Skip to main content
. 2020 Sep 8;9:e57732. doi: 10.7554/eLife.57732

Figure 4. Loss of junctional CLDN5 occurs after EphrinB2 blockade in vitro but does not lead to breakdown of LEC junctions in vivo.

(A) Top: Western blot of EphrinB2-Fc-precipitated EphB4 HDLECs for phosphotyrosine (4G10) and EphB4. Cells were treated with control (siCTRL) or EphrinB2 siRNA (siEFNB2), and stimulated with pre-clustered EphrinB2-Fc or control Fc. Total cell lysates (TCL) were immunoblotted for EphB4 and actin. Bottom: Quantification of relative EphB4 phosphorylation (adjusted to total precipitated EphB4 protein) from four independent experiments, showing EphrinB2-dependent basal EphB4 activity in unstimulated LECs. (B) Immunofluorescence of HDLECs treated with EphrinB2 blocking antibody (B11) for 3 hr using antibodies against VE-cadherin (green) and CLDN5 (red). Single channel images are depicted in grey. EphrinB2 blockade induces junction remodeling and removal of junctional CLDN5. (C) Quantification of HDLEC monolayer permeability to 40 kDa FITC-dextran showing increase upon B11 treatment for 3 hr (n = 10 replicates per condition from three independent experiments). (D) Quantification of junctional CLDN5 immunofluorescence in VE-cadherin+ junctions (n = 6–7 images from three independent experiments). (E) Immunofluorescence of HDLECs transfected with CTRL, CLDN5 or CDH5 (VE-cadherin) siRNA for 48 hr. Staining for CLDN5 (green) and VE-cadherin (grey) show disruption of VE-cadherin+ junctions upon CLDN5 siRNA transfection. Note, CLDN5+ junctions are not affected upon VE-cadherin silencing. (F) Quantification of HDLEC monolayer permeability to 40 kDa FITC-dextran showing minor and moderate increase upon CDH5 or CLDN5 silencing, respectively (n = 7–8 replicates per condition from two independent experiments). (G) Whole-mount immunofluorescence of mesenteric collecting vessels and valves in P11 Cldn5flox;Prox1-CreERT2 mice using antibodies against CLDN5 and VE-cadherin. Note efficient CLDN5 depletion apart from a few CLDN5 hot spots and altered morphology but no disintegration of the junctions. Data in A represent mean ± s.e.m. p value, One-sample t-test. Data in C, D, F represent mean ± s.e.m. p value, Two-tailed unpaired Student’s t-test. Source data for panels (A,C,D,F) are provided. Scale bars: 100 μm (B, E, G mesenteric valves), 50 μm (G mesenteric collecting vessel).

Figure 4—source data 1. Quantification of the effects of EphrinB2 inhibition on EphB4 phosphorylation, endothelial monolayer permeability and junctional CLDN5.

Figure 4.

Figure 4—figure supplement 1. Loss of junctional CLDN5 after long-term EphrinB2 inhibition in primary LECs.

Figure 4—figure supplement 1.

Immunofluorescence of HDLECs treated with EphrinB2 blocking antibody (B11) for 16 hr using antibodies against VE-cadherin (red) and CLDN5 (green). Single channel images are depicted in grey. Note almost complete removal of junctional CLDN5 and disruption of VE-cadherin+ junctions (arrows) in B11-treated cells. Scale bar: 50 µm.
Figure 4—figure supplement 2. The effect EFNB2 silencing on CLDN5 and VE-cadherin.

Figure 4—figure supplement 2.

(A) Western blot analysis of total cell lysates (TCL) of HDLECs treated for 48 hr with control (siCTRL) or EFNB2 siRNA (siEFNB2). Note reduced CLDN5 but unchanged VE-cadherin levels in EFNB2-silenced cells. Actin was used as a loading control. Western blot is shown as representative of 3 independent experiments. (B) Immunofluorescence of HDLECs transfected with CTRL or EFNB2 siRNA for 48 hr. Staining for CLDN5 and VE-cadherin shows removal of junctional CLDN5 and disruption of VE-cadherin+ junctions (arrows) upon EFNB2 silencing. Scale bar: 50 μm.
Figure 4—figure supplement 3. Silencing of CDH5 and CLDN5 in HDLECs.

Figure 4—figure supplement 3.

Western blot analysis of total cell lysates (TCL) of HDLECs treated for 72 hr with control (siCTRL), CDH5 siRNA (siCDH5) and CLDN5 siRNA (siCLDN5). Western blot is shown as representative of 2 independent experiments.
Figure 4—figure supplement 4. Generation of conditional Cldn5 knock-out mice.

Figure 4—figure supplement 4.

(A) Schematic of the Cldn5 wild type allele, targeted ‘Knockout-First’ allele, conditional allele (floxed) and deletion allele. The ‘Knockout-First’ allele Cldn5tm1a(EUCOMM)Wtsi contains an IRES:lacZ trapping cassette and a floxed promoter-driven neo cassette inserted into the (single) coding exon of the Cldn5 gene, disrupting Cldn5 gene function. Flp convert the ‘Knockout-First’ allele to a conditional allele. Cre deletes the floxed 3’ sequence of the coding exon generating a null allele. The locations of PCR genotyping primers are indicated in light blue. (B) PCR genotyping of wild type and conditional floxed Cldn5 mice. (C) Whole-mount immunofluorescence of mesenteries of E18.5 Cldn5flox/flox;PGK-Cre (i.e. germline Cldn5 homozygous) and Cre negative littermate embryos showing lack of CLDN5 expression in the mutant. Scale bar: 100 μm.