It is generally accepted that Barrett epithelium can progress through a metaplasia-dysplasia-carcinoma sequence, but the mechanism is unclear. As such, Barrett esophagus (BE) is considered a premalignant condition, increasing a patient's risk of developing esophageal cancer 50- to 100-fold, compared with the general population (1). Barrett-associated adenocarcinoma (BAA) is a highly lethal esophageal tumor with an overall 5-year survival rate of 10% to 15% in most Western populations (2).
Gastroesophageal reflux disease (GERD) is estimated to occur in 20% to 40% of adult Americans (3). In most patients, the damaged squamous epithelium heals by regeneration of more squamous cells. Only a minority of patients with GERD actually develop BE, which is a consequence of continuous reflux esophagitis, induced by hydrochloric acid (HCl) and bile acids, that heals through an alternative metaplastic pathway, in which intestinal-type columnar cells replace reflux-damaged cells (3). BE is defined by the replacement of normal esophageal squamous mucosa with columnar epithelium containing goblet cells (1–3). Although the precise mechanisms involved in the pathogenesis of BE have not been completely elucidated, both gastric acid and bile reflux have been implicated in its development (4–6). However, not every patient with BE has a history of GERD. Epidemiological studies have shown that only 8% to 12% of patients with GERD develop metaplasia (7); BE is detected in only 10% to 15% of patients with GERD; and approximately 30% of patients with BE do not have long-standing GERD (2). Other studies have shown that reflux symptom duration, frequency, or severity were not reliable predictors for the presence of BE in adult patients (2).
Eosinophilic esophagitis (EoE) represents an inflammatory disease of the esophagus characterized by eosinophilic infiltration of the esophageal mucosa, which is frequently associated with atopic disorders (8,9). EoE is increasingly reported in the medical literature, both in adults and in children (10). The diagnosis of EoE is confirmed in symptomatic patients after the exclusion of GERD, by either documenting a negative pH study or by esophageal biopsies demonstrating the persistence of at least 15 eosinophils per high-power field, even after 8 weeks of twice-daily proton-pump inhibitor therapy (9). Patients with EoE frequently have symptoms that mimic GERD, and patients with EoE can be misdiagnosed as having GERD (8). However, the mediators implicated in EoE-induced esophagitis are not HCl or bile acids but the by-product of eosinophil occupation and release of eosinophil granule proteins into esophageal epithelium, such as major basic protein (MBP) and eosinophil peroxidase (EPO) (11,12). The long-term consequence of prolonged exposure of esophageal squamous epithelium to these EoE-associated inflammatory mediators has not been previously described. This leads to the question: could EoE, like GERD, be a precursor of a premalignant condition, like BE, in some patients?
p27Kip1 is a member of the Cip/Kip family of cyclin-kinase inhibitors implicated in the control of the G1 (resting) to S phase (dividing) cell cycle transition (13). p27 inhibits or inactivates the cyclin-dependent kinases that are responsible for phosphorylating substrates required for cell cycle transitions, and appears to be a direct target of mitogenic and antimitogenic stimulation (13–15). As such, p27 has been implicated as a tumor suppressor protein, and decreased expression and activity of p27 has been reported in colon, breast, prostate, lung, brain, and gastric cancers (15). In Helicobacter pylori–associated disease, altered expression of p27 has been observed during both active infection and posteradication (16,17). Unlike many other tumor suppressors, the p27 gene is rarely mutated or silenced and p27 inactivation occurs primarily via posttranslational modification, where increased degradation, mislocalization to the cytoplasm, or increased sequestration by cyclin D1 removes p27 from its nuclear targets (13–16). In BAA, p27 levels are reduced or it are cytoplasmically mislocalized, a finding that appears to be associated with poor prognosis (14,15,18,19).
To determine whether mediators of EoE, like mediators of GERD, caused the development of a premalignant condition, such as BE, we examined p27 because this is a well-documented marker of dysplasia (15). We determined whether exposure to the components of GERD refluxate, bile acid and HCl, or eosinophil granule proteins affected p27 levels and/or subcellular localization in normal esophageal cells. We also determined whether altered p27 expression increased the proliferative rate of esophageal epithelial cells, potentially predisposing them to the development of dysplasia. We also sought to determine whether exposure to eosinophil granule proteins may have similar effects on p27 expression and localization as bile acid and HCl.
MATERIALS AND METHODS
Human esophageal epithelial (HET-1A) cells, purchased from American Type Culture Collection (Manassas, VA), is a human esophageal squamous epithelial cell line immortalized by transfection of the SV40 T antigen early region gene. HET-1A cells were grown in HAMS-F12 media (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (Hyclone Laboratories, Logan, UT), 10 μg/mL human insulin (Sigma, St Louis, MO), 100 U/mL penicillin G, and 100 g/mL streptomycin (Invitrogen). Cells were harvested with 0.05% trypsin (Invitrogen). SEG-1 is a human Barrett esophageal adenocarcinoma cell line (a gift from Dr David G. Beer, University of Michigan, Ann Arbor). SEG-1 cells were grown in Dulbecco minimal essential media (Invitrogen) supplemented with 10% fetal bovine serum, 100 U/mL penicillin G and 100 μg/mL streptomycin (Invitrogen). Cells were harvested with 0.25% trypsin (Invitrogen).
Cell Viability Studies
Cells (105) were plated in 6-well tissue culture dishes. Bile acid, acidified media, or eosinophil granule proteins were added to the cells as described in detail for each experimental condition (as described below). At different times after plating, cells were counted using a hemocytometer.
Chenodeoxycholic acid (CDC; Sigma) was dissolved in dimethylsulfoxide, stored in aliquots at 4°C, and freshly thawed for each experiment. CDC was added to cell culture media at the time of the experiment. Incubations with increasing CDC concentrations (50–300 μmol/L) and increasing exposure times (2–24 hours) were performed.
Media was acidified to pH 2.0 or pH 4.0 using 12 N HCl. Cells were exposed to acidified media for 60 seconds and washed with phosphate-buffered saline (PBS) before recovering in the presence of pH 7.0 media at 37°C for 24 hours.
Eosinophil Granule Proteins
MBP stock vial of 1.1 mg/mL eosinophil granule proteins were (a gift from Dr Hirohito Kita, Mayo Clinic, Rochester, MN) in 0.025 mmol/L sodium acetate buffer, pH 4.3 with 150 mmol/L sodium chloride, and EPO stock vial of 0.83 mg/mL in PBS were stored in aliquots at −80°C. Proteins were added to media at the time of the experiment. HET-1A cells were incubated with MBP at 5 to 10 μg/mL or EPO at 0.8 to 8 μg/mL for 24 hours at 37°C before analysis. The concentrations used for MBP and EPO were based on levels shown to be cytotoxic to guinea pig tracheal epithelium and below the range reported in patients with asthma or hypereosinophilia syndromes (20,21).
RNA Extraction, cDNA Synthesis, and Reverse Transcription Polymerase Chain Reaction
Cells untreated and treated with bile acid were harvested, placed in Trizol (Invitrogen), and stored at −80°C. RNA was extracted using the standard Trizol method. Five micrograms of total RNA were subjected to a reverse transcription reaction to generate cDNA, according to the manufacturer's protocol (Bio-Rad iScript cDNA synthesis kit; Bio-Rad, Hercules, CA). This cDNA was diluted in nuclease-free water. Two microliters of cDNA were amplified in a 20-μL reaction mixture containing 1× iQ SYBR Green Supermix (Bio-Rad) and 0.5 μmol/L primers. Products were visualized by electrophoresis on a 1.2% agarose gel stained with ethidium bromide.
Real-Time PCR Quantification of mRNA Expression
Real-time polymerase chain reactions were performed in triplicate with 2 μL of cDNA, 1× iQ SYBR Green Supermix (Bio-Rad), and 0.5 μmol/L primers against caudal-type homeobox transcription factor 2 (CDX2), FP: 5′ GAA GGA GTT TCA CTA CAG TCG CTA CA-3′; RP: 5′-CAG ATT TTA ACC TGC CTC TCA GAG A-3′) and glyceraldehyde-3-phosphate dehydrogenase, in a final reaction volume of 20 μL.
To analyze p27 protein levels, immunoblot analysis was performed as previously described (22). p27 antibodies were isolated from rabbits immunized with a C-terminal peptide (New England Peptide, Gardner, MA) or were a gift from Joan Massagué Memorial Sloan-Kettering Cancer Center, New York). Anti-actin antibodies (A2066) were from Sigma-Aldrich.
HET-1A cells were grown on coverslips. Cells were treated for 15 minutes with 2 mL of cold 4% paraformaldehyde-PBS, pH 7.4, then washed 3 times with room-temperature PBS. Cells were blocked for 1 hour at room temperature with 0.2% Triton X-100–5% bovine serum albumin (BSA) in PBS. Cells were incubated with primary antibody against p27 (anti-mouse p27 antibody, DCS-72.F6, Neomarkers, Fremont, CA) in 0.2% Triton-1% BSA-PBS overnight at 4°C. Three washes in PBS were performed, followed by incubation with fluorescently labeled goat anti-mouse secondary antibody (Invitrogen) at a 1:1000 dilution in 0.1% Triton X-100–1% BSA in PBS in a light-protected humidified chamber for 1 hour. After 1 wash with PBS, the nuclei were stained for 15 minutes using bisBenzimide (Sigma). After 3 PBS washes, coverslips were mounted on rectangular glass slides with Fluoromount solution and images were captured on a fluorescent microscope.
Flow Cytometric Analysis
For fluorescence-activated cell sorter assays, cells from each experimental condition were fixed in 95% ethanol for at least 1 hour at 4°C and stained with propidium iodide (PI) for 30 minutes at 37°C, then analyzed on a FACScan apparatus (Becton-Dickinson, Franklin Lakes, NJ).
CDX2 RT-PCR Analysis
HET-1A and SEG-1 cells were treated with bile acid as described, and analysis of CDX2 mRNA expression served as a positive control for the effect of bile acid on a viable esophageal epithelial cell line. CDX2 is an intestinal specific transcription factor important in early differentiation and maintenance of intestinal epithelial cells. It is poorly expressed in nontransformed cells, but CDX2 protein expression has been detected in Barrett epithelium and in esophageal epithelium with esophagitis (5,6). Other investigators have demonstrated that CDX2 mRNA expression was increased in HET-1A and SEG-1 cells after bile acid exposure (4). Our results showed that CDX2 mRNA expression was upregulated in both HET-1A and SEG-1 cells following exposure to increasing concentration of CDC for 24 hours (data not shown). CDX2 mRNA expression was normalized to glyceraldehyde-3-phosphate dehydrogenase mRNA expression. These results confirm that the effect of bile acid treatment on HET-1A and SEG-1 cells was consistent with previously reported results (4).
Bile Acid (CDC)
Treatment of HET-1A cells with increasing concentrations of CDC or increased exposure time to 300 μmol/L CDC resulted in reduced cell numbers as measured by cell counting, suggesting that prolonged bile acid treatment killed the cells (Fig. 1A). When cells were treated with lower concentrations of CDC (100 μmol/L), less cell death was detected. FACS analysis to determine cell cycle phase demonstrated that 300 μmol/L CDC treatment reduced the G1 content, suggesting that these cells were proliferating more rapidly (Fig. 1B, left). When the lower, less cytotoxic concentrations of CDC (50–100 μmol/L) were analyzed, the %G1 content continued to decrease up to 96 hours (Fig. 1B, middle, right), suggesting that these cells were proliferating even more rapidly. To directly examine p27 expression, lysates from the CDC-treated cells were assayed by immunoblot analysis using p27 antibodies (Fig. 1C), which showed that p27 levels in whole cell lysate were unchanged in the presence of CDC.
To examine the subcellular localization of p27, indirect immunofluorescence microscopy (IF) using p27 antibodies was performed (Fig. 1D, green). Nuclear bisbenzimide staining (blue) was also performed. In untreated (no tx) and dimethylsulfoxide-treated cells, p27 was predominantly nuclear, colocalizing with bisbenzimide; however, weak cytoplasmic staining was also detected. Treatment with 100 μmol/L CDC treatment for 24 hours appeared to increase the p27 cytoplasmic staining, but 300 μmol/L CDC treatment resulted in markedly increased p27 cytoplasmic staining (Fig. 1D, left). Increasing duration of exposure to 300 μmol/L CDC demonstrated that increased mislocalization of p27 occurred with time (Fig. 1D, right).
Similar analysis was performed with HET-1A cells treated with acidified media (pH 2–pH 4). HET-1A cells were pulsed for 60 seconds with media acidified to pH 2 or pH 4, followed by a 24-hour recovery period in normal media (pH 7). Cell counting experiments demonstrated decreased viability of HET-1A cells following exposure to decreased pH (Fig. 2A); however, 60% of cells were still viable following exposure to pH 2. Exposure to acidified media did not affect cell cycling, as demonstrated by PI staining and FACS analysis to determine cell cycle phase (Fig. 2B). Immunoblot analysis of HET-1A lysates with p27 antibodies demonstrated that p27 levels were unchanged following exposure to acidified media (Fig. 2C). However, similar to the results seen with bile acid treatment, IF microscopy showed an increase in cytoplasmic p27 following exposure to acidified media (Fig. 2D). The cytoplasmic mislocalization appeared to be more significant in cells treated with pH 2 compared with pH 4 (Fig. 2D).
Eosinophil Granule Proteins
HET-1A cells were treated with the eosinophil granule proteins, MBP, and EPO. In contrast to the results seen with bile acid or acidified media treatment, addition of MBP or EPO did not decrease viability of HET-1A cells, as seen in cell counting experiments (Fig. 3A). MBP or EPO treatment did not alter cell cycling as determined by PI staining and FACS analysis (Fig. 3B). p27 levels from treated lysates were unchanged as seen by immunoblot analysis with p27 antibodies (Fig. 3C). However, similar to bile acid or HCl treatment, we observed an increase in cytoplasmic p27 by IF microscopy following a 24-hour incubation with either MBP or EPO (Fig. 3D). This suggested that even short exposure to MBP or EPO was sufficient to alter the subcellular distribution of p27.
The BAA SEG-1 cell line was also treated with bile acid, HCl, MBP, or EPO (data not shown). Although we did see decreased viability and reduced cell numbers following bile acid exposure, no changes in p27 protein levels by immunoblot or subcellular localization by IF microscopy were seen in SEG-1 cells after exposure to bile acid, acidified media, or eosinophil granule proteins. Cell cycling as seen by PI staining and FACS analysis was also unaffected in each experimental condition for SEG-1 cells.
In the present study, we established a cell culture model of esophagitis induced by exposure to bile acid, HCl, or eosinophil granule proteins. All 3 mediators of esophagitis caused mislocalization of p27 from the nucleus to the cytoplasm, suggesting that p27 is a target of extracellular assault in esophageal epithelial cells. To our knowledge, this is the first study showing that products of eosinophilic degranulation, implicated in the pathogenesis of EoE, can cause p27 mislocalization. Mislocalization of p27 removes it from its nuclear targets, eliminating its inhibitory effect on cyclin-dependent kinases (cdk) and increases the risk of unregulated cell proliferation. The reduction in nuclear p27 levels could increase cdk2 activity and permit passage across the G1 checkpoint normally regulated by this protein. p27 may also have a cytoplasmic role because it has been shown to directly interact with and inhibit RhoA (23). This suggests that the net effect of mislocalization to the cytoplasm may help coordinate cell cycle progression with cell migration.
Mislocalization of p27 has been shown to be an important prognostic marker in BAA (14,15,18,19). We wanted to determine whether inactivation of p27 may be an early insult in the dysplasia-metaplasia-carcinoma pathway, due to chronic GERD or EoE. A chronically increased proliferative rate in esophageal epithelial cells would favor the acquisition of additional cancer-predisposing mutations and serve as a first step to cancer development. Although bile acids and HCl have been linked to the development of BE (19,24), eosinophil granule proteins implicated in EoE have not. Recently published case reports described the simultaneous detection of both BE and EoE in 2 pediatric patients (25) and 1 adult patient (26). Whether these are incidental findings or whether an association exists between EoE and BE is unclear. Our results suggest that additional studies are needed to investigate a potential risk of esophageal dysplasia in patients with long-term EoE.
In the present study, a maximum 24-hour incubation time was chosen for the exposure of esophageal epithelial cells to eosinophil granule proteins because the stability of these proteins in culture beyond 24 hours was unknown. Future studies will attempt to address the influence of prolonged incubations of eosinophil proteins on cell viability and proliferation. Further development of a cell culture model of EoE would use a Transwell (Sigma) noncontact coculture of AML14.3D10 cells (eosinophil cell line) with HET-1A cells, similar to that previously used by Furuta et al (27) in studying the effect of eosinophils on T84 colonic epithelial cells. We plan to explore the role of interleukin-5 by adding this cytokine to AML14.3D10 cell media exposed to HET-1A cells via a semipermeable membrane using the Transwell model. This will allow us to study any differences in HET-1A cell viability and proliferation in the presence and absence of interleukin-5 exposure to the AML14.3D10 cells. This coculture system would permit long-term and repeated exposure of HET-1A cells to eosinophil degranulation products, in an attempt to mimic chronic inflammation from eosinophil granule proteins.
Patients with BE experience longer and more frequent episodes of esophageal acid exposure at pH <2 (6). Dvorak et al (6) showed that a 60-second exposure of HET-1A cells to either pH 2 or 4 resulted in the production of reactive oxygen species, which may cause DNA damage. We did not observe any changes in cell cycling of HET-1A cells after exposure to acidified media or eosinophilic degradation proteins, but future studies with longer and repeated cycles of exposure to these conditions may translate into effects on cell cycling that are similar to those observed after CDC exposure.
Unlike HET-1A cells, bile acid and acidified media decreased cell viability of SEG-1 cells, but did not affect p27 protein levels, subcellular localization, or cell cycle status (data not shown). Eosinophil granule proteins also did not have any effects on SEG-1 cells. Because the SEG-1 cell line was derived from transformed esophageal BAA cells, it may have acquired mutations that prevent effects on p27 levels or subcellular localization in response to esophagitis mediators.
Similar to other investigators, we showed that CDC upregulates the expression of CDX2 in HET-1A and SEG-1 cells (4). CDX2 is an intestinal specific transcription factor important in early differentiation and maintenance of intestinal epithelial cells, which is not expressed in normal mature human esophageal epithelial cells. CDX2 protein expression has been reported in BE and in esophageal epithelium with pre-Barrett esophagitis (5,7). We do not know whether CDX2 upregulation is a phenomenon specific to CDC exposure and/or whether p27 mislocalization contributes to its reexpression. No previous association has been reported between CDX2 and EoE, and thus it will be important to investigate this relation.
Other investigators have implicated gene pathways influential in the development of Barrett epithelium, and have potentially linked these pathways to the pathogenesis of EoE. Milano et al (28) reported that in both human and rat tissues, the bone morphogenic protein (BMP) pathway is activated in esophagitis and BE. Similar to CDX2, genes in the BMP pathway are important in intestinal cell development. They note that upon incubation of primary esophagus squamous cell culture with BMP-4, there was a shift in the cytokeratin expression pattern consistent with columnar epithelium. They also found that BMP-4–treated squamous cells showed a gene expression profile similar to that of cultured BE cells (28). Mulder et al (11) discussed the role of fibroblast growth factor 9 (FGF9) in the pathogenesis of EoE. FGF9 was increased in biopsies of EoE patients, which correlated with basal hyperplasia; recombinant human FGF9 caused proliferation of HET-1A cells; and recombinant human FGF9 altered the downstream proliferation genes, BMP-2 and BMP-4. In particular, FGF9 caused decreased BMP-2 and increased BMP-4 expression. The findings by Milano et al (28) and Mulder et al (11) both support the need for more research into the natural history of EoE and whether there is a potential link to BE.
The components of esophageal refluxate, bile acid and HCl, caused mislocalization of p27 from the nucleus to the cytoplasm. Mislocalization of p27 may disrupt its normal cdk-inhibitory function in the nucleus and increase its poorly defined cytoplasmic functions. This change may serve as an early marker of increased cell proliferation and suggests that chronic GERD may increase the risk of developing BE and BAA. Similar effects on p27 subcellular localization were seen in HET-1A cells after EPO or MBP exposure, suggesting that EoE-affected cells may follow a similar dysplastic pathway. Additional experiments in tissue culture and animal models are warranted to determine whether there is a connection between mediators of EoE and dysplasia.
We thank Dr Ken Setchell (Cincinnati Children's Medical Center, Cincinnati, OH) for guidance on bile acid preparation. We are also grateful to Dr David G. Beer (University of Michigan, Ann Arbor) for the SEG-1 cell line and Dr Hirohito Kita (Mayo Clinic, Rochester, MN) for generous donation of eosinophil granule proteins. p27 antibodies were a gift from Joan Massagué at Memorial Sloan-Kettering Cancer Center, New York, NY. We also acknowledge Sarah Nataraj (SUNY Downstate, Brooklyn, NY) for assistance in flow cytometric analysis and Dr Maja Nowakowski (SUNY Downstate, Brooklyn, NY) for guidance in fluorescent microscopy.
1. Schuchert MJ, Luketich JD. Barrett's esophagus — emerging concepts and controversies. J Surg Oncol 2007; 95:185–189.
2. Pondugula K, Wani S, Sharma P. Barrett's esophagus and esophageal adenocarcinoma in adults: long-term GERD or something else? Curr Gastroenterol Rep 2007; 9:468–474.
3. Souza RF, Krishman K, Spechler SJ. Acid, Bile, and CDX: the ABCs of making Barrett's metaplasia. Am J Physiol Gastrointest Liver Physiol 2008; 295:211–218.
4. Hu Y, Williams VA, Gellerson O, et al
. The pathogenesis of Barrett's esophagus: secondary bile acids upregulate intestinal differentiation factor CDX2 expression in esophageal cells. J Gastrointest Surg 2007; 11:827–834.
5. Kazumori H, Ishihara S, Rumi MAK, et al
. Bile acids directly augment caudal related homeobox Cdx2 expression in oesophageal keratinocytes in Barrett's epithelium. Gut 2006; 55:16–25.
6. Dvorak K, Fass R, Dekel R, et al
. Esophageal acid exposure at pH<2 is more common in Barrett's esophagus patients and is associated with oxidative stress. Dis Esophagus 2006; 19:366–372.
7. Krishnadath KK. Novel findings in the pathogenesis of esophageal columnar metaplasia or Barrett's esophagus. Curr Opin Gastroenterol 2007; 23:440–445.
8. Blanchard C, Wang N, Rothenberg ME. Eosinophilic esophagitis: pathogenesis, genetics, and therapy. J Allergy Clin Immunol 2006; 118:1054–1059.
9. Furuta GT, Liacouras CA, Collins MH, et al
. Eosinophilic esophagitis in children and adults: a systematic review and consensus recommendations for diagnosis and treatment. Gastroenterology 2007; 133:1342–1363.
10. Chebade M, Sampson HA. Epidemiology and etiology of eosinophilic esophagitis. Gastrointest Endoscopy Clin N Am 2008; 18:33–44.
11. Mulder DJ, Pacheco I, Hurlbut DJ, et al
. FGF9-induced proliferative response to eosinophilic inflammation in oesophagitis. Gut 2009; 58:166–173.
12. Gupta SK. Noninvasive markers of eosinophilic esophagitis. Gastrointest Endoscopy Clin N Am 2008; 18:157–167.
13. Blain SW, Scher HI, Cordon-Dardo C, et al
as a target for cancer therapeutics. Cancer Cell 2003; 3:111–115.
14. Philipp-Staheli J, Payne SR, Kemp CJ. p27Kip1: regulation and function of a haploinsufficient tumor suppressor and its misregulation in cancer. Exp Cell Res 2001; 264:148–168.
15. Chu IM, Hengst L, Slingerland JM. The Cdk inhibitor p27
in human cancer: prognostic potential and relevance to anticancer therapy. Nat Rev 2008; 8:253–267.
16. Kim SS, Meitner P, Konkin TA, et al
. Altered expression of Skp2, c-Myc and p27
proteins but not mRNA after H. pylori
eradication in chronic gastritis. Mod Pathol 2006; 19:49–58.
17. Kuzushita N, Rogers AB, Monti NA, et al
. p27kip1 deficiency confers susceptibility to gastric carcinogenesis in Helicobacter pylori
-infected mice. Gastroenterology 2005; 129:1544–1556.
18. Ellis FH, Xu X, Kulke MH, et al
. Malignant transformation of the esophageal mucosa is enhanced in p27
knockout mice. J Thorac Cardiovasc Surg 2001; 122:809–814.
19. Ellis FH, Loda M. p27
and Barrett's esophagus: a review. Dis Esophagus 2004; 17:113–117.
20. Frigas E, Loegering DA, Gleich GJ. Cytotoxic effects of the guinea pig eosinophil major basic protein on tracheal epithelium. Lab Invest 1980; 42:35–43.
21. Gleich GJ, Frigas E, Loegering DA, et al
. Cytotoxic properties of the eosinophil major basic protein. J Immunol 1979; 123:2925–2927.
22. Blain SW, Montalvo E, Massague J. Differential interaction of the cyclin-dependent kinase (Cdk) inhibitor p27
with cyclin A-Cdk2 and cyclin D2-Cdk4. J Biol Chem 1997; 272:25863–25872.
23. Assoian RK. Stopping and going with p27Kip1. Dev Cell 2004; 6:458–459.
24. Singh SP, Lipman J, Goldman H, et al
. Loss or altered subcellular localization of p27
in Barrett's associated adenocarcinoma. Cancer Res 1998; 58:1730–1735.
25. Francalanci P, De Angelis P, Minnei F, et al
. Eosinophilic esophagitis and Barrett's esophagus: an occasional association or an overlap disease? esophageal ‘double trouble’ in two children. Digestion 2008; 77:16–19.
26. Wolfsen HC, Hemminger LL, Achem SR. Eosinophilic esophagitis and Barrett's esophagus with dysplasia. Clin Gastroenterol Hepatol 2007; 5:A18.
27. Furuta GT, Nieuwenhuis E, Karhausen J, et al
. Eosinophils alter colonic epithelial barrier function: role for major basic protein. Am J Physiol Gastrointest Liver Physiol 2005; 289:G890–G897.
28. Milano F, Van Baal J, Buttar N, et al
. Bone morphogenic protein 4 expressed in esophagitis induces a columnar phenotype in esophageal squamous cells. Gastroenterology 2007; 132:2412–2421.