p120-Catenin prevents neutrophil transmigration independently of RhoA inhibition by impairing Src dependent VE-cadherin phosphorylation

Pilar Alcaide, Roberta Martinelli, Gail Newton, Marcie R. Williams, Alejandro Adam, Peter A. Vincent, Francis W. Luscinskas


Leukocyte transendothelial migration (TEM) is regulated by several signaling pathways including Src family kinases (SFK) and the small RhoGTPases. Previous studies have shown that vascular endothelial-cadherin (VE-cad) forms a complex with β-,γ-, and p120-catenins and this complex disassociates to form a transient gap during leukocyte TEM. Additionally, p120-catenin (p120-1A) overexpression in human umbilical vein endothelial cells (HUVEC) stabilizes VE-cad surface expression, prevents tyrosine phosphorylation of VE-cad, and inhibits leukocyte TEM. Based on reports showing that p120 overexpression in fibroblasts or epithelial cells inhibits RhoA and activates Rac and Cdc42 GTPases, and on other reports showing that RhoA activation in endothelial cells is necessary for leukocyte TEM, we reasoned that p120 overexpression inhibited TEM through inhibition of RhoA. To test this idea, we overexpressed a mutant p120 isoform, p120-4A, which does not interact with RhoA. p120-4A colocalized with VE-cad in HUVEC junctions and enhanced VE-cad surface expression, similar to overexpression of p120-1A. Interestingly, overexpression of either p120-4A or p120-1A dramatically blocked TEM, and overexpression of p120-1A in HUVEC did not affect RhoA basal activity or activation of RhoA and Rac induced by thrombin or ICAM-1 crosslinking. In contrast, biochemical studies revealed that overexpression of p120-1A reduced activated pY416-Src association with VE-cad. In summary, p120 overexpression inhibits neutrophil TEM independently of an effect on RhoA or Rac and instead blocks TEM by preventing VE-cad tyrosine phosphorylation and association of active Src with the VE-cad complex.

  • inflammation
  • leukocyte recruitment
  • adherens junctions
  • host defense
  • vascular endothelial-cadherin

transendothelial cell migration (TEM) of leukocytes is an essential component of the inflammatory response. The endothelial molecular mechanisms regulating TEM, however, remain incompletely understood. Vascular endothelial-cadherin (VE-cad) is a transmembrane protein expressed in vascular endothelium at cell-cell borders where it forms a complex with β-catenin, γ-catenin (plakoglobin), and p120-catenin (p120) and interacts with the actin cytoskeleton through an as yet incompletely understood mechanism involving α-catenin (reviewed in Ref. 41). Previous studies (16) have shown VE-cad plays an essential role in vascular morphogenesis, endothelial barrier function, proliferation and survival, and cell alignment in response to fluid shear stress.

The current dogma is that the VE-cad complex acts as the gatekeeper during leukocyte transmigration based on in vivo (32) and in vitro (4, 7, 8, 33, 38) models of inflammation. Initial reports (4, 8, 30, 33, 36) showed that adherent leukocytes caused a transient gap in VE-cad at cell-cell borders through which leukocytes then migrated and the gaps resealed. The mechanisms that control VE-cad dynamic behavior during TEM have not been identified, but a previous study (29) suggest that phosphorylation of tyrosine 658 and 731 in the cytoplasmic tail of VE-cad limits its association with p120-catenin and β-catenin, respectively, and thus formation of the VE-cad complex. Based on these data, VE-cad tyrosine phosphorylation was envisioned to cause the VE-cad complex to dissociate and uncouple from the cytoskeleton (29). p120-catenin (p120-1A) is a major substrate of Src family kinases (23) and is a cytosolic scaffold protein that binds to cytoplasmic domains of the cadherin protein family (31). p120 regulates the cycle of adherens junction formation and disassembly in endothelium and epithelium (reviewed in Ref. 5).

In a previous study, we (4) examined the hypothesis that leukocyte TEM at endothelial cell junctions, termed parcellular TEM, involved rapid endocytosis of VE-cad to form transient gaps through which leukocytes then migrated. The experimental approach was to stabilize VE-cad surface expression at junctions by overexpression of p120-GFP fusion protein. Surprisingly, we found that VE-cad endocytosis was not involved in gap formation during TEM. Instead, tyrosine phosphorylation of the VE-cad cytoplasmic tail was a critical factor to initiate VE-cad gap formation, and p120 overexpression dramatically reduced VE-cad phosphorylation and VE-cad protein turnover under resting conditions and also prevented ICAM-1 crosslinking-induced phosphorylation of tyrosine 658 in the VE-cad cytoplasmic tail and blocked TEM. In a murine model of inflammation overexpression of p120 in the lung attenuated neutrophil infiltration, and loss of barrier function induced by Escherichia coli lipopolysaccharide (LPS) instillation (42). In addition, nonphosphorylatable VE-cad mutants of Y658E or Y731E prevented neutrophil TEM (4, 7). Transgenic mice in which VE-cad was replaced by a VE-cad-α-catenin fusion protein, thereby stabilizing adherens junctions in vivo, were completely resistant to VEGF and histamine-induced permeability increase, and neutrophil or lymphocyte recruitment into inflamed microvessels in cremaster, lung, and skin was reduced (32). However, studies (10, 26) using NIH 3T3, an adenoma cell line, or adipocytes indicated that p120 overexpression inhibited RhoA while activating the related GTPases Rac1 and Cdc42 and resulted in impaired cytoskeletal functions. Overexpression of p120 in bovine aortic endothelial cells resulted in inhibition of RhoA and correlated with a reduction in barrier function and a dramatic shape change to a dendritic-like morphology (22). Both RhoA and Rac are involved in leukocyte TEM (reviewed in Refs. 3, 14, 39). Inhibition of endothelial RhoA reduced T-cell and monocyte-like cell U937 TEM across endothelium (2, 40). Based on these observations, p120 could play two different roles in endothelium during leukocyte TEM: p120-1A overexpression may prevent tyrosine phosphorylation of VE-cad through inhibition of Src kinase activity, or it can inhibit small RhoA GTPases, or both.

In this study, we compared the effects of two p120 isoforms in an in vitro TEM assay, the endogenous p120 isoform (p120-1A) and a mutant p120 isoform (p120-4A), which lacks both domains required to bind RhoA (a NH2-terminal coiled-coil domain and lysines-622, 623 to alanine substitutions in the regulatory domain; Refs. 10, 13, 19). Our results show that inhibition of RhoA or overexpression of p120-1A or p120-4A prevented leukocyte TEM and that overexpression of either isoform of p120 had a greater inhibitory effect on leukocyte TEM than RhoA inhibition. Furthermore, transduction of p120-1A in human umbilical vein endothelial cells (HUVEC) did not alter basal or thrombin-induced RhoA activity, and ICAM-1 crosslinking induced activation of RhoA or Rac. Instead, our data indicate that overexpression of p120 inhibits the association of activated pY416-Src with VE-cad and partially reduced expression of the phosphoinositide 3-kinase (PI3K) p110-α isoform, which has been implicated in leukocyte TEM (12), without affecting RhoA or Rac activation.



Human recombinant TNF-α was purchased from PeproTech (Rocky Hill, NJ), and hec-1 [nonblocking MAb to human VE-cad (6) a gift from Dr. William Muller, Weill Medical College, Cornell University, NY] was purified IgG and was conjugated to Alexa 568 (Molecular Probes, Eugene, OR). Antibodies (all as purified IgG) were as follows: p120 (BD Biosciences, San Jose, CA); GFP (Abcam, Cambridge, MA); α-catenin and ZO-1 (Zymed, San Francisco, CA); β-catenin (RDI, Flanders, NJ); ICAM-1 and Hu5/3 (4); major histocompatibility complex-Class I (MHC-I), W6/32 (4) JAM-A, 1H2A9 (34); phospho-VE-cad-Tyr658 (Biosource, Camarillo, CA); anti-β-actin, α-thrombin, and PP2, a Src inhibitor (Sigma Aldrich, St. Louis, MO); P-Src-416 and Src (Cell signaling, Danvers, MA) and p110α (Millipore, Billerica, MA); and phalloidin-Alexa Fluor 546 (Invitrogen, Carlsbad, CA). RhoA and Rac G-lisas, Rotekin RBD beads, Abs to RhoA, and C3Cb peptide were from Cytoskeleton (Denver, CO).


HUVEC (subculture 2) were seeded on glass coverslips precoated overnight with fibronectin (5 μg/ml; BD Biosciences, San Jose, CA). Human polymorphonuclear neutrophils (PMNs; >95% pure) were isolated from whole blood drawn from healthy volunteers, kept at 8°C and used immediately. Blood was drawn and handled according to protocols for protection of human subjects approved by the Brigham and Women's Hospital Institutional Review Board, and all volunteer subjects gave written informed consent.

Adenovirus production and cell infection.

Adenovirus encoding p120-1A GFP, p120-4AK GFP or GFP alone have been described previously (19). HUVEC were plated to be at 80–90% confluence at the time of infection, infected 24 h after plating with varying doses of Adv vector, and cultured for another day. Analysis by flow cytometry measured the GFP fluorescence. Virus infection at optimized concentrations did not affect endothelial activation as determined by induction of endothelial cell adhesion molecule expression (Fig. 2C).

PMN transmigration assay under shear flow conditions.

The live cell fluorescence microscopy flow model has been described previously (33). HUVEC monolayers on glass coverslips were activated with TNF-α (10 ng/ml, 4 h) and pretreated with C3Cb (2 h, 1 μg/ml) or with PP2 (10 μM, 30 min, 37°C) where indicated, and VE-cad was immunolabeled with fluorescent Hec-1-Alexa568 MAb (10 min, 1 μg/ml). The coverslips were inserted into the flow chamber, and PMN (1 × 106/ml) were drawn across HUVEC at 1.0 dyn/cm2 for 10 min to allow for adhesion and transmigration to occur.

Image acquisition and analysis.

Live cell imaging of leukocyte TEM was performed using a digital imaging system coupled to a Nikon TE2000 inverted microscope as detailed previously (34). The number of accumulated leukocytes was determined by counting the total number of adhered and transmigrated cells per field (differential interference contrast 20×/0.75 NA objective). The percent TEM was calculated as (total transmigrated leukocytes) ÷ (total adhered + transmigrated leukocytes) × 100.

Quantitation of VE-cad fluorescence in endothelium.

The analysis by fluorescence microscopy and NIH ImageJ software was as described previously (4). Briefly, VE-cad was quantified in 30 different cell-cell junctions for each condition in images corresponding to three different fields of view using a differential interference contrast 20×/0.75 NA objective. Values were normalized by dividing each by the value corresponding to VE-cad (red) intensity in uninfected cells.

Immunoprecipitation and Western blotting.

Transduced or control HUVEC were surface biotinylated (biotinylation kit; Amersham, Arlington Heights, IL) and cultured for the indicated times or lysed directly without biotinylation as previously described (4). Equal aliquots of lysate were immunoprecipitated with the indicated MAbs, and samples resolved on 10% SDS-PAGE, transferred to nitrocellulose membranes, and probed with streptavidin-horseradish peroxidase or with the appropriate MAb in the case of nonbiotinylated HUVEC.

GTPase pull downs and G-lisa studies.

G-lisa kits and Rotekin RBD bead pull downs were used following the manufacturer's instructions (Cytoskeleton).

Electrical cell-substrate impedance sensor.

HUVEC monolayer permeability was determined by measuring changes in electrical resistance [electrical cell-substrate impedance sensor (ECIS), Applied BioPhysics; Ref. 1]. Briefly, HUVEC (80,000 cells) were seeded onto ECIS 10E cultureware (0.8 cm2/well) precoated with 0.1% gelatin and incubated for 24 h before infecting with adenoviruses. The electrical impedance across the monolayer was measured at 1 V, 4,000 Hz with current flowing through 10 small gold electrodes per well plus 1 large counter electrode, using the culture medium as the source of electrolytes. Impedance was monitored for 48 h after adenovirus infection by the lock-in amplifier, stored, and then used to calculate resistance and capacitance by the manufacturer's software. Data are presented as a plot of total resistance vs. time.

Statistical analysis.

All data are presented as means ± SD for n = 3 or more unless otherwise indicated and student's t-test was used to determine significance (P values < 0.05, <0.01, <0.001) unless ANOVA analysis is indicated.


Inhibition of RhoA in TNF-α -activated HUVEC results in decreased neutrophil TEM.

Robust overexpression of p120 has been demonstrated to result in changes in the activity of RhoA in a variety of epithelial cells (10, 26), and in bovine endothelial cells it impairs barrier function (22), and in human endothelial cells inhibits leukocyte TEM. RhoA is involved in cytoskeleton remodeling, which occurs during leukocyte TEM. To evaluate the role of RhoA in leukocyte TEM, we performed 10-min transmigration assays under defined shear flow conditions using TNF-α-activated HUVEC monolayers pretreated with C3TCb, a RhoA cell permeable inhibitor peptide. Inhibition of RhoA resulted in a significant inhibition of TEM without affecting PMN adhesion (Fig. 1, A and B) or altering endothelial cell monolayer morphology. Overexpression of p120 dramatically reduced TEM but did not affect PMN adhesion, consistent with our previous findings (4). These data indicate that RhoA activity in endothelium is required for leukocyte TEM and that overexpression of p120 can inhibit TEM.

Fig. 1.

Inhibition of RhoA in human umbilical vein endothelial cells (HUVEC) inhibits polymorphonuclear neutrophils (PMNs) transmigration. HUVEC were transduced with p120GFP-Adv or treated with the RhoA inhibitor C3TCb peptide (1 μg/ml, 2 h) and stimulated with TNF-α (10 ng/ml, 4 h). Transendothelial migration (TEM; A) and PMN accumulation (B) were determined in each category and are shown as the relative %control. Data are pooled from 3 different experiments and represent means ± SD. P values are indicated, comparing each bar with control HUVEC. **P < 0.01, ***P < 0.001 by Student's t-test.

Characterization of p120-4A GFP expression in HUVEC.

To determine whether p120 regulation of RhoA occurs in human endothelium, we expressed a p120 mutant isoform recombinant construct, p120 4AK-GFP (p120-4AGFP), that prevents binding to and inhibition of RhoA (10, 13, 19). p120-4AGFP contains an alanine substitution of two lysines (K622,623) and a NH2-terminal truncation. We characterized p120-4AGFP expression and its effects on the VE-cad complex in HUVEC by epifluorescence microscopy. p120-4AGFP localized to cell-cell junctions, thus mimicking the localization of endogenous p120 (Fig. 2A). Both endogenous isoforms of p120 catenin (p120/p100) and p120-4AGFP are expressed in HUVEC and associate with VE-cad by coimmunoprecipitation analysis (Fig. 2B). Overexpression of p120-4AGFP did not alter the surface expression of the TNF-α-induced adhesion molecules E-selectin, ICAM-1, or VCAM-1 (Fig. 2C). These findings indicate that p120-4AGFP mimics the localization of wild-type (p120-1AGFP) and endogenous p120 in HUVEC monolayers.

Fig. 2.

Characterization of p120-4A overexpression in HUVEC. Confluent HUVEC were infected with GFP, p120-1AGFP, or p120-4AGFP adenoviruses as described in materials and methods. A: p120-4AGFP colocalizes with endogenous p120 at cell junctions. Monolayers were fixed with 10% buffered formalin, permeabilized, and stained with anti-p120 MAb. Representative fields were examined by epifluorescence microscopy and show junctional distribution of p120-4AGFP and colocalization with endogenous p120 at cell junctions. B: vascular endothelial-cadherin (VE-cad) was immunoprecipitated from the HUVEC lysates and the material was immunoblotted for p120 to show endogenous isoforms of p120/p100 and p120-GFP exogenous isoforms (identified by asterisk) or with an anti-GFP MAb to detect expression of p120-GFP isoforms. Representative blot from 3 different experiments is shown. C: HUVEC were transduced with p120-4AGFP, treated with TNF-α during 4 h, stained with the indicated MAb, and analyzed by flow cytometry. Data represent 3–5 independent experiments.

p120-4A-GFP overexpression increases VE-cad at the cell-cell junctions and enhances its surface expression.

Because p120 selectively controls VE-cad turnover at cell-cell borders in endothelial cells (43), we evaluated if this function by p120 was dependent on interactions with RhoA by expressing the p120-4AGFP mutant. Overexpression of p120-4AGFP resulted in increased levels of expression of VE-cad at cell junctions (Fig. 3, A and B) similar to previous findings in human dermal microvascular endothelial cells (HDMVECs) (22). With regards to turnover, the reduction of -surface biotinylated VE-cad in GFP-transfected HUVEC cells was ∼65% after 6 h whereas little change VE-cad surface expression was detected in either p120-1AGFP- or p120-4AGFP-transduced HUVEC after 6 h, suggesting that the sustained expression of VE-cad at the cell surface induced by p120 overexpression is independent of RhoA activity (Fig. 3C). The optimal concentration of each virus led to an infection efficiency of 80–90% in monolayers, and the levels of other molecules that localize to endothelial cell junctions, such as JAM-A, ZO-1, and α- and β-catenins, were not affected (data not shown), indicating that the mutant isoform p120-4AGFP can mimic the functions of WT p120-1AGFP. RhoA interactions with p120, therefore, are not involved in regulation of VE-cad surface expression. We also determined the effect of the optimal concentration of both p120-GFP constructs in HUVEC permeability. Both p120-1AGFP and p120-4AGFP decreased monolayer permeability as measured by an increase in endothelial monolayer transelectrical resistance (Fig. 3D), indicating that the enhancement of transelectrical resistance observed in HUVEC after overexpression of p120 is also independent of any effect on RhoA activity, as previously described for human dermal microvascular endothelial cells (19).

Fig. 3.

p120-4A overexpression increases VE-cad at the cell-cell junctions. HUVEC were infected with GFP, p120-4AGFP, or p120-1AGFP adenoviruses. A and B: Hec-1-Alexa 568 MAb was used to detect VE-cad in each monolayer. Intensity of VE-cad at the junctions was quantified in live HUVEC by analyzing 30 separate cell-cell junctions including representative junctions of the heterogeneous population for each condition (in duplicate) for each experiment performed. C: HUVEC transduced with the indicated Adv were surface biotinylated and lysed at time 0 or cultured for 2, 4, or 6 h before lysis and immunoprecipitation with anti-VE-cad MAb Hec-1 and Western blotted with streptavidin. A representative blot from two studies is shown. Optical density (OD) values calculated by densitometry are shown (at bottom) as the ratio of OD VE-cad divided by OD of β-actin for each of the 3 transduced HUVEC conditions. D: transduced HUVEC or noninfected (NI) were seeded and the electrical impedance was monitored as described in methods. Data are presented as a plot of total resistance vs. time and are the average of 3 independent experiments performed under identical conditions.

p120 inhibits TEM independent of an effect on endothelial RhoA activity.

Based on our findings that both inhibition of endothelial RhoA and p120 overexpression block PMN transmigration, we tested whether overexpression of the p120-4AGFP mutant also blocked TEM. Expression of either p120-4AGFP or p120-1AGFP significantly inhibited TEM without a significant effect on leukocyte adhesion (Fig. 4, A and B). This result indicates that p120 overexpression inhibits TEM independent of an effect on adhesion or association with RhoA.

Fig. 4.

p120 overexpression inhibits TEM independently of an effect on RhoA activity. HUVEC were infected with GFP-Adv, p120-4AGFP-Adv, and p120-1AGFP-Adv. A: PMN transmigration under flow conditions is shown as percent relative to GFP-Adv control. PMN remained bound at the cell-cell junctions on monolayers overexpressing p120-4AGFP, whereas PMN disappeared from the cell-cell junctions and efficiently transmigrated across HUVEC transduced with GFP-Adv. Values represent the means ± SD of 5 different experiments using PMN of different donors. **P < 0.01, ***P < 0.001 by Student's t-test. B: total accumulation of PMN was similar in all adenovirus transduced HUVEC monolayers. C: HUVEC monolayers were lysed, and subjected to Western blot for phospho-VE-cad-Tyr658, total-VE-cad, and β-actin. Representative blot is shown from 2 experiments performed. Graph represents normalized values obtained by densitometry analysis, indicating the relative absorbance of phospho-VE-cad-Y658 with respect to total VE-cad for each condition and normalized to the GFP-Adv values. β-actin is shown as a loading control. *P < 0.05 by Student's t-test. D–F: HUVEC monolayers were transduced with GFP or p120-1AGFP, treated with thrombin (2 U/ml) for 1.5 min (D) or 10 min (E and F) in the presence of the RhoA inhibitor peptide C3TCb (E), lysed, and subjected to immunoprecipitation using Rotekin beads (D and E). Data are a representative experiment from 4 studies performed. F: HUVEC were treated with thrombin, fixed, labeled with phalloidin-alexa 568, and imaged using fluorescence microscopy. Data shown represent of 3 studies performed.

We (4) previously showed that p120 regulates TEM by inhibition of tyrosine phosphorylation of the VE-cad cytoplasmic domain triggered by ICAM-1 engagement. Next, we evaluated if overexpression of p120-4AGFP affected phosphorylation of VE-cad at Tyr658 under baseline conditions. Overexpression of both p120 isoforms resulted in a reduction in the levels of VE-cad phosphorylation in resting cells that was concurrent with an increase in total VE-cad protein level. We calculated the ratio P-VE-cad to total VE-cad and showed that it is significantly reduced by overexpression of either isoform of p120 (Fig. 4C), suggesting that p120 overexpression induced a net increase in the amount of nonphosphorylated VE-cad (and a decrease in phosphorylated VE-cad) independent of RhoA association with p120. Given this striking result, we next examined the effects of p120-1AGFP overexpression on RhoA activation. Thus confluent HUVEC monolayers were stimulated with thrombin, a potent RhoA activator (9). We found that p120 overexpression had no effect on GTP-bound (active) RhoA as assayed by Rhotekin pull-down (Fig. 4D). As a control, preincubation with an inhibitor of RhoA (C3TCb peptide) completely inhibited thrombin-induced RhoA activation (Fig. 4E). In a second assay, few actin stress fibers were present in resting HUVEC and p120 overexpression did not prevent thrombin-induced stress fiber formation, a process known to be mediated by RhoA activation (Fig. 4F). Because VE-cad tyrosine phosphorylation is induced downstream of ICAM-1 engagement during TEM (4, 17, 25, 37), we also evaluated whether ICAM-1 crosslinking triggered RhoA activation and if this action was altered by p120 overexpression as determined by G-lisa assays as an alternative method for measuring RhoA activation. Anti-ICAM-1-coated beads induced activation of RhoA that peaked at 10 min in 16-h TNF-α-stimulated HUVEC and overexpression of p120 did not alter this response (Fig. 5A). We also evaluated by G-lisa whether p120 overexpression was linked to an alteration in the activation of Rac, which is activated in other cell types by p120 overexpression (26). We found that ICAM-1 crosslinking led to activation of Rac that peaked after 4 min in TNF-α-treated Adv-GFP HUVEC, and this effect was not altered by p120 overexpression (Fig. 5B). Thus p120 overexpression in HUVEC does not affect the GTPases RhoA or Rac, supporting the conclusion that p120 regulates PMN transmigration independently of an effect on RhoA or Rac activation.

Fig. 5.

Overexpression of p120-1AGFP in HUVEC does not affect RhoA or Rac1 activation induced by ICAM-1 crosslinking. HUVEC were transduced with GFP or p120-1AGFP, treated with TNF-α for 16 h and incubated with media or anti-ICAM-1 MAb coated beads for 10 min (A) or 4 min (B). HUVEC were then lysed and subjected to the indicated G-lisas for RhoA and Rac following manufacturer's instructions. Values represent the means ± SD from 3–5 different experiments *P < 0.05 by Student's t-test.

p120 overexpression prevents the association of pY416-src with VE-cad and reduces PI3K p110α expression.

Because p120 overexpression inhibited TEM through a mechanism involving VE-cad phosphorylation (4), which is independent of RhoA activity as demonstrated in Fig. 4C, and Src is involved in leukocyte TEM and contributes to VE-cad tyrosine phosphorylation (44, 45), we investigated whether p120 overexpression reduces VE-cad tyrosine phosphorylation through a mechanism involving Src inhibition. Treatment of TNF-α stimulated HUVEC with the Src inhibitor PP2 (18) resulted in a 69% inhibition in TEM of PMN compared with carrier (DMSO)-treated cells (Fig. 6A) and in a significant inhibition of P-src association with VE-cad, as detected by coimmunoprecipitation studies and immunoblot using pY416-Src MAb that detects activated Src (28) (Fig. 6B). Interestingly, coimmunoprecipitation studies also revealed that pY416-Src is associated with VE-cad in HUVEC after crosslinking with control beads (anti-MHC-I beads) or anti-ICAM-1 beads and that overexpression of p120 inhibited significantly pY416-Src association with VE-cad in both situations. Notably, a significant amount of pY416-Src is constitutively associated with VE-cad in HUVEC, and crosslinking with anti-ICAM-1 beads did not further enhance p416-Src association with VE-cad, although p120 overexpression still inhibited this association during anti-ICAM-1 crosslinking.

Fig. 6.

Both PMN transmigration and P-Src association with VE-cad are inhibited by PP2 or p120-1A overexpression. A: TNF-α-stimulated HUVEC were treated with PP2 (10 μM, 30 min at 37°C) and TEM studies were performed under flow conditions (1 dyn/cm2) as described in materials and methods. Values represent (# of transmigrated cells divided by the total number of cells bound to the monolayer) × 100 and are the means ± SD of 2 different experiments performed in duplicate. P values compare each bar with Media and DMSO. ***P < 0.001 (Student's t-test). B: HUVEC were treated with PP2 as in A, lysed, and immunoprecipitated with antibody for VE-cad, and the immunoprecipitated material was transferred and blotted for pY416Src or total Src. Blot is representative of 2 independent experiments. C: HUVEC were transduced with GFP or p120-1AGFP, treated with TNF-α (16 h), and incubated with control anti-MHC-I beads or anti-ICAM-1 beads for 10 min, lysed, and subjected to immunoprecipitation and immunoblot as in B. A representative blot is shown from 5 independent experiments. Graph represents normalized values obtained by densitometry analysis corresponding to phospho-Src divided by total Src. *P < 0.05, **P < 0.01 (Student's t-test).

A recent study (12) reported that PI3K isoform p110α contributes to endothelial monolayer permeability and leukocyte TEM through Pyk-2 activation and association with the VE-cad complex, and increased Rac1 and RhoA activity. Interestingly, in our hands overexpression of p120-1A reduced p110α levels by 30% in TNF-α activated HUVEC (Fig. 7A). In our system, however, we were unable to consistently detect p110α association with the VE-cad complex in HUVEC transduced with GFP or p1201A Adv (Fig. 7B). Thus we cannot determine its contribution to inhibition of TEM by p120 overexpression.

Fig. 7.

Overexpression of p120-1AGFP in HUVEC inhibits the expression of the phosphoinositide 3-kinase (PI3K) isoform p110α. A: HUVEC were transduced with GFP or p1201A-GFP as described in materials and methods, lysed, and immunoblotted for p110α, stripped, and blotted for β-actin to serve as the loading control. Representative blot is shown from n = 7 independent experiments with different HUVEC preparations. Bar graph depicts the average OD values normalized to β-actin, and then to GFP control for each independent experiment. *P < 0.05 by Student's t-test. B: HUVEC were immunoprecipitated with antibodies to VE-cad or p110α, lysed, and immunoblotted for VE-cad, p110α, or p120, where indicated. Data represent 2 independent experiments.


The process of leukocyte TEM is complex and known to involve increased RhoA activity (2, 40). p120 overexpression in epithelial cells and fibroblasts results in inhibition of RhoA and activation of Rac and dramatic morphological alterations (10, 26). p120 overexpression in HUVEC and human microvascular endothelial cells results in inhibition of leukocyte TEM without morphological alterations. We (4) have previously shown that p120 regulates leukocyte TEM in association with changes in VE-cad tyrosine phosphorylation. We show here that the block in leukocyte TEM by p120 is independent of an effect on RhoA and instead correlates with a reduction in pY416-Src association with the VE-cad complex. The evidence in support of this conclusion is as follows. Specific inhibition of RhoA using the inhibitor C3TCb resulted in decreased TEM but not to the same extent as p120 overexpression. We took advantage of the mutant p120-4AGFP isoform, which does not interact with RhoA, and characterized its expression in HUVEC. p120-4AGFP associated with VE-cad and colocalized with VE-cad and endogenous p120 at the cell-cell junctions. p120-4AGFP also extended the lifetime of VE-cad surface expression to a similar degree as the p120-1AGFP isoform and like p120-1A isoform did not alter TNF-α-induced expression of ICAM-1, E-selectin, or VCAM-1 adhesion molecules. Overexpression of p120-4AGFP blocked TEM and Y658 VE-cad phosphorylation, indicating that p120 overexpression inhibits TEM and VE-cad tyrosine phosphorylation in a RhoA independent fashion. We also found that p120 overexpression in HUVEC did not alter RhoA activation or stress fiber formation induced by thrombin, contrary to what p120 does in other cell types, or RhoA and Rac activation induced by ICAM-1 crosslinking. Since inhibition of RhoA is required to activate Rac in most systems, it is not surprising that p120 overexpression did not affect Rac activation. These findings led us to look for an alternative mechanism to explain p120 inhibition of TEM. We found that pharmacological inhibition of Src by PP2 resulted in significantly decreased leukocyte TEM that was less but comparable to that of p120 overexpression, and importantly, PP2 also significantly inhibited activation of Src and its association with the VE-cad complex, to a similar extent as p120-1A overexpression. Thus these data suggest a model in which p120 overexpression impairs ICAM-1-induced Src activation and subsequent downstream Src tyrosine phosphorylation of the VE-cad complex, which contributes to TEM inhibition.

Initially, we hypothesized that p120 overexpression was inhibiting RhoA activity to explain decreased TEM, but our current studies with the mutant isoform p120-4AGFP showed this is not the case. It is important to note that the level of p120 expression necessary to inhibit TEM does not affect RhoA activation induced by thrombin or by ICAM-1 engagement in HUVEC and does not alter cell morphology or increase HUVEC monolayer permeability. This contrasts to data generated in fibroblasts, epithelial cells, and bovine pulmonary artery endothelial cells in which p120 overexpression inhibits RhoA and provokes a dendritic-like cell morphology. However, the level of p120 required in those cell types to inhibit RhoA is well above the amount used in our studies. Indeed, p120 expressed at higher levels used in this study does induce a change in cell morphology and increased monolayer permeability (leaky). Despite the reports that p120 overexpression inhibits RhoA in other cell systems, other processes regulated by p120 overexpression do not seem to be mediated through RhoA: p120 regulates endothelial cell permeability independently of RhoA (19), p120 regulates clathrin-dependent endocytosis of VE-cad independently of RhoA (15), and here we show that p120 regulates TEM independently of RhoA.

Our finding (42) that overexpression of p120 reduces leukocyte recruitment is consistent with a recent study in which p120-catenin overexpression results in reduced inflammation and permeability in a LPS-induced murine lung model of inflammation. In contrast, knockdown of p120 in this model resulted in increased neutrophil recruitment, enhanced production of inflammatory cytokines TNF-α and IL-6, loss in lung barrier function, and increased mortality in response to LPS compared with control animals. These data implicate p120 as a modulator of the immune function during inflammation, and the mechanism used by p120 to modulate leukocyte influx and permeability in this LPS model might be similar to the mechanism we propose here in TNF-α activated HUVEC monolayers in vitro. Indeed, we (35) previously reported that HUVEC cells treated with LPS, TNF-α, or IL-1β showed similar induction of ICAM-1 expression.

Several studies (4, 7, 12, 38) have suggested that Src, Pyk-2, and possibly other kinases can phosphorylate cadherins, and in particular, VE-cad during leukocyte adhesion and ICAM-1 engagement. Inhibition of Src kinases by PP2 also results in significant inhibition of TEM (∼60%; Fig. 6A), mostly likely because it directly, or indirectly, interferes with Src mediated VE-cad phosphorylation and thus increases the total amount of nonphosphorylated VE-cad. Our data showing that PP2 prevents the association of pY416-Src with VE-cad are in line with this idea. More importantly, p120 overexpression results in less activated Src associated with VE-cad during ICAM-1 engagement, further supporting the idea that p120 can regulate pY416-Src association with VE-cad, and in this way, regulate VE-cad phosphorylation and leukocyte passage. Although phosphorylation events mediated by Src and other kinases take place at the endothelial junctions during TEM, the association of pY416-Src with VE-cad was not enhanced by ICAM-1 crosslinking, indicating there may be additional membrane pools of activated Src kinases, or alternatively, that different Src kinases are involved because c-Src, Yes, Fyn, and Lyn are expressed in endothelium (20) and that these Src kinases regulate other signaling events during TEM.

Another possibility is that different pools of the VE-cad complex exist at cell-cell junctions, one stable and another less stable, and that p120 overexpression may significantly shift the balance towards an increase in the more stable pool of VE-cad complex (e.g., nonphosphorylated VE-cad) and such regions are resistant to VE-cad gap formation and TEM. In support of this speculation, others (7, 33) have reported that cells within a confluent HUVEC monolayer are motile, and as a consequence cell-to-cell junctions break and reform frequently (i.e., remodel existing junctions). Of particular note Huveneers et al. (21) recently reported that vinculin colocalized with remodeling VE-cad complexes to a much greater degree than “stable VE-cad”, and they termed these remodeling VE-cad-vinculin complexes functional adherens junctions (FAJs). These authors (24) presented evidence that FAJs attach to F-actin bundles at cell-cell junctions that contain vinculin, a protein that localizes to focal adhesions and has mechanosensory properties. Treatment of HUVEC with vascular endothelial growth factor, TNF-α, or thrombin induced the remodeling of “stable junctions” into FAJs. In human pulmonary endothelial cells, paxillin and focal adhesion kinase, two components of the focal adhesions, have been shown to promote barrier function in response to challenge with oxidized 1-palmitoyl-2-arachinoyl-sn-glycero-3-phosphocholine (OxPAPC). This effect was mediated, in part, by association of paxillin with VE-cad and p120, which coimmunoprecipitated in a complex containing vinculin. However, direct association of vinculin with p120 or VE-cad was not detected by immunoprecipitation studies (11). On the other hand, a recent publication (27) has evaluated the role of several focal adhesion proteins in neutrophil transmigration under shear flow conditions. The results showed a selective loss of paxillin and focal adhesion kinase in focal adhesions in the proximity of transmigrating PMN, whereas vinculin expression was not affected. Although this would argue against vinculin playing a role in junctional remodeling during TEM, given the complexities of TEM and barrier regulation, we believe a more comprehensive study is necessary to evaluate whether p120 affects vinculin interactions with the VE-cad complex and AJ remodeling during TEM.

A recent study (12) showed that small interfering RNA silencing of the PI3K p110α isoform reduced the levels of phosphorylated Y731 and Y685 in VE-cad and also reduced the level of phospho-Pyk2 kinase, and the activity of Rac1 and RhoA, whereas Src activity was unchanged. These biochemical data correlated with an increase in endothelial baseline permeability (less leak) and a reduction in TNF-α induced permeability and modest reduction in TEM (∼35%) of human THP-1 cells and T lymphoblast. In additional studies, silencing of Pyk-2, Rac1, or both Pyk-2 and Rac1 partially restored the TNF-α induced permeability increase and leukocyte TEM. Thus this study suggests p110α regulates permeability and leukocyte TEM through Pyk-2 activation and association with the VE-cad complex and increased Rac and RhoA activity. We observed (Fig. 7A) that overexpression of p120-1A reduced p110α levels by 30% in our assay system. While the level of inhibition of p110α expression is modest when overexpressing p120, we speculate it could contribute to the inhibition of P-src association with VE-Cad and explain why overexpression of p120 is more effective than inhibition of Src by PP2 in the TEM assay. Taken together, these data suggest that inhibition of leukocyte TEM by p120 overexpression and p110α silencing are additive, but mechanistically distinct (see the model in Fig. 8).

Fig. 8.

p120 regulation of PMN TEM across HUVEC-mediated VE-cad phosphorylation by effects on Src family kinases. A: TNF-α activated endothelial cells express high levels of cell surface adhesion molecules including ICAM-1 and rely on junctional VE-cad in complex with p120 and P-Src family kinases (active) for maintenance of endothelial cell-cell junctions. B: when p120 is overexpressed, and VE-cad protein levels at cell-cell borders increase and are stabilized (longer half-life) because of decreased phosphorylation, decreased association with P-Src and increased association with p120. C: adhesion of PMN to TNF-α-activated endothelium results in clustering of ICAM-1 around the PMN and its subsequent downstream signaling in the endothelial cell, including P-Src association with the VE-cad complex and phosphorylation of VE-cad, resulting in destabilization of the VE-cad complex at cell-cell junctions. These signaling events lead to PMN TEM. D: TEM is inhibited by p120 overexpression due to increased stability and decreased tyrosine phosphorylation of VE-cad at the cell-cell junction. In addition, PI3K isoform p110α has been shown to contribute to endothelial monolayer permeability and leukocyte TEM (12). Small interfering (si)RNA silencing of p110α reduced the levels of phosphorylated Y731 and Y685 in VE-cad and reduced phospho-Pyk2 kinase levels, and Rac1 and RhoA, whereas Src activity was unchanged. While the level of inhibition of p110α expression is modest in HUVEC overexpressing p120 (Fig. 7), we speculate it could contribute to the inhibition of P-src association with VE-Cad and explain why overexpression of p120 is more effective than inhibition of Src by PP2 in the TEM assay. In summary, these data suggest inhibition of leukocyte TEM by p120 overexpression and p110α silencing are additive but mechanistically distinct.

In summary, we proposed a model in which overexpression of p120 in human endothelial cells inhibits neutrophil transmigration by blocking the association of activated pY416-Src with VE-cad (Fig. 8 and its legend) and by a reduction in expression of PI3K p110-α isoform, which has been implicated in leukocyte TEM by another report (7, 12, 38).


This study was supported by National Heart, Lung, and Blood Institute Grants HL-36028 (FWL), K99-HL-097406 (PA), and R01-HL-077870 (PAV).


No conflicts of interest, financial or otherwise, are declared by the author(s).


Author contributions: P.A., P.A.V., and F.W.L. conception and design of research; P.A., R.M., G.N., M.R.W., A.A., and F.W.L. performed experiments; P.A., R.M., G.N., A.A., and F.W.L. analyzed data; P.A., G.N., M.R.W., A.A., P.A.V., and F.W.L. interpreted results of experiments; P.A. and A.A. prepared figures; P.A. and F.W.L. drafted manuscript; P.A., M.R.W., P.A.V., and F.W.L. edited and revised manuscript; P.A., R.M., G.N., M.R.W., A.A., P.A.V., and F.W.L. approved final version of manuscript.


We thank Kay Case and Vannessa Davis at the Center for Excellence in Vascular Biology, Brigham and Women's Hospital, for providing well-characterized HUVEC in these studies.

Current address for P. Alcaide: Molecular Cardiology Research Institute, Tufts Medical Center, and Tufts University School of Medicine, Boston, MA, 02111.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
View Abstract