Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Oncogene. Author manuscript; available in PMC 2013 April 18.
Published in final edited form as:
PMCID: PMC3310953

Dietary zinc deficiency fuels esophageal cancer development by inducing a distinct inflammatory signature


Chronic inflammation is implicated in the pathogenesis of esophageal squamous cell cancer (ESCC). The causes of inflammation in ESCC, however, are undefined. Dietary zinc-deficiency (ZD) increases the risk of ESCC. We have previously shown that short-term ZD (6 weeks) in rats induces overexpression of the proinflammatory mediators S100a8 and S100a9 in the esophageal mucosa with accompanying esophageal epithelial hyperplasia. Here we report that prolonged ZD (21 weeks) in rats amplified this inflammation that when combined with non-carcinogenic low doses of the environmental carcinogen N-nitrosomethylbenzylamine (NMBA) elicited a 66.7% (16/24) incidence of ESCC. With zinc-sufficiency NMBA produced no cancers (0/21) (P<0.001). At tumor endpoint, the neoplastic ZD esophagus as compared with zinc-sufficient esophagus had an inflammatory gene signature with upregulation of numerous cancer-related inflammation genes (CXC and CC chemokines, chemokine receptors, cytokines, and Cox-2) in addition to S100a8 and S100a9. This signature was already activated in the earlier dysplastic stage. Additionally, time-course bioinformatics analysis of expression profiles at tumor endpoint and prior to NMBA exposure revealed that this sustained inflammation was due to ZD rather than carcinogen exposure. Importantly, zinc replenishment reversed this inflammatory signature at both the dysplastic and neoplastic stages of ESCC development, and prevented cancer formation. Thus, the molecular definition of ZD-induced inflammation as a critical factor in ESCC development has important clinical implications with regard to development and prevention of this deadly disease.


Esophageal cancer is the eighth most common cancer worldwide (with 482,000 new cases) and the sixth most common cause of death from cancer (with 407,000 deaths estimated in 2008) (1). Esophageal squamous cell carcinoma (ESCC) is the predominant subtype. Because of an absence of early symptoms, ESCC is frequently diagnosed at an advanced stage of the disease. Although cancer mortality rates have declined worldwide since the mid 1980s, ESCC remains a deadly cancer with a 5-year survival of only 10%. Thus, insights into its pathogenesis are critical in order to devise new preventions, earlier diagnostics, and novel treatment strategies.

Epidemiological and clinical studies show that chronic inflammation predisposes to many types of cancer, including ESCC (2-4). In this regard, persistent inflammation in the esophagus is a frequent occurrence in populations at high risk for ESCC (5). The challenge is to determine the cause of this chronic inflammation and its role in ESCC development.

The major risk factors for ESCC are chronic alcohol consumption, tobacco use, nutritional deficiencies, and exposure to environmental carcinogens such as N-nitrosomethylbenzylamine (NMBA) (6, 7). In particular, zinc (Zn)-deficiency (ZD) is implicated in the pathogenesis of ESCC in many populations (8, 9), including persons with chronic alcohol consumption (10). Abnet et al. (8) provided the strongest evidence of an association between dietary ZD and ESCC in a high incidence area by establishing an inverse relationship between Zn concentration in biopsy samples and the subsequent risk of developing ESCC. Although Zn is present in a large variety of foods, its content is low in most, with the exception of red meat and seafood. Accordingly, individuals subsisting largely on a cereal diet are likely to be Zn-deficient. In the US 10% of the population is estimated to ingest less than 50% of the recommended daily allowance for Zn (11). In the developing world, dietary ZD may affect more than 2 billion people (12). Because Zn is required for the activity of many enzymes, for proper immune function, and for the conformation of many transcription factors that control cell proliferation, apoptosis, and signaling pathways (13-15), ZD predisposes to disease by adversely affecting these processes.

The environmental carcinogen NMBA is widely used to induce esophageal tumors in rodents. An early study reported that NMBA induces a 63% incidence of ESCC in nutritionally complete rats after a cumulative dose of 50 mg/kg body weight (2.5 mg/kg weekly for 20 weeks) (16). A typical esophageal tumor bioassay in chemoprevention studies entails weekly administration of low doses of NMBA for 15 weeks (cumulative dose = 7.5 mg/kg), producing a 100% incidence of squamous papillomas but without ESCC development (17). The tumorigenicity of NMBA in rodents is the results of the bioactivation of the carcinogen by esophageal cytochrome P450 isozymes to produce a DNA-methylating agent, leading to the formation of the mutagenic DNA adduct O6-methylguanine (18), and a high prevalence of mutations in Ha-Ras and p53 genes in papillomas (19).

Our ZD rat esophageal tumor model that combines dietary ZD with exposure to the environmental carcinogen NMBA (20) mimics aspects of human ESCC in high-risk populations (6, 8, 9). We have shown in prior studies that weanling rats on a ZD diet for ~6 weeks develop increased cell proliferation and changes in gene expression in the squamous epithelium, including overexpression of Cox-2 (21). The hyperplastic ZD rat esophagus is highly sensitive to NMBA (20), displaying a 93% incidence of esophageal papillomas (but no ESCC) after a single exposure to a low NMBA dose (2 mg/kg). Whether dietary ZD could promote ESCC development in tumor bioassays by low NMBA doses has not been determined.

DNA microarray analysis using genome arrays provides a powerful tool to understand how ZD might affect gene expression to predispose to esophageal carcinogenesis. Using a rat genome array, we reported that short-term ZD (6 weeks) alone (without NMBA) induces a distinct gene signature in the esophageal mucosa with upregulation of two proinflammation-genes S100 calcium binding protein a8 and a9 (S100a8/a9), associated with esophageal epithelial hyperplasia. Zn was shown to modulate the interaction of S100A8 with its receptor for advanced glycation end products (RAGE) and the downstream nuclear factor (NF)-kB-COX-2 signaling pathway (22).

We now ask whether prolonged ZD might amplify the inflammatory program to provide a microenvironment conducive to ESCC development on exposure to low carcinogen doses, thus serving as a faithful model for conditions that lead to human ESCC. For this, we conducted a long-term tumor bioassay in Zn-modulated rats using low doses of NMBA. In parallel, we performed transcriptome profiling of esophageal mucosa at tumor endpoint and during tumor development in order to correlate gene expression changes with tumor progression.


Prolonged ZD combined with low doses of NMBA elicits ESCC

To determine if exposure to non-carcinogenic low doses of NMBA (17) in a ZD environment leads to ESCC, a tumor study in Zn-modulated rats was performed. As shown in the experimental plan (Figure 1a), ZD, ZR (Zn-replenished), and ZS (Zn-sufficient) rats were given 3 intragastric doses of NMBA (2 mg/kg at 6 weeks intervals, cumulative dose = 6 mg/kg). At 5 weeks after the first dose and at tumor endpoint (15 weeks post first dose), esophagi were evaluated for tumor incidence and the mucosa-derived RNA for gene expression profiling. We found that the ZD esophageal epithelium was already dysplastic, whereas the ZS and ZR esophagus remained non-proliferative at 5 weeks after the first NMBA dose (Figure 1b). At endpoint ZD rats showed significantly higher tumor incidence and multiplicity (papillomas and carcinomas) than ZS rats (Figure 1c). Histological examination revealed that 66.7% (16/24) of ZD rats harbored ESCC with prominent inflammatory cell infiltrates in the stroma, whereas none of the ZS rats (0/21) or ZR rats (0/23) developed a cancer (Figures 1b and c, P < 0.001). In fact, all tumors in ZS and ZR rats were only papillomas. Thus, ZR prevented ESCC development. These data establish for the first time that in the presence of prolonged dietary deficiency of Zn, non-carcinogenic doses of NMBA can induce ESCC.

Figure 1
Zn-deficiency induces distinct gene expression patterns in Zn-modulated rat esophagus in the dysplastic and neoplastic stages of ESCC development. (a) Study design: weanling rats were fed ZD and ZS diets for 6 weeks. Then, the rats were given 3 intragastric ...

A distinct inflammatory signature accompanies ESCC development

To identify genes critical in ESCC development, we performed comparative transcriptomics analysis of esophageal mucosa from zinc-modulated rats (n = 4 rats/group/time point) at the early dysplastic (5 weeks) and later neoplastic stages (endpoint) using Affymetrix Rat 230 2.0 Genome GeneChip. Hierarchical clustering analyses of 30,000 transcripts (Figure 1d) revealed that the ZD neoplastic and dysplastic esophagus had distinct gene expression patterns when compared with their ZS counterparts. ZR led to a global reversal of expression patterns at both time points (Figure 1d) accompanied by reversal of the dysplastic and neoplastic phenotype (Figure 1b).

Using a cutoff point of q-value < 5% and ≥ 2-fold difference, we identified 751 upregulated and 686 downregulated transcripts in the neoplastic ZD esophagus (Dataset 1a). Many of the upregulated genes were hallmarks of cancer-related inflammation (chemokines, cytokines, prostaglandins, and S100a8/a9) implicated in the development of human cancers (23, 24). The list of inflammation genes (Table 1) includes CXC and CC chemokines that are chemoattractants for neutrophils and macrophages (Cxcl5, up 84-fold; Cxcl2, up 38-fold; Cxcl3, up 9-fold; and Cxcl1, up 7-fold, Ccl2 [also known as Mcp-1], up 2.4-fold); CXC and CC chemokine receptors Ccr1 (up 7-fold), Ccr2 (up 4-fold), and Cxcr4 (up 6-fold), an important molecule in cancer progression/metastasis that is overexpressed in most cancers (25); prostaglandin-endoperoxide synthase 2 (Ptgs2, also known as Cox-2, up 41-fold), commonly overexpressed in human cancers and involved in mediating inflammation through the synthesis of prostaglandins (26); interleukin-1b Il1b (up 12-fold), a key inflammatory cytokine (27); nuclear factor-kb1 (Nfkb1, up 2-fold) that is a link between inflammation and cancer development/progression (28); hypoxia inducible factor 1, alpha (Hif1a, up 2-fold) that is responsible for the majority of Hif-1-induced gene expression changes with hypoxia and closely associated with inflammation (29), and finally the proinflammatory mediators S100a8 and S100a9 (both up 2-fold) that play a key role in inflammation-associated cancers (24) and are induced by short-term ZD in esophagus (22). Almost all of these inflammation genes were already expressed in dysplastic ZD esophagus (5 weeks) (Table 1, Dataset 1b), albeit at lower levels than at tumor endpoint. These data showed that prolonged ZD (21 weeks) and 3 low doses led to an amplified inflammatory gene repertoire in the esophagus, with overexpression of numerous cancer-related inflammation genes in addition to the proinflammation mediators S100a8 and S100a9 that were induced by short-term ZD (6 weeks) (22). This amplified inflammatory program accompanied ESCC development. Importantly, ZR reversed the inflammatory gene signature at both the dysplastic and neoplastic stages of ESCC development (Dataset 2), and blocked cancer development (Figure 1c).

Table 1
Summary of upregulated inflammation genes in ZD esophagus during progression to ESCC

Validation of array data

To validate the microarray results, we performed quantitative real-time RT-PCR analysis on 15 inflammation genes (Cxcl5, Cxcl3, Cxcl2, Cxcl1, Ptgs2, Il1b, Cxcr4, Ccl2, Ccr1, Ccr2, Hif1a, Tlr2, Tlr4, S100a8, and S100a9) at endpoint and 10 genes (Cxcl5, Cxcl3, Cxcl2, Cxcl1, Ptgs2, Il1b, Cxcr4, Hif1a, S100a8, and S100a9) at 5 weeks. The qRT-PCR data confirmed that these inflammation genes were upregulated by ZD and restored to normal levels by ZR (Figure 2a). Additionally, enzyme-linked immunosorbent assay (ELISA) and immunoblot (Figure 2b) analyses showed that the Zn-modulating effects observed at the transcript levels for Il1b, Cxcl5, NF-kB p65, Cox-2, S100a8, S100a9, Cxcl2, and Hif1a were also reflected at the protein level.

Figure 2
Zn-deficiency induces an inflammatory gene signature. (a) Validation of microarray data by qRT-PCR. (b) ELISA assay (IL1B, CXCL5, NF-κB p65; n = 3 rats/group); and Immunoblot (COX-2, S100A8, S100A9, CXCL2, HIF1A, n = 3 rats/group). (c) IPA uncovers ...

To investigate the temporal and spatial localization of key inflammation markers NF-kB p65, COX-2, S100A8, and S100A9 in ZD esophagus during cancer development, immunohistochemistry (IHC) was performed (Figure 3a, endpoint; Figure S1, 5 weeks). Both the ZD dysplastic esophagus (5 weeks) and the ESCC (endpoint) showed high proliferative activity with abundant PCNA (an endogenous cell proliferation marker)-positive nuclei and strong cytoplasmic staining of the squamous cell tumor marker KRT14 (30) and strong cytoplasmic staining for all 4 inflammation markers. By contrast, both the ZR and ZS esophagus showed only basal cell proliferation, with weak and diffuse immunostaining of the same inflammation markers. These data demonstrated that the ZD esophagus acquires a highly proliferative and inflammatory phenotype during ESCC development that is reversed by ZR.

Figure 3
Immunohistochemical analysis of cell proliferation and inflammation markers in ESCC development. (a) ZD esophagus at 5 weeks and at tumor endpoint. Also shown are the ZR and ZS esophagus at tumor endpoint (ZR and ZS data at 5 weeks are shown in Figure ...

Chemotaxis is a key biological process among the upregulated genes

We next performed gene ontology analyses using DAVID (Database for Annotation, Visualization and Integrated Discovery) resources (31) to define the biological significance of the genes differentially expressed in the ZD esophagus. At both tumor endpoint and at the earlier dysplastic stage (5 weeks) significantly over-represented biological processes were found only among the upregulated genes (Table 2). In particular, the involved biological processes included “chemotaxis” at the tumor endpoint with 8 inflammation genes (Cxcl3, Ccr1, Cxcl2, Cxcl1, S100a8, Cxcl5, S100a9, and Il1b, P=8.34E-08) and also at the early dysplastic stage with 4 inflammation genes (Cxcl2, S100a8, Cxcl5, S100a9). This finding is consistent with the concept that chemotaxis in the tumor microenvironment is an essential component during progression to malignancy (32). Thus, the DAVID data support the conclusion that these inflammatory chemokines (Cxcl5, Cxcl3, Cxcl2, Cxcl1), cytokine (Il1b), and S100a8/a9 are relevant to the ZD-driven development of ESCC.

Table 2
Overrepresented biological processes in ZD esophagus at tumor endpoint and at the early dysplastic stage

IL1B is an anchor in an inflammation cancer network

To understand the interactions of the differentially expressed genes in neoplastic ZD epithelia, we performed an Ingenuity Pathway Analysis (IPA). We identified an inflammatory cancer network of 35 genes with Il1b as an anchor (Figure 2c). Fifty-seven percent of the genes (20 of 35) were upregulated, including Ptgs2 (Cox-2), Cxcl5, Cxcl2, Cxcl1, S100a8, and S100a9. Il1b has direct connectivity to Ptgs2 and indirect connectivity to S100a8, S100a9, Cxcl5, Cxcl2, Cxcl1, and Cxcl3). Because Il1b plays a central role in inflammation (33) and the progression of gastric inflammation and cancer in transgenic mice overexpressing human IL1B (34), our data that Il1b showed connectivity to these upregulated inflammation mediators predict interactions among these genes in pathways that promote ESCC development.

NMBA tumorigenicity does not induce inflammation

To determine whether the tumorigenic effect of NMBA in ZS animals that induced only squamous papillomas (Figure 1c) was accompanied by inflammation, we performed longitudinal (time-course) gene expression comparisons in ZS and ZD rats. We first compared the expression profile of ZD esophagus at tumor endpoint vs our previously published profile of ZD esophagus at 6 weeks of diet without NMBA (22) (Dataset 3a). This result was then compared with the profile obtained from the ZS esophagus at endpoint vs the published profile of ZS esophagus at 6 weeks of diet without NMBA (22) (Dataset 3b). Using a cutoff point of q-value <5% and ≥4-fold difference, our Venn diagram (Figure 1e) showed that ZD esophagus had 171 up- and 51 downregulated genes, ZS esophagus had 62 up- and 9 downregulated genes, ZD and ZS esophagus had 23 common upregulated but no common downregulated genes. While 16 of the 171 upregulated genes (9.3%) in ZD esophagus were inflammation genes as listed in Table 1, none of the 62 upregulated genes in ZS esophagus were inflammation genes (Supplementary table 1). In addition, our DAVID bioinformatics (31) showed that significantly over-represented processes were only found among the upregulated genes in ZD esophagus but not in ZS esophagus. In particular, the most significantly upregulated biological process in ZD esophagus was “defense response” with 34 genes that included many inflammation genes (Cxcl1, Ptgs2, Ccr1, Cxcl2, Tlr2, Cxcl5, Il17f, Il1b, Ccr5, Ccr2, P-value = 1.94E-22) (Table 3). Thus, time-course analysis revealed that NMBA tumorigenicity that induced only papillomas in ZS rats did not cause inflammation, whereas ZD with NMBA led to overexpression of numerous inflammation genes and thus, to cancers.

Table 3
Overrepresented biological processes in ZD and ZS esophagus at tumor endpoint as compared with their counterpart prior to NMBA dosing

Long-term ZD alone causes overexpression of numerous inflammation genes

To determine whether long-term ZD (23 weeks) without NMBA could amplify the inflammation that was induced by short-term ZD (6 weeks) (22), we used qRT-PCR to analyze the expression of 15 selected inflammation genes (discovered by microarray, Table 1). At 23 weeks, serum Zn levels were significantly lower in ZD (42 μg/100 ml, 95% CI = 39 to 45 mg/100 ml) than ZS rats (88 μg/100 ml, 95% CI = 85 to 91 mg/100 ml) (P < 0.0001). We found that long-term ZD resulted in an amplified inflammatory gene repertoire with upregulation of 12 inflammation genes (Cxcl5, Cxcl3, Cxcl2, Cxcr4, Il1b, Ptgs2, Ccr2, and Ccl2, S100a8, S100a9, Hif1a, and Tlr4) (Figure 2d), as compared with only 4 genes (S100a8, S100a9, Hif1a, and Tlr4) in short-term ZD esophagus (22). The expression level, however, was not as strong as in neoplastic ZD esophagus (Figure 2e). Immunoblot (Figure 2f) and IHC (Figure 3b) analyses showed that long-term ZD induced overexpression of COX-2, NF-kB p65, S100A8, S100A9, NF-kB p65, and HIF1A protein in the highly proliferative esophageal epithelium that showed intense and abundant immunostaining of PCNA and KRT14 protein. Additionally, chronic inflammation was characterized by the frequent presence of inflammatory cells in the stroma (Figure 3b; H&E). These data showed that prolonged ZD per se induced a highly inflammatory microenvironment that was conducive to ESCC development upon exposure to NMBA.


Chronic inflammation is central to the pathogenesis of many human cancers, including ESCC. The etiologic agents of chronic inflammation associated with cancers are infectious agents (virus and bacteria), chemical and physical agents such as silica particles, asbestos fibers, cigarette smoke, carcinogens, ultraviolet light, gastric acids, and urinary catheters (2, 3). While microbial infection is implicated in the pathogenesis of ~15% of human cancers, the causes of chronic inflammation in most cancers including ESCC are poorly defined. Previously, we have shown that short-term ZD (6 weeks) that causes esophageal epithelial hyperplasia is accompanied by overexpression of 2 proinflammation mediators S100a8/S100a9 (22). The present data establish that prolonged ZD (21 weeks) amplifies this inflammation in the esophagus by causing overexpression of numerous inflammation genes in addition to S100a8/S100a9, thereby providing an inflammatory microenvironment conducive to cancer development. Indeed, in the presence of ZD-induced chronic inflammation, non-carcinogenic low doses of NMBA elicited a 66.7% (16/24) incidence of ESCC. Conversely, with zinc-sufficiency NMBA produced no cancers (0/21). Importantly, our bioinformatics data confirmed that this sustained inflammation was due to ZD rather than carcinogen exposure (Table 3, Supplementary table 1).

The transcriptomics analyses showed that at tumor endpoint the neoplastic ZD esophagus had an inflammatory gene signature that was already activated in the dysplastic stage of ESCC development, with upregulation of CXC and CC chemokines and receptors (e.g. Cxcl5, -2, -3, -1, Ccl-2, Cxcr4) as the most abundant category, as well as Il1b, Cox-2, S100a8/a9, and transcription factors Nfkb1 and Hifla (Table 1). Importantly, these same inflammatory mediators CXCL5, CXCL2, CXCR4, IL1B, COX-2, S100A8, and S100A9 are all overexpressed in human ESCC (35-38). Initially defined as factors that recruit leukocytes to tumors, the complex network of chemokines and their receptors have emerged as key components of cancer-related inflammation that play critical roles in tumor biology by promoting tumor cell survival, proliferation, and metastasis (23, 28, 39). In this regard,, downregulation of CXCL5 by RNA interference was reported to inhibit tongue SCC development, indicating that CXCL5 plays an important role in squamous cell carcinogenesis (40).

The transcription of many cancer-associated chemokine genes (CXCL1, -2, -3, -5, -8, -9, 10, 12, and CCL-2) (41), COX-2 (42), and S100A8/A9 (43) are modulated by the NF-kB family of transcription factor. Expression of NF-kB, in turn, is induced by IL1B (44). Because IL1B plays a central role in inflammation (33) and NF-kB is a link between inflammation and cancer (28), our IPA finding that predicts a cancer network among the upregulated inflammation genes with IL1B as a central anchor strongly suggests that these ZD-induced inflammation genes interact in cancer pathways that promote ESCC development (Figure 2c).

Perhaps most significantly, the present study shows that replenishing Zn after carcinogenic exposure reverses the inflammatory gene signature and prevents ESCC development. Recently, we reported that Zn supplementation in nutritionally complete rodents also inhibits forestomach and tongue carcinogenesis through attenuation of inflammation (45, 46). It is likely that the effects of Zn represent a continuum in its action in attenuating inflammation when Zn is deficient, as well as when it is sufficient. Thus, oral Zn replenishment or supplementation in individuals at high risk for ESCC may reduce the risk of developing this deadly disease in humans.

In conclusion, the present data detail a new model of ESCC development the molecular basis by which dietary ZD promotes ESCC and by which ZS and ZR prevent cancer development. Our study suggests that deficiency of Zn in the diet is a likely etiologic agent of chronic inflammation in the esophagus that contributes to ESCC development in humans. Because ESCC is prevalent in many regions in the world, the definition of this molecular role of ZD-induced inflammation as a critical factor in ESCC development has clear clinical implications with regard to development and prevention of this deadly disease.

Materials and methods

Animals and diet

Male weanling male Sprague-Dawley rats (50 ± 5 g) were obtained from Taconic Laboratory (Germantown, NY). Custom-formulated egg-white based diets were obtained from Harlan Teklad (Madison, WI). ZD and ZS diets were identical except for the Zn content, which was 3 ppm and 65 ppm, respectively.(20) NMBA was purchased from Midwest Research Institute (Kansas City, MO).

Experimental design

Animal studies were carried out according to approved institutional approved protocols. The study design is shown in Figure 1a. Weanling rats were fed a ZD or ZS diet for 6 weeks to establish esophageal proliferation in ZD rats (20). ZD rats were fed ad libitum and ZS rats were pair-fed to ZD animals to match the decreased food consumption of ZD rats. Then, all rats were administered the 1st intragastric dose of NMBA (2 mg/kg body weight). A day later, half of the ZD rats were switched to a ZS diet and became ZR; the remaining ZD and ZS rats continued on their diets, thus forming the ZD, ZS, and ZR groups. Five weeks after the 1st NMBA dose, 14 rats per group were sacrificed for tumor evaluation and gene expression profiling studies. The 2nd and 3rd NMBA dose was administered at 6 and 12 weeks after the 1st NMBA dose. At 21 weeks (15 weeks after the 1st NMBA dose, and 21 weeks after ZD diet), the animals (24 ZD, 21 ZS, and 23 ZR rats) were sacrificed for tumor evaluation and related gene expression profiling studies.

Long-Term Zn deficiency Study

Weanling male rats were fed a ZD or pair-fed a ZS diet for 23 weeks (10 ZS and 19 ZD rats). At study conclusion, esophagus was isolated as described above.

Esophageal Epithelia Preparation

Esophagi were cut into 2 portions. One portion was fixed in 10% buffered formalin for histological and immunohistochemical studies. Esophageal epithelium was prepared from the remaining portion by using a blade to remove the submucosal and muscularis layers, snap-frozen in liquid nitrogen, and stored at –80°C (20).

RNA preparation

Total RNA was prepared from individual esophagus by using TRIZOL reagent (Invitrogen, Carlsbad, CA). The integrity of RNA samples will be determined by the Agilent 2100 Bioanalyzer to ensure the samples display intact ribosomal RNA 18S and 28S bands.

Gene expression profiling

Expression profiling of esophageal epithelia from NMBA-treated ZD, ZR, and control ZS rats (n=4/group) was performed at 5 weeks after the first NMBA dose and at tumor endpoint, using Affymetrix Rat Genome 230 2.0 GeneChip (Affymetrix, Santa Clara, CA) that contains more than 31,000 probe sets that represent >30,000 transcripts and variants from ~28,000 genes. Five μg of total RNA were reverse transcribed into cDNA followed by in vitro transcription and labeling to produce biotin-labeled cRNA (IVT Labeling Kit). The labeled cRNA was hybridized to the array according to the manufacturer's protocol, performed by the Ohio State University Comprehensive Cancer Center Microarray Facility. The distribution of fluorescent material on the oligonucleotide probes sets is obtained using GeneArray Scanner.

Expression data analysis

The normalization of microarray data was performed by using Affy (47), a Bioconductor package for the R statistical programming language ( Analysis of Microarrays Software (SAM, version 3.0) was used to identify differentially expressed mRNAs. Based on the False Discovery Rate, the cutoff for significance is set so that the estimated q-value is between zero and 5%.

Gene ontology (GO) and pathway analyses

DAVID bioinformatics (31) was used to identify relevant biological processes/functions from expression data captured by transcriptome analysis. The DAVID Gene Functional Classification Tool generates a gene-to-gene similarity matrix based shared functional annotation using over 75,000 terms from 14 functional annotation sources.

IPA software ( was used to analyze probable network/pathway. For each data set, the selected genes were uploaded into the IPA application. Networks were then algorithmically generated based on gene-gene connectivity. Each gene identifier was mapped to its corresponding object in Ingenuity's Knowledge Base. These molecules, called Network Eligible molecules, were overlaid onto a global molecular network developed from information contained in Ingenuity's Knowledge Base. Networks of Network Eligible Molecules were then algorithmically generated based on their connectivity.

Quantitative real-time RT-PCR

qRT-PCR was performed using pre-designed probes (Applied Biosystems), PSMB6 as the normalizer, and the comparative Ct method. The analysis used the ABI 7300 Real-time PCR System (Applied Biosystems, Foster City, CA).

Immunoblot analysis

Samples of 20-50 μg of total protein are separated by 4-14% SDS-polyacrylamide gel electrophoresis and transferred onto a polyvinylidene difluoride membrane. After incubation with respective primary antibodies: S100A8 (SC-8113, Santa Cruz Biotechnology, Santa Cruz, CA), S100A9 (AF2065, R&D Systems, Minneapolis, MN), COX-2 (160106, Cayman Chemical, Ann Arbor, MI), HIF1A (Ab-1, Abcam, Cambridge, MA), the membrane was incubated with appropriate peroxidase-conjugated secondary antibodies. Immunodetection was by Pierce enhanced chemiluminescence substrate (Thermo Scientific, Waltham, MA). GAPDH (CB1001, Calbiochem, San Diego, CA) was used as a loading control. Intensity of protein bands was quantified by GS-800 calibrated densitomter ((BIO-RAD, Hercules, CA).

Enzyme-linked immunosorbent assay (ELISA)

Esophageal mucosal whole-cell extracts were prepared using a nuclear extract kit (40010, Active Motif, Carlsbad, CA). The expression of IL1B, CXCL5, and NF-kB p65 was quantified by ELISA system (IL1B, DY501, R&D; CXCL5, ELR-LIX-001C, RayBiotech, Norcross GA, and NF-kB p65, 40096, Active Motif) according to the manufacturer's instructions.

Immunohistochemistry (IHC)

Formalin-fixed, paraffin-embedded tissues were deparaffinized, and rehydrated in graded alcohols. IHC was carried out as previously described (20, 22), using primary antibodies for proliferating cell nuclear antigen (PCNA) (clone PC-10, Ab-1, Neomarker (Thermo Scientific), KRT14 (NCL-LL002, Novocastra, Buffalo Grove, IL), COX-2 (160106, Cayman Chemical), S100A8 (T-1032, BMA, Augst, Switzerland), S100A9 (AF2065, R&D), and NF-kB p65 (Ab-7970, Abcam). Protein was localized by incubation with 3-amino-9-ethylcarbazole substrate-chromogen (AEC) (Dako, Carpinteria, CA) or 3,3′-diaminobenzidine tetrahydrochloride (DAB) (Sigma-Aldrich, St. Louis, MO).

Serum Zn measurement

Blood was collected from the retro-orbital venous plexus of rats after anesthesia with isoflurane (GE Healthcare). Serum zinc analysis was by atomic absorption spectrometry, using AAnalyst 400 (Perkin Elmer, Waltham, MA).

Statistical analysis

Data on tumor multiplicity and serum Zn were analyzed by analysis of variance (ANOVA). Differences among the groups were assessed using post hoc t-tests with the Tukey-HSD multiple comparisons. Confidence intervals (CI) were calculated using standard t-statistics and an alpha-level of 0.05. Differences in tumor/ESCC incidence were assessed by Fisher's exact test. Statistical tests were 2-sided and were considered significant at p<0.05. Statistical analysis was performed using R (

Supplementary Material

Dataset 1

Dataset 2

Dataset 3

Figure S1

Supplementary Table 1


We thank Dr. Kay Huebner (The Ohio State University) for reading and critical discussion of the manuscript. We thank Shao-gui Wan's assistance with animal care and validation of the array data. The microarray data were submitted to ArrayExpress (Accession number: E-MTAB-428). This work was supported by NIH grants R01CA118560 (L.Y.Y.F.) and R01CA115965 (C.M.C.).


Classification: Cancer Biology

This article contains supplementary information.

Conflict of interest

The authors declare no conflict of interest.


1. Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer. 2010 Dec 15;127(12):2893–917. [PubMed]
2. Balkwill F, Mantovani A. Inflammation and cancer: back to Virchow? Lancet. 2001 Feb 17;357(9255):539–45. [PubMed]
3. Coussens LM, Werb Z. Inflammation and cancer. Nature. 2002 Dec 19-26;420(6917):860–7. [PMC free article] [PubMed]
4. Karin M. Nuclear factor-kappaB in cancer development and progression. Nature. 2006 May 25;441(7092):431–6. [PubMed]
5. Mandard AM, Hainaut P, Hollstein M. Genetic steps in the development of squamous cell carcinoma of the esophagus. Mutat Res. 2000 Apr;462(2-3):335–42. [PubMed]
6. Magee PN. The experimental basis for the role of nitroso compounds in human cancer. Cancer Surv. 1989;8(2):207–39. [PubMed]
7. Yang CS. Research on esophageal cancer in China: a review. Cancer Res. 1980 Aug;40(8 Pt 1):2633–44. [PubMed]
8. Abnet CC, Lai B, Qiao YL, Vogt S, Luo XM, Taylor PR, et al. Zinc concentration in esophageal biopsy specimens measured by x-ray fluorescence and esophageal cancer risk. J Natl Cancer Inst. 2005 Feb 16;97(4):301–6. [PubMed]
9. Kmet J, Mahboubi E. Esophageal cancer in the Caspian littoral of Iran: initial studies. Science. 1972 Feb 25;175(24):846–53. [PubMed]
10. Poschl G, Seitz HK. Alcohol and cancer. Alcohol Alcohol. 2004 May-Jun;39(3):155–65. [PubMed]
11. Ho E, Ames BN. Low intracellular zinc induces oxidative DNA damage, disrupts p53, NFkappa B, and AP1 DNA binding, and affects DNA repair in a rat glioma cell line. Proceedings of the National Academy of Sciences of the United States of America. 2002 Dec 24;99(26):16770–5. [PubMed]
12. Prasad AS, Kucuk O. Zinc in cancer prevention. Cancer Metastasis Rev. 2002;21(3-4):291–5. [PubMed]
13. Vallee BL, Falchuk KH. The biochemical basis of zinc physiology. Physiol Rev. 1993 Jan;73(1):79–118. [PubMed]
14. Rink L, Haase H. Zinc homeostasis and immunity. Trends Immunol. 2007 Jan;28(1):1–4. [PubMed]
15. Berg JM, Shi Y. The galvanization of biology: a growing appreciation for the roles of zinc. Science. 1996 Feb 23;271(5252):1081–5. [PubMed]
16. Stinson SF, Squire RA, Sporn MB. Pathology of esophageal neoplasms and associated proliferative lesions induced in rats by N-methyl-N-benzylnitrosamine. J Natl Cancer Inst. 1978 Dec;61(6):1471–5. [PubMed]
17. Stoner GD, Gupta A. Etiology and chemoprevention of esophageal squamous cell carcinoma. Carcinogenesis. 2001 Nov;22(11):1737–46. [PubMed]
18. Pegg AE. Methylation of the O6 position of guanine in DNA is the most likely initiating event in carcinogenesis by methylating agents. Cancer Invest. 1984;2(3):223–31. [PubMed]
19. Lozano JC, Nakazawa H, Cros MP, Cabral R, Yamasaki H. G-->A mutations in p53 and Ha-ras genes in esophageal papillomas induced by N-nitrosomethylbenzylamine in two strains of rats. Mol Carcinog. 1994 Jan;9(1):33–9. [PubMed]
20. Fong LY, Nguyen VT, Farber JL. Esophageal cancer prevention in zinc-deficient rats: rapid induction of apoptosis by replenishing zinc. J Natl Cancer Inst. 2001 Oct 17;93(20):1525–33. [PubMed]
21. Fong LY, Zhang L, Jiang Y, Farber JL. Dietary zinc modulation of COX-2 expression and lingual and esophageal carcinogenesis in rats. J Natl Cancer Inst. 2005 Jan 5;97(1):40–50. [PubMed]
22. Taccioli C, Wan SG, Liu CG, Alder H, Volinia S, Farber JL, et al. Zinc replenishment reverses overexpression of the proinflammatory mediator S100A8 and esophageal preneoplasia in the rat. Gastroenterology. 2009 Mar;136(3):953–66. [PMC free article] [PubMed]
23. Mantovani A, Allavena P, Sica A, Balkwill F. Cancer-related inflammation. Nature. 2008 Jul 24;454(7203):436–44. [PubMed]
24. Gebhardt C, Nemeth J, Angel P, Hess J. S100A8 and S100A9 in inflammation and cancer. Biochem Pharmacol. 2006 Nov 30;72(11):1622–31. [PubMed]
25. Balkwill F. Cancer and the chemokine network. Nat Rev Cancer. 2004 Jul;4(7):540–50. [PubMed]
26. Williams CS, Mann M, DuBois RN. The role of cyclooxygenases in inflammation, cancer, and development. Oncogene. 1999 Dec 20;18(55):7908–16. [PubMed]
27. Voronov E, Shouval DS, Krelin Y, Cagnano E, Benharroch D, Iwakura Y, et al. IL-1 is required for tumor invasiveness and angiogenesis. Proceedings of the National Academy of Sciences of the United States of America. 2003 Mar 4;100(5):2645–50. [PubMed]
28. Karin M, Greten FR. NF-kappaB: linking inflammation and immunity to cancer development and progression. Nat Rev Immunol. 2005 Oct;5(10):749–59. [PubMed]
29. Eltzschig HK, Carmeliet P. Hypoxia and inflammation. N Engl J Med. 2011 Feb 17;364(7):656–65. [PubMed]
30. Cintorino M, Tripod SA, Santopietro R, Antonio P, Lutfi A, Chang F, et al. Cytokeratin expression patterns as an indicator of tumour progression in oesophageal squamous cell carcinoma. Anticancer Res. 2001 Nov-Dec;21(6A):4195–201. [PubMed]
31. Huang da W, Sherman BT, Lempicki RA. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc. 2009;4(1):44–57. [PubMed]
32. Roussos ET, Condeelis JS, Patsialou A. Chemotaxis in cancer. Nat Rev Cancer. 2011;11(8):573–87. [PubMed]
33. Samad TA, Moore KA, Sapirstein A, Billet S, Allchorne A, Poole S, et al. Interleukin-1beta-mediated induction of Cox-2 in the CNS contributes to inflammatory pain hypersensitivity. Nature. 2001 Mar 22;410(6827):471–5. [PubMed]
34. Tu S, Bhagat G, Cui G, Takaishi S, Kurt-Jones EA, Rickman B, et al. Overexpression of interleukin-1beta induces gastric inflammation and cancer and mobilizes myeloid-derived suppressor cells in mice. Cancer Cell. 2008 Nov 4;14(5):408–19. [PMC free article] [PubMed]
35. Wang B, Hendricks DT, Wamunyokoli F, Parker MI. A growth-related oncogene/CXC chemokine receptor 2 autocrine loop contributes to cellular proliferation in esophageal cancer. Cancer Res. 2006 Mar 15;66(6):3071–7. [PubMed]
36. Zimmermann KC, Sarbia M, Weber AA, Borchard F, Gabbert HE, Schror K. Cyclooxygenase-2 expression in human esophageal carcinoma. Cancer Res. 1999 Jan 1;59(1):198–204. [PubMed]
37. Kaifi JT, Yekebas EF, Schurr P, Obonyo D, Wachowiak R, Busch P, et al. Tumor-cell homing to lymph nodes and bone marrow and CXCR4 expression in esophageal cancer. J Natl Cancer Inst. 2005 Dec 21;97(24):1840–7. [PubMed]
38. Kumar A, Chatopadhyay T, Raziuddin M, Ralhan R. Discovery of deregulation of zinc homeostasis and its associated genes in esophageal squamous cell carcinoma using cDNA microarray. Int J Cancer. 2007 Jan 15;120(2):230–42. [PubMed]
39. Allavena P, Germano G, Marchesi F, Mantovani A. Chemokines in cancer related inflammation. Exp Cell Res. 2011 Mar 10;317(5):664–73. [PubMed]
40. Miyazaki H, Patel V, Wang H, Edmunds RK, Gutkind JS, Yeudall WA. Down-regulation of CXCL5 inhibits squamous carcinogenesis. Cancer Res. 2006 Apr 15;66(8):4279–84. [PubMed]
41. Richmond A. Nf-kappa B, chemokine gene transcription and tumour growth. Nat Rev Immunol. 2002 Sep;2(9):664–74. [PMC free article] [PubMed]
42. Yamamoto K, Kitayama W, Denda A, Morisaki A, Kuniyasu H, Inoue M, et al. Suppressive effects of a selective cyclooxygenase-2 inhibitor, etodolac, on 4-nitroquinoline 1-oxide-induced rat tongue carcinogenesis. Exp Toxicol Pathol. 2004 Dec;56(3):145–51. [PubMed]
43. Nemeth J, Stein I, Haag D, Riehl A, Longerich T, Horwitz E, et al. S100A8 and S100A9 are novel nuclear factor kappa B target genes during malignant progression of murine and human liver carcinogenesis. Hepatology. 2009 Oct;50(4):1251–62. [PubMed]
44. Hiscott J, Marois J, Garoufalis J, D'Addario M, Roulston A, Kwan I, et al. Characterization of a functional NF-kappa B site in the human interleukin 1 beta promoter: evidence for a positive autoregulatory loop. Mol Cell Biol. 1993 Oct;13(10):6231–40. [PMC free article] [PubMed]
45. Sun J, Liu J, Pan X, Quimby D, Zanesi N, Druck T, et al. Effect of zinc supplementation on N-nitrosomethylbenzylamine-induced forestomach tumor development and progression in tumor suppressor-deficient mouse strains. Carcinogenesis. 2011 Mar;32(3):351–8. [PMC free article] [PubMed]
46. Fong LY, Jiang Y, Rawahneh ML, Smalley KJ, Croce CM, Farber JL, et al. Zinc supplementation suppresses 4-nitroquinoline 1-oxide-induced rat oral carcinogenesis. Carcinogenesis. 2011 Apr;32(4):554–60. [PMC free article] [PubMed]
47. Bolstad BM, Irizarry RA, Astrand M, Speed TP. A comparison of normalization methods for high density oligonucleotide array data based on variance and bias. Bioinformatics. 2003 Jan 22;19(2):185–93. [PubMed]