Neurotensin-induced Proinflammatory Signaling in Human Colonocytes Is Regulated by β-Arrestins and Endothelin-converting Enzyme-1-dependent Endocytosis and Resensitization of Neurotensin Receptor 1*

Background: Neurotensin induces proinflammatory responses in human colonic epithelial cells. Results: β-Arrestins and endothelin-converting enzyme 1 regulate neurotensin receptor 1-mediated inflammatory signaling in human colonocytes. Conclusion: Neurotensin-induced proinflammatory responses depend on β-arrestins and are regulated by receptor recycling. Significance: This is a previously unrecognized pathway for regulating neurotensin-induced colonic inflammatory responses. The neuropeptide/hormone neurotensin (NT) mediates intestinal inflammation and cell proliferation by binding of its high affinity receptor, neurotensin receptor-1 (NTR1). NT stimulates IL-8 expression in NCM460 human colonic epithelial cells by both MAP kinase- and NF-κB-dependent pathways. Although the mechanism of NTR1 endocytosis has been studied, the relationship between NTR1 intracellular trafficking and inflammatory signaling remains to be elucidated. In the present study, we show that in NCM460 cells exposed to NT, β-arrestin-1 (βARR1), and β-arrestin-2 (βARR2) translocate to early endosomes together with NTR1. Endothelin-converting enzyme-1 (ECE-1) degrades NT in acidic conditions, and its activity is crucial for NTR1 recycling. Pretreatment of NCM460 cells with the ECE-1 inhibitor SM19712 or gene silencing of βARR1 or βARR2 inhibits NT-stimulated ERK1/2 and JNK phosphorylation, NF-κB p65 nuclear translocation and phosphorylation, and IL-8 secretion. Furthermore, NT-induced cell proliferation, but not IL-8 transcription, is attenuated by the JNK inhibitor, JNK(AII). Thus, NTR1 internalization and recycling in human colonic epithelial cells involves βARRs and ECE-1, respectively. Our results also indicate that βARRs and ECE-1-dependent recycling regulate MAP kinase and NF-κB signaling as well as cell proliferation in human colonocytes in response to NT.

GPCR trafficking also participates in signal transduction. GPCRs at the plasma membrane can activate MAP kinases by G protein-dependent PKA or PKC and by G protein-mediated transactivation of the EGF receptor. However, G protein-mediated signaling at the plasma membrane is rapidly attenuated by ␤ARR-mediated desensitization mechanisms. ␤ARRs are also scaffold proteins that recruit certain GPCRs and MAP kinases to endosomes and thereby mediate a second wave of ␤ARR-dependent and G protein-independent signaling from endosomes (30). ␤ARRs interact with c-Src, c-Raf-1, MEK1, ERK2, and JNK3 (30 -34) and recruit MAP kinases to form "signalosomes" facilitating GPCR signaling (35)(36)(37)(38). When compared with G protein-dependent ERK activation, ␤ARR-associated ERK activation is more persistent, depends on ␤ARRs expression, and takes place in the cytosol (20,39,40).
Endothelin-converting enzyme (ECE-1) plays an important role in the endosomal trafficking and signaling of neuropeptide/hormone GPCRs (41)(42)(43)(44). ECE-1 exists in four isoforms (isoforms a-d), all of which share a common catalytic domain (45). All four isoforms are present in early endosomes, although ECE-1a and ECE-1c are mainly localized at the plasma membrane (45,46), whereas ECE-1b and ECE-1d are predominantly in endosomes (46,47). ECE-1 degrades certain neuropeptides, such as substance P and calcitonin gene-related peptide (42), in the acidified endosomal environment and thereby destabilizes the peptide/receptor/␤ARR signalosome (48). This destabilization has a dual effect on signal transduction. In the case of the substance P neurokinin-1 receptor and the calcitonin gene-related peptide receptor, endosomal ECE-1 allows the receptors, freed from ␤ARRs, to recycle, which mediates resensitization of G protein-mediated signaling at the plasma membrane. However, endosomal ECE-1 disrupts the substance P/neurokinin 1 receptor/␤ARR signalosome and thereby terminates ␤ARR-dependent ERK2 signaling from endosomes (48).
Sustained proinflammatory signaling responses represent a major component of NTR1 function that could be attenuated by disrupting receptor resensitization or intracellular signaling. Therefore, in this study, using nontransformed colonic epithelial cells as an in vitro model, we investigated the role of ␤ARRs and ECE-1, the key mediators in NTR1 trafficking for NT-stimulated proinflammatory signaling and proliferation. Our results suggest that both proinflammatory and proliferative responses in colonic epithelial cells are regulated by ␤ARRs and ECE-1dependent recycling of NTR1.
Western Blot Analysis-After appropriate treatment, NCM460-NTR1 were washed by ice-cold PBS twice and then incubated with radiolabeled immunoprecipitation assay buffer containing protease inhibitors and supplemented with PMSF and sodium orthovanadate (Santa Cruz Biotechnology) for 5 min. The cell lysates were centrifuged (12,000 rpm for 15 min at 4°C), and supernatants were analyzed by Western blot analysis. Proteins from equal amounts of cell extracts were separated in 10% SDS-PAGE and transferred to nitrocellulose membranes (Bio-Rad) at 70 V for 2 h at 4°C. Membranes were blocked by 5% nonfat dried milk in TBS (Boston Bioproducts, Ashland, MA) supplemented with 0.05% Tween 20 for 1 h. Appropriate antibodies were incubated with the membranes overnight at 4°C, washed with TBS/Tween 20, and incubated with secondary antibodies conjugated with horseradish peroxidase. Signals from target proteins were detected with SuperSignal chemiluminescent substrate (Pierce). Western blot bands were quantified by densitometry using Multi Gauge V3.1 (Fuji).
NF-B ELISA-Following appropriate treatment, nuclei-enriched portions of NCM460-NTR1 were isolated with Nuclei EZ prep nuclei isolation kit (Sigma-Aldrich) according to the manufacturer's instructions. The cells were washed with icecold PBS twice and lysed with Nuclei EZ lysis buffer. The nuclei were centrifuged (500 ϫ g for 5 min at 4°C) and resuspended in Nuclei EZ storage buffer. The supernatant (cytosolic portion) was stored at Ϫ80°C for further analysis. Translocation of p65 into the nucleus was detected by TransAM NF-B (Active Motif, Carlsbad, CA) as described by the manufacturer. In brief, 5 g of nuclei-enriched cell extracts were allowed to bind to the oligonucleotide containing the NF-B consensus site coated in 96-well plates. The p65 in nuclear extracts was detected by primary antibodies against p65 and secondary antibodies conjugated with horseradish peroxidase. The signals were detected at A 405 nm using a 96-well plate reader.
IL-8 ELISA-NCM460-NTR1 cells were exposed to NT for 6 h, and IL-8 in cell supernatants was measured by Duo-Set ELISA for IL-8 (R & D Systems, Minneapolis, MN) according to the manufacturer's instructions.
Luciferase Activity-The NF-B-driven luciferase reporter construct was purchased from Clontech and was used as previously described (15). A reporter construct with the IL-8 promoter (nucleotides Ϫ1481 to ϩ40) was used to measure the transcription activity of IL-8 as described (15). Either one of the reporter constructs, a control luciferase construct (pRL-TK; Promega, Madison, WI) and siRNAs (control and ␤ARRs) were transfected into NCM460-NTR1 using Lipofectamine 2000. NCM460-NTR1 were seeded at 1 ϫ 10 5 cells/well in 24-well plates 24 h prior to transfection. After 48 h, transfected cells were serum-fasted (16 h) and pretreated with pharmacological inhibitors where appropriate, followed by stimulation with NT for 1.5 h. Firefly and Renilla luciferase cell activities were detected using dual luciferase reporter assay system (Promega). The relative luciferase activities were calculated by normalizing the firefly luciferase activity with that from Renilla luciferase. The results were presented as the relative luciferase activity (means Ϯ S.E.) from at least three independent sets of experiments, each with five replicated measurements.
Cell Proliferation Assay-NCM460-NTR1 were seeded at 1 ϫ 10 4 cells/well in 96-well culture plates 24 h prior to incubation with 10 or 100 nM NT for 24 h in M3:D medium supplemented with 1% FBS and antibiotics. Cell proliferation assay was performed with CellTiter 96 AQueous nonradioactive cell proliferation assay (Promega) and cell proliferation ELISA, BrdU (colorimetric) (Roche Applied Science) according to the manufacturer's instructions.
Statistical Analysis-All of the results were derived from at least three sets of repeated experiments. The results are expressed as the means Ϯ S.D. The data were analyzed with Student's t tests. In all statistical comparisons, p Ͻ 0.05 was used to indicate significant differences.

RESULTS
␤-Arrestins Are Internalized with NTR1 to Endosomes-Previous studies identified NTR1 as class B GPCR that forms stable interactions with ␤ARR1 and ␤ARR2 in endosomes (21,22). ␤ARR1 and ␤ARR2 initiate internalization in NTR1 endocytosis by docking with the phosphorylated NTR1 (22). To determine whether ␤ARR1 and ␤ARR2 traffic with NTR1 to endosomes, we localized ␤ARR1, ␤ARR2, and NTR1 by confocal microscopy. Plasmids expressing either ␤ARR1 or ␤ARR2 tagged with GFP were transfected into NCM460-NTR1 cells 48 h prior to NT exposure, and NTR1 was localized with a specific antibody. As shown in Fig. 1, the majority of NTR1 before stimulation was localized at the plasma membrane, whereas ␤ARR1 and ␤ARR2 were distributed evenly throughout the cytoplasm. Following 30 min of NT stimulation, NTR1 was co-localized with ␤ARR1 and ␤ARR2 in endosomes (Fig. 1,  A and B, arrows). Thus, ␤ARR1 and ␤ARR2 were translocated with NTR1 to endosomes following stimulation with NT.
NTR1 Translocates to ECE-1 Containing Early Endosomes upon NT Stimulation-We next examined whether endogenous ECE-1, the metalloendopeptidase capable of degrading NT (56), was present in early endosomes in NCM460-NTR1 cells. ECE-1 was localized using an antibody to the ECE-1b and ECE-1d isoforms (42). In unstimulated cells, ECE-1 immunoreactivity co-localized with EEA1, a specific marker for early endosomes ( Fig. 2A), confirming the presence of endogenous ECE-1 in early endosomes, as previously shown in rat kidney epithelial cells (42). We next exposed NCM460-NTR1-GFP to NT for 30 min and localized NTR1-GFP and ECE-1 immunoreactivity by confocal microscopy. In unstimulated conditions, NTR1 was localized mainly at the plasma membrane, whereas ECE-1 was localized in endosomes (Fig. 2B). However, upon NT stimulation, NTR1 co-localized with ECE-1 (Fig. 2B, arrows), suggesting that this receptor translocates to early endosomes containing ECE-1 following NT exposure.
ECE-1 Degrades NT at Acidic pH-NT is an ECE-1 substrate (56). By analogy with previous studies of the role of ECE-1 in regulating endosomal trafficking and signaling of certain class B GPCRs (41-44), we hypothesized a similar role for ECE-1 in regulating NTR1. To assess whether ECE-1 could degrade NT within endosomes or at the plasma membrane, we incubated NT with recombinant human ECE-1 in buffer with physiological pH of early endosomes (pH 5.5) or of extracellular fluid (pH 7.4) and assessed degradation by HPLC. As shown in representative HPLC chromatograms, NT degradation products eluted at ϳ20 min and were detected only after incubation at endosomal pH (Fig. 3B) and not at extracellular fluid pH (Fig. 3, A-C). Time course experiments confirmed that ECE-1 degraded NT only in acidic, endosomal conditions (Fig. 3D). Analysis by mass spectrometry revealed that ECE-1 degraded NT at three sites, which would be expected to inactivate this neuropeptide/hormone (Fig. 3E). Thus, ECE-1 can degrade NT at the acidic pH of endosomes but not at the neutral pH of extracellular fluid.

ECE-1 Mediates Receptor
Recycling-In the case of the substance P and calcitonin gene-related peptide receptors, ECE-1 inhibition or knockdown causes retention of the receptors and   ␤ARRs in early endosomes and thereby prevents receptor recycling and resensitization of plasma membrane signaling (41,42). Bafilomycin A1, a vacuolar H ϩ ATPase inhibitor similarly prevents recycling because it impedes ECE-1 activity, which requires endosomal acidification. To examine whether the translocation of NTR1 from early endosomes also requires endogenous ECE-1 activity and endosomal acidification, NCM460-NTR1 cells were treated with the protein synthesis inhibitor cycloheximide (100 nM), ECE-1 inhibitor SM19712 (10 M), and the endosomal acidification inhibitor bafilomycin A1 (100 nM) 30 min prior to NT exposure (100 nM, 1 h) as previously described (20). The cells were then washed twice with ice-cold HBSS and allowed to recover in NT-free medium. The majority of NTR1 was recovered from early endosomes at 6 h after NT removal (Fig. 4A). Our results show that cycloheximide did not affect receptor endocytosis or NTR1 translocation to plasma membrane (Fig. 4B). However, inhibition of ECE-1 activity by 10 M SM19712 and inhibition of endosomal acidification by 100 nM bafilomycin A1 reduced translocation of NTR1 to plasma membrane, as shown by the retention of NTR1 in endosomes 6 h after recovery (Fig. 4, C and D). Our results suggest that recovery of NTR1 from early endosomes depends on endogenous ECE-1 activity in acidified endosomal vesicles.
ECE-1 Inhibition and ␤ARR1 and ␤ARR2 Gene Silencing Attenuate NTR1-dependent MAP Kinase Activation-The effectiveness of cell desensitization depends on the rates of receptor endocytosis and recycling. Although we have shown that NTR1 is recycled from ECE-1-containing acidified early endosomes in NCM460-NTR1, agonist-stimulated endocytosis of the NTR1 is mediated by ␤ARRs (21-23). NT stimulates MAP kinase signaling, especially ERK-dependent signaling (9,13), and activates NF-B (14,15). Both these pathways are involved in NT-mediated proinflammatory and cell proliferative responses (15,57), and ␤ARRs are also implicated in MAP kinase signaling of endocytosed GPCR (38). Therefore, we investigated whether ECE-1 and ␤ARRs contribute to NTstimulated MAP kinase signaling in colonocytes. NCM460-NTR1 cells were treated with siRNA to knockdown expression of ␤ARR1 or ␤ARR2. Fig. 5 (A and B) shows that the transcription levels of ␤ARR1 and ␤ARR2 in transfected cells are significantly reduced compared with cells transfected with scrambled siRNA. To examine whether ECE-1-mediated recycling contributes to ERK1/2 signaling, we treated cells with the ECE-1 inhibitor SM-19712. All the experiments were repeated 5 times and densitometric analysis was performed. Inhibition of ECE-1 activity also reduced ERK1/2 phosphorylation upon NT stimulation (Fig. 5C). On the other hand, although previous studies suggested that both ␤ARR1 and ␤ARR2 participate in GPCR internalization (22,23), our Western blot analysis showed that only ␤ARR1, but not ␤ARR2 silencing significantly reduced ERK1/2 phosphorylation upon 100 nM NT stimulation, whereas the basal phosphorylation levels were similar (Fig.  5C). Thus, NT-induced ERK1/2 signaling can be regulated by ␤ARR1 expression and ECE-1-dependent NTR1 receptor recycling and resensitization.
Next, we tested whether ␤ARR expression and ECE-1 activity influence the activity of JNK. We found that NT-induced JNK activation was attenuated by inhibition of NTR1 recycling with SM-19712 and by ␤ARR1 and ␤ARR2 silencing (Fig. 5D). Because the ability of NT to activate JNK has not previously reported, we examined the pathophysiologic consequences of this response in human colonocytes. NT enhanced cell proliferation (58,59), whereas increased JNK phosphorylation is linked to cell proliferation (60). NCM460-NTR1 cells were incubated in the presence or absence of NT for 24 h, and cell proliferation was assessed by the assay. As shown in Fig. 5E, NT (10 nM) increased cell proliferation, and co-incubation with the JNK inhibitor JNK(AII) reversed the effect. To confirm this response, we also used BrdU incorporation colorimetric assay. We found that NT-associated (10 nM, 24 h) increased BrdU incorporation was inhibited by JNK(AII) (by ϳ20%, n ϭ 5, p Ͻ 0.05). Because JNK phosphorylation is associated with inflammatory responses, we determined its involvement in NT-induced IL-8 secretion. NCM460-NTR1 cells were transfected with a plasmid encoding IL-8 promoter-driven luciferase or a control luciferase plasmid 48 h prior to NT stimulation in the presence or absence of JNK(AII). We found no significant difference in IL-8 promotor-driven luciferase activity between JNK(II)-treated and control cells (Fig. 5F). The reduction in luciferase activity in cells pretreated with the NF-B inhibitor caffeic acid phenethyl ester prior to NT exposure (Fig. 5F), confirmed that IL-8 transcription in NCM460-NTR1 cells involved NF-B activation (15). These results indicate that, at least in After washing with HBSS, the cells were allowed to recover in NT-free medium for 6 h. In unstimulated cells, NTR1 were distributed at the plasma membrane in all groups. NT-exposed NCM460-NTR1 treated with SM19712 and bafilomycin A1 show reduced NTR1 on the cell surface after recovery. Scale bars, 10 M.

NTR1 Internalization and Recycling Regulates NT Signaling
colonocytes, NTR1-associated JNK activation is linked to cell proliferation, but not to NF-B-driven IL-8 transcription.
ECE-1 Inhibition and ␤ARR1 and ␤ARR2 Gene Silencing Attenuate NT-induced NF-B Activation-We next investigated whether ECE-1 and ␤ARRs regulated NT-induced NF-B activation and subsequent NF-B-dependent IL-8 transcription. Western blot analysis of NCM460-NTR1 cells stimulated with 100 nM NT showed that p65 phosphorylation was attenuated by gene silencing of both ␤ARRs (Fig. 6A), and IB-␣ phosphorylation was attenuated by gene silencing of ␤ARR2 (Fig.   FIGURE 5. Inhibition of NTR1 endocytosis and ECE-1-dependent resensitization attenuated MAP kinase signaling. A and B, NCM460-NTR1 were transfected with scrambled siRNA (si-Control) and siRNA specific for ␤ARR1 (si-␤ARR1) or siRNA specific for ␤ARR2 (si-␤ARR2) 48 h prior to NT treatment. Transcription levels of ␤ARR1 and ␤ARR2 were compared against the control group and examined by quantitative PCR. C, NCM460-NTR1 cells were preincubated with SM19712 (10 M) or transfected with siRNA 48 h and subsequently treated with NT (100 nM) for 5 min. Equal amounts of proteins were subjected to Western blot analysis using anti-phospho-ERK1/2 antibody (p-ERK1/2) or anti-␤-tubulin to ensure equal loading. D, NCM460-NTR1 cells were preincubated with SM19712 (10 M) or transfected with siRNA 48 h and subsequently treated with NT (100 nM) for 40 min. Equal amounts of proteins were subjected to Western blot analysis using anti-phospho-JNK antibody (p-JNK) or anti-␤-tubulin to ensure equal loading. E, NCM460-NTR1 cells were pretreated with or without JNK(AII) (10 M) before incubated with NT for 24 h. Cell proliferation was assessed by microculture tetrazolium assay according to manufacturer's instructions (n ϭ 5). F, NCM460-NTR1 cells were transfected with plasmids expressing with IL-8 promoter-driven luciferase and plasmids expressing control luciferase construct 48 h before NT (100 nM) treatment for 1.5 h in the presence or absence of JNK(AII) and caffeic acid phenethyl ester (CAPE). Luciferase activity of each group was measured with luminometer. *, p Ͻ 0.05 when compared with intergroup NT-treated control. #, p Ͻ 0.05 when compared with intragroup NT-treated control. FIGURE 6. Inhibition of NTR1 endocytosis and ECE-1-dependent resensitization attenuates NF-B signaling. A and B, NCM460-NTR1 cells were treated with SM19712 (10 M) 30 min prior to NT treatment or transfected with siRNA specific for scrambled siRNA (si-Control) and siRNA specific for ␤ARR1 (si-␤ARR1) or siRNA specific for␤ARR2 (si-␤ARR2) 48 h prior to NT treatment. The cells were treated with NT (100 nM) for 1.5 h. Equal amounts of proteins were subjected to Western blot analysis using anti-phospho-p65 antibody (p-p65) (A) or anti-phospho-IB␣ antibody (p-IB␣) (B) or anti-␤tubulin to ensure equal loading. C, NCM460-NTR1 cells were subjected to the same treatment as above. The nuclei from different treatment groups were isolated and quantified. Equal amounts of nuclear lysates were used to quantify the concentration of p65 by ELISA. D, NCM460-NTR1 cells were transfected with plasmids expressing NF-B-driven luciferase and control luciferase 48 h prior to NT treatment. The cells were treated as stated above, and the luciferase activity was measured by luminometer. E, NCM460-NTR1 cells were pretreated with SM19712 (10 M) for 30 min or transfected with scrambled siRNA (si-Control) and siRNA specific for ␤ARR1 (si-␤ARR1) or ␤ARR2 (si-␤ARR2) 48 h before NT (100 nM) treatment for 6 h. Culture media were collected, and IL-8 secretion was analyzed with ELISA. F, NCM460-NTR1 cells were treated with SM19712 and siRNA transfection as mentioned above and underwent NT (100 nM) stimulation for 2 h. Transcription levels of TNF-␣ and IL-1␤ of different treatment groups were analyzed by quantitative PCR. *, p Ͻ 0.05 when compared with treatment control. 6B). Transcription of NF-B-dependent proinflammatory cytokines involves translocation of p65 to the nucleus and binding of the corresponding p65 promoter sequences to cytokine genes. We next isolated nuclear extracts from cells transfected with ␤ARR1 and ␤ARR2 siRNAs, before or after treatment with the ECE-1 inhibitor SM19712 and measured p65 by ELISA. Our results showed a significant reduction in NT-induced p65 levels in SM19712-treated cells as well as in ␤ARR2-silenced cells, whereas in contrast, ␤ARR1 silencing slightly increased p65 nuclear translocation (Fig. 6C). Interestingly, SM19712 reduced p65 phosphorylation at basal conditions and in response to NT (Fig. 6C). Moreover, NT-stimulated NF-Bdriven luciferase activity and IL-8 secretion were reduced after SM19712 treatment or ␤ARR2 silencing (Fig. 6, D and E). We next examined the outcome of ␤ARR1 and ␤ARR2 silencing in the expression of other inflammatory genes by NT. Quantitative PCR analysis performed on cDNA collected from NCM460-NTR1 after 2 h of NT exposure demonstrated reduced TNF-␣ and IL-1␤ mRNA levels in ␤ARR2-silenced cells, whereas ␤ARR1 silencing only reduced TNF-␣ mRNA (Fig. 6F).

DISCUSSION
Previous studies have shown that activated NTR1 internalizes by a ␤ARR-dependent process (21)(22)(23)(24). We have reported that NTR1 couples to proinflammatory pathways in human colonocytes, including NF-B and MAP kinase activation and secretion of neutrophil chemoattractants, such as IL-8 (14,15). However, the fate of NT/NTR1 complex following NT stimulation in colonocytes is not known, and whether NT/NTR1 internalization and recycling are associated with NT-induced proinflammatory and proliferative responses in these cells remains to be elucidated. Our present results in human colonocytes show that NT induces trafficking of NTR1, ␤ARR1, and ␤ARR2 to early endosomes containing ECE-1 and that NTR1 then slowly recycles back to the plasma membrane. ␤ARRs are required for the proinflammatory and proliferative signaling of NT, including activation of ERK1/2, JNK, and p65 and generation of IL-8. By degrading NT in acidified endosomes, ECE-1 mediates recycling of NTR1, where resensitization of plasma membrane signaling is necessary for the sustained proinflammatory actions of NT. Thus, we propose a mechanism (Fig. 7) in which NT-induced proinflammatory signaling responses involve a ␤ARRs-dependent endosomal mechanism and sustained NT-induced MAP kinase and NF-B activation require ECE-1-dependent NTR1 receptor recycling.
Step 4, free NTR1 is transported back to the plasma membrane in recycling endosomes for resensitization.
Structural and in vivo evidence indicate that ␤ARR1 and ␤ARR2 are, to some extent, functionally redundant (63,64). However, we observed a stronger attenuation in NTR1-dependent ERK and JNK phosphorylation after ␤ARR1 silencing compared with ␤ARR2 knockdown (Fig. 5, B-D). In contrast, NF-B p65 phosphorylation and IL-8 production were more strongly reduced in ␤ARR2-silenced colonocytes compared with ␤ARR1-silenced cells (Fig. 6, A, E, and F). Moreover, NTinduced IB-␣ phosphorylation, NF-B-driven luciferase activity, and p65 translocation were only reduced after ␤ARR2 silencing (Fig. 6, B-D). These results suggest that NTR1 signaling is ␤ARRs-dependent and that ␤ARR1 and ␤ARR2 may play different roles in promoting inflammatory signal transduction pathways in the context of NTR1 activation.
Our results suggest that only ␤ARR1 but not ␤ARR2 is required for NT-induced activation of ERK1/2, yet ␤ARR2 appears to be involved in NT-induced NF-B activation. Because evidence indicates that in many signaling pathways NF-B activation involves ERK1/2, our finding of an independent association of ␤ARR1 and ␤ARR2 with ERK1/2 and NF-B signaling pathways may appear contradictory. However, recent studies have identified ERK-independent NF-B activation in rheumatoid synovial fibroblasts and hepatocellular HepG2 cells (65,66). Although our results suggest that ␤ARR2associated NF-B activation is ERK-independent, the mechanism(s) involved in the participation of ␤ARRs in NTR1-associated MAP kinase and NF-B activation in colonocytes are not entirely clear. Moreover, ␤ARR1 and ␤ARR2 have distinct roles in other systems. For example, previous studies have shown that ␤ARR1 can form signalsomes with MAP kinases, such as c-Src, ERK, and JNK3 (30,(32)(33)(34), whereas ␤ARR2 recruits CARMA3, a scaffold protein involved in lysophosphatidic acidinduced NF-B activation and subsequent IL-6 production (67). Moreover, the NF-B inhibitor IB-␣ co-immunoprecipitates with ␤ARR2 in mammalian cells and a yeast two-hybrid system (62,68). We have also shown that NT-induced ERK activation is Ras-GTPase-related and NF-B activation is Rho-GTPase-dependent (9,14), although both signaling cascades are involved in IL-8 secretion (9). Therefore, additional studies are required to further elucidate the involvement of the two ␤-arrestin isoforms in proinflammatory signaling in colonocytes.
We demonstrated that in human NCM460-NTR1 colonocytes, NTR1 was translocated to ECE-1-containing acidified early endosomes along with ␤ARR1 and ␤ARR2 (Figs. 1 and 2). Endosomal ECE-1 regulates intracellular signaling at two levels. ECE-1 promotes receptor recycling and sensitization of plasma membrane signaling of receptors for substance P and calcitonin gene-related peptide (42,44). On the other hand, endosomal ECE-1 also terminates signaling by destabilizing the endosomal MAP kinase signalosome after ligand degradation in acidified early endosomes (48). Our results indicate that recycling of NTR1 promoted by ECE-1 may be more important for sustained NT-induced proinflammatory signaling (Figs. 4, 5, C and D, and 6). Similarly to NTR1, neurokinin-1 receptors, the receptor of substance P, is a class B GPCR (21). When stimulated by high ligand concentrations, both NTR1 and neurokinin-1 receptors are translocated to perinuclear sorting endo-somes by Rab5a, which is present in early endosomes and remains in perinuclear sorting endosomes for more than 60 min (25,69), whereas when low concentrations of substance P are used, neurokinin-1 receptor is rapidly recycled from early endosomes (69). It is likely that in human colonocytes, inhibition of ECE-1 activity promotes the translocation of NTR1 to perinuclear sorting endosomes and depletes NTR1 on plasma membrane (25), thus reducing NT-associated signaling.
In the present study, we found that NT induced JNK phosphorylation via a pathway that involved NTR1 internalization and recycling, whereas inhibition of NT-induced JNK activation leads to reduced cell proliferation in colonocytes (Fig. 5, D  and E). Increased JNK phosphorylation is evident in the inflamed colonic mucosa of patients with inflammatory bowel disease (70,71), whereas NTR1 has been associated with mucosal healing following colitis (7), suggesting that JNK activation in response to NT in colonocytes may be involved in the pathophysiology of mucosal healing.
In summary, NT-induced MAP kinase signaling and NF-B activation in human nontransformed colonocytes is regulated by NTR1 internalization and recycling. Both MAP kinase and NF-B activities are ␤ARRs-dependent, and rapid recycling of internalized NTR1 from early endosomes by ECE-1 is essential for signaling transduction. Further investigation is needed to address whether NTR1 internalization and recycling participate in the proinflammatory mechanisms by which NTR1 modulates colitis.