Squamous cell carcinoma antigen (SCCA) is up-regulated during Barrett’s carcinogenesis and predicts esophageal adenocarcinoma resistance to neoadjuvant chemotherapy

Squamous Cell Carcinoma Antigen (SCCA) is consistently overexpressed in many different solid tumors, and has been associated with both tumor aggressiveness and chemoresistance. No data, however, is currently available on SCCA expression during esophageal Barrett's carcinogenesis, nor on SCCA expression's role on esophageal adenocarcinoma chemoresistance. The SCCA immunohistochemical expression was assessed in a series of 100 biopsy samples covering the whole histological spectrum of Barrett's oncogenesis. Squamous native mucosa was characterized by a moderate to strong cytoplasmic and nuclear SCCA expression in suprabasal, medium, and superficial layers. On the other hand, almost half of the considered lesions did not express SCCA; the other half featured weak to moderate SCCA expression. The relationship between SCCA protein expression and tumor response to neoadjuvant chemotherapy was assessed in 90 esophageal adenocarcinoma specimens (40 biopsy and 50 surgery specimens), stratified according to Mandard tumor regression grade. As observed in other settings, the presence of SCCA expression clustered in the group of tumors characterized by a lower responsiveness to neoadjuvant treatments. The present results suggest an involvement of SCCA in a subset of Barrett-related tumors, and prompt to consider the SCCA-protein expression as response-predictive marker of neoadjuvant therapy in esophageal adenocarcinomas.


INTRODUCTION
Barrett's esophagus is a complication of longstanding gastro-esophageal reflux and is a major risk for esophageal adenocarcinoma, one of the leading causes of cancer-related death worldwide [1][2][3][4].
The recent advances in the therapeutic approaches for gastroesophageal tumors has significantly improved Research Paper www.impactjournals.com/oncotarget the curative resection rates, and both the disease-free and the overall survival [5][6][7]. Nevertheless, no biomarker is available to reliably predict tumor chemosensitivity.
SCCA-1 confers to cancer cells resistance to induced apoptosis by different mechanisms, including inhibition of lysosomal cathepsin proteases, JNK pathway or p38 activation [8]. More recently, it has been demonstrated the role of SCCA-1 in inhibiting reactive oxygen species (ROS) generation and cell death through its inhibitory interaction with respiratory Complex I [14]. SCCA expression has never been investigated in the different histology phenotypes of the multistep Barrett's carcinogenetic process. We first analyzed SCCA expression by immunohistochemistry on a series of Barrett-related lesions. We also focused on the relationship between SCCA expression and tumor response to neoadjuvant chemotherapy, as assessed by Mandard scale Tumor Regression Grade (TRG) [17].
By exploring the NCBI-GEO database, we analyzed two further independent studies, which both showed a significant down-regulation of SCCA-1 expression in BM versus N (p=0.016 for matched samples and p<0.001, respectively) [18,19] ( Figure 1B Figure 1C). A similar significant downregulation of SCCA-2 in BM versus N was observed in both studies available at the NCBI GEO database and previously characterized for SCCA-1 expression (p=0.010 and p<0.001, respectively) ( Figure 1D). Normal esophageal mucosa showed moderate to strong cytoplasmic and nuclear SCCA expression in suprabasal, medium, and superficial layers (i.e., spinous and granular layers), whereas basal layers were SCCA negative ( Figure 2). As observed in the array database meta-analysis for both SCCA isoforms, SCCA expression was significantly down-regulated in the Barrett-related lesions (p=ns). However, a comparable SCCA expression prevalence was observed among groups, with around 40-60% of samples showing a 1+ SCCA expression. Most of these samples expressing SCCA showed a faint to moderate cytoplasmic immunoreaction. Two BM samples showed a strong 2+ cytoplasmic immunoreaction. No significant relationship was observed among SCCA expression and the intensity of the inflammatory infiltrate in BM and neoplastic lesions.

SCCA-expressing tumors are less sensitive to neoadjuvant chemotherapy
Since SCCA-1 overexpression has been associated to a worst prognosis and chemoresistance in several different tumor types [8,[14][15][16]24], we assessed SCCA expression in a series of BAc specimens selected according to their postneoadjuvant chemotherapy Mandard status ( Figure 3A).
The presence of SCCA expression (both 1+ and 2+) as both assessed in pre-neoadjuvant therapy endoscopy biopsies (n=40; 20 TRG1/2 and 20 TRG5) or post-therapy surgical specimens (n=50; 25 TRG2 and 25 TRG5) was significantly associated to a lower sensitivity to neoadjuvant treatments (both p<0.001). Of note, all the observed 2+ cases clustered in the Tumor Regression Grade 5 (TRG5) group. Two out of five 2+ SCCA cases showed a strong cytoplasmic (and focally nuclear) heterogeneous reaction ( Figure 3B).

DISCUSSION
In Western countries, the rising incidence of esophageal adenocarcinoma is associated with an overall 5-year survival rate lower than 20% [1,25]. Currently, neoadjuvant chemo(radio)therapy followed by surgical resection offers the best clinical outcome for locally advanced tumors, with around 25% of complete response rate [5,[26][27][28]. Moreover, the pathologic complete response is associated with lower recurrence, lower metastatic potential, and longer overall survival [29]. No reliable prognostic biomarker, however, is currently available to identify those tumors that may benefit of neoadjuvant therapeutic protocols.
This study focuses on SCCA expression in Barrett's carcinogenesis. Under normal conditions, SCCA expression is reliably expressed in both the spinous and the granular layers of normal squamous epithelium (tongue, esophagus, tonsil, cervix uterine, vagina, Hassal's corpuscles of the thymus and some areas of the skin) [30,31]. As expected, also in the present study, native esophageal mucosa consistently featured SCCA expression.
In both microarray data metanalysis and immunohistochemistry, we observed a significant reduction in SCCA expression in Barrett's related lesions in comparison to normal mucosa. Of note, however, a significant proportion of the considered cases consistently retained a weak to moderate protein expression, with no significant phenotypic differences among SCCA-positive and SCCA-negative cases.
A similar SCCA downregulation has been demonstrated in the esophageal squamous cell carcinoma (ESCC) setting in both invasive and intraepithelial lesions by proteomic analysis [32]. These data further emphasize the importance of SCCA expression maintenance in a subset of esophageal tumors. On the other hand, a significant overexpression in circulating SCCA-2 mRNA levels were observed among ESCC patients [33], but the cell source (i.e., SCCA-expressing tumors and/or normal www.impactjournals.com/oncotarget esophageal mucosa damaged by neoplastic cells) should be further explored.
Increased SCCA-1 expression has been recently associated to inhibition of protein turnover, unfolded protein response, and activation of NF-kB signaling [34]. Of note, NF-κB pathway has been associated with BE-related cancer progression, being up-regulated in BM and BAc samples [35]. In this setting, it regulates the expression of many pro-inflammatory and growth regulatory cytokines including tumour necrosis factor-α(TNF-α). Confirming these data, TNF-α expression has been shown to increase along the metaplasia-dysplasiacarcinoma sequence, leading to an increase in the protooncogene MYC via a β-catenin mediated pathway [36]. A SCCA-1-mediated (both direct and indirect) MYC overexpression has recently been demonstrated in hepatocellular carcinoma [37]. Overall, these data suggest a possible role for SCCA-1 in the neoplastic transformation of a distinct sub-group of Barrett's mucosa.  (Original magnifications 4x, 10x, and 20x). www.impactjournals.com/oncotarget SCCA-1 is highly expressed in the spinous and granular layers of normal skin, which is histologically similar to what observed in the esophageal mucosa. Katagiri and colleagues recently described that SCCA-1 levels are significantly up-regulated in lesional skin compared to their normal sun-protected skin [38]. Furthermore, subjects with high levels of SCCA-1 in the epidermis were more susceptible to barrier disruption by external stimuli, and this was accompanied with a further subsequent increase of SCCA-1. This model could be translated into esophageal pathophysiology: the gastroesophageal reflux will lead to a chronic inflammatory insult to the esophageal mucosa with a concurrent SCCA-1 up-regulation and disruption of the esophageal mucosa barrier to external stimuli, with an accelerated process of mucosa metaplastic transformation. On the other hand, metaplastic epithelia are SCCA-1 negative. As observed in liver carcinogenesis, under longstanding inflammation the metaplastic Barrett-epithelia would acquire a progressive increase in SCCA-1 expression, which would sustain the oncogenic process within the Barrett's mucosa [39].
Recent evidence linked SCCA-1 overexpression to chemoresistance in several different tumor types [8,[14][15][16]24]. Based on such features, in a series of BAc cases we explored the relationship between SCCA expression and the tumor response to neoadjuvant chemotherapy (as assessed by Mandard score). Of interest, most of tumors with a SCCA expression clustered in the TRG5 group, demonstrating a significant association between the SCCA expression and a lower responsiveness to neoadjuvant treatments. Similar data have been observed in epithelial ovarian cancers exposed to platinum-based chemotherapy regimens [16,40] and in breast carcinomas treated with anthracycline-based neoadjuvant chemotherapy [15]. In addition, it has recently been demonstrated that SCCA-1 protects from oxidative damage by chemotherapeutics through inhibition of mitochondrial respiratory complex I [14].
In conclusion, this study originally demonstrated a significant downregulation of the SCCA expression in the different phenotypic lesions included in the Barrett's carcinogenesis. In esophageal adenocarcinoma, the significant association between SCCA-expression and a lower responsiveness to neoadjuvant chemotherapy, potentially qualifies the protein expression as a prognosticpredictive marker to select patients and define more effective and specific interventional approaches.

Ethics statement
All the histopathological samples were retrospectively collected (2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015) from the files of the Surgical Pathology Unit of the Department of Medicine (DIMED) at the University of Padua and from the Veneto Region's multicenter Barrett Esophagus Registry (EBRA, Padua Unit) [41]. Investigation has been conducted in accordance with the ethical standards, according to national and international guidelines, and has been approved by the authors' institutional review board (418/04/CE). To test SCCA expression predictive impact on platinum-based neoadjuvant chemotherapy, a series of 90 esophageal adenocarcinoma specimens stratified according to Mandard tumor regression grade (TRG) was considered [17]. Because of the high intra-pathologists variability in assessing Mandard score [38], only tumors characterized by TRG1/2 (i.e., TRG1= tumor complete regression; TRG2= fibrosis with scattered tumor cells) and TRG5 (i.e., TRG5= tumor without changes of regression) were selected. In particular two different series were analyzed: i) forty pre-therapy BAc biopsy samples selected according to the final TRG score on the surgically treated adenocarcinoma (20 TRG1/2 and 20 TRG 5); ii) fifty post-neadjuvant therapy surgicallytreated adenocarcinomas (25 TRG2 and 25 TRG 5). In the latter group, only TRG2 cases were considered because TRG1 esophagectomies lack by definition tumor cells to be analyzed.

Array database meta-analysis
The Oncomine database and gene microarray analysis tool, a repository for published microarray data (www.oncomine.org) [42], was explored (15 May 2016) for SCCA-1and SCCA-2 mRNA expression in esophageal native mucosa, BM and BAc samples. Oncomine algorithms, which enable multiple comparisons among different studies, were used for the statistical analysis of the differences in mRNA expression between the aforementioned comparisons.
The NCBI-GEO repository of published array data (http://www.ncbi.nlm.nih.gov/geo/) and the GEO2R microarray analysis tool were explored (15 May 2016) to assess SCCA-1and SCCA-2 expression in Barrett-related lesions (using the keywords: SCCA-1 or serpinB3, and esophagus). Studies already considered in the Oncomine analysis were excluded from the NCBI-GEO analysis.

Statistical analysis
Differences between groups were tested by applying the (paired) t-test, the modified Kruskal-Wallis nonparametric test for trend, and the Mann Whitney test, as appropriate. P values <0.05 were considered significant. The statistical analysis was performed using STATA software (Stata Corporation, College Station, TX).