The RhoA guanine nucleotide exchange factor, LARG, mediates ICAM-1-dependent mechanotransduction in endothelial cells to stimulate transendothelial migration

Abstract

RhoA-mediated cytoskeletal rearrangements in endothelial cells (ECs) play an active role in leukocyte transendothelial cell migration (TEM), a normal physiological process in which leukocytes cross the endothelium to enter the underlying tissue. Although much has been learned about RhoA signaling pathways downstream from ICAM-1 in ECs, little is known about the consequences of the tractional forces that leukocytes generate on ECs as they migrate over the surface before TEM. We have found that after applying mechanical forces to ICAM-1 clusters, there is an increase in cellular stiffening and enhanced RhoA signaling compared with ICAM-1 clustering alone. We have identified that leukemia-associated Rho guanine nucleotide exchange factor (LARG), also known as Rho GEF 12 (ARHGEF12) acts downstream of clustered ICAM-1 to increase RhoA activity, and that this pathway is further enhanced by mechanical force on ICAM-1. Depletion of LARG decreases leukocyte crawling and inhibits TEM. To our knowledge, this is the first report of endothelial LARG regulating leukocyte behavior and EC stiffening in response to tractional forces generated by leukocytes.

Figure 1

Mechanical force on ICAM-1 increases cellular stiffening. Magnetic beads coated with ICAM-1 mAb were added to a monolayer of TNF-treated HMVECs. Magnetic tweezers were used to apply pulses of force to individual beads and bead movement recorded with high-speed video. (A) Typical displacement of a bead bound to ICAM-1. Arrows denote displacement distance (Top). A diagram of the 160 pN force regimen used (3s of force with 5s recovery for 5 pulses) (Lower). Percentage bead displacement in response to sequential pulses of force for ECs plated on collagen (B) or fibronectin (C). For D–F, the ECs were plated on collagen. (D) Bead displacements on HMVECs treated with specified inhibitors for 30 min followed by 2 pulses of force. (E) Bead displacement on HMVECs and HMVECs treated with miRNA to inhibit RhoA expression with or without rescue with myc-RhoA. (F) Western blotting confirms RhoA KD and myc-RhoA re-expression. (B–E) Quantification of bead displacement with each pulse normalized to the first pulse. Asterisks shows p-value of statistical significance compared to the control (*, p≤0.05; **, p≤0.01). The means ± SEM of ≥9 independent bead pulls are shown.

Figure 2

Mechanical force on ICAM-1 increases RhoA activity and MLC phosphorylation. Magnetic beads coated with mAb against ICAM-1 (A, C, and D) or MHC class I (B and C) were added for 15 min to a monolayer of TNF-treated HUVECs and ~10 pN force was applied with a ceramic magnet placed above the cells for 1 min. (A and B) Using GST-RBD, RhoA. GTP was isolated and detected by immunoblotting. (C) Phase contrast images of EC monolayers 15′ after beads were added and washed 2x with media before fixing. (D) Lysates were immunoblotted for total MLC or MLC phosphorylated on Thr18/Ser19. Graphs show quantification of RhoA activity (A and B) or pMLC levels (D) from ≥3 independent experiments. Graphs show the means ± SEM. Asterisk shows p-value of statistical significance compared to control (*, p≤0.05; **, p≤0.01).

Figure 3

LARG is activated downstream of ICAM-1 clustering alone and enhanced with mechanical force. TNF-treated HUVECs were treated with mAb-coated beads. (A–F) GEF activity was determined by affinity purification via GST-RhoA^G17A and detected by immunoblotting for the specified GEF, LARG (A and F), p190RhoGEF (B), p115 (C), GEFH-1 (D), PDZ-RhoGEF (E). (G) LARG was immunoprecipitated and immunoblotted for phosphotyrosine and LARG. (H) Active LARG was detected by sedimentation with GST-RhoA^G17A in the presence of SU6656. For all experiments, a representative blot of ≥2 independent experiments is shown. Graphs show the means ± SEM. Asterisk shows p-value of statistical significance compared to control by t test (*, p≤0.05; **, p≤0.01).

Figure 4

Confirmation of LARG KD. HMVECs were treated with control or LARG shRNA lenti-virus for 48 h and selected for with 2.5 ng/ml puromycin for 24 h. (A) EC lysates were immunoblotted with the indicated pAb. (B) Western blotting shows that ICAM-1 expression before and after TNF-treatment is not affected by LARG KD. (C) Electrical impedance was used to measure monolayer integrity for HMVECs plated at high density for 72 h. No significant difference was found in impedance values after control or LARG KD. n=3 independent experiments preformed in triplicate wells.

Figure 5

LARG mediates EC response to mechanical force on ICAM-1 and affects neutrophil crawling and TEM. HUVECs were treated with control (B and C) or LARG (A and C) shRNA lenti-virus for 48 h and selected with 2.5 ng/ml puromycin for 24 h, then TNF-treated overnight. (A and B) RhoA activity was determined by immunoblotting after ICAM-1 clustering with or without force in HUVECs (left) and quantified (right). The means ± SEM of ≥4 independent experiments are shown. Asterisk shows p-value of statistical significance by t test (*, p≤0.05). (C) The stiffness of HMVECs was measured using magnetic tweezers and magnetic beads coated with ICAM-1 mAb. (D) Relative displacement of magnetic beads coated with ICAM-1 mAbs was measured in control HMVECs or in HMVECs in which LARG expression had been knocked down. The means ± SEM of N≥15 independent bead pulls are shown. Asterisk shows p-value of statistical significance by t test (p≤0.01). (E and F) Neutrophils were added to a monolayer of TNF-treated HMVECs after LARG expression had been knocked down. (E) Neutrophils were imaged as they migrated over the HMVEC monolayer surface and their velocity was measured using tracking software. Data are the average of 3 experiments with ≥15 neutrophils measured per experiment. (F) The passage of neutrophils across a confluent EC monolayer was measured using transwell tissue culture inserts. Data are the average of 3 experiments each performed in duplicate. The means ± SEM are graphed. Asterisk shows p-value of statistical significance (*, p≤0.05; **, p≤0.01).