The Small GTPase RhoB Regulates TNFα Signaling in Endothelial Cells

The inflammatory response of endothelial cells triggered by cytokines such as TNFα and IL1β plays a pivotal role in innate immunity. Upon pro-inflammatory cytokine stimulation, endothelial cells produce chemokines and cytokines that attract and activate leukocytes, and express high levels of leukocyte adhesion molecules. This process is mediated by intracellular signaling cascades triggered by activation of e.g. the TNFα receptor (TNFR) that lead to the activation of the NFκB transcription factor and of MAP kinases, which in turn activate inflammatory gene transcription. We found that the small GTPase RhoB was strongly and rapidly upregulated in primary human endothelial cells by TNFα, IL1β and LPS. We subsequently investigated the role of RhoB in the regulation of TNFR signaling in endothelial cells by silencing RhoB expression with siRNA. We provide evidence that the TNFα-induced activation of p38 MAP kinase is strongly dependent on RhoB, but not on RhoA, while JNK activation is regulated by both RhoB and RhoA. Consistent with the important role of p38 MAP kinase in inflammation, we demonstrate that loss of RhoB impairs TNFα-induced ICAM-1 expression and reduces cell production of IL6 and IL8. In addition, we show that RhoB silencing alters the intracellular traffic of TNFα after endocytosis. Since RhoB is a known regulator of the intracellular traffic of membrane receptors, our data suggest that RhoB controls TNFα signaling through the regulation of the TNFR traffic.


Introduction
Tumor necrosis factor a (TNFa) is a pleiotropic pro-inflammatory cytokine that plays a pivotal role in the innate immune response to infection and tissue injury. Vascular endothelial cells respond to TNFa by upregulating the expression of cytokines and chemokines, such as IL-6 and IL-8, and of endothelial leukocyte adhesion molecules, such as VCAM-1, ICAM-1 and E-selectin [1]. These molecules enable TNFa-activated endothelial cells to attract, activate and recruit circulating leukocytes, which subsequently extravasate to reach the site of infection or injury. The inflammatory program induced by TNFa is a result of intracellular signaling triggered by the TNFa-receptor (TNFR) [2,3]. Upon ligand binding, TNFR trimerizes and recruits TRAF-2 (TNFRassociated factor 2) and RIP1 (receptor interacting protein 1) to its cytoplasmic death domain. The formation of this signaling complex leads to the activation of the transcription factor NFkB and of the MAP kinases JNK and p38. Subsequently, the TNFR is rapidly endocytosed and eventually degraded in the lysosomes [4,5]. However, TNFR internalization is clearly not only a mechanism of receptor downregulation but also of signaling compartmentalization, providing temporal and spatial regulation of the diverse signaling cascades triggered by the activated receptor [6]. While signaling from the TNFR leading to NFkB activation takes place at the plasma membrane, there is compelling evidence that TNFR pro-apoptotic signaling occurs on endosomes [4,7]. In addition, several molecules involved in TNFR signaling are found on the surface of endosomal and lysosomal compartments [8,9]. Finally, one study has demonstrated that internalization of the TNFR from the plasma membrane is a required step for the activation of p38 and JNK MAP kinases [10].
RhoB is a short-lived Rho GTPase whose expression is inducible by a variety of stimuli including growth factors, such as EGF and PDGF [11] and stress stimuli such as DNA-damaging drugs, UV irradiation and reactive oxygen species [12,13]. RhoB is 83% identical to RhoA, a constitutively expressed GTPase and a wellestablished regulator of actomyosin-based contractility and of seruminduced transcription. Although these two GTPases bind to a similar set of proteins in solution, their non-overlapping intracellular distribution provides specificity to their respective actions [14]. Whereas RhoA is cytosolic and translocates to the plasma membrane upon activation, RhoB localizes to endosomes/multivesicular bodies [15]. Multivesicular bodies are primarily involved in the sorting of membrane proteins for their delivery to lysosomes for degradation. Consistently, RhoB regulates the sorting and degradation of growth factor and cytokine receptors [16][17][18][19][20][21]. In agreement with the role of Rho GTPases as critical regulators of actin dynamics, RhoB appears to control vesicle traffic through the regulation of actin polymerization on endosomes [22,23], possibly through the recruitment and activation of Diaphanous proteins [23,24].
Inflammatory cytokines such as tumor necrosis factor a (TNFa and interleukin 1b (IL1b activate endothelial cells by inducing multiple intracellular signaling pathways that regulate gene expression. The small GTPase RhoB is a short-lived protein encoded by an immediate-early gene that is rapidly activated in response to a wide variety of stimuli including growth factors, UV radiation and oxidative stress [11][12][13]. Here we report that RhoB protein is rapidly upregulated in primary human endothelial cells by TNFa, IL1b and bacterial lipopolysaccharide (LPS). We have addressed the role of RhoB in TNFR signaling by using siRNA-mediated knockdown of RhoB. We present evidence that RhoB is essential for the activation of p38 and JNK MAP kinases, but not NFkB, by TNFa. Finally, we show that RhoB silenced cells accumulate endocytosed TNFa pointing to a defect in traffic kinetics and/or receptor sorting. In summary, our data suggest that RhoB has as role in TNFa signaling through the regulation of TNFR intracellular traffic.

Cell culture
Pooled primary human umbilical vein endothelial cells (HUVEC, Lonza) were seeded on fibronectin-coated culture flasks and maintained in EGM-2 medium (Lonza) in a humidified atmosphere of 95% air and 5% CO2 at 37uC.

Reagents and antibodies
Recombinant human TNFa, IL1b, IFNc, VEGF and TGFb were from R&D Systems. Bacterial lipopolysaccharide (LPS) was from Sigma.The pharmacological inhibitors cycloheximide, MG132 and N-acetyl-cysteine (NAC) were from Sigma. SC-514, SP600125 and SB230580 were from Calbiochem. Rabbit and mouse anti-RhoB, mouse anti-RhoA, mouse anti-phospho-ERK, mouse anti-phospho-JNK, goat anti-VCAM-1 and mouse anti-NFkB p50 were obtained from Santa Cruz Biotechnology. Rabbit anti-phospho-p38 and mouse anti-IkBa were obtained from Cell Signaling. Mouse anti-a-tubulin (DM1A) was from Sigma. Mouse anti-EEA1 and mouse anti-RhoGDI were from BD Biosciences. Secondary antibodies labeled with Alexa488 or Alexa568 for immunofluorescence were purchased from Invitrogen. Secondary antibodies labeled with horseradish peroxidase for immunoblotting were from Amersham.

Immunoblotting and phosho-MAP kinase arrays
Cells were lysed in cold NP-40 buffer (1% NP40, 100 mM NaCl, 100 mM MgCl 2 , 10% glycerol, 50 mM Tris pH 7.4) containing a cocktail of phosphatase and protease inhibitors (Thermo Scientific). Protein content of lysates was quantified with the Precision Red Advanced Protein Assay Reagent (Cytoskeleton). Equal protein concentrations were loaded in SDS-PAGE gels and analyzed by Western blotting. Equal loading was additionally controlled by detection of Rho-GDI and atubulin. The pixel density of each band was determined with ImageJ and values corrected by the corresponding atubulin intensities. These values  were normalized to those of siRNA control-transfected cells. Nonparametric one-way ANOVA Tukey test was used to evaluate statistical significance of at least 3 independent experiments. A two-way ANOVA with Bonferroni post-test was used to evaluate statistical significance when the effects of several siRNA effects were compared at different time points on TNF stimulation. Results are expressed as mean 6 SEM (*p,0.05; **p,0.01; ***p,0.001).
Human phospho-mitogen activated kinase protein antibody arrays were purchased from R&D Systems and used according to manufacturer instructions. Briefly, cell lysates were incubated with the antibody arrays overnight at 4uC. After washing, arrays were incubated with a mixture of phospho-site specific biotinylated antibodies for 2 hours at room temperature. Bound biotinylated antibodies were detected with HRP-streptavidin. Arrays were developed in ECL and exposed to X-Omat films. Digital scans of the films were analyzed for pixel density with ImageJ. Averaged background values corresponding to the negative controls were subtracted from the values of each spot.
Values for duplicates on the array were averaged and represented as a percentage of the pixel density of the positive controls included in the array.

RhoB activity assay
Rho-GTP pulldown assays were performed as previously described [25]. Cells were lysed on ice in lysis buffer containing 50 mM Tris pH 7.6, 500 mM NaCl, 1% Triton X-100, 0.1% SDS, 0.5% deoxycholate, 10 mM MgCl 2 , 100 and a cocktail of protease inhibitors (Sigma). Lysates were clarified by centrifugation at 14,0006g 5 min and incubated with Glutathion Stransferase (GST)-Rho binding domain beads. After washing four times in 50 mM Tris pH 7.6, 150 mM NaCl, 1% Triton X-100, 0.5 mM MgCl 2 , 100 mM orthovanadate, with protease inhibitors, bound Rho proteins were solubilized with SDS-sample buffer and analyzed by SDS-PAGE with specific antibodies for RhoB. RhoB and RhoA knock-down with siRNA Cells were transfected with a pool of 3 siRNA duplexes for RhoB (sc-29472, Santa Cruz Biotechnology, 59-39 CCCUUG-UCUGUAACAUAGAAs(siRhoB#1); CCACACUUGUACGC-UGUAA(siRhoB#2); CCAGUGGUACUUCUACUAA(siRhoB#3) or with either siRNA#1 or #2. siRNA duplex for RhoA (sc-29471, Santa Cruz Biotechnology, 59-39 GGCAGAGAUAUG-GCAAACA). As control we used a non targeting 20-25 nucleotide siRNA designed as a negative control (Control siRNA-A, sc-37007, Santa Cruz Biotechnology). Transfection of HUVEC was performed using the siRNA transfection reagent and medium from Santa Cruz Biotechnology according to the manufacturer's protocol. Briefly, HUVEC were seeded the day previous to transfection in EGM-2 without antibiotics. For transfection, cells were washed in transfection medium and incubated with a mixture of siRNA and transfection reagent for 5 hours at 37uC and 5% CO 2 . Cells were then rinsed and incubated for a period of 48 to 72 hours before stimulation and analysis.

TNFa endocytosis
To analyse TNFa endocytosis and intracellular traffic, cells were incubated with biotin-labelled human TNFa (R&D Systems) for 1 hour at 4uC followed by 30 minutes incubation with FITCavidin (R&D Systems) also at 4uC. After washing, cells were either immediately fixed in 3.7% formaldehyde or transferred to 37uC to allow TNFa internalization and fixed after different incubation times for immunofluorescence analysis. Flow cytometry analysis was performed as indicated above but cells were detached from the dish with trypsin-EDTA (Lonza) after each incubation time. After addition of Trypsin Neutralizer Solution (Lonza), cells were kept on ice. Flow cytometry analysis of FITC-positive cells was performed in a FACS Canto (BD Biosciences).

Immunofluorescence
Cells seeded on fibronectin-coated glass coverslips were fixed in formaldehyde 3.7%, permeabilized with 0.1% Triton X-100 and blocked with PBS containing 0.5% bovine serum albumin (PBS-BSA). Primary antibodies were incubated for 1 hour at room temperature followed by 30 minutes incubation with secondary antibodies. All antibodies were diluted in PBS-BSA. Images of stained cells were collected with a LSM510 confocal microscope (Zeiss). Images were analyzed for quantitative co-localization using Zen 2009 software (Zeiss).

Cytokine ELISA
Cytokine levels of IL6 and IL8 were measured in the supernatants of HUVEC transfected with Control, RhoB or RhoA siRNAs and stimulated with TNFa for 4 and 16 hours using commercially available enzyme-linked immunosorbent assay (ELISA) kits (PeliKine Compact TM human ELISA kits, Sanquin, Amsterdam, The Netherlands) as previously described [26]. The plates were read in an ELISA reader (Labsystems Multiskan Multisoft, Helsinki, Finland) at 450 nm, with 540 nm as a reference.

Statistical analyses
Statistically significant differences between data means were determined by two-tailed paired Student's t test using Excel software (Microsoft); P,0.05 was considered statistically significant.

Results
We found that the pro-inflammatory mediators TNFa, IL1b and bacterial LPS (lipopolysaccharide) potently stimulated RhoB expression in primary human endothelial cells, while RhoA expression was unchanged ( Figure 1A). In contrast, other endothelial stimuli such as interferon c (INFc, transforming growth factor b (TGFb and vascular endothelial growth factor (VEGF) had little effect on RhoB expression ( Figure 1A).
To assess whether, in addition to RhoB protein levels, TNFa also increases the levels of activated RhoB, we precipitated GTP-Rho with the Rho-binding domain of rhotekin coupled to Sepharose beads [25] and detected RhoB by western blotting. GTP-RhoB could be detected in lysates of cells stimulated with TNFa ( Figure 1B), which indicates that TNFa increases the level of activated RhoB in endothelial cells.
Next, we sought to determine the mechanism by which TNFa induced RhoB up-regulation. Two mechanisms regulate RhoB levels in cells: increased synthesis and protein stabilization [12,[27][28][29]. To discriminate between these two possibilities, we used inhibitors of protein synthesis (cycloheximide) and of proteasomal degradation (MG132). We incubated HUVEC with TNFa alone for 4 hours, in order to accumulate RhoB, and subsequently added cycloheximide and/or MG132 for an additional 2 hours. Cycloheximide incubation caused the complete depletion of RhoB whereas MG132 increased RhoB levels when compared to cells stimulated with TNFa alone ( Figure 1C). These results indicate that TNFa upregulates RhoB protein synthesis and that newly synthesized RhoB is rapidly degraded by the proteasome. Consistently, inhibition of both synthesis and degradation after TNFa stimulation resulted in RhoB levels comparable to those in unstimulated cells ( Figure 1C). To estimate the half-life of RhoB in resting and in TNFa-stimulated cells, the kinetics of RhoB degradation was examined. Endothelial cells were first stimulated with TNFa for 4 hours to accumulate RhoB in cells. Then, protein synthesis was inhibited by the addition of cycloheximide for 1, 2 and 4 hours and RhoB levels were analyzed by western blotting. A progressive loss of RhoB was observed correlating with the duration of cycloheximide treatment in both control and TNFa-treated cells ( Figure 1D). We plotted the band intensity as a percentage of the RhoB present in the absence of cycloheximide and fitted the data points using a one-phase exponential decay function ( Figure 1D). The estimated values for the half-life of RhoB was 2.3 hours in unstimulated cells and 1.7 hours in TNFastimulated cells. Thus, TNFa does not significantly change the half-life of RhoB, suggesting that TNFa does not enhance RhoB protein stability.
TNFa activates gene transcription through the activation of both NFkB and MAP kinase pathways [3]. To assess the involvement of these pathways in the upregulation of RhoB by TNFa, we tested different pharmacological inhibitors of NFkB and p38 and JNK MAP kinases [30]. Although the inhibitor of NFkB, sc-514 and the ROS scavenger N-acetyl-cysteine (NAC) impaired RhoB induction by TNFa, the largest effect was observed with the JNK inhibitor SP600125 ( Figure 1E).
To explore the role of RhoB in TNFa-induced inflammation we examined the two main signaling pathways triggered by TNFa,; NFkB and MAPK, after silencing RhoB expression using siRNAmediated knock-down with a pool of 3 different siRNAs or with single siRNAs (#1 and #2) from this pool individually (Figure 2A). Activation of NFkB by TNFa results in the phosphorylation and subsequent degradation of inhibitory IkB proteins [3]. Neither RhoB, nor RhoA downregulation affected the breakdown of the IkBa chain following TNFa stimulation ( Figure 2B and Figure S1A). Consistent with these results, RhoB silencing did not prevent NFkB nuclear translocation, as determined by immunofluorescent staining for the p65 NFkB subunit ( Figure 2C). Thus, RhoB does not regulate NFkB activation by TNFa.
We next used anti-phospho MAP kinase antibody arrays to analyze changes in phosphorylation of various cellular serine/ threonine kinases in endothelial cells transfected with RhoB siRNA or control siRNA before and after TNFa stimulation. The results of this analysis showed that TNFa-induced phosphorylation of JNK (JNK1, JNK2), p38a and heat shock protein 27 (hsp27, a substrate of the p38 MAP kinase pathway) is abrogated in RhoBdeficient cells ( Figure 3A). In support of the validity of the MAP kinase array analysis, RhoB silencing also reduced the levels of phosphorylated Akt and GSK3b, as previously described in endothelial cells [31] and in keratinocytes [32]. We further confirmed the effects of RhoB silencing on the levels of phosphorylated MAP kinases by traditional western blotting. Similar to the results with phospho-MAP kinase arrays, RhoB knock-down resulted in decreased levels of phospho-JNK, phospho-p38 and phospho-hsp27 upon TNFa stimulation ( Figure 3B and Figure S1B). In contrast, RhoB silencing had no effect on ERK activation, indicating that RhoB only regulates stress-activated MAPKs. To test for the specificity of RhoB in p38 activation, we examined p38 activation in RhoA-silenced cells ( Figure S1B). These data suggest that both RhoB and RhoA are required for TNFa-induced JNK activation while RhoB specifically regulates p38 activation. Interestingly, we found that blocking RhoA expression upregulates the cellular levels of RhoB by approximately 6-fold in unstimulated and 3-fold in TNFastimulated cells. RhoB knock-down only moderately increases RhoA levels by less than 2-fold.
MAPKs p38 and JNK regulate TNFa-induced pro-inflammatory gene expression [33,34]. Since our data show that RhoB regulates MAPK activation by TNFa, we next tested whether RhoB is required for the pro-inflammatory response of endothelial cells. To this end, we analyzed the expression of VCAM-1 and ICAM-1, as well as of the production of IL8 and IL6 (Figure 4). RhoB silencing significantly reduced TNFa-induced ICAM-1 expression. Although VCAM-1 levels appeared slightly reduced, the changes were not statistically significant ( Figure 4A). In addition, RhoB silencing diminished the endothelial production of IL8 and IL6 ( Figure 4B). Thus, RhoB appears to be required for optimal expression of pro-inflammatory molecules by endothelial cells upon TNFa stimulation.
RhoB modulates a variety of signal transduction pathways through the regulation of receptor traffic [16][17][18][19][20][21]., therefore we addressed the question whether RhoB also controls intracellular traffic of the TNFR. RhoB localized to vesicles in endothelial cells stimulated with TNFa ( Figure 5A, upper panels). This distribution is similar to that found in unstimulated cells treated with proteasome inhibitor in order to accumulate RhoB to detectable levels ( Figure 5A, lower panels). Thus, TNFa does not appear to change the subcellular distribution of RhoB, which localizes to EEA1-positive early endosomes ( Figure 5B) [23,35].
Upon TNFa binding, the TNFR is internalized into endosomes, which subsequently fuse with trans-Golgi network-derived vesicles, and is finally transported to lysosomes [4]. To address the question whether RhoB regulates TNFR traffic we used biotin-TNFa/FITC-streptavidin as previously described [4]. Cells were incubated at 4uC with biotin-TNFa followed by incubation with FITC-streptavidin. Subsequently, internalization of biotin-TNFa/FITC-streptavidin complexes was allowed by transfer of the cells to 37uC for various periods of time. Detection of RhoB shows that internalized TNFa colocalizes with RhoB ( Figure 5C), suggesting that the TNFR traffics through RhoB-positive endosomes.
We then studied TNFa internalization upon RhoB silencing by immunofluorescence ( Figure 6A) and flow cytometry ( Figure 6B) in a time-course experiment. No apparent differences were observed in the amount of biotin-labelled TNFa bound to the cell surface between control and RhoB-deficient cells, which indicates that RhoB does not control surface expression of the TNFR. Also after 10 and 30 minutes of internalization, similar intracellular amounts of TNFa were found in control and RhoB-negative cells ( Figure 6A). However, after 90 minutes, RhoB-deficient cells contained a larger amount of TNFa than control cells ( Figure 6A), suggesting that RhoB plays a role in TNFR traffic. We performed the same experiment and quantitatively analyzed cell-associated fluorescence by flow cytometry ( Figure 6B). These results recapitulate those found by confocal microscopic analysis. The absence of TNFa-positive cells at time point 0 can be explained by the fact that cells are suspended by trypsinization, which is likely to cause the loss of membrane-bound TNFa.

Discussion
Rho GTPases are key signaling components controlling the inflammatory response elicited by pro-inflammatory cytokines [36]. In our study, we examined the role of RhoB in the inflammatory response elicited by TNFa in primary human endothelial cells.
We show here that pro-inflammatory mediators such as TNFa, IL1b and LPS upregulate RhoB expression, whereas RhoA appears to be constitutively expressed. Our data suggest that TNFa increases RhoB protein synthesis without affecting protein stability. The half-life of RhoB in unstimulated endothelial cells was of ,2 hours, as previously shown in other cell types, and of 1.7 hours in TNFa-stimulated cells, indicating that TNFa does not promote RhoB protein stabilization. Our data suggest that TNFa induces the transcriptional activation of the immediate-early gene encoding RhoB mainly through a JNK-dependent pathway, previously involved in the transcriptional upregulation of RhoB [37,38].
Following our initial observation that TNFa potently stimulates RhoB protein synthesis and activity, we argued that RhoB might have a role in TNFa-dependent signaling. First, we show that TNFa increases active GTP-RhoB in endothelial cells. This increase may be a consequence of the increase in RhoB protein rather than of enhanced RhoB activation by TNFa-regulated guanine nucleotide-exchange factors (GEFs). Although the exchange factor that activates RhoB in endothelial cells is currently unknown, recent work showed that GEF-H1 mediates LPSinduced RhoB activation in dendritic cells [39]. Future studies will determine if GEF-H1 is also involved in RhoB activation by TNFa in endothelial cells.
The enhanced activity of RhoB in TNFa-stimulated cells suggested a role for RhoB in TNFR signaling. We studied the activation of the two main signaling cascades triggered by TNFa; the NFkB and MAPK pathways, after RhoB silencing with siRNA. To control for specificity of RhoB action, we silenced RhoA, a close member of the Rho GTPase subfamily of Ras GTPases. Our studies revealed that RhoB does not regulate NFkB activation by TNFa. However, activation of p38 MAP kinase by TNFa is critically dependent on RhoB, while both RhoB and RhoA are required for JNK activation. To the best of our knowledge, this is the first study on the specific role of endogenous RhoA or RhoB in the activation of p38 and JNK by TNFa in primary human endothelial cells. Our findings are supported by previous studies using Rho-targeting toxins or Rho mutants [40,41]. Inhibition of all Rho isoforms (RhoA, B and C) with C3 toxin was shown to impair TNFa-induced p38 activation in endothelial cells [40]. In addition, expression of active mutants of RhoA and RhoB was shown to activate JNK in 293T cells [41]. Thus, RhoB regulates TNFa-dependent activation of stressactivated MAPKs in endothelial cells.
Given the central role that JNK and p38 play in inflammatory responses elicited by TNFa, our findings led to the hypothesis that RhoB has a pro-inflammatory role in endothelial cells. Further proof for this hypothesis was obtained upon examination of various pro-inflammatory proteins regulated by the p38 and/or JNK pathways. We show that RhoB silencing impairs TNFainduced production of IL6 and IL8 and significantly reduces the expression of ICAM-1. Consistently, TNFa-induced activation of p38 was previously found to be critical for the expression of proinflammatory molecules such as TNFa, IL6, IL8 and cyclooxygenase-2 (COX-2) [42][43][44][45][46][47][48]. Similarly, p38 was previously shown to regulate ICAM-1 expression in endothelial cells [33,34]. Collectively, our data strongly suggest that RhoB participates in the pro-inflammatory response of endothelial cells to TNFa through the regulation of p38 activation.
RhoB is an endosomal GTPase that regulates endosome dynamics through the recruitment of actin-polymerizing proteins of the formin family [23,24]. Accordingly, RhoB controls the endocytic traffic and signaling of growth factor and chemokine receptors [16][17][18][19][20][21]. We assessed whether RhoB also regulates TNFR traffic by analyzing TNFa internalization in RhoBdeficient cells. Although we could not detect the TNFR by immunofluorescence due to high background and low specific signal of anti-TNFR antibodies, we are confident the intracellular traffic of TNFa reflects that of TNFa/TNFR complexes [4]. First, we show that RhoB localizes to endosomes and that internalized TNFa traffics through RhoB-positive endosomes. RhoB silencing causes the intracellular accumulation of endocytosed TNFa whereas, in control cells TNFa disappears in time. These data show that RhoB is involved in the regulation of the intracellular trafficking of TNFa and suggest that RhoB is required for the sorting of the TNFR to the degradative pathway, in a similar manner as previously described for the CXCR2 receptor [19].
Even though we do not yet have direct proof, we speculate that RhoB participates in the activation of MAPKs by TNFa through the regulation of TNFR traffic. In support of our hypothesis, EEA-1 positive endosomes carry activated MAP kinases [49,50] and inhibition of receptor endocytosis hampers downstream activation of these kinases, suggesting that kinase activation takes place in an intracellular endocytic compartment [51]. Specifically, TNFR internalization is required for the activation of MAPK and Akt but not for IkBa degradation [10], suggesting that TNFR activates MAP kinases from an intracellular compartment whereas NFkB activation occurs at the plasma membrane. Similarly, TNFRinduced caspase activation takes place on endosomes [4]. Finally, endocytic compartments have been involved in the TNFadependent expression of cytokines and adhesion molecules [8,52].
In summary, our study shows that RhoB is critically required for the inflammatory response of endothelial cells to TNFa, likely through MAP kinase activation downstream of the TNFR. In addition, our data suggest that RhoB may regulate TNFR signaling through its regulation of TNFR endocytic traffic kinetics and/or of receptor sorting. Figure S1 (A) Cells transfected with a pool of 3 RhoB siRNAs, with a RhoA siRNA or with siRNA control were stimulated or not with TNFa for 30 minutes and IkBa was detected by western blotting of cell lysates. a-Tubulin was detected as control for equal protein loading; (B) Cells transfected with siRNAs mentioned above were stimulated with TNFa for 30 minutes. Subsequently, phospho-p38 and phospho-JNK were detected by western blotting of cell lysates. a-Tubulin was detected as control for equal protein loading. (TIF)