erlotinib promotes endoplasmic reticulum stress-mediated injury in the intestinal epithelium.docx

Erlotinib promotes endoplasmic reticulum stress-mediated injuryin the intestinal epitheliumErlotinib promotes endoplasmic reticulum stress-mediated injury in theintestinal cpithcliumLu Fana, 1, Lingna Hua, 1, Baofang Yanga, Xianying Fanga, Zhc Gaoa, Wanshuai Lia, Yang Suna, Yan Shena, Xuefeng Wua, Yongqian Shub, Yanhong Gub,?, Xudong Wua,?, Qiang Xua, ?aState Key Laboratory of Pharmaceutical Biotechnology, School of Life Sciences, Nanjing University, 22 Hankou Road, Nanjing 210093, ChinabDepartment of Clinical Oncology, The First Affi1iated Hospital of Nanjing Medical University, 140 Hanzhong Road, Nanjing 210029, Chinaa bstractarticleinf oArticle history: Received 6 November 2021 Revised 10 April 2021Accepted 15 April 2021Available online 24 April 2021Keywords:ErlotinibDiarrhcaERstressE-cadherinApoptosisErlotinib, apopulardrugfortreatingnon-smalice11lungcancer(NSCLC ),causesdiarrheainapproximately55%ofpatients receiving this drug. In thepresentstudy, we found that erlotinibinduced barrierdysfunction inratsmall intestine epithelial cells (IEC-6) by increasing epithelial permeabi1ity and down-regulating E-cadherin. ThemRNA levels of various pro-inflammatory cytokines (Il-6, 11-25 and Il-17f) were increased after erlotinib treat-ment in IEC-6 cells. Erlotinib concentration- and time-dependently induced apoptosis and endoplasmic reticu-lum (ER) stress in both IEC_6 and human colon epithelial cells (CCD 841 CoN). Intestinal epithelial injury wasalso observed in male C57BL/6J mice administrated with erlotinib. Knockdown of C/EBP homologous protein(CHOP) with small interference RNA partially reversed erlotinib-induced apoptosis, production of IL-6 anddown-regulation of E-cadherin in cultured intestinal epithelial cells. Tn conclusion, erlotinib caused ER stress-mediated injury in the intestinal epithelium, contributing to its side effects of diarrhea in patients. ? 2021 Elsevier Inc. All rights reserved. IntroductionErlotinib is an orally administered, low molecular weight, quinazolinc-bascd agent (chemical structure of erlotinibwasshowninSupplernentaryFig. 1A)whichselectivelyandreversiblyinhibitsepider-malgrowth factor receptor (EGER) kinaseactivity(Hidalgo etal., 2001). Erlotinib is currently approved for the treatment of advanced ormetastatic resistant non-smal 1 cell lung cancer (NSCLC) and for use incombination with gemcitabine to treat advanced, unresectable, or met-astatic pancreatic cancer (Bareschinoet al., 2007). However, sideeffectsincludingskin rash, diarrhea, and pulmonary toxicity often limit thcap-plication of erlotinib, particularly for long-term treatment. Diarrhea isthe second most common side effect of erlotinib. The National Cancerinstitute of Canada Clinical Trials Group Study BR. 21 documented that55% of patients treated with erlotinib had diarrhea, compared to 19%of patients on placebo. In the Iressa Pan-Asia Study (TPASS), 47% of pa-tients treated with erlotinib developed diarrhea (Nguyen and Neal, 2021). It has been reported that severe diarrhea caused by erlotinibcanresul tinfluidandelectrolytelosses, whichma5 1 eadtodehydration, electrolyte imbalancesandrenalinsufficiency(Melosky,2021).Epider-malgrowthfactor(EGF)hasbeenshowntoinhibitCa2+-dependentCl?transport in T84 human colonic epithelial cells. EGFR inhibitors maycause diarrhea by blocking this inhibitory loop and causing chloride se-cretion (Uribe et al., 1996). Still, the mechanisms by which erlotinib de-stroys the intestinal epithelium remain unclear.The intestinal epithelium forms a protective barrier between the lu-minal contents and the external environment.Breakdown of the gutbarrier is associated with increased intestinal permeabi1ity and gut in-flammation (Krzyzaniak et al., 2021). Epithelial cells interconnect byan array of intercellular adhesion complexes including adherens junc-tions (AJs), tight junctions (TJs) and desmosomes (Pcrez-Morenoet al., 2003). E-cadhcrin is a major component of AJs, impaired expres-sion of which in the small intestine and colon has been linked to dis-turbed intestinal homeostasis and barrier function (Schneider et al. , 2021). The endoplasmic reticulum (ER) is a principal site for protein syn-thesis and folding, calcium storage and signaling, and it is highly sensi-tive to alterations in calcium homeostasis and environmentalperturbations (Lin et al., 2021). Physiologic or drug-induced disruptionof protein folding causes misfolded, aggregated, or unassembled pro-teins to accumulate in the ER lumen, triggering a response called ERstress (Kaufman, 1999). The ability of cells to respond to perturbationsinERfunction, or 'ERstress'virus protease inhibitor disrupts epithelial barrier integrity through ac-tivation of ER stress and unfolded protein response in intestinal epithe-lial cells (Wu et al., 2021). However, no reports concern ER stress anderlotinib-induced diarrhea. In the present study, we found for the first time that erlotinib in-duced injury and dysfunction in the intestinal epithelium through ERstress. Materials and methodsMaterials. Antibodies against ATF4 (#11815), phospho-eIF2 a (#3398)and eIF2 a (#5324) were purchased from Cell Signaling Technology (Bever 1 y, MA). /nnexin V - fluorescein isothiocyanate (FTTC)/propidiumiodide (PT) kit was purchased from BD Biosciences (San Jose, CA). ELISAkitforinterleukin-6(IL-6)waspurchasedfromR&DSystems(Minneap-olis,MN).Antibodies against E-cadherin (sc-8426), Z0-1 (sc-10804), ATF6 (sc-22799), XBP-1 (sc-7160), CHOP (sc-575) and P -actin (sc-47778) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). FITC - dextran, dimethylsulfoxide(DMSO)andinsulinwereobtain-ed from Sigma Chemical Co. (St. Louis, MO). Protease inhibitor cocktai Iwas obtained from Roche Technology (Basel, Switzerland). Erlotinibwas obtained from Genentech Inc. (South San Francisco, CA). Cell culture and erlotinib treatment. Rat small intestine epithelial cells(IEC-6) were maintained in DMEM (Life Technologies, Carlsbad, CA)supplemented with 100 mU/ml of insulin and 10% fetal bovine serum(Life Technologies, Carlsbad, CA) under a humidified 5% (v/v) C02at-mosphere at 37 0 C. CCD 841 CoN human colon epithelial cells were in-cubated in MEM (Life Technologies, Carlsbad, CA) containing 10% fetalbovine serum (Life Technologies, Carlsbad, CA) in a 5% C02atmosphereat 37 ° C. Erlotinib was dissolved in DMSO to a concentration of 20 n)M(stock solution) and stored at ?20 ° C. Permeability measurement. Cells were seeded into 12-well transwellplates (0.4-u m pore diameter, Corning-Costar Corp., Cambridge, MA)and grown to 100% confluency. The cells were incubated with 0. 1%DMSO and 5, 10 or 20 u M erlotinib for 24 h, respectively. FITC - dextran(MW, 4. 4kDa)wasdissolvedinculturemediumandusedatafinalcon-centration of 2.2 mg/ml in the apical cell compartment. After 4 h of in-cubation, 100 u 1 aliquots were obtained fromthemediumofbasal chamber. Fluorescencewasmeasuredwi thaf1uorescencespectrometer(exc i tat i on490 nm;emission520nm). Cytotoxicityassay.Thecytotoxicactivityofer1otinibwasexami nedby3-(4, 5-dimethyl thiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT)assay and Lactate Dehydrogenase (LDH) assay. Cells were treated witherlotinib for indicated time periods in 96-well plates. For MTT assay, MTT solution (4 mg/ml in phosphate buffered saline PBS) was added(20 u 1/well) into each well and incubated for 4 h at 37 ° C. Then the su-pernatant was discarded and the purple formazan crystals were dis-solved in 200 u 1 of DMSO for 5 min. The plates were read on anautomated microplate spectrophotometer (Sunrise, Tecan, Austria) at570 nm. For LDH assay,thesupernatantwasharvested and thereleasedLDH was detected. Total LDH was detected after cells were lysed. Cyto-toxicity was evaluated by the released LDH/total cell LDH. Isolation and culture of primary rat intestinal cells. Primary rat intestinalcells were isolated from male rats and cultured as described (Campbell, 2021).Reverse transcription and real-time quantitative PCR (Q-PCR). RNA sam-ples were treated by DNase and subjected to real-time quantitativePCR. First-strand cDNAs were generated by reverse transcription usingoligo(dT) (Hidalgo ct al., 2001). Quantitative PCR was performed withthc ABI Prism 7000 sequence detection system (Applied Biosystems, Foster City, CA) using SYBR Green I dye (Biotiuni, Inc., Hayward, CA), and threshold cycle numbers were obtained using ABI Prism 7000 SDSsoftware version 1.0. The cDNA amplification was performed for 30 cy-clesusingtheprimerslistedinTablelwiththefollowingsettings:94° Cfor 5 min, 94 ° C for 30 s, 58 ° C for 30 s, 72 ° C for 30 s, and 72 ° C forlO min. Flow cytometry. Annexin V - FITC/PI double staining assay was per-formed to detect apoptosis. Cel Is were harvested and resuspended inAnnexin-V binding buffer. The suspension was incubated with 2. 5 u lof Annexin V - FITC and 2 n 1 of PI for 10 min at room temperature inthe dark, followed by cytometric analysis (EPICS XL, Beckman Coulter, Fullerton, CA) within 30 min of staining. Samples were analyzed usinga I?ACSCalibur flow cytometer. ELISA. Cytokine (IL-6) was measured with ELISA kit (R6000B,M6000B)from R&D systems (Minneapolis, MN) according to themanufacturer * sprotocol. Western blot. Proteins were extracted in lysis buffer (30 nunol/1 Tris, pH 7. 5, 150 mmol/I sodium chloride, 1 mmol/I phenyl methyl sulfonyl fluoride, 1 mmol/I sodiumorthovanadate,1%NonidetP-40,10%glycer-ol, andproteaseinhibitorcocktai1). TheproteinswerethenseparatedbySDS-PAGEandelectr ophoretically transferrcdontopolyvinylidenefluo-ride membranes. The membranes were incubated with primary anti-bodies overnight at 4 C, and then incubated with a horserad i shperox i dase-coupledsecondary antibody for 1 h at roomtemperature. Detectionwasperformedusinga LumiGLOchemiluminescentsubstratesystem(KPL,GuiIdford, UK). Smal1 interfering RNA (siRNA) transfection. C/EBP homologous protein(CHOP) siRNA sequence used in CCD 841 CON cells was: 5, -UUCAUCUGAAGACAGGACCUCUUGC-3' . Luciferase siRNA was used as describedbefore (Wang et al. , 2021). CHOP siRNA and nonspecific siRNA used inIEC-6 cells were purchased from Guangzhou RiboBio Co., Ltd. (Guang-zhou, China). Cells were cultured to 60% confluency in a 6-well plateand transfected with luciferase siRNA or CHOP siRNA using Lipofecta-mine 2000 (Life Technologies, Carlsbad, CA) for 30 h. Then cells wereTable1Primersused for amplification. GenePrimer sequenceCdhl(rat)5, -ATCTAAAGCTTCACAAGCTGGA-3' 5' -TGATCTGTGACTGTGACCACTA-3" 5' -CCATCTTTGGACCGATTG CTG-3' 5' -TAATGCCCGAGCTCCGATG-3z 5' -GACCTGCCTTGGTGTCTGTGC-3z 5' -AGCAACCACACCAGCTACAA T-3Z 5' -CCGGAGAGGAGACTTCACAG-3r 5' -CAGAATTGCCATTGCACAAC-3z 5' -CCCGCTTTGACTGTGTACT-3 '5' -TGAGTACCAGCGGGATCTTCTC-3' 5' -CTTACCCAGATGCTGTCCC-3' 5' -GATTCAAGTCCCTGTCCAAC-3" 5' -GGCATTGCTCTCAATGACAA-3f 5' -AGGGCCTCTCTCTTGCTCTC-3r 5' -CTGCAAGAGACTTCCATCCAG-3z 5 '-AGTGGTATAGACAGGTCTGTTGG-3z 5' -ACAGGGACTTGAATCGGGTC-3; 5' -TGGTAAAGTGGGACGGAGTTG-3; 5' -TGCTACTGTTGATGTTGGGAC-37 5' -CAGAAATGCCCTGGTTTTGGT-3; 5' -AATGGATTTGGACGCATTGGT-3; 5' -TTTGCACTGGTACGTGTTGAT-3z 5' -CGAGAGCTACACGTTCACGG-3z 5' -GGGTGTCGAGGGAAAAATAGG-3; 5 '-ACTCACCTCTTCAGAACGAATTG-3z 5' -CCATCTTTGGAAGGTTCAGGTTG-3z 5' -TGTGGGCATCAATGGATTTGG- 3' 5' -ACACCATGT/TTCCGGGTCAAT-3z Tjpl (rat)Itjbl(rat) T1-6(rat)Socs3 (rat) 11-25(rat)Gapdh (rat)11-6 (mouse)11-25 (mouse)Il-17f (mouse)Gapdh (mouse)CDH1 (human)IL-6 (human)Gapdh (human)46L. Fan et al. / Toxicology and Applied Pharmacology 278 (2021) 45 - 52treated with either 20 u M erlotinib or 0. 1% DMSO as control for another24 h.Immunofluorescence microscopy. Cells on glass cover slips were washedwith PBS andfixed with4%paraformaldehydeonicefor 30 min, follow-ed by permeabilization with 0. 1% Triton X-100 for 20 min. Cells wereblocked with PBS containing 10% serum for 1 h at room temperatureand then incubated with primary antibody (E-cadherin specific anti-body) overnight at 4 ° C. Slides were washed three times with PBS, andincubated with anti-rabbit IgG conjugated with FITC (Invitrogen, Carls-bad, CA) for 1 h at room temperature. Nuclei were counterstained with2 u g/ml 4, 6-diamidino-2-phenylindole (DAPI) for 1 min. The fluores-cent signals were detected with a mercury 1 amp (Olympus, Tokyo, Japan) and analyzed by Image Pro Plus 6. 0 (Media Cybernetics, Inc., Be-thesda, MD). Mice. Male C57BL/6J mice, 6-8 weeks old, werepurchasedfrom theEx-perimental Animal Center of Jiangsu Province (Nanjing, China). Animalwelfarc and experimental procedures were carried out strictly in accor-dance with the Guide for the Care and Use of Laboratory Animals (USNII1) and inaccordance with ethical regulationsof ouruniversity. Allef-forts were made to minimize animals' sufferingand to reduce the num-ber of animals used. Body weight and clinical signs of disease wererecorded beginning on day 1. The mice were divided random!y into 4groups (n = 8 per group). Mice were orally administrated with vehicle(distilled water) or various concentrations of erlotinib for 8 weeks. Then the mice were anesthetized with sodium amobarbital (15 mg/kgbody weight, intraperitoneal injection). The intestines were removedand part were fixed in 10% formalin,the others were snap-frozen in liq-uidnitrogen andhomogenized in 10 u1/mgbuffer(200 mMNaCl, 4 mMEDTA, 10 mM Tris, 10% (w/v) glycine, protease inhibitor cocktail, pH 7. 4) for the measurement of cytokines with ELISA. TUNEL staining. TUNEL staining was performed using aMEBSTAIN Apo-ptosis Kit Direct (Bio-med Tech, Beijing, China) according to themanufacturer, s protocol. This method detected nucleosome-sized DNAfragments by tailing their 3' -OH ends with digoxigenin nucleotidesusing terminal deoxynucleotidyl transferase (TdT). TUNEL-pos i t i vece11s were counted using fluorescence microscopy. Immunohistochemistry Immunohistochemical staining was performedwitha two-step EnVisionmethod. Alltissueswerefixed inlO% bufferedformalin, embedded in paraffin, and cut into 4-u m sections. Afterdeparaffinization, heat-induced antigen retrieval techniques wereused. Endogenous peroxidase activity was then blocked with 0. 5%H202. After washing in Tris-buffered saline (TBS), the sections werestained with primary antibodies at 4 ° C overnight. Then, the scctionswcrcwashedinTBSagain, andincubatedwiththcsecondaryantibodicsEnvision? (Dako, Glostrup, Denmark). The cell nuclei were restainedwith hematoxylin after diaminobenzidine (DAB) showed color. Histological analysis. Formalin-fixed, paraffin-embedded intestinal tis-sues were sectioned at 4- u m thickness, and sections were stained withhematoxylin and eosin (H&E) to assess: 1) leukocyte infiltration andvascular congestion and 2) erosion and anabrosis of epidermal cells. Thehistologicalscoreswereassessedfromlto4. Finaldataweretheav-erage scores of each animal in the same group. High scores mean morcinflamination. Statistical analysis. Data were expressed as the mean ± standard devia-tion (SD) of three independent experiments. The Stude