During sepsis, endothelial barrier dysfunction contributes to cardiovascular failure, mainly through the release of oxidative metabolites by penetrant leukocytes. We reported the non-muscular isoform of myosin light chain kinase (nmMLCK) playing a pivotal role in endotoxin shock injury associated with oxidative and nitrative stresses, and vascular hyporeactivity. The present study was aimed at understanding the molecular mechanism of lipopolysaccharide (LPS)-induced vascular alterations as well as studying a probable functional association of nmMLCK with nuclear factor κ-light-chain enhancer of activated B cells (NF-κB). Aortic rings from mice were exposed in vitro to LPS and, then, vascular reactivity was measured. Human aortic endothelial cells (HAoECs) were incubated with LPS, and interaction of nmMLCK with NF-κB was analysed. We provide evidence that nmMLCK deletion prevents vascular hyporeactivity induced by in vitro LPS treatment but not endothelial dysfunction in the aorta. Deletion of nmMLCK inhibits LPS-induced NF-κB activation and increases nitric oxide (NO) release via induction of inducible NO synthase (iNOS) within the vascular wall. Also, removal of endothelium prevented both NF-κB and iNOS expression in aortic rings. Among the proinflammatory factors released by LPS-treated endothelial cells, interleukin-6 accounts for the induction of iNOS on smooth muscle cells in response to LPS. Of particular interest is the demonstration that, in HAoECs, LPS-induced NF-κB activation occurs via increased MLCK activity sensitive to the MLCK inhibitor, ML-7, and physical interactions between nmMLCK and NF-κB. We report for the first time on NF-κB as a novel partner of nmMLCK within endothelial cells. The present study demonstrates a pivotal role of nmMLCK in vascular inflammatory pathologies.
As nmMLCK plays a pivotal role in endotoxin shock injury associated with oxidative and nitrative stresses and vascular hyporeactivity, we decided to achieve better understanding of the molecular mechanism of LPS-induced vascular alterations and nmMLCK, using nmMLCK−/− mice.
Our study revealed that nmMLCK deletion prevents vascular hyporeactivity induced by in vitro LPS treatment but not endothelial dysfunction in the aorta. In addition, we found that deletion of nmMLCK inhibits NF-κB activation and the increase of NO release via induction of iNOS within the vascular wall.
We provide evidence that in human aortic endothelial cells, lipopolysaccharide-induced NF-κB activation occurs via physical interactions between nmMLCK with NF-κB, suggesting that nmMLCK is a potential therapeutic target for treatment of septic shock.
During severe sepsis, the cardiovascular system adopts a high cardiac output–low peripheral resistance haemodynamic profile with a vascular component involving dilatation of resistance arteries . In the long term, sepsis is associated with multiple organ failure as a consequence of the combination of redistribution of blood, microvascular failure, constriction, obstruction and permeability changes, leading to failure of oxygen delivery . Data collected from clinical studies and different models of endotoxaemia suggest that lipopolysaccharide (LPS) belongs to the most toxic constituent of bacterial endotoxin, a cell-associated product of Gram-negative bacteria that is capable of causing fever, shock and organ injury in mammals . The effects of LPS are achieved mainly through activation of toll-like receptor 4 and subsequent promotion of signalling of nuclear factor κ-light-chain enhancer of activated B cells (NF-κB), enabling the expression of several critical genes involved in the pathogenesis of septic shock including inducible nitric oxide synthase (iNOS), leading to an overproduction of nitric oxide (NO) . The latter plays a major role in vascular hyporeactivity in different animal blood vessels and in small vessels from patients with septic shock. Under these circumstances, NO can be generated from either macrophages or vascular cells.
Among vascular cells, the endothelium is the primary target of LPS-induced vascular inflammation. The release of proinflammatory cytokines by LPS through NF-κB activation modulates, in vivo, the opening of tight junctions by increasing intracellular calcium, leading to endothelial cell contractility, formation of intracellular gaps through rearrangement of the actin cytoskeleton and profound vascular leakiness . There is a great deal of evidence to show that the non-muscular isoform of myosin light chain kinase (nmMLCK) plays a significant role in the maintenance of endothelial barrier function by controlling the permeability of tight junctions and leukocyte transmigration [6,7].
When compared with smooth muscle MLCK (MLCK108), nmMLCK differs both structurally and functionally. Wainwright et al.  have shown that mice knocked out for nmMLCK, but still expressing MLCK108, are less susceptible to endotoxin-induced acute lung injury. In addition, this strain shows enhanced survival during subsequent mechanical ventilation, suggesting a role for nmMLCK in pulmonary inflammation. We reported that deletion of nmMLCK enhanced survival after either intraperitoneal injection of LPS or caecal ligation puncture. LPS-induced vascular hyporeactivity to vasoconstrictor agents was completely prevented in aortas from nmMLCK−/− mice. This was associated with a decreased up-regulation of NF-κB expression and activity, iNOS expression and level of oxidative stress in the vascular media. Furthermore, an LPS-induced increase of NO production in the circulation and tissues (including heart, liver and lung), which correlated with an increased expression of iNOS, was also reduced in nmMLCK−/− mice .
The present study was aimed at better understanding of the molecular protection mechanism of nmMLCK deletion against LPS-induced vascular alterations. Changes in vascular reactivity on LPS stimulation were conducted in vitro to exclude the contribution of in vivo circulating parameters responsible for leukocyte infiltration and/or cytokine and chemokine release. It has been reported that nmMLCK contains amino acid sequence motifs associated with subcellular targeting or protein–protein interactions in the proteome [10,11]. This domain of the enzyme plays a role as a cellular organizer, providing integration among diverse proteins, including cytoskeletal proteins . The hypothesis of a probable functional association of MLCK, including nmMLCK, with proteins belonging to the NF-κB family, which might be responsible for vascular hyporeactivity, was also tested.
All animal experimentation was performed in accordance with institutional guidelines and protocols were approved by the French Animal Care Committee in accordance with European regulations (CEEA.PdL 2012.94). This study was performed in male 10- to 12-week-old nmMLCK+/+ and nmMLCK−/− mice, generated as previously described . Before any sacrifice and organ dissections, mice were anaesthetized with ketamine (100 mg/kg intraperitoneally) mixed with medetomidine (50 μg/kg intraperitoneally).
Aortic ring preparation
Thoracic aortas from nmMLCK+/+ and nmMLCK−/− mice were dissected, and cleaned of fat and connective tissue. Each aorta was cut into two segments; the endothelial layer was mechanically removed in one part and then cut into segments of 2 mm. Aortic rings were then incubated in a medium of RPMI (Roswell Park Memorial Institute)/Ham's F12 (1:1), 50 units/ml of penicillin, 50 units/ml of streptomycin, 4 mM glutamine (Lonza) supplemented with 10% FBS with or without LPS (E. coli, serotype 055:B5, 100 μg/ml, Sigma-Aldrich) for 20 h.
After 20 h of incubation mice aortas were mounted on a wire myograph (Aarhus). Mechanical activity was recorded isometrically using a force transducer. The experiments were performed at 37°C with physiological saline of the following millimolar composition (NaCl 130, NaHCO3 14.9, KCl 3.7, KH2PO4 1.2, MgSO4·7H2O 1.2, CaCl2·H2O 1.6, glucose 11) continuously bubbled with 95% O2 and 5% CO2. The functionality of the endothelium was assessed by the ability of acetylcholine (Sigma-Aldrich) to induce relaxation as previously described . Concentration–response curves were constructed by cumulative application of 5-hydroxytryptamine (5-HT or serotonin, Sigma-Aldrich) to vessels with and without functional endothelium in the absence or presence of the NO synthase inhibitor, L-nitro-arginine (L-NA, Sigma-Aldrich).
In another set of experiments, arteries with endothelium were precontracted to 80% of maximal contraction with the thromboxane A2 analogue, 9,11-dideoxy-9α,11α-methaneopoxy prostaglandin F2α (PGF2α) (U-46619, Merck Chemicals Ltd). When stable contraction was reached, endothelium-dependent relaxation was assessed by cumulative addition of acetylcholine in order to construct a concentration–response curve.
NO spin trapping and electronic paramagnetic resonance studies
Aortas taken from nmMLCK+/+ and nmMLCK−/− mice were harvested and incubated with or without LPS (100 μg/ml) for 20 h at 37°C in a culture medium composed of RPMI/Ham's F12 (1:1), 50 units/ml of penicillin, 50 units/ml of streptomycin, 4 mmol/l glutamine and 10% FBS (Lonza). Detection of NO production was performed using a previously described technique with ferrous diethyldithiocarbamate (DETC, Sigma-Aldrich) as a spin trap . Vessels were placed in 24-well clusters filled with 250 μl of Kreb's solution, and then treated with 250 μl of colloid Fe(DETC)2 and incubated (37°C, 1 h). These studies were performed on a tabletop x-band spectrometer Miniscope (Magnettech). Recordings were made at 77K, using a Dewar flask. Instrument settings were 10 mW of microwave power, 1 mT of amplitude modulation, 100 kHz of modulation frequency, 60 s of sweep time and 5 scans.
Staining and imaging by confocal microscopy
After fixation, aorta sections with endothelium were incubated (2 h, room temperature) in blocking buffer (5% non-fat dry milk in physiological buffer solution). After three washes, tissue sections were incubated overnight (4°C) with monoclonal iNOS antibody (1:100, Transduction Laboratories). Three washes were followed by incubation (1 h, room temperature) with secondary mouse fluorescence-labelled antibody Fluoprobes FP 546 (1:100, Interchim).
In another set of experiments, thoracic aortas were dissected from nmMLCK+/+ and nmMLCK−/− mice, and cleaned of fat and connective tissue. The thoracic aortas were then cut into two segments and the endothelium removed in one of them. Aortic rings were then incubated in the medium RPMI/Ham's F12 (1:1), 50 units/ml of penicillin, 50 units/ml of streptomycin, 4 mM glutamine supplemented with 10% FBS with or without LPS for 20 h, and then iNOS and NF-κB labelling (1:500, Abcam) was performed as described above.
Western blot analysis
After LPS treatment, crushed aortas with intact endothelium were homogenized and lysed. 40 μg of protein was separated by 4–12% gel electrophoresis. In addition, protein lysates from human aortic endothelial cells (HAoECs) (40 μg) were subjected to SDS/PAGE using 4–12% gels. After electrophoresis, proteins were transferred to nitrocellulose membranes and probed with rabbit polyclonal antibody for iNOS (1:1000, Santa Cruz Biotechnology), endothelial NOS (eNOS, 1:1000, BD Pharmingen), anti-phospho-eNOS (p-eNOS) Ser 1177 (1:1000, Cell Signaling), NF-κB (1:1000, Abcam), phosphorylated p65 subunit of NF-κB (1:1000, Santa Cruz Biotechnology), inhibitor of κB (IκBα; 1:1000, US Biological) and phosphorylated IκBα (1:1000, US Biological). After washing, bound antibodies were detected with a secondary peroxidase-conjugated anti-rabbit antibody (1:20 000) or anti-mouse antibody (1:5000). The same membrane was used to determine β-actin or α-tubulin expression using a polyclonal antibody (1:5000, Sigma-Aldrich). The bands were visualized using the enhanced chemiluminescence system (ECL) and quantified by densitometry.
Culture of human aortic endothelial cells
The primary HAoECs (PromoCell GmbH) were cultured in Endothelial Growth Medium MV2 containing SupplementMix provided by PromoCell. At 90% of confluence, the medium was removed and the cells were treated with LPS (10 μg/ml) in the absence or presence of ML-7 (5 μM) for 3 h.
Anti-nmMLCK antibody (5 μl/400 μl; Santa Cruz Biotechnology) was incubated with protein A/G Sepharose beads (Santa Cruz Biotechnology) at 4°C for 2 h to form a beads–antibody complex. Then, HAoEC protein extracts were added at 4°C overnight to allow antibody recognition of MLCK. Subsequently, the tubes containing the immunoprecipitates were centrifuged and the supernatants aspirated. Pellets were washed twice with 100 μl of lysis buffer diluted to 1:5 with PBS. Gel-loading buffer (15 μl) was then added to the pellets. The tubes were heated for 10 min at 70°C, after which MLCK proteins were separated on SDS/4–12% PAGE. The gel was transblotted to perform Western blot analysis for nmMLCK and NF-κB proteins.
Electrophoretic mobility shift assays
Electrophoretic mobility shift assay (EMSA) was used to determine the activation of the p65 subunit of NF-κB in HAoECs untreated or treated with LPS (10 μg/ml, 3 h) in the absence and presence of the MLCK inhibitor, ML-7 (5 μM). Briefly, the reaction mixtures (20 μl) containing 10 μg of nuclear extracts were incubated with 100 nM double-stranded oligonucleotide probes in reaction buffer for 20 min at room temperature. Samples were subjected to electrophoresis in 4% non-denaturing PAGE. At the end, the gel was incubated with Sybr Green diluted in TBE buffer (89 mM Tris base, 89 mM boric acid, 1 mM EDTA, pH 8.0), for 20 min at room temperature. After two washes with water, the stain was visualized using UV on the imager Infinity VX2 (Vilber Lourmat). Histograms show densitometric analysis of band shifts. The oligonucleotide probes used with NF-κB fixation domain were: TTGTTACAAGGGGACTTTCCGCTGGGGACTTTCCAGGGAGGC.
Evaluation of cytokines and chemokines released by HAoECs
Cytokine and chemokine secretions were measured in the supernatants obtained after 20 h of treatment with LPS of HAoECs using the Human Inflammation Antibody Array 3 (RayBiotech Inc.), which detects 40 cytokines and chemokines simultaneously (see Supplementary Figure S1). Array membranes were processed following the manufacturer's recommendations. The signal intensity was measured on the Vilber Lourmat imager, and the resulting images were analysed using Image J software. To compare the luminescence intensities among the samples, the background staining was subtracted and the data were normalized to the positive controls on the same membrane.
Culture and stimulation of human aortic smooth muscle cells
Primary human aortic smooth muscle cells (Gibco) were cultured in Medium 231 (Gibco) containing Smooth Muscle Growth Supplement (Gibco) and 1% penicillin/streptomycin. Human aortic smooth muscle cells were stimulated for 20 h in the presence of TAK-242 (1 μM, Merck Chemicals Ltd), with conditioned medium from HAoECs treated for 20 h with LPS (10 μg/ml) in the absence or presence of ML-7 (5 μM). TAK-242, a toll-like receptor inhibitor, was used to avoid the direct effect of LPS on smooth muscle cells. Then, proteins were extracted and protein lysates (15 μg) were subjected to Western blot analysis for iNOS expression as indicated above. In another set of experiments, human aortic smooth muscle cells were pre-incubated with either an interleukin-6 (IL-6) receptor-neutralizing antibody (0.4 μg/ml; R&D Systems) or a control isotype immunoglobulin G (IgG) at the same concentration for 30 min, then cells were treated with conditioned medium as described above.
Results are expressed as means±S.E.M.s for the number n. For animal experiments, n represents the number of mice. Vascular reactivity was compared using a two-way ANOVA with repeated measurements. Unpaired Student's t-tests were used for Western blots. The levels of NO were compared using one-way ANOVA followed by a Newman–Keuls multiple comparison post-hoc test. In all cases P<0.05 was considered to be significant. The differences are illustrated as such (*P<0.05 and **P<0.01).
Deletion of nmMLCK does not protect against endothelial dysfunction induced by
in vitro LPS treatment
Acetylcholine produced a concentration-dependent relaxation in the aorta with functional endothelium from both nmMLCK+/+ (Figure 1A) and nmMLCK−/− (Figure 1B) mice. After incubation with LPS (100 μg/ml, 20 h), relaxation in response to acetylcholine was significantly decreased in aortic rings from either nmMLCK+/+ or nmMLCK−/− mice compared with aortic rings of the two strains incubated with the medium alone (Figure 1).
Deletion of nmMLCK does not protect against endothelial dysfunction induced by
in vitro LPS treatment
Deletion of nmMLCK prevents vascular hyporeactivity induced by LPS incubation
Concentration–response curves to 5-HT for nmMLCK+/+ (Figure 2A) and nmMLCK−/− (Figure 2B) aortic rings, with functional endothelium incubated with medium, were indistinguishable. Although LPS exposure decreased vascular reactivity to the agonist in aortic rings taken from nmMLCK+/+ mice (Figure 2A), LPS incubation did not affect the response to 5-HT in vessels from nmMLCK−/− mice (Figure 2B). Thus, the deletion of nmMLCK completely prevented the vascular hyporeactivity induced by LPS incubation in aortic rings with endothelium. In contrast, LPS incubation did not significantly modify the response to 5-HT in aortas with no functional endothelium, taken from either nmMLCK+/+ (Figure 2C) or nmMLCK−/− (Figure 2D) mice. To investigate the mechanisms involved in vascular hyporeactivity induced by LPS, the involvement of NO was evaluated by studying the effect of the NO synthase inhibitor L-NA in mice aortic rings in response to 5-HT. L-NA significantly increased the contraction in response to 5-HT in arteries from nmMLCK+/+ mice incubated with LPS towards the response obtained in aortas not incubated with LPS (Figure 2E). In contrast, L-NA did not significantly alter the response to the same agonist in nmMLCK−/− aortic rings treated with LPS (Figure 2F).
Deletion of nmMLCK prevents vascular hyporeactivity induced by LPS incubation
It should be noted that LPS treatment did not modify contraction induced by potassium chloride depolarization, U46619 or their combination, in vessels from the two strains, independent of the presence or absence of functional endothelium (see Supplementary Figure S2).
Deletion of nmMLCK prevents LPS-induced increase of NO production and iNOS expression in the aorta
For a direct demonstration of the link between the potentiating effect of L-NA in aortic rings from nmMLCK+/+ mice incubated with LPS and increased production of NO measurement by electron paramagnetic resonance (EPR) was conducted. As shown in Figure 3(A), LPS-treated aortic rings from nmMLCK+/+ mice exhibited a marked increase of NO production compared with controls. In contrast, LPS incubation could not significantly enhance NO production in aortas from nmMLCK−/− mice when compared with control aortas taken from either nmMLCK+/+ or nmMLCK−/− mice.
Deletion of nmMLCK prevents LPS-induced increase of NO production, iNOS expression and NF-κB activation in the aorta
As L-NA and EPR cannot discriminate all NOS enzyme contributions, immunostaining of iNOS was conducted. LPS induced marked staining of iNOS in the medial and adventitial layers of aortas taken from nmMLCK+/+ mice, but not in those taken from nmMLCK−/− mice (Figure 3B). These results are consistent with the hypothesis that deletion of nmMLCK reduces LPS-induced increase of NO production and iNOS expression in the aorta.
Deletion of nmMLCK prevents LPS-induced NF-κB activation
Stimulation of the LPS receptor, toll-like receptor 4, induces NF-κB activation. The NF-κB family members are heterodimers of p65/RelA and p50/NF-κB, but only the p65 subunit has transactivation domains capable of initiating transcription on proinflammatory genes that code for various cytokines and enzymes such as iNOS. The latter is responsible for the overproduction of circulating and tissue NO, including by the aorta, and is involved in vascular hyporeactivity to vasoconstrictor agonists after LPS treatment. Therefore, we assessed the effect of the deletion of nmMLCK on LPS-induced NF-κB activation within the aorta. For this purpose, Western blotting of the p65 subunit of NF-κB was used as an index of activation of NF-κB. Also, Western blotting of phosphorylated IκBα was conducted to confirm NF-κB activation. LPS induced a significant increase of the p65/RelA subunit of NF-κB expression in aortas taken from nmMLCK+/+ mice but not in those from nmMLCK−/− mice (Figure 3C). Furthermore, LPS treatment significantly increased the expression of p-IκBα in vessels from nmMLCK+/+ mice (Figure 3D). IκBα phosphorylation was greater in aortas from nmMLCK−/− compared with nmMLCK+/+ mice. It is of interest that LPS incubation did not produce a further increase in phosphorylation of IκBα in aortas from nmMLCK−/− mice. These results are consistent with the hypothesis that deletion of nmMLCK reduces LPS-induced NF-κB activation in vitro.
To establish whether the endothelium plays a crucial role in NF-κB and iNOS activation by LPS, we analysed the expression of these two proteins in endothelium-intact and endothelium-denuded aortic rings from nmMLCK+/+ mice. As shown in Figures 4(A)–4(C), LPS induced an increase in both NF-κB and iNOS expression in vessels with endothelium. In addition, NF-κB expression was observed in the intimal layer of the aorta (Figure 4A), whereas iNOS expression was increased in the media and adventitia (Figures 4B and 4C). Furthermore, in endothelium-denuded aortic rings, LPS failed to induce NF-κB and iNOS expression. These results suggest that activation of the NF-κB pathway takes place at the endothelium, and the subsequent induction of iNOS at the smooth muscle and adventitial layers is probably due to the release of proinflammatory factors from the endothelium.
LPS fails to induce NF-κB and iNOS expression in endothelium-denuded aortic rings
Then, to verify this hypothesis we next incubated human smooth muscle cells with the conditioned medium of HAoECs treated with LPS and iNOS activation was evaluated. In addition, to avoid the possibility that LPS used to treat HAoECs directly activates smooth muscle cells, these cells were treated with the toll-like receptor 4 inhibitor TAK-242 (1 μM). As illustrated in Figure 5(A), conditioned medium from LPS-treated HAoECs increased iNOS expression in human smooth muscle cells. Pharmacological inhibition of nmMLCK activity with ML-7 did not significantly affect iNOS expression in both control and LPS-treated endothelial cells. Next, we identified the factors secreted by HAoECs in the absence or presence of LPS. LPS treatment enhanced secretions of IL-6 (Figure 5B), regulated on activation, normal T-cell expressed and secreted (RANTES), C-X-C motif chemokine (CXCL-10), granulocyte–macrophage colony-stimulating factor (GM-CSF), monocyte chemotactic protein 1 (MCP-1) and chemokine (C-C motif) ligand 1 (CCL1) (see Supplementary Figure S3). To verify whether IL-6 is able to induce iNOS, we analysed the effects of the conditioned medium from LPS-treated HAoECs on vascular smooth muscle cells in the presence of a neutralizing IL-6 receptor antibody (Figure 5C). Under these conditions, conditioned medium failed to increase iNOS expression, suggesting that released IL-6 from LPS-treated HAoECs accounts for the increase of iNOS on smooth muscle cells. Altogether, these results indicate that LPS activates endothelial cells to release proinflammatory mediators, including IL-6, leading to the induction of iNOS in vascular smooth muscle cells.
Secreted IL-6 from human aortic endothelial cells in response to LPS induces iNOS in human vascular smooth muscle cells
Inhibition of MLCK activity prevents LPS-induced increase in NF-κB activation
The role of MLCK activity in NF-κB-dependent transactivation was investigated in HAoECs activated by LPS using the MLCK-selective inhibitor, ML-7. Stimulation of the LPS receptor, toll-like receptor 4, induces NF-κB activation, for which NF-κB requires the degradation of IκB proteins; this, in turn, depends on IκB phosphorylation and ubiquitin-dependent proteasome degradation . Therefore, we explored phosphorylated NF-κB (p-p65) and p-IκBα proteins in HAoECs exposed to LPS. HAoECs treated with LPS (10 μg/ml, 3 h) exhibited increased phosphorylation of p65 (Figure 6A) and IκBα (Figure 6B). ML-7 alone did not significantly modify phosphorylation of p65 and IκBα. It is interesting that LPS treatment was not able to increase the phosphorylation of both proteins in the presence of ML-7.
Inhibition of MLCK activity prevents LPS-induced increase in NF-κB activation
nmMLCK interacts physically with NF-κB
It is known from the literature that nmMLCK can make protein–protein interactions with such proteins as the macrophage migration inhibitor factor (MIF) . This interaction was found in endothelial cells of the bovine pulmonary artery treated with tumour necrosis factor α (TNF-α) and is time dependent. It has been reported that signal transducer and activator of transcription 5 (STAT5) negatively regulates NF-κB activation of MLCK to modulate MLC phosphorylation, which protects intestinal epithelial cell barrier function to promote wound healing . On the basis of these previous studies, we examined the possibility that interaction of nmMLCK with NF-κB contributes to NF-κB activation, and further experiments were conducted on HAoECs treated or not treated with LPS (10 μg/ml, 3 h). After incubation of HAoECs with or without LPS, the cells were placed in lysis buffer and protein fractions were obtained. The nmMLCK represented 90% of the total MLCK in the HAoEC extract. Potential NF-κB–nmMLCK complexes were co-immunoprecipitated from lysates with anti-MLCK antibody, and then the proteins were visualized using the anti-NF-κB antibody. However, potential NF-κB–nmMLCK complexes were co-immunoprecipitated from lysates with antibody anti-NF-κB, after immunoblotting with the anti-MLCK. It is interesting that NF-κB was found in the immunoprecipitated extract from control HAoECs and those treated with LPS (Figure 7A). Conversely, nmMLCK co-immunoprecipitated with the p65 subunit of NF-κB (Figure 7B). These results demonstrate for the first time a protein–protein interaction between nmMLCK and NF-κB. Finally, an immunoprecipitation study revealed that the physical association between nmMLCK and NF-κB was not modified in the presence of ML-7, suggesting that such an association was not controlled by a change in nmMLCK activity (Figures 7A and 7B).
nmMLCK interacts physically with NF-κB
We provide evidence that nmMLCK deletion prevents vascular hyporeactivity but not endothelial dysfunction induced by in vitro LPS treatment. Deletion of nmMLCK inhibits NF-κB activation and the increase of NO release via induction of iNOS within the vascular wall. In addition, NF-κB activation at the endothelial level induces release of proinflammatory factors, including IL-6, that act on smooth muscle cells, leading to iNOS expression. Of particular interest is the demonstration that LPS-induced NF-κB activation occurs via an increase in MLCK activity and a physical interaction between nmMLCK and NF-κB. We reported that deletion of nmMLCK enhanced survival after either intraperitoneal injection of LPS or caecal ligation, puncture-induced death . In addition, vascular hyporeactivity induced by in vivo injection of LPS is completely prevented in aortas from nmMLCK−/− mice. In the present study, we wanted to ensure that the protective effect of nmMLCK deletion occurs directly on the vascular wall, independent of the contribution of in vivo circulating parameters responsible for leukocyte infiltration and/or cytokines, and chemokine release when LPS was intraperitoneally injected. This is particularly important inasmuch as a lung injury model for nmMLCK is essential for neutrophil transmigration independent of the myosin II regulatory light chain, the only known substrate of nmMLCK.
In this case, nmMLCK-mediated activation of β2-integrins through Pyk2 links β2-integrin signalling to the actin motile machinery of neutrophils . The data underscored two major points: deletion of nmMLCK did not affect endothelial dysfunction but prevented vascular hyporeactivity in response to 5-HT in aortas incubated with LPS. The mechanisms involved in LPS-mediated impairment of endothelium-dependent vasodilatation include inhibition of eNOS enzymatic activity and/or down-regulation of eNOS expression . In cultured endothelial cells, LPS and various cytokines decrease agonist-mediated NO release . In the present study, LPS treatment did not affect eNOS expression, but it decreased eNOS activation as shown by its ability to reduce eNOS phosphorylation at the activated site (see Supplementary Figure S4). In our previous study, we found that nmMLCK plays a role in the release of endothelial factors (i.e. NO and cyclo-oxygenase products) in response to flow in small mesenteric arteries, but the enzyme is not involved in endothelium-dependent vasodilatation in the aorta . Therefore, it is expected that deletion of nmMLCK would not have an impact on endothelial dysfunction induced by LPS in terms of the correction of reduced endothelial NO release in aortas.
With regard to vascular reactivity, in vitro incubation of aortas from nmMLCK+/+ mice with LPS reduced contraction in response to 5-HT, but not potassium chloride, U46619 or their combination, in the presence, but not the absence, of functional endothelium. This indicates that, under these conditions, LPS did not induce hyporeactivity in response to vasoconstrictor agents using signalling pathways different to those of 5-HT. Furthermore, vascular hyporeactivity to 5-HT was less pronounced than that obtained in response to phenylephrine in aortas taken from mice treated with an intraperitoneal injection of LPS in vivo . The difference is probably related to the effect of in vivo circulating parameters responsible for leukocyte infiltration and/or cytokine and chemokine release. Of note is the fact that the presence of functional endothelium was essential to observe vascular hyporeactivity under the experimental conditions used. Furthermore, in endothelium-denuded aortic rings, LPS failed to induce NF-κB activation and the subsequent iNOS induction in the media layer. These data suggest that vasodilator factors released from the endothelium are implicated in the inflammatory response associated with NF-κB activation and NO release via induction of iNOS within the vascular wall on LPS treatment. Several proinflammatory factors are released by LPS-treated HAoECs, including IL-6, RANTES, CXCL-10, GM-CSF, MCP-1 and CCL1. This is reinforced by the fact that conditioned medium from HAoECs treated with LPS induced iNOS in cultured human vascular smooth muscle cells. In addition, blockade of the IL-6 receptor on human vascular smooth muscle cells abrogates conditioned medium-evoked induction of iNOS. It is of interest that deletion of nmMLCK prevented LPS-induced vascular hyporeactivity in vivo  and in vitro (the present study). As nmMLCK is mainly present in the endothelium, one can advance the hypothesis that its deletion abrogates the release of endothelial factors (cytokines or chemokines) that promote impairment of vascular contraction.
The mechanism by which nmMLCK interacts with pathway(s), leading to the release of such factors as NF-κB and its related protein, IκBα, was investigated in HAoECs. As expected, nmMLCK was expressed in HAoECs, and LPS induced an increased expression of the p65 subunit of NF-κB and the phosphorylation of IκBα, as an index of NF-κB activation. NmMLCK is recognized as a highly multifunctional protein and it serves as a critical regulator of the paracellular space, and actively regulates the movement of leukocytes and soluble factors between the two compartments [17,22]. Indeed, nmMLCK is known to interact with different binding partners such as actin and cortactin, but also with caspase-3 and the regulatory protein p60Src . With regard to cytokine release, an interaction between endothelial nmMLCK and MIF has been reported and this might provide a link for nmMLCK activation, NF-κB-dependent transcription and E-selectin expression in TNF-α-stimulated bovine lung endothelium . Also, a functional association between poly(ADP)-ribose polymerase 1 (PARP-1) and NF-κB is necessary for LPS-induced NF-κB activation and cytokine release in macrophages . Recently, it has been reported that STAT5 negatively regulates NF-κB activation of the epithelial MLCK to modulate MLC phosphorylation, which protects the function of the intestinal epithelial cell barrier to promote wound healing .
Thus, in the present study, we tested the hypothesis that nmMLCK might be functionally associated with an NF-κB complex in endothelial cells. Co-immunoprecipitation experiments demonstrate that interactions between nmMLCK and NF-κB occur in HAoECs in non-treated cells and after LPS stimulation. Deletion of nmMLCK alters these interactions and might participate in the reduced capacity of the endothelium to release factors implicated in vascular hyporeactivity on LPS treatment.
In addition to the functional associations between nmMLCK and NF-κB, changes in nmMLCK activity might also be involved for LPS-dependent endothelial NF-κB activation. Such an effect has been demonstrated in bovine pulmonary endothelial cells for TNF-α . Wadgaonkar et al.  found that TNF-α-induced cytoskeletal rearrangement, driven by nmMLCK activity, is necessary for TNF-α-dependent NF-κB activation and amplification of pro-survival signals. The most recent report demonstrates that TRPC6-Ca2+ signalling-dependent activation of nmMLCK in endothelial cells is required to increase both lung vascular permeability and interaction of MyD88 and IRAK4 components of LPS/toll-like receptor 4 signalling. This machinery, being required for NF-κB activation, leads to inflammation, the two key features of acute lung injury . Also, it has recently been shown that nmMLCK controls nuclear translocation and transcriptional activity of p65 NF-κB in endothelial cells . It is interesting that, in the present study, the ability of LPS to increase phosphorylation of the p65 subunit of both NF-κB and IκBα was prevented in the presence of the MLCK inhibitor, ML-7. These data indicate that the mechanism by which LPS promotes NF-κB activation is regulated at least partially by the activity of nmMLCK. This effect might partly explain the decreased susceptibility to endothelial injury by LPS in nmMLCK−/− mice, and its consequences on vascular walls such as reduced vascular hyporeactivity or reduced inflammation through increase of iNOS and NO production via NF-κB activation.
In summary, we have defined for the first time specific interactions of nmMLCK with NF-κB, and provided a possible mechanism to explain the protective effect of either nmMLCK deletion or nmMLCK inhibition against the LPS-induced deleterious effect on the endothelium that leads to vascular hyporeactivity and inflammation. The present study and our previous work, showing that nmMLCK is involved in lethal complications , as well as in the vascular reactivity changes associated with endotoxin shock in vivo, demonstrate a pivotal role of this enzyme in vascular inflammatory pathologies. These findings help to design new therapeutic strategies for the treatment of septic shock, based on nmMLCK, allowing the effective down-regulation of oxidative and nitrative stress associated with this life-threatening disease.
S. Recoquillon, N. Carusio, A.-H. Lagrue-Lakhal and S. Tual-Chalot performed the experiments; S. Recoquillon and N. Carusio, who contributed equally to this work, analysed and interpreted the data; S. Recoquillon, A. Filippelli, R. Andriantsitohaina and M.C. Martinez drafted the manuscript; R. Andriantsitohaina and M.C. Martinez conceived of and designed the study.
The authors thank Mireille Wertheimer, and the staff of Service Commun d’Animalerie Hospitalo-Universitaire, for the care of animals. The nmMLCK−/− mice were provided by D.M. Watterson, Center for Drug Discovery and Chemical Biology, Northwestern University, Chicago, IL, U.S.A.
This work was supported by grants from Institut National de Santé et Recherche Médicale and the Agence Nationale de la Recherche (ANR-12-BSV1–0024–01). A.-H.L.-L. is a recipient of a post-doctoral fellowship from Univaloire Angers, and S.R. and S.T.-C. are recipients of doctoral fellowships from the French Education Ministry.
5-hydroxytryptamine or serotonin
chemokine (C-C motif) ligand 1
C-X-C motif chemokine
electrophoretic mobility shift assay
endothelial nitric oxide synthase
electron paramagnetic resonance
granulocyte–macrophage colony-stimulating factor
human aortic endothelial cell
inhibitor of κB
inducible nitric oxide synthase
monocyte chemotactic protein 1
macrophage inhibitory factor
nuclear factor κ-light-chain enhancer of activated B cells
non-muscular myosin light-chain kinase
regulated on activation, normal T-cell expressed and secreted
Roswell Park Memorial Institute
signal transducer and activator of transcription 5
tumour necrosis factor α