SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References

LMX1A is epigenetically inactivated in cervical cancer. However, the expression and methylation status of LMX1A in gastric cancer tissues remains unknown. In the present study, we found that the expression of LMX1A was significantly decreased in gastric cancer tissues compared with normal tissues. A statistically significant inverse association was found between the LMX1A methylation status and expression of LMX1A in tumor tissues (= 0.008). Restoration of LMX1A induced cell apoptosis and suppressed anchorage-independent growth, suggesting LMX1A may be a potential biomarker for gastric cancer. (Cancer Sci 2011; 102: 361–366)

The LIM homeobox transcription factor 1, alpha (LMX1A) gene maps to 1q24.1 and is one of the group of LIM homeobox-containing genes that encode LIM-homeodomain (LIM-HD).(1,2) It is proven that LMX1A is a critical regulator of cell-fate decisions using genetic fate mapping in wild-type and LMX1A (−/−) mice.(3) LMX1A also plays a pivotal role in the mDA differentiation of human embryonic stem (hES) cells. In the developing cerebellum, loss of LMX1A completely abolishes roof plate induction in the spinal cord.(4) Recently, evidence for the role of LMX1A in cancers has been found. For example, LMX1A was identified as a metastasis suppressor in cervical cancer.(5) It is well known that hypermethylation of CpG islands in their promoter regions is an important mechanism for loss of function of several tumor suppressor genes.(6,7) Like most tumor suppressors, LMX1A is methylated more frequently in squamous cell carcinoma tissues than in the normal cervix.(8) It was also reported that the methylation of LMX1A genes correlated with recurrence and overall survival of ovarian cancer patients.(8) Methylation of LMX1A was observed in a colon cancer cell line (HCT-116), and was demethylated in the DKO cell line genetically disrupted in DNMT1 and DNMT3B.(9)

To characterize the role of LMX1A in gastric cancer, we examined a series of primary gastric cancers and control tissues for LMX1A expression and promoter methylation. Our results confirm frequent downregulation of LMX1A expression in primary gastric cancers and identify hypermethylation of the LMX1A gene as a common epigenetic aberration in these tumors.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References

Patient samples and cell lines.  Tumor samples were collected from surgical specimens from 50 patients with gastric cancer at the Department of Pathology, Minhang District Central Hospital, Shanghai, China. Non-tumor samples from the macroscopic tumor margin were isolated at the same time and used as the matched adjacent non-neoplastic tissues (>5 cm). Tissue samples were collected, immediately snap frozen in liquid nitrogen and stored at −80°C until RNA extraction. Informed consent was obtained for all samples and the institutional review board of Minhang District Central Hospital approved the study. All patients obtained a confirmed diagnosis of gastric carcinoma after resection. The five different established human gastric cancer cell lines used in the present study (AGS, SGC-7901, MKN28, MKN45, BCG-823) were maintained in DMEM with 10% FBS and 1% streptomycin/penicillin antibiotics. The gastric cancer cell lines were incubated for 96 h with 1 μmol/L 5-aza-2′-deoxycytidine (DAC; Sigma, St. Louis, MO, USA), with the medium changed every day. Treated cells were harvested for analysis 2 days after the procedure.

Combined bisulfite restriction analysis and bisulfite sequencing.  Genomic DNA (2 μg) was modified with sodium bisulfite using EpiTect Bisulfite kit (Qiagen, Valencia, CA, USA). The methylation status was analyzed by bisulfite genomic sequencing and combined bisulfite restriction analysis (COBRA). The fragment covering 39 CpG sites from the cystatin M promoter region was amplified from bisulfite-modified DNA. The primers for bisultite sequencing (BSP) were designed using Methprimer software. Amplified bisulfite-sequencing PCR products were cloned into pMD18-T simple vector (Takara, Kyoto, Japan); six independent clones were sequenced. COBRA was carried out by overnight digestion of the PCR product at 60°C with the restriction enzyme BstUI (New England BioLabs, Ipswich, MA, USA), which has the recognition sequence 5′-CGCG-3′. The resultant DNA fragments were electrophoresed on agarose gels and stained with ethidium bromide. The proportion of methylated (M) versus unmethylated (U) product (digested versus undigested) was quantitated using a densitometer to determine the density of methylation. The percent methylation was calculated as follows: M/(M + U) × 100.

Semiquantitative polymerase chain reaction.  Total RNA was extracted from gastric cancer or normal tissues using TRIZOL (Invitrogen, Carlsbad, CA, USA). Reverse transcription was carried out using Moloney Murine Leukemia Virus Reverse Transcriptase (M-MLV RT; Promega, Foster City, CA, USA). The polymerase chain reaction was performed using a SYBR Green kit following the manufacturer’s protocol (Takara). A 7900HT Fast Real-time PCR System (Applied Biosystems) was used for testing. Following the protocol of the manufacturer, the amount of LMX1A expression, normalized to a human actin endogenous reference is given by: 2−ΔΔCt. Real-time RT-PCR was repeated at least three times for each specimen and the mean was obtained.

Immunohistochemical analysis.  Human tissues were embedded in paraffin and cut into 5-μm sections onto glass slides. The pathological types of gastric cancer were determined by the Department of Pathology of Minhang District Central Hospital. These slides were treated by microwaving in 10 mM citric acid buffer for 20 min to retrieve antigenicity. After antigen retrieval, the peroxidase activity was blocked by 3% H2O2 for 10 min and the sections were incubated with 10% normal goat, followed by incubation with anti-LMX1A antibodies (1:100; Sigma) at 4°C overnight. The sections were then incubated with horseradish peroxidase-labeled goat anti-mouse antibodies (Dako) for 30 min at room temperature, and the signal was visualized with diaminobenzidine-chromogen. The slides were then counterstained with Harris Hematoxylin (Sigma). The sections were scored as previously described.(6) The following categories were used for scoring: intensity of staining, none (0), mild (1), moderate (2), and strong (3); percentage of positive staining, <5% (0), 5–25% (1), 25–50% (2), and >50% (3). Combining the intensity and percentage of staining resulted in the following score: 0–1, negative (−); and 2–6, positive (+).

Construction of plasmids and stable cell line generation.  For construction of pCMV4-flag-LMX1A, the LMX1A cDNA was generated by reverse transcription PCR and ligated into pCMV-flag vector (Invitrogen). For transfection experiments, the AGS cells were plated into six-well plates 24 h before transfection. The cells were transfected with 5 μg/well of empty pCMV-flag or pCMV4-flag-LMX1A using Superfect (Qiagen) according to the manufacturer’s instructions. For 48 h after transfection, the cells were passaged at 1:5 and cultured in medium supplemented with G418 at 500 μg/mL for 4 weeks. One AGS clone overexpressing LMX1A (AGS/LMX1A) was selected for further study, with LMX1A expression verified by western blot. As a control group, cells stably transfected with an empty vector pCMV-flag were also generated (AGS/vector).

Western blot analysis.  Protein from treated cell lines was extracted by mammalian protein extraction reagent (Pierce, Rockford, IL, USA) supplemented with protease inhibitors cocktail (Sigma). Fifty-microgram protein samples were resolved by 10% SDS-PAGE and then transferred to PVDF membranes. Autoradiograms were quantified by densitometry (Quantity One software; Bio-Rad, Hercules, CA, USA). Actin-specific antibody was used for the loading control. Mouse monoclonal anti-flag (1:1000; Sigma), and mouse monoclonal anti-actin (1:1000; Abcam) were used.

Cell growth assay and wound healing assay.  AGS/LMX1A and AGS/vector cells were seeded into 96-well plates. The Cell Counting Kit-8 (Dojindo, Kumamoto, Japan) was used to determine relative cell growth. For the wound healing assay, AGS/LMX1A and AGS/vector cells were seeded at 5 × 104 in six-well plates, resulting in a confluent monolayer, and maintained in serum-free media. Each well of cells was scratched with the tip of a 200 AL pipette tip. Forty-eight hours following the scratch, the extent of “wound healing” was observed microscopically.

Evaluation of apoptosis and cell cycle.  Apoptosis was detected by flow cytometric (FCM) analysis of annexin V staining. Annexin V–FITC versus propidium iodide (PI) assay was performed as previously reported.(10) Briefly, adherent cells were harvested and suspended in the Annexin-binding buffer (1 × 106 cells/mL). Thereafter, cells were incubated with Annexin V–FITC and PI for 15 min at room temperature in the dark and immediately analyzed using flow cytometry. For the cell cycle assay, the cells were harvested by trypsinization, washed three times with ice-cold PBS and fixed with 70% ethanol overnight at 4°C. The fixed cells were rehydrated in PBS and subjected to PI/RNase staining followed by fluorescence-activated cell sorter scan (FACS) analysis (Becton Dickinson, Mountain View, CA, USA).

Anchorage-independent growth.  For the anchorage-independent growth assay, 1 × 103 cells were plated in 0.3% low melting point agar/growth medium onto 6 cm dishes with a 0.6% agar underlay. After 4 weeks, colonies that were >1 mm in diameter were counted.

Tumorigenecity assay.  The tumors were established by s.c. injection of 1 × 106 AGS/LMX1A and AGS/vector cells into BALB/c nude mice. Ten mice were included in each group in all experiments. Tumors sizes were estimated using the equation V = 4/3π × L/2 × (W/2)2, where L is the mid-axis length and W is the mid-axis width.

Statistical analysis.  Fisher’s exact tests were performed to evaluate the significance of the differences between the frequencies of the LMX1A promoter hypermethylation status of the various tissue categories. A P value < 0.05 was considered statistically significant.

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References

Methylation analysis of the LMX1A gene in gastric cancer tissues.  Methylation of the promoter region of LMX1A in gastric cancer tissues was determined by COBRA. The COBRA analysis is a well-characterized method for DNA methylation studies and has been done largely before. The area of the CpG-rich region around the promoter of the LMX1A gene, which spanned 39 CpG sites, and the BstUI recognition sequences are shown in Figure 1a. Promoter methylation of the LMX1A gene was frequent in the gastric cancer tissues, with 41 of 50 (82%) samples positive. For the paired adjacent non-tumor tissue samples, 13 of 50 samples were positive (26%) (< 0.001). Representative examples of the gel analysis of COBRA are shown in Figure 1b. In this COBRA study, the proportion of methylated (M) versus unmethylated (U) products (digested versus undigested) was semiquantitated using a densitometer. The percent methylation was calculated as follows: M/(M + U) × 100.

image

Figure 1.  (a) A map of the CpG islands in relation to the promoter of the LMX1A. The locations of sense and antisense primers used for bisulfite-sequencing PCR are indicated with brackets. The cutting sites of BstUI are indicated with underline. (b) Representative combined bisulfite restriction analysis (COBRA) results of the LMX1A promoter methylation in gastric cancer tissues. The PCR products from bisulfite treated DNA were digested with BstUI, which generated digested bands on full or partial digestions. Digested fragments correspond to methyalted DNA. Case numbers are shown on top. Normal tissue, 3, was proven to be fully unmethylated by bisulfate genomic sequencing, therefore each experiment was conducted with DNA from N3 and DNA treated with SssI methyltransferase as negative and positive controls. (c) The line within each box represents the median ΔΔCt value, ΔΔCt = ΔCt(Actin)−ΔCt(LMX1A). The upper and lower edges of each box represent the 75th and 25th percentile, respectively. The upper and lower bars indicate the highest and lowest values determined, respectively. P < 0.001, two-sided t test. (d) An inverse relationship between LMX1A promoter methylation and mRNA expression in gastric cancer tissues.

Download figure to PowerPoint

Association of the LMX1A promoter methylation with transcriptional gene silencing.  LMX1A mRNA expression levels in each of the gastric cancer tissues were assessed relative to the normal tissues using real-time RT-PCR analysis. The results showed that LMX1A was significantly downregulated in tumor tissues compared with normal tissues (< 0.001) (Fig. 1c). To elucidate whether aberrant methylation of LMX1A is associated with loss of LMX1A expression, the samples were divided into two different groups, according to the percentage of methylation obtained from COBRA, as <50% and >50%. Correlation of the promoter methylation with LMX1A expression showed a significantly lower expression level in the tumors with >50% methylation compared with the tumors with <50% methylation (Fig. 1d). To further validate this result, we detected the expression of LMX1A in the protein level by using immunohistochemistry. Overall, LMX1A was absent in 35 of 50 carcinomas (70%) and 5 of 50 (10%) in normal gastric cancer samples. Representative examples of LMX1A protein expression in gastric cancer samples are shown in Figure 2. Six of the nine unmethylated carcinoma samples (66.7%) demonstrated positive staining and 32 of 41 methylated carcinoma samples (78%) showed loss of expression of LMX1A. Thus, the immunostaining results were strongly correlated (P = 0.008) with LMX1A methylation status (Table 1). These results suggested that LMX1A promoter methylation correlated with loss of LMX1A expression.

image

Figure 2.  Immunohistochemical staining for LMX1A with anti-LMX1A in cancerous and normal tissues. Abundant expression of LMX1A was observed in normal mucosa, and weak expression of LMX1A was detected in the primary cancer tissue.

Download figure to PowerPoint

Table 1.   LMX1A methylation status and protein expression in gastric cancer samples
 MethylationNonmethylationP value
  1. Statistical analysis was evaluated with Fisher’s exact tests.

Protein expression
+9 (22%)6 (66.7%) 
32 (78%)3 (33.3%)0.008

Restoration of LMX1A expression by the demethylation agent.  To confirm that this loss of expression was because of the LMX1A promoter methylation, five gastric cancer cell lines were incubated in the presence or absence of the DAC, and the methylation status and LMX1A mRNA expression were analyzed by bisulfite DNA sequencing and real-time RT-PCR, respectively. As shown in Figure 3a, most CpG dinucleotides were methylated in the gastric cancer cell lines. We examined the role of methylation in the silencing of LMX1A. To confirm that CpG methylation is indeed responsible for the silencing of LMX1A, we treated these heavily methylated and silenced cell lines with DAC, a methyltransferase inhibitor. As shown in Figure 3b, LMX1A expression was markedly induced after the treatment of DAC in all cell lines. Bisulfite DNA sequencing of the gastric carcinoma cell lines confirmed the promoter methylation status with or without treatment of DAC (Fig. 3c).

image

Figure 3.  (a) Methylation patterns of individual bisulfite-sequenced clones of the LMX1A promoter in gastric cancer cell lines. Black and white areas represent the percentage of methylated and unmethylated CpG sites, respectively, out of the colonies sequenced for each case. (b) Expression of LMX1A in gastric cancer cell lines treated with or without demethylation agent DAC as determined by RT-PCR. (c) An illustrative fragment of the sequencing electropherogram is shown for AGS cells treated with or without DAC; the CpG sites are underlined.

Download figure to PowerPoint

Exogenous expression of LMX1A alters gastric cancer cell growth, apoptosis and cell cycle.  The more frequent silencing of LMX1A by methylation in gastric cancer than in adjacent gastric mucosa suggests a potential tumour suppressor role of this gene. To assess whether LMX1A affects the biological behavior of gastric cancer cells, we established a LMX1A expression plasmid and AGS/LMX1A, which stably express LMX1A. The expression of LMX1A in AGS/LMX1A cells was confirmed by western analysis (Fig. 4a). Cell growth was analyzed using Kit-8 assays, which were performed in triplicate in 96-well plates. Compared with the AGS and AGS/vector cells, the proliferation of AGS/LMX1A cells was inhibited (Fig. 4b) (< 0.05).

image

Figure 4.  (a) Western blots of total protein from AGS, AGS/vector and AGS/LMX1A were probed with anti-flag (top) and anti-actin (bottom, loading control). (b) Cellular proliferation was measured by CCK-8 assay in AGS, AGS/vector and AGS/LMX1A cells. Points represent the mean ± standard deviation (SD) of three independent experiments. (c) Exogenous expression of LMX1A induces cell cycle G1 arrest in gastric cancer cells. Cell cycle progression of AGS, AGS/vector and AGS/LMX1A cells was determined by FACS analysis. Data are representative of three independent experiments. (d) The apoptotic percentage of the AGS, AGS/vector and AGS/LMX1A cells was detected by flow cytometric analysis. The values above each bar represent the fraction of annexin V+/PI− and annexin V+/PI+. Values are expressed as the mean ± SD of three replicate experiments.

Download figure to PowerPoint

To further understand the mechanisms by which cell proliferation is affected, flow cytometry was performed to analyze the cell cycle phase distribution of these cells. The cell cycle progression of AGS/LMX1A cells was stalled at the G1 phase with a significant decrease in S and G2/M phases compared with AGS and AGS/vector cells (Fig. 4c). We sought to evaluate whether the effect of cell growth was related to apoptosis. Seventeen ± 4% of AGS/LMX1A cells were stained positively with Annexin V as determined by FCM, whereas there were no significant changes in AGS or AGS/vector cells (Fig. 4d) (< 0.05).

Exogenous expression of LMX1A reduces cell motility, colony formation in vitro and tumor growth in mice.  Following growth in serum-free media, confluent dishes of AGS/vector and AGS/LMX1A clones were scratched with a 200 AL pipette tip. Twenty-four hours later, AGS/vector cells had fully invaded the resulting “wound,” whereas AGS/LMX1A had not noticeably moved into the scratch region (Fig. 5a).

image

Figure 5.  (a) Exogenous expression of LMX1A results in reduced migratory potential of AGS cells in a wounding assay. (b) Examples of a colony assay after 4-week selection. Bottom left, detailed images from the top panels (magnification, ×400). Cells were stained with Giemsa and the colonies were counted. The graph represents the means ± standard deviations of three independent experiments. (c) Effect of LMX1A on the growth of AGS cells in nude mice. Tumors in nude mice 1 month after injection of 106 AGS/vector or AGS/LMX1A cells are shown.

Download figure to PowerPoint

We further showed the growth inhibitory features of LMX1A reintroduction in a colony formation assay and nude models. We observed that LMX1A re-expression revealed tumor suppressor activity; there was lower colony formation compared with the control vector (< 0.05) (Fig. 5b). We examined the ability of AGS/LMX1A cells to form tumors compared with AGS/vector. All mice injected with AGS/vector cells and 5 of 10 mice injected with AGS/LMX1A cells developed palpable tumors. Tumors derived from AGS/vector cells were 2–2.5 times larger than that derived from AGS/LMX1A cells (< 0.001) (Fig. 5c).

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References

Gastric cancer is the second most frequent cause of death from cancer in the world.(11,12) The molecular mechanisms underlying gastric carcinogenesis are not fully understood, although several factors, including a low intake of vegetables and fruit, as well as Helicobacter pylori infection, have been found to be related with gastric carcinogenesis.(13) In addition to genetic changes, epigenetic alterations such as DNA methylation and histone modifications can result in heritable gene silencing without changes to genetic sequences and are recognized as important causes of cancer.(14,15)

LMX1A is one of the group of LIM homeobox-containing genes that encode LIM-homeodomain (LIM-HD),(1) and is known to play an important role in developmental events. Although much remains unknown about the function of LMX1A in cancers, LMX1A has been identified as a metastasis suppressor in cervical cancer.(5) Like most tumor suppressors, hypermethylation of the LMX1A promoter is found in ovarian and cervical cancers.(5,8) It was also reported that the methylation of LMX1A genes correlated with recurrence and overall survival of ovarian cancer patients.(8) Methylation of LMX1A was observed in a colon cancer cell line (HCT-116), and was demethylated in the DKO cell line genetically disrupted in DNMT1 and DNMT3B.(9)

At present, the LMX1A methylation status remains unclear in gastric cancer. Here, we analyzed DNA methylation in the LMX1A gene in 50 patients with gastric cancer. We found the promoter methylation of the LMX1A gene was frequent in the gastric cancer tissues compared with the paired adjacent non-tumor tissue samples, indicating that it is a comment event in gastric cancer. The results of immunohistochemistry showed that LMX1A was significantly downregulated in tumor tissues compared with normal tissues. Correlation of the promoter methylation with LMX1A expression showed a significantly lower expression level in the tumors with hypermethylation compared with the tumors with low methylation. These results suggested that the LMX1A promoter methylation correlated with a loss of LMX1A expression.

Analysis of gastric cancer cells indicated the LMX1A promoter methylation correlated with a loss of LMX1A expression and LMX1A expression was restored in cells cultured on the DNA demethylating agent DAC. From a functional standpoint, we next wanted to examine whether epigenetic inactivation of LMX1A inhibited growth suppression in gastric cancer cells. Ectopic expression of LMX1A in gastric cancer cell line, AGS, induced cell apoptosis, and suppressed cell proliferation and anchorage-independent growth in vitro. Exogenous expression of LMX1A reduces cell motility, colony formation in vitro and tumor growth in mice. In cervical cancer, LMX1A inhibits cancer invasion and metastasis partly through inhibition of the epithelial–mesenchymal transition (EMT).(5) The EMT was originally defined as the morphogenetic change of epithelial cells to fibroblast-like cells, and is considered to be a potential mechanism of cancer progression.(16) In our future studies, we will try to figure out the exact mechanism by LMX1A fulfilling its tumor suppressor role in gastric cancer.

In conclusion, epigenetic inactivation of LMX1A is a common event, and forced expression of LMX1A-induced cell apoptosis, and suppressed anchorage-independent growth, suggesting LMX1A may be a potential biomarker for gastric cancer patients.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References

The authors thank Yongping Zhang and Mingming Qiao for cell preparation. This research was supported by National Natural Science Foundation of Shanghai (grant no. 10ZR1426300).

Disclosure Statement

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References

No conflicts of interest exist for any of the authors.

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References