MicroRNA Expression Profiling in the Histological Subtypes of Barrett's Metaplasia

Abstract

OBJECTIVES:The histological definition of Barrett's esophagus (BE) is debated, particularly regarding the phenotype of its metaplastic columnar epithelium. Histologically proven intestinal metaplasia (IM) was the sine qua non condition for a diagnosis of BE but, more recently, non-intestinalized (i.e., cardiac gastric-type; GM) columnar metaplasia has been re-included in the spectrum of Barrett's histology. MicroRNAs modulate cell commitment, and are also reportedly dysregulated in Barrett's carcinogenesis. This study investigates miRNA expression in the histological spectrum of esophageal columnar metaplastic changes, specifically addressing the biological profile of GM vs. IM.METHODS:A study was performed to discover microRNA microarray in 30 matching mucosa samples obtained from 10 consecutive BE patients; for each patient, biopsy tissue samples were obtained from squamous, GM and intestinalized epithelium. Microarray findings were further validated by qRT-PCR analysis in another bioptic series of 75 mucosa samples.RESULTS:MicroRNA profiling consistently disclosed metaplasia-specific microRNA signatures. Six microRNAs were significantly dysregulated across the histological phenotypes considered; five of them (two overexpressed (hsa-miR-192; -miR-215) and three under-expressed (hsa-miR-18a*; -miR-203, and -miR-205)) were progressively dysregulated in the phenotypic sequence from squamous to gastric-type, to intestinal-type mucosa samples.CONCLUSIONS:A consistent microRNA expression signature underlies both gastric- and intestinal-type esophageal metaplasia. The pattern of microRNA dysregulation suggests that GM may further progress to IM. The clinico-pathological implications of these molecular profiles prompt further study on the "personalized" cancer risk associated with each of these metaplastic transformations.

Citation: Clinical and Translational Gastroenterology (2013) 4, e34; doi:10.1038/ctg.2013.5

& 2013 the American College of Gastroenterology All rights reserved 2155-384X/13 http://www.nature.com/ctg

Web End =www.nature.com/ctg

Matteo Fassan, MD1,2, Stefano Volinia, MD3, Jeff Palatini, PhD3, Marco Pizzi, MD1, Cecilia Fernandez-Cymering, PhD3, Mariangela Balistreri, BD1, Stefano Realdon, MD4, Giorgio Battaglia, MD4, Rhonda Souza, MD5, Robert D. Odze, MD, FACG, FRCPc6, Giovanni Zaninotto, MD7, Carlo M. Croce, MD3 and Massimo Rugge, MD, FACG1,4

OBJECTIVES: The histological denition of Barretts esophagus (BE) is debated, particularly regarding the phenotype of its metaplastic columnar epithelium. Histologically proven intestinal metaplasia (IM) was the sine qua non condition for a diagnosis of BE but, more recently, non-intestinalized (i.e., cardiac gastric-type; GM) columnar metaplasia has been re-included in the spectrum of Barretts histology. MicroRNAs modulate cell commitment, and are also reportedly dysregulated in Barretts carcinogenesis. This study investigates miRNA expression in the histological spectrum of esophageal columnar metaplastic changes, specically addressing the biological prole of GM vs. IM.

METHODS: A study was performed to discover microRNA microarray in 30 matching mucosa samples obtained from 10 consecutive BE patients; for each patient, biopsy tissue samples were obtained from squamous, GM and intestinalized epithelium. Microarray ndings were further validated by qRT-PCR analysis in another bioptic series of 75 mucosa samples. RESULTS: MicroRNA proling consistently disclosed metaplasia-specic microRNA signatures. Six microRNAs were signicantly dysregulated across the histological phenotypes considered; ve of them (two overexpressed (hsa-miR-192; -miR-215) and three under-expressed (hsa-miR-18a*; -miR-203, and -miR-205)) were progressively dysregulated in the phenotypic sequence from squamous to gastric-type, to intestinal-type mucosa samples.

CONCLUSIONS: A consistent microRNA expression signature underlies both gastric- and intestinal-type esophageal metaplasia. The pattern of microRNA dysregulation suggests that GM may further progress to IM. The clinico-pathological implications of these molecular proles prompt further study on the personalized cancer risk associated with each of these metaplastic transformations. Clinical and Translational Gastroenterology (2013) 4, e34; doi:http://dx.doi.org/10.1038/ctg.2013.5

Web End =10.1038/ctg.2013.5 ; published online 16 May 2013

Subject Category: Esophagus

INTRODUCTION

Barretts esophagus (BE) is dened as the metaplastic replacement of native esophageal squamous mucosa by columnar epithelia.1,2 The columnar esophageal metaplasia is considered the cancerization eld in which esophageal Barretts adenocarcinoma develops.35

The clinico-pathological denition of BE is widely disputed.69 Based on its original description, BE should include any replacement of native esophageal epithelium by glandular mucosa.10 The elective link between the prevalence of intestinal metaplasia (IM) and the risk of Barretts adenocarcinoma has since led the denition of BE to be operatively restricted to columnar intestinal-type metaplasia alone.11,12

On the other hand, recent evidence of a close relationship between esophageal adenocarcinoma and cardiac-type mucosa would support a histogenetic role for non-intestinalized epithelia in Barretts carcinogenesis as well, meaning that columnar, non-intestinalized metaplasia should be re-included in the spectrum of Barretts disease.1315

MicroRNA Expression Proling in the Histological Subtypes of Barretts Metaplasia

These conicting denitions of BE are not just a matter of semantics: any inconsistencies in the diseases histological assessment result in major variations in the estimates of its incidence, and this ultimately affects the cost-benet balance of any Barretts adenocarcinoma secondary prevention strategy.16

MicroRNAs (miRNAs or miR) are a class of short non-coding RNAs involved in modulating gene expression by targeting messenger RNAs (mRNAs). Depending on their complementarity with target mRNAs, miRNAs can either block their translation or hasten their degradation.1719

Several reports point to miRNAs as molecules involved in each step of Barretts carcinogenesis,17,1925 but no studies

have so far provided any comprehensive miRNA expression proling in esophageal intestinalized vs. non-intestinalized columnar metaplasia. This study aimed to further characterize the molecular prole of the different metaplastic phenotypes considered candidates for inclusion in the spectrum of Barretts mucosa.

1Department of Medicine (DIMED), Surgical Pathology and Cytopathology Unit, University of Padua, Padua, Italy; 2Department of Surgical, Oncological and Gastroenterological Sciences (DiSCOG), General Oncology Unit, University of Padua, Padua, Italy; 3Comprehensive Cancer Center, Ohio State University, Columbus, Ohio, USA; 4Istituto Oncologico Veneto - IOV-IRCCS, Padua, Italy; 5Department of Medicine, University of Texas Southwestern Medical Center and VA North Texas Health Care System, Dallas, Texas, USA; 6Department of Pathology, Brigham and Womens Hospital, Boston, Massachusetts, USA and 7Department of Surgical Oncology and Gastroenterology Sciences (DiSCOG), Surgery Unit, University of Padua, Padua, ItalyCorrespondence: Massimo Rugge, MD, FACG, Department of Medicine (DIMED), Surgical Pathology and Cytopathology Unit, University of Padua, Istituto Oncologico Veneto-IRCCS; Via Aristide Gabelli, 61, 35121 Padua, Italy. E-mail: mailto:[email protected]

Web End [email protected] Received 4 October 12; revised 21 February 13; accepted 29 March 13

miRNAs and Esophageal Metaplasia

Fassan et al.

2

MATERIALS AND METHODS

Patients. The cases considered in the present study were collected retrospectively from the les of the Veneto Regions multicenter Barretts Esophagus Registry (EBRA (http://www.esofagodibarrett.it

Web End =www. http://www.esofagodibarrett.it

Web End =esofagodibarrett.it ); Padua Unit, Surgical Pathology and Cytopathology Unit at Padua University).26 All patients had endoscopically conrmed Z3 cm long segments of columnar metaplasia in the distal esophagus and had undergone endoscopy/biopsy according to the Seattle protocol (i.e., four-quadrant biopsies obtained from every 2 cm of meta-plastic mucosa).27 Only patients whose non-intestinalized and intestinalized metaplastic lesions coexisted within the same quadrant were considered; mosaic patterns of BE were excluded. No cases of epithelial dysplasia/neoplasia, or cases obtained from patients being followed-up endoscopically for epithelial dysplasia/neoplasia, were considered. The institutes ethical regulations concerning research on human tissues were followed. Original slides or serial sections (46 mm thick) obtained from archival parafn-embedded tissue samples (H&E, Alcian-PAS) were jointly re-assessed by two pathologists (MF and MP); where their opinions differed, a third GI-specialist pathologist (MR) was consulted.

Histopathology. Metaplastic lesions were classied as nonintestinalized or intestinalized based on the presence of goblet cells, assessed by Alcian-PAS staining.

Non-intestinalized metaplastic lesions were further classied as follows:

(i) cardiac gastric-type (GM), when only non-goblet muco-secreting columnar cells were found at histology;

(ii) oxyntic-type, when the histological phenotype was that of native corpus/fundus gastric glands. These lesions are currently considered as inlet patches of oxyntic ectopia, and were excluded for the purposes of the present study;8

(iii) multilayered epithelium (MLE), dened as multilayered, attened squamoid epithelium overlaid by columnar

mucus-producing, non-intestinalized cells. It has been suggested that this lesion is a putative early stage in the development of BE.28

Depending on goblet cells density, cases of intestinalized metaplasia were further classied as having a low goblet density (IM / intestinal commitment in o50% of

glandular structures) or high goblet density (IM intestinal

commitment in 450% of glandular structures).14,15

Sample selection. A total of 105 mucosa samples obtained at biopsy from 58 BE patients (age 64.28.5 years, range 5477; all Caucasian males) were considered, and used as shown in Table 1.

A rst discovery set, used in the miRNA microarray study, concerned 10 histologically proven long-segment BE patients (mean age 61.86.9, range 5476; all Caucasian males). Three biopsies were used for each patient (Table 1):(i) a squamous epithelium sample obtained no less than 3 cm away from any type of metaplastic mucosa; (ii) a GM sample obtained no o2 cm away from the gastroesophageal junction;

and (iii) an IM sample obtained no o2 cm away from the gastroesophageal junction and comprising 450% of goblet cells.

A validation set, used in the qRT-PCR study, consisted of 75 biopsy samples obtained from 48 cases of histologically proven long-segment BE (patients were all Caucasian males; mean age 65.58.3, range 5777). The following histologically validated tissue samples were considered (Table 1): (i) squamous esophageal epithelium obtained no o3 cm away from the metaplastic mucosa 15 cases; (ii) MLE 15 cases;

(iii) GM obtained no o2 cm away from the gastroesophageal junction 15 cases; (iv) IM / obtained no o2 cm away

from the gastroesophageal junction and comprising o50% of

goblet cells 15 cases; and (v) IM obtained no o2 cm away

from the gastroesophageal junction and comprising 450% of

goblet cells 15 cases.

In all cases, lesions were microdissected manually, based on the original H&E slides, from formalin-xed, parafn-embedded samples to obtain at least 80% of target cells.

Table 1 Schematic diagram of the present study

Discovery set Validation set

miRNA Microarray Study (326 miRNA genes)

qRT-PCR Study (7 miRNA genes)

10 Long-segment BE patients(Number of biopsy samples considered per patient)

48 Long-segment BE patients(Number of biopsy samples collected from the series considered)

1 Biopsy sample of squamous epithelium (Z3 cm away from any type of metaplastic mucosa)

15 Biopsy samples of squamous epithelium (Z3 cm away from any type of metaplastic mucosa)

15 Biopsy samples of multilayered epithelium

1 Biopsy sample of gastric-type mucosa (Z2 cm away from the GEJ)

15 Biopsy samples of gastric-type mucosa (Z2 cm away from the GEJ)15 Biopsy samples of intestinalized mucosa with a low goblet cell density (IM / ; o50% goblet cells; Z2 cm away from the GEJ)

1 Biopsy sample of intestinalized mucosa with a high goblet cell density (450% goblet cells; Z2 cm away from the GEJ)

15 Biopsy samples of intestinalized mucosa with a high goblet cell density (450% goblet cells; Z2 cm away from the GEJ)

Abbreviations: BE, Barretts esophagus; GEJ, gastroesophageal junction; miRNA, microRNA.

A discovery set of 30 biopsy samples was used in the miRNA microarray study, and a validation set of 75 biopsy samples was used in the qRT-PCR study.

Clinical and Translational Gastroenterology

miRNAs and Esophageal Metaplasia Fassan et al.

3

miRNA microarray. Formalin-xed, parafn-embedded biopsy samples were deparafnized and underwent total RNA extraction using the RecoverAll kit according to the manufacturers instructions (Ambion Inc, Austin, TX). RNA labeling and hybridization on miRNA microarray chips were done as described in detail elsewhere.25,29 Briey, 5 mg of total

RNA from each sample were reverse-transcribed using biotin end-labeled random-octamer oligonucleotide primer. Biotin-labeled complementary DNA was hybridized on an Ohio State University custom miRNA microarray chip (OSU_CCC version4.0) containing B1,100 miRNA probes, including 326 human and 249 mouse miRNA genes, plus 10 control genes, spotted in duplicate. The hybridized chips were washed and processed for biotin-containing transcript detection by streptavidin-Alexa 647 conjugate and scanned on an Axon 4000B microarray scanner (Axon Instruments, Sunnyvale, CA).

Statistical and bioinformatic analyses. Microarray images were analyzed using GENEPIX PRO 6.0 (Axon Instruments). Average values of the replicate spots of each miRNA were background subtracted, normalized using quantiles enabling a comparison between chips, and further analyzed. The microarray data are deposited in the Gene Expression Omnibus at the National Center for Biotechnology Information (GEO: GSE24839). The miRNAs that were expressed differently in the different esophageal lesions were identied using a random-variance t-test, which is an improvement over the standard separate t-test, because it enables information on within-class variation to be shared among genes without assuming that all genes have the same variance. Genes were considered statistically signicant if their P value was o0.001; a stringent signicance threshold was used to limit the number of false-positive ndings. A linear regression model using normalized log2-transformed miRNA expression values was applied to test signicant dysregulated miRNAs in the different metaplastic lesions, and P values were adjusted for multiple testing using FDR correction. Only FDRo0.0005 and Po0.0001 were considered. Only mature miRNAs that were differently expressed are reported.

Quantitative real-time polymerase chain reaction (qRTPCR). The NCode miRNA qRT-PCR method (Invitrogen,

Carlsbad, CA) was used to detect and quantify mature miRNAs on Applied Biosystems (Foster City, CA) qRT-PCR instruments according to the manufacturers instructions.25,30

Normalization was done with the small nuclear RNA U6B (RNU6B; Invitrogen). All real-time reactions, including no-template controls and real-time minus controls, were run in a GeneAmp PCR 9700 thermocycler (Applied Biosystems). Gene expression levels were quantied using the ABI Prism 7900HT Sequence Detection System (Applied Biosystems). Comparative real-time PCR was performed in triplicate, including no-template controls. The fold difference for each sample was obtained using the equation 2 dCt, where Ct is the threshold cycle and dCt stands for the Ct average sample genethe Ct average RNU6B. Differential expression was tested using the two-sided t-test.

RESULTS

miRNAs are dysregulated in Barretts metaplasia. To identify the miRNA proles that are dysregulated in columnar Barretts mucosa, a miRNA microarray analysis was performed on a discovery set of 30 matching biopsy samples obtained from 10 long-segment BE patients (Table 1). The miRNA microarray analysis was performed using a validated custom microarray platform.29,31

Different miRNA expression proles were identied by comparing the different metaplastic phenotypes (Table 2). Eleven miRNAs were found dysregulated (Po0.001) in

GM by comparison with squamous epithelium (Figure 1a), and six in IM by comparison with squamous epithelium (Figure 1b). All six IM-associated miRNAs were shared with the GM signature (Table 2).

In GM samples, six miRNAs were signicantly down-regulated (i.e., hsa-miR-18a*, hsa-miR-205, hsa-miR-203, hsa-miR-20a, hsa-miR-106a, and hsa-miR-20b) and ve were upregulated (i.e., hsa-miR-611, hsa-miR-145, hsa-miR-6625*, hsa-miR-192, and hsa-miR-215). The IM signature disclosed ve downregulated miRNAs (hsa-miR-18a*, hsamiR-205, hsa-miR-203, hsa-miR-20a, and hsa-miR-106a) and one upregulated miRNA (hsa-miR-215).

The different expression of six miRNAs was assessed in the squamous vs. non-intestinalized vs. intestinalized

Table 2 Differently expressed miRNAs in Barretts metaplastic lesions

miRNA S/GM/IM sequence S vs. GM S vs. IM

P FDR-adjusted P Fold change P FDR-adjusted P Fold change P FDR-adjusted P

hsa-miR-18a* o0.0001 o0.0001 0.00 o1e-07 o1e-07 0.01 1.0e-07 3.8e-05 hsa-miR-205 o0.0001 o0.0001 0.01 o1e-07 o1e-07 0.02 3.0e-07 5.6e-05 hsa-miR-203 o0.0001 0.0005 0.21 1.5e-05 0.001 0.21 1.2e-03 0.015 hsa-miR-20a 0.23 4.1e-03 0.017 0.26 6.3e-03 0.047 hsa-miR-106a 0.32 8.9e-03 0.031 0.27 8.6e-03 0.053 hsa-miR-20b 0.46 8.3e-05 0.004 hsa-miR-611 1.65 8.9e-03 0.031 hsa-miR-145 2.57 3.5e-05 0.002 hsa-miR-625* 2.58 3.4e-03 0.016 hsa-miR-192 o0.0001 0.0001 5.29 2.7e-05 0.002

hsa-miR-215 o0.0001 o0.0001 12.90 6.6e-06 0.001 16.81 2.4e-03 0.022 hsa-miR-194 o0.0001 0.0001

Abbreviations: FDR, false discovery rate; GM, gastric metaplasia; IM, intestinal metaplasia; miRNA, microRNA; S, squamous epithelium.

Clinical and Translational Gastroenterology

miRNAs and Esophageal Metaplasia Fassan et al.

4

Figure 1 miRNA expression is altered in esophageal metaplastic lesions. (a) miRNA signicantly dysregulated (Po0.001) in gastric metaplasia (right panel) by comparison with squamous esophageal epithelium (left panel). (b) miRNA signicantly dysregulated (Po0.001) in intestinal metaplasia (right panel) by comparison with squamous esophageal epithelium (left panel). Rows represent individual genes; columns represent individual tissue samples. The gray scale indicates transcript levels below, equal to, or above the mean (white, gray, and black, respectively); the scale represents the intensity of gene expression (log2 scale ranges between 3 and 3).

phenotypes (Figure 2 and Table 2; logistic regression, FDR o0.0005). Three miRNAs were found increasingly upregulated (i.e., hsa-miR-215, hsa-miR-192, and hsa-miR-194), and three were increasingly downregulated (i.e., hsa-miR-18a*, hsa-miR-205, and hsa-miR-203).

qRT-PCR validation. To validate the results of the micro-array analysis, qRT-PCR analysis was performed on an independent series of 75 endoscopic biopsy samples obtained from 48 long-segment BE patients (Table 1). The analysis included 15 samples of MLE, which is recognized as an early-intermediate form of columnar metaplasia with both squamous and columnar features. A set of low-density IM samples was also considered to test the inuence of the prevalence of goblet cells on miRNA assessment.

Seven miRNA dysregulations were validated (Figure 3 and Table 3), i.e.,: (i) ve miRNAs in the squamous to GM to IM sequence (hsa-miR-18a*, hsa-miR-205, hsa-miR-203, hsamiR-192, and hsa-miR-215; Table 2); and (ii) two miRNAs shared by the squamous vs. GM, and by the squamous vs. IM proles (hsa-miR-20a and hsa-miR-106a).

As for the comparisons between squamous and GM, and between squamous and IM, qRT-PCR results were consistent with those obtained by miRNA microarray analysis. As expected, hsa-miR-20a and hsa-miR-106a showed no signicant difference between GM and IM. It is noteworthy that:(i) MLE showed a miRNA dysregulation comparable with the picture seen in columnar metaplastic lesions; and (ii) goblet cell density did not signicantly affect the results (Figure 3, Table 3).

DISCUSSION

It has recently been reported that the incidence of cancer among BE patients is lower than was previously believed,16,32,33 but at least three potential biases may

signicantly affect the assessment of the cancer risk associated with BE: (i) the histological denition of Barretts metaplasia; (ii) the signicant variability in the endoscopic diagnostic approach, including the biopsy sampling protocols; and (iii) inconsistencies in endoscopistpathologist interactions. The most important factor concerns the unequivocal histological denition of the Barretts mucosa phenotype, which is crucial to a consistent identication of the population at higher neoplastic risk, and to a personalized secondary cancer prevention strategy.

In addition to the well-established relationship between cancer and esophageal IM, recent studies have associated neoplastic progression with non-intestinalized metaplasia too.1216 Elucidating this point is challenging, however: results obtained in experimental models are only partially applicable to humans, and information obtained in clinical trials is strongly affected by a signicant variability in histological assessments and endoscopic follow-up protocols.

Unlike most RNA molecules, miRNAs are long-living in vivo and very stable in vitro.17,28,34 These structural characteristics

allow for miRNA testing in formalin-xed, parafn-embedded tissue samples, which is essential to link specic biological signatures with well-established histological phenotypes.17

miRNA expression proling thus has the potential for histologically distinguishing between and classifying different lesions, and several reports have demonstrated the excellent reproducibility of miRNA expression proling in (archival) formalin-xed, parafn-embedded tissue samples.35

Several studies have focused on miRNA dysregulation in Barretts carcinogenesis1925 and specic miRNA expression

signatures have been associated with cancer progression,2125

whereas the molecular proling of BE-related metaplastic changes had never been investigated.

The present ndings show that: (i) miRNA dysregulation occurs early in the morphogenesis of Barretts mucosa and is at least partially responsible for its columnar metaplastic

Clinical and Translational Gastroenterology

miRNAs and Esophageal Metaplasia Fassan et al.

5

Figure 2 MicroRNA (miRNA) expression is altered in the progression from squamous epithelium to intestinal metaplasia. miRNA was signicantly (FDRo0.001) dysregulated in the progression from squamous epithelium to gastric metaplasia to intestinal metaplasia. Rows represent individual genes; columns represent individual tissue samples. Pseudo-colors indicate transcript levels below, equal to, or above the mean (green, black, and red, respectively); the scale represents the intensity of gene expression (log2 scale ranges between 3 and 3).

Figure 3 qRT-PCR analysis for dysregulated miRNAs in metaplastic lesions. A total of 75 biopsy samples were considered, comprising: 15 squamous mucosa, 15 multilayered epithelium (MLE), 15 gastric metaplasia cardiac-type (GM), 15 low-density intestinal metaplasia (IM / ), and 15 high-density IM. Two microRNAs (miRNAs;

hsa-miR-192 and hsa-miR-215) were signicantly upregulated in the metaplastic tissue by comparison with the squamous epithelium, whereas ve miRNAs (hsa-miR-18a, hsa-miR-20a, hsa-miR-106a, hsa-miR-203, and hsa-miR-205) were downregulated. Rows represent individual genes; columns represent different lesion classes. Pseudo-colors indicate transcript levels below, equal to, or above the mean (green, black, and red, respectively); the scale represents the log2 difference between the mean expression levels seen in the metaplastic lesions and squamous epithelium. Numerical values are given in Table 3.

transformation; (ii) similar miRNA dysregulations are behind both non-intestinalized and intestinalized columnar metaplasia, and this supports a basic biological consistency of the pathway leading to metaplastic columnar changes (with and without a goblet component); and (iii) miRNA signatures (Tables 2 and 3) show a progressive dysregulation along the path from squamous to nonintestinalized to intestinalized metaplasia, supporting the hypothesis of a progressive transformation from a nonintestinalized (earlier) to an intestinalized (more advanced) Barretts phenotype.

The present results are in keeping with previous reports on miRNA dysregulation in esophageal IM, in which we showed a signicant downregulation of hsa-miR-203/-miR-205, and a concomitant upregulation of hsa-miR-192/-miR-215.2123,25

These miRNAs have also been associated with the whole Barretts carcinogenic cascade,2023,25 further reinforcing the

role of their dysregulation in the molecular natural history of Barretts disease.hsa-miR-203 is known to target the transcription factor p63, and is therefore involved in maintaining the squamous commitment of different stratied epithelia.3639 The reported

Clinical and Translational Gastroenterology

miRNAs and Esophageal Metaplasia

Fassan et al.

6

Table 3 miRNA expression tested by qRT-PCR analysis in the different metaplastic lesions

MLE GM IM /- IM P (S vs. GM) P (S vs. IM) P (GM vs. IM) hsa-miR-18a* 2.50.3 3.20.3 3.10.2 3.60.3 o0.01 o0.01 NS

hsa-miR-20a 1.10.1 1.00.1 1.20.1 1.10.1 0.04 0.04 NS

hsa-miR-106a 2.80.3 3.00.3 3.00.2 2.90.2 o0.01 o0.01 NS

hsa-miR-192 2.20.2 3.00.4 2.90.3 3.60.3 o0.01 o0.01 0.02 hsa-miR-215 4.20.5 5.20.4 5.20.4 5.60.5 o0.01 o0.01 NS hsa-miR-203 2.20.3 3.50.2 6.70.6 7.10.6 o0.01 o0.01 o0.01

hsa-miR-205 1.90.2 3.40.3 8.10.8 12.10.9 o0.01 o0.01 o0.01

Abbreviations: FDR, false discovery rate; GM, gastric metaplasia; IM, intestinal metaplasia; IM / , low-density intestinal metaplasia; MLE, multilayered epithelium;

NS, not statistically signicant; S, squamous epithelium.

Data are expressed as log2(lesion)-log2(native esophageal epithelium)s.e.

downregulation of hsa-miR-203 is basically consistent with the loss of the native squamous phenotype, and matches with the emerging columnar morphology.

The concomitant upregulation of two p53-induced miRNAs like hsa-miR-192 and hsa-miR-215 could be in response to the genotoxic stress caused by chronic gastroesophageal acid reux.40,41

The progressive upregulation of hsa-miR-194 in the squamous to GM to IM sequence is consistent with its biological function: this miRNA is involved in the commitment and maturation of intestinal epithelia, and it is regulated by the hepatocyte nuclear factor 1a.42 The role of hsa-miR-194 overexpression in establishing the intestinal phenotype warrants further investigation in in vivo models.

A novel nding in the present study is the similarity in the expression proles of MLE and overt metaplastic lesions (GM and both low- and high-density IM). The trend of miRNA dysregulation also suggests a multistep metaplastic transformation of the esophageal mucosa, evolving from the native squamous epithelium to GM columnar cells, and then to a full intestinal phenotype.

Future efforts should focus on assessing similarities and differences in miRNA signatures between gastric/esophageal IM and normal intestinal mucosa, and between normal gastric cardia and cardiac-type esophageal metaplasia. This could help us to identify novel biomarkers of metaplastic transformation for use in clinical practice when it comes to deciding secondary prevention strategies.

In conclusion, this mRNA proling study disclosed similar molecular dysregulations in both non-intestinalized and intestinalized columnar esophageal metaplasia, supporting the impression that any type of columnar transformation is part of the biological spectrum of Barretts mucosa. Further prospective studies, also on different ethnic groups, should seek to assess the cancer risk associated with the different Barretts mucosa phenotypes.

CONFLICT OF INTEREST

Guarantor of the article: Massimo Rugge, MD, FACG. Specic author contributions: All authors approved the nal version of the manuscript. Study concept and design:M. Fassan and M. Rugge; data acquisition: M. Balistreri,G. Battaglia, M. Fassan, and M. Pizzi, M. Rugge; data analysis and interpretation: M. Fassan, C. Fernandez-Cymering,J. Palatini, and S. Volinia; drafting of the manuscript: C.M. Croce,M. Fassan, R.D. Odze, M. Rugge, R Souza, and G. Zaninotto.

Financial support: This research was partially supported by the AIRC grant Veneto Region, 2008. The funding agency had no role in the design and performance of the study.

Potential competing interests: None.

Ethical approval: Only material that was not required for diagnostic purposes was used and all patients signed an informed consent approved by the Padua University Hospital Review Board, which allows researchers to use excess material for research purposes.

Acknowledgements. The microarray data are deposited in the Gene Expression Omnibus at the National Center for Biotechnology Information (GEO:GSE24839). We thank Frances Coburn for text editing, and Cristiano Lanza and Vanni Lazzarin for technical assistance.

Study Highlights

WHAT IS CURRENT KNOWLEDGE

| The denition of Barretts mucosa is controversial, particularly regarding the phenotype of its metaplastic columnar epithelium.

| Recent studies have re-included non-intestinalized columnar metaplasia in the histological spectrum of

Barretts mucosa.

| MicroRNAs control gene expression by targeting messenger RNAs, and have recently been found dysregulated in Barretts carcinogenesis.

WHAT IS NEW HERE

| miRNA dysregulation occurs early in esophageal metaplastic transformation.

| miRNA expression proling shows similar molecular dysregulations in both non-intestinalized and intestinalized columnar esophageal metaplasia, suggesting that any type of columnar transformation should be included in the biological spectrum of Barretts mucosa.

| The Barretts adenocarcinoma risk associated with intestinalized vs. non-intestinalized BE should be specically assessed, and BE follow-up strategies decided accordingly.

1. American Gastroenterological AssociationSpechler SJ, Sharma P, Souza RF et al. American Gastroenterological Association medical position statement on the management of Barretts esophagus. Gastroenterology 2011; 140: 10841091.

2. Spechler SJ, Sharma P, Souza RF et al. American Gastroenterological Association technical review on the management of Barretts esophagus. Gastroenterology 2011; 140: e18e52.

Clinical and Translational Gastroenterology

miRNAs and Esophageal Metaplasia Fassan et al.

7

3. Odze RD. Barrett esophagus: histology and pathology for the clinician. Nat Rev Gastroenterol Hepatol 2009; 6: 478490.

4. Paulson TG, Reid BJ. Focus on Barretts esophagus and esophageal adenocarcinoma. Cancer Cell 2004; 6: 1116.

5. Fitzgerald RC. Molecular basis of Barretts oesophagus and oesophageal adenocarcinoma. Gut 2006; 55: 18101820.

6. Riddell RH, Odze RD. Denition of Barretts esophagus: time for a rethinkis intestinal metaplasia dead? Am J Gastroenterol 2009; 104: 25882594.

7. Rugge M, Fassan M, Battaglia G et al. Intestinal or gastric? The unsolved dilemma of Barretts metaplasia. Hum Pathol 2009; 40: 12061207.

8. Fiocca R, Mastracci L, Milione M et al. Microscopic esophagitis and Barretts esophagus: the histology report. Dig Liver Dis 2011; 43(Suppl 4): S319S330.

9. Fassan M, Lanza C, Lazzarin V et al. The original sin of oesophageal mucosa. Dig Liver Dis 2011; 43: 246.

10. Barrett NR. The oesophagus lined by columnar epithelium. Gastroenterologia 1956; 86: 183186.

11. Jankowski JA, Harrison RF, Perry I et al. Barretts metaplasia. Lancet 2000; 356: 20792085.12. Vakil N, van Zanten SV, Kahrilas P et al. The Montreal denition and classication of gastroesophageal reux disease: a global evidence-based consensus. Am J Gastroenterol 2006; 101: 19001920.

13. Takubo K, Aida J, Naomoto Y et al. Cardiac rather than intestinal-type background in endoscopic resection specimens of minute Barrett adenocarcinoma. Hum Pathol 2009; 40: 6574.

14. Hahn HP, Blount PL, Ayub K et al. Intestinal differentiation in metaplastic, nongoblet columnar epithelium in the esophagus. Am J Surg Pathol 2009; 33: 10061015.

15. Liu W, Hahn H, Odze RD et al. Metaplastic esophageal columnar epithelium without goblet cells shows DNA content abnormalities similar to goblet cell-containing epithelium. Am J Gastroenterol 2009; 104: 816824.

16. Rugge M, Fassan M, Cavallini F et al. Re: Risk of malignant progression in Barretts esophagus patients: results from a large population-based study. J Natl Cancer Inst 2012; 104: 17711772.

17. Fassan M, Croce CM, Rugge M. miRNAs in precancerous lesions of the gastrointestinal tract. World J Gastroenterol 2011; 17: 52315239.

18. Croce CM. Causes and consequences of microRNA dysregulation in cancer. Nat Rev Genet 2009; 10: 704714.

19. Kan T, Meltzer SJ. MicroRNAs in Barretts esophagus and esophageal adenocarcinoma. Curr Opin Pharmacol 2009; 9: 727732.

20. Feber A, Xi L, Luketich JD et al. MicroRNA expression proles of esophageal cancer. J Thorac Cardiovasc Surg 2008; 135: 255260.

21. Kan T, Sato F, Ito T et al. The miR-106b-25 polycistron, activated by genomic amplication, functions as an oncogene by suppressing p21 and Bim. Gastroenterology 2009; 136: 16891700.

22. Yang H, Gu J, Wang KK et al. MicroRNA expression signatures in Barretts esophagus and esophageal adenocarcinoma. Clin Cancer Res 2009; 15: 57445752.

23. Wijnhoven BP, Hussey DJ, Watson DI et al. MicroRNA proling of Barretts oesophagus and oesophageal adenocarcinoma. Br J Surg 2010; 97: 853861.

24. Bansal A, Lee IH, Hong X et al. Feasibility of mcroRNAs as biomarkers for Barretts Esophagus progression: a pilot cross-sectional, phase 2 biomarker study. Am J Gastroenterol 2011; 106: 10551063.

25. Fassan M, Volinia S, Palatini J et al. MicroRNA expression proling in human Barretts carcinogenesis. Int J Cancer 2011; 129: 16611670.

26. Zaninotto G, Minnei F, Guirroli E et al. The Veneto Regions Barretts Oesophagus Registry: aims, methods, preliminary results. Dig Liver Dis 2007; 39: 1825.

27. Levine DS, Blount PL, Rudolph RE et al. Safety of a systematic endoscopic biopsy protocol in patients with Barretts esophagus. Am J Gastroenterol 2000; 95: 11521157.

28. Glickman JN, Spechler SJ, Souza RF et al. Multilayered epithelium in mucosal biopsy specimens from the gastroesophageal junction region is a histologic marker of gastroesophageal reux disease. Am J Surg Pathol 2009; 33: 818825.

29. Baffa R, Fassan M, Volinia S et al. MicroRNA expression proling of human metastatic cancers identies cancer gene targets. J Pathol 2009; 219: 214221.

30. Fassan M, Pizzi M, Battaglia G et al. Programmed cell death 4 (PDCD4) expression during multistep Barretts carcinogenesis. J Clin Pathol 2010; 63: 692696.

31. Volinia S, Calin GA, Liu CG et al. A microRNA expression signature of human solid tumors denes cancer gene targets. PNAS 2006; 103: 22572261.

32. Hvid-Jensen F, Pedersen L, Drewes AM et al. Incidence of adenocarcinoma among patients with Barretts esophagus. N Engl J Med 2011; 365: 13751383.

33. Bhat S, Coleman HG, Yousef F et al. Risk of malignant progression in Barretts esophagus patients: results from a large population-based study. J Natl Cancer Inst 2011; 103: 10491057.

34. van Baal JW, Verbeek RE, Bus P et al. microRNA-145 In Barretts oesophagus: regulating BMP4 signalling via GATA6. Gut 2012; 62: 664675.

35. Streichert T, Otto B, Lehmann U. MicroRNA expression proling in archival tissue specimens: methods and data processing. Mol Biotechnol 2012; 50: 159169.

36. Yi R, OCarroll D, Pasolli HA et al. Morphogenesis in skin is governed by discrete sets of differentially expressed microRNAs. Nat Genet 2006; 38: 356362.

37. Tran N, McLean T, Zhang X et al. MicroRNA expression proles in head and neck cancer cell lines. Biochem Biophys Res Commun 2007; 358: 1217.

38. Lena AM, Shalom-Feuerstein R, Rivetti di Val Cervo P et al. miR-203 Represses stemness by repressing DeltaNp63. Cell Death Differ 2008; 15: 11871195.

39. Yuan Y, Zeng ZY, Liu XH et al. MicroRNA-203 inhibits cell proliferation by repressing DNp63 expression in human esophageal squamous cell carcinoma. BMC Cancer 2011;

11: 57.40. Georges SA, Biery MC, Kim SY et al. Coordinated regulation of cell cycle transcripts by p53-Inducible microRNAs, miR-192 and miR-215. Cancer Res 2008; 68: 1010510112.

41. Braun CJ, Zhang X, Savelyeva I et al. p53-Responsive micrornas 192 and 215 are capable of inducing cell cycle arrest. Cancer Res 2008; 68: 1009410104.

42. Hino K, Tsuchiya K, Fukao T et al. Inducible expression of microRNA-194 is regulated by HNF-1alpha during intestinal epithelial cell differentiation. RNA 2008; 14: 14331442.

Clinical and Translational Gastroenterology is an open-access journal published by Nature Publishing Group. This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-sa/3.0/

Web End =http://creativecommons.org/licenses/by-nc-sa/3.0/

Clinical and Translational Gastroenterology