Introduction
Esophageal carcinoma is one of the ten common human malignancies. The World Health Organization (WHO) reported in 2008 that there were 482 000 new cases of esophageal cancer and this disease led to 407 000 deaths worldwide each year [
1]. In China, approximately 90% of esophageal cancers are squamous cell carcinomas (ESCCs), predominantly in the male population. The latest data from the Chinese Cancer Registry Annual Report 2011 reveals that the ESCC incidence is 20.85/100 000 and the mortality is 16.24/100 000, ranking the fourth leading cause of deaths from cancer [
2].
It is difficult to diagnose early stage ESCC, because the majority of early-stage esophageal cancers is asymptomatic. At the time of diagnosis, more than 50% of patients have unresectable tumors or radiographically visible metastases, and the overall five-year survival rate is always 10% or so for many years [
3,
4]. Early detection and early diagnosis, which would allow earlier clinical intervention, will contribute to improve the survival and quality of life for ESCC patients. Moreover, significant molecular biological changes likely occur long before the morphological changes of a tumor lesion. Thus it will be important to investigate molecular alterations in tumors and precancerous lesions of the esophagus to identify biomarkers with high sensitivity and specificity for early detection or classification of ESCC.
Current situation of ESCC diagnosis and classification
The development of ESCC correlates with a phenotypic spectrum including esophagitis, basal cell hyperplasia (BCH), squamous dysplasia (atypical hyperplasia), cancer
in situ, and invasive cancer [
5]. Dysplasia is recognized as the precancerous lesion of the esophagus, which is associated with various degrees of cell pleomorphism and structural disturbances without invasion of the lamina propria. It is usually divided into three types: mild dysplasia (md), moderate dysplasia (Md), and serious dysplasia (Sd) or carcinoma
in situ (CIS). The first two types are referred to as low-grade squamous intraepithelial lesion (LSIL), whereas the last one as high-grade squamous intraepithelial lesion (HSIL). All the confirmed serious dysplasia is treated by endoscopic mucosal resection as the same as to carcinoma
insitu. But based on the omnidirectional progression to cancer, regression or stabilization, mild and moderate dysplasia is clinically subjected to only regular follow-up. However, a long-term follow-up study revealed that within 3.5 years of the initial diagnosis 5% of the patients with md, 27% with Md, and 67% with Sd/CIS developed esophageal cancer [
6], and within 13.5 years, to 24%, 50%, and 74%, respectively [
7]. Obviously, there exist over-treatment in patients with serious dysplasia and under-treatment in those with mild and moderate dysplasia to a certain extent. Another gray area of clinical management involves early esophageal cancer, which refers to a tumor that has invaded to the lamina propria or submucosal layer of the epithelial mucosa without lymph node metastasis. It is difficult to judge the depth of malignant lesions and evaluate lymphatic metastasis by preoperative examination. Nevertheless, invasive depth and lymphatic metastasis status are extremely important to determine whether a patient requires surgery or minimally invasive treatment.
Currently, early diagnosis technologies for ESCC include two categories: (1) Imaging technology, such as the esophageal barium swallow test, computerized tomography (CT), and magnetic resonance imaging (MRI), and (2) Endoscopic techniques, such as high-magnification and high-resolution endoscopy, chromoendoscopy (Lugol’s iodine staining, toluidine blue dying, and toluidine blue & Lugol’s iodine double staining), narrow-spectrum imaging endoscopy, autofluorescence imaging endoscopy, confocal laser endoscopy (CLE), optical continuous X-ray tomographic endoscopy, and endoscopic ultrasound (EUS). Esophageal exfoliocytology analysis by non-endoscopic capsule sponge device could be used to the primary screening before endoscopic examination. All these techniques have distinct advantages in their ability to identify precancerous or cancerous lesions, but are difficult to predict the canceration tendency of dysplasia and diagnose lymphatic metastasis for early esophageal cancer. This situation suggests that more accurate methods are needed to supplement the current diagnostic procedures. Research on molecular changes in precancerous lesions and early-stage ESCCs could find effective biomarkers that can be used to identify the risk of patients with dysplasia, to predict the infiltration depth of tumors and to estimate lymph node metastasis and vascular invasion.
Molecular alterations in ESCC and clinical relevance
Studies on the molecular alterations in esophageal lesions are focused on chromosomes, nucleic acids and proteins, including cytogenetic abnormalities, amplifications, deletions, mutations and fusions of genomic DNAs, upregulation, downregulation and variation splicing of RNAs, and expression level, molecular weight, subcellular localization, and post-translational modifications of proteins.
Molecular alterations at the chromosomal/genomic level
By fluorescence
in situ hybridization (M-FISH) with chromosome-specific centromere DNA probes, Yao
et al. analyzed the chromosomal aneuploidies in 124 ESCCs and dysplasia from 113 patients. The authors found copy number gains of chromosomes 3, 8, 10, 12, 17 and 20 in tumors with frequencies of 80.9% (93/115), 81.0% (94/116), 70.5% (79/112), 75.9% (85/112), 68.7% (79/115) and 82.8% (48/58). The aneuploidy of these six chromosomes was also frequently observed in both early ESCCs and precancerous lesions, the rates of which were 80.0% (12/15), 93.8% (15/16), 71.4% (10/14), 64.3% (9/14), 75.0% (12/16) and 63.6% (7/11) in early-stage carcinomas, and 62.5% (5/8), 75% (6/8), 62.5% (5/8), 87.5% (7/8), 87.5% (7/8) and 100% (3/3) in dysplasia, respectively [
8]. In another study, Kang
et al. observed the loss of the Y chromosome in 61.2% of the male ESCC patients. The combination of the four chromosome probes for 3, 8, 10 and 20 detected 74.5% of ESCC, and the combination of 3, 8, 20 and Y detected in 85.0% of tumors occurred in male patients [
9].
With the technique of single nucleotide polymorphism (SNP) arrays, Chattopadhyay
et al. investigated endoscopic biopsy specimens from 20 India patients and showed DNA gains at chromosomal loci 1p36.13-p36.12, 1p21.1, 1q21.1-q44, 2p25.3-p25, 2q14.1-q14.2, 3q28, 3q29, 4p15.2-p15.1, 4q21.23-q21.3, 5p15.2-p12, 5q11.2-q12.1, 6p25.3-q11.1, 7p21.1-p15.3, 9q13-q34.13, 10q21.3, 11p13-p11.2, 11q12.3-q14.1, 11q22.1-q25, 12p13.1-p12.3, 17q21.2-q21.31, 18q11.2, 20p13-p11.21, and losses at 1p36.32-p36.31, 1p36.21-p36.13, 3p26.3-p14.3, 5q32, 6p21.32-p21.2, 6q13-q14.1, 6q23.3-q24.2, 8p23.2-p21.3, 8p21.1-p12, 10p15.3-p11.21, 12q15, 13q12.11-q34, 16q13, 17q21.31-q21.33, 18q21.2 and 22q11.21 [
10]. Shi
et al. performed an analysis of array comparative genomic hybridization (array CGH) on a screening set of 44 tumor tissues from Chinese ESCC patients, and found that the most frequent alterations of copy number were gains of 19p13.3-q13.43, 11q13.1-q13.4, 20p13-q13.33, 3q24-q29, 22q11.21-q12.1, and losses of 4p16.3-q35.2, 13q12.11-q34, 18p11.32-q23 and chromosome Y. High-level amplifications were detected at nine chromosome regions including 3q27.1, 7p11.2, 8q21.11, 8q24.21, 11q13.3, 12q15-q21.1, 14q11.2, 18q11.2 and 19q13.11-q13.12. Two homozygous deletion regions were identified on 4q34.3-q35.1 and 9p21.3 [
11]. Univariate analysis showed that gain of 11q13.2 and loss of 7q34 was each associated with short overall survival (OS). In multivariate analysis, 11q13.2 gain was identified as independent prognostic factor. Copy number increase of CPT1A located on 11q13.2 was confirmed significantly correlated with short OS, and especially identified as an independent prognostic marker in two validation sets (HR, 1.683; 95% CI, 1.133 to 2.500,
P = 0.010; HR, 2.488, 95% CI, 1.235 to 5.013,
P = 0.011).
Microsatellite changes may lead to the alterations of oncogenes and tumor suppressor genes. Liu
et al. analyzed 16 microsatellite markers on nine chromosomal regions in tumors and precancerous tissue resected from 34 ESCC by fluorescently labeled PCR. They showed that the overall frequency of loss of heterozygosity (LOH) significantly increased as the pathological status of the resection specimens changed from low-grade dysplasia (LGD) to high-grade dysplasia (HGD) and ESCC. LOH in eight microsatellite loci (D3S1597, D3S2452, D3S1285, D4S174, D5S2501, D9S125, D13S153, and D17S786) was detected only in HGD and ESCC [
12]. He
et al. detected LOH of 35 microsatellite markers in tumorous and precancerous lesions of the esophagus. The markers with high frequency of LOH in both tumor and dysplasia of the same patient were subjected to further detection in iodine-unstained biopsy samples from the population screening in ESCC high-incidence region. They found common alterations at D3S3644, D3S1768, D3S3040, D3S4542,
RPL14, D9S169, D13S171 and D13S263 in both cancer tissue and precancerous lesions around tumors. Interestingly, LOH of D3S3644, D3S1768, D3S3040, D3S4542,
RPL14 and D13S263 was also observed in iodine-staining abnormal lesions from the population screening and the occurrence rates were increased with pathological severity, suggesting that detection of the above six markers in combination with iodine staining might contribute to the diagnosis for patients on preclinical and preneoplastic phase of the disease and to the prediction for the risk of ESCC development [
13].
Molecular alterations at the protein level
Techniques for detecting protein alterations associated with human cancer include enzyme-linked immunosorbent assay (ELISA), immunohistochemistry (IHC), immunocytochemistry (ICC), radioimmunoassay (RIA), and mass spectrometry (MS). Incorporation of digital histological technology allows IHC to acquire objective and quantitative data on protein detection in tumor tissue.
Lin
et al. summarized 214 proteins with expression changes in ESCC tissue and sera reported from 1977 to 2006. Among them, approximately 20 proteins were altered during the early stages of ESCC, including ALCAM, Cox2, FHIT, GnT-V, MMP7, MMP9, MMP13, p27, p63, Periplakin, RARbeta2, Rras2, Smad6, Smad7, TF-antigen and Tp53. Dysregulated expression of Caveolin-1, CEA, CXCR4, E2F1, Galectin-3, HGF, Laminin-5 gamma 2, MDM2, MMP-26, P63, RhoA and Syndecan-1 was associated with ESCC metastasis, and alteration of Cyclin D1, E-cadherin, TP53 and VEGF predicted a poor outcome [
14]. Recently, more other proteins dysregulated in ESCC have been found. For example, Ishikawa
et al. analyzed 265 ESCC samples and found that antigen 6 complex locus K (LY6K) was significantly overexpressed in tumor tissue when compared with normal epithelium and that LY6K overexpression was associated with poor prognosis of patients (
P = 0.0278) [
15]. Koinuma
et al. investigated 305 ESCC samples and showed that Opa-interacting protein 5 (OIP5) was overexpressed relative to normal tissue, and also significantly associated with poor prognosis of patients (
P = 0.0168) [
16]. Huang
et al. detected 299 ESCC from a high-risk area of China and observed that MCM2 might be a candidate marker for the diagnosis of dysplasia with sensitivity of 91.3% and specificity of 61.8%. They further found that esophageal exfoliocytology examination using a non-endoscopic capsule sponge device combined with the detection of MCM2 expression was able to increase the positive identification rate of ESCC to an extent. The authors therefore proposed that MCM2 immunostaining combined with surface brushing could be useful in screening patients at high risk of cancer in mass surveys [
17]. Abedi-Ardekani
et al. detected 91 non-tumoral esophageal biopsies from patients with biopsy-proven ESCC and 103 gastrointestinal clinic patients with no endoscopic or biopsy evidence of ESCC (control subjects) from a rural region in north-eastern Iran by immunohistochemical staining. They showed that PAH expression was significantly higher in non-tumoral esophageal epithelia from patients with ESCC than from control subjects, suggesting that PAH might be useful for the risk prediction and the prevention of ESCC [
18]. Proteomic analysis of ESCC tissue showed the dysregulated expression of multiple proteins [
19], in which the alterations of calreticulin (CRT), GRP, annexin V, M2-PK, prosaposin (PSAP), plectin 1 (PLEC1) and protein disulfide isomerase A4 (PDIA4) were further confirmed by immunohistochemistry [
20,
21]. In addition, several proteins including human leukocyte antigen (HLA)-G, endothelin (ET), dikkopf-1 (DKK1) and pepsinogens (PGs) have been reported to be altered in serum or plasma samples of ESCC patients as compared to the normal healthy controls. Lower serum PGI/II ratio showed a dose-response association with increased risk of esophageal squamous dysplasia. And elevation of plasma big ET-1 level is an independent prognostic factor for short survival in patients with ESCC (
P = 0.003) [
22-
25].
miRNA alterations
Guo
et al. reported a significant difference in the expression of seven miRNAs between tumors and adjacent normal tissue, suggesting that these miRNAs may be potential biomarkers for ESCC [
26]. Zhang
et al. investigated plasma samples of 290 ESCC and 140 age- and sex-matched controls by using Solexa sequencing technology and quantitative reverse-transcription PCR (RT-qPCR), and derived a panel of seven plasma miRNA markers (miR-10a, miR-22, miR-100, miR-148b, miR-223, miR-133a and miR-127-3p) that can reliably predict ESCC at a relative early stage. The area under the ROC curve for these seven miRNAs ranged from 0.817 to 0.949 [
27]. Recent studies showed that the serum levels of miR-31 and miR-1322 were expressed significantly higher in ESCC tissue and serum samples than in healthy controls with the ROC AUC of 0.902 and 0.847, and patients with high-levels of serum miR-31 had a poorer prognosis in relapse-free survival (
P = 0.001) and tumor-specific survival (
P = 0.005), whereas the levels of serum miR-1322 were positively correlated with TNM staging (
P<0.001) [
28,
29]. There are currently not yet studies investigating alterations in miRNA expression in precancerous lesions and early ESCC.
Epigenetic alterations
Zheng
et al. analyzed serum DNA samples from 70 patients with ESCC and found that runt-related transcription factor 3 (
RUNX3) hypermethylation was associated with advanced tumor stage and lymph node metastasis [
30]. Lima
et al. presented another investigation that the promoter methylation in Trefoil factors family 1 (
TFF1) promoter methylation may be an early change during esophageal carcinogenesis by comparing 106 ESCC cases and 27 controls [
31]. Kaz
et al. documented that 24 genes harbored abnormal levels of methylation in ESCC, in which the alterations of
CDKN2A,
MGMT,
HIN-1 MLH1,
RARb2,
CDH1,
DAPK and
FHIT might be potential biomarkers for the precancerous risk assessment and early diagnosis of the disease [
32]. Other epigenetic alterations including histone modifications and methylation may also have potential clinical implications [
33]. Although the promoter methylation in secreted frizzled-related protein 2 (
SFRP2) and Wnt inhibitory factor-1 (
WIF1) was observed in ESCC tissue and cell lines, the clinical relevance was however indistinct [
34,
35].
Genetic polymorphisms and susceptibility of ESCC
Although environmental factors represent one of the primary causes of ESCC, relatively few individuals are susceptible to ESCC even when exposed to a similar environment, suggesting that individual susceptibility factors may play an important role in the development of ESCC. Lin
et al. identified genetic polymorphisms in multiple genes correlated with susceptibility to ESCC, such as
P53,
MDM2,
CASP8,
12-LOX,
COX-2 and
S100A14 [
36-
40]. Another study of 9053 ESCCs and 13 283 controls from Chinese Han and Uygur-Kazakh revealed that genetic polymorphisms of
PLCE1 and
C20orf54 could have biological implications for ESCC in these populations [
41]. Genetically susceptible individuals present a relatively higher probability of developing cancer. Research on genetic polymorphisms in ESCC will be important for the identification of populations that are susceptible to ESCC.
Conclusions and future directions
A large amount of data has accumulated regarding the possible development of biomarkers for the molecular classification, early diagnosis of ESCC and the risk assessment of precancerous lesions. Early molecular alterations in ESCC in esophageal squamous cell carcinomas are summarized in Table 1, potential markers for diagnosis markers in ESCC in Table 2, and those associated with prognosis of ESCC patients are summarized in Tables 3 and 4.
It is worthwhile pointing out that none of molecular biomarkers currently can be used in routine clinical practice to determine prognosis or prediction, despite growing knowledge of the biology of ESCC. There might exist more or less discrepancies between the above different investigations concerning experiment conditions, detection reagent resources, populations tested, etc. Future studies require better standardization for experiments, larger samples, multicenter verification, prospective observations.
Compliance with ethics guidelines
Li Shang and Mingrong Wang declare that they have no conflict of interest. This manuscript is a review article and does not involve a research protocol requiring approval by the relevant institutional review board or ethics committee.
Higher Education Press and Springer-Verlag Berlin Heidelberg
PDF
(164KB)
