to the same family of enzymes known as 2-oxoglutarate (2OG)-dependent oxygenases (2OG-DO) and, consequently, can be collectively inhibited by drugs such as 2,4 pyridinedicarboxylic acid (2,4 PDCA) (Gilkes et al., 2013; Rose et al., 2011). Consistent with the notion that increased abundance of 2OG-DO in embryonic RASA1-deficient BEC is responsible for the impaired folding and export of collagen IV, 2,4 PDCA rescued collagen IV retention and BEC apoptosis in induced global RASA1-deficient embryos (Chen et al., 2019). Based on these findings, we asked whether 2,4 PDCA could also rescue LVV and VV development in the absence of RASA1. As in previous drug rescue experiments, 2,4 PDCA was administered with TM to E12.5 Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos and on successive days thereafter until embryo harvest at E16.5 (Fig. 7). 2,4 PDCA was able to rescue the impaired development of LVV in the absence of RASA1 (Fig. 7A). Following drug treatment, the majority of LV-BV junctions contained 2 LVV (Fig. 7B). Furthermore, 2,4 PDCA restored export of collagen IV by LVV EC (Fig. 7D). In addition, 2,4 PDCA almost completely rescued VV development. Following drug treatment, the majority of examined BV junctions contained normal IJV, EJV and SCV VV (Fig. 7C).

Fig. 7.

Partial rescue of LVV development in induced EC-specific RASA1-deficient embryos by 2,4 PDCA. (A-D) Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos were administered TM and 2,4 PDCA at E12.5 and 2,4 PDCA every day thereafter until embryo harvest at E16.5. Longitudinal sections of the neck region (dorsal to ventral from top to bottom) were stained with the indicated antibodies. In A, arrows and arrowheads indicate LVV and VV, respectively. (B,C) Percentage of LV-BV junctions from Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos from A that contain the indicated number of LVV (B) and categories of VV (C, see Fig. 2D). Rasa1fl/fl n=11, Rasa1fl/fl Cdh5ert2cre n=14. Note that data from Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos treated with TM alone from Fig. 2C,E (red and black bars) is also plotted for comparison. (D) Normal deposition of collagen IV in LVV of Rasa1fl/fl Cdh5ert2cre embryos treated with 2,4 PDCA. Right panel shows magnification of boxed area on left. EJV, external jugular vein; IJV, internal jugular vein; LS, lymphatic sac; SCV, subclavian vein; SVC, superior vena cava.

Fig. 7.

Partial rescue of LVV development in induced EC-specific RASA1-deficient embryos by 2,4 PDCA. (A-D) Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos were administered TM and 2,4 PDCA at E12.5 and 2,4 PDCA every day thereafter until embryo harvest at E16.5. Longitudinal sections of the neck region (dorsal to ventral from top to bottom) were stained with the indicated antibodies. In A, arrows and arrowheads indicate LVV and VV, respectively. (B,C) Percentage of LV-BV junctions from Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos from A that contain the indicated number of LVV (B) and categories of VV (C, see Fig. 2D). Rasa1fl/fl n=11, Rasa1fl/fl Cdh5ert2cre n=14. Note that data from Rasa1fl/fl and Rasa1fl/fl Cdh5ert2cre embryos treated with TM alone from Fig. 2C,E (red and black bars) is also plotted for comparison. (D) Normal deposition of collagen IV in LVV of Rasa1fl/fl Cdh5ert2cre embryos treated with 2,4 PDCA. Right panel shows magnification of boxed area on left. EJV, external jugular vein; IJV, internal jugular vein; LS, lymphatic sac; SCV, subclavian vein; SVC, superior vena cava.

RASA1 is necessary for the maintenance of central VV in adults

To determine whether RASA1 is necessary for the maintenance of central VV in adults, we administered TM to adult littermate Rasa1fl/fl and Rasa1fl/fl Ubert2cre mice and examined VV function 3 months later. As a first approach, we examined central venous pressure that would be expected to be elevated in mice as a consequence of impaired VV function (Eberhardt and Raffetto, 2014). Central venous pressure was monitored by telemetry throughout the 24-h daily cycle for 6 days for each examined mouse (Fig. 8A). In these experiments Rasa1fl/fl Ubert2cre mice showed increased central venous pressure compared with Rasa1fl/fl controls, consistent with impaired VV function.

Fig. 8.

A required role for RASA1 in the maintenance of central VV function in adult mice. (A) TM was administered to adult Rasa1fl/fl and Rasa1fl/fl Ubert2cre mice and central venous blood pressure was measured by telemetry 3 months later. Left shows the mean+s.e.m. of the central venous pressure for three individual mice of each genotype based upon continuous monitoring over 6 days (see Materials and Methods). Right shows the mean+s.e.m. of the mean central venous pressure of individual mice on the left. *P<0.05; Student's 2-sample t-test. (B) Image of a dissected cervical vein from a Prox1-eGFP mouse showing positions of Pin, Pout and Psn micropipettes in low pressure back-leak assays (top) and the fluorescent eGFP-positive VV (bottom). (C) Low pressure back-leak tests of cervical VV from adult Rasa1fl/fl Prox1-eGFP and Rasa1fl/fl Ubert2cre Prox1-eGFP mice administered TM 3 months previously. Rasa1fl/fl Prox1-eGFP n=9, Rasa1fl/fl Ubert2cre Prox1-eGFP n=8. Note leakage of Rasa1fl/fl Ubert2cre Prox1-eGFP valves throughout the Pout pressure range. (D) Whole-mount images of representative cervical veins from Rasa1fl/fl Prox1-eGFP and Rasa1fl/fl Ubert2cre Prox1-eGFP mice in C. Vessels were stained with an anti-CD31 antibody before mounting. (E) Mean+s.e.m. of the number of EC per VV leaflet. Rasa1fl/fl Prox1-eGFP n=8, Rasa1fl/fl Ubert2cre Prox1-eGFP n=6. **P<0.01, Student's 2-sample t-test.

Fig. 8.

A required role for RASA1 in the maintenance of central VV function in adult mice. (A) TM was administered to adult Rasa1fl/fl and Rasa1fl/fl Ubert2cre mice and central venous blood pressure was measured by telemetry 3 months later. Left shows the mean+s.e.m. of the central venous pressure for three individual mice of each genotype based upon continuous monitoring over 6 days (see Materials and Methods). Right shows the mean+s.e.m. of the mean central venous pressure of individual mice on the left. *P<0.05; Student's 2-sample t-test. (B) Image of a dissected cervical vein from a Prox1-eGFP mouse showing positions of Pin, Pout and Psn micropipettes in low pressure back-leak assays (top) and the fluorescent eGFP-positive VV (bottom). (C) Low pressure back-leak tests of cervical VV from adult Rasa1fl/fl Prox1-eGFP and Rasa1fl/fl Ubert2cre Prox1-eGFP mice administered TM 3 months previously. Rasa1fl/fl Prox1-eGFP n=9, Rasa1fl/fl Ubert2cre Prox1-eGFP n=8. Note leakage of Rasa1fl/fl Ubert2cre Prox1-eGFP valves throughout the Pout pressure range. (D) Whole-mount images of representative cervical veins from Rasa1fl/fl Prox1-eGFP and Rasa1fl/fl Ubert2cre Prox1-eGFP mice in C. Vessels were stained with an anti-CD31 antibody before mounting. (E) Mean+s.e.m. of the number of EC per VV leaflet. Rasa1fl/fl Prox1-eGFP n=8, Rasa1fl/fl Ubert2cre Prox1-eGFP n=6. **P<0.01, Student's 2-sample t-test.

To examine central VV function directly, cervical veins were dissected from mice and VV function was tested in low pressure back-leak assays. In these assays, veins were trimmed to contain a single VV and were cannulated at both ends to allow manipulation of intraluminal pressure upstream (Pin) and downstream (Pout) of the valve. For back-leak assays, Pin was held constant at 0.5 cm H2O and Pout was elevated over a physiological range from 0.5 to 10 cm of H2O. The ability of valves to resist backflow in response to elevated pressure downstream of the valve was measured with the use of a servo null micropipette (Psn) inserted through the vessel wall into the upstream lumen (Fig. 8B). Cervical VV from Rasa1fl/fl mice were able to resist transfer of elevated downstream pressure to the upstream vessel lumen starting at around 4 cm of H2O (Fig. 8C). In contrast, cervical VV from Rasa1fl/fl Ubert2cre mice were unable to prevent backflow throughout the entire 0.5 to 10 cm H2O range (Fig. 8C). Thus, cervical VV are functionally impaired in Rasa1fl/fl Ubert2cre mice.

Impaired LV valve function in adult induced RASA1-deficient mice is explained by reduced cellularity of LV valve leaflets resulting in shortening of leaflets and failed LV valve closure (Lapinski et al., 2017). Therefore, we examined the cellularity of cervical VV leaflets in induced RASA1-deficient and control mice. As determined with the use of the Prox1-eGFP transgene, which allows ready identification and enumeration of VV EC, induced loss of RASA1 in adult mice resulted in significantly fewer EC in cervical VV leaflets compared with controls (Fig. 8D,E). Therefore, RASA1 is necessary to maintain numbers of EC in VV leaflets, which accounts for impaired VV function in induced RASA1-deficient mice.

In this study, we show that RASA1 performs a non-redundant role in the development of LVV and the development and maintenance of VV. These findings extend our knowledge of the function of this RasGAP in both blood and lymphatic vascular systems. A role for RASA1 in the continued development of LVV and VV, but not their initial specification, parallels our earlier reported finding regarding the role of RASA1 in LV valve development (Lapinski et al., 2017). LVV and VV development was also impaired in embryos in which RASA1 R780Q alone was expressed in EC. However, the impairment in LVV and VV development in these embryos was less than in embryos that lacked RASA1 in EC. This suggests that RASA1 regulates LVV and VV development through mechanisms that are independent of its GAP activity as well as mechanisms that are dependent upon its GAP activity. For LVV development, this is consistent with the finding that, although an MAPK inhibitor was able to rescue valve development, rescue was incomplete. Moreover, for VV development, the same MAPK inhibitor was largely unable to rescue valve development in the absence of RASA1. Previously, we found that the RASA1 R780Q form of RASA1 was able to partially rescue the impaired function of adult LV valves resulting from induced loss of RASA1 (Lapinski et al., 2017). Thus, the time to development of LV valve dysfunction in adult mice induced to express RASA1 R780Q alone is substantially delayed compared with adult mice induced to lose RASA1 completely. Which Ras-independent signaling pathways contribute to LVV and VV development and maintenance of mature LV valves remains to be determined. In addition to its GAP domain, RASA1 contains several modular-binding domains including SH2, SH3, PH and C2 homology domains (King et al., 2013). Potentially, through interaction with different signaling moieties, one or more of these binding domains could link RASA1 to distinct signaling pathways that control valve development and maintenance.

Abnormalities of LVV are first apparent at E15.5 following disruption of Rasa1 at E12.5. At this time, LEC and BEC layers of forming LVV detach from one another. In addition, intracellular accumulation of collagen IV can be detected in EC of valves at this time point. Intracellular accumulation of collagen IV is ultimately observed in all EC in which Rasa1 is disrupted before E15.5 and results in their apoptotic death, either as a result of anoikis (detachment from the substratum) or induction of an unfolded protein response (Chen et al., 2019). We propose that impaired export of collagen IV from EC of induced RASA1-deficient LVV, VV and LV valves is responsible for failed valve development. An inability of RASA1-deficient valve forming EC to export collagen IV for deposition in the developing ECM core of valves would be expected to result in EC detachment from the valve and/or induction of a unfolded protein response (UPR). Blocked export of collagen IV from RASA1-deficient embryonic EC and hemorrhage can be rescued by 4PBA, which promotes folding of collagen IV in the ER. Similarly, in the current study, 4PBA largely rescued the impasse in LVV development and partially rescued the block in VV development in the absence of RASA1. These findings support the concept that collagen IV is retained within EC of developing RASA1-deficient LVV and VV because it is improperly folded. Our earlier proteomic studies of RASA1-deficient BEC identified increases in the abundance of all detectable collagen IV-modifying 2OG-DO enzymes in these cells (Chen et al. 2019). Evidence that this increase is responsible for collagen IV misfolding and retention in BEC was provided by the finding that drugs such as ethyl-3,4-dihydroxybenzoic acid (EDHB) and 2,4 PDCA, which broadly inhibit this class of enzymes, rescued collagen IV export from BEC and prevented hemorrhage. Similarly, as shown here, 2,4 PDCA rescued both LVV and VV development in the absence of RASA1. This is consistent with a model in which an increased abundance of 2OG-DO in EC of LVV and VV results in collagen IV retention in the ER, which in turn is responsible for blocked development of these valves in the absence of RASA1.

Induced loss of RASA1 in adult mice has not previously been reported to result in any spontaneous BV phenotypes (Lapinski et al., 2012). This can be explained on the grounds that deposition of collagen IV in BM primarily occurs during developmental angiogenesis. Collagen IV is an inherently stable molecule, thus obviating a requirement for EC to continue to engage in high rate collagen IV synthesis in postnatal life (Shoulders and Raines, 2009). A requirement of RASA1 for VV maintenance in adults represents an exception to this and parallels a requirement of RASA1 in the maintenance of LV valves in adults (Lapinski et al., 2017). Dysfunction of LV valves in induced RASA1-deficient mice can be explained by gradual loss of LEC from LV valve leaflets resulting in leaflet shortening below a threshold level that is necessary to prevent fluid backflow when closed leaflets overlap. Similarly, we show here that VV dysfunction is associated with reduced cellularity of VV leaflets. EC in LV valves and VV would be subject to higher shear stress forces than EC in LV and BV walls to the extent that continued high rate synthesis of collagen IV may be necessary for these cells to remain attached to the valve leaflet (Bazigou and Makinen, 2013; Bazigou et al., 2014). Indeed, it is established that EC synthesize higher amounts of collagen IV when subject to shear stress in vitro (Yamane et al., 2010). In this regard, induced loss of RASA1 in adult LV valve and VV EC may result in their detachment from leaflets or apoptotic death consequent to induction of a UPR, which may be responsible for the leaflet shortening over time.

Last, based upon the studies reported herein, it is conceivable that in CM-AVM, acquired somatic second hit mutations of the RASA1 gene in LVV-forming and VV-forming cells or their precursors could result in impaired development and function of affected LVV and VV in these patients. Whether or not CM-AVM patients show abnormalities of LVV and VV function is currently unknown.

Mice

Rasa1fl/fl and Rasa1fl/R780Q mice with and without Ubert2cre transgenes have been previously described (Lapinski et al., 2007, 2012, 2017; Lubeck et al., 2014). Cdh5ert2cre mice were obtained from Cancer Research UK (Wang et al., 2010). Prox1-eGFP mice were obtained from Dr Young Kwon Hong at the University of Southern California. Rasa1fl/fl Cdh5ert2cre and Rasa1fl/fl Prox1-eGFP mice with and without Cdh5ert2cre and Ubert2cre were generated through cross-breeding. All mice were on a mixed 129S6/SvEv×C57BL/6 genetic background. All experiments performed with mice were in compliance with University of Michigan and University of Missouri guidelines and were approved by the respective university committees on the use and care of animals.

LVV and VV specification and development

Pregnant Rasa1fl/fl mice carrying Rasa1fl/fl embryos with and without Cdh5ert2cre and pregnant Rasa1fl/fR780Q mice carrying Rasa1fl/fl, Rasa1fl/R780Q and Rasa1R780Q/R780Q embryos with and without Ubert2cre were given two intraperitoneal (i.p.) injections of TM on consecutive days starting at E10.5, E12.5 or E14.5 of gestation. For some pregnant mice, AZD6244 (Selleckchem; 0.05 mg/g body weight per injection), 4PBA (Sigma; 0.25 mg/g body weight per injection), or 2,4 PDCA (Sigma; 0.1 mg/g body weight per injection) was injected i.p. at the same time as TM and on subsequent days until embryo harvest. Embryos were harvested at different times and fixed in 4% paraformaldehyde overnight. For embryos that were harvested at E14.5 and later, an abdominal incision was made and blood was drained from embryos before fixation.

Immunohistochemistry on sections was performed as previously described (Geng et al., 2016). Fixed embryos were embedded in OCT solution (Sakura) and 12 μm thick cryosections were prepared in the frontal orientation using a cryotome (Thermo Fisher Scientific, HM525 NX). Primary antibodies used were: goat anti-VEGFR3 (R&D Systems, AF743, 1:300), rabbit anti-PROX1 (AngioBio, 11-002, 1:500), rat anti-CD31 (BD Pharmingen, 553370, 1:500), goat anti-PROX1 (AF2727, R&D Systems, 1:500), goat anti-collagen IV (1340-01, Southern Biotech, 1:100), rabbit anti-pERK (D13.14.4E, Cell Signaling Technology, 1:100), and mouse anti-fibronectin (FN-3E2, Sigma, 1:50). Secondary antibodies were Alexa Fluor 488-donkey anti-goat (705-545-147), Alexa Fluor 488 AffiniPure F (ab′)2-donkey anti-rabbit (711-546-152), Alexa Fluor 488 AffiniPure F (ab′)2-donkey anti-goat (705-546-147), Alexa Fluor 488 AffiniPure F (ab′)2-donkey anti-goat IgG (705-546-147), Alexa Fluor 594 AffiniPure F (ab′)2-donkey anti-mouse (715-586-151), Alexa Fluor 594 AffiniPure F (ab′)2-donkey anti-goat (705-586-147), Alexa Fluor 594 AffiniPure F (ab′)2-donkey anti-rabbit (715-586-152), Alexa Fluor 594 AffiniPure F (ab′)2-donkey anti-rat (712-586-150), Cy3-donkey anti-rabbit (711-165-152) and Cy5-donkey anti-rat (712-175-150) (all Jackson ImmunoResearch Laboratories, 1:500). The antibodies were diluted in blocking buffer [PBS+ 0.1% Triton X-100+0.1% bovine serum albumin (BSA)] and immunohistochemistry was performed using standard protocols. After mounting, the sections were visualized using an Eclipse 80i microscope (Nikon) equipped with a Zyla sCMOS camera (Andor Technology) and analyzed using NIS-Elements BR software (Nikon). Several consecutive sections were analyzed to determine the presence or absence of LVVs and VVs.

SEM was performed according to our previous protocol (Geng et al., 2016; Geng and Srinivasan, 2018). Briefly, 500 µm sections were prepared using a vibratome (Leica, VT 1000S). Sections were fixed in 2% glutaraldehyde in 0.1 M cacodylate buffer for 2 h. After washing profusely in 0.1 M cacodylate buffer, the sections were post fixed in 1% osmium tetroxide in 0.1 M cacodylate buffer for 2 h and subsequently dehydrated in a graded ethanol series. The sections were further dehydrated in hexamethyldisilazane and allowed to air-dry overnight. Dry sections were sputter-coated with Au/Pd particles (Med-010 Sputter Coater by Balzers-Union) and observed under a Quanta SEM (FEI) at an accelerating voltage of 20 KV.

Central venous pressure

Three-month old littermate Rasa1fl/fl and Rasa1fl/fl Ubert2cre mice were given two i.p. injections of TM (0.05 mg/g body weight per injection, dissolved in corn oil) on consecutive days. One week later, central venous pressure was determined by radiotelemetry using an implantable microminiaturized electronic monitor (PA-C10; Data Sciences International) as previously described (Whitesall et al., 2004) with minor modifications. Briefly, the catheter of the device was passed into the right EJV and the transducer was placed in the abdominal cavity. After surgery, mice were individually housed in cages atop receiver pads allowing for real-time measurements of venous pressure, heart rate and activity. To minimize the influence of fluctuations in venous pressure resulting from changes in respiration and activity, for each mouse an average central venous pressure was calculated based upon readings taken every 10 s for 15 min every 3 h for 6 days total.

Vein dissection and cannulation

Three-month old littermate Rasa1fl/fl Prox1-eGFP mice with and without Ubert2cre were given two i.p. injections of TM as above. After 12 weeks, mice were anesthetized with sodium pentobarbital (60 mg/kg, i.p.) and placed on a heating pad. The skin under the chin and upper chest was shaved and a ventral incision was made to expose the underlying cervical and EJV in association with superior cervical afferent LV. The LV and attached connective tissue were removed and the EJV was cut centrally, held with fine forceps and pulled caudally while trimming along the sides of the vessel and its major branch, the cervical vein (and sometimes also other minor branches), until a segment ∼3-4 mm in length was obtained. Some vessels were unbranched (Fig. 8B) but in most cases the cervical vein joined the EJV immediately distal to the valve. After cutting the distal ends of the main vessel and any branches, the entire segment was removed and placed in room temperature Krebs-BSA solution [146.9 mM NaCl; 4.7 mM KCl; 2 mM CaCl2·2H2O; 1.2 mM MgSO4; 1.2 mM NaH2PO4·H2O; 3 mM NaHCO3; 1.5 mM sodium-Hepes; 5 mM D-glucose; 0.5% BSA (pH 7.4 at 37°C)] in a dissection dish with a 1 cm layer of Sylgard (Dow Corning). The mouse was then euthanized with an intracardiac injection of KCl (0.2 M). After pinning the segment with pieces of 40 µm wire, the remaining connective tissue and fat were removed and the segment was transferred to a 3 ml cannulation chamber. The proximal and distal ends were cannulated onto 90 µm glass micropipettes filled with Krebs-BSA and tied with 12-0 suture. After pressurization, most of the red blood cells were flushed out and any side branches were identified and tied with 12-0 suture. The cannulation chamber, with attached pipette holders and vessel, were transferred to the stage of an inverted microscope, where the segment was heated to 37°C and perfused with Krebs buffer (0.5 ml/min). Inflow and outflow pressures were set initially using standing fluid reservoirs and the axial length was adjusted to remove slack with luminal pressure briefly set to 10 cm H2O. The segment was then equilibrated for 30 min at 2 cm H2O luminal pressure. Development of spontaneous tone confirmed vessel viability but the venous valve tests described below were performed after perfusion of the segment with Ca2+-free Krebs (Krebs with 3 mM EDTA replacing CaCl2·2H2O) for at least 20 min to eliminate spontaneous tone. The vessel image was digitized using a fire-wire camera (model A641FM; Basler) and inner diameter was continuously tracked using a custom computer algorithm (Davis et al., 2011). All valve tests were recorded as AVI files, with embedded pressure and diameter data, for later replay and valve tracking, as needed.

Valve back-leak assays

To conduct tests of valve function, the pressure control for each cannulation pipette was switched from the reservoirs to a two-channel pump controller (Cardiovascular Research Institute, Texas A&M University) driven through a D-A interface by a LabVIEW program (National Instruments) running under Windows 7. A servo-nulling micropipette with tip diameter ∼3 µm was inserted through the vessel wall on the upstream side of the valve to measure the local pressure (Psn) at that site (Davis et al., 2011). Psn, inflow pressure (Pin), outflow pressure (Pout) and diameter were recorded at 30 Hz using a model PCI 6030e A-D interface (National Instruments). After micropuncture, Pin and Pout were raised simultaneously from 0.5 to 10 cm H2O in order to verify the calibration of the servo-nulling system. To assess the ability of a closed valve to prevent pressure back leak, Pin was set to 0.5 cm H2O and Pout was elevated ramp-wise from 0.5 to 10 cm H2O over a 2-min period. The servo-nulling pipette detected any rise in pressure upstream from the valve. The Pout ramp was repeated three times. For data analysis, Psn and Pout values were binned in 1 cm H2O intervals for each ramp, using a LabVIEW program, to obtain average values of Psn and Pout for plotting and statistical tests.

Venous valve whole-mount staining

Vessels from VV function assays were fixed in 1% paraformaldehyde overnight, blocked by incubation in PBS/10% donkey serum/0.3% Triton X-100 and incubated overnight with rat anti-CD31 in PBS/10% donkey serum/0.3% Triton X-100. Vessels were subsequently incubated with donkey anti-rat Ig coupled to Alexa Fluor 594 (Jackson ImmunoResearch, 1:500) in PBS for 2 h before viewing on a Leica SP5 X confocal microscope. PROX1-positive EC in VV leaflets were enumerated using Imaris software (Bitplane).

Statistical analysis

P values were calculated using two-tailed Student's 2-sample 2-sided t-test.

We acknowledge the assistance of the Michigan Medicine Physiology Phenotyping Core for measurement of central venous pressure in mice.

Author contributions

Conceptualization: D.C., X.G., P.E.L., M.J.D., R.S.S., P.D.K.; Methodology: D.C., X.G., P.E.L., M.J.D., P.D.K.; Formal analysis: D.C., X.G., M.J.D., R.S.S., P.D.K.; Investigation: D.C., X.G., P.E.L., M.J.D.; Writing - original draft: P.D.K.; Writing - review & editing: D.C., X.G., P.E.L., M.J.D., R.S.S.; Supervision: P.D.K.; Project administration: P.D.K.; Funding acquisition: R.S.S., P.D.K.

Funding

This work was supported by the National Institutes of Health [HL120888 and HL146352 to P.D.K., HL131652 and HL133216 to R.S.S., HL120867 to M.J.D., and GM103441 to X.G. (PI: Dr McEver)]. Deposited in PMC for release after 12 months.

Bazigou
,
E.
and
Makinen
,
T.
(
2013
).
Flow control in our vessels: vascular valves make sure there is no way back
.
Cell. Mol. Life Sci.
70
,
1055
-
1066
.
Bazigou
,
E.
,
Lyons
,
O. T. A.
,
Smith
,
A.
,
Venn
,
G. E.
,
Cope
,
C.
,
Brown
,
N. A.
and
Makinen
,
T.
(
2011
).
Genes regulating lymphangiogenesis control venous valve formation and maintenance in mice
.
J. Clin. Invest.
121
,
2984
-
2992
.
Bazigou
,
E.
,
Wilson
,
J. T.
and
Moore
,
J. E.
Jr
. (
2014
).
Primary and secondary lymphatic valve development: molecular, functional and mechanical insights
.
Microvasc. Res.
96
,
38
-
45
.
Bos
,
J. L.
,
Rehmann
,
H.
and
Wittinghofer
,
A.
(
2007
).
GEFs and GAPs: critical elements in the control of small G proteins
.
Cell
129
,
865
-
877
.
Buday
,
L.
and
Downward
,
J.
(
2008
).
Many faces of Ras activation
.
Biochim. Biophys. Acta
1786
,
178
-
187
.
Burrows
,
P. E.
,
Gonzalez-Garay
,
M. L.
,
Rasmussen
,
J. C.
,
Aldrich
,
M. B.
,
Guilliod
,
R.
,
Maus
,
E. A.
,
Fife
,
C. E.
,
Kwon
,
S.
,
Lapinski
,
P. E.
,
King
,
P. D.
, et al. 
(
2013
).
Lymphatic abnormalities are associated with RASA1 gene mutations in mouse and man
.
Proc. Natl. Acad. Sci. USA
.
110
,
8621
-
8626
.
Butler
,
J.
,
Watson
,
H. R.
,
Lee
,
A. G.
,
Schuppe
,
H.-J.
and
East
,
J. M.
(
2011
).
Retrieval from the ER-golgi intermediate compartment is key to the targeting of c-terminally anchored ER-resident proteins
.
J. Cell. Biochem.
112
,
3543
-
3548
.
Chen
,
D.
,
Teng
,
J. M.
,
North
,
P. E.
,
Lapinski
,
P. E.
and
King
,
P. D.
(
2019
).
RASA1-dependent cellular export of collagen IV controls blood and lymphatic vascular development
.
J. Clin. Invest.
129
,
3545
-
3561
.
Chioran
,
A.
,
Duncan
,
S.
,
Catalano
,
A.
,
Brown
,
T. J.
and
Ringuette
,
M. J.
(
2017
).
Collagen IV trafficking: The inside-out and beyond story
.
Dev. Biol.
431
,
124
-
133
.
Choi
,
I.
,
Chung
,
H. K.
,
Ramu
,
S.
,
Lee
,
H. N.
,
Kim
,
K. E.
,
Lee
,
S.
,
Yoo
,
J.
,
Choi
,
D.
,
Lee
,
Y. S.
,
Aguilar
,
B.
, et al. 
(
2011
).
Visualization of lymphatic vessels by Prox1-promoter directed GFP reporter in a bacterial artificial chromosome-based transgenic mouse
.
Blood
117
,
362
-
365
.
Davis
,
M. J.
,
Rahbar
,
E.
,
Gashev
,
A. A.
,
Zawieja
,
D. C.
and
Moore
,
J. E.
Jr
. (
2011
).
Determinants of valve gating in collecting lymphatic vessels from rat mesentery
.
Am. J. Physiol. Heart Circ. Physiol.
301
,
H48
-
H60
.
de Wijn
,
R. S.
,
Oduber
,
C. E. U.
,
Breugem
,
C. C.
,
Alders
,
M.
,
Hennekam
,
R. C. M.
and
van der Horst
,
C. M. A. M.
(
2012
).
Phenotypic variability in a family with capillary malformations caused by a mutation in the RASA1 gene
.
Eur. J. Med. Genet.
55
,
191
-
195
.
Eberhardt
,
R. T.
and
Raffetto
,
J. D.
(
2014
).
Chronic venous insufficiency
.
Circulation
130
,
333
-
346
.
Eerola
,
I.
,
Boon
,
L. M.
,
Mulliken
,
J. B.
,
Burrows
,
P. E.
,
Dompmartin
,
A.
,
Watanabe
,
S.
,
Vanwijck
,
R.
and
Vikkula
,
M.
(
2003
).
Capillary malformation-arteriovenous malformation, a new clinical and genetic disorder caused by RASA1 mutations
.
Am. J. Hum. Genet.
73
,
1240
-
1249
.
Geng
,
X.
and
Srinivasan
,
R. S.
(
2018
).
Correlative fluorescence and scanning electron microscopy to study lymphovenous valve development
.
Methods Mol. Biol.
1846
,
85
-
96
.
Geng
,
X.
,
Cha
,
B.
,
Mahamud
,
M. R.
,
Lim
,
K.-C.
,
Silasi-Mansat
,
R.
,
Uddin
,
M. K. M.
,
Miura
,
N.
,
Xia
,
L.
,
Simon
,
A. M.
,
Engel
,
J. D.
, et al. 
(
2016
).
Multiple mouse models of primary lymphedema exhibit distinct defects in lymphovenous valve development
.
Dev. Biol.
409
,
218
-
233
.
Geng
,
X.
,
Cha
,
B.
,
Mahamud
,
M. R.
and
Srinivasan
,
R. S.
(
2017
).
Intraluminal valves: development, function and disease
.
Dis. Model. Mech.
10
,
1273
-
1287
.
Gilkes
,
D. M.
,
Chaturvedi
,
P.
,
Bajpai
,
S.
,
Wong
,
C. C.
,
Wei
,
H.
,
Pitcairn
,
S.
,
Hubbi
,
M. E.
,
Wirtz
,
D.
and
Semenza
,
G. L.
(
2013
).
Collagen prolyl hydroxylases are essential for breast cancer metastasis
.
Cancer Res.
73
,
3285
-
3296
.
Henkemeyer
,
M.
,
Rossi
,
D. J.
,
Holmyard
,
D. P.
,
Puri
,
M. C.
,
Mbamalu
,
G.
,
Harpal
,
K.
,
Shih
,
T. S.
,
Jacks
,
T.
and
Pawson
,
T.
(
1995
).
Vascular system defects and neuronal apoptosis in mice lacking ras GTPase-activating protein
.
Nature
377
,
695
-
701
.
Ishikawa
,
Y.
and
Bächinger
,
H. P.
(
2013
).
A molecular ensemble in the rER for procollagen maturation
.
Biochim. Biophys. Acta
1833
,
2479
-
2491
.
Jeanne
,
M.
,
Jorgensen
,
J.
and
Gould
,
D. B.
(
2015
).
Molecular and genetic analyses of collagen type IV mutant mouse models of spontaneous intracerebral hemorrhage identify mechanisms for stroke prevention
.
Circulation
131
,
1555
-
1565
.
King
,
P. D.
,
Lubeck
,
B. A.
and
Lapinski
,
P. E.
(
2013
).
Nonredundant functions for Ras GTPase-activating proteins in tissue homeostasis
.
Sci. Signal.
6
,
re1
.
Kuo
,
D. S.
,
Labelle-Dumais
,
C.
,
Mao
,
M.
,
Jeanne
,
M.
,
Kauffman
,
W. B.
,
Allen
,
J.
,
Favor
,
J.
and
Gould
,
D. B.
(
2014
).
Allelic heterogeneity contributes to variability in ocular dysgenesis, myopathy and brain malformations caused by Col4a1 and Col4a2 mutations
.
Hum. Mol. Genet.
23
,
1709
-
1722
.
Lapinski
,
P. E.
,
Bauler
,
T. J.
,
Brown
,
E. J.
,
Hughes
,
E. D.
,
Saunders
,
T. L.
and
King
,
P. D.
(
2007
).
Generation of mice with a conditional allele of the p120 Ras GTPase-activating protein
.
Genesis
45
,
762
-
767
.
Lapinski
,
P. E.
,
Kwon
,
S.
,
Lubeck
,
B. A.
,
Wilkinson
,
J. E.
,
Srinivasan
,
R. S.
,
Sevick-Muraca
,
E.
and
King
,
P. D.
(
2012
).
RASA1 maintains the lymphatic vasculature in a quiescent functional state in mice
.
J. Clin. Invest.
122
,
733
-
747
.
Lapinski
,
P. E.
,
Lubeck
,
B. A.
,
Chen
,
D.
,
Doosti
,
A.
,
Zawieja
,
S. D.
,
Davis
,
M. J.
and
King
,
P. D.
(
2017
).
RASA1 regulates the function of lymphatic vessel valves in mice
.
J. Clin. Invest.
127
,
2569
-
2585
.
Lapinski
,
P. E.
,
Doosti
,
A.
,
Salato
,
V.
,
North
,
P.
,
Burrows
,
P. E.
and
King
,
P. D.
(
2018
).
Somatic second hit mutation of RASA1 in vascular endothelial cells in capillary malformation-arteriovenous malformation
.
Eur. J. Med. Genet.
61
,
11
-
16
.
Lubeck
,
B. A.
,
Lapinski
,
P. E.
,
Bauler
,
T. J.
,
Oliver
,
J. A.
,
Hughes
,
E. D.
,
Saunders
,
T. L.
and
King
,
P. D.
(
2014
).
Blood vascular abnormalities in Rasa1(R780Q) knockin mice: implications for the pathogenesis of capillary malformation-arteriovenous malformation
.
Am. J. Pathol.
184
,
3163
-
3169
.
Macmurdo
,
C. F.
,
Wooderchak-Donahue
,
W.
,
Bayrak-Toydemir
,
P.
,
Le
,
J.
,
Wallenstein
,
M. B.
,
Milla
,
C.
,
Teng
,
J. M. C.
,
Bernstein
,
J. A.
and
Stevenson
,
D. A.
(
2016
).
RASA1 somatic mutation and variable expressivity in capillary malformation/arteriovenous malformation (CM/AVM) syndrome
.
Am. J. Med. Genet. A
170
,
1450
-
1454
.
Mao
,
M.
,
Alavi
,
M. V.
,
Labelle-Dumais
,
C.
and
Gould
,
D. B.
(
2015
).
Type IV collagens and basement membrane diseases: cell biology and pathogenic mechanisms
.
Curr. Top. Membr.
76
,
61
-
116
.
Meissner
,
M. H.
,
Eklof
,
B.
,
Smith
,
P. C.
,
Dalsing
,
M. C.
,
DePalma
,
R. G.
,
Gloviczki
,
P.
,
Moneta
,
G.
,
Neglén
,
P.
,
O'Donnell
,
T.
,
Partsch
,
H.
, et al. 
(
2007a
).
Secondary chronic venous disorders
.
J. Vasc. Surg.
46
Suppl.
,
S68
-
S83
.
Meissner
,
M. H.
,
Gloviczki
,
P.
,
Bergan
,
J.
,
Kistner
,
R. L.
,
Morrison
,
N.
,
Pannier
,
F.
,
Pappas
,
P. J.
,
Rabe
,
E.
,
Raju
,
S.
and
Villavicencio
,
J. L.
(
2007b
).
Primary chronic venous disorders
.
J. Vasc. Surg.
46
Suppl. S,
S54
-
S67
.
Munger
,
S. J.
,
Geng
,
X.
,
Srinivasan
,
R. S.
,
Witte
,
M. H.
,
Paul
,
D. L.
and
Simon
,
A. M.
(
2016
).
Segregated Foxc2, NFATc1 and Connexin expression at normal developing venous valves, and Connexin-specific differences in the valve phenotypes of Cx37, Cx43, and Cx47 knockout mice
.
Dev. Biol.
412
,
173
-
190
.
Revencu
,
N.
,
Boon
,
L. M.
,
Mulliken
,
J. B.
,
Enjolras
,
O.
,
Cordisco
,
M. R.
,
Burrows
,
P. E.
,
Clapuyt
,
P.
,
Hammer
,
F.
,
Dubois
,
J.
,
Baselga
,
E.
, et al. 
(
2008
).
Parkes Weber syndrome, vein of Galen aneurysmal malformation, and other fast-flow vascular anomalies are caused by RASA1 mutations
.
Hum. Mutat.
29
,
959
-
965
.
Revencu
,
N.
,
Boon
,
L. M.
,
Mendola
,
A.
,
Cordisco
,
M. R.
,
Dubois
,
J.
,
Clapuyt
,
P.
,
Hammer
,
F.
,
Amor
,
D. J.
,
Irvine
,
A. D.
,
Baselga
,
E.
, et al. 
(
2013
).
RASA1 mutations and associated phenotypes in 68 families with capillary malformation-arteriovenous malformation
.
Hum. Mutat.
34
,
1632
-
1641
.
Ricard-Blum
,
S.
(
2011
).
The collagen family
.
Cold Spring Harb. Perspect. Biol.
3
,
a004978
.
Rose
,
N. R.
,
McDonough
,
M. A.
,
King
,
O. N. F.
,
Kawamura
,
A.
and
Schofield
,
C. J.
(
2011
).
Inhibition of 2-oxoglutarate dependent oxygenases
.
Chem. Soc. Rev.
40
,
4364
-
4397
.
Rubinfeld
,
H.
and
Seger
,
R.
(
2005
).
The ERK cascade: a prototype of MAPK signaling
.
Mol. Biotechnol.
31
,
151
-
174
.
Sabine
,
A.
,
Agalarov
,
Y.
,
Maby-El Hajjami
,
H.
,
Jaquet
,
M.
,
Hägerling
,
R.
,
Pollmann
,
C.
,
Bebber
,
D.
,
Pfenniger
,
A.
,
Miura
,
N.
,
Dormond
,
O.
, et al. 
(
2012
).
Mechanotransduction, PROX1, and FOXC2 cooperate to control connexin37 and calcineurin during lymphatic-valve formation
.
Dev. Cell
22
,
430
-
445
.
Sevick-Muraca
,
E. M.
and
King
,
P. D.
(
2014
).
Lymphatic vessel abnormalities arising from disorders of Ras signal transduction
.
Trends Cardiovasc. Med.
24
,
121
-
127
.
Shoulders
,
M. D.
and
Raines
,
R. T.
(
2009
).
Collagen structure and stability
.
Annu. Rev. Biochem.
78
,
929
-
958
.
Srinivasan
,
R. S.
and
Oliver
,
G.
(
2011
).
Prox1 dosage controls the number of lymphatic endothelial cell progenitors and the formation of the lymphovenous valves
.
Genes Dev.
25
,
2187
-
2197
.
Wang
,
Y.
,
Nakayama
,
M.
,
Pitulescu
,
M. E.
,
Schmidt
,
T. S.
,
Bochenek
,
M. L.
,
Sakakibara
,
A.
,
Adams
,
S.
,
Davy
,
A.
,
Deutsch
,
U.
,
Lüthi
,
U.
, et al. 
(
2010
).
Ephrin-B2 controls VEGF-induced angiogenesis and lymphangiogenesis
.
Nature
465
,
483
-
486
.
Wennerberg
,
K.
,
Rossman
,
K. L.
and
Der
,
C. J.
(
2005
).
The Ras superfamily at a glance
.
J. Cell Sci.
118
,
843
-
846
.
Whitesall
,
S. E.
,
Hoff
,
J. B.
,
Vollmer
,
A. P.
and
D'Alecy
,
L. G.
(
2004
).
Comparison of simultaneous measurement of mouse systolic arterial blood pressure by radiotelemetry and tail-cuff methods
.
Am. J. Physiol. Heart Circ. Physiol.
286
,
H2408
-
H2415
.
Yamane
,
T.
,
Mitsumata
,
M.
,
Yamaguchi
,
N.
,
Nakazawa
,
T.
,
Mochizuki
,
K.
,
Kondo
,
T.
,
Kawasaki
,
T.
,
Murata
,
S.
,
Yoshida
,
Y.
and
Katoh
,
R.
(
2010
).
Laminar high shear stress up-regulates type IV collagen synthesis and down-regulates MMP-2 secretion in endothelium. A quantitative analysis
.
Cell Tissue Res.
340
,
471
-
479
.

Competing interests

The authors declare no competing or financial interests.

Supplementary information