Arpin is expressed in endothelial cells.

(A) End-point RT-PCR for ARPIN in the human microvascular endothelial cell line (HMEC-1) and human umbilical vein endothelial cells (HUVEC), 7SL was used as housekeeping gene (top); and Arpin in the mouse brain endothelial cell line bEnd.3 and primary murine lung endothelial cells (MLEC); Actb was used as a housekeeping gene (bottom). A representative gel of three independent experiments is shown.

(B) Representative western blot for arpin in HUVEC and HMEC-1. MW= molecular weight bands. The graph shows the quantification of the relative pixel intensity of the arpin band normalized to γ-tubulin as a loading control (n=4). Data are represented as mean ± standard error of the mean (SEM); ns: non-significant; two-tailed student’s t-test.

(C) Representative immunostaining of arpin and VE-Cadherin in HUVEC to analyze arpin localization at cell-cell junctions (40x objective; scale bar = 20 μm). Magnified views of boxed regions are shown (4.3 digital zoom; scale bar = 5 μm). The graphs show the dashed line scans performed with ImageJ software along the white lines in the magnified images. Profiles of the pixel intensity means of arpin and VE-Cadherin across junctions (1), along mature junctions (2) and along immature junctions (3) are depicted. 25 images each were analyzed from three independent experiments.

(D) Representative immunostaining of arpin and PECAM-1 in postcapillary venules of mouse cremaster muscles to analyze arpin localization at cell-cell junctions in vivo (40x objective; scale bar = 20 μm). Magnified views of dashed boxed regions are shown (3.3x digital zoom; scale bar = 10 μm). The graphs show the line scans performed with ImageJ software along the white lines in the magnified images. Profiles of the pixel intensity means of arpin and PECAM-1 across junctions (1), and along mature junctions (2) are depicted. 15 venules each were analyzed from four mice.

Arpin is downregulated by TNFα.

(A) qRT-PCR for the Arp2/3 complex inhibitors ARPIN, AP1AR and PICK1 using cDNA from HUVEC treated or not with TNFα for the indicated times (n=4). Data are shown as relative expression (fold change) normalized to the housekeeping gene 7SL.

(B) Representative western blot for arpin in HUVEC treated or not with TNFα for the indicated times (n=5). MW= molecular weight bands. ICAM-1 was used as a positive control of the induction of inflammation. The graph shows the mean pixel intensity of arpin bands normalized to tubulin as loading control and to the control condition (set to 1).

(C) Representative immunostaining of arpin and VE-Cadherin together with phalloidin F-actin staining in HUVEC treated or not with TNFα for 4 h (40x Objective, scale bar = 20 μm). Magnifications on the right show the loss of arpin at junctions after TNFα treatment (orange arrows). The graphs show arpin pixel intensity quantification after treatment normalized to the average of ctrl HUVECs: top-left, total arpin in 17 images each was analyzed from four independent experiments; top-right, junctional and actin cytoskeletal arpin in 19 images each were analyzed from three independent experiments; and bottom graph: Pearson’s correlation analysis showing that arpin is inversely correlated to central actin fiber density (Pearson’s correlation coefficient, r; 49 cells were analyzed from three independent experiments).

(D) Representative immunostaining for arpin (blue) and PECAM-1 (red) in postcapillary venules treated or not with TNFα at the indicated times. PECAM-1 and MRP14 (neutrophils) stainings (right) are shown as positive control for the induction of inflammation (40x objective, scale bar = 20 μm). The top graph shows total arpin pixel intensity quantification after TNFα treatment normalized to the average of control venules (12-14 venules were analyzed from four mice in each group). The bottom graph shows junctional and non-junctional arpin quantification after TNFα treatment normalized to the average of control venules (12-14 venules were analyzed from four mice in each group).

All data are represented as mean ± SEM; *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001; two-tailed student’s t-test.

Arpin depletion induces hyperpermeability and irregular junction patterns.

(A) Representative Western blots for WAVE2, WASP, the Arp2/3 subunit ArpC5 and arpin in lysates of control (shCtrl) and arpin-depleted (shArpin) HUVEC. The graph shows the mean pixel intensities of the protein bands normalized to the respective control bands. All bands were normalized to γ-tubulin as loading control (n=3).

(B) Paracellular flux of 150 kDa FITC-dextran across confluent control and arpin-depleted HUVEC monolayers cultured on transwell filters (0.4μm pore size) untreated or treated with TNFα for 18 h. Data are represented as relative permeability normalized to control HUVEC set to 1 (n=9 in three independent experiments).

(C) Representative Western blots for VE-Cadherin, β-catenin, vinculin, and arpin in lysates of control and arpin-depleted HUVEC. The graph shows the mean pixel intensities of the protein bands normalized to the respective control bands (set to 1). All bands were normalized to γ-tubulin as a loading control (n=3).

(D) Representative Western blots for ZO-1, claudin-5 and arpin in lysates of control and arpin-depleted HUVEC. γ-Tubulin was used as a loading control. The graph shows the quantification of the relative pixel intensity of ZO-1 and claudin-5 bands normalized to the untreated control (set to 1) and γ-tubulin (n=3).

(E) Immunostaining of VE-Cadherin and β-catenin in control and arpin-depleted HUVEC (40x objective, scale bar = 20 μm; dashed boxes indicate magnified areas that highlight changes in junctional architecture; 3.3 digital zoom, scale bars = 5 μm). Representative images of four independent experiments are shown. Arrows depict linear and mature junctions in shctrl HUVEC and interrupted and gap junction formation and junction internalization in shArpin HUVEC.

All data are represented as mean ± SEM; ns: non-significant; *p<0.05; ***p<0.001; ****p<0.0001; two-tailed student’s t-test.

Arpin depletion induces actin filament formation.

(A) Representative F-actin staining using phalloidin of control and arpin-depleted HUVEC (40x objective, scale bar = 20μm). The graph shows phalloidin pixel intensity quantification normalized to the average of control HUVEC. 18 images were analyzed in each group from four independent experiments.

(B) Representative immunostaining of VE-Cadherin with F-actin staining using phalloidin (top) in control and arpin-depleted HUVEC. The method of central actin fiber quantification using the phalloidin signal is depicted on the bottom as reported before (Garcia Ponce et al. 2016) (63x objective, scale bar = 5 μm). Yellow dashed lines delineate an individual cell. The graph shows the mean central actin fiber density quantification along the orange lines. 109 cells were analyzed in each group from 3 independent experiments.

(C) Representative Western blot for total actin in lysates of control and arpin-depleted HUVEC. The graph shows the mean pixel intensity of the actin bands normalized to the control and to γ-tubulin as a loading control (n=4).

(D) Representative Western blots for cortactin, coronin1B, p-cofilin, cofilin and arpin in lysates of control and arpin-depleted HUVEC. The graph shows the mean pixel intensity of the protein bands normalized to the respective control bands (set to 1). All bands were normalized to γ-tubulin as a loading control (n=4).

All data are represented as mean ± SEM; ns: non-significant; **p<0.01; ****p<0.0001; two-tailed student’s t-test.

Inhibition of Arp2/3 complex in arpin-depleted HUVEC does not rescue hyperpermeability and actin cytoskeleton alterations.

(A) Immunostaining of VE-Cadherin with F-actin staining using phalloidin in control and arpin-depleted HUVEC treated with 100 μM of the Arp2/3 inhibitor CK-666 or vehicle (40x objective, scale bar = 20 μm). Representative images of three independent experiments are shown.

(B) Total phalloidin pixel intensity quantification normalized to the average of control HUVEC. 9 images were analyzed in each group from three independent experiments.

(C) Central actin fiber density quantification done as described in figure 4B. 100 cells were analyzed in each group from three independent experiments.

(D) Paracellular flux assays using confluent control and arpin-depleted HUVEC monolayers on 0.4 µm pore transwell filters treated with 100 μM CK-666 or vehicle (n=7-9 from three independent experiments). Data are normalized to control HUVEC treated with vehicle set to 1.

All data are represented as mean ± SEM; *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001; two-tailed student’s t-test.

Arpin depletion increases the formation of actin stress fibers.

(A) Representative Western blots for pMLC (Ser19), MLC and arpin in lysates of control and arpin-depleted HUVEC. The graph shows the mean pixel intensity of the protein bands normalized to the respective control band (set to 1). All bands were normalized to γ-tubulin as a loading control (n=4).

(B) Representative Western blots for pMYPT1 (Thr696), pMYPT1 (Thr853), MYPT1, and arpin in lysates of control and arpin-depleted HUVEC. The graph shows the mean pixel intensity of the protein bands normalized to the respective control band (set to 1). All bands were normalized to γ-tubulin as a loading control (n=4).

(C) Representative Western blots for ROCK1, mDia, ZIPK, and arpin in lysates of control and arpin-depleted HUVEC. The graph shows the mean pixel intensity of the protein bands normalized to the respective control band (set to 1). All bands were normalized to γ-tubulin as a loading control (n=4).

All data are represented as mean ± SEM; ns: non-significant; *p<0.05; **p<0.01; ***p<0.001; two-tailed student’s t-test.

ROCK1/2 and ZIPK inhibition partially rescues increase in stress fibers and permeability in arpin-depleted HUVEC.

(A) Representative immunostaining of VE-Cadherin with F-actin staining using phalloidin in control and arpin-depleted HUVEC treated or not with 10 μM of the ROCK1/2 inhibitor Y27632 (40x objective, scale bar = 20μm).

(B) Phalloidin pixel intensity quantification normalized to the average of control HUVEC. 14 images were analyzed in each group from three independent experiments.

(C) Central actin fiber density quantification of at least 140 cells in each group from three independent experiments.

(D) Paracellular flux assays using confluent control and arpin-depleted HUVEC monolayers on 0.4 µm pore transwell filters treated with 10 μM Y27632 or vehicle (n=10-12 from four independent experiments).

(E) Representative immunostaining of VE-Cadherin with F-actin staining using phalloidin in control and arpin-depleted HUVEC treated with 10 μM of the ZIPK inhibitor HS38 or vehicle (40x objective; scale bar = 20μm).

(F) Phalloidin pixel intensity quantification normalized to the average of control HUVEC treated with the vehicle. 13 images were analyzed in each group from three independent experiments.

(G) Central actin fiber density quantification of at least 125 cells in each group from three independent experiments.

(H-I) Paracellular flux assays using confluent control and arpin-depleted HUVEC monolayers on 0.4 µm pore transwell filters treated with (H) 10 μM HS38 or vehicle (n=9-12 in four independent experiments) or (I) treated or not with 15 ng/mL TNFα and with 10 μM HS38 or vehicle (7-8 in three independent experiments).

All immunofluorescences are representative of three independent experiments. All data are represented as mean ± SEM; ns: non-significant; *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001; two-tailed student’s t-test.

Arpin-deficient mice are viable but show increased vascular permeability in the lung and skin.

(A) Arpin mouse gene showing exon 3, which was deleted using the CRISP/Cas9 technology to generate the complete arpin-/- mouse model.

(B) Representative genotypification PCR of mice carrying the WT (arpin+/+), heterozygous (arpin+/-), and deleted (KO, arpin -/-) alleles.

(C) Representative Western blots for arpin in lysates of the indicated organ tissues from arpin+/+ and arpin-/- mice (n=3).

(D) Permeability assays in the lungs. 150 kDa FITC-dextran was injected into arpin+/+ and arpin-/- mice via the cannulated artery carotid. Animals were sacrificed, perfused, and lungs collected and homogenized in PBS. The homogenized tissue containing the leaked FITC-dextran was quantified using a fluorometer. (5 arpin+/+ and 8 arpin-/- mice were analyzed). Data are represented as mean ± SEM; *p<0.05; two-tailed student’s t-test with Welch’s correction.

(E) Miles assays to determine vascular permeability in the back skin of arpin+/+ and arpin-/- mice under basal conditions and after treatment with histamine, 10 μM ROCK1/2 inhibitor Y27632, 10 μM ZIPK inhibitor HS38, 100 μM Arp2/3 complex inhibitor CK-666, or PBS only as control. Data are shown as relative permeability normalized to the untreated arpin+/+ group set to 1 (n = at least 7 mice in each group), and represented as means ± SEM. **p<0.01; ***p<0.001; ****p<0.0001; two-tailed student’s t-test.

(F) Representative images of the hematoxylin and eosin staining of lung tissues of arpin+/+ (images 1 and 2) and arpin-/- (images 3-6) mice. Image 1 shows normal histology (10x objective). Image 2 shows normal structure of alveoli (*) and the blood vessels (arrowheads) without any pathology (40x objective, images 2-6). Image 3 shows some areas with alveolar volume reduced (*) in arpin-/- mice. Image 4 shows discrete interalveolar hemorrhages (arrow heads). Image 5 shows congestion and dilatation of the capillaries (arrow heads). Image 6 shows interalveolar edema (*) and microhemorrhage (arrowhead). The histological score of the lung tissue is shown in the graph below. A score of 0 indicates no inflammation; 1 low inflammation; 2 moderate inflammation; 3 high inflammation (9 images from arpin+/+ and 12 images from arpin-/- mice were analyzed). Data are represented as mean ± SEM; *p<0.05; two-tailed student’s t-test.

(G) Representative immunostaining for PECAM-1 and F-actin staining using phalloidin in cryosections from lungs of arpin+/+ and arpin-/- mice. Images in the bottom shown endothelial cell (EC) F-actin extracted using PECAM-1 as a templete of ECs using Imaris software (40x objective, scale bar = 50μm). Graph shows mean F-actin pixel intensity quantification in arpin-/- lungs normalized to the average of images from arpin+/+ lungs (20 images were analyzed from three mice in each group). Data are represented as mean ± SEM; **p<0.01; two-tailed student’s t-test.

Current working model for arpin functions in ECs.

Under basal conditions arpin is located throughout the cell and enriched at cellular junctions. During inflammation, arpin is downregulated causing the formation of actomyosin actin stress fibers, junction disruption, and increased permeability. Question marks indicate the hitherto unknown mechanism of how arpin controls the activity of ROCK and ZIPK to induce the formation of contractile actin stress fibers.

Arpin is reduced by IL-1β.

(A) Representative Western blot for PICK1 in HUVEC treated or not with TNFα for the indicated times (n=3). The graph shows the mean pixel intensity of PICK1 bands normalized to tubulin as loading control and to the control condition (set to 1).

(B) Representative Western blot for arpin in HUVEC treated or not with IL-1β for the indicated times. ICAM-1 was used as a positive control of the induction of inflammation. The graph shows the quantification of the relative pixel intensity of arpin bands normalized to the untreated control (set to 1) and γ-tubulin as a loading control (n=4).

(C) Representative Western blot for PICK1 in HUVEC treated or not with IL-1 β for the indicated times (n=3). The graph shows the mean pixel intensity of PICK1 bands normalized to tubulin as loading control and to the control condition (set to 1).

(D) Representative immunostaining of VE-Cadherin and arpin with F-actin staining using phalloidin in HUVEC treated or not with IL-1β (Scale bar = 20μm). The upper left graph shows total arpin pixel intensity quantification normalized to the average of control HUVEC (top-left; n=11 in three independent experiments). The upper right graph shows junctional and actin cytoskeletal arpin quantification in control and IL-1β treatment HUVEC (n=11-15 in each group in three independent experiments). Tho graph in the bottom shows that arpin is inversely correlated to central actin fiber density (Pearson’s correlation coefficient, r, n=45 cells).

All data are represented as mean ± SEM; *p<0.05; **p<0.01; ***p<0.001; two-tailed student’s t-test.

Arpin does not participate in junction assembly.

(A) Representative immunostaining of VE-Cadherin in control and arpin-depleted HUVEC in calcium-switch assays over time (63x objective; scale Bar = 20 μm). The graph shows the quantification of junctional length, calculated in imageJ software, per number of cells (n=3).

All data are represented as mean ± SEM; *p<0.05; **p<0.01; ****p<0.0001; two-way ANOVA.

Arpin deficiency increases vascular permeability in skin and alters venular junction architecture in cremaster muscle.

(A) The graph shows the average weight of arpin+/+ and arpin -/- mice over time (at least 3 arpin+/+ and arpin-/- mice were weighted at each time point).

(B) Representative Western blots for the indicated proteins in lysates of lung tissue from arpin+/+ and arpin-/- mice. The graph shows the quantification of the relative pixel intensity of the protein bands normalized to γ-tubulin as a loading control (n=3).

(C) Representative immunostaining for arpin and VE-Cadherin in postcapillary venules of arpin+/+ and arpin-/- mice. (n=3; dashed boxes indicate magnified areas, 40x objective; scale Bar = 20 μm). Arrows in the magnifications to the right show the altered junction architecture in the absence of arpin (4x digital zoom; scale Bar = 5 μm).

(D) Representative immunostaining for ZO-1 in postcapillary venules of arpin+/+ and arpin -/- mice. (n=3; Dashed box indicate delineate magnified area, 40x objective; Scale Bar = 20μm), images from the right show junction architecture (4x digital zoom; Scale Bar = 10μm).

Mating statistics of arpin+/- mice

Genotyping of 126 littermates from breedings of heterozygous (arpin+/-) mice are shown. χ2 test probability is shown. Deviations from Mendelian rules are non-significant

Hemograms of arpin+/+ and arpin-/- mice.