The potassium channel K2P2.1 shapes the morphology and function of brain endothelial cells via actin network remodeling

Abstract

K_2P2.1 (gene: Kcnk2), a two-pore-domain potassium channel, regulates leukocyte transmigration across the blood-brain barrier by a yet unknown mechanism. We demonstrate that Kcnk2^-/- mouse brain microvascular endothelial cells (MBMECs) exhibit an altered cytoskeletal structure and surface morphology with increased formation of membrane protrusions. Cell adhesion molecules cluster on those protrusions and facilitate leukocyte adhesion and migration in vitro and in vivo. We observe downregulation of K_2P2.1 and activation of actin modulating proteins (cofilin 1, Arp2/3) in inflamed wildtype MBMECs. In the mechanosensitive conformation, K_2P2.1 shields the phospholipid PI(4,5)P₂ from interaction with other actin regulatory proteins, especially cofilin 1. Consequently, after stimulus-related K_2P2.1 downregulation and dislocation from PI(4,5)P₂, actin rearrangements are induced. Thus, K_2P2.1-mediated regulatory processes are essential for actin dynamics, fast, reversible, and pharmacologically targetable.

Fig. 1. K_2P2.1 modulation influences T cell migration in the context of neuroinflammation in vivo.

a Scheme of the experimental setup. Adoptive EAE was induced by injecting GFP⁺ Th17 cells (green syringe) intravenously into Rag2^−/−cgn^−/− mice. At disease score of 2, mice were anesthetized, equipped with a carotid artery catheter and prepared for intravital two-photon microscope imaging of the brainstem. Blood brain vessels were visualized by rhodamine-labeled dextran (red syringe). K_2P2.1 inhibition was performed using spadin. b 3D-reconstructions with blood-vessels (red), CD4⁺ T cells (green) were calculated using two-photon imaging derived XYZ-stacks. Representative endoluminal crawling sequence (10 min) depicted from the two perspectives of endoluminal (upper panel) and perivascular (lower panel), respectively. c Representative sequence showing a CD4⁺ T cell (green) extravasating the blood vessel (lumen, red) shown from an extravascular sight. d The same blood vessel (red) was recorded for 30 min before (left), after spadin treatment (center) and after spadin treatment and blocking ICAM1 (right) with tracks from CD4⁺ T cells extravasating the vessel (green). e Mean number of CD4⁺ T cells extravasating the blood vessel in baseline (basel., dots) recordings (N = 5) and after spadin and either isotype antibody (cubes, N = 7) or anti-ICAM1 injection (N = 6). f EAE disease course of endothelial cell-specific Kcnk2^−/− (Kcnk2^fl/flxTie2^cre+, blue) and control (Kcnk2^fl/flxTie2^cre−, white) mice with (gray cubes) or without (black dots) spadin treatment (N = 8). g Analysis of EAE disease courses; Area under the curve (AUC) is shown for the different groups (N = 8). Data are presented as mean ± SEM. N representing the number of individual mice. Statistical analysis using (e) 1-way ANOVA + Bonferroni correction and (f, g) 2-way ANOVA + Bonferroni correction with *p < 0.05 and ***p < 0.001. Exact p-values are listed in the Source Data file.

Fig. 2. K_2P2.1 expression pattern and cell adhesion in vitro.

a mRNA expression levels of Kcnk2 in WT MBMECs with or without TNFα and IFNγ stimulation (500 U/ml, each, white) and after wash out of TNFα/IFNγ (blue). Expression levels were normalized to untreated cells. Statistics were calculated with ΔC_T values (N = 4–6). b Mean fluorescence intensities (MFI, left) and representative histograms (right) of K_2P2.1 in WT MBMECs with or without TNFα/IFNγ stimulation (500 U/ml, each) determined by flow cytometry (N = 6). c Experimental setup for flow conditions. MBMECs were seeded into µ-slides (ibidi®); stimulated WT CD4⁺ T cells were applied under low physiological flow (0.25 dyn/cm²). d Total number of T cells adhering to WT (white, black dots) and Kcnk2^−/− (gray, gray cubes) endothelial cells within 30 min of acquisition (N = 12–16). N representing the number of individual MBMEC preparations. Exact N-numbers for each condition are listed in the Source Data file. Data are presented as mean ± SEM. Statistical analysis using (a, b) Kruskal-Wallis test + Dunn’s multiple comparison correction and (d) 1-way ANOVA + Bonferroni correction with *p < 0.05, **p < 0.01 and ***p < 0.001. Exact p-values are listed in the Source Data file.

Fig. 3. Kcnk2^−/− MBMECs show significantly altered morphology with pro-adhesive protrusions.

a Immunofluorescence staining of untreated and inflamed WT and Kcnk2^−/− MBMECs for f-actin (phalloidin, blue) and ICAM1 (red). White boxes in left panels represent the magnified area in the right panels. Orthogonal view is shown from Z-Stack images. Scale bars represent 25 µm and 5 µm. b 3D surface plot of images (ImageJ) shown in (a). Z-stacks (30 images) were used for analysis of the 3D profile of the cells, 5% signal intensity was set as minimum, 60% as maximum for projection. Smoothing was applied for better visualization. ICAM1 signals are shown in red, phalloidin in blue. c AFM images of WT (left two panels) and Kcnk2^−/− (right two panels) MBMECs. First row shows raw data images used for automated protrusion detection, which are highlighted as green spots in the second row. Z-profile indicated on the upper right panel ranges from 0 µm to 1.5 µm. Third row represents protrusions scaled to the cell surface. Z-profile ranges from 0 nm to 300 nm. Scale bars represent 10 µm. d Protrusion counts (left; N = 3) and volume (right; N = 3) of AFM images shown in (c) of WT (white, black dots) and Kcnk2^−/− (gray, gray cubes) MBMECS. N representing the number of individual MBMEC preparations. e AFM-based single-cell force spectroscopy of stimulated CD4⁺ T cell approached towards the indicated MBMEC monolayer. Contact time 2 s, constant force of 1.5 nN, retraction speed 10 µm/s. Maximum pulling force (left) and adhesion energy (right) in T cell – MBMEC (WT (white, black dots) and Kcnk2^−/− (gray, gray cubes)) adhesion measurements using CellHesion^® (N = 4−8, n = 17−19; each data point represents the mean value of at least 20 force-distance curves measured with one T cell on different MBMECs.). N and n representing the number of individual MBMEC preparations and individual T cells, respectively. Exact N and n-numbers for each condition are listed in the Source Data file. All data are shown as mean +/- SEM. Statistical analysis using (d) 1-way ANOVA + Bonferroni correction and (e) Kruskal-Wallis test + Dunn’s multiple comparison correction with *p < 0.05, **p < 0.01 and ***p < 0.001. Exact p-values are listed in the Source Data file.

Fig. 4. Kcnk2^−/− influences protein expression levels involved in cytoskeletal regulation and cell morphology.

a Volcano Plot of differentially regulated proteins of Kcnk2^−/− versus WT MBMECs, blue = p < 0.05, magenta = p < 0.05 and log2 fold change (FC) > 2 (N = 5). b Tree map of enriched GO-Terms of differentially expressed proteins in naïve Kcnk2^−/− MBMECs versus WT MBMECs. GO-Terms were summarized by REVIGO analysis. c Network analysis of differentially regulated pathways, proteins and respective interaction partners in naïve Kcnk2^−/− MBMECs versus WT MBMECs. Red circles indicate upregulation, blue circles downregulation in Kcnk2^−/− MBMECs. Yellow borders indicate an ion-dependency of respective proteins. Interaction of proteins are shown according to STRING analysis. Biological processes are annotated within the respective circles. d Representative AFM (atomic force microscopy) images of WT and Kcnk2^−/− MBMECs. Color code depicts height of observed structures (0−500 nm), scale bars represent 5 µm. e Calculations of the fibrosity index from AFM images in (d) of WT (white, black dots) and Kcnk2^−/− (gray, gray cubes) MBMECs (N = 3). f AFM-based measurements of cortical stiffness of WT (white, black dots) and Kcnk2^−/− (gray, gray cubes) MBMECs (N = 4, n = 45–59; each data point represents the mean value of 6 force-distance curves of one MBMEC cell). N and n representing the number of individual MBMEC preparations and individual MBMEC cells, respectively. Exact N and n-numbers for each condition are listed in the Source Data file. All data are shown as mean +/- SEM. Statistical analysis using (a) two-tailed student’s t-test + Benjamini-Hochberg correction for multiple testing and (e, f) 1-way ANOVA + Bonferroni correction with *p < 0.05 and ***p < 0.001. Exact p-values are listed in the Source Data file.

Fig. 5. K_2P2.1 depletion leads to altered expression of cytoskeleton regulators.

a Venn Diagram of significantly (p < 0.05) differentially regulated genes of inflamed WT (gray), untreated Kcnk2^−/−(red) and inflamed Kcnk2^−/− (orange) MBMEC (to untreated WT MBMECs). Numbers of overlapping transcripts are shown within the respective circles. The actin depolymerizing factor cofilin 1 (Cfl1) is differentially expressed in inflamed WT, as well as untreated and inflamed Kcnk2^−/− MBMECs (N = 4). b−d qRT-PCR data of (b) Cofilin1 (Cfl1), (c) LIM domain kinase 1 (Limk1) and (d) Slingshot homolog 1 (Ssh1) expression in WT (white, black dots) and Kcnk2^−/− (gray, gray cubes) MBMECs upon inflammation (TNFα/IFNγ). Statistics were calculated with ΔC_T values (N = 3–4). e Representative immunofluorescence staining of untreated and inflamed (TNFa/IFNγ for 3 h and 6 h) WT and Kcnk2^−/− MBMECs for Cfl1 (green) or phosphorylated Cfl1 (pCfl1, green). DAPI is shown in blue. Scale bar represents 50 µm. f Quantification of Cfl1 and pCfl1 in untreated and TNFα/IFNγ treated WT and Kcnk2^−/− MBMECs by immunofluorescence staining shown in (e). Normalized data (to DAPI) was used to calculate ratios between Cfl1 and pCfl1 (N = 4). N representing the number of individual MBMEC preparations. Exact N-numbers for each condition are listed in the Source Data file. All data are shown as mean +/- SEM. Statistical analysis using (a−d, f) 1-way ANOVA + Bonferroni correction with *p < 0.05, **p < 0.01 and ***p < 0.001. Exact p-values are listed in the Source Data file.

Fig. 6. Interaction of K_2P2.1 and Cfl1 with PI(4,5)P₂.

a Representative immunofluorescence images of untreated and TNFα/IFNγ treated (for 3 h, 6 h, 24 h) Kcnk2^−/− and WT MBMECs. Actin is stained in red using phalloidin, nucleus in blue (DAPI). White boxes indicate the magnified image depicted on the right side of each panel. Scale bars represent 25 µm and 5 µm, respectively. b, c Quantification of actin stress fiber diameter (b) and length (c) in images from (a) of WT (white, black dots) and Kcnk2^−/− (gray, gray cubes) MBMECs (N = 4). d Representative immunofluorescence staining of untreated and inflamed (TNFα/IFNγ for 3 h, 6 h and 24 h) WT and Kcnk2^−/− MBMECs using a proximity ligation assay (PLA). Red dots indicate interaction of Cfl1 with PI(4,5)P₂ in the PLA, DAPI is shown in blue. Scale bar represents 25 µm. e Quantification of dots in the PLA shown in (d). Positive events per image are depicted (N = 4). f Representative immunofluorescence staining of untreated and inflamed (TNFα/IFNγ for 3 h, 6 h and 24 h) WT MBMECs using a PLA. Untreated Kcnk2^−/− MBMECs were used as control for the PLA. Red dots indicate interaction of K_2P2.1 with PI(4,5)P₂ in the PLA, DAPI is shown in blue. Scale bar represents 25 µm. g Quantification of dots in the PLA shown in (f). Positive events per image are depicted (N = 5). N representing the number of individual MBMEC preparations. All data are shown as mean +/- SEM. Statistical analysis using (b−g) 1-way ANOVA + Bonferroni correction with *p < 0.05, **p < 0.01 and ***p < 0.001. Exact p-values are listed in the Source Data file.

Fig. 7. Knock-down of Cfl1 decreases number of adherent T cells under inflammatory conditions.

a Representative immunofluorescence staining of untreated and inflamed (TNFα/IFNγ for 24 h) WT and Kcnk2^−/− MBMECs, transfected with 240 nM of non-target control (NTC) or Cfl1 SMARTpool. Cfl1 is shown in green, DAPI in blue, scale bar represents 50 µm. b mRNA expression levels of Cfl1 in untreated and inflamed (TNFα/IFNγ for 24 h) WT (white, black dots) and Kcnk2^−/− (gray, gray cubes) MBMECs, transfected with 240 nM of non-target control (NTC) or Cfl1 SMARTpool assessed by qRT-PCR. Data are shown as n-fold change, normalized to the respective NTC WT MBMEC control condition. Statistics were calculated with ΔC_T values (N = 3). c Untreated and inflamed (TNFα/IFNγ for 24 h) WT and Kcnk2^−/− MBMECs, transfected with 240 nM of non-target control (NTC) or Cfl1 SMARTpool MBMECs were seeded into µ-slides (ibidi®); stimulated WT CD4⁺ T cells were applied under low flow (0.25 dyn/cm²). Total number of T cells adhering to the endothelial cells within 30 min of acquisition (N = 8-10). N representing the number of individual MBMEC preparations. Exact N-numbers for each condition are listed in the Source Data file. Data are presented as mean ± SEM. Statistical analysis using (b, c) 2-way ANOVA + Bonferroni correction with *p < 0.05, **p < 0.01 and ***p < 0.001. Exact p-values are listed in the Source Data file.

Fig. 8. Model of interaction mechanisms between K_2P2.1 and the actin cytoskeleton.

a K_2P2.1 (blue) is present as a homodimer in the plasma membrane. No interaction of the C-terminal domain of K_2P2.1 with the membrane in the closed state of the channel. Dephosphorylated E306 and S333 ensure membrane targeting of K_2P2.1 and provide potential interaction sites with the actin cytoskeleton (black lines). Additionally, binding to actin takes place by a hydrophobic region (red). b Gated state of K_2P2.1 is induced by binding of positively charged amino acids (+) in the C-terminal domain to PI(4,5)P₂. K_2P2.1 is thereby sensitive to stretch-activation. Actin turnover in close vicinity to this membrane patch is regulated in space and time by anchoring K_2P2.1 to PI(4,5)P₂. This binding prevents the interaction of PI(4,5)P₂ and Cfl1 which in turn leads to a dampened Arp2/3(green) and Cfl1 (yellow) activity. Upon mechanical, stretch-induced, activation of K_2P2.1, actin depolymerization is activated. The actin cytoskeleton stays in a tightly regulated and controlled state. c Downregulation of K_2P2.1 after internalization or knock-out results in an altered regulation of actin turnover causing membrane protrusion induction at the cell periphery and formation of actin stress fibers inside the cell body. This effect is mediated by the disrupted interaction of K_2P2.1 with PI(4,5)P₂ (orange) and thereby the possible binding of Cfl1. Arp2/3 induces protrusion formation and exclusion of Cfl1 and Arp2/3 in the cell body stabilize actin stress fibers. K_2P2.1 can be found in close association to actin stress fibers. d Overview of K_2P2.1-mediated processes in MBMECs as described in (a−c) and the impact on ICAM1 expression, distribution and T cell (purple) adhesion and transmigration at the blood-brain barrier.