| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Human Cancer Biology |
Authors' Affiliations: 1 Cancer Research UK Institute for Cancer Studies, University of Birmingham; 2 Sandwell and West Birmingham Hospitals NHS Trust, Birmingham, United Kingdom; 3 Academic Unit of Pathology, University of Sheffield Medical School, Sheffield, United Kingdom; and 4 Iron Metabolism Laboratory, Queensland Institute of Medical Research, PO Royal Brisbane Hospital, Brisbane, Queensland, Australia
Requests for reprints: Chris Tselepis, Cancer Research UK Institute for Cancer Studies, University of Birmingham, Vincent Drive, Birmingham B15 2TH, United Kingdom. Phone: 44-121-414-2972; Fax: 44-121-627-2384; E-mail: c.tselepis{at}bham.ac.uk.
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
Experimental Design: Perls' staining was used to examine iron deposition in tissue. mRNA expression in samples of Barrett's metaplasia matched with esophageal adenocarcinoma and samples of Barrett's metaplasia without evidence of adenocarcinoma were examined by real-time PCR. Semiquantitative immunohistochemistry was used to examine cellular localization and protein levels. The effect of iron loading on cellular proliferation and iron transporter expression was determined in esophageal cell lines OE33 and SEG-1 using a bromodeoxyuridine assay and real-time PCR, respectively.
Results: In the progression of Barrett's metaplasia to adenocarcinoma, there was overexpression of divalent metal transporter 1 (DMT1), transferrin receptor 1, duodenal cytochrome b, ferroportin, and H-ferritin, and these changes were associated with increased iron deposition. Overexpression of DMT1 was further associated with metastatic adenocarcinoma. Iron loading OE33 and SEG-1 cells caused increased cellular proliferation, which was associated with increased H-ferritin and decreased transferrin receptor 1 and DMT1 expression.
Conclusions: Progression to adenocarcinoma is associated with increased expression of iron import proteins. These events culminate in increased intracellular iron and cellular proliferation. This may represent a novel mechanism of esophageal carcinogenesis.
In addition to the plethora of proteins implicated in the malignant progression of Barrett's metaplasia to adenocarcinoma, including c-MYC (7), there is evidence suggesting that iron is important in the development of esophageal cancer (8–14). The involvement of iron in esophageal carcinogenesis is exemplified by a series of studies using an animal model in which rats underwent an esophagoduodenal anastomosis to mimic reflux (10–12). Rates of esophageal adenocarcinoma 30 weeks after surgery were 
10-fold higher in animals supplemented with iron than in nonsupplemented animals. The cellular role of iron in carcinogenesis has been attributed to its potential to induce reactive oxygen species, DNA adducts, and cellular proliferation and repress cell adhesion (15–18).
Iron is principally absorbed in the duodenum where dietary ferric iron is reduced and transported into the enterocyte by the proteins duodenal cytochrome b (DCYTB) and divalent metal transporter 1 (DMT1), respectively (19–21). The iron can then be either stored in an inert form bound to ferritin (22), used by the cell, or exported from the enterocyte via a pathway that requires the ferroxidase hephaestin (23, 24) and the basolateral iron transporter ferroportin (25). Iron is transported in the serum bound to transferrin, which interacts with transferrin receptor 1 (TfR1) on the plasma membrane of cells that take up iron. The iron/transferrin complex is internalized by receptor-mediated endocytosis and iron released (26). In esophageal keratinocytes, the main route of iron acquisition is likely to be through the transferrin receptor–mediated endocytic pathway. Luminal iron absorption in the esophagus is highly unlikely.
A role for iron in carcinogenesis is further strengthened by our own studies showing that in the human colonic adenoma to carcinoma sequence there is a modulation in the expression of iron transport proteins, particularly an overexpression of DCYTB, DMT1, and TfR1 and a reduction in hephaestin (15). These changes culminate in increased intracellular iron import and reduced iron efflux leading to increased intracellular iron, which we believe is crucial in driving Wnt signaling, a common oncogenic signaling pathway (27).
Our hypothesis is that the modulation of iron transport proteins, culminating in cellular iron accumulation, represents a mechanism of carcinogenesis central to all cancers. Thus, the primary aim of this study was to characterize the expression of the iron import (DCYTB, DMT1, and TfR1), export (ferroportin and hephaestin), and storage (H-ferritin) proteins in normal esophagus, Barrett's metaplasia, and esophageal adenocarcinoma and to further determine the effect of increased intracellular iron on esophageal cellular proliferation and the expression of iron transport proteins.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Patient tissue
Esophageal adenocarcinoma resection specimens. Samples of esophageal adenocarcinoma (n = 29), some of which were matched with Barrett's metaplasia (n = 11) from the same resection specimen, were collected during surgery and each tissue specimen was divided in two for RNA extraction and pathologic confirmation.
Endoscopic specimens. Samples of long-segment (
3 cm) Barrett's metaplasia (n = 13), defined as columnar mucosa with intestinal-type goblet cells, with matched normal gastric fundal mucosa and esophageal squamous mucosa from the same patient were collected during endoscopy. Each tissue specimen was divided in two for RNA extraction and immunohistochemistry. Any patients with Barrett's metaplasia with evidence of dysplasia or adenocarcinoma were excluded from this study.
Archived tissue. Paraffin sections of normal esophagus (n = 10), Barrett's metaplasia from patients with no evidence of esophageal dysplasia or adenocarcinoma (BM-ve; n = 20), Barrett's metaplasia with adenocarcinoma in the same section (BM+ve; n = 20), Barrett's with low-grade dysplasia (n = 10), Barrett's with high-grade dysplasia (n = 20), and esophageal adenocarcinoma (n = 20) were identified within the archived tissue bank, Department of Pathology, Queen Elizabeth's Hospital Birmingham, and processed for immunohistochemistry.
Real-time PCR
Real-time PCRs were done using 18S rRNA as an internal standard (PE Biosystems, Roche) and primers to DMT1 [iron response element (IRE)-containing isoform], TfR1, DCYTB, ferroportin, hephaestin, H-ferritin, and MYC as previously described (7, 15). All analyses were done on the same pool of cDNA samples.
Immunohistochemistry
Immunohistochemistry was done as previously described (15) using microwave antigen retrieval and rabbit polyclonal antibodies to (a) DCYTB (1:200, clone 834; a kind gift from Prof. A. McKie), (b) DMT1 (1:3,000; ADI), (c) ferroportin (1:200, clone 3566; a kind gift from Prof. A. McKie), (d) H-ferritin (1:1,000, F-5012; Sigma), and (e) hephaestin (1:50, HEPH11-A; ADI) or a mouse monoclonal antibody to TfR1 (1:40, clone 10F11; Novocastra). Small bowel mucosa was included as a positive control, and omission of primary antibody and the use of rabbit serum alone and isotype controls were all used as negative controls. In addition, cellular iron localization was assessed by 3,3'-diaminobenzidine–enhanced Prussian blue staining as previously described (15). The slides were scored by a previously described method for intensity of staining and percentage of epithelial cell staining (28). All sections were scored independently by three observers (S.H., K.R., and C.T.).
Tissue microarrays
A separate set of esophageal carcinoma specimens (n = 76) collected between 2000 and 2005, for which clinicopathologic data were available, was used to evaluate the association between iron transporter levels and prognostic factors. The prognostic factors included T stage, differentiation status, vascular invasion, and nodal involvement. Each tumor is represented on the microarray in triplicate to provide representative sampling of the source tissue. Immunohistochemistry and scoring were done as described above.
Western blotting
Western blotting was done as described previously with antibodies to either ferroportin (1:1,000) or H-ferritin (1:2,500; ref. 15). A mouse monoclonal antibody to cytokeratin 19 (1:2,000, cloneA53-B/A2.26; Merck Chemicals Ltd.) was used for normalization of epithelial protein loading. Immunoreactive bands were then subjected to densitometry using NIH Image 1.62 software.
Cell culture
Cell lines derived from esophageal adenocarcinomas [OE33 (29) and SEG-1 (30)] were routinely cultured in DMEM (Life Technologies) with 10% FCS supplemented with 100 units/mL penicillin and 0.1 mg/mL streptomycin. On reaching 70% confluence, lines were challenged with either growth medium alone (control) or iron loading medium (growth medium supplemented with 100 µmol/L FeSO4 and 10 µmol/L sodium ascorbate) for between 1 and 24 h. FeSO4 (100 µmol/L) was chosen as this source of iron and concentration has been used in previous studies (15, 27). Furthermore, this concentration of iron was shown to be optimal for esophageal cellular proliferation (data not shown). At the end of the time period, cells were used for intracellular iron determination, RNA extraction, or cell proliferation assays. All experiments were done in triplicate on three separate occasions.
Bromodeoxyuridine incorporation assay
A colorimetric cell proliferation ELISA was done according to the manufacturer's instructions (Roche Applied Science). Briefly, cells were labeled with bromodeoxyuridine followed by fixation and incubation with anti-bromodeoxyuridine peroxidase, and the immune complex was then detected using a 3,3',5,5'-tetramethylbenzidine substrate with the reaction product assessed at 370 nm.
Ferrozine assay
Nonhaem iron was assayed as previously described (15) and cellular iron content was expressed as nmoles of iron per mg protein. Protein concentrations were assessed by the bicinchoninic acid protein assay according to the manufacturer's instructions (Pierce).
Statistics
All data are presented as mean ± 1 SE. Statistical significance was calculated using Mann-Whitney tests for analysis of immunohistochemical staining, paired t tests for mRNA analysis where multiple samples were obtained from individual patients, and unpaired Student's t tests for analysis of the in vitro experiments. To assess the association between c-MYC and iron transport proteins, linear regression analysis was done on log-transformed mRNA fold change values, and the R2 value was determined to show the extent to which expression of TfR1 and H-ferritin was correlated to MYC (31). A Bonferroni adjustment was applied to P values attained from multiple comparisons during the analysis of prognostic factors (32). In univariate analyses, significance was accepted at P 
0.05 following Bonferroni adjustment, and P 0.0083 was considered significant. All analyses were done using Statistical Package for the Social Sciences version 10.0 (SPSS, Inc.).
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
|
Immunolocalization of iron metabolism proteins in archived tissue
To complement the mRNA analyses, we did semiquantitative immunohistochemistry on archived paraffin tissue specimens from patients with a range of esophageal pathologies (Figs. 1
and 2
). In addition to cellular localization being determined, all staining was scored (Table 2
) as described above.
|
|
|
No discernible DMT1 immunoreactivity was observed in normal esophageal epithelium unlike in Barrett's metaplasia where staining was evident on the apical surface epithelium. However, no immunoreactivity was observed in the deeper lying crypts. In the majority of columnar cells with low- and high-grade dysplasia, there was cytoplasmic staining, which became more widespread in adenocarcinoma (Fig. 1). Semiquantitative analysis showed a statistical increase in DMT1 immunoreactivity in esophageal adenocarcinoma compared with nondysplastic Barrett's metaplasia (P = 0.001; Table 2).
TfR1, like DMT1, was not detected in normal esophagus. In Barrett's metaplasia, there was lateral membranous immunoreactivity in all columnar cells, whereas in high-grade dysplastic Barrett's metaplasia and in adenocarcinoma there was an abundance of both membranous and cytoplasmic immunoreactivity (Fig. 1). Semiquantitative analysis revealed that there was significantly higher TfR1 immunoreactivity in high-grade dysplastic Barrett's metaplasia and also in adenocarcinoma compared with nondysplastic Barrett's metaplasia (P = 0.0001 and 0.0001, respectively; Table 2).
Iron export and storage proteins. Neither ferroportin nor hephaestin was detected in normal squamous mucosa but there was evidence of immunoreactivity for both proteins in Barrett's metaplasia. Localization of ferroportin was mostly supranuclear on the surface epithelium with diminishing reactivity down into the crypts. Hephaestin was localized in a diffuse cytoplasmic manner throughout all columnar cells with immunoreactivity more intense toward the basal pole of the columnar cells. In adenocarcinoma, whereas there was an abundance of discrete supranuclear and vesicular ferroportin immunoreactivity, there was only weak diffuse immunoreactivity for hephaestin (Fig. 2). Semiquantitative analysis showed that ferroportin was overexpressed in adenocarcinoma compared with nondysplastic Barrett's metaplasia (P = 0.0002), whereas hephaestin expression showed no statistical difference in expression when compared with nondysplastic Barrett's metaplasia.
The storage protein H-ferritin was expressed in normal squamous esophagus in a distribution similar to that of DCYTB. Immunoreactivity was mostly confined to the basal layer and was predominantly cytoplasmic in all columnar cells. In nondysplastic and dysplastic Barrett's metaplasia, weak diffuse immunoreactivity was observed in all cellular compartments, whereas in adenocarcinoma there was an abundance of cytoplasmic immunoreactivity (Fig. 2). Semiquantitative analysis showed that H-ferritin was overexpressed in adenocarcinoma compared with nondysplastic Barrett's metaplasia (P = 0.04; Table 2).
Thus, in the progression from normal esophagus through Barrett's metaplasia to adenocarcinoma, there is an overexpression of iron import proteins (DCYTB, DMT1, and TfR1), iron export protein (ferroportin), and iron storage protein (H-ferritin). However, there was no alteration in hephaestin expression.
Tissue microarray. We further sought to determine whether changes in protein expression were associated with prognostic factors, such as stage of disease and nodal involvement, by using a tissue array of esophageal adenocarcinomas with known outcome (Table 3 ). The tissue microarray analysis showed no significant changes in staining intensity for any of the iron transport proteins analyzed with respect to T stage, tumor differentiation status, presence of vascular invasion, age, or gender. However, DMT1 expression was significantly associated with lymph node involvement (P = 0.0068).
|
|
|
|
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
During the progression from Barrett's metaplasia to adenocarcinoma, we showed that components of the main cellular iron import machinery (TfR1, DMT1, and DCYTB) were overexpressed, as was the iron storage protein H-ferritin. Surprisingly, the major iron export protein ferroportin was also overexpressed under these conditions, although it seemed to be cytoplasmic in localization, suggesting a lack of functionality. These data are comparable with our previous findings in colorectal carcinogenesis (15).
Using a semiquantitative method, we further showed that these changes in iron transport proteins were not just evident in late-stage disease but that two of the proteins, TfR1 and DCYTB, were induced in the transition from low- to high-grade Barrett's dysplasia. Akin to our previous data in colon carcinogenesis, DMT1 overexpression was associated with more aggressive disease, as shown by the correlation of DMT1 expression and nodal metastasis. Whether changes in iron transport proteins can drive metastasis is unclear; however, it is possible that elevated intracellular iron caused by increased expression of iron import protein DMT1 mediates a repression in the cell adhesion molecule E-cadherin, a molecule that is commonly silenced in cellular invasion and metastasis. In support of this, we have previously shown iron-mediated E-cadherin repression in a colorectal model system (15).
We believe that it is highly unlikely that the overexpression of these iron transport proteins represents increased luminal iron uptake. Rather, we predict that their overexpression, coupled to their described localization, in adenocarcinoma suggests that the main route of iron acquisition is through transferrin receptor–mediated endocytosis. The only anomaly to this is DCYTB, which has no role in transferrin receptor–mediated endocytosis. We describe it to be cytoplasmic in localization, suggesting that neither is it involved in luminal iron uptake in adenocarcinoma. However, the fact that we observed DCYTB in the very basal layer of the stratified squamous esophagus suggests that it may have a function other than in iron absorption, such as cellular proliferation.
Thus, increased transferrin receptor–mediated endocytosis in the face of a potential functional loss of iron export, as suggested by cytoplasmic localization of ferroportin, might represent a mechanism by which iron could accumulate within esophageal epithelial cells. Indeed, we showed increased stainable iron and increased H-ferritin expression in tissue sections of Barrett's metaplasia and adenocarcinoma. If these adenocarcinoma cells are iron rich, one might predict reduced TfR1 and DMT1 expression because elevated iron will reduce the interaction of iron regulatory protein (IRP) with 3' IRE and consequently reduce mRNA stabilization (33, 34). However, we have observed paradoxical overexpression of TfR1 and DMT1 that suggests an aberration in the IRP/IRE system. Alternatively, these proteins may be modulated by other factors and that, despite normal IRP/IRE iron sensing, the balance is skewed toward induced expression. One such factor could be c-MYC, an oncogene reported to induce TfR1 and IRP2, while repressing H-ferritin expression (31, 35). Interestingly, we previously showed c-MYC to be induced by acidified bile acids and overexpressed in the esophageal metaplasia-dysplasia-adenocarcinoma sequence (7). We compared the mRNA expression of MYC and TfR1 in both Barrett's metaplasia and adenocarcinoma tissue and showed a positive correlation in adenocarcinoma but not in Barrett's metaplasia. This may suggest that in adenocarcinoma the balance is skewed toward MYC-mediated TfR1 induction rather than an IRE/IRP-mediated TfR1 repression.
Iron loading esophageal adenocarcinoma cell lines resulted in the predicted reduction in TfR1 and DMT1 mRNA and increase in H-ferritin protein expression. Iron loading SEG-1 cells also resulted in a significant reduction in ferroportin protein expression. Because DMT1 has a 5' IRE, increasing cellular iron would be expected to result in increased ferroportin expression as reported in some cell types, such as macrophages (36). However, our data are consistent with other studies that suggest that iron does not regulate intestinal ferroportin expression in an IRE-predicted manner. This is exemplified by several studies showing elevated ferroportin expression in the duodenum of iron-deficient mice (25, 37).
Irrespective of the molecular mechanism by which adenocarcinoma cells become iron loaded, the cellular downstream effects are likely to be multiple. In this and previous studies, iron loading is important in cellular proliferation (15, 17, 27, 36). Historically, this has been attributed to iron being crucial in the activity of ribonucleotide reductase, an iron-containing enzyme responsible for the synthesis of deoxyribonucleotides (38). However, more recently, the phosphatase CDC14A, which has been shown to directly regulate the activity of several cell cycle proteins, is an IRE-containing protein and thus regulated by iron (39). In addition, we have recently shown that iron can induce Wnt signaling, which culminates in oncogene activation and cellular proliferation (27). In support of a role for iron in proliferation, studies by Le and Richardson (40) have shown that reducing intracellular iron levels by chelation blocks cellular proliferation and induces differentiation through several cell cycle proteins, including NDRG1.
As well as modulating cellular proliferation, iron is likely to induce, through Fenton reaction chemistry, reactive oxygen species and oxidative stress (16). In vitro studies have clearly shown that iron can induce oxidative damage to DNA, protein, and lipid. Furthermore, murine models of esophageal adenocarcinoma suggest that iron-mediated oxidative stress in concert with reflux is central to esophageal carcinogenesis, as exemplified by marked overexpression of oxidative stress–responsive genes (9–12).
What remains unclear following this study is how these cellular and molecular alterations are linked to the well-characterized risk factors for esophageal adenocarcinoma (male sex, obesity, gastroesophageal reflux, and the presence of Barrett's metaplasia; ref. 1). Our previous studies have shown that c-MYC can be induced by components of the refluxate (i.e., acidified bile acids; ref. 7). Thus, because c-MYC is a trigger for changes in TfR1, H-ferritin, and IRP2, this would suggest that ablating reflux may represent a strategy for reducing intracellular iron and malignant progression.
In summary, we have shown changes in iron transport proteins in the malignant progression of Barrett's metaplasia to adenocarcinoma and speculate that this may act to increase cellular iron acquisition. Whether this represents an essential molecular aberration in esophageal carcinogenesis or whether it reflects the increased metabolic rate and mitotic activity of cancer tissue requires delineation.
|
Footnotes |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Note: T. Iqbal and C. Tselepis contributed equally to this work.
Received 5/ 3/07; revised 10/11/07; accepted 10/11/07.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
