Tuning of Liver Sieve: The Interplay between Actin and Myosin Regulatory Light Chain Regulates Fenestration Size and Number in Murine Liver Sinusoidal Endothelial Cells

Abstract

Liver sinusoidal endothelial cells (LSECs) facilitate the efficient transport of macromolecules and solutes between the blood and hepatocytes. The efficiency of this transport is realized via transcellular nanopores, called fenestrations. The mean fenestration size is 140 ± 20 nm, with the range from 50 nm to 350 nm being mostly below the limits of diffraction of visible light. The cellular mechanisms controlling fenestrations are still poorly understood. In this study, we tested a hypothesis that both Rho kinase (ROCK) and myosin light chain (MLC) kinase (MLCK)-dependent phosphorylation of MLC regulates fenestrations. We verified the hypothesis using a combination of several molecular inhibitors and by applying two high-resolution microscopy modalities: structured illumination microscopy (SIM) and scanning electron microscopy (SEM). We demonstrated precise, dose-dependent, and reversible regulation of the mean fenestration diameter within a wide range from 120 nm to 220 nm and the fine-tuning of the porosity in a range from ~0% up to 12% using the ROCK pathway. Moreover, our findings indicate that MLCK is involved in the formation of new fenestrations-after inhibiting MLCK, closed fenestrations cannot be reopened with other agents. We, therefore, conclude that the Rho-ROCK pathway is responsible for the control of the fenestration diameter, while the inhibition of MLCK prevents the formation of new fenestrations.

Keywords: MLC phosphorylation; MLCK; ROCK; actin; fenestration; liver sinusoidal endothelial cells; myosin regulatory light chain; non-muscle myosin II; scanning electron microscopy (SEM); structured illumination microscopy (SIM).

Figure 1

Immunofluorescence images of untreated LSEC collected using SIM. pMLC (a–e) and ppMLC (f–j). Individual black and white channels show actin (phalloidin-Atto647N, (c,h), cell membrane (CellMask Orange), (d,i), and pMCL/ppMLC staining (e,j)). Coloured images show merged channels. The selected areas of sieve plates (white boxes) were digitally magnified (b,g). The antibody is located on actin fibres at the edges of a sieve plate, and not within a sieve plate. ppMLC was also concentrated in the nuclear area. Image size: 40.96 µm × 40.96 µm, inset size: 4.1 µm × 4.1 µm. A gamma of 0.4 was used for the actin and cell membrane for better visualization of the data.

Figure 2

The effect of ROCK and MLCP inhibition on fenestrations in LSEC. (A) Schematic presentation of the mechanism of ROCK inhibition by Y27632. MLC is partially dephosphorylated (faded-red colour) by direct ROCK inhibition by Y27632. Phosphorylation by MLCK is still possible, but MLCP cannot be phosphorylated (deactivated) by ROCK, causing additional dephosphorylation. (B) Representative SEM image of LSEC treated with 10 µM Y27632+CytB. The inset highlights two groups of normal-sized and <100 nm-sized fenestrations. (C) Fenestration diameter distribution (fenestrations measured: ctrl-7843, Y27632-36358, Y27632+CytB-54492, cytB-17230). The proportion of fenestrations >200 nm is 17%, 12%, 8%, and 11% for ctrl, Y27632, Y27632+CytB, and CytB, respectively. In addition, the cumulative effect of 10 µM Y27632 with CytB resulted in the formation of a second group of fenestrations of diameters < 100 nm. (D) A change in the porosity relative to the control is presented (n = 3). The ROCK inhibition resulted in a significant increase in porosity; the effect is cumulative with CytB. (E) A schematic presentation of MLCP inhibition by CalA. MLC is phosphorylated (vivid-red colour) by ROCK and MLCK and no dephosphorylation pathway (via MLCP) is available. (F) A representative SEM image of LSEC treated with 100 nM CalA+CytB. The inset highlights fenestrations with diameters > 500 nm (presented SEM images B and F are in the same scale). (G) Fenestration size distribution after MLCP inhibition (the number of fenestrations measured: ctrl-6157, CalA-3467, CalA+CytB-1103). The proportion of fenestrations > 200 nm increased with the appearance of a second peak of 206 nm. Pre-treatment with CytB (30 min) resulted in mean 67% (n = 3) fenestration diameter >200 nm. (H) Change in porosity relative to the control (n = 3). MLCP inhibition resulted in a significant decrease in the porosity down to zero (no fenestrations identified) for 100 nM; the defenestrating effect of CalA is partially reduced by 30 min pre-treatment with CytB. (I) Fluorescence signal for pMLC and ppMLC measured using SIM for individual well-spread (fenestrated) LSEC. * p < 0.01, relative to control. The mean porosity of the control varied from 5.5% to 7.5%.

Figure 3

The SIM projection images (phalloidin-Atto647N actin staining) of representative LSEC treated with the inhibitors. The selected area (white square) on LSEC periphery was digitally magnified (inset). All images size: 40.96 µm × 40.96 µm, insets: 6.2 µm × 6.2 µm.

Figure 4

The effect of direct MLCK inhibition and calcium/calmodulin-dependent MLCK inhibition on LSEC fenestrations. (A) A schematic presentation of the ML-7 effect. MLC is partially dephosphorylated (faded-red colour) by direct MLCK inhibition. Phosphorylation by Rho/ROCK pathway still occurs. (B) A representative SEM image of LSEC treated with 20 µM of ML-7. The inset highlights a small sieve plate with enlarged fenestrations. (C) Fenestration diameter distribution (fenestrations measured: ctrl-22618, 1 µM ML-7-8283, 5 µM ML-7-12560, 10 µM ML-7-4630, 20 µM ML-7-870, 10 µM ML-7 (rinsed)-6607). The proportion of fenestrations > 200 nm is 7% (ctrl), 7% (1 µM ML-7), 15% (5 µM ML-7), 24% (10 µM ML-7), 38% (20 µM ML-7), and 14% (10 µM ML-7, rinsed). (D) Changes in the porosity relative to the control are presented (n = 3–6). MLCK inhibition by ML-7 resulted in a significant and dose-dependent decrease in porosity; the effect is not reduced by CytB. (E) A schematic presentation of calcium/calmodulin-dependent inhibition of MLCK. MLCK activation by PKC and PKG still occurs. MLC phosphorylation remains unchanged. (F) Representative SEM image of LSEC treated with 20 µM KN93+CytB. The inset highlights fenestrations gathered in the sieve plate, showing a similar response to the effect of CytB alone (presented SEM images are in the same scale). (G) Fenestration size distribution after calcium/calmodulin inhibition (number of fenestrations measured: ctrl-6341, KN93-5306, KN93+CytB-5830, cytB-35677). The proportion of fenestrations > 200 nm is similar to the control (ctrl-14%, KN93-14%, KN93+CytB-11%, CytB-4%). (H) A change in porosity relative to the control is presented (n = 3). The calcium/calmodulin inhibition resulted in a significant decrease in porosity; the effect of KN93 does not hamper the increase in porosity caused by CytB. (I) A fluorescence signal for pMLC and ppMLC measured using SIM for individual well-spread (fenestrated) LSEC. * p < 0.01, relative to the control.

Figure 5

The effect of myosin heavy chain inhibition by 20 µM blebbistatin. (A) Schematic presentation of the blebbistatin effect—myosin cannot exert a force on actin, as the myosin motor domain is blocked. The use of blebbistatin uncouples the dependence of pMLC activation on actomyosin contraction. (B) Representative SEM image of LSEC treated with 20 µM blebbistatin. (C) Fenestration diameter distribution (number of fenestrations measured: ctrl-5402, 20 µM blebbistatin-26017). The proportion of fenestrations > 200 nm is 12% (ctrl) and 13% (20 µM blebbistatin). (D) A change in porosity relative to the control is presented (n = 2). (E) A fluorescence signal for pMLC and ppMLC measured using SIM for individual well-spread (fenestrated) LSEC.

Figure 6

SIM projection images of a representative LSEC (control, untreated cells). Myosin VIIa/MYO7A rabbit antibody (1:100), secondary donkey anti-rabbit, Alexa 488 (1:100), cell membrane-CellMask Orange, actin-phalloidin-Atto647N. Image size: 40.96 µm × 40.96 µm.

Figure 7

A diagram showing the effect of selected inhibitors on the regulation of MLC phosphorylation and the observed effect on fenestrations in LSEC. n.c.—no change, ↓—fenestration number (fenestr. no.)/diameter (fenestr. diam.) decrease, ↑—fenestration number/diameter increase.

Figure 8

Schematic representation of the action of selected inhibitors used to test the influence of phosphorylation of regulatory myosin light chain (MLC) on fenestrations in LSEC. Phosphorylated MLC (pMLC, glowing red) activates the myosin motor domain that exerts a contractile force on actin. The contractile force generation on actin can be inhibited by blebbistatin independently of MLC phosphorylation. Cytochalasin B (CytB) disturbs actin fibres and therefore weakens the contractile force of the actomyosin complex. KN93, ML-7, and Y27632 inhibit MLC phosphorylation in different ways. Calyculin A (CalA) inhibits MLC phosphatase (MLCP) and prevents MLC dephosphorylation. At the bottom of the figure, it is also highlighted that higher levels of MLC phosphorylation (monophosphorylation (pMLC) and diphosphorylation (ppMLC)) are associated with more prominent actin stress fibres.