Abstract
Transient receptor potential vanilloid 1 (TRPV1) is a calcium-selective ion channel expressed in human lung cells. We show that activation of the intracellular subpopulation of TRPV1 causes endoplasmic reticulum (ER) stress and cell death in human bronchial epithelial and alveolar cells. TRPV1 agonist (nonivamide) treatment caused calcium release from the ER and altered the transcription of growth arrest- and DNA damage-inducible transcript 3 (GADD153), GADD45α, GRP78/BiP, ATF3, CCND1, and CCNG2) in a manner comparable with prototypical ER stress-inducing agents. The TRPV1 antagonist N-(4-tert-butylbenzyl)-N′-(1-[3-fluoro-4-(methylsulfonylamino)-phenyl]ethyl)thiourea (LJO-328) inhibited mRNA responses and cytotoxicity. EGTA and ruthenium red inhibited cell surface TRPV1 activity, but they did not prevent ER stress gene responses or cytotoxicity. Cytotoxicity paralleled eukaryotic translation initiation factor 2, subunit 1 (EIF2α) phosphorylation and the induction of GADD153 mRNA and protein. Transient overexpression of GADD153 caused cell death independent of agonist treatment, and cells selected for stable overexpression of a GADD153 dominant-negative mutant exhibited reduced sensitivity. Salubrinal, an inhibitor of ER stress-induced cytotoxicity via the EIF2αK3/EIF2α pathway, or stable overexpression of the EIF2α-S52A dominant-negative mutant also inhibited cell death. Treatment of the TRPV1-null human embryonic kidney 293 cell line with TRPV1 agonists did not initiate ER stress responses. Likewise, n-benzylnonanamide, an inactive analog of nonivamide, failed to cause ER calcium release, an increase in GADD153 expression, and cytotoxicity. We conclude that activation of ER-bound TRPV1 and stimulation of GADD153 expression via the EIF2αK3/EIF2α pathway represents a common mechanism for cytotoxicity by cell-permeable TRPV1 agonists. These findings are significant within the context of lung inflammatory diseases where elevated concentrations of endogenous TRPV1 agonists are probably produced in sufficient quantities to cause TRPV1 activation and lung cell death.
Lung cell damage causes acute respiratory distress and contributes to the pathogenesis of chronic lung diseases (Knight and Holgate, 2003). Evidence suggests that the transient receptor potential vanilloid type-1 receptor (TRPV1, capsaicin receptor, VR1; Hs. 268606) may be a mediator of lung pathologies caused by xenobiotic toxicants and endogenous agonists as well as a therapeutic target for treating and/or preventing lung disorders (Jia et al., 2005; Szallasi et al., 2006).
TRPV1 is widely expressed in the respiratory tract, including nasal mucosal cells (Seki et al., 2006), C-fiber neurons and airway smooth muscle cells (Mitchell et al., 2005; Watanabe et al., 2005), and alveolar and bronchial epithelial cells (Veronesi et al., 1999; Reilly et al., 2003; Agopyan et al., 2004). TRPV1 is selectively activated by capsaicin, the primary pain-producing chemical in hot peppers, and a variety of exogenous and endogenous respiratory toxicants, including anandamide (Van Der Stelt and Di Marzo, 2004), products of arachidonic acid metabolism by lipoxygenases (Hwang et al., 2000), H2S (Trevisani et al., 2005), ethanol (Trevisani et al., 2004), acids (Tominaga et al., 1998; Ricciardolo et al., 2004), and particulate pollutants (Veronesi et al., 1999; Agopyan et al., 2004). Capsaicin and other TRPV1 agonists are routinely used to study the TRPV1 pharmacology and have proven instrumental in defining the physiological roles of TRPV1 in the lung and other organs. Here, we use capsaicin to elucidate toxicological phenomena associated with TRPV1 activation in lung cells.
Capsaicin is used clinically to induce cough (Morice et al., 2001) and to treat rhinitis (van Rijswijk and Gerth van Wijk, 2006). However, numerous case reports have described adverse respiratory effects and death in humans following exposures to concentrated capsaicinoid aerosols (Heck, 1995; Steffee et al., 1995; Billmire et al., 1996). In animal models, high doses of capsaicin cause acute respiratory and cardiovascular failure, independently of the route of administration (Glinsukon et al., 1980). Inhalation of capsaicinoids by rats causes lung inflammation and widespread damage to tracheal, bronchial, and alveolar cells (Reilly et al., 2003). In vitro studies with human bronchial epithelial cells have demonstrated two principal outcomes associated with TRPV1 activation: proinflammatory cytokine (interleukin-6 and interleukin-8) production and oncotic cell death (Reilly et al., 2003, 2005). Cytokine synthesis and cell death were inhibited by TRPV1 antagonists that prevented calcium release from the endoplasmic reticulum (ER) and included LJO-328, SC0030, and 5-iodo-RTX. Conversely, inhibition of the cell surface population of TRPV1 using EGTA, ruthenium red, and calcium-free media only prevented cytokine responses.
In mammalian cells, depletion of ER calcium initiates a homeostatic stress response program termed ER stress. ER stress is generally initiated by a reduction in protein processing efficiency in the ER, and its roles in human diseases and xenobiotic toxicities have been reviewed (Cribb et al., 2005; Schroder and Kaufman, 2005; Zhang and Kaufman, 2006). ER stress is predominantly regulated by three sensors: activating transcription factor 6 (ATF6; Hs. 492740), eukaryotic initiation factor 2α kinase-3 (EIF2αK3 or PERK; Hs. 591589), and ER to nucleus signaling 1 and 2 (a.k.a. IRE1α and β; Hs. 133982 and Hs. 592041) (Schroder and Kaufman, 2005). Activation of one or more of these proximal sensors is dependent upon the type of cellular stress. For example, the prototypical ER stress-inducing agent thapsigargin preferentially activates the “translational branch” involving EIF2αK3. Activated EIF2αK3 catalyzes the phosphorylation of cytosolic EIF2α (Hs. 151777) (Lu et al., 2004; Boyce et al., 2005). Heterodimerization of EIF2α-P with EIF2β promotes ATF4 translation (Hs. 496487) and inhibits the translation of “nonessential” genes (Wek et al., 2006). ATF4 translocates to the nucleus where it modulates the expression of a subset of stress-response genes that include ATF3, GADD153, CCND1, and BiP/GRP78 (see Table 1 for UniGene identifications numbers). Phosphorylation of EIF2α is considered protective (Lu et al., 2004; Boyce et al., 2005), but increased expression of GADD153, as a consequence of EIF2α phosphorylation, causes cell cycle arrest at G1/S and cell death (Oyadomari and Mori, 2004).
In this study, we tested the hypothesis that activation of the intracellular ER subpopulation of TRPV1 by prototypical and endogenous TRPV1 agonists would disrupt ER calcium homeostasis and activate EIF2αK3-dependent ER stress responses to cause cytotoxicity. The data obtained from this work imply that a common mechanism of cytotoxicity exists for cell-permeable TRPV1 agonists and that conditions that promote TRPV1 activation in vivo (e.g., inflammation, inhalation of polluted air) may promote lung pathologies through TRPV1- and EIF2αK3-dependent procytotoxic ER stress pathways.
Materials and Methods
Chemicals. Structures of the TRPV1 agonists and antagonists used in this study are shown in Fig. 1. Nonivamide (n-vanillyl-nonanamide), sulfinpyrazone, dithiothreitol (DTT), H2O2, ruthenium red, EGTA, benzylamine-HCl, and nonanoyl chloride were purchased from Sigma-Aldrich (St. Louis, MO). LJO-328 was generously provided by Dr. Jeewoo Lee (Seoul National University, Seoul, Korea). Thapsigargin and 5-iodo-resiniferatoxin were purchased from Axxora Life Sciences, Inc. (San Diego, CA). Salubrinal (EIF2α-inhibitor) was purchased from Calbiochem (San Diego, CA). PCR primers were purchased from Integrated DNA Technologies (Coralville, IA). n-Benzylnonanamide was synthesized by reacting benzylamine-HCl and nonanoyl chloride in 0.1 M NaOH and collecting the precipitate. Product structure was verified by liquid chromatography-tandem mass spectrometry (m/z 248) and 1H and 13C NMR. Purity was estimated to be ∼98% by high-performance liquid chromatography/UV analysis (230 nm). See supplemental data for chemical analysis data. All other chemicals and reagents were purchased from established suppliers.
Cell Culture. BEAS-2B human bronchial epithelial cells (CRL-9609) were purchased from American Type Culture Collection (Manassas, VA). TRPV1-overexpressing cells were generated as described previously (Reilly et al., 2003). BEAS-2B and TRPV1-overexpressing cells were cultured in LHC-9 media (BioSource International, Camarillo, CA). Normal human bronchial epithelial (NHBE) cells, a primary cell line, were purchased from Cambrex Bio Science Walkersville, Inc. (Walkersville, MD) and cultured in BEGM media. Human embryonic kidney (HEK)-293 human embryonic kidney (CRL-1573) and A549 human lung carcinoma (CCL-185) cells were purchased from American Type Culture Collection and were cultured in Dulbecco's modified Eagle's medium:F-12 containing 10% fetal bovine serum (Hyclone Laboratories, Logan, UT). Culture flasks for BEAS-2B and TRPV1-overexpressing BEAS-2B cells were coated with LHC basal media fortified with 30 μg/ml collagen, 10 μg/ml fibronectin, and 10 μg/ml bovine serum albumin. Cells were maintained between 30 and 90% maximal density and were subcultured every 2 to 4 days.
Fluorometric Calcium Flux Assays. TRPV1-overexpressing cells were used to evaluate calcium flux. Flux in BEAS-2B, A549, and NHBE cells was not detectable. Functional evidence provided here and in previous studies (Reilly et al., 2003, 2005; Johansen et al., 2006) demonstrates that the TRPV1-overexpressing cells model responses of BEAS-2B and other lung cells when treated with diverse TRPV1 agonists, with the exceptions that TRPV1-dependent calcium flux is quantifiable and dose responses for TRPV1 agonists are shifted to lower concentrations. To assay calcium flux, TRPV1-overexpressing cells were subcultured into 96-well culture plates and grown to ∼90% maximal density. Cells were loaded with the fluorogenic calcium indicator Fluo-4-acetoxymethyl ester (2.5 μM) (Invitrogen, Carlsbad, CA) for 90 min at room temperature (∼22°C) in LHC-9 media containing 200 μM sulfinpyrazone. Cells were washed and incubated for an additional 20 min at room temperature to permit methyl ester hydrolysis and activation of Fluo-4. Changes in cellular fluorescence in response to agonist and antagonist treatments were assessed microscopically on cell populations 1 min after treatments using methods described previously (Reilly et al., 2005; Johansen et al., 2006). ER calcium flux was evaluated by pretreating cells with 2.5 μM thapsigargin for 5 min followed by addition of 2.5 μM nonivamide. Calcium flux due to cell surface TRPV1 activity was assessed by treating cells with nonivamide in calcium-free media containing 50 μM EGTA and 250 μM ruthenium red. Differences in fluorescence responses observed between the treatments and controls were used to assess the relative contribution of ER-bound and cell surface TRPV1 in total calcium flux initiated by agonists. Data are expressed as -fold change in fluorescence intensity.
Cytotoxicity Assays. Cells were subcultured into Multiwell plates and allowed to reach ∼90% confluence. The cells were treated for 24 h with various agonists and antagonists prepared in the appropriate culture media without fetal bovine serum. Cell viability was assessed using the Dojindo cell counting kit-8 (Dojindo Laboratories, Gaithersburg, MD), according to the supplier's recommendations. Loss of cell monolayer integrity due to treatment with toxic TRPV1 agonists was confirmed microscopically. Toxicity data are expressed as the percentage of remaining viable cells relative to untreated controls, calculated using the absorbance ratio of the formazan dye product generated from the Dojindo reagent.
RT-PCR Analysis. Cells were subcultured into 25-cm2 cell culture flasks, grown to a density of ∼90%, and treated with TRPV1 agonists and antagonists. Total RNA was extracted from cells using the RNeasy RNA isolation kit (QIAGEN, Valencia, CA), and 2.5 μg of total RNA was transcribed into cDNA using PolyT and Superscript III (Invitrogen). cDNA corresponding to GADD153, GADD45α, ATF3, CCND1, CCNG2, BiP/GRP78, and β-actin was amplified by PCR from 1 μl of the cDNA synthesis reaction using the primers listed in Table 1 and GoTaq Green PCR Master Mix (Promega, Madison, WI). The PCR program consisted of an initial 2 min incubation at 94°C and 28 cycles of 94°C (30 s), 55°C (30 s), and 72°C (30 s). A final extension period of 10 min at 72°C followed. PCR products were resolved on 1% SB agarose gels, and images were captured using a Gel-Doc imaging system (Bio-Rad, Hercules, CA). Product quantification was achieved by determining the band intensities for each PCR product relative to β-actin, the internal PCR control, using the Gel Doc density analysis tools in the Quantity One software (Bio-Rad). Experiments were reproduced a minimum of three times on different passages of cells.
Cloning of ER Stress Gene cDNA. The full-length cDNA for human GADD153, ATF3, EIF2α, and ATF4 were amplified from BEAS-2B cells using Phusion GC-rich PCR Super Mix (New England Biolabs, Ipswich, MA). The following primers were used: GADD153 (+) 5′-CACCATGGCAGCTGAGTCATTGCCTTTC-3′ and (–) 5′-TGCTTGGTGCAGATTCACCATTC-3′, ATF3 (+)5′-CACCATGATGCTTCAACACCCAG-3′, and (–)5′-ATACTGAAGCTGCAGGCACTC-3′, EIF2α (+) 5′-CACCATGCCGGGTCTAAGTTGTAG-3′ and (–) 5′-ATCTTCAGCTTTGGCTTCCATTTC-3′, and ATF4 (+) 5′-CACCATGACCGAAATGAGCTTCCTG-3′ and (–) 5′-GGGGACCCTTTTCTTCCCCCTTG-3′. These primers incorporated a 5′-CACC sequence immediately before the ATG start site to permit directional cloning into the pcDNA3.1-V5/His6 mammalian expression vector (Invitrogen) and eliminated the stop codon to allow for epitope tagging with V5-His6. An expression plasmid for p58IPK (pcDNA1-p58IPK) was generously provided by Dr. Michael G. Katze (University of Washington, Seattle, WA). The pMaxGFP expression vector was purchased from Amaxa Biosystems (Gaithersburg, MD). All clones were sequence verified by comparison to the appropriate GenBank sequences. Plasmids used in the transient transfection assays were simultaneously purified using the QIAGEN Plasmid DNA Midi-Prep kit and further purified using the GeneElute HP Plasmid Mini-Prep kit (Sigma-Aldrich).
Site-Directed Mutagenesis. The GADD153-L134A/L141A (Matsumoto et al., 1996) and EIF2α-S52A (Srivastava et al., 1998) dominant-negative mutants were constructed using the QuikChange XL site-directed mutagenesis kit (Stratagene, La Jolla, CA) and the following primers: GADD153-L134A/L141A (+) 5′-GGCACAGGCAGCTGAAGAGAATGAACGGGCCAAGCAGG-3′ and (–)5′-CCTGCTTGGCCCGTTCATTCTCTTCAGCTGCCTGTGCC-3′ and EIF2αS52A (+) 5′-CTTCTTAGTGAATTAGCCAGAAGGCG-3′ and (–)5′-GGATACGCCTTCTGGCTAATTCACTA-3′.
Transient Overexpression Assays and Stable Overexpressing Cell Lines. A459 cells respond to TRPV1 agonists similar to BEAS-2B, NHBE, and TRPV1-overexpressing cells, with the exception that they exhibit slightly reduced sensitivity to agonists due to lower levels of TRPV1 expression (Reilly et al., 2003). A549 cells were used as transfection hosts to evaluate the protoxic effects of ER stress-induced gene products in lung cells, because they exhibited reproducibility in transfection efficiency and limited toxicity due to transfection reagents. Transfection efficiency typically reached ∼80% using A549 cells versus ∼5 to 10% with BEAS-2B cells, or <1% using NHBE cells. This level of transfection was necessary to evaluate the effects of ER stress genes on cell populations. A549 cells were subcultured into 48-well cell culture plates and grown to a density of ∼70 to 80%. Cells were washed with Opti-MEM media and transfected for ∼18 h using Lipofectamine 2000 (Invitrogen) at a ratio of 3:1 lipid/plasmid DNA. After transfection, cells were washed with Opti-MEM and allowed to grow for an additional 24 h. Cell viability was assessed as described above. All experiments were performed in triplicate and were normalized to control cells transfected with equal quantities of the pMaxGFP plasmid.
Stably overexpressing cell lines were generated by culturing transfected A549 cells in media fortified with 600 μg/ml G-418 (Geneticin; Invitrogen) for ∼3 weeks. Resistant foci were isolated and expanded in selective media. Individual clones were screened for overexpression of the target genes by assaying for V5-His6 expression by RT-PCR and subsequently used for cytotoxicity screening.
Western Blotting. BEAS-2B cells were grown to ∼90% maximal density in 25-cm2 flasks. Before treatment cells were cultured in fresh media for 2 h. Cells were treated for 0, 1, 2, 4, and 8 h, rinsed with phosphate-buffered saline, and immediately lysed on ice using 20 mM HEPES, pH 7.5, containing 150 mM NaCl, 1% Triton X-100, 1 mM EDTA, 10 mM sodium pyrophosphate, 100 mM sodium fluoride, 17.5 mM β-glycerophosphate, 1 mM phenylmethylsulfonyl fluoride, 4 mg/ml aprotinin, and 2 mg/ml pepstatin A. The lysates were clarified by centrifugation at ∼20,000g for 15 min at 4°C, and the concentration of protein was determined using the BCA protein assay (Pierce Chemical, Rockford, IL). Fifty micrograms of soluble protein from each sample then was resolved on a 10% NuPAGE gel (Invitrogen) and subsequently transferred to polyvinylidene difluoride membrane. The blots were probed for EIF2α-P using a rabbit polyclonal IgG fraction specific to EIF2α-pS52 (BioSource, International) according to supplier protocols. GADD153 expression was determined using an anti-GADD153 antibody from Biolegend (San Diego, CA) and the protocol provided by the supplier.
Statistical Analysis. Statistical testing used the paired t tests and ANOVA with post hoc testing using Dunnett's test to determine significance. A 95% confidence interval was used as the limit for significance. Specific details on statistical analyses are presented in the figure legends.
Results
Treatment of TRPV1-overexpressing cells with 2.5 μM nonivamide produced marked increases in cytosolic calcium due to release of calcium from ER stores (Fig. 2A). EGTA and ruthenium red cotreatment had little to no effect on calcium flux, but cotreatment with LJO-328 or prior depletion of ER calcium stores with thapsigargin completely prevented calcium flux. n-Benzylnonanamide failed to elicit ER calcium release at 2.5 μM (Fig. 2A) or at concentrations up to 25 μM (data not shown). Treatment of TRPV1-overexpressing cells with 1 μM nonivamide caused an approximate 50% loss in cell viability after a 24-h period (Fig. 2B). Cell death corresponded to a loss of monolayer consistency (data not shown) and was inhibited by LJO-328 cotreatment, but not by EGTA and ruthenium red. n-Benzylnonanamide did not cause cell death, consistent with a lack of TRPV1 activation.
Analysis of collective genetic responses in TRPV1-overexpressing and BEAS-2B cells exposed to 1 and 100 μM (LC50 concentrations) nonivamide, respectively, for 4 h, in the presence or absence of LJO-328, by microarray yielded preliminary insight into cellular processes that constituted the cell death process (see Supplemental Data, explanation of microarray data and microarray data). Increased expression of GADD153, GADD45α, ATF3, CCNG2, and BiP/GRP78 mRNA and a decrease in CCND1 mRNA were observed, and these responses were validated by RT-PCR (Fig. 3A). Cotreatment of cells with the TRPV1 antagonist LJO-328 prevented changes in gene expression, whereas little to no inhibition was observed using EGTA and ruthenium red (Fig. 3A). Treatment of BEAS-2B cells with the prototypical ER stress-inducing agents thapsigargin and DTT produced similar changes in the expression of GADD153, GADD45α, ATF3, CCND1, CCNG3, and BiP/GRP78 (Fig. 3B) mRNA.
BEAS-2B cells treated with 100 and 200 μM nonivamide also exhibited a shift in the relative amount of EIF2α-P and an increase in the expression of GADD153 mRNA and protein (Fig. 4, A–C). EIF2α phosphorylation and GADD153 expression was inhibited by LJO-328, but not by EGTA and ruthenium red (Fig. 4A). The kinetic and dose-dependent features of GADD153 induction and EIF2α-P accumulation paralleled cytotoxicity (Fig. 4, A–C). The highest levels of EIF2α phosphorylation and GADD153 protein were detected at ∼8 h with 200 μM nonivamide. For GADD153, increases in mRNA in BEAS-2B cells was maximal at ∼4 h and occurred at concentrations >150 μM. Similar responses were observed using the TRPV1-overexpressing cells, but maximal increases in protein and mRNA were observed with a dose of 1 to 2 μM (data not shown).
Transient overexpression of GADD153 in A549 cells produced an approximate 50% loss in cell viability relative to pMaxGFP-transfected control cells in the absence of cytotoxic TRPV1 agonists (Fig. 5A). Transient overexpression of ATF4, which stimulates GADD153 transcription, also produced ∼20% cell death. GADD153-L134A/L141A, ATF3, or p58IPK were not cytotoxic. Transient cotransfection of A549 cells with ATF3 and GFP (10:1) yielded a high proportion of viable GFP-expressing cells 48 h after the transfection procedure (Fig. 5B). No ethidium bromide (EtBr)-stained nuclei were observed in these cells, indicating cellular integrity. Conversely, very few cells transfected with GADD153 and GFP (10:1) survived, whereas those that remained attached to the culture dish exhibited intense nuclear staining with EtBr. These data were consistent with a loss of cell viability, cell membrane integrity, and oncotic cell death, as reported previously for BEAS-2B and A549 cells treated with capsaicin (Reilly et al., 2003).
Inhibition of cytotoxicity using dominant-negative forms of EIF2α (EIF2α-S52A) and GADD153 (GADD153-L134A/L141A) was also evaluated (Fig. 6). Figure 6A shows that both the EIF2α-S52A- and GADD153-L134A/L141A-overexpressing A549 cells were less susceptible to cytotoxicity by nonivamide. Likewise, the addition of salubrinal to treatment solutions containing 1 or 100 μM nonivamide inhibited cell death in TRPV1-overexpressing and BEAS-2B cells with a maximal effect between 2.5 and 5 μM (Fig. 6B). Salubrinal inhibits EIF2αK3-induced cytotoxicity (Lu et al., 2004; Boyce et al., 2005)
Induction of the proapoptotic/oncotic ER stress-induced gene GADD153 was also compared in TRPV1-overexpressing, BEAS-2B, A549, and NHBE lung cells as well as HEK-293 cells (Table 2). All four lung cell types express TRPV1, but HEK-293 cells do not. Significant (6- to 8-fold) GADD153 mRNA induction was observed following 4-h treatment of BEAS-2B, TRPV1-overexpressing, A549, and NHBE cells with LC50 concentrations of nonivamide, resiniferatoxin, and anandamide, but not with n-benzylnonanamide. Interestingly, n-benzylnonanamide inhibited cell death caused by nonivamide in the TRPV1-overexpressing cells at concentration ratios >5:1 (data not shown). Induction of GADD153 transcription was attenuated by LJO-328 in all cells types exhibiting a response as well as by 5-iodo-RTX in the TRPV1-overexpressing line. GADD153 induction was not observed in HEK-293 cells treated with nonivamide or resiniferatoxin.
Discussion
Previous studies of TRPV1 and the effects of its agonists on cultured lung cells and in animal models of airway injury support the hypothesis that TRPV1 is a mediator of lung injury and inflammation (Reilly et al., 2003, 2005; Vargaftig and Singer, 2003; Li et al., 2005; Trevisani et al., 2005; Bhatia et al., 2006; Geppetti et al., 2006). However, precise molecular mechanisms of cell death have not been established.
Quantitation of calcium flux in TRPV1-overexpressing cells demonstrated that 85 to 90% of functional TRPV1 existed in the ER membrane (Fig. 2A). Selective inhibitors of TRPV1 and treatments that reduced the passage of calcium ions from extracellular sources into cells (Fig. 2, A and B) confirmed previous data demonstrating a correlation between ER calcium release and cytotoxicity in TRPV1-overexpressing cells (Reilly et al., 2005). Although calcium flux was not detected in BEAS-2B, NHBE, or A549 cells, results presented here demonstrate that the TRPV1-overexpressing cells model the TRPV1 agonist-induced effects in these cell types.
cDNA microarray analysis (see supplemental data, explanation of microarray data and microarray data) demonstrated that TRPV1 activation was associated with changes in the expression of several prototypical ER stress genes in lung cells. Comparisons between gene expression changes elicited by nonivamide in the presence and absence of LJO-328 and EGTA/ruthenium red (Fig. 3A) and changes elicited by the prototypical ER stress inducing-agents thapsigargin and DTT (Fig. 3B) support our conclusion that TRPV1 activation causes ER stress. Furthermore, ER stress proceeded via pathways similar to those activated by thapsigargin and DTT (Schroder and Kaufman, 2005).
ER stress responses are compensatory responses. Up-regulation of specific gene products through dedicated signaling pathways, coupled with cell cycle arrest and a temporary halt of general transcription and translation, are coordinated processes that have evolved to help cells overcome inefficiencies in protein processing (Schroder and Kaufman, 2005). Alterations in ER processing efficiency occur with nutrient deprivation, viral infection, disruption of cellular redox state, changes in ER folding environment (e.g., alterations in calcium homeostasis, redox state), expression of unstable polymorphic variant proteins, and toxicant exposures (Cribb et al., 2005; Schroder and Kaufman, 2005). If cells cannot compensate for a specific stress, they die.
ER stress-induced cell death has been primarily attributed to the expression of GADD153 following EIF2αK3 activation (Matsumoto et al., 1996; McCullough et al., 2001; Oyadomari and Mori, 2004). GADD153 inhibits cell proliferation by reducing the expression of CCND1, and it causes cell death by sequestering the antiapoptotic Bcl-2 protein and inhibiting nuclear factor-κB and Akt/protein kinase B-mediated cytoprotective processes (McCullough et al., 2001; Hu et al., 2004; Hyoda et al., 2006). The balance between cell death and survival ultimately depends upon the level of GADD153 expression and the coexpression of other pro- and anticytotoxic gene products that participate in ER stress responses.
Treatment of BEAS-2B cells with nonivamide promoted the phosphorylation of EIF2α at serine 52 (Fig. 4A). This was indicative of EIF2αK3 activation. EIF2α phosphorylation was associated with increased expression of GADD153 expression (Figs. 3A and 4, A and B). Increased concentrations of EIF2α-P and GADD153 correlated with the onset of cell death in BEAS-2B cells, as determined using dose- and temporal-response correlations with protein and mRNA (Fig. 4, A–C). These trends were reproduced using the TRPV1-overexpressing line. EIF2α phosphorylation and GADD153 expression were attenuated by LJO-328, but not by EGTA or ruthenium red. n-Benzylnonanamide, a pharmacologically inactive nonivamide analog, did not promote ER calcium release or induce GADD153 expression in BEAS-2B or any other cells tested, and it was nontoxic at concentrations equal to or in 2-fold excess of nonivamide (Fig. 2, A and B; Table 2). These data support our conclusion that TRPV1 activation promotes cytotoxicity via activation of EIF2αK3, phosphorylation of EIF2α, and expression of GADD153.
To substantiate the role of GADD153 in cell death, we cloned this gene and transiently transfected A549 cells with the expression construct. Performing transient transfection studies in the BEAS-2B and NHBE cells were hampered by variable transfection efficiency and high levels of toxicity due to transfection reagents. As such, we used A549 cells as the model for these experiments. We have previously shown that A549 cells respond to TRPV1-agonists similar to BEAS-2B cells (Reilly et al., 2003). Cells transfected with GADD153 exhibited reduced viability due to loss of cells from the culture wells (Fig. 5, A and B). Cytotoxicity and cell loss relative to controls were not observed with GADD153-L134A/L141A, ATF3, or p58IPK, but toxicity was observed with ATF4. These results were consistent with the established roles of these proteins (Schroder and Kaufman, 2005). Specifically, ATF3 and p58IPK limit ER stress responses by inhibiting ATF4-dependent gene transcription and the phosphorylation of EIF2α by EIF2αK3, respectively. Conversely, ATF4 promotes GADD153 transcription, and GADD153 is procytotoxic. Additional support for GADD153 as the ultimate mediator of cytotoxicity was obtained by treating A549 cells that stably overexpressed the GADD153-L134A/L141A dominant-negative mutant (Matsumoto et al., 1996). Overexpression of GADD153-L134A/L141A markedly reduced cytotoxicity caused by nonivamide (Fig. 6A). Data in Figs. 5 and 6 imply that GADD153 was the primary cause of cytotoxicity in lung cells treated with TRPV1 agonists.
The effects of modifying the EIF2αK3/EIF2α signaling were also evaluated. Two approaches were used: stable overexpression of the EIF2α-S52A dominant-negative mutant in A549 cells (Srivastava et al., 1998) and pharmacological stabilization of EIF2α-P in BEAS-2B and TRPV1-overexpressing cells using salubrinal (Boyce et al., 2005). Interestingly, squelching of EIF2α phosphorylation (Fig. 6A) and inhibition of EIF2α dephosphorylation (Fig. 6B) protected cells from toxicity. Initially, these data seemed contradictory, but literature supports a dual role for EIF2α-P in regulating cell survival and death during ER stress. Thus, the results in Fig. 6, A and B, highlight this dual effect of the EIF2αK3/EIF2α pathway. However, the molecular basis for these antithetical responses remains enigmatic.
We also investigated whether ER stress represented a common mechanism of cytotoxicity for structurally diverse TRPV1 agonists. Table 2 shows that transcriptional activation of GADD153 occurred in BEAS-2B, A549, NHBE, and TRPV1-overexpressing cells treated with LD50 concentrations of nonivamide, resiniferatoxin, and anandamide. As predicted, TRPV1 agonists failed to induce GADD153 expression in the TRPV1-null HEK293 cell line (Table 2). Likewise, n-benzylnonanamide failed to elicit GADD153 expression, confirming the direct link between TRPV1 activation, GADD153 expression, and cell death. This conclusion was also supported by the inhibition of GADD153 expression by LJO-328 and 5-iodo-RTX (Table 2). The inability of 5-iodo-RTX to completely inhibit GADD153 expression in the BEAS-2B cell line was consistent with our previous findings that 5-iodo-RTX (like capsazepine) causes cytotoxicity at elevated concentrations (Reilly et al., 2003, 2005).
Collectively, the results presented by this study support the following mechanism of cytotoxicity for TRPV1-agonists in lung (and possibly other) cells. First, activation of intracellular TRPV1 leads to a decrease in ER calcium content, an accumulation of unfolded/partially folded proteins in the ER lumen, and an overall decrease in protein processing efficiency. As a result, EIF2αK3 is activated resulting in the phosphorylation of EIF2α and an increase in the expression of ATF4, GADD153, and other ER stress-related genes. Ultimately, increased transcription and expression of GADD153 causes cell death.
The translational facets of the results presented in this study are 2-fold. First, the near uniform response elicited by structurally diverse TRPV1 agonists in all four lung cell types suggests that this mechanism of toxicity is applicable to many other TRPV1 agonists. Specifically, environmental TRPV1 agonists that promote lung inflammation and injury (e.g., particle pollutants) and endogenous TRPV1 agonists (e.g., leukotrienes, H2S) that are produced during inflammation or infection may also cause lung cell death and tissue damage via the EIF2αK3-dependent ER stress pathway. As such, future clinical research targeting TRPV1 and/or the EIF2αK3-dependent ER stress pathways may prove beneficial in the treatment and/or prevention diverse respiratory maladies. Second, our results indicate that the effects of a TRPV1 ligand on a cell will depend upon both the relative subcellular distribution of TRPV1 and the relative permeability of the ligand. Hence, it must be stressed that the subcellular location of TRPV1 should be established and multiple TRPV1 agonists and antagonists (preferably not capsazepine) should be used in future research studies evaluating the role of TRPV1 in specific biological outcomes. Although we have not specifically tested whether cell-impermeable agonists of TRPV1 (e.g., pH or environmental particle pollutants) exhibit different mechanisms of cytotoxicity, evidence supports this hypothesis. Specifically, inhibition of the cell surface TRPV1 in lung cells has no effect on cytotoxicity by TRPV1 agonists, despite inhibition of proinflammatory cytokine synthesis (Reilly et al., 2005), and sensory neurons, which primarily express TRPV1 on the cell surface, are protected against cytotoxicity by inhibiting cellular influx of calcium (Wood et al., 1988).
Overall, these results provide novel insight into mechanisms by which diverse exogenous and endogenous TRPV1 agonists affect lung cell physiology. These findings provide fundamental knowledge that will facilitate future basic science and clinical research on TRPV1 in an array of physiological and pharmacological models, including models of acute lung injury and inflammatory lung injury.
Acknowledgments
We thank Dr. Manivannan Ethirajan (Department of Medicinal Chemistry, University of Utah, Salt Lake City, UT) for assistance with n-benzylnonanamide synthesis and Dr. David Ron (Skirball Institute, New York University Medical Center, New York City, NY) for helpful suggestions.
Footnotes
-
This study was supported by National Heart, Lung, and Blood Institute Grant HL069813.
-
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
-
doi:10.1124/jpet.107.119412.
-
ABBREVIATIONS: TRPV1, transient receptor potential vanilloid 1 (a.k.a. VR1 or the capsaicin receptor); ER, endoplasmic reticulum; LJO-328, N-(4-tert-butylbenzyl)-N′-(1-[3-fluoro-4-(methylsulfonylamino)phenyl]ethyl)thiourea; 5-iodo-RTX, 5-iodo-resiniferatoxin; ATF, activating transcription factor; EIF2α, eukaryotic translation initiation factor 2, subunit 1 (α, 35kDa); GADD153, growth arrest- and DNA damage-inducible transcript 3 (a.k.a. DDIT3 and CHOP); GADD45α, growth arrest and DNA-damage-inducible, α (a.k.a. DDIT1); BiP/GRP78, glucose-regulated protein, 70 kDa (a.k.a. HSPA5); CCND1, cyclin D1; CCNG2, cyclin G2; EIF2α-P, phosphorylated eukaryotic translation initiation factor 2, subunit 1; DTT, dithiothreitol; PCR, polymerase chain reaction; NHBE, normal human bronchial epithelial; HEK, human embryonic kidney; RT, reverse transcription; ANOVA, analysis of variance; GFP, green fluorescent protein; EtBr, ethidium bromide; EGFP, enhanced green fluorescent protein; SC0030, N-(4-tert-butylbentyl)-N′-[3-fluoro-4-(methylsulfonylamino)benzyl]thiourea.
-
↵ The online version of this article (available at http://jpet.aspetjournals.org) contains supplemental material.
-
↵1 Current affiliation: Jacobs Dugway Team, Life Sciences Testing Facility, Dugway, Utah.
- Received January 3, 2007.
- Accepted February 28, 2007.
- The American Society for Pharmacology and Experimental Therapeutics