TNF-α induces complex signaling events in endothelial cells (ECs), leading to inflammatory gene transcription and junctional permeability increases. This study examined the activation of RhoA and Rho kinase induced by TNF-α in primary human pulmonary microvascular ECs and its role in regulating EC responses to TNF-α. TNF-α induced a time-dependent activation of RhoA and Rho kinase in these ECs. TNF-α also induced activation of JNK that peaked at 15 min and lasted for at least 3 h. Inhibition of Rho kinase using a specific pharmacological inhibitor, Y27632, prevented TNF-α-induced early and late JNK activation. Inhibition of RhoA protein expression using small-interfering RNA, however, did not prevent TNF-α-induced Rho kinase activation or JNK activation. Studies using MAPK kinase 4 (MKK4) small-interfering RNA showed that MKK4 was not required for TNF-α-induced early JNK activation and that Rho kinase modulated early JNK activation through MKK4-independent mechanisms. Rho kinase, however, modulated TNF-α-induced late JNK activation mainly through MKK4-dependent mechanisms. Activation of Rho kinase was required for JNK-dependent IL-6 secretion induced by TNF-α. Moreover, inhibition of Rho kinase prevented TNF-α-induced cytoskeletal changes and permeability increases. Inhibition of JNK activation, however, did not prevent TNF-α-induced cytoskeletal changes, suggesting that Rho kinase did not modulate cytoskeletal changes through JNK activation. Therefore, Rho kinase plays important roles in EC responses to TNF-α by regulating permeability increases and JNK-dependent IL-6 production during pulmonary inflammation.
Acute inflammatory responses in the lung are modulated by local production of early response cytokines such as TNF-α, and pulmonary microvascular endothelial cells (ECs)3 are primary targets of these cytokines. Responses elicited in these ECs by TNF-α include changes in the F-actin cytoskeleton, increases in junctional permeability, and transcriptional up-regulation of multiple inflammatory genes. Transcriptional regulation in these cells by TNF-α is mainly mediated by the activation of two transcription factors: NF-κB and AP-1 (1, 2, 3, 4). NF-κB is rapidly activated as a result of the ligation of TNF-α receptors, leading to transcriptional up-regulation of adhesion molecules and chemokines such as ICAM-1 and IL-8. TNF-α-induced AP-1 up-regulation requires JNK activation (5). The essential role of JNK in modulating pulmonary inflammatory responses is supported by studies showing that inhibition of JNK prevents LPS-induced neutrophil influx and permeability increases in the lung (6, 7).
TNF-α-induced signaling occurs as a result of the ligation of TNF-α receptors, TNFR1 and R2. The majority of TNF-α-mediated signaling occurs through TNFR1, although TNFR2 also signals and mediates some of the responses induced by TNF-α both in vitro and in vivo (8, 9, 10). TNF-α itself is a homotrimer, and binding of TNF-α to TNFR1 results in reorganization of TNFR1, dissociation of the inhibitory protein silencer of death domains from the receptor, and recruitment of complex 1 composed of three adaptor molecules: TNFR-associated death domain, receptor-interacting protein, and TNFR-associated factor 2 (TRAF2) (1, 2, 3, 4). This in turn initiates activation of the NF-κB pathway. Complex 1 then dissociates from the receptor and complex 2—composed of two adaptor molecules Fas-associated death domain (FADD) and caspase 8 recruited to the receptor–initiates proapoptotic signaling pathways. TNF-α also induces activation of MAPK that often exhibit an early and robust activation phase (within 30 min) and a more prolonged and less robust activation phase (for hours). The early and late phase of MAPK activation is regulated by different mechanisms and may have different functions (11), although the mechanisms proximal to the receptor leading to MAPK activation are not completely understood.
Recent studies indicate that TNF-α also induces activation of Rho family small GTPases (12). Rho family small GTPases play important roles in organizing the F-actin cytoskeleton and the junctional proteins. Rho (RhoA, B, C), Rac (Rac1, 2, 3), and Cdc42 are the best-characterized members of the Rho family small GTPases (13, 14, 15). All Rho GTPases exist in either the GDP-bound inactive form or the GTP-bound active form. The activity of Rho GTPases is regulated by Rho guanine nucleotide exchange factors, Rho GTPase-activating proteins, and Rho GDP dissociation inhibitor (GDI). The majority of the Rho GTPases are kept in the inactive state by binding to Rho GDI in the cytosol (14, 15). Rho GDI binds the GDP-bound Rho, Rac, and Cdc42 and prevents their association with the membrane. Dissociation of Rho GDI allows GDP-bound Rho GTPases to become associated with the membrane, where Rho guanine nucleotide exchange factors stimulate the activation of Rho GTPases. Several molecules can displace Rho GDI from GDP-bound Rho GTPases (13). This may represent a mechanism through which Rho GTPases can be delivered to particular sites within a cell in response to a stimulus.
Activation of Rho is required for TNF-α-induced cytoskeletal changes through its downstream targets including Dia 1 and Rho kinase, and inhibition of Rho kinase prevents TNF-α-induced increases in albumin fluxes across human pulmonary artery ECs and inhibits LPS-induced lung permeability increases in mice (16, 17). A major substrate of Rho kinase is myosin phosphatase target subunit (MYPT-1), and phosphorylation of MYPT-1 by Rho kinase leads to inhibition of myosin phosphatase activity, resulting in increased myosin L chain phosphorylation and actin-myosin-mediated contractility (18). Rho kinase activation can also be induced through Rho-independent mechanisms. Binding of lipid messengers, including arachidonic acid or sphingosine phosphorylcholine, to Rho kinase or proteolytic cleavage of Rho kinase by caspase 3 or granzyme B can efficiently activate Rho kinase independent of Rho (18).
In addition to its well-described function in regulating the F-actin cytoskeleton, the Rho/Rho kinase pathway may serve other biological functions. In particular, recent studies demonstrate that several downstream targets of Rho regulate JNK activation, raising the possibility that Rho and/or the Rho kinase pathway may also regulate JNK-dependent responses (19, 20). This study seeks to determine whether the RhoA or Rho kinase pathway regulates inflammatory responses by modulating JNK activation in response to inflammatory cytokines such as TNF-α. More specifically, this study asked: 1) whether RhoA or Rho kinase is activated by TNF-α in human pulmonary microvascular ECs; 2) whether RhoA or Rho kinase regulates TNF-α-induced JNK activation and the underlying mechanisms; 3) whether RhoA or Rho kinase regulates JNK-dependent inflammatory cytokine expression such as IL-6 production.
Materials and Methods
Cells, pharmacological inhibitors, and small-interfering RNA (siRNA)
Human pulmonary microvascular ECs (Cambrex) were grown on dishes precoated with 4 μg/ml fibronectin. Confluent ECs were treated with either vehicle (PBS containing 0.1% BSA) or 20 ng/ml recombinant human TNF-α for 5 min to 24 h by adding vehicle or a stock solution directly to EC cultures (21).
The role of Rho kinase, JNK, or PI3K was examined using specific pharmacological inhibitors (Calbiochem). To evaluate the role of Rho kinase, ECs were pretreated with either vehicle (H2O) or 20 μM Y27632 for 30 min (19, 22, 23, 24). To evaluate the role of JNK, ECs were pretreated with either vehicle (0.1% DMSO) or 40 μM SP600125 for 30 min (25). To evaluate the role of PI3K, ECs were pretreated with either vehicle (0.1% DMSO) or 20 μM LY294002 for 30 min (26).
siRNA targeting MAPK kinase 4 (MKK4), RhoA, or a control siRNA was obtained as a pool of four siRNA duplexes from Dharmacon. ECs were treated for 4 h with 100 nM MKK4 siRNA or 10 nM RhoA siRNA or their corresponding control siRNA premixed with 6 μg/ml lipofectin (21). The cells were then incubated with normal culture medium for 72 h before experiments.
The activity of Rho kinase was evaluated by measuring the phosphorylation of a major Rho kinase substrate, myosin phosphatase targeting subunit 1 (MYPT-1) at threonine 696 and 853 (Upstate Biotechnology) (27, 28, 29). MYPT-1 phosphorylation by Rho kinase decreases the activity of myosin phosphatase, thereby resulting in increased myosin L chain phosphorylation and actin-myosin interaction (30). The activity of JNK, p38, AKT, and MKK4 was evaluated by measuring the amount of phosphorylated and active form of these kinases: phosphorylated JNK (threonine 183/tyrosine 185), p38 (threonine 180/tyrosine 182), AKT (serine 473), and MKK4 (serine 257/threonine 261) (Cell Signaling). Following cell treatment, the proteins were solubilized using SDS buffer, and the level of phosphorylated MYPT-1, JNK, p38, AKT, or MKK4 was examined by immunoblot (21).
Evaluation of the activity of RhoA
Activation of RhoA was evaluated using a pull-down assay kit according to the manufacturer’s instruction (Cytoskeleton). Cells were lysed using a Triton X-100 lysis buffer containing 50 mM Tris (pH 7.5), 10 mM MgCl2, 0.5 M NaCl and 1% Triton X-100. The samples were incubated with rhotekin-Rho-binding domain (RBD)-coated agarose beads for 1 h at 4°C, and the beads were washed two times. The amount of RhoA associated with the beads (GTP-RhoA) and the amount of RhoA left in the supernatant (GDP-RhoA) were examined by immunoblotting using an Ab that recognizes RhoA (Santa Cruz Biotechnology).
Subcellular distribution of RhoA, Rho GDI, F-actin, and vascular endothelial (VE) cadherin
The distribution of RhoA and Rho GDI was evaluated by immunocytochemistry using an Ab that recognizes RhoA (Santa Cruz Biotechnology) and Rho GDI (Santa Cruz Biotechnology). The coverslips were mounted using anti-fade mounting medium (Vector Laboratories). The slides were examined using a Zeiss LSM 510 META scanning confocal microscope and the images were acquired using the manufacturer’s software.
To examine the distribution of VE cadherin and F-actin, cells were fixed with 1% paraformaldehyde and washed. VE cadherin was detected using a mouse anti-human VE cadherin Ab (BD Pharmingen) and a fluorescein-conjugated goat anti-mouse IgG. F-actin was detected using rhodamine-phalloidin (21). The images were acquired using a Zeiss LSM 510 META scanning confocal microscope.
Measurement of EC permeability to fluxes of dextran
The fluxes of dextran across ECs were evaluated using transwell chambers (21). ECs were seeded onto transwell inserts (6.5-mm diameter and 0.4-μm pore size) coated with 4 μg/ml fibronectin. The cells were pretreated with either vehicle or 20 μM Y27632, followed by stimulation with 20 ng/ml TNF-α or buffer. One hour after TNF-α treatment, 250 μg of FITC-conjugated dextran (10 kDa) was added to the top well, and 100-μl aliquots of samples were taken from the bottom chamber every 15 min for 2 h. The amount of dextran in the samples was measured using a fluorescence plate reader and calculated after constructing a standard curve. Each time after taking a sample, 100 μl of medium was added back to the bottom chamber. For each time point, the amount of dextran present in the bottom wells as well as the amount in all the samples removed for measurement was calculated, added up, and plotted. In every experiment, fibronectin-coated filter inserts without ECs were included, and the amount of dextran in the bottom wells in each sample was expressed relative to the amount of dextran in the filter-alone samples at 2 h.
Measurement of IL-6 release
IL-6 release was examined by measuring the amount of IL-6 present in culture supernatant using an ELISA according to the manufacturers’ instructions (eBioscience). After EC treatment, the supernatant samples were collected and added to 96-well plates precoated with an anti-IL-6 capture Ab. After overnight incubation, the wells were incubated with an anti-IL-6 detection Ab conjugated with biotin for 1 h, HRP-conjugated avidin for 30 min, and HRP substrate for 15 min. The reaction was stopped with 1 N H2SO4, and absorbance at 450 nm was measured using an ELISA plate reader. The concentration of the samples was calculated after constructing a standard curve, and all samples were analyzed in duplicates.
Data were analyzed using the Student t test or one-way ANOVA followed by post-hoc comparisons (Least Significant Difference test). A p value <0.05 was considered significant. The data are expressed as the mean value ± SEM.
TNF-α induces activation of RhoA and Rho kinase in human pulmonary microvascular ECs
The activity of RhoA following TNF-α treatment in human pulmonary microvascular ECs was examined using a rhotekin RBD pull-down assay. In confluent and untreated ECs, the majority of total cell RhoA was in the GDP-bound and inactive state (Fig. 1,A). TNF-α induced activation of RhoA in a time-dependent manner, and a significant increase was detected within 5 min (Fig. 1,A). The activity of RhoA then decreased, but a late activation of RhoA was induced 3 h following TNF-α treatment (Fig. 1 A).
The inactive, GDP-bound Rho GTPases are kept in the cytosol complexed with their endogenous inhibitor, Rho GDI (13). The displacement of Rho GDI allows Rho to translocate to the plasma membrane, where activation of Rho occurs (13). The effect of TNF-α on the subcellular localization of RhoA and Rho GDI was thus examined. In cells not treated with TNF-α, RhoA is distributed in the cytoplasm in a pattern similar to Rho GDI. TNF-α treatment for 5 min did not induce changes in the distribution of RhoA and Rho GDI that could be detected by confocal microscopy (Fig. 1,B). By contrast, TNF-α treatment for 3 h induced focal accumulation of RhoA at EC edges, where Rho GDI was absent (Fig. 1 B, arrows).
Active RhoA exerts its effects by binding to its effector molecules such as Rho kinase and inducing their activation. Rho kinase in turn phosphorylates its substrates such as MYPT-1 at threonine 696 and 853. Phosphorylation of MYPT-1 inactivates myosin L chain phosphatase, leading to myosin L chain phosphorylation, cytoskeletal reorganization, and stress fiber formation (30). To determine whether Rho kinase is also activated by TNF-α, the phosphorylation state of MYPT-1 was measured. TNF-α induced a significant increase in the amount of phosphorylated MYPT-1 at threonine 853 (Fig. 2,A), and to a lesser degree, at threonine 696 (Fig. 2,B). This increase was completely blocked by pretreatment with a specific Rho kinase inhibitor, Y27632, indicating that MYPT-1 phosphorylation is mediated by Rho kinase and that the phosphorylation state of MYPT-1 is indicative of Rho kinase activity (Fig. 2).
TNF-α-induced Rho kinase activation is required for JNK activation
Recent studies demonstrate that several downstream targets of Rho, including Rho kinase, may regulate JNK activation (19, 20). Thus, experiments were performed to determine whether TNF-α-induced activation of Rho kinase is required for JNK activation. TNF-α treatment of human pulmonary microvascular ECs induced an early and robust JNK activation that peaked by 15 min and a prolonged and less robust JNK activation that persisted for at least 6 h (Fig. 3, A and B). Pretreatment with 20 μM Y27632 inhibited early JNK activation and completely prevented late phase activation (Fig. 3, A and B). Activation of p38, however, was not affected by pretreatment with Y27632 (Fig. 3, A and B). In addition, TNF-α-induced ICAM-1 expression in these cells was not inhibited by Y27632 (data not shown).
Because Rho kinase activation can occur independent of RhoA, the following experiments were performed to determine whether RhoA is required for TNF-α-induced Rho kinase activation and JNK activation. Treatment with RhoA siRNA prevented RhoA protein expression (Fig. 3,C). RhoA siRNA, however, did not inhibit TNF-α-induced Rho kinase activation or JNK activation (Fig. 3 C). These data indicate that activation of Rho kinase induced by TNF-α does not require RhoA and that Rho kinase activation, but not RhoA, modulates TNF-α-induced JNK activation.
To determine whether this modulation of JNK activation by Rho kinase is specific toward TNF-α-induced signaling, the effect of Y27632 on JNK activation induced by IL-1β was also examined. IL-1β induced JNK activation in a similar time course as TNF-α: IL-1β treatment for 15 min, 30 min, 1 h, 3 h, and 6 h increased JNK phosphorylation by 48.1- ± 8.3-, 8.4- ± 3.8-, 5.1- ± 2.5-, 2.3- ± 0.5-, 1.1- ± 0.4-fold when compared with cells treated with buffer alone (n = 4). IL-1β-induced JNK activation at 15 min or 3 h was also inhibited by Y27632 (Fig. 3 D).
Although multiple MAPKKKs can be activated by TNF-α leading to JNK activation, MKK4 and MKK7 are the only two MAP kinase kinase kinase (MKKs) leading to JNK activation. In pulmonary microvascular ECs, TNF-α induced time-dependent MKK4 activation, which could be completely blocked by pretreatment with Y27632 (Fig. 4 A).
To determine whether Rho kinase-regulated JNK activation is dependent on MKK4, ECs were treated with MKK4 siRNA to inhibit MKK4 expression, and the effect of Y27632 on JNK activation was examined. The MKK4 siRNA blocked MKK4 protein expression, and prevented MKK4 activation induced by TNF-α (Fig. 4,B). Interestingly, TNF-α-induced early JNK activation (by 15 min) was not inhibited by MKK4 siRNA, and treatment with Y27632 similarly inhibited early JNK activation even in the presence of the MKK4 siRNA (Fig. 4, B and C). However, TNF-α-induced late JNK activation (by 3 h) was largely inhibited by the MKK4 siRNA, indicating that late JNK activation required MKK4. Because MKK4 signaling was completely inhibited by the Rho kinase inhibitor, these data indicate that Rho kinase modulated TNF-α-induced late JNK activation largely through MKK4-dependent mechanisms.
Rho kinase negatively regulates the PI3K/AKT pathway, which is not required for modulating JNK activation
A recently identified substrate of Rho kinase is phosphatase and tensin homolog (PTEN) (31). Activation of PI3K generates PI(3, 4, 5)P3, which in turn leads to activation of AKT (32). PTEN dephosphorylates PI(3, 4, 5)P3, thereby turning off the PI3K pathway (32). Phosphorylation of PTEN by Rho kinase leads to PTEN activation and subsequent inhibition of the PI3K/AKT pathway (31). In ECs, AKT phosphorylates and activates endothelial NO synthase, leading to NO production (33, 34). Endogenously produced NO has been shown to down-regulate JNK activation (35, 36, 37). Therefore, experiments were performed to determine whether Rho kinase regulates JNK activation by inhibiting a negative regulator of the JNK pathway, the PI3K/AKT/NO pathway.
To determine whether Rho kinase modulates JNK activation through the PI3K/AKT/NO pathway, the effect of Y27632 on TNF-α-induced JNK activation in the presence or absence of a PI3K inhibitor, LY294002, was examined. In the absence of a PI3K inhibitor, pretreatment with Y27632 inhibited JNK activation induced by TNF-α, as has been shown in Fig. 3. In the presence of a PI3K inhibitor, LY294002, pretreatment with Y27632 still inhibited JNK activation induced by TNF-α (data not shown). Similarly, in the presence of a NO inhibitor, l-NMA, Y27632 inhibited JNK activation induced by TNF-α (data not shown). These data indicate that although Rho kinase negatively regulates the PI3K/AKT pathway, this is not the mechanism through which Rho kinase modulates JNK activation.
TNF-α-induced IL-6 secretion requires Rho kinase activation
TNF-α-induced JNK activation is required for AP-1-dependent IL-6 expression (5, 11). A time-course study showed that TNF-α-induced IL-6 secretion from ECs that could be detected as early as 3 h and lasted for at least 24 h (data not shown). In addition, measurement of IL-6 in the cell lysates and in the culture supernatants showed that the amount of IL-6 present in the supernatant accounted for 80–90% of total IL-6 after 3 h of TNF-α treatment (data not shown). To determine whether TNF-α-induced IL-6 production requires activation of the JNK pathway and whether modulation of the JNK pathway by the Rho kinase inhibitor modulates IL-6 production, the effect of a JNK inhibitor, SP600125, and Y27632 was examined. Pretreatment with SP600125 completely prevented JNK activation induced by TNF-α (Fig. 5,A). Pretreatment with SP600125 prevented TNF-α-induced early IL-6 release, indicating that TNF-α-induced IL-6 release is JNK dependent (Fig. 5,B). Inhibition of MKK4 expression by siRNA, which only inhibited late JNK activation induced by TNF-α, had little effect on IL-6 release: in cells treated with control siRNA, TNF-α stimulation for 3 h increased IL-6 release from 7.4 ± 1.3 to 22.2 ± 4.1 pg/ml (n = 4, p < 0.05), while in cells treated with MKK4 siRNA, IL-6 release similarly increased from 8.2 ± 3.3 to 24.0 ± 3.9 pg/ml (n = 4, p < 0.05). Pretreatment with Y27632, which inhibited both the early and late JNK activation, prevented TNF-α-induced IL-6 release (Fig. 5 B). By contrast, inhibition of RhoA protein expression did not inhibit IL-6 release. In cells treated with control siRNA, TNF-α stimulation for 3 h increased IL-6 release from 6.4 to 21.2 pg/ml, while in cells treated with RhoA siRNA, IL-6 release increased from 8.5 to 36.5 pg/ml (data represent averaged values from two separate experiment). Therefore, activation of Rho kinase is required for TNF-α-induced IL-6 release, while RhoA is not.
Rho kinase activation modulates TNF-α-induced cytoskeletal changes and permeability increases independent of its effect on JNK activation
TNF-α also induces cytoskeletal changes and permeability increases in cultured human pulmonary microvascular ECs that are apparent after 1–3 h of TNF-α treatment (21). TNF-α treatment caused changes in the F-actin cytoskeleton in these ECs that included thickening of F-actin bundling along the EC periphery (Fig. 6,A). Pretreatment with 20 μM Y27632 completely prevented these changes (Fig. 6 A).
The effect of Y27632 on TNF-α-induced permeability increases was examined by measuring the fluxes of dextran across an EC monolayer. TNF-α was added to ECs 1 h before measuring dextran fluxes. In ECs pretreated with vehicle, TNF-α caused a significant increase in the dextran fluxes that was detectable by 2 h following TNF-α treatment. Pretreatment with 20 μM Y27632 lowered the baseline permeability to dextran fluxes and completely prevented TNF-α-induced permeability increases (Fig. 6 B).
Because TNF-α-induced Rho kinase activation was required for JNK activation, and JNK activation has been shown to modulate the actin cytoskeleton (6, 7), whether Rho kinase modulates cytoskeletal changes in ECs through its effect on JNK activation was examined. Pretreatment of ECs with the JNK inhibitor, SP600125, however, did not prevent TNF-α-induced cytoskeletal changes in ECs (Fig. 6 C). These data indicate that Rho kinase modulates cytoskeletal changes in these ECs independent of its effect on the JNK pathway.
Activation of Rho kinase plays important roles in regulating innate immune responses. Inhibition of Rho kinase in vivo prevents neutrophil recruitment and edema formation in the lungs during LPS pneumonia (17). Although Rho kinase is well-recognized for its role in modulating the actin cytoskeleton in many cell types during an inflammatory response, this study provides evidence that Rho kinase also modulates JNK activation and JNK-dependent IL-6 production in pulmonary microvascular ECs in response to TNF-α stimulation. This may represent an important mechanism through which activation of Rho kinase modulates an inflammatory response.
TNF-α induced rapid activation of RhoA in primary human pulmonary microvascular ECs that could be detected within 5 min. This early onset of RhoA activation suggests a role for the direct engagement of the TNF-α receptor(s). In smooth muscle cells, some of the TNFR1 is localized in lipid rafts. TNF-α ligation induces recruitment of adaptor molecules TNFR-associated death domain, TRAF2, and receptor-interacting protein as well as RhoA into lipid rafts, and this compartmentalization is required for TNF-α-induced RhoA activation (12). This suggests that early RhoA activation occurs proximal to the TNF receptor(s) and requires recruitment of adaptor molecules to the receptor(s). Lipid rafts are beyond the detection limit of conventional microscopy including confocal microscopy, and this may explain why early RhoA activation induced by TNF-α in these ECs was not associated with changes in RhoA distribution (Fig. 1 B). Interestingly, a late phase of RhoA activation was also detected at 3 h, and this late RhoA activation was associated with the accumulation of RhoA in focal areas along the EC edges. These areas were devoid of Rho GDI, an endogenous inhibitor of Rho. The accumulation of RhoA was focal, consistent with the activity assay showing that RhoA activity increased by only 2-fold.
Rho kinase activation was also induced by TNF-α. Rho kinase is kept in the inactive form through intramolecular interactions, and binding to GTP-RhoA, arachidonic acid or sphingosine phosphorylcholine, or proteolytic cleavage of the autoinhibitory domain by caspase-3 or perforin can result in Rho kinase activation (38). Our data using siRNA showed that RhoA was not required for Rho kinase activation induced by TNF-α. In addition, inhibition of Rho kinase activation by Y27632 prevented TNF-α-induced JNK activation, while inhibition of RhoA expression using siRNA did not. These data indicate that TNF-α-induced Rho kinase activation does not require RhoA, and that Rho kinase, but not RhoA, modulates JNK activation. Because Y27632 did not prevent TNF-α-induced p38 activation and ICAM-1 expression, the effect of Y27632 on JNK activation is unlikely due to a direct effect on the TNF-α receptor. Moreover, Y27632 inhibited JNK activation induced by IL-1β, further suggesting that Y27632 did not inhibit JNK activation by modulating the receptor itself.
In many cell types, TNF-α induces a time-dependent activation of the JNK pathway. An early phase of JNK activation occurs by 5–30 min that is transient and robust. This is followed by a less robust phase that lasts for hours (11). Rho kinase activation is required for both the early and late phase of JNK activation, and the Rho kinase activity is similarly elevated at both the early and late phase of JNK activation. These data suggest that while Rho kinase is required for JNK activation, the activity of Rho kinase alone does not dictate the level of JNK activation, and other mechanisms also play a role. Indeed, recent studies suggest that different mechanisms mediate the early and late JNK activation in a cell type-specific manner and that early and late JNK activation likely serve different biological functions (39). Early activation of JNK occurs as a result of the assembly of signaling complexes at the TNF-α receptor and requires the adaptor protein TRAF2 (40). Through mechanisms that remain to be defined, sequential activation of MAP3K isoforms (such as MLK, ASK1, and TAK1), MKK isoforms (MKK4 and MKK7), and JNK is induced. After the initial and robust JNK activation, several pathways are induced by TNF-α that result in prolonged JNK activation at much lower levels compared with the peak levels. These mechanisms include the induction of MAPK phosphatase and cellular FLIP that serve to inhibit JNK activation, and the production of reactive oxygen species (ROS) that mediates JNK activation through inactivation of MAPK phosphatases as well as activation of a MAP3K isoform, ASK1 (41, 42, 43). NF-κB negatively regulates ROS production by up-regulating antioxidants such as the ferritin H chain (44, 45), while JNK positively regulates ROS production (46). Therefore, at later times of TNF-α stimulation, Rho kinase activation regulates MKK4 activation and subsequent JNK activation. However, all the negative mechanisms are also induced to keep the total JNK activity elevated at a much lower level than the peak level.
MKK4 and MKK7 are the two MKKs that activate JNK in response to TNF-α (47). Although MKK4 can also phosphorylate and activate p38 under certain conditions, MKK7 is a specific activator of JNK (48). Single genetic deletion of MKK4 or MKK7 in mice results in early embryonic death, indicating that MKK4 and MKK7 serve nonredundant functions in vivo (49, 50). MKK4 and MKK7 can be differentially activated by different stimuli and recruit different scaffold molecules to activate JNK (48, 51). Although MKK4 is required for optimal JNK activation induced by TNF-α, MKK7 is the major MKK leading to early JNK activation (47). Our studies using MKK4 siRNA are consistent with these previous studies, but also identified a role for MKK4 in TNF-α-induced late JNK activation. These data, along with the observations that MKK4 activation completely depended on Rho kinase, indicate that Rho kinase modulated early vs late JNK activation through different mechanisms.
This study links the Rho kinase pathway to the JNK pathway and suggests that Rho kinase likely regulates JNK-dependent gene transcription. TNF-α-induced gene transcription in many cell types is mediated by transcription factors, most importantly, NF-κB and AP-1. TNF-α-induced AP-1 up-regulation requires JNK (5), and both the early phase and late phase of JNK activation is required for AP-1 activation and transcription of inflammatory genes such as IL-6 (11). Therefore, Rho kinase likely modulates AP-1-dependent gene transcription by regulating the JNK activation induced by TNF-α and IL-1β. Indeed, inhibition of Rho kinase reduced IL-6 secretion induced by TNF-α. The effect of these inhibitors was only partial, likely reflecting the fact that the IL-6 promoter contains binding sites for other transcription factors such as NF-κB (52). Interestingly, the Rho kinase inhibitor and the JNK inhibitor induced similar reduction in IL-6 secretion, despite the fact that the Rho kinase inhibitor only partially inhibited early JNK activation. Because IL-6 deficiency and JNK inhibition reduce inflammatory responses in the lungs, Rho kinase may regulate inflammatory responses in vivo by regulating the JNK-IL-6 pathway (6, 7, 17, 53).
Independent of JNK regulation, Rho kinase was required for TNF-α-induced cytoskeletal changes and permeability increases. Although Rho kinase activation occurs within minutes of TNF-α stimulation, TNF-α-induced permeability increases require hours to develop (Fig. 6). This suggested that increases in Rho kinase activity alone are not sufficient to induce cytoskeletal changes and permeability increases. The subcellular accumulation of RhoA following TNF-α stimulation for 3 h suggests that RhoA likely modulates actin cytoskeletal changes and junctional proteins in a site-specific manner at later times of TNF-α stimulation, leading to permeability increases.
In summary, this study demonstrates that TNF-α induced early activation of the RhoA and Rho kinase pathway in human pulmonary microvascular ECs. Rho kinase activation, but not RhoA, regulated TNF-α-induced early and late JNK activation that had a differential requirement for MKK4. Activation of Rho kinase was required for JNK-dependent IL-6 secretion, and modulated cytoskeletal changes and permeability increases in a JNK-independent manner. Therefore, Rho kinase plays important roles in endothelial responses to TNF-α by regulating IL-6 production and permeability increases.
The confocal studies were performed using the imaging core facility in the Department of Pediatrics at Case Western Reserve University and the Herbert Irving Comprehensive Cancer Center Optical Microscopy Facility at Columbia University.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by National Institutes of Health Grant HL070009 (to Q.W.).
Abbreviations used in this paper: EC, endothelial cell; GDI, GDP dissociation inhibitor; MYPT-1, myosin phosphatase target subunit-1; siRNA, small-interfering RNA; TRAF, TNFR-associated factor; RBD, rhotekin-Rho-binding domain; PTEN, phosphatase and tensin homolog; ROS, reactive oxygen species; MKK, MAPK kinase; VE, vascular endothelial.