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
  9. Supporting Information

Whole genome-scale integrated analyses of exon array and array-comparative genomic hybridization are expected to enable the identification of unknown genetic features of cancer cells. Here, we evaluated this approach in 22 gastric and colorectal cancer cell lines, focusing on protein kinase genes and genes belonging to the cadherin–catenin family. Regarding alternative splicing patterns, several cancer cell lines predominantly expressed isoform 1 of protein kinase A catalytic subunit beta (PRKACB). Paired gastric cancer specimens demonstrated that isoform 1 of PRKACB was a novel cancer-related variant transcript in gastric cancers. In addition, the exon array analysis clearly identified exon 3 or exon 3–4 skipping in catenin beta 1, a short intron insertion with exon 9 skipping in CDH1, and a deletional transcript of CDH13. These abnormal transcripts were shown to have arisen from small genomic deletions. Meanwhile, an integrated analysis of 11 gastric cancer cell lines revealed that four cell lines amplified fibroblast growth factor receptor 2, with truncated forms observed in two of the cell lines. Gene amplification, and not the truncated form, was found to determine the sensitivity to a fibroblast growth factor receptor inhibitor, indicating that our cell line panel might be useful for cell-based evaluations of specific inhibitors. Using an integrated analysis, we identified several abnormal transcripts and genomic alterations in gastric and colorectal cancer cells. Our approach might enable genetic changes to be identified more efficiently, and the present results warrant further investigation using clinical samples and integrated analyses. (Cancer Sci 2012; 103: 221–227)

Recent advancements in the field of array technology over the past several years have enabled alternative splicing and abnormal transcripts to be explored at the whole genome level. Accordingly, the identification of cancer-related transcripts has been intensively investigated at the whole genome level in lung cancer, glioblastoma, thymic tumors and colorectal cancer using exon arrays, yielding novel splice variants, hyper-splicing signatures and overexpressed variants.(1–4) In contrast, whole genome array-comparative genomic hybridization (array-CGH) has enabled a higher resolution and probe density and is expected to make possible the identification of relatively small genomic copy number changes caused by gene amplifications and deletions that were previously undetectable. The array-CGH tool is also producing promising data for cancer research and the diagnosis, classification, and outcome prediction of different malignancies.(5)

In relation to abnormal transcripts caused by genomic changes, mutations and small genomic deletions can disrupt or create splice sites, change other consensus sequences, or affect splicing silencers or enhancers, such as ATM and BRCA1.(6) The detection of these transcripts containing exon-level abnormalities is thought to be difficult using conventional methods, and unknown genetic changes might exist. Therefore, we hypothesized that the use of only an exon analysis or an array-CGH might be insufficient, and that a simultaneous analysis might contribute to the further identification of abnormal transcripts in cancer cells.

In the present study, we performed a whole genome exon array and array-CGH analysis in gastric and colorectal cancer cells with the objective of identifying abnormal transcripts at the exon level.

Materials and Methods

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

Cell lines and cultures and sample preparations.  The method used in this section is described in Data S1.

Clinical samples.  The endoscopic biopsy samples were obtained from gastric cancer and paired non-cancerous lesions of gastric mucosa. The samples were immediately placed in an RNA stabilization solution (Isogen, Nippongene, Tokyo, Japan) and stored at −80°C. This analysis was approved by the Institutional Review Board of the National Cancer Center Hospital, and written informed consent was obtained from all the patients.

Exon array.  Two micrograms of total RNA was used as the starting material. rRNA was first removed using the RiboMinus Human/Mouse Transcriptome Isolation Kit (Invitrogen, Carlsbad, CA, USA) and cDNA synthesis was performed using the GeneChip WT cDNA Synthesis Kit (Affymetrix, Santa Clara, CA, USA). The cDNA was fragmented and labeled with biotin using the GeneChip WT Terminal Labeling Kit (Affymetrix). Biotinylated targets were hybridized onto a GeneChip Human Exon 1.0 ST Array (Affymetrix) according to the manufacturer’s instructions. The array was washed and stained in Fluidics Station 450 (Affymetrix) and scanned to generate a CEL file using the GeneChip Scanner 3000 (Affymetrix) and GeneChip Operating Software version 1.4.

Array-based comparative genomic hybridization.  The Genome-wide Human SNP Array 6.0 (Affymetrix) was used to perform array-CGH on genomic DNA from each of the colorectal cancer cell lines and MKN74 according to the manufacturer’s instructions. The GeneChip Human Mapping 250K Nsp Array (Affymetrix) was used to perform array-CGH on genomic DNA from each of the gastric cancer cell lines, with the exception of MKN74, according to the manufacturer’s instructions. The method is described in detail in Data S1.

RT-PCR, sequencing and colony formation assay.  The method used in this section is described in Data S1.

In vitro growth inhibition assay.  The growth-inhibitory effects of the fibroblast growth factor receptor (FGFR) inhibitor PD173074 (Sigma-Aldrich, St. Louis, MO, USA) were examined using an MTT assay, as previously described.(7)

siRNA study.  The methods used in this section have been previously described.(8)

Statistical analysis.  All exon array data were analyzed using Partek Genomic Suite 6.4 software (Partek, St. Louis, MO, USA). The robust multi-array average algorithm was used for the exon-level intensity analysis. Exon-level data were filtered to include only those probe sets that were included within the “Core Meta-Probeset.” The software calculated the differences between the cancer cell lines and the normal samples for each probe set on the transcript. The alt-splice score was the minimum P-value from the Z-test of each probe set’s difference against the remaining probe sets. A low Alt-splice score indicates that at least one probe set behaved differently from the rest. The Alt-Splice score was used to screen the transcript variants. In the array-CGH analysis, sample-specific copy number changes were analyzed using Partek Genomic Suite 6.4 software (Partek) and Affymetrix Genotyping Console ver.3.0.2 (Affymetrix) for the 40 samples of Human HapMap JPT.

Results

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

Cβ1 isoform of protein kinase A catalytic subunit beta is a novel cancer-related variant transcript.  To evaluate the use of an integrated analysis involving an exon array and an array-CGH, we applied a focused-gene approach against cancer-related and relatively annotated genes, including 480 protein kinase genes and 68 cadherin–catenin genes in gastric and colorectal cancer cell lines.

Cyclic adenosine monophosphate-dependent protein kinase is a signaling molecule that is important for a variety of cellular functions.(9) Alternatively spliced transcript variants encoding distinct isoforms have been found in protein kinase cAMP-dependent catalytic beta (PRKACB).(10,11) Interestingly, an exon analysis showed that isoform 1 of PRKACB (Cβ1 isoform) are frequently expressed in gastric and colorectal cancer cell lines, but not in normal mucosa (Fig. 1A). Among colorectal cancers, WiDr, LoVo and normal colonic mucosa predominantly expressed the Cβ2 isoform, whereas the CaR-1, DLD-1, CoCM-1, RCM-1, OUMS-23 and COLO320 DM cells expressed the Cβ1 isoform. Meanwhile, the HCC56, COLO201 and SW837 cell lines expressed very low levels. Similar expression patterns were observed in gastric cancer and normal gastric mucosa. RT-PCR confirmed these results (Fig. S1A). The RT-PCR study and the densitometrical analysis confirmed that the Cβ1 isoform was predominantly expressed in 32 gastric cancer specimens, compared with paired non-cancerous mucosa specimens in RT-PCR (P = 0.0012, Figs 1B and S1B). The expression of the Cβ2 isoform of PRKACB was not significantly different between the cancer and paired gastric mucosa specimens (P = 0.21). These results demonstrated that the Cβ1 isoform of PRKACB is a novel cancer-specific variant transcript in gastric cancer. We examined the biological functions of the Cβ1 isoform of PRKACB using specific siRNA for Cβ1 in DLD-1 and CaR-1 cell lines. The knockdown of the Cβ1 isoform significantly reduced cellular proliferation in both the DLD-1 and CaR-1 cell lines (Fig. 1C). In addition, a colony assay revealed that the knockdown of the Cβ1 isoform significantly reduced colony formation in both cell lines (Fig. 1D). These results suggest that the Cβ1 isoform, which is overexpressed in cancer cells, is involved in cellular proliferation in cancer cells. Because a recent study has demonstrated that specific PRKACB isoforms can properly recruit activated p75NTR to lipid rafts and determine p75NTR bioactivity, these different isoforms of PRKACB in cancer and normal cells might occur during carcinogenesis and might perform specific biological functions.(12)

image

Figure 1.  Cancer-related alternative splicing variants of protein kinase cAMP-dependent catalytic beta (PRKACB). (A) Exon array analysis showing the exon level mRNA expressions of PRKACB in 22 gastric and colorectal cancer cell lines and one gastric and one colorectal normal mucosa specimen. The log2-transformed values indicate the expression levels of each exon. The schematic diagram in the upper panel shows each exon of PRKACB (Cβ1 and Cβ2 isoforms). Note that normal gastric and colonic tissues and some cancer cell lines predominantly expressed the Cβ2 isoform, but several cancer cell lines predominantly expressed the Cβ1 isoform. p1–2: primer set. GC: 11 gastric cancer cell lines and one normal gastric mucosa specimen. CRC: 11 colorectal cancer cell lines and one normal colonic mucosa specimen. Cβ1, Cβ2: isoforms of PRKACB. (B) RT-PCR analysis for Cβ1 and Cβ2 isoforms of PRKACB in 32 paired gastric cancer and non-cancerous gastric mucosa specimens. (C) Knockdown of Cβ1 isoform of PRKACB using specific siRNA for Cβ1 and resulting cellular proliferation in DLD-1 and CaR-1 cell lines. Two different sequences of Cβ1-targeting siRNA (CB1#1 and CB1#2; 10 nM each) were used. Con-si, control siRNA. (D) Evaluation of colony formation in siRNA-treated DLD-1 and CaR-1 cell lines.

Download figure to PowerPoint

Identification of truncated transcripts caused by small genomic deletions.  Alterations of the cadherin–catenin cell adhesion system are considered to be a cause of gastric and colorectal cancers; therefore, we analyzed genes involved in this system. We found a novel truncated transcript of the CDH13/cadherin 13/H-cadherin gene. An exon analysis of CDH13 showed that the expression levels of exons 3–5 were extremely downregulated in COLO201 cells compared with other cell lines (Fig. 2A). Next, using array-CGH analysis, we evaluated the CDH13 locus and identified a small genomic deletion (approximately 0.4 Mbp) involving exons 3–5 in COLO201 cells (Fig. 2B). RT-PCR and sequencing confirmed the truncated transcript (Figs 2C and S2). Genomic PCR of exon 4 of CDH13 revealed a heterozygous deletion (data not shown).

image

Figure 2.  Truncated transcripts of CDH13 caused by small genomic deletion. (A) Exon array analysis showing the exon level mRNA expressions of CDH13 in COLO201 (lacking exons 3–5), DLD-1 and a normal colonic mucosa specimen. The log2-transformed values indicate the expression levels of each exon. The schematic diagram in the upper panel shows each exon of CDH13. p1: primer set. (B) Array-comparative genomic hybridization analysis for CDH13 locus in 11 colorectal cancer cell lines. The loss of genomic copy number is shown as bars extending toward the minus side of the baseline. (C) Detection of truncated mutant transcripts using RT-PCR for COLO201 (lacking exon 3–5), DLD-1 (normal) and normal colonic mucosa (normal). The schematic diagram shows each exon of CDH13. p1: primer set.

Download figure to PowerPoint

Mutations in the CDH1/E-cadherin gene are a well-documented cause of hereditary diffuse gastric cancer.(13) The exon analysis easily detected both the downregulation of exon 9 and an intron insertion in HSC-44 and its subline 44As3 (Fig. 3A). RT-PCR and sequencing confirmed these abnormal transcripts (Figs 3B and S3). Genomic PCR showed a 10-base deletion at the exon–intron boundary of the CDH1 gene (exon 9 and intron 9) in HSC-44 and 44As3 (data not shown). Meanwhile, the exon analysis clearly demonstrated the presence of a truncated catenin beta 1 (CTNNB1) transcript lacking exons 3–4 as a result of a small genomic deletion in HSC-39 cells (Fig. 4A). In addition, we found another type of truncated CTNNB1 transcript lacking exon 3 in HCC56. RT-PCR confirmed the deleted transcript (Fig. 4B). Genomic PCR demonstrated that the deletion of these transcripts was caused by a homozygous genomic deletion (data not shown). These results indicated that this integration analysis could detect exon-level mRNA abnormalities involving genomic deletions.

image

Figure 3.  Identification of mutant transcripts with a short intron insertion and exon skipping in CDH1/E-cadherin. (A) Exon array analysis showing the exon level mRNA expressions of CDH1 in 22 gastric and colorectal cancer cell lines and one gastric and one colorectal normal mucosa specimen. The log2-transformed values indicate the expression levels of each exon. The schematic diagram in the upper panel shows each exon of CDH1. p1–3: primer set. GC: 11 gastric cancer cell lines and one normal gastric mucosa specimen. CRC: 11 colorectal cancer cell lines and one normal colonic mucosa specimen. Note that exon 9 expression was downregulated and intron expression was observed in HSC-44 and 44As3 cells. int, intron. (B) Detection of mutant transcripts using RT-PCR and microfluidics-based electrophoresis for HSC-44 and 44As3 (lacking exon 9 or short intron insertion) and other control cell lines. The schematic diagram shows each exon of CDH1. p1–3: primer set.

Download figure to PowerPoint

image

Figure 4.  Identification of truncated mutant transcripts of CTNNB1. (A) Exon array analysis showing the exon level mRNA expression of CTNNB1 in 22 gastric and colorectal cancer cell lines and one gastric and one colorectal normal mucosa specimen. The log2-transformed values indicate the expression levels of each exon. The schematic diagram in the upper panel shows each exon of CTNNB1. p1–3: primer set. GC: 11 gastric cancer cell lines and one normal gastric mucosa specimen. CRC: 11 colorectal cancer cell lines and one normal colonic mucosa specimen. (B) Detection of truncated mutant transcripts using RT-PCR for HSC-39 (lacking exons 3–4), HCC56 (lacking exon 3) and other control cell lines. p2 primer were used as a positive control. The schematic diagram shows each exon of CTNNB1. p1–3: primer set. (C) DNA copy number analysis for major oncogenes and tumor suppressor genes in 11 gastric and 11 colorectal cancer cell lines using array-comparative genomic hybridization analysis. Gene amplification (≥10 copies and 10–3.6 copies) and deletion (≤0.5 copies and 1.2–0.5 copies) are shown by the indicated colors.

Download figure to PowerPoint

Whole genome analysis of alternatively spliced transcript variants and gene copy number profiles.  Next, we screened for cancer-specific spliced transcripts in each cancer cell line and compared the results with those for a normal mucosa sample using exon array data obtained at the whole genome level (Data S2 and S3). Fourteen and ten transcripts were identified as recurrently altered genes observed in more than six cell lines among the 11 gastric cancer cell lines and the 11 colorectal cancer cell lines, respectively (Data S4). The whole genome gene copy number profiles were also examined for 22 cancer cell lines (Data S5, Figs S4 and S5). In addition, we analyzed and identified overlapping genes between the results of an exon array (Alt-Splice Score) and an array-CGH (copy number changes) analysis at the whole genome level. The identified genes are listed according to the number of overlapping cell lines and are shown in Table 1. In the list, the FGFR2 gene tops the list of overlapping genes, suggesting that the results of the whole genome level approach were consistent with those of the focused-gene approach. Interestingly, the second-ranked gene, KIAA1797, was deleted in approximately 30% of the genomic region and expressed a truncated transcript, because the KIAA1797 gene locus was located at the end of the deleted region, including CDKN2A, in SNU-16 cells (data not shown). However, some genes identified using the focused-gene approach, including CDH1, CDH13 and CTNNB1, were not selected using the whole genome level analysis. These results indicate that an integrated analysis of the exon array and an array-CGH analysis might be better performed using both a whole genome level approach and a focused-gene approach.

Table 1.   List of overlapping genes with altered exon expressions and copy numbers
Gene symbolNameRefSeqTranscript IDNo. of cell linesAlt-splice scoreCopy numberCancer
  1. No. of cell lines, overlapping cell lines with altered exon expressions (Alt-Splice Score < 0.001) and copy numbers (amplification, >3.6 copies and deletion, <1.2 copies). Amp, gene amplification; CRC, colorectal cancer; Del, gene deletion; GC, gastric cancer.

FGFR2Fibroblast growth factor receptor 2NM_022970331004130.00011AmpGC
KIAA1797KIAA1797NM_017794316460130.00053DelGC
ASAP1ArfGAP with SH3 domain, ankyrin repeat and PH domain 1NM_018482315342820.00006AmpGC
ATP8A1ATPase, type 8A, member 1NM_006095276737820.00021DelGC
CNDP2CNDP dipeptidase 2NM_018235379376020.00051DelGC
DENND4CDENN/MADD domain containing 4CNM_017925316422120.00041DelGC
GAS6Growth arrest-specific 6NM_000820350282920.00047AmpGC
ITPR2Inositol 1,4,5-trisphosphate receptor, type 2NM_002223344815220.00019Amp/DelGC
NFRKBNuclear factor related to kappaB binding proteinNM_006165339807620.00040Amp/DelGC
PCDHGC5Protocadherin gamma subfamily C, 5NM_018929283253320.00012DelGC
PCM1Pericentriolar material 1NM_006197308781320.00011DelGC
PEX1Peroxisomal biogenesis factor 1NM_000466306119120.00025AmpGC
PHLPPPH domain and leucine rich repeat protein phosphatase 1NM_194449379148220.00024DelGC
PIGNPhosphatidylinositol glycan anchor biosynthesis, class NNM_176787381108620.00009DelGC
RGS12Regulator of G-protein signaling 12NM_198227271602520.00045DelGC
SHROOM3Shroom family member 3NM_020859273206820.00021DelGC
SMARCA2SMARC, subfamily a, member 2NM_003070315994620.00027DelGC
SPG7Spastic paraplegia 7NM_199367367404820.00019DelGC
TAF4TAF4 RNA polymerase IINM_003185391271820.00071Amp/DelGC
TLN1Talin 1NM_006289320474420.00055Amp/DelGC
UBA6Ubiquitin-like modifier activating enzyme 6NM_018227277171820.00023DelGC
WDR7WD repeat domain 7NM_052834378944220.00039DelGC
KIAA1967KIAA1967NM021174308959730.00029DelCRC
MYCBP2MYC binding protein 2NM_015057351849620.00045Amp/DelCRC
TH1LTH1-like (Drosophila)NM_198976389127820.00071AmpCRC
WDR7WD repeat domain 7NM_052834378944220.00045DelCRC

Gene copy number profile using array-comparative genomic hybridization analysis.  We evaluated changes in the gene copy number, including gene amplification/deletion and chromosomal gain/loss, in a set of 35 definitive oncogenes and tumor suppressor genes by referring to a previous report.(14) The gene copy number profiles for a panel of 22 gastric and colorectal cancer cell lines are shown in Figure 4(C). A gain in the copy number was frequently observed in MYC (41%, 9/22), FGFR2 (18%, 4/22), MET (14%, 3/22), KRAS (14%, 3/22) and SRC (9%, 2/22). Meanwhile, a loss of copy number was observed in CDKN2A (41%, 9/22), FHIT (27%, 6/22), FGFR1 (14%, 3/22) and RB1 (14%, 3/22). A loss of copy number for KIT and PDGFRA, located side-by-side on chromosome 4q12, was unexpectedly observed in two gastric cancer cell lines (HSC-43 and MKN45). Interestingly, MET and FGFR2 were exclusively amplified only in gastric cancers. The results also suggested that copy number changes occur more frequently in gastric cancers than in colorectal cancers.

Fibroblast growth factor receptor 2 amplification, C-terminus truncation and sensitivity to fibroblast growth factor receptor inhibitor.  We previously reported that FGFR2-amplified gastric cancer cell lines were markedly sensitive to a small molecule inhibitor with a kinase inhibitory effect on FGFR, and some of the FGFR2-amplified cell lines dominantly expressed a C-terminus truncated FGFR2.(15) In this study, we evaluated the truncated-FGFR2 transcripts using an exon array analysis (Fig. 5A). The exon array demonstrated that HSC-39, HSC-43, SNU-16 and TU-KATOIII cells overexpressed FGFR2, and amplification was also confirmed using an array-CGH (Fig. 5B). Notably, a C-terminus truncated form of FGFR2 lacking exon 17 was observed dominantly in HSC-43 and partially in TU-KATOIII. RT-PCR for exons 16–17 and exons 2–3 (control) confirmed these results (Fig. 5C). Array-CGH data also showed that the FGFR2-amplicon included the following genes in HSC-39 (FGFR2), HSC-43 (BRWD2, FGFR2 and ATE1), SNU-16 (PPAPDC1A, BRWD2, FGFR2 and ATE1) and TU-KATOIII (PPAPDC1A, BRWD2 and FGFR2). As expected, only FGFR2-amplified cell lines, which dominantly express the truncated form (HSC-43) or the wild-type form (SNU-16), exhibited a remarkable sensitivity (approximately 100-fold) to the FGFR tyrosine kinase inhibitor PD173074, indicating that the major determinant of sensitivity to the FGFR inhibitor was FGFR2 amplification, and not truncation, in the case of FGFR2 (Fig. 5D).

image

Figure 5. FGFR2 amplification and marked sensitivity to fibroblast growth factor receptor (FGFR) tyrosine kinase inhibitor in gastric cancer cell lines. (A) Exon array analysis showing the exon level mRNA expression of FGFR2 in 11 gastric and cancer cell lines and one gastric normal mucosa specimen. The log2-transformed values indicate the expression levels of each exon. The schematic diagram in the upper panel shows each exon of FGFR2. Note that FGFR2 is overexpressed in HSC-39, HSC-43, SNU-16 and TU-KATOIII cells (log2 scale), and a C-terminus truncation of FGFR2 is observed in HSC-43 and partially in TU-KATOIII. p1–2: primer set. (B) Array-comparative genomic hybridization analysis of FGFR2 locus in 10 gastric cancer cell lines. A gain of genomic copy number is shown as a bar extending toward the plus side of the baseline (HSC-39, HSC-43, SNU-16 and TU-KATOIII cells). (C) The mRNA expression levels of FGFR2 were detected using RT-PCR and the following primer (p1, C-terminus region; p2, common region). (D) Growth inhibitory effect of the FGFR tyrosine kinase inhibitor PD173074 in FGFR2-amplified (HSC-43 and SNU-16) and non-amplified (44As3 and 58As1) gastric cancer cell lines.

Download figure to PowerPoint

These results for FGFR2 amplification and drug sensitivity suggested that an integrated analysis involving an exon array and array-CGH is useful for the rapid and efficient identification of factors that determine sensitivity to molecular-targeted drugs and for evaluating abnormal transcripts that are amplified.

Discussion

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

Normal gastric and colonic tissues expressed the Cβ2 isoform of PRKACB; however, we found that gastric cancer specimens predominantly expressed the Cβ1 isoform (Fig. 1). The large difference in the expressions of variants between gastric cancer cells and paired non-cancerous gastric mucosa suggests that the predominant expression of the Cβ1 isoform increases during carcinogenesis or is involved in the malignant phenotype. Because these cancer-specific alternative splice variants might be useful therapeutic targets in cancer treatment, the significance of cancer-related isoforms of PRKACB warrants further biological investigation.

We found that CTNNB1, CDH13 and CDH1 expressed mutated transcripts that were caused by small genomic deletions in gastric and colorectal cancer cells. CDH13 is considered to be a tumor suppressor gene, and its expression is downregulated by the hypermethylation of the promoter region in breast, ovarian and lung cancers.(16) COLO201 cells expressed novel mutated transcripts lacking exons 3–5 as a result of a small genomic deletion (approximately 0.4 Mb, Fig. 2A). Because this deletion was too small to detect using an array-CGH in a standard manner, it would have been overlooked if the exon expression data had not been available (Fig. 2A,B). Collectively, an integrated analysis involving a whole genome exon array and an array-CGH might be a promising and efficient approach for identifying mutated transcripts on a whole genome scale, even when the mutated transcripts are produced by relatively small and exon-level genomic deletions in cancer cells.

Recent successful clinical developments of molecular-targeted drugs, largely targeting constitutively activated oncogenes such as tyrosine kinases, show that mechanisms enabling a gain-of-function in oncogenes could be promising therapeutic targets.(17) In this study, drug sensitivity and gene amplification were highly correlated (Fig. 5), indicating that cell-based evaluations using a genetically well-characterized cell line panel for uncharacterized compounds might be useful for identifying factors that determine sensitivity.

In conclusion, we found several abnormal transcripts and genomic alterations in gastric and colorectal cancer cells using an integrated analysis involving a whole genome exon array and array-CGH. Our approach might enable discoveries in cancer genetics to be made more efficiently than with the use of conventional methods and warrants further investigation involving integrated analyses of clinical samples. Our results also suggest that the combination of a whole genome level approach and a focused-gene approach might be an effective strategy.

Acknowledgments

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

This study was supported by the Third-Term Comprehensive 10-Year Strategy for Cancer Control and a Grant-in-Aid for Cancer Research from the Ministry of Health, Labour and Welfare (22-9 and 22-15). We thank Miss Tomoko Kitayama for her technical assistance.

Disclosure Statement

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

All authors declare no financial support or relationship that may pose conflict of interest.

References

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

Supporting Information

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

Data S1. Materials and Methods in detail.

Data S2. Exon array analysis for gastric cancer cell lines.

Data S3. Exon array analysis for colorectal cancer cell lines.

Data S4. Exon array analysis for recurrently altered genes.

Data S5. Array-comparative genomic hybridization analysis for gastric and colorectal cancer cell lines.

Fig. S1. Expression levels of PRKACB1 isoform.

Fig. S2. Sequencing analysis for CDH13 transcripts.

Fig. S3. Sequencing analysis for CDH1 transcripts.

Fig. S4. Comparative genomic hybridization (CGH) analysis for gastric cancer cell lines.

Fig. S5. Comparative genomic hybridization (CGH) analysis for colorectal cancer cell lines.

FilenameFormatSizeDescription
CAS_2132_sm_DataS1.doc40KSupporting info item
CAS_2132_sm_DataS2.xls298KSupporting info item
CAS_2132_sm_DataS3.xls274KSupporting info item
CAS_2132_sm_DataS4.xls97KSupporting info item
CAS_2132_sm_DataS5.xls3614KSupporting info item
CAS_2132_sm_fS1.tiff356KSupporting info item
CAS_2132_sm_fS2.tiff311KSupporting info item
CAS_2132_sm_fS3.tiff375KSupporting info item
CAS_2132_sm_fS4.tiff191KSupporting info item
CAS_2132_sm_fS5.tiff183KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.