FAK regulates tension transmission to the nucleus and endothelial transcriptome independent of kinase activity

Abstract

The mechanical environment generated through the adhesive interaction of endothelial cells (ECs) with the matrix controls nuclear tension, preventing aberrant gene synthesis and the transition from restrictive to leaky endothelium, a hallmark of acute lung injury (ALI). However, the mechanisms controlling tension transmission to the nucleus and EC-restrictive fate remain elusive. Here, we demonstrate that, in a kinase-independent manner, focal adhesion kinase (FAK) safeguards tension transmission to the nucleus to maintain EC-restrictive fate. In FAK-depleted ECs, robust activation of the RhoA-Rho-kinase pathway increased EC tension and phosphorylation of the nuclear envelope protein, emerin, activating DNMT3a. Activated DNMT3a methylates the KLF2 promoter, impairing the synthesis of KLF2 and its target S1PR1 to induce the leaky EC transcriptome. Repleting FAK (wild type or kinase dead) or inhibiting RhoA-emerin-DNMT3a activities in damaged lung ECs restored KLF2 transcription of the restrictive EC transcriptome. Thus, FAK sensing and control of tension transmission to the nucleus govern restrictive endothelium to maintain lung homeostasis.

Keywords: CP: Cell biology; DNA methylation; DNMT3a; FAK; KLF2; S1PR1; emerin; endothelial cell; intracellular tension; stiffness.

Figure 1.. FAK regulation of nuclear mechanotransduction controls chromatin accessibility of transcription factors

(A) Young’s modulus (E) in a single control or FAK-depleted ECs using atomic force microscopy (AFM) by nanoscale indentation at 1 m/s. N = 40 cells/group, pooled from three independent experiments (n = 3, unpaired t test). (B and C) Collagen gel contraction assay in control and FAK-depleted ECs. A representative photograph is shown in (B), while (C) shows the quantification of gel contraction from three independent experiments (n = 3, unpaired t test). (D) Young’s modulus (E) in FAK-depleted ECs following 1-h treatment with 20 μM blebbistatin or 10 μM Y-27632. The data were pooled from three independent experiments (one-way ANOVA with Tukey’s post hoc test). (E) Young’s modulus in FAK-depleted EC transducing vector, WT-FAK, KD-FAK, and DN-RhoA cDNAs performed as in (A). N = 40 cells/group, pooled from three independent experiments (n = 3, one-way ANOVA with Tukey’s post hoc test). (F and G) Workflow of indicated cDNAs delivery in ECs of EC-FAK-null and control mice using liposomes (F) and lung edema (G) assessment (n = 5 mice/group, one-way ANOVA with Tukey’s post hoc test). (H and I) Schematic of EC sorting (CD45⁻CD31⁺dsRed⁺) from lungs of control tdTomato and tdTomato-EC-FAK^−/− mice for ATAC-seq analysis (H). Snapshot of genomic loci obtained from Integrative Genomic Viewer, showing chromatin-accessible peaks at the transcription start sites (TSSs) for KLF2, SOX17, and FOXP1 in FAK^fl/fl and EC-FAK^/lung ECs (I). (J and K) TEER in control (sh), KLF2-depleted ECs (shKLF2) (J), or FAK-depleted EC transducing vector or KLF2 cDNA (K). The data were pooled from four experiments (n = 4 wells/group) (unpaired t test for J, one-way ANOVA with Tukey’s post hoc test for K). (L) Lung edema in control or EC-FAK-null mice following indicated gene delivery as described in (F) and (G) (n = 6 mice/group, one-way ANOVA with Tukey’s post hoc test). Data show individual values along with mean ± SEM. Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, not significant. Please see Figure S1.

Figure 2.. FAK upregulates KLF2 by suppressing DNA methylation of KLF2 promoter

(A) Schematic of KLF2 promoter containing CpG island (−343 to −524) from the transcription start site. (B and C) (B) Bisulfite sequencing of whole-genome DNA from control or FAK-depleted ECs to quantify KLF2 promoter methylation using quantification tool for methylation analysis (QUMA). Methylated CpG, dark circles; unmethylated CpG, clear circles. (C) A pie-chart of quantified methylated and unmethylated CpG (n = 3). (D and E) Methylation-specific PCR (MS-PCR) of bisulfite-converted genomic DNA from control or FAK-depleted ECs. A representative DNA agarose gel of amplified products is shown in (D), while (E) shows the quantification of the amplified products (n = 3, unpaired t test). (F) Chromatin immunoprecipitation (ChIP) assay followed by qPCR of amplified MEF2 binding on KLF2 promoter in control or FAK-depleted ECs (n = 3, unpaired t test). (G and H) DNMT activity in FAK^fl/fl and EC-FAK^−/− lungs (G) or control and FAK-depleted ECs (H) (n = 3, unpaired t test). (I and J) Stochastic optical reconstruction microscopy (STORM) imaging of single nuclei in control versus FAK-depleted ECs performed using GE OMX-SR Super-Resolution microscope. A representative micrograph is shown in (I), while (J) shows the quantification of 5-mC fluorophore molecules in 18 nuclei from control or FAK-depleted ECs pooled from three independent experiments (n = 3, unpaired t test). Scale bar: 5 μm. (K) DNMT activity in control, FAK-depleted, DNMT3a-depleted, and DNMT3a- and FAK-depleted ECs (n = 3; one-way ANOVA with Tukey’s post hoc test). Data show individual values with mean ± SEM. Statistical significance: **p < 0.01, ##p < 0.01, ****p < 0.0001; ns, not significant.

Figure 3.. DNMT3a compromises lung vascular homeostasis in FAK-deleted ECs

(A–D) Protocol for inhibiting DNMT3a in lungs of WT mice (A) and assessment of DNMT activity (n = 3) (B), lung KLF2 mRNA expression (n = 3) (C), and lung edema (n = 5 mice/group) (D) (one-way ANOVA with Tukey’s post hoc test). (E) FAK and KLF2 mRNA expression in control, FAK-depleted, DNMT3a, and FAK+DNMT3a-depleted ECs using GAPDH as an internal control (n = 3, one-way ANOVA with Tukey’s post hoc test). (F–H) Schematic of EC-FAK^−/−/DNMT3a^−/− (double knockout) mouse generation (F). (G) Lung edema in the indicated mouse group (n = 6 mice/group, one-way ANOVA with Tukey’s post hoc test). (H) mRNA expression of indicated genes using GAPDH as an internal control (n = 3, one-way ANOVA with Tukey’s post hoc test). (I–K) Schematic for theaflavin-3,3′-digallate (TF-3) treatment (I), TEER (J), and Young’s modulus (K). The data were pooled from four experiments (n = 4 wells/group), and N = 40 cells/group in (K) pooled from three independent experiments (one-way ANOVA with Tukey’s post hoc test in J and K). Data show scatter along with mean ± SEM. Statistical significance: *p < 0.05, #p < 0.05, **p < 0.01, ##p < 0.01, ***p < 0.001, ###p < 0.001, ****p < 0.0001; ns, not significant.

Figure 4.. Transcriptome analysis identifies S1PR1 as the gene of interest in maintaining vascular integrity downstream of FAK

(A–C) Volcano plot (A) of differential genes from bulk RNA-seq analysis of ECs sorted from FAK^fl/fl and EC-FAK^−/− mouse lungs. S1PR1 mRNA (B) or protein (C) in FAK⁺ and FAK⁻ ECs using GAPDH as an internal mRNA control and actin as a loading control. Numbers indicate densitometric analysis (n = 3, unpaired t test). (D) Lung edema 30 min after S1P (1 mg/kg, i.v.) administration in FAK^fl/fl or EC-FAK^−/− mice (n = 5 mice/group, one-way ANOVA with Tukey’s post hoc test). (E and F) TEER in control and FAK-depleted ECs at baseline and after adding 1 μM S1P. A representative trace from a single experiment with four wells/group is shown in (E), while (F) shows quantification pooled from four experiments (n = 4 wells/group) (one-way ANOVA with Tukey’s post hoc test). (G–I) Representative TEER trace in FAK-depleted ECs expressing vector or GFP-tagged WT-S1PR1 cDNA after addition of 1 μM S1P (n = 4 wells/group) is shown in (G), while (H) shows quantification pooled from four sets of experiments (n = 4 wells/group). (I) A representative immunoblot of S1PR1 expression (n = 3). Actin was used as a loading control. The densitometric analysis could not be done as the GFP vector was undetected. (J) Lung edema in FAK^fl/fl and EC-FAK^−/− mice receiving vector or GFP-tagged WT-S1PR1 cDNA as in Figure 1F (n = 6 mice/group, one-way ANOVA with Tukey’s post hoc test). Data are shown as individual values along with mean ± SEM. Statistical significance: *p < 0.05, **p < 0.01, ****p < 0.0001.

Figure 5.. KLF2 regulates S1PR1 expression downstream of FAK

(A and B) Schematic representation of S1PR1 promoter with three KLF2 binding sites (−341, −501, and −937 from TSS) (A) and luciferase activity (B) in ECs co-transfecting KLF2 cDNA with WT- or mutated-S1PR1 luciferase promoter (which lacks three KLF2 binding sites) (n = 3; one-way ANOVA with Tukey’s post hoc test). (C and D) S1PR1 mRNA (C) or protein expression (D) in control (shCtr) or KLF2-depleted ECs (shKLF2) using GAPDH as an internal control for mRNA and actin as a loading control. Numbers indicate densitometric analysis (n = 3, unpaired t test). (E) A representative TEER trace in control or KLF2-depleted ECs after S1P (1 μM) addition. The experiments were repeated four times. (F) FAK and S1PR1 mRNA expression in FAK-depleted ECs, transducing KLF2 cDNA using GAPDH as an internal control (n = 3, one-way ANOVA with Tukey’s post hoc). (G and H) A representative TEER in control or FAK-depleted EC transducing vector or KLF2 cDNA (G). (H) Quantification of TEER pooled from four sets of experiments (n = 4 wells/group) (one-way ANOVA with Tukey’s post hoc test). (I) Young’s modulus measurement in control or FAK-depleted ECs transducing vector or KLF2 cDNA using AFM by nanoscale indentation at 1 m/s. N = 40 cells/group, pooled from three independent experiments (n = 3, one-way ANOVA with Tukey’s post hoc test). (J and K) S1PR1 mRNA in control, FAK-depleted, DNMT3a-depleted, and FAK+DNMT3a-depleted ECs (J) or following DNMT3a inhibition with 1 μM TF-3 for 4 h (K). GAPDH was used as an internal control (n = 3, one-way ANOVA with Tukey’s post hoc test). (L) S1PR1 mRNA in FAK^fl/fl and EC-FAK^−/− mice 24 h after receiving 5-aza-2′-deoxycytidine (AZA) (0.5 mg/kg, i.v.) or TF-3 (1 mg/kg, i.v.). GAPDH was used as an internal control (n = 3, one-way ANOVA with Tukey’s post hoc test). Data scatter along with mean ± SEM. Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, not significant. Please see Figure S7.

Figure 6.. FAK regulates nuclear mechanotransduction by suppressing emerin activity

(A and B) A representative micrograph showing emerin localization in the nuclei of control or FAK-depleted ECs (A). Insets in (A) are magnified ×3. Scale bar: 20 μm for unmagnified images and 60 μm for magnified images. (B) A uniform area was drawn around the nuclear periphery in control or FAK-depleted ECs, and fluorescence intensity was quantified. Data are representative of at least n = 12 from experiments independently repeated three times (unpaired t test). (C) cSrc-induced tyrosine phosphorylation in control or FAK-depleted ECs treated with Y-27632 (10 μM) as in Figure 1D or cSrc inhibitor (saracatinib, 100 nM, 2 h). An immunoblot with cSrc pan antibody was used to control protein loading. A representative blot is shown from experiments that were independently repeated three times. Numbers indicate densitometric analysis (n = 3, one-way ANOVA with Tukey’s post hoc test). (D) Emerin tyrosine phosphorylation in the nuclear fraction from control or FAK-depleted ECs after treatment with indicated inhibitors. Immunocomplexes wereimmunoblotted with p-Tyr (PY99/PY29, 1:250 dilution) to assess emerin tyrosine phosphorylation over total emerin and lamin. A representative immunoblot is shown from experiments that were repeated three times. Numbers indicate densitometric analysis (n = 3, one-way ANOVA with Tukey’s post hoc test). (E and F) Emerin reorganization in FAK-depleted ECs transducing WT or phospho-deficient emerin mutant (Y74F/Y95F) was determined as in (A). Scale bar: 20 μm for unmagnified images and 60 μm for magnified images. The fluorescence intensity of emerin was assessed as in (B) (data pooled from three individual experiments, one-way ANOVA with Tukey’s post hoc test). Data points show individual values along with mean ± SEM. Statistical significance: *p < 0.05, ***p < 0.001, ****p < 0.0001; ns, not significant.

Figure 7.. Suppression of emerin activity subverts restrictive EC fate

(A–E) DNMT activity (A), KLF2 and S1PR1 mRNA (B), and intercellular gaps measured by immunostaining with anti-VE-Cadherin antibody (C–E) in FAK-depleted ECs transducing WT or phosphor-defective emerin. GAPDH was used as an internal control for RNA in (B). Intercellular gap (n = 3) (D) and VE-cadherin fluorescence intensity (E) (n = 3, one-way ANOVA with Tukey’s post hoc test for A, B, D, and E). Scale bar: 20 μm. (F and G) TEER (F) and Young’s modulus (G) are in control or FAK-depleted ECs, transducing WT or phospho-deficient emerin mutants. (G) Quantification of TEER pooled from four sets of experiments (n = 4 wells/group). N = 40 cells/group in (G) pooled from three independent experiments (one-way ANOVA with Tukey’s post hoc test for F and G). Data expressed as mean ± SEM. Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, not significant.