Operatively induced chronic reflux in rats: A suitable model for studying esophageal carcinogenesis? Caroline Gronnier, Emilie Bruyère, Guillaume Piessen, Nicolas Briez, Jérôme Bot, David Buob, Emmanuelle Leteurtre, Isabelle van Seuningen, Christophe Mariette

To cite this version:

Caroline Gronnier, Emilie Bruyère, Guillaume Piessen, Nicolas Briez, Jérôme Bot, et al.. Operatively induced chronic reflux in rats: A suitable model for studying esophageal carcinogenesis?. Surgery, Elsevier, 2013, ￿10.1016/j.surg.2013.05.029￿. ￿hal-02905869￿

HAL Id: hal-02905869 https://hal.archives-ouvertes.fr/hal-02905869 Submitted on 30 Sep 2020

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. P-annotatePDF-v11

INSTRUCTIONS ON THE ANNOTATION OF PDF FILES

To view, print and annotate your article you will need Adobe Reader version 9 (or higher). This program is freely available for a whole series of platforms that include PC, Mac, and UNIX and can be downloaded from http://get.adobe.com/reader/. The exact system requirements are given at the Adobe site: http://www.adobe.com/products/reader/tech-specs.html. Note: if you opt to annotate the file with software other than Adobe Reader then please also highlight the appropriate place in the PDF file. PDF ANNOTATIONS Adobe Reader version 9 Adobe Reader version X and XI When you open the PDF file using Adobe Reader, the To make annotations in the PDF file, open the PDF file using Commenting tool bar should be displayed automatically; if Adobe Reader XI, click on ‘Comment’. not, click on ‘Tools’, select ‘Comment & Markup’, then click If this option is not available in your Adobe Reader menus on ‘Show Comment & Markup tool bar’ (or ‘Show then it is possible that your Adobe Acrobat version is lower Commenting bar’ on the Mac). If these options are not than XI or the PDF has not been prepared properly. available in your Adobe Reader menus then it is possible that your Adobe Acrobat version is lower than 9 or the PDF has not been prepared properly.

This opens a task pane and, below that, a list of all (Mac) Comments in the text. These comments initially show all PDF ANNOTATIONS (Adobe Reader version 9) the changes made by our copyeditor to your file.

The default for the Commenting tool bar is set to ‘off’ in version 9. To change this setting select ‘Edit | Preferences’, then ‘Documents’ (at left under ‘Categories’), then select the option ‘Never’ for ‘PDF/A View Mode’.

(Changing the default setting, Adobe version 9)

HOW TO... Action Adobe Reader version 9 Adobe Reader version X and XI Insert text Click the ‘Text Edits’ button on the Click the ‘Insert Text’ icon on the Comment Commenting tool bar. Click to set the cursor tool bar. Click to set the cursor location in the text location in the text and simply start typing. The and simply start typing. The text will appear in a text will appear in a commenting box. You may commenting box. You may also cut-and-paste text also cut-and-paste text from another file into the from another file into the commenting box. Close commenting box. Close the box by clicking on ‘x’ in the top right-hand corner. the box by clicking on ‘_’ in the top right-hand corner. Replace text Click the ‘Text Edits’ button on the Click the ‘Replace (Ins)’ icon on the Commenting tool bar. To highlight the text to be Comment tool bar. To highlight the text to be replaced, click and drag the cursor over the text. replaced, click and drag the cursor over the text. Then simply type in the replacement text. The Then simply type in the replacement text. The replacement text will appear in a commenting box. replacement text will appear in a commenting box. You may also cut-and-paste text from another file You may also cut-and-paste text from another file into this box. To replace formatted text (an into this box. To replace formatted text (an equation for example) please Attach a file (see equation for example) please Attach a file (see below). below). Remove text Click the ‘Text Edits’ button on the Click the ‘Strikethrough (Del)’ icon on the Commenting tool bar. Click and drag over the text Comment tool bar. Click and drag over the text to to be deleted. Then press the delete button on be deleted. Then press the delete button on your your keyboard. The text to be deleted will then be keyboard. The text to be deleted will then be struck through. struck through. Highlight text/ make a Click on the ‘Highlight’ button on the Click on the ‘Highlight Text’ icon on the comment Commenting tool bar. Click and drag over the text. Comment tool bar. Click and drag over the text. To To make a comment, double click on the make a comment, double click on the highlighted highlighted text and simply start typing. text and simply start typing. Attach a file Click on the ‘Attach a File’ button on the Click on the ‘Attach File’ icon on the Commenting tool bar. Click on the figure, table or Comment tool bar. Click on the figure, table or formatted text to be replaced. A window will formatted text to be replaced. A window will automatically open allowing you to attach the file. automatically open allowing you to attach the file. To make a comment, go to ‘General’ in the A graphic will appear indicating the insertion of a ‘Properties’ window, and then ‘Description’. A file. graphic will appear in the PDF file indicating the insertion of a file. Leave a note/ comment Click on the ‘Note Tool’ button on Click on the ‘Add Sticky Note’ icon on the the Commenting tool bar. Click to set the location Comment tool bar. Click to set the location of the of the note on the document and simply start note on the document and simply start typing. Do typing. Do not use this feature to make text edits. not use this feature to make text edits. HOW TO... Action Adobe Reader version 9 Adobe Reader version X and XI Review To review your changes, click on the ‘Show’ Your changes will appear automatically in a list below the Comment tool bar. Navigate by button on the Commenting tool clicking on a correction in the list. Alternatively, bar. Choose ‘Show Comments List’. Navigate by double click on any mark-up to open the clicking on a correction in the list. Alternatively, commenting box. double click on any mark-up to open the commenting box. Undo/delete To undo any changes made, use the right click To undo any changes made, use the right click change button on your mouse (for PCs, Ctrl-Click for the button on your mouse (for PCs, Ctrl-Click for the Mac). Alternatively click on ‘Edit’ in the main Mac). Alternatively click on ‘Edit’ in the main

Adobe menu and then ‘Undo’. You can also Adobe menu and then ‘Undo’. You can also delete edits using the right click (Ctrl-click on delete edits using the right click (Ctrl-click on the Mac) and selecting ‘Delete’. the Mac) and selecting ‘Delete’.

SEND YOUR ANNOTATED PDF FILE BACK TO ELSEVIER Save the annotations to your file and return as instructed by Elsevier. Before returning, please ensure you have answered any questions raised on the Query Form and that you have inserted all corrections: later inclusion of any subsequent corrections cannot be guaranteed. FURTHER POINTS x Any (grey) halftones (photographs, micrographs, etc.) are best viewed on screen, for which they are optimized, and your local printer may not be able to output the greys correctly. x If the PDF files contain colour images, and if you do have a local colour printer available, then it will be likely that you will not be able to correctly reproduce the colours on it, as local variations can occur. x If you print the PDF file attached, and notice some ‘non-standard’ output, please check if the problem is also present on screen. If the correct printer driver for your printer is not installed on your PC, the printed output will be distorted.

ARTICLE IN PRESS

1 63 2 64 3 65 4 Operatively induced chronic reflux 66 5 67 6 in rats: A suitable model for studying 68 7 69 8 70 9 esophageal carcinogenesis? 71 10 72 11 Q2 Caroline Gronnier, MD,a,b,c Emilie Bruyere , PhD,a,b,c Guillaume Piessen, MD, PhD,a,b,c 73 12 Nicolas Briez, MD,a,b,c Jerome Bot, MD,a,b,c David Buob, MD,b,d 74 13 Emmanuelle Leteurtre, MD, PhD,a,b,d Isabelle Van Seuningen, PhD,a,b and 75 14 Christophe Mariette, MD, PhD,a,b,c Lille, France 76 15 77 16 78 17 Background. The mechanisms of esophageal reflux leading to esophageal adenocarcinoma (EA) remain 79 18 poorly understood. This study appraises critically an operatively induced chronic reflux rat model. 80 19 Methods. We randomized 108 Sprague-Dawley rats into 2 experimental groups; one was performing 81 20 esophagoduodenal (ED) anastomosis with or without gastrectomy to induce duodeno-esophageal reflux 82 21 (DER group; n = 63), and the other involved duodeno-gastro-esophageal reflux (DGER group; n = 45). 83 22 Control groups included (i) Roux-en-Y esophagojejunal anastomosis, (ii) laparotomy alone, (iii) subtotal 84 23 gastrectomy to induce duodenogastric reflux (DGR group), and (iv) the same procedure as in the DGER 85 24 group plus proton pump inhibition (PPI group). The underwent histologic and molecular 86 25 analyses. 87 26 Results. The prevalence of Barrett’s esophagus (BE), , and EA in the experimental groups was 88 27 41%, 7%, and 11%, respectively. Histologic and molecular analyses in groups DER, DGER, and 89 28 DGR suggested that BE occurred through de novo intestinal and proximal migration of 90 29 duodenal cells. No distant metastases were identified. The molecular characteristics of both BE and EA 91 30 were similar to humans. BE was more common, and dysplasia and EA less frequent in the DER group 92 31 when compared with the DGER group (44% vs 24% [ P = .038] and 7% vs 25% [ P = .012], 93 32 respectively). Compared with the DGER group, carcinogenic sequence occurred less frequently in the 94 33 PPI-treated group (P = .019). 95 34 Conclusion. Despite pathophysiologic differences with humans, the rat model of esophagoduodenostomy 96 35 97 reproduces accurately histologic and molecular lesions in the carcinogenetic sequence of BE and allowed 36 98 us to identify novel, tumor-associated proteins that may be potential biomarkers and new therapeutic 37 j j j 99 38 targets in EA. (Surgery 2013; : - .) 100 39 101 40 From Inserm, UMR837,a Jean-Pierre Aubert Research Center, Team 5 ‘‘Mucins, epithelial differentiation 102 41 and carcinogenesis,’’ the Universite Lille Nord de France,b the Department of Digestive and Oncological 103 42 Surgery,c University Hospital Claude Huriez, Centre Hospitalier Regional et Universitaire de Lille, and 104 43 the Centre de Biologie-Pathologie,d Department of , Centre Hospitalier Regional et Universitaire 105 44 de Lille, Lille, France 106 45 107 46 108 47 DESPITE ADVANCES IN MULTIMODAL THERAPY, the prog- influence of duodeno-gastric-esophageal reflux 109 48 nosis for invasive esophageal adenocarcinoma (DGER).2 The cytotoxic mechanisms of reflux 110 49 (EA) remains poor.1 EA is thought to develop in leading to initiation and progression of BE remain 111 50 Barrett’s esophagus (BE), following the low- to poorly understood. The hypothesis has been made 112 51 113 52 high-grade dysplasia sequence and under the that pH changes in DGER caused by PPI may 114 53 induce cytoxicity related to the bile acid refluxed 115 54 and may play a role in the increase in incidence 116 Accepted for publication May 16, 2013. 55 of EA incidence in Western countries.3 117 56 Reprint requests: Christophe Mariette, MD, PhD, Professor of Because of the long delay in the progression of 118 57 Surgery, Department of Digestive and Oncological Surgery, 119 University Hospital Claude Huriez, Centre Hospitalier regional the carcinogenic sequence in humans, an efficient 58 120 et Universitaire, Place de Verdun F-59037, Lille, France. E-mail: animal model of BE and EA better our under- 59 121 [email protected]. standing of the mechanisms involved. Various 60 4-10 122 61 0039-6060/$ - see front matter surgical models of BE/EA have been reported. 123 62 Ó 2013 Mosby, Inc. All rights reserved. None showed superiority. Discrepant results have 124 http://dx.doi.org/10.1016/j.surg.2013.05.029 been reported regarding the time of development

SURGERY 1 FLA 5.2.0 DTD SEC CODE: OC n YMSY3448_proof 12 June 2013 8:13 am ce ARTICLE IN PRESS 2 Gronnier et al Surgery j 2013

125 of each step in the sequence of carcinogenesis.11 transected 4 cm distal to the duodenojejunal 185 126 Whether the metaplasia in these models is de junction and sectioned proximal to the ligature. 186 127 187 128 novo, originating from the esophageal glands, is The distal segment was anastomosed to the distal 188 129 derived by entrapment of duodenal mucosa to esophagus via a Roux-en-Y (RY) esophagojejunos- 189 130 the esophageal wall during the anastomosis, or tomy and an entero-entero anastomosis $15 cm 190 131 occurs by creeping substitution (proximal migra- distal to the esophagojejunostomy using a 1-layer, 191 132 tion of duodenal cells through the anastomosis) running 7.0 polypropylene suture. 192 133 remains unclear.8,12,13 Doubt has been cast on Laparotomy group. We performed laparotomy 193 134 whether these models reflect the development of (LAP) only (external control). 194 135 8 195 136 BE and EA. Moreover, many investigators have Duodenogastric reflux group. We performed a 196 137 also reported adenosquamous or pure squamous partial gastrectomy respecting the gastroesopha- 197 138 cell carcinoma, questioning the model’s ability to geal junction, with a proximal gastroduodenal 198 139 elucidate mechanisms of tumorigenesis.9,13 Finally, anastomosis. This duodenogastric reflux (DGR) 199 140 the respective roles of duodenal and gastroduo- group, in which the esophagus was adjacent to but 200 141 denal reflux in the model remain unsolved.5,7 not part of the anastomosis, acted as a control for 201 142 These questions raise the validity of the animal the duodenoesophageal reflux (DER) group to 202 143 203 model as a reliable tool in understanding the evaluate if the origin of esophageal intestinal 144 8,12,13 204 145 human carcinogenic sequence. metaplasia was de novo, originating from the 205 146 The aims of our study were to evaluate at esophagus, or originating from the duodenum 206 147 histologic and molecular levels the (i) value of through entrapment of duodenal mucosa into 207 148 the rat model, and (ii) respective roles of the esophageal wall during the ED or creeping of 208 149 the duodenal and gastroduodenal reflux, in the duodenal cells across the anastomosis as a healing 209 150 BE/EA carcinogenetic sequence. process. The stomach was anastomosed to the 210 151 211 152 duodenum via a gastroduodenostomy with 10 212 Q1 MATERIAL AND METHODS 153 interrupted sutures of 7.0 polypropylene. 213 154 Six-week-old, male Sprague-Dawley rats (Charles Proton pump inhibitor group. Rats that underwent 214 155 River, L’Arbresle, France), were housed under the same procedure as in the DGER group 215 156 standard laboratory conditions. Operations were received Esomeprazole (Astra-Zeneca, Dunkerque, 216 157 performed under general anesthesia, using France), at a dose of 5 mmol/kg via daily gavage 217 158 xylazine (12 mg/kg) and ketamine (80 mg/kg). from day 1 after the operative procedure. This 218 159 219 160 Through a midline laparotomy, operative proce- group was used to evaluate the suppression of the 220 161 dures were performed to induce esophageal acidic in the gastric reflux in operatively induced 221 162 reflux. The experimental protocol received DGER rats to evaluate its role on reflux-induced 222 163 approval from the Veterinary Ethics Committee lesions. 223 164 (CREEA, authorization no.59-350177). Tissue specimens. All animals were weighed 224 165 Experimental groups (Fig 1). DGER group.We weekly. Rats that became ill or lost >15% of their 225 [F1-4/C] performed an esophagoduodenal (ED) anasto- weight were killed. Rats that died before the 226 227 mosis without gastrectomy to study the impact scheduled date were autopsied. 228 166 of the combination of gastric and duodenal Animals were killed under general anesthesia to 167 229 reflux on esophageal mucosa. The esophagus was recover tissue samples at 20 weeks for BE lesions, 230 168 11 169 transected proximal to the gastroesophageal junc- 30 weeks for dysplasia, and 50 weeks for EA. As a 231 170 tion, and anastomosed to the duodenum with consequence of the intermediate results obtained 232 171 mucosa-to-mucosa apposition with 8 interrupted for the last 55 rats operated, the killing of these 233 172 sutures of 7.0 polypropylene. animals was delayed to 60, 70, and 80 weeks, to 234 235 173 Duodeno-esophageal reflux group. We performed an test the hypothesis that a longer duration of 174 236 ED anastomosis with gastrectomy to evaluate the exposure to reflux could increase the incidence 237 175 1 176 impact of the duodenal reflux on esophageal of each step of the carcinogenic sequence. After 238 177 mucosa. A total gastrectomy was performed after injection of the euthanizing agent T61, the 239 178 ligation of left gastric and short gastric vessels, and thoracic and abdominal cavities were inspected 240 179 the duodenum just distal to the pylorus. The distal with an en bloc removal of the esophagus. After 241 180 esophagus was anastomosed as in DGER group. longitudinal opening, the esophagus was sectioned 242 181 243 Control groups. Roux-en-Y esophagojejunal anasto- into slices of 2- to 3-mm thickness. In rats with 244 182 mosis group. We performed an esophagojejunal (EJ) visible esophageal tumor, the liver and the lungs 183 245 184 anastomosis with jejunal limb without reflux were removed to perform a histologic analysis 246 (internal control). The proximal jejunum was looking for micrometastases.

FLA 5.2.0 DTD SEC CODE: OC n YMSY3448_proof 12 June 2013 8:13 am ce ARTICLE IN PRESS Surgery Gronnier et al 3 Volume j, Number j

247 Histologic analysis. Histologic analyses were pH-metry. A measurement of intragastric pH 309 248 performed by 2 expert pathologists in a double- was performed using a pH meter orion 2 star 310 249 311 250 blinded manner (EL, DB). Tissues were fixed in (Thermoscientific, Brebieres, France), in 10 312 251 10% (w/v) buffered formaldehyde, embedded in randomly selected rats of the DGER, LAP, and 313 252 paraffin, cut at 4 mm thickness, and applied on PPI groups after aspiration through a gastrotomy 314 253 SuperFrost slides (Menzel-Glaser, Braunschweig, to assess the persistence of acid secretion (DGER 315 254 Germany). Slides were then stained with hemat- group) or its suppression (PPI group). 316 255 oxylin and eosin, Safran, and Astra blue. Lesions RNA extraction and purification. Total RNA was 317 256 were defined as (i) BE, specialized columnar isolated and purified from tissues using the 318 257 € 319 258 epithelium with cellular characteristics mimicking NucleoSpin RNA L kit (Macherey-Nagel, Duren, 320 259 human BE (intestinal-like cells and Alcian blue- Germany) as described in the manufacturer’s 321 260 positive goblet cells); (ii) dysplasia, lesions with protocol. The RNA quantity was determined by 322 261 an increase of the nuclear-cytoplasmic ratio, measuring the optical density at 260 nm with 323 262 nuclear atypia, partial loss of cell polarity, and a NanoDrop 1000 spectrophotometer (Thermo 324 263 increase in mitotic figures, and (iii) EA, predom- Scientific, Pittsburgh, Penn). 325 264 inant component or pure EA with invasion Quantitative reverse transcriptase-polymerase 326 265 327 through the basement membrane to surrounding chain reaction. Total RNA (0.5 mg) was used to 266 2 328 267 tissues, tumor stroma reaction, increase in nuclear prepare cDNA using the RT First Strand 329 268 atypia, and mitotic figures. Kit (SABioscience, Valencia, Calif). Quantitative 330 269 pH-metry. A measurement of intragastric pH reverse transcriptase-polymerase chain reaction 331 270 was performed using a pH meter orion 2 star (qRT-PCR) was performed on cDNA (1 mL), 332 271 (Thermoscientific, Brebieres, France), in 10 using specific primers for TFF2 (PPR48691A), 333 272 randomly selected rats of the DGER, LAP, and TFF3 (PPR52672A), Muc1 (PPR51463A), Muc4 334 273 335 274 PPI groups after aspiration through a gastrotomy (PPR47950A), Nfkb1 (PPR42746A), and Pik3cg 336 275 to assess the persistence of acid secretion (DGER (PPR56854A) synthesized by SABioscience and 337 2 276 group) or its suppression (PPI group). the RT Fast SYBR Green Master Mixes/fluores- 338 277 RNA extraction and purification. Total RNA was cein qPCR Master Mixes. For each of those 339 278 isolated and purified from tissues using the primers, the standard curve was made to deter- 340 279 NucleoSpin RNA L kit (Macherey-Nagel, Duren,€ mine the best concentration of cDNA to use. 341 280 Germany) as described in the manufacturer’s Amplification was made in triplicate for each sam- 342 281 343 282 protocol. The RNA quantity was determined by ple, and the internal control used was the lactate 344 283 measuring the optical density at 260 nm with dehydrogenase A (Ldha) gene (PPR56603B). 345 284 a NanoDrop 1000 spectrophotometer (Thermo Amplification was made using the CFX96 thermo- 346 285 Scientific, Pittsburgh, Penn). cycler (BioRad, Marnes-la-Coquette, France): 1 347 286 Quantitative reverse transcriptase-polymerase cycle of 10 minutes at 958C, followed by 45 cycles 348 287 chain reaction. Total RNA (0.5 mg) was used to of 15 seconds at 958C and 1 minute at 608C. 349 288 prepare cDNA using the RT2 First Strand RT-PCR. Total RNA (1 mg) was used to prepare 350 289 m 351 290 Kit (SABioscience, Valencia, Calif). Quantitative cDNA using oligod(T) (1 L) and recombinant 352 291 reverse transcriptase-polymerase chain reaction Retro-Transcriptase Moloney Murine Leukemia 353 292 (qRT-PCR) was performed on cDNA (1 mL), Virus (1 mL; Promega, Charbonnieres, France). 354 293 using specific primers for TFF2 (PPR48691A), PCR was performed on cDNA (5 mL), using specific 355 294 TFF3 (PPR52672A), Muc1 (PPR51463A), Muc4 pairs of primers: b-actin (sense: ATATCGCTGC 356 295 (PPR47950A), Nfkb1 (PPR42746A), and Pik3cg GCTCGTCGTCGACAA; anti-sense: AACACAGCC 357 296 (PPR56854A) synthesized by SABioscience and TGGATGGCTACGTACAT), cyclin D1 (sense: TGA 358 297 2 359 298 the RT Fast SYBR Green Master Mixes/fluorescein CTGCCGAGAAGTTGTG; anti-sense: GAGGGTG 360 299 qPCR Master Mixes. For each of those primers, the GGTTGGAAATG), ErbB1 (sense: AGTGGTC 361 300 standard curve was made to determine the best CTTGCAAACTTGG; anti-sense: TTAACTCAAGCT 362 301 concentration of cDNA to use. Amplification was GCCTCGCC), cyclo-oxygenase-2 (sense: AGTAT 363 302 made in triplicate for each sample, and the internal CAGAACCGCATTGCC; anti-sense: TAAGGTTT 364 303 control used was the lactate dehydrogenase CAGGGAGAAGCG), PI3K (sense: GAAGCCATT 365 304 A(Ldha) gene (PPR56603B). Amplification was GAGAAGAAAGGA; anti-sense: GAGGTGTTCAG 366 305 k 367 306 made using the CFX96 thermocycler (BioRad, TATTATCAGAGC), NF- B (sense: GAAGAAGCGA 368 307 Marnes-la-Coquette, France): 1 cycle of 10 minutes GACCTGGAG; anti-sense: TCCGGAACACAATG 369 308 at 958C, followed by 45 cycles of 15 seconds at GCCAC), Smad4 (sense: CATTCCTGTGGCTTC 370 958C and 1 minute at 608C. CACAA; anti-sense: GACTGATGGCTGGAGCTATT)

FLA 5.2.0 DTD SEC CODE: OC n YMSY3448_proof 12 June 2013 8:13 am ce ARTICLE IN PRESS 4 Gronnier et al Surgery j 2013

371 433 372 434 373 435 374 436 375 437 376 438 377 439 378 440 379 441 380 442 381 443 382 444 383 445 384 446 385 447 386 448 387 449 388 450 389 451 390 452 391 453 392 454 393 455 394 456 395 457 396 458 397 459 398 460 399 461 400 462 401 463 402 464 403 465 404 466 405 467 406 468 407 469 408 470 409 471 410 472 411 473 412 474 413 475 414 476 415 477 416 478 417 479 418 480 419 481 420 482 421 483 422 484 423 485 424 486 425 487 426 488 427 489 428 490 429 491 Fig 1. Macroscopic aspect of specimens classified according the operation group and histologic observation of the 430 492 431 esophagus after operation using hematoxylin and eosin, Safran, and Astra blue stainings. The control groups with 493 432 no reflux---(A) LAP and (B) RY groups---showed thin and noninflamed esophagus. (C) In the DGER group, the 494 esophagi were dilated and inflamed, sometimes with the presence of a macroscopic tumor always localized in the distal

FLA 5.2.0 DTD SEC CODE: OC n YMSY3448_proof 12 June 2013 8:13 am ce ARTICLE IN PRESS Surgery Gronnier et al 5 Volume j, Number j

495 and vascular endothelial growth factor (sense: ACG European Collection of Cell Culture and cultured 557 496 AAA GCG CAA GAA ATC CC; anti-sense: TTA ACT as described in Mariette et al.14 558 497 559 498 CAA GCT GCC TCG CC; MWG-Biotech, Ebersberg, Small interfering RNA assays. Parental OE33 560 499 Germany). For each gene, analyses were made from cells were transfected with small interfering RNA 561 500 $3 different rats. Rat b-actin was used as the (siRNA) from Dharmacon (Epsom, United 562 501 internal control. PCR products (20 mL) were Kingdom) following the protocol described by 563 502 separated on a 1.5% (w/v) agarose gel containing Piessen et al.15 564 503 ethidium bromide run in 1X Tris Borate EDTA Proliferation. Cell proliferation was analyzed by 565 504 buffer. The gene of interest/b-actin ratio counting cells on a Malassez counting chamber. 566 505 567 506 was calculated after scanning DNA bands OE33 cells transfected with siRNA were plated on 568 507 with GelAnalyst-GelSmart software (Claravision, 24-well plates then counted during 3 days after 569 508 Verrieres Le Buisson, France). transfection. 570 509 Microarray analysis. Complementary RNA was Migration. Cell migration was studied by a 571 510 synthesized from total RNA and purified using the test. Cells were plated on 96-well 572 511 TrueLabeling-AMP 2.0 Kit and the ArrayGrade plates until confluence; a ‘‘wound’’ was then 573 512 cRNA Cleanup Kit respectively (SABioscience), performed using the 96-pin WoundMaker (Essen 574 513 575 together with biotin-UTP (Roche, Neuilly-sur- Bioscience, Ann Arbor, Mich). Plates were placed 514 8 576 515 Seine, France) according to the manufacturer’s in an incubator with 5% of CO2 at 37 C into the 577 516 instructions. Complementary RNA (2 mg) was Incucyte device (Essen Bioscience). Every other 578 517 incubated with Oligo GEArrays (SABioscience)--- hour, a picture of the wound was taken and the 579 518 Rat Cancer PathwayFinder, Rat Tumor Metastasis, percentage of wound closure was then every 580 519 Rat Cell Cycle, and Rat PI3K-AKT signaling pathway 12 hours, from 0 to 48 hours. 581 520 microarrays---all designed for profiling the expres- Statistical analysis. The primary objective was 582 521 $ 583 522 sion of 113 genes. Expression level of each gene was the occurrence of 1 element of the carcinoge- 584 523 detected by chemoluminescence using ECF netic sequence (either BE, dysplasia, or EA). Based 585 524 substrate on a Storm 860 scanner (GE Healthcare on preliminary experiments and the literature, we 586 525 Buckinghamshire, UK; IFR114/IMPRT, U. Lille 2). hypothesized that with a power of 80% and an a 587 526 Signals were analyzed finally with the GEArray value of 5%, $45 animals per experimental group 588 527 Expression Analysis Suite (SABioscience). (DGER and DER groups) were required to observe 589 528 . Protein expression was an incidence of 40% in 1 group versus 75% in the 590 529 591 530 studied using either manual or automatic immuno- other group (bilateral comparison of 2 binomial 592 531 histochemistry (IHC). Automatic IHC with an proportions). Because intermediate analysis 593 532 automated 8 immunostainer (ES, Ventana Medical showed a mortality rate of #60%, the number of 594 533 System, Illkirch Graffenstaden, France). Antibodies rats required per group was 120. Statistical analysis 595 534 were purchased from Cell Signaling (Danvers, Mass; was performed using SPSS version 15.0 software 596 535 cyclin D1[2926 at 1/50e]), Santa Cruz Biotechnology (SPSS, Chicago, Ill). Data are shown as prevalence 597 536 [Santa Cruz, Calif; MCM6 [sc-9845 at 1/100e], Mmp3 and mean values (standard deviation). Continuous 598 537 599 538 (sc-31074 at 1/100e), Mmp7 (sc-8832 at 1/50e), data were compared by means of the Mann– 600 539 Mmp10 (sc-6842 at 1/200e), and Timp1 (sc-6834 at Whitney U test and ordinal data by the 601 540 1/100e)]. Positive controls were included by staining Chi-square test or Fisher exact test as appropriate. 602 541 normal rat tissues known to express the protein All statistical tests were 2-sided. 603 542 of interest and negative controls were run with 604 543 13 D-PBS instead of primary antibodies. RESULTS 605 544 Cell culture. The OE33 EA esophageal adeno- Global evaluation. Among 285 rats operated, 606 545 607 546 carcinomatous cells were purchased from the 120 were included in the DGER group and 120 in 608 547 609 548 610 549 = 611 550 portion of esophagus exposed to the reflux. The histologic observations showed (D) esophagitis (original magnifica- 612 3 551 tion, 200). (E) Esophageal adenocarcinoma with the presence of mucus lakes stained in blue (black arrow; original 613 3 3 552 magnification, 200) compared with (D, insert) normal esophagus (original magnification, 200). (F) Operative 614 553 procedures and compositions of reflux for each group. DGER group, duodeno-gastro-esophageal reflux group allows 615 554 the induction of a mixed gastric and duodenal reflux in the distal portion of the esophagus; DER group, duodeno- 616 555 esophageal reflux only in the distal portion of the esophagus; RY group, internal control; LAP group, external 617 556 control: DGR group, control for the DER group and PPI groups, which allows inhibition of acidic component of the 618 gastric reflux.

FLA 5.2.0 DTD SEC CODE: OC n YMSY3448_proof 12 June 2013 8:13 am ce ARTICLE IN PRESS 6 Gronnier et al Surgery j 2013

619 Table I. Number of rats analyzed: Histologic analysis of resected esophagus of rats, analyzed according to 681 620 the treatment group 682 621 683 622 Experimental groups Control groups 684 623 685 Group DGER group DER group DGR group PPI group RY group LAP group 624 686 625 Reflux composition Gastric Duodenal Duodenal Gastric nonacid No reflux No reflux 687 626 duodenal duodenal 688 627 No. of rats 63 45 6 10 7 10 689 628 Histologic results* 690 629 Esophagitis 63 (100%) 45 (100%) 6 (100%) 10 (100%) 0 (0%) 0 (0%) 691 630 BE 23 (36%) 21 (47%) 1 (16%) 1 (10%) 0 (0%) 0 (0%) 692 631 Dysplasia 7 (11%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 693 632 694 EA 8 (13%) 4 (9%) 1 (16%) 0 (0%) 0 (0%) 0 (0%) 633 695 634 *Some rats may exhibit concomitant steps of the carcinogenetic sequence. 696 635 BE, Barrett’s esophagus; DER, duodeno-esophageal reflux; DGER, duodeno-gastro-esophageal reflux; DGR, duodeno-gastric reflux; EA, esophageal 697 636 adenocarcinoma; LAP, laparotomy; PPI, proton pump inhibitor; RY, Roux-en-Y. 698 637 699 638 700 639 the DER group. The overall mortality rate was 53% 32 animals exhibited visible tumor growth localized 701 640 and decreased during the study period from 68% on the external surface of the esophagus (Fig 1, C). 702 641 to 37%; 55 rats died from respiratory failure owing Histologic analysis in the experimental groups 703 642 to massive aspiration (none in the no-reflux showed at least 1 step of the carcinogenetic 704 643 groups). Other causes of death were malnutrition sequence in 50% of rats (Table I). Prevalence of ½T1 705 644 (loss of weight > 15%; n = 62), peritonitis (n = 9), BE, dysplasia, and cancer were 41% (n = 44), 8% 706 645 707 646 intraoperative complications (n = 20), and (n = 9), and 11% (n = 12), respectively. Histologic 708 647 unknown causes (n = 5). Number of survivors lesions were always observed in the distal esoph- 709 648 analyzed was 63 in the DGER group, 45 in DER agus, which was the region with maximal exposure 710 649 the group, 6 in the DGR group, 10 in the PPI to reflux. Among 32 macroscopic tumors, 711 650 group, 7 in the RY group, and 3 in the LAP group. only 12 were histologically proven carcinomas 712 651 Nutritional status. The body weight loss was (Fig 1, E), the 20 remaining cases showing only a 713 652 10 ± 4% and was maximal after 3 weeks. After an granulomatous inflammatory reaction. The posi- 714 653 715 654 initial postoperative decrease, the curve of animal tive predictive value of a macroscopic examination 716 655 body weight joined progressively the one of the in predicting a cancerous lesion was only 38%. 717 656 control group (data not shown), suggesting that This finding suggests that exposure to chronic 718 657 nutritional status did not interfere with the results reflux promotes the development of elements of 719 658 observed. the carcinogenic sequence in half of the animals.16 720 659 pH-metry. The mean values of the intragastric pH Histologic tumor characteristics. Among the proven 721 660 were 2.7 ± 0.4, 2.8 ± 0.4, and 5.1 ± 0.4 in the DGER, carcinomas, 8 were well-differentiated, mucinous 722 661 723 662 LAP, and PPI groups, respectively. This underlines adenocarcinomas (malignant infiltrating glands 724 663 the expected persistence of acid gastric secretion associated with lakes of extracellular mucus; 725 664 in the DGER group and a decrease in acid Fig 1, E). Five rats developed adenosquamous 726 665 secretion after PPI administration (P = .019), carcinomas combining components of squamous 727 666 which validates the models used to generate both carcinoma and adenocarcinoma, but in all carci- 728 667 duodenal and gastroduodenal refluxes. nomas, adenocarcinoma was the major compo- 729 668 Reflux lesions. In the LAP and RY groups nent. Some of these cases exhibited vascular 730 669 731 670 (Fig 1, F), all rats had normal esophagus on macro- invasion, confirming their malignant nature. 732 671 scopic (Fig 1, A and B) and microscopic examina- Neither lymph node nor distant metastases were 733 672 tion (Fig 1, D). In the DGER, DER, DGR, and PPI identified macroscopically during necroscopies. 734 673 groups (Fig 1, F), the esophagus was dilated and In rats with visible esophageal tumor growth, a 735 674 markedly shortened in all rats, and microscopic systematic, histologic analysis of the liver and the 736 675 analysis showed esophagitis (Fig 1, D). These re- lungs did not exhibit any micrometastatic dissemi- 737 676 sults attest the efficacy of this chronic reflux model nation. Even if features of local esophageal 738 677 739 678 to induce esophagitis. malignancy were present, these results challenge 740 679 Steps of the carcinogenetic sequence. In the LAP and the ability of these tumors to metastasize. 741 680 RY groups (Fig 1, A and B), no lesions of the carcino- Characterization of the intestinal lesions. The IHC 742 genetic sequence were found. In the reflux groups, analysis showed that intestine-specific markers

FLA 5.2.0 DTD SEC CODE: OC n YMSY3448_proof 12 June 2013 8:13 am ce ARTICLE IN PRESS Surgery Gronnier et al 7 Volume j, Number j

743 803 744 804 745 805 746 806 747 807 748 808 749 809 750 810 751 811 752 812 753 813 754 814 755 815 756 816 757 817 758 818 759 819 760 820 761 821 762 822 763 823 764 824 765 825 766 826 767 827 768 828 769 829 770 830 771 Fig 2. Histologic analyses showing the presence of intestinal tissue in the esophagus exposed to reflux. (A) Immuno- 831 772 histochemistry (IHC) using an anti-MUC2 antibody. (B) Lysozyme staining shows presence of villi (arrows). 832 773 (C) IHC carried out with an anti-MUC4 antibody shows presence of Brunner’s glands (arrow). (D) Typical aspects of 833 774 entrapment characterized by a short segment of Barrett’s esophagus surrounded by squamous mucosa. IHC staining 834 775 was performed using an anti-MUC4 antibody. (A, original magnification, 3200; B, original magnification, 3400; 835 776 C, D, original magnification, 3100). 836 777 837 778 838 such as Villin, Cdx-2 (data not shown), and Muc2 entrapment or creeping. To validate these, a histo- 839 [F2-4/C] (Fig 2, A), were expressed in BE, whereas they were logic reassessment of rats with BE was performed. 840 779 never expressed in normal esophagus (data not A short segment of BE surrounded by squamous 841 780 shown). Expression of cytokeratins 4 (CK4) and mucosa corresponding to a typical aspect of 842 781 14 (CK14), known to participate in the cytoskel- entrapment was found in 1 rat (Fig 2, D). Second, 843 782 eton formation and usually found in normal another control group was constructed by per- 844 783 845 esophageal epithelium, in the cytoplasm of the forming a subtotal gastrectomy and gastroduo- 846 784 suprabasal cells for CK4 and of the basal cells for denal anastomosis (DGR group) in which the 785 847 786 CK14, was also assessed. They were expressed in esophagus did not form part of the anastomosis. 848 787 the histologically normal rat esophagus without As presented in Table I, esophagitis was present 849 788 expression in BE lesions (data not shown). These in all these rats, confirming the presence of reflux. 850 789 results confirm the differentiation of normal One rat developed a mucinous adenocarcinoma 851 790 esophageal cells submitted to reflux into an arising directly from the anastomotic site and 852 791 853 intestinal-type phenotype mimicking human BE. located between the remaining stomach and lower 854 792 Origin of the intestinal mucosa. To investigate the esophagus with surrounding BE lesions. In the 793 855 794 origin of BE lesions observed in the reflux groups, absence of histologic lesions in the esophagus, 856 795 we went expanded our histologic analysis. Villi and this group did not allow us to eliminate the 857 796 Brunner€ glands were present in the esophageal hypothesis of creeping. Expression pattern of 858 797 wall, 2 histologic characteristics of normal genes encoding trefoil factors (TFF) in BE had 859 798 duodenum, usually not present in human BE been shown previously to differ from that in 860 799 17 861 (Fig 2, B and C). These results suggest that BE normal jejunal mucosa in an EJ rat model, 862 800 lesions in the ED rat model may not be true BE suggesting that BE lesions represented de novo 801 863 802 lesions originating de novo from the esophagus, BE and not creeping; therefore, we studied 864 but may be caused by mechanically induced the expression of TFF-2 and TFF-3 mRNA

FLA 5.2.0 DTD SEC CODE: OC n YMSY3448_proof 12 June 2013 8:13 am ce ARTICLE IN PRESS 8 Gronnier et al Surgery j 2013

865 927 866 928 867 929 868 930 869 931 870 932 871 933 872 934 873 935 874 936 875 937 876 938 877 939 878 940 879 941 880 942 881 943 882 944 883 945 884 946 885 947 886 948 887 949 888 950 889 Fig 3. Gene expression in control squamous esophagus (Roux-en-Y [RY] and [LAP] groups) and in esophageal 951 890 adenocarcinoma (EA). (A) Upregulation of Muc1 and Muc4 in EA compared with RY by quantitative reverse 952 891 transcriptase-polymerase chain reaction (qRT-PCR). Expression levels are normalized to Ldha and compared 953 892 with the RY group. (B) Neo-expression of Smad4 and Cox-2 in EA compared with RY and LAP groups by RT-PCR. 954 893 (C) Upregulation of cyclin D1, ErbB1, NF-kB, and Vegf in EA compared with control groups using RT-PCR. The internal 955 894 control used was bactin.(D) Activation of Pi3k and NF-kb associated with the tumor progression in esophagus subjected Q3 956 895 to reflux by qRT-PCR. Expression levels are normalized to Ldha and compared with the RY control. (E) Proliferation 957 896 958 and (F) migration assays on OE33 cells transfected with PI3K or NF-kB small interfering RNA (siRNA). Controls are 897 959 made by a transfection with a non-targeting (NT) siRNA or with the transfection agent alone (mock). *P # .05; 898 960 **P # .005. 899 961 900 962 901 963 902 by qRT-PCR in BE and normal duodenum. The well as at the protein level (data not shown). We 964 903 expression of TFF-2 was more prominent in BE evaluated also expression of transcription factors 965 904 compared with duodenum (1,030-fold), whereas associated with esophageal tumorigenesis21 or 966 905 expression of TFF3 was more uniform in the implicated in MUC1 or MUC4 regulation by 967 906 tissues and was not discriminant (1.6 fold; data bile acids,15,20 (Fig 3, B, C and D). A neo- 968 907 969 not shown). Altogether, these data suggest that, expression of Smad4 (transforming growth factor- 908 b 970 909 in this ED model, BE lesions may occur through pathway) and cyclo-oxygenase-2 (inflammation) 971 910 various concomitant mechanisms, including de was observed in EA compared with RY controls 972 911 novo intestinal metaplasia and proximal migration (Fig 3, B). An increase of the cell-cycle factor cyclin 973 912 of duodenum. D1 and the epidermal growth factor receptor 974 913 Characterization of the esophageal adenocarcinoma- ErbB1 (cell proliferation), nuclear factor (NF)-kB 975 914 tous lesions. We sought to evaluate the similarities of (inflammation), and vascular endothelial growth 976 915 977 916 tumors found in humans and in this model. The factor (angiogenesis) was also found (Fig 3, C). 978 917 expression of genes and proteins associated By qRT-PCR, we confirmed the activation of 2, cen- 979 18,19 918 with EA tumorigenesis, and/or implicated tral signaling pathways in EA tumor progression: 980 919 in mediation of the effects of bile acid in EA20 NF-kB and PI3K (Fig 3, D). Hnf-1a and Hnf-4a 981 920 was compared in normal esophageal mucosa (no- transcription factors were not expressed in the 982 921 reflux control groups) and in EA. Having esophagus of the RY and LAP groups, whereas 983 922 previously shown that membrane-bound mucins nuclear expression was found in EA lesions. 984 923 985 MUC1 and MUC4 are overexpressed in human Nuclear expression of P63 protein was found in 924 15,20 986 925 EA and that they are regulated by bile acids, the esophageal epithelium of RY and LAP groups 987 926 we showed that genes encoding Muc1 and Muc4 and was lost in EA (data not shown). Nuclear 988 ½F3 mucins were overexpressed in EA (Fig 3, A), as expression of Ki-67 (proliferation) was observed

FLA 5.2.0 DTD SEC CODE: OC n YMSY3448_proof 12 June 2013 8:13 am ce ARTICLE IN PRESS Surgery Gronnier et al 9 Volume j, Number j

989 1049 990 1050 991 1051 992 1052 993 1053 994 1054 995 1055 996 1056 997 1057 998 1058 999 1059 1000 1060 1001 1061 1002 1062 1003 1063 1004 1064 1005 1065 1006 1066 1007 1067 1008 1068 1009 1069 1010 1070 1011 1071 1012 1072 1013 1073 1014 1074 1015 1075 1016 1076 1017 1077 1018 1078 1019 1079 1020 1080 1021 Fig 4. Expression of genes associated with tumorigenesis in esophagus subjected to reflux. (A) Expression of genes 1081 1022 using microarrays. Average of 5 different rats showing esophageal adenocarcinoma lesions. (B) Expression of proteins 1082 1023 (immunohistochemistry) for which genes were up-regulated in microarrays studies. Normal squamous esophagus 1083 1024 (original magnification, 3100); esophageal adenocarcinoma (original magnification, 3100; insert, original magnifica- 1084 1025 tion, 3200). 1085 1026 1086 1027 1087 1028 in the basal layer of the normal esophageal (Fig 5). Decreased proliferation was observed ½F5 1088 1029 epithelium in the RY and LAP groups and in EA for the 3 KD cell lines as soon as day 2 for 1089 1030 lesions (data not shown). After studying genes MCM6-KD cells (D2, P = .0309; D3, P = .0019) 1090 1031 1091 known to participate in esophageal carcinogenesis, and as of day 3 for CHEK2-KD and TSG101-KD 1032 < 1092 we then searched for new factors using a transcrip- cells (D3, P .0056 for both). No difference in 1093 [F4-4/C] tomic approach (Fig 4). As above, genes known to cell migration was found in CHEK2-, TSG101-, or 1094 1033 participate in proliferation (cyclin D1) and metas- MCM6-KD cells compared with controls (Fig 5). 1095 1034 tasis formation or invasion (Mmp3, -7, -9, -10, -13, These data indicate that tumor progression is 1096 1035 the metalloproteinase-inhibitor Timp1) were upre- induced by chronic reflux with activation of key 1097 1036 gulated in EA. Conversely, tumor suppressor genes signaling pathways and expression of tumor- 1098 1037 1099 (Tp53, Tsg101, Pten) and genes participating associated proteins in a similar way in the DGER 110 0 1038 in (CHEK2, Caspase 9) were down- and DER groups. These results demonstrate that 1039 1101 1040 regulated (Fig 4, A). Alteration of gene expression BE and EA present similar molecular characteris- 1102 1041 was confirmed at the protein level by IHC for some tics in both experimental groups compared with 1103 1042 of the markers (Fig 4, B). We then focused on 3 humans. 1104 1043 genes---CHEK2, TSG101, and MCM6---not yet Role of exposure time in esophageal carcino- 1105 1044 known to participate in esophageal tumor genesis. The prevalence of histologic lesions did 1106 1045 1107 progression, but known to play roles in epithelial not differ between the different times of exposure 1108 1046 tumorigenesis.21-23 Using small interfering RNA to reflux (esophagitis, P = .612; BE, P = .195; 1047 1109 1048 approach, mRNA expression of CHECK, TSG101, dysplasia and EA, P = .741). No differences were 1110 and MCM6 was knocked down (KD) in OE33 cells found when comparing time of reflux exposure

FLA 5.2.0 DTD SEC CODE: OC n YMSY3448_proof 12 June 2013 8:13 am ce ARTICLE IN PRESS 10 Gronnier et al Surgery j 2013

1111 1173 1112 1174 1113 1175 1114 1176 1115 1177 1116 1178 1117 1179 1118 1180 1119 1181 1120 1182 1121 1183 1122 1184 1123 1185 1124 1186 1125 1187 1126 1188 1127 1189 1128 1190 1129 1191 1130 1192 1131 1193 1132 1194 1133 1195 1134 1196 1135 1197 1136 1198 1137 Fig 5. Proliferation and migration assays on OE33 cells transfected with MCM6, CHEK2, or TSG101 siRNA. Controls 1199 1138 are made by a transfection with a non-targeting (NT) siRNA or with the transfection agent alone (mock). *P # .05; EQ1 1200 1139 **P # .005. 1201 1140 1202 1141 1203 1142 to promote $1 step of the carcinogenetic PPI (n = 10) and in the DGER groups (n = 19), 1204 1143 sequence between the DGER and DER groups with esophagitis present in all rats. At least 1 step 1205 1144 (median times, 50 [range, 22–68] vs 50 [22–83] of the carcinogenetic sequence was observed 1206 1145 1207 ½ 1146 T2 weeks; P = .789; Table II). Subgroup analysis in more frequently in the DGER group (58%) 1208 1147 the DGER and the DER groups showed similar compared with the PPI group (10%; P = .019). 1209 1148 results (P = .741). These findings suggest that These results suggest that the acidic reflux may 1210 1149 esophageal carcinogenesis is not linked to favor progression of the carcinogenic sequence 1211 1150 exposure time to reflux and that delaying toward EA. 1212 1151 necroscopy to >50 weeks is useless. 1213 1152 Role of the reflux composition on esophageal DISCUSSION 1214 1153 1215 carcinogenesis. The proportion of rats who Our rat model using either ED or EJ anasto- 1154 $ 1216 1155 developed 1 step of the carcinogenic sequence mosis is considered the most promising model for 1217 1156 did not differ between the 2 groups (49% vs 51%; reproducing histologic lesions mimicking the 1218 1157 P = .841). BE lesions were more frequent in the human EA carcinogenic sequence,3 but few groups 1219 1158 DER group compared with the DGER group have characterized the lesions that develop at both 1220 1159 (44 vs 24%; P = .038). Prevalence of dysplasia the histologic and molecular levels.8,13,24,25 This 1221 1160 and/or EA lesions was less in the DER group lack of in-depth characterization of this frequently 1222 1161 1223 1162 (7%, vs 25%; P = .012). These results suggest that used model led us to determine whether the rat 1224 1163 the duodenal component of the reflux may play model reliably and accurately reproduces the 1225 8,12 1164 an initiating role in the carcinogenic sequence, carcinogenic sequence as occurs in humans. 1226 1165 whereas duodenogastric juice may lead to We demonstrated herein that the ED rat model 1227 1166 progression of the carcinogenic sequence to EA. of reflux allows the development of metaplasia-like 1228 1167 Impact of acid component of gastric reflux. To lesions (BE) and its transformation into dysplasia 1229 1168 determine the role of acid component in these and EA with similar histologic and molecular 1230 1169 1231 1170 findings, the control PPI group (n = 10) was characteristics described in humans. We showed 1232 1171 constructed. Macroscopic and histologic results that the duodenal component of the reflux plays a 1233 1172 are presented in Table I. At 30 weeks, the esopha- crucial role in initiating the carcinogenic 1234 geal macroscopic appearance was similar in the sequence, whereas duodenogastric juice may lead

FLA 5.2.0 DTD SEC CODE: OC n YMSY3448_proof 12 June 2013 8:13 am ce ARTICLE IN PRESS Surgery Gronnier et al 11 Volume j, Number j

1235 Table II. Distribution of histologic lesions according to the duration of reflux exposure in rats 1297 1236 1298 1237 Time of exposure to reflux (in weeks) 1299 1238 21–30 30–50 50–83 1300 1239 1301 1240 Reflux group DGER group DER group DGER group DER group DGER group DER group 1302 1241 1303 Histologic results 1242 1304 1243 Esophagitis only 6 (50%) 2 (25%) 11 (55%) 9 (64%) 15 (48%) 11 (48%) 1305 1244 BE only 3 (25%) 5 (62%) 4 (20%) 4 (28%) 8 (26%) 11 (48%) 1306 1245 Dysplasia or EA 3 (25%) 1 (13%) 5 (25%) 1 (8%) 8 (26%) 1 (4%) 1307 1246 Total 12 (100%) 8 (100%) 20 (100%) 14 (100%) 31 (100%) 23 (100%) 1308 1247 BE, Barrett’s esophagus; EA, esophageal adenocarcinoma. 1309 1248 1310 1249 1311 1250 to its progression to EA. Many discrepancies, The incidence of EA in the present study was 1312 1251 however, have been highlighted in the literature low (13%) with some adenosquamous tumors 1313 1252 when comparing human observations with occurring despite avoiding the use of nitrosa- 1314 1253 1,5-10 1315 1254 previous animal models, including the devel- mines, a co-carcinogen shown to be responsible 1316 1255 opment of benign tumor and adenosquamous for high rates of a squamous cell tumoral compo- 1317 27 1256 tumors observed rarely in humans, a low rate of nent. A review has discussed the reproducibility 1318 1257 EA tumorigenesis despite deliberate esophageal of the rat model28 with regard to (i) the variability 1319 1258 exposure to extensive gastroesophageal reflux, of the operative procedures, with a greater rate of 1320 1259 and no apparent impact of duration of the EA after EJ than after ED anastomosis,3,13 possibly 1321 1260 exposure on the incidence of BE/EA. The absence because of a lesser volume of refluxate in the latter 1322 1261 29 1323 1262 of any tumoral dissemination to regional lymph type of anastomosis, (ii) the various durations of 1324 1263 nodes or distant organs also questions whether reflux, (iii) the use of a variety of rat species, and 1325 1264 the induced malignancies behave in an aggressive (iv) the suboptimal pathologic analysis of induced 1326 1265 manner. It has also been questioned whether lesions.8 1327 1266 creeping of the duodenal mucosa or entrapment Analysis of the transcriptome allowed us to 1328 1267 of the jejunal mucosa may be the mechanisms to identify genes that had been shown to previously 1329 1268 be involved in the development of BE rather be important in human EA, corroborating the 1330 1269 8,9 15,30,31 1331 1270 than a de novo process. As reported by others, validity of this model. Moreover, this analysis 1332 1271 the high postoperative mortality despite the allowed us to identify novel, potential biomarkers 1333 1272 well-known expertise of the surgical team raised involved in esophageal carcinogenesis. First, the 1334 1273 ethical concerns and questioned the reprodu- S100a4 gene, whose protein is known to have many 1335 1274 cibility of such modeling. roles in cancer cell properties, was upregulated.32 1336 1275 In humans, BE metaplasia is thought to result We reported previously that altered properties of 1337 1276 from the abnormal differentiation of esophageal OE33 esophageal cancer cells deficient for MUC4 1338 1277 1339 stem cells or from the transdifferentiation of 1 mucin correlated with that of the calcium-binding 1278 2 1340 1279 mature esophageal cell type into another type of protein s100A4, suggesting that regulation of 1341 26 1280 mature cell. The gene expression profile of TFF S100A4 by the membrane-bound mucin MUC4 1342 1281 in a rodent model with EJ anastomosis suggests would lead to decreased proliferation and migration 1343 1282 that BE occurs de novo and is not owing to of esophageal cancer cells. 1344 1283 proximal migration of jejunal mucosa.17 Brunner MCM6 and CHEK2, both involved in cell 1345 1284 glands in the DRG group and observation of short proliferation, were up regulated in our transcrip- 1346 1285 1347 1286 segments of BE surrounded by squamous mucosa tomic study, and activity on cell proliferation was 1348 1287 suggest that BE-like lesions in the rat model could confirmed in vitro. MCM6 has been shown 1349 1288 arise also from duodenal mucosa through entrap- previously to be involved in meningioma, and its 1350 1289 ment or creeping in accordance with previous expression was correlated with clinical outcomes 1351 1290 studies.8,9 Only a formal lineage tracing would and grade of the tumor.21 As for CHEK2, a gene 1352 1291 allow to solve this question definitely. Contrary to mutation has been associated with an increased 1353 1292 our results, some investigators7,9,11 have suggested risk of breast cancer.22 Altogether, our results and 1354 1293 1355 1294 that the time of exposure to reflux may determine those in the literature suggest strongly that 1356 1295 the type and incidence of lesions that develop; MCM6 and CHEK2 may play an important roles in 1357 1296 however, the number of rats operated on in the properties of esophageal adenocarcinomatous 1358 each of those studies was relatively small.6,8,11 cells as well as in esophageal carcinogenesis.

FLA 5.2.0 DTD SEC CODE: OC n YMSY3448_proof 12 June 2013 8:13 am ce ARTICLE IN PRESS 12 Gronnier et al Surgery j 2013

1359 Tsg101 was shown previously to decrease to identify potential tumor-associated proteins that 1421 1360 proliferation in breast cancer.23 Accordingly, our may prove to be new biomarkers and new 1422 1361 1423 1362 results in esophageal cancer cells deficient for therapeutic targets in EA. This rat ED anastomosis 1424 1363 Tsg101 showed decreased proliferation, suggesting model offers a unique opportunity to study the 1425 1364 that Tsg101 is a new and interesting marker for respective roles of the duodenal and gastroduo- 1426 1365 further study in that type of cancer. denal contents in the BE carcinogenic sequence. 1427 1366 The respective contributions of the duodenal 1428 1367 and gastroduodenal reflux on the development of C.G. and E.B. equally contributed to the work. I.V.S. 1429 1368 BE and EA remain controversial.5,7,9,12 We wanted and C.M. equally contributed to the work. Supported by 1430 1369 1431 to evaluate the impact of the gastric and the La Fondation pour l’Avenir and La Fondation pour la 1370 1432 duodenal contents on the carcinogenic sequence. Recherche Medicale (Equipe FRM 2009). Emilie 1371 Bruyere is the recipient of a PhD Fellowship from Le 1433 1372 The pH of gastric juice has been shown previously 1434 Centre Hospitalier Regional Universitaire de Lille and 1373 to be neutralized when an EJ anastomosis is 1435 Region Nord-Pas de Calais. Isabelle Van Seuningen is 1374 33 1436 performed without gastrectomy ; therefore, we the recipient of a ‘‘Contrat Hospitalier de Recherche 1375 6,28 1437 chose to perform an ED anastomosis. We Translationnelle’’/CHRT 2010, AVIESAN. 1376 showed that absence of gastric juice (DER group) 1438 1377 1439 € 1378 was associated with significantly earlier lesions of The authors thank Dr Jean-Noel Freund (Inserm U782, 1440 1379 BE, whereas combined gastric and duodenal reflux Strasbourg) for scientific discussions. We thank also Dr 1441 1380 (DGER group) was associated with more advanced William B Robb for critical reading of this manuscript. 1442 1381 lesions (dysplasia and EA). A control group of rats This work was supported by grants from CHRU de Lille, 1443 1382 exposed to mixed reflux and receiving pharmaco- Region Nord-Pas-de-Calais, La Fondation de l’Avenir and 1444 1383 logic acid suppression (PPI group) confirmed this La Fondation pour la Recherche Medicale (EQUIPE FRM 1445 1384 mechanism with a dramatic decrease in the 2009) and AVIESAN (CHRT 2010, IVS). 1446 1385 1447 1386 number of early lesions. Confirming data reported 1448 1387 previously using a cellular model of EA, these 1449 REFERENCES 1388 results suggest a distinct impact of the duodenal 1450 1389 reflux according to pH environment,34 which 1. Mariette C, Piessen G, Briez N, Gronnier C, Triboulet JP. 1451 Oesophagogastric junction adenocarcinoma: which thera- 1390 1452 suggests that PPIs may prevent initiation of the peutic approach? Lancet Oncol 2011;12:296-305. 1391 1453 carcinogenic sequence with a protective effect. 2. Hamilton SR, Smith RR. The relationship between 1392 1454 These results are in accordance with a recently columnar epithelial dysplasia and invasive adenocarcinoma 1393 1455 35 36 arising in Barrett’s esophagus. Am J Clin Pathol 1987;87: 1394 published study, but contrary to another in 1456 which a PPI was administered subcutaneously on 301-12. 1395 3. Moore KH, Barry P, Burn J, Falk G. Adenocarcinoma of the 1457 1396 alternative days without verification of acid rat esophagus in the presence of a proton pump inhibitor: a 1458 1397 suppression by pH measurements. pilot study. Dis Esophagus 2001;14:17-22. 1459 1398 Gastric content may act as a catalyst for the 4. Levrat M, Lambert R, Kirshbaum G. Esophagitis produced 1460 1399 carcinogenic process favoring tumor development. by reflux of duodenal contents in rats. Am J Dig Dis 1962; 1461 1400 When comparing our data with published results 7:564-73. 1462 1401 5. Ireland AP, Peters JH, Smyrk TC, DeMeester TR, Clark GW, 1463 1402 from previous, similar rat studies comparing DGER Mirvish SS, et al. Gastric juice protects against the develop- 1464 1403 and DER, 3 of 4 studies did not show any ment of esophageal adenocarcinoma in the rat. Ann Surg 1465 1404 differences between groups regarding the 1996;224:358-70. 1466 1405 occurrence BE or EA but with small number of 6. Goldstein SR, Yang GY, Curtis SK, Reuhl KR, Liu BC, Mirvish 1467 1406 animals enrolled and lesions not related to SS, et al. Development of esophageal metaplasia and 1468 adenocarcinoma in a rat surgical model without the use of 1407 reflux.7,9,12 The fourth study suggested a protective 1469 1408 a carcinogen. Carcinogenesis 1997;18:2265-70. 1470 effect of gastric juice against the development 7. Fein M, Peters JH, Chandrasoma P, Ireland AP, Oberg S, 1409 5 1471 Ritter MP, et al. Duodenoesophageal reflux induces 1410 of EA. 1472 esophageal adenocarcinoma without exogenous carcin- 1411 For ethical reasons, we chose to include a 1473 ogen. J Gastrointest Surg 1998;2:260-8. 1412 limited number of rats in the control groups. In 1474 8. Buskens CJ, Hulscher JB, van Gulik TM, Ten Kate FJ, van 1413 1475 the PPI group, we performed a preliminary Lanschot JJ. Histopathologic evaluation of an animal model 1414 analysis at 30 weeks. Because the results were for Barrett’s esophagus and adenocarcinoma of the distal 1476 1415 significant, it did not seem ethically valid to esophagus. J Surg Res 2006;135:337-44. 1477 1416 continue the study. 9. Miwa K, Segawa M, Takano Y, Matsumoto H, Sahara H, Yagi 1478 1417 1479 In conclusion, despite pathophysiologic differ- M, et al. Induction of oesophageal and forestomach 1418 carcinomas in rats by reflux of duodenal contents. 1480 1419 ences with humans, the rat ED model reproduces Br J Cancer 1994;70:185-9. 1481 1420 accurately histologic and molecular lesions of 10. Li H, Walsh TN, O’Dowd G, Gillen P, Byrne PJ, Hennessy 1482 the BE carcinogenetic sequence and allowed us TP. Mechanisms of columnar metaplasia and squamous

FLA 5.2.0 DTD SEC CODE: OC n YMSY3448_proof 12 June 2013 8:13 am ce ARTICLE IN PRESS Surgery Gronnier et al 13 Volume j, Number j

1483 regeneration in experimental Barrett’s esophagus. Surgery 24. Bruyere E, Jonckheere N, Frenois F, Mariette C, Van Seunin- 1545 1484 1994;115:176-81. gen I. The MUC4 membrane-bound mucin regulates esoph- 1546 1485 11. Pera M, Brito MJ, Poulsom R, Riera E, Grande L, Hanby A, ageal cancer cell proliferation and migration properties: 1547 1486 et al. Duodenal-content reflux esophagitis induces the Implication for S100A4 protein. Biochem Biophys Res Com- 1548 1487 development of glandular metaplasia and adenosquamous mun 2011;413:325-9. 1549 1488 carcinoma in rats. Carcinogenesis 2000;21:1587-91. 25. Su Y, Chen X, Klein M, Fang M, Wang S, Yang CS, et al. 1550 1489 12. Oberg S, Wenner J, Johansson J, Walther B, Willen R. Phenotype of columnar-lined esophagus in rats with esoph- 1551 1490 Barrett esophagus: risk factors for progression to dysplasia agogastroduodenal anastomosis: similarity to human Bar- 1552 1491 and adenocarcinoma. Ann Surg 2005;242:49-54. rett’s esophagus. Lab Invest 2004;84:753-65. 1553 1492 13. Bonde P, Sui G, Dhara S, Wang J, Broor A, Kim IF, et al. 26. Eberhard D, Tosh D. Transdifferentiation and metaplasia as 1554 1493 Cytogenetic characterization and gene expression profiling a paradigm for understanding development and . 1555 1494 in the rat reflux-induced esophageal tumor model. J Thorac Cell Mol Life Sci 2008;65:33-40. 1556 1495 Cardiovasc Surg 2007;133:763-9. 27. Mirvish SS. Studies on experimental animals involving sur- 1557 14. Mariette C, Piessen G, Leteurtre E, Hemon B, Triboulet JP, gical procedures and/or nitrosamine treatment related to 1496 1558 Van Seuningen I. Activation of MUC1 mucin expression by the etiology of esophageal adenocarcinoma. Cancer Lett 1497 1559 bile acids in human esophageal adenocarcinomatous cells 1997;117:161-74. 1498 1560 and tissues is mediated by the phosphatidylinositol 3-kinase. 28. Attwood SE, Harrison LA, Preston SL, Jankowski JA. Esoph- 1499 1561 Surgery 2008;143:58-71. ageal adenocarcinoma in ‘‘mice and men’’: back to basics!. 1500 15. Piessen G, Jonckheere N, Vincent A, Hemon B, Ducourou- Am J Gastroenterol 2008;103:2367-72. 1562 1501 ble MP, Copin MC, et al. Regulation of the human mucin 29. Sato T, Miwa K, Sahara H, Segawa M, Hattori T. The sequen- 1563 1502 MUC4 by taurodeoxycholic and taurochenodeoxycholic tial model of Barrett’s esophagus and adenocarcinoma 1564 1503 bile acids in oesophageal cancer cells is mediated by hepa- induced by duodeno-esophageal reflux without exogenous 1565 1504 tocyte nuclear factor 1alpha. Biochem J 2007;402:81-91. carcinogens. Anticancer Res 2002;22:39-44. 1566 1505 16. Macke RA, Nason KS, Mukaisho K, Hattori T, Fujimura T, 30. Song S, Guha S, Liu K, Buttar NS, Bresalier RS. COX-2 in- 1567 1506 Sasaki S, et al. Barrett’s esophagus and animal models. duction by unconjugated bile acids involves reactive oxygen 1568 1507 Ann N Y Acad Sci 2011;1232:392-400. species-mediated signalling pathways in Barrett’s oesoph- 1569 1508 17. Oh DS, DeMeester SR, Dunst CM, Mori R, Lehman BJ, agus and oesophageal adenocarcinoma. Gut 2007;56:1512- 1570 1509 Kuramochi H, et al. Validation of a rodent model of 21. 1571 1510 Barrett’s esophagus using quantitative gene expression 31. Langer R, Von Rahden BH, Nahrig J, Von Weyhern C, Re- 1572 1511 profiling. Surg Endosc 2009;23:1346-52. iter R, Feith M, et al. Prognostic significance of expression 1573 1512 18. Fitzgerald RC. Molecular basis of Barrett’s oesophagus and patterns of c-erbB-2, p53, p16INK4A, p27KIP1, cyclin D1 1574 1513 oesophageal adenocarcinoma. Gut 2006;55:1810-20. and epidermal growth factor receptor in oesophageal 1575 1514 19. Zagorowicz E, Jankowski J. Molecular changes in the adenocarcinoma: a tissue microarray study. J Clin Pathol 1576 1515 progression of Barrett’s oesophagus. Postgrad Med J 2007; 2006;59:631-4. 1577 83:529-35. 32. Sherbet GV, Lakshmi MS. S100A4 (MTS1) calcium binding 1516 1578 20. van der Sluis M, Melis MH, Jonckheere N, Ducourouble MP, protein in cancer growth, invasion and metastasis. Anti- 1517 1579 Buller€ HA, Renes I, et al. The murine Muc2 mucin gene is cancer Res 1998;18:2415-21. 1518 1580 transcriptionally regulated by the zinc-finger GATA-4 33. Fein M, Fuchs KH, DeMeester TR, Peters JH, Wittmann D, 1519 1581 transcription factor in intestinal cells. Biochem Biophys Weig M. Evaluation of the intestinal microflora in the rat 1520 Res Commun 2004;325:952-60. model for esophageal adenocarcinoma. Dis Esophagus 1582 1521 21. Gauchotte G, Vigouroux C, Rech F, Battaglia-Hsu SF, 2000;13:39-43. 1583 1522 Soudant M, Pinelli C, et al. Expression of minichromo- 34. Kaur BS, Ouatu-Lascar R, Omary MB, Fang M, Wang S, Yang 1584 1523 some maintenance MCM6 protein in meningiomas is CS, et al. Bile salts induce or blunt cell proliferation in Bar- 1585 1524 strongly correlated with histologic grade and clinical rett’s esophagus in an acid-dependent fashion. Am J Physiol 1586 1525 outcome. Am J Surg Pathol 2012;36:283-91. Gastrointest Liver Physiol 2000;278:1000-9. 1587 1526 22. Vahteristo P, Bartkova J, Eerola H, Syrjakoski K, Ojala S, 35. Miyashita T, Shah FA, Marti GP, Wang J, Bonde P, Gibson 1588 1527 Kilpivaara O, et al. A CHEK2 genetic variant contributing MK, et al. Rabeprazole impedes the development of 1589 1528 to a substantial fraction of familial breast cancer. Am J reflux-induced esophageal cancer in a surgical model. Dig 1590 1529 Hum Genet 2002;71:432-8. Dis Sci 2011;56:1309-14. 1591 1530 23. Zhang Y, Song M, Cui ZS, Li CY, Xue XX, Yu M, et al. Down- 36. Nasr AO, Dillon MF, Conlon S, Downey P, Chen G, Ireland 1592 1531 regulation of TSG101 by small interfering RNA inhibits the A, et al. Acid suppression increases rates of Barrett’s 1593 1532 proliferation of breast cancer cells through the MAPK/ERK esophagus and esophageal injury in the presence of 1594 1533 signal pathway. Histol Histopathol 2011;26:87-94. duodenal reflux. Surgery 2012;15:382-90. 1595 1534 1596 1535 1597 1536 1598 1537 1599 1538 1600 1539 1601 1540 1602 1541 1603 1542 1604 1543 1605 1544 1606

FLA 5.2.0 DTD SEC CODE: OC n YMSY3448_proof 12 June 2013 8:13 am ce Our reference: YMSY 3448 P-authorquery-v9

AUTHOR QUERY FORM

Journal: YMSY

Article Number: 3448

Dear Author,

Please check your proof carefully and mark all corrections at the appropriate place in the proof (e.g., by using on-screen annotation in the PDF file) or compile them in a separate list. Note: if you opt to annotate the file with software other than Adobe Reader then please also highlight the appropriate place in the PDF file. To ensure fast publication of your paper please return your corrections within 48 hours.

For correction or revision of any artwork, please consult http://www.elsevier.com/artworkinstructions.

Any queries or remarks that have arisen during the processing of your manuscript are listed below and highlighted by flags in the proof.

Location Query / Remark: Click on the Q link to find the query’s location in text in article Please insert your reply or correction at the corresponding line in the proof If there are any drug dosages in your article, please verify them and indicate that you have done so by initialing this query Q1 Please check heading levels throughout to ensure that they are correct. Q2 Please confirm that given names and surnames have been identified correctly. Q3 For figure 3A/D, the resolution is too low to be used. Please provide better quality figure of 300 dpi. EQ1 Figure 5: ‘***’ not defined at caption. Kindly check and provide.

Please check this box or indicate your approval if you have no , corrections to make to the PDF file

Thank you for your assistance.