Results
A total of 60 patients were divided into three groups: 20 patients as a control group (who had normal esophageal mucosa with no esophagogastroduodenoscopic or histopathological lesions), 20 patients with EE without Barrett’s esophagus and microscopic adenocarcinoma and 20 patients with macroscopic and histopathological adenocarcinoma. The mean age ± SD of the control group was 51.5±11.3 years. The mean age ± SD of the EE group was 56.6±10.2 years. The mean age ± SD of the EAC group was 58.6±12.4 years.
For the three groups genetic analysis was used to investigate the expression of 84 key genes (Table I) encoding enzymes known or predicted to modify genomic DNA and histones to regulate chromatin accessibility and therefore gene expression.
Upregulated and downregulated genes in the EAC and control group are summarized in Table II. AURKA, AURKB, NEK6 were expressed at significantly higher levels in the EAC than in the control group. MBD2 was expressed significantly lower in the EAC than in the control group. Fig. 1 is a scatter plot of the log base 10 of the hybridization intensity of each gene in the two groups [x-axis, control group; y-axis, the EAC (group 1) group]. The middle line indicates a fold-change (2−ΔCt) of 1. The top and the bottom lines indicate the desired fold-change in gene expression threshold. Expression of 80 key genes was unchanged, showing no significant difference in expression between the two groups. The three points above the top line indicate upregulated (AURKA, AURKB, NEK6) genes. The one point under the bottom line represents a down-regulated (MBD2) gene.
Upregulated and downregulated genes in the EE and control group are summarized in Table III. Seven genes (AURKA, AURKC, HDAC9, NEK6, HDAC8, SETD5, SETD7) were upregulated and AURKA, AURKC, HDAC9, NEK6 were expressed at significantly higher levels in EE than in the control group. There were no downregulated genes in the two groups. Fig. 2 is a scatter plot of the log base 10 of the hybridization intensity of each gene in the two groups [x-axis, control group; y-axis, EE (group 2) group]. The middle line indicates a fold-change (2–ΔCt) of 1. The top and the bottom lines indicate the desired fold-change in gene expression threshold. Expression of 77 key genes was unchanged with no significant difference in gene expression between the two groups. The seven points above the top line represent upregulated (AURKA, AURKC, HDAC9, NEK6, HDAC8, SETD5, SETD7) genes.
Upregulated and downregulated genes in EAC and EE are summarized in Table IV. There was no significant difference in gene upregulation between the two groups. Two genes (MBD2 and MYSM1) were downregulated and MBD2 was significantly downregulated in EAC compared to the EE group. Fig. 3 is a plot of the log base 10 of the hybridization intensity of each gene in the two groups (x-axis represents group 1, EAC, and the y-axis shows group 2, EE). The middle line indicates a fold-change (2−ΔCt) of 1. The top and bottom lines indicate the desired fold-change in gene expression threshold. The expression of 82 key genes was unchanged and no significant difference in gene upregulation was noted between the two groups. The two points under the bottom line represent downregulated (MBD2, MSYM1) genes.
The NEK6 and AURKA genes were significantly upregulated in the EAC and EE groups compared to the control group.
The correlation between the AURKA gene and the stage of EAC is summarized in Fig. 4A. There were more patients in the stage 1 group than in the stage 2 and stage 3 groups (p<0.05).
The correlation between the NEK6 gene and the stage of EAC is summarized in Fig. 4B. Again, there were more patients classified in the stage 1 group than in stage 2 and stage 3 groups (p<0.05).
The correlation between expression of the AURKA gene and the Los Angeles classification of erosive esophagitis are summarized in Fig. 5A. Expression of the AURKA gene was found to be elevated dependent on the grade of EE according to the Los Angeles Classification.
The correlation between expression of the NEK6 gene and the Los Angeles classification of EE is summarized in Fig. 5B. The patients with overexpression of the NEK6 gene were more prevalent in the LA Grade B than LA Grade A group (p>0.05).
A clustergram analysis based on differentially expressed genes between the three groups is shown in Fig. 6. The cluster-gram creates a heat map with dendrograms to show genes that are co-regulated. The color saturation reflects the magnitude of the change in gene expression. Green squares represent lower gene expression in the experimental samples (ratios less than 1); black squares represent genes equally expressed (ratios near 1); red squares represent higher than control levels of gene expression (ratios greater than 1); gray squares indicate insufficient or missing data.
Discussion
Understanding the epigenetic structure of carcinomas provides important information on carcinogenesis. Expanding the information on the molecular biology of cancer may result in better follow-up of precancerous and cancer lesions (13,16). Epigenetic changes in stem cells provide important information concerning cancer aetiology, and epigenetic alterations such as gene expression have been used as biomarkers in recent years (17).
In the present study, three genes were overexpressed in the EAC group, with AURKA, AURKB and NEK6 being significantly more highly expressed than levels in the control group. Four genes (AURKA, AURKC, HDAC9, NEK6) were significantly more highly expressed in the EE group than levels in the control group. AURKA and NEK6 genes were significantly more highly expressed in the EE group and in the EAC group than levels in the control group.
Recent studies have shown that the aurora kinase family, polo-like kinase family and NIMA (never in mitosis gene A) kinase family control cell cycle (18,19). Aurora kinases are activated through autophosphorylation at the activation loop unlike most kinases within the cell (20) and NimA promotes mitotic chromosome condensation through phosphorylation of histone H3 at serine 10 and may compose the nuclear membrane division during mitotic exit (21).
Enzymes in the Aurora kinase family are encoded by the AURKA (also called AIK/ARK1/AURA/AURORA2/BTAK/MGC34538/STK15/STK6/STK7) gene, which is localized on 20q13.2 (22,23). These enzymes play very important roles in mitosis and meiosis for healthy cell proliferation. AURKA is a serine/threonine kinase acting as a regulator of centro-some function/duplication, mitotic entry, and bipolar spindle assembly (24). AURKA protein levels and kinase activity are low in the G1/S phase; accumulate during G2/M and decrease rapidly following mitosis (25). AURKA is an important kinase-encoding gene involved in centrosome duplication and distribution; its overexpression leads to centrosome amplification, chromosomal instability and aneuploidy in several cancer types (26,27). AURKA overexpression has been found in numerous tumor cells and tissues including gastric cancer, breast cancer, colorectal cancer, bladder cancer, pancreatic cancer, ovarian cancer, prostate cancer and esophageal squamous-cell carcinoma, esophageal adenocarcinoma and Barrett’s esophagus (27,28). Dar et al demonstrated overexpression of mitotic kinase encoding gene in upper gastrointestinal adenocarcinomas through the immunohistochemical analysis of 130 tumors. This overexpression was more prevalent in gastroesophageal junction adenocarcinomas and lower in esophageal, Barrett-related adenocarcinomas (BAS) than in antrum and body gastric adenocarcinomas. They also found that the expression of AURKA caused an anti-apoptotic effect in gastrointestinal cancer cells with drug-induced apoptosis in an in vitro model (27). Rugge et al found that AURKA immunostaining increased significantly along with the Barrett’s carcinogenesis, from Barrett’s mucosa even without metaplasia towards Barrett’s adenocarcinoma (24). AURKA appears to play an important role in the carcinogenesis of esophageal adenocarcinoma and will be an important target for surveillance, diagnosis, treatment and prognosis in Barrett’s esophagus. The positive relationship between Barrett’s esophagus and AURKA is important for our results since in this study we demonstrated that AURKA is upregulated in erosive esophagitis and the AURKA gene was found to be elevated dependent on the grade of erosive esophagitis based on the Los Angeles Classification. GERD can yield to complications such as erosive esophagitis and strictures; furthermore, it can cause Barrett’s esophagus, which can progress to adenocarcinoma (29). We believe that ascertaining whether the AURKA gene may be used as an early marker in erosive esophagitis toward the development of EAC is crucial.
NIMA is another gene found to be related to cell cycle dysfunction when it is overexpressed or underexpressed. NIMA (never in mitosis gene a)-related kinase 6 (NEK6; also called SID6-1512) is localized on chromosome 9q33-34 and is a serine/threonine kinase that belongs to the Neks (NIMA-related kinases) family, which has been implicated in mitosis control (30). Yin et al previously found that human Nek6 is required for metaphase-anaphase transition during cell cycle progression (31). It is believed that interfering with Nek6 function causes mitotic arrest and triggers apoptosis (32,33). The negative mutant form of Nek6 was found to induce spindle defects, abnormal chromosome segregation, mitotic arrest and apoptosis (34). Overexpression of Nek6 was shown in hepatocellular carcinoma as compared with the adjacent normal tissue as an evidence of its antiapoptotic effect (18). Overexpression of NEK6 was associated with histological grade, level of α feto protein and poor prognosis. NEK6 was shown to mediate human cancer cell transformation and was proposed as a potential cancer therapeutic marker in a previous study (34). Takeno et al stated that NEK6 is a potential marker of gastric cancer regardless of stage and since conventional staging of tumors are not adequate to predict individual prognosis, genetic analyses of tumor tissues may provide better opportunities to predict disease outcome for each individual and may even predict response to therapy (35). The authors selected seven focus genes showing a 2-fold change; four had not been previously evaluated for the association with gastric tumors. NEK6 was one these four new genes (35). They concluded that mapping of gene expression data on large sample numbers helped to identify two novel candidate genes, INHBA and NEK6, that are promising potential markers of gastric cancer. Nassirpour et al revealed that the protein level and kinase activity of Nek6 are highly elevated in a variety of malignant human cancers including breast, uterus, colon, stomach, ovary, lung, kidney, rectum, thyroid, cervix, prostate, pancreas, small intestine cancer cells, and knockdown of Nek6 resulted in reduction of tumors in a nude mouse xenograft model (34). They concluded that since inhibition of NEK6 specifically induces cell death in tumor cells and not in normal tissues, NEK6 inhibitors are a better therapeutic option with lower side effects than cytotoxic antitumor agents.
This is the first study investigating the impact of NEK6 in esophageal adenocarcinoma. Our data show the significant upregulation of NEK6 in erosive esophagitis and esophageal adenocarcinoma. NEK6 was more prevalent in samples with Los Angeles classification B than A in erosive esophagitis demonstrating that NEK6 and AURKA are more evident in more severe forms of esophagitis.
Erosive esophagitis is chronic damage of the esophagus caused by acid, pepsin and biliary salts. Environmental insults cause genetic and epigenetic alterations and they affect the expression of tumor-progenitor genes. Chronic injury is a major cause of cancer even though it is not inherently mutagenic (16). There are studies with large number of patients and long follow-up periods for GERD patients investigating whether they are at risk of developing esophageal adenocarcinoma. In a Swedish nationwide case-control study, gastroesophageal reflux and obesity were identified as strong and independent risk factors for esophageal adenocarcinoma. The risk increased with duration and severity of reflux symptoms and with increasing body mass index (36). Erichsen et al performed a nationwide cohort study in Denmark using data from 33,849 GERD patients and concluded that erosive but not non-erosive reflux disease has an impact on the development of adenocarcinoma emphasizing inflammation as an important factor in carcinogenesis (37).
In our study, in addition to AURKA and NEK6, HDAC9 and AURKC were significantly upregulated in EE compared to the control group. These genes were not expressed in EAC. Wu et al stated that AURKC was overexpressed in inflamed cervical tissue specimens and HDAC inhibitors are therapeutic for several inflammatory conditions (38,39). Do these genes have a role in addition to defect of the defense mechanism of the esophagus in patients with EE in gastroesophageal disease?
MBD2 was significantly downregulated in EAC compared to EE and the control group. MBD2 is a member of the MBD protein family. MBD2 binds to methylated promoter CpG islands and acts as a methylation-dependent transcriptional repressor (40). MBD2 has a role in the activation of methylated and unmethylated genes (42). Expression of MBD2 was particularly low in brain tumors, immune thrombocytopenia and in colorectal and gastric carcinomas (42–44) as found in our study. The reasons for the loss of MBD2 expression and the functional consequences are unknown (44). Thus, MBD2 should be studied further in relation to its association with esophageal adenocarcinoma.
This is the first study concerning the epigenetic chromatin histone modification in EE and EAC patients. Compared to the other genes investigated, AURKA and NEK6 were notably upregulated in EAC and EE. AURKA was proven to be associated with EAC in previous studies, and we found similar results for AURKA. This is the first study investigating the impact of NEK6 in esophageal adenocarcinoma, and our data revealed the significant upregulation of NEK6 in erosive esophagitis and esophageal adenocarcinoma. Our study is also the first study aiming to detect the presence of genetic upregulation in EE, a lesion which is not considered to be a precancerous lesion. These results pave the way for future studies with larger numbers of patients and longitudinal studies with longer follow-up periods. In a recent study (7), we found that low prevalence of Barrett’s esophagus was demonstrated in a Western Turkish population. Based on these data, we intend to explore expression of AURKA and NEK6 genes in the future at the national level, using another group with Barrett’s esophagus.
It is hoped that future studies may address the following questions: i) Is erosive esophagitis a precancerous lesion and once detected, is surveillance required? ii) Can AURKA and NEK6 be used as screening tests for esophageal adeno-carcinoma in erosive esophagitis? iii) What are the ranges of AURKA and NEK6 overexpression predictive of the prognosis and the outcome of antitumor therapy? iv) Can AURKA and NEK6 be used as therapeutic targets?
In conclusion, we demonstrated overexpression of AURKA and NEK6 in erosive esophagitis and esophageal adenocarcinoma in a Turkish population. Understanding the molecular pathophysiology of the disease will aid in elucidating the steps for diagnosis, therapy and prognosis. AURKA and NEK6 are two promising genetic markers for erosive esophagitis and esophageal adenocarcinoma.