Mechanosensitive FHL2 tunes endothelial function via microtubule-actomyosin crosstalk

Figures

• 

  Figure 1. Identification of flow-sensitive gene expression profiles in endothelial cells. (A) Schematic showing the interplay between gene expression and endothelial mechanotransduction that facilitates multi-scale tissue adaptation, morphology, and function. In particular, hypercontractility of tissues through RhoGTPase mechano-signalling, cell-cell adhesion formation, and gene expression, collectively orchestrate tissue functions including permeability. (B) TeloHAECs (human aortic endothelial cells) subjected to two flow profiles: healthy laminar unidirectional flow (UF) and athero-prone disturbed flow (DF). Brightfield images of cells under UF and DF are shown. ****P < 1 × 10−15. (C) Graph shows the aspect ratio (ratio of the major to the minor axis of a cell) of cells under UF and DF. In the box-and-whisker plot, the box extends from the 25th to the 75th percentile, the whiskers show the minimum and maximum values, and the line within the box represents the median. N = 3 independent experiments (n = 361 for UF, and n = 425 for DF). ***P = 0.0004. (D) From brightfield images, the alignment of cells under UF and DF are calculated by local gradient orientation, and represented as the…

• 

  Figure EV1. Actin- and adhesion- related differentially expressed genes in response to flow. (A) Actin-related genes present within the top 10 BP, CC or MF most enriched actin-related GO terms which are significantly different between endothelial cells under UF or DF are shown. The size of the dot reflects the −log10(adjusted P value). (B) Adhesion-related genes present within the top 10 BP, CC or MF most enriched adhesion-related GO terms which are significantly different between endothelial cells under UF or DF are shown. The size of the dot reflects the -log10(adjusted P value). (C) Dotplot showing all genes encoding for LIM domain proteins which were detected in endothelial cells under DF or UF. Genes are ordered by adjusted P value and log2FC. The size of the dots indicates absolute log2FC values. The left-most dot represents the most significantly upregulated gene in DF, and the right-most dot is the most significantly downregulated gene in DF encoding for a LIM domain protein (orange = Padj <0.05 & log2FC >0, purple = Padj <0.05 & log2FC <0, yellow = Padj = > 0.05). Genes of the LIM family that are not expressed in TeloHAECs are also shown.

• 

  Figure 2. FHL2 expression and localisation to stress fibres is enriched in cells under athero-prone DF. (A) Partial ligation of the left carotid artery of the aortic arch in mice (indicated with a red band). The right carotid artery is labelled. 48 h post ligation surgery, mice were euthanised and RNA was extracted from the intima of both carotid arteries (with the contralateral artery as control). A cross-section of a blood vessel is shown, with tunica intima consisting of the endothelial layer. (B) Graph shows the fold change expression of genes with respect to the expression in control (Ctl) arteries, as measured by qPCR. Two flow-sensitive genes nitric oxide synthase 3 (Nos3) and Krüppel-like Factor 2 (Klf2) are shown as positive controls/validations. The fold change in Fhl2 expression is depicted. N = 3 mice. *P = Ctl vs. Nos3, Klf2 is 0.025, 0.0130 respectively, **P = Ctl vs Fhl2 is 0.0056. (C) Confocal images of cross sections of contralateral and ligated mouse arteries stained for FHL2, imaged along with Elastin. Zoomed-in images of FHL2 staining from each arterial section are shown, with ‘L’ indicating the lumen and endothelial side. N = 2 mice (n = 6 sections per condition). (D)…

• 

  Figure EV2. FHL2 expression and localisation in endothelial cells. (A) Western blot of lysates from TeloHAECs subjected to UF and DF for 24 h. Samples were analysed by immunoblotting for FHL2 and GAPDH (loading control). (B) Normalised FHL2 intensity in UF and DF cells. N = 3 independent experiments. Dots represent experimental means, whiskers show the minimum and maximum values, and the line within represents the median. (C) Western blot of lysates from Ctl (-dox) and FHL2-overexpressing (+dox; FHL2 OE) cells. Samples were analysed by immunoblotting for FHL2 and tubulin (loading control). FHL2 blotting reveals two bands - endogenous FHL2 (~32 kDa) and overexpressed m-Emerald tagged FHL2 protein (~60 kDa). (D) Graph shows the ‘total FHL2 expression levels’ in Ctl and FHL2 OE cells, defined by the sum of the intensities of both bands (FHL2 endogenous + overexpression) normalised to the respective loading control. N = 3 independent experiments. Dots represent experimental means, whiskers show the minimum and maximum values, and the line within represents the median. (E) Immunostaining of FHL2 and actin, and visualisation of mEmerald-FHL2 tagging in Ctl and FHL2-OE cells. (F) Ctl and FHL2 OE…

• 

  Figure 3. FHL2 overexpression drives focal adherens junction formation, similar to DF cell clusters. (A) Inverted contrast confocal images of TeloHAECs under UF and DF, stained for actin and vascular endothelial cadherin, VE-Cadherin (to label adherens junctions). Three cell junction types were annotated using VE-Cadherin staining: i. linear junctions, ii. reticular junctions, iii. focal adherens junctions. Examples of each junction type are shown as zoomed-in inverted contrast images on the bottom. (B–D) Graphs showing the aspect ratio (ratio of the major to the minor axis of a cell) vs. the fraction of the cell perimeter occupied by linear junctions (B), focal adherens junctions (C), reticular junctions (D). N = 3 independent experiments (n = 356 for UF, 417 for DF). ns: P value not significant, ****P < 0.0001. (E) Using measurements of aspect ratio and fractions of the three junction types in (B–D), cells were clustered using k-means, with k = 2. Two clusters and the corresponding data points are shown. N > 3 independent experiments (n = 1179 cells clustered into cluster 1 or 2). (F) Inverted contrast confocal images of control (Ctl; -dox) and FHL2-overexpressing (OE; +dox) TeloHAECs…

• 

  Figure EV3. Junction morphology analysis and unbiased clustering of cell populations under varying flow profiles. (A) Graphs showing the aspect ratio (ratio of the major to the minor axis of a cell) vs. the fraction of the cell perimeter occupied by linear junctions (A), focal adherens junctions (B), reticular junctions (C). n = 66 for No Flow, 106 for UF, 149 for DF. (D) Using measurements of aspect ratio and fractions of the three junction types in Fig. 3B–D, cells were clustered using k-means, with k = 2. Two clusters and the corresponding data points are shown. N > 3 independent experiments (n = 1179 cells clustered into Cluster 1 or Cluster 2). Graphs showing the normalised aspect ratio vs. the fraction of the cell perimeter occupied by focal adherens junctions. Each cell is represented as a dot, and all the WT cells under UF and DF from across various experiments were clustered and are represented on the graph. Cells are colour coded based on whether they were correctly categorised into Clusters 1 or 2, based on the experimental condition. All the grey dots show that the cells that were under UF or DF were clustered into UF-like Cluster 1 or DF-like Cluster 2, respectively. The pink…

• 

  Figure EV4. Knockdown of FHL2 in endothelial tissue response to DF. (A) Western blot of lysates from control (Ctl; -dox) and FHL2-knockdown (shFHL2; +dox) TeloHAECs. Samples were analysed by immunoblotting for FHL2 and tubulin (loading control). (B) Normalised FHL2 intensity in Ctl and shFHL2 cells. N = 4 independent experiments. Dots represent experimental means, whiskers show the minimum and maximum values, and the line within represents the median. (C) Inverted contrast images of Ctl and shFHL2 cells immunostained for actin and FHL2. (D) Fluorescence intensity of FHL2 normalised to actin intensity of each cell, in Ctl and shFHL2 TeloHAECs. N = 5 independent experiments (n = 187 for Ctl and 151 for shFHL2). (E) Graph shows the aspect ratio (ratio of the major to the minor axis of a cell) of control and shFHL2 (+dox) cells under DF. N = 2 independent experiments (n = 171 for Ctl, 225 for shFHL2). (F) From brightfield images, the alignment of Ctl and shFHL2 cells under DF are calculated by local gradient orientation and represented as the probability distribution of cells with the flow alignment angles (centred around 0°). Shaded error bars representing standard deviation are shown. N = 1…

• 

  Figure EV5. Knockdown of FHL2 partially reverses focal adherens junction morphology without any effects on control conditions. (A) Inverted contrast confocal images of control (Ctl; -dox) and shFHL2 TeloHAECs under DF, stained for phalloidin and VE-Cadherin. Annotated junctions are depicted as in Fig. 3A (purple: linear junctions, yellow: focal adherens junctions, orange: reticular junctions). (B–D) Graphs showing the aspect ratio (ratio of the major to the minor axis of a cell) vs. the fraction of the cell perimeter occupied by linear junctions (B), focal adherens junctions (C), reticular junctions (D). N = 2 independent experiments (n = 171 for DF Ctl, 225 for DF shFHL2). The DF Ctl means were pooled from 5 experiments for statistical tests. (E) Inverted contrast confocal images of control (Ctl; -dox) and shFHL2 TeloHAECs under UF, stained for phalloidin and VE-Cadherin. Annotated junctions are depicted as in Fig. 3A (purple: linear junctions, yellow: focal adherens junctions, orange: reticular junctions). (F–H) Graphs showing the aspect ratio (ratio of the major to the minor axis of a cell) vs. the fraction of the cell perimeter occupied by linear junctions (F), focal adherens junctions…

• 

  Figure EV6. Clustering analysis of cell populations in UF and DF revealed that FHL2 OE cells are highly represented in DF-like cluster. (A, B) UF Ctl and FHL2 OE cells were clustered (unbiased) into the pre-defined UF-like Cluster 1 (C1) or DF-like Cluster 2 (C2). Graphs showing the normalised aspect ratio vs. the fraction of the cell perimeter occupied by focal adherens junctions in (A) Ctl, (B) FHL2 OE cells. (A) UF Ctl cells completely belong to UF-like Cluster 1. (B) FHL2 OE cells were clustered into the predefined clusters, and the circles represent the dominant cluster (in this case, DF-like Cluster 2; C2), while the ‘x’ depicts the minority cluster (in this case, UF-like Cluster 1; C1). (C, D) DF Ctl and shFHL2 cells were clustered (unbiased) into the pre-defined UF-like Cluster 1 (C1) or DF-like Cluster 2 (C2). Graphs showing the normalised aspect ratio vs. the fraction of the cell perimeter occupied by focal adherens junctions in (C) Ctl, (D) shFHL2 cells. (C) DF Ctl cells almost completely belong to DF-like Cluster 2 (circles represent the dominant cluster, in this case, DF-like Cluster 2, C2; while the ‘x’ depicts the minority cluster (in this case, UF-like Cluster 1; C1). (D) DF…

• 

  Figure 4. FHL2 binding to stress fibres is essential for promoting athero-prone phenotypes. (A) Schematic of the four phenylalanine point mutations created in FHL2, to generate FHL2 F1-4A (mutations previously described in (Sun et al, 2020)). (B) Inverted contrast brightfield images of TeloHAECs overexpressing mEmerald tagged- FHL2 WT or FHL2 F1-4A. (C) Graph shows the aspect ratio (ratio of the major to the minor axis of a cell) of control (-dox) and F1-4A overexpressing (OE; +dox) TeloHAECs under UF and DF. N = 2 independent experiments (n = 143 for Ctl and 167 for F1-4A OE). ns: P value (=0.64) not significant. (D) From brightfield images, the alignment of control and F1-4A overexpressing cells under UF are calculated by local gradient orientation, and represented as the probability distribution of cells with the flow alignment angles (centred around 0°). Shaded error bars representing standard deviation are shown. N = 3 independent experiments (n = 34 for UF Ctl, and n = 39 for UF F1-4A OE; fields of view, with >100 cells/field). ns: P value (=0.22) not significant. (E) Inverted contrast confocal images of control and F1-4A overexpressing TeloHAECs under UF, stained for phalloidin and…

• 

  Figure EV7. FHL2 mutant and clustering analysis shows its similarity with UF-like Cluster 1. (A) Inverted contrast images of FHL2-overexpressing (FHL2 OE; +dox) and FHL2-F1-4A-overexpressing (F1-4A OE; +dox) cells immunostained for VE-Cadherin. mEmerald tagging of FHL2 protein is visualised. (B) Western blot of lysates from control (Ctl; -dox) and F1-4A OE (+dox) cells. Samples were analysed by immunoblotting for FHL2 and tubulin (loading control). FHL2 blotting reveals two bands - endogenous FHL2 (~ 32 kDa) and overexpressed m-Emerald tagged FHL2-F1-4A mutant protein (~60 kDa). N = 2 independent experiments. (C) Graph showing the aspect ratio (ratio of the major to the minor axis of a cell) vs. the fraction of the cell perimeter occupied by reticular junctions. N = 2 independent experiments (n = 143 for UF Ctl, 167 for UF F1-4A OE). One data point belonging to UF Ctl (aspect ratio 18.25, linear 0.62, focal adherens 0.01) used in the analyses is excluded from the graph for representation purposes. The UF Ctl means were pooled from 5 experiments for statistical tests. (D) Inverted contrast confocal images of control (Ctl; -dox) and FHL2-F1-4A-overexpressing (F1-4A OE; +dox) TeloHAECs under…

• 

  Figure 5. FHL2 expression promotes actomyosin contractility. (A) Inverted contrast images of TeloHAECs immunostained for phalloidin (to label actomyosin stress fibres) and phosphorylated (serine 19) myosin light chain (pMLC). (B) Graph shows the normalised pMLC levels (pMLC intensity/actin intensity) by immunostaining in control (Ctl; -dox) and FHL2-knockdown cells (shFHL2; +dox). N = 3 independent experiments (n = 125 for Ctl, 119 for shFHL2). ****P < 0.0001. (C) Western blot of lysates from Ctl and shFHL2 cells. Samples were analysed by immunoblotting for FHL2, pMLC, MLC, and tubulin (loading control). For representative purposes, this blot is also depicted in Fig. EV8A. (D) Normalised pMLC intensity (pMLC/MLC intensity) in Ctl and shFHL2 cells from western blots. N = 4 independent experiments. **P = 0.0023. (E) Traction stress maps a representative Ctl (-dox) and FHL2-knockdown (shFHL2; +dox). Vectors indicate direction and magnitude of the stress. (F) Strain energy of each cell (in Joules) for Ctl (-dox) and FHL2-knockdown (shFHL2; +dox) cells shown. N = 2 independent experiments (n = 64 for Ctl, 75 for shFHL2). ****P < 0.0001. (G) Schematic showing that the levels of FHL2 in cells…