To identify tumor markers and differentiation markers for lung adenocarcinoma (AdC), we analysed expression profiles of 14 500 genes against three cases of type II alveolar epithelial cells, bronchiolar epithelial cells, and bronchial epithelial cells, respectively, and 10 cases of AdC cells isolated by laser capture microdissection. Hierarchical clustering analysis indicated that AdC cells and noncancerous lung epithelial cells are significantly different in their expression profiles, and that different sets of differentiation markers are expressed among alveolar, bronchiolar and bronchial epithelial cells. Nine genes were identified as being highly expressed in AdC cells, but not expressed in noncancerous lung epithelial cells. Sixteen genes were identified as differentiation markers for lung epithelial cells. Real-time RT–PCR analysis of 45 lung AdC cases further revealed that expression of four tumor markers in AdC cells was significantly higher than that in noncancerous lung cells and that expression of ten differentiation markers was retained in a considerable fraction of lung AdC cases. Five tumor markers and seven differentiation markers were not expressed in peripheral blood cells. Similarities and differences in expression profiles between normal epithelial cells from different lung respiratory compartments and AdC cells demonstrated in this study will be informative for the molecular diagnosis of lung AdC.
Lung cancer is the leading cause of cancer death in many countries, and in recent years, adenocarcinoma (AdC) has become the most common type. Since a majority of lung AdC patients are discovered in an advanced stage, difficulties in detection of early stage AdCs and lack of effective therapies for advanced stage AdCs have been recognized as major causes of the high mortality of this disease. Even with the introduction of computed tomographic (CT) screening for lung cancer, the overall 5-year survival rates of patients with small-sized (⩽2 cm in diameter) AdC and those with stage IA AdC are approximately 80% (Naruke et al., 2001; Takamochi et al., 2004). Therefore, it has been assumed that a subset of these patients already had occult metastases at the time of primary surgery that were undetectable by current tumor-node-metastasis (TNM) staging methods (Sobin, 2002). For this reason, identification of lung AdC specific molecular markers has been thought to be essential for the development of novel ways of effective screening and more accurate TNM staging (Hosch et al., 2001; Wang et al., 2002). In particular, identification of a set of tumor markers and differentiation markers specific for lung AdC will be highly useful for detection of lung AdC cells in peripheral blood (PB), bone marrow and lymph nodes. Development of a more sensitive and specific method for detection of micrometastasis will lead to a more accurate TNM staging resulting in the more accurate prediction of patients' prognoses and design of postoperative therapeutic approaches.
Up to the present, several tumor markers, such as carcino-embryonic antigen (CEA) and cancer testis (CT) antigens, have been studied for detection of lung AdC (D'Cunha et al., 2002; Egland et al., 2002; Kufer et al., 2002; Takamochi et al., 2004). Several differentiation markers, such as cytokeratin (CK) 19, have been also studied for detection of lung AdC. However, most of them are not specific for lung AdC but for epithelial cells or cancers of epithelial cell origin in general. Even for tissue-specific differentiation markers, such as the surfactant pulmonary-associated proteins A, B and C (SFTPA2, SFTPB and SFTPC, respectively) and thyroid transcription factor-1 (TTF-1), sensitivity as well as specificity is not high enough to use for molecular diagnosis (Chhieng et al., 2001; Zamecnik and Kodet, 2002). Thus, it is absolutely necessary to search for more specific and highly sensitive markers that will be useful for molecular diagnosis of lung AdC.
Recently, GeneChip oligonucleotide microarray analysis has been introduced to identify many additional molecular markers (Segal et al., 2004), and several groups have reported microarray-based subclassifications of lung AdC (Bhattacharjee et al., 2001; Garber et al., 2001; Miura et al., 2002; Virtanen et al., 2002; Tomida et al., 2004). However, it is still unclear what genes are useful for molecular diagnosis of lung AdC in clinics. The origins of AdC cells are thought to be alveolar, bronchiolar and bronchial epithelial cells (Ten Have-Opbroek et al., 1997; Otto, 2002; Borczuk et al., 2003). Thus, we thought it was important to elucidate similarities and differences of AdC cells in comparison with several types of lung epithelial cells for gene expression profiling for the identification of novel and useful molecular markers specific for lung AdC. For this purpose, a panel of type II alveolar epithelial cells, bronchiolar epithelial cells and bronchial epithelial cells as well as that of lung AdC cells isolated by laser capture microdissection (LCM) (Bonner et al., 1997; Player et al., 2004) was subjected to oligonucleotide microarray analysis of 14 500 genes. By combination of LCM and microarray analyses and by confirmation with quantitative RT–PCR analysis of candidate genes, we identified several genes that will be useful for molecular diagnosis of AdCs.
Laser capture microdissection of lung adenocarcinoma cells and noncancerous lung epithelial cells
Ten cases of AdCs were microdissected by the LCM method, and histologically, four of them were classified as being well differentiated (WD1∼WD4), three were moderately differentiated (MD1∼MD3), and the remaining three were poorly differentiated (PD1∼PD3). Three cases of type II alveolar epithelial cells (Alv1∼Alv3), bronchiolar epithelial cells (Bio1∼Bio3) and bronchial epithelial cells (Bia1∼Bia3), respectively, were also obtained by the LCM method as representatives of noncancerous epithelial cells from three different lung respiratory compartments. Pairs of Alv1/Bio1, Alv2/WD1, Alv3/PD1 and Bio3/Bia1 were obtained from the same patients, respectively, while other samples were obtained from different patients. Microscopic visualization of representative tissue sections are shown in Figure 1. Type II alveolar epithelial cells and bronchiolar epithelial cells as well as AdC cells were microdissected as previously described (Kobayashi et al., 2004). Bronchi are encircled with cartilage, while bronchioles are not. Thus, bronchial epithelial cells were microdissected from epithelia neighboring cartilage. Nonciliated columnar cells are called Clara cells and line the distal bronchi and the bronchioles. Ciliated bronchiolar or bronchial epithelial cells were microdissected together with Clara cells.
Total RNA was extracted from the cells, and the quality, quantity and purity of RNAs were assessed by RT–PCR analysis of the surfactant pulmonary-associated protein B (SFTPB), Clara-cell-specific 10-kDa protein (SCGB1A1) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) genes (Kobayashi et al., 2004). SFTPB mRNA is known to be expressed in both type II alveolar epithelial cells and bronchiolar epithelial cells, whereas SCGB1A1 mRNA is known to be expressed in Clara cells in bronchiolar and bronchial epithelia but not in type II alveolar epithelial cells (Broers et al., 1992; Xu et al., 1998; Barth et al., 2000; Borczuk et al., 2003). As expected, SFTPB specific PCR products were detected in both type II alveolar epithelial cells and bronchiolar epithelial cells, and SCGB1A1 specific products were detected in both bronchiolar and bronchial epithelial cells (Figure 2a). The quality and quantity of RNAs extracted from all samples, including 10 AdC samples, were confirmed by RT–PCR of the GAPDH gene. The results indicated that the quality of RNA from epithelial cells as well as AdC cells was well preserved after LCM and good enough for microarray analysis.
Oligonucleotide microarray analysis of lung adenocarcinoma cells and noncancerous lung epithelial cells
RNAs were amplified by the TALPAT method (Aoyagi et al.,. 2003), and hybridized to the U133A geneChip that contain 22 283 probe sets corresponding to 14 500 genes. The percent of probe sets judged as ‘present calls’ ranged from 25.6 to 35.7% in noncancerous lung epithelial cells and from 27.0 to 41.7% in lung AdC cells, indicating that the number of genes expressed is not significantly different between noncancerous epithelial cells and AdC cells (Table S1). Signal intensities for gene expression were then scaled to a target intensity of 1000 by Affymetrix Microarray Suite 4.0 software. Figure 2b shows microarray data for the expression of the SFTPB, SCGB1A1 and GAPDH genes in the corresponding samples used for RT–PCR analysis. The results of microarray analysis were well compatible with those of RT–PCR analysis. We then classified all 19 samples by hierarchical clustering analysis to examine how gene expression profiles were different among the samples. For the analysis, expression signals were first filtered by absolute calls, and among 22 283 probe sets, 7446 sets that showed ‘absent calls’ for all 19 samples were excluded. The analysis using the remaining 14 837 probe sets classified the samples into several subgroups (Figure 3). Nine noncancerous epithelial cell samples and 10 AdC cell samples were first separated to different branches. Among the nine epithelial cell samples, type II alveolar epithelial cells were then branched from bronchiolar and bronchial epithelial cells. Finally, three bronchiolar epithelial cell samples and three bronchial epithelial cell samples were divided into two different groups. However, 10 AdC cell samples did not evenly branch according to the differentiation status of pathological classification. These results indicated that differences in gene expression profiles were more evident between noncancerous epithelial cells and AdC cells than among alveolar, bronchiolar and bronchial epithelial cells, that gene expression profiles among alveolar, bronchiolar and bronchial epithelial cells are significantly different to each other, and that expression profiles do not always correlate with pathological classification of AdC cells.
Identification of tumor markers for lung adenocarcinoma
To identify genes whose expression is high in lung AdC cells and not detected in noncancerous lung epithelial cells, we compared gene expression profiles of lung AdC cells with those of alveolar, bronchiolar and bronchial epithelial cells. Among 14 837 probe sets that showed ‘present calls’ for at least one of the 19 samples, 2372 sets, which showed ‘absent calls’ for all nine epithelial cell samples, were considered as being not expressed in epithelial cells. Among the 2372 sets, a number of probe sets showed high signal intensities in AdC cells. To further define genes significantly and commonly overexpressed in AdC cells, probe sets that showed signal intensities of >3000 in at least three AdC samples were selected (Table 1). Nine genes of 10 probe sets, SGNE1, COL11A1, STK6, TFPI2, ATP10B, TM4SF4, GAGED2, MCM6 and TOP2A, were selected under this criterion. There was no gene that showed overexpression in all 10 AdC cases, instead, those genes showed overexpression in three to five of them.
To confirm the overexpression of the nine genes in lung AdC, we examined the expression of these genes in 45 cases of macrodissected primary lung AdC cells and 14 cases of macrodissected noncancerous lung cells by real-time RT–PCR analysis. To make a comparison of expression levels for each gene among the cases easier, mean levels of expression for each gene in 45 cases of AdCs were adjusted to 1000 and the relative values for expression were compared among the cases. Mean levels of expression in AdC cells were significantly higher (P<0.05) than those in noncancerous lung cells in four genes of SGNE1, COL11A1, GAGED2 and TOP2A (Table 2, Figure S1). In particular, the mean levels of COL11A1 and GAGED2 expression in noncancerous lung cells was more than 40 times lower than those in AdC cells. Thus, it was concluded that four genes of SGNE1, COL11A1, GAGED2 and TOP2A were significantly overexpressed in lung AdC cells.
Identification of differentiation markers for lung adenocarcinoma
We next attempted to identify genes that are differentially expressed among three different anatomic regions of lung epithelial cells, because those genes are strong candidates for differentiation markers of alveolar, bronchiolar and bronchial epithelial cells, respectively. Among 12 465 probe sets that showed ‘present calls’ in at least one of the nine noncancerous epithelial cell samples, 3089 showed ‘present calls’ in all the nine samples, and thus were considered as being expressed in any region of epithelial cells analysed. In contrast, 356 probe sets showed cell type specific patterns of gene expression among alveolar, bronchiolar and bronchial epithelial cells (Table S2). The remaining 9020 probe sets did not show any correlation between their present/absent calls and cell type specificities.
Among the 356 probe sets, 121 probe sets showed ‘present calls’ only in alveolar cells, and ‘absent/marginal calls’ in bronchiolar/bronchial epithelial cells, while 40 probe sets showed ‘present calls’ in alveolar/bronchiolar epithelial cells, and ‘absent/marginal calls’ in bronchial epithelial cells, suggesting that genes for these 161 probe sets were expressed predominantly in alveolar/bronchiolar epithelial cells. On the other hand, 154 probe sets showed ‘present calls’ in bronchiolar/bronchial epithelial cells, but ‘absent/marginal calls’ in alveolar epithelial cells, and another 22 probe sets showed ‘present calls’ only in bronchial epithelial cells, and ‘absent/marginal calls’ in alveolar/bronchiolar epithelial cells, suggesting that genes for these 176 probe sets were expressed predominantly in bronchiolar/bronchial epithelial cells. The number of probe sets that were positive only in bronchiolar epithelial cells was 14, and the number of probe sets positive in both alveolar and bronchial epithelial cells and negative in bronchiolar epithelial cells was only five.
However, if signal intensities of >5000 were considered as being highly expressed, only 16 genes were judged as being expressed in a cell-type specific manner (Table 3). There were five genes that were expressed only in alveolar epithelial cells. Three other genes of four probe sets were expressed in alveolar/bronchiolar epithelial cells and not in bronchial epithelial cells. Thus, these eight genes were judged as being strong candidates for differentiation markers for alveolar/bronchiolar epithelial cells. As expected, SFTPA2, SFTPB and SFTPC were classified into this group (Otto, 2002). However, it was noted that five genes of CLDN5, RARRES2, RAFTLIN, CLST11240 and DF were shown to be expressed only in type II alveolar epithelial cells, indicating that these genes could be unique and noble markers for type II alveolar epithelial cells. None of the probe sets were selected as being present only in bronchiolar or bronchial epithelial cells under this criterion. Instead, there were eight genes that showed positive in bronchiolar/bronchial epithelial cells and negative in alveolar epithelial cells, suggesting that these eight genes were differentiation markers for bronchiolar/bronchial epithelial cells. As expected, SCGB1A1 was judged as being highly expressed in bronchiolar and bronchial lung epithelial cells (Broers et al., 1992). None of the probe sets were selected as being ‘present’ in alveolar/bronchial epithelial cells but absent/marginal in bronchiolar epithelial cells. Thus, in total, 16 genes of 17 probe sets were selected as being differentiation markers for lung epithelial cells.
Subclassification of lung adenocarcinomas according to the expression profiles of differentiation makers
Most lung AdCs are thought to originate from type II alveolar epithelial cells or bronchiolar epithelial cells and often retain their phenotypes, in particular in well-differentiated type AdCs. Thus, it was interesting to classify the 10 AdC samples based on the expression profiles of the 17 probes selected above. As shown in Figure 4, all the 10 AdC cases were classified into a group of type II alveolar epithelial cells. In particular, one of four well differentiated types (WD3) was classified in the same branch as three cases of type II alveolar epithelial cells, and two other cases of well differentiated types (WD1 and WD4) were also classified into the same sub-branch as type II alveolar epithelial cells. Thus, although pathological classification of differentiation stages did not always correlate with the classification based on the expression profiles of the 17 probes, well differentiated types of AdC often retained the expression profiles of type II alveolar epithelial cells.
Expression of differentiation markers in lung adenocarcinoma cells
We next examined the expression of the 16 genes in the same set of macrodissected lung AdC cells and noncancerous lung cells by real-time RT–PCR analysis (Table 4). Although we failed to design primer sets for the amplification of mRNAs specific to two genes (Affy. ID numbers 205382 and 222271 in Table 3), 14 other genes were successfully analysed by this method. To make a comparison of the differences in relative expression levels for each gene among samples analysed, and of those between lung AdCs and noncancerous lung cells for differentiation marker genes with those for tumor marker genes easier, mean levels of expression for each gene in 45 cases of AdCs were also adjusted to 1000 and the relative values for expression were compared among samples (Table 4). Interestingly, the mean value of MUC4 in lung AdCs was higher than that in noncancerous lung cells. The result indicated that expression of the MUC4 gene was retained in a large fraction of lung AdC cases. Alternatively, it is also possible that the MUC4 gene was aberrantly overexpressed in a subset of lung AdCs. In either case, the result suggested that the MUC4 gene could be useful as a differentiation marker whose expression is retained in lung AdCs. Expression of three other genes, SFTPC, SFTPB and FABP6, was retained relatively frequently in lung AdCs, since the mean values were not significantly different between lung AdCs and noncancerous lung cells. However, in the remaining 10 genes, mean values in lung AdCs were significantly lower than those in noncancerous lung cells, suggesting that those genes were downregulated in a large fraction of lung AdCs.
To further evaluate whether expression of these genes in AdC cells correlates with the histological differentiation of AdC cells, levels of mRNA expression for each gene were compared among 45 AdC cases (Table 5). In this table, to compare between the expression levels of each gene and histological differentiation of the AdC cases, cases with expression over the mean levels of 14 cases of noncancerous lung cells are indicated by plus marks. For CLDN5, CLST11240, SCGB1A1 and DKFZp586M1120, none of the 45 AdC cases showed higher expression levels than noncancerous lung cells, thus, these genes were deleted from this table. The other 10 genes showed high expression levels in various fractions of AdC cases. SFTPA2, SFTPB and SFTPC showed high expression levels only in the well differentiated type, whereas RAFTLIN, LOC123872 and DFKZp434B227 showed high expression levels in the well and moderately differentiated types. Interestingly, MUC4 and FABP6 showed high levels not only in well and moderately differentiated types but also in the poorly differentiated type, and DNAI2 showed a high level only in a case of the moderately differentiated type and RARRES2 showed high levels in one case of moderately and poorly differentiated types, respectively. Thus, these genes were expressed to various extents and irrespective of differentiation status in lung AdCs.
Expression of lung adenocarcinoma markers in peripheral blood cells
We next examined the expression of nine tumor marker genes in five cases of PB cells. Expression of three genes, COL11A1, TM4SF4 and GAGED2, was not detected and expression of two other genes, SGNE1 and TFPI2, was extremely low in PB cells (Table 2). On the other hand, the remaining four genes, STK6, ATP10B, MCM6 and TOP2A, were expressed at considerable levels in PB cells. We next examined the expression of 14 differentiation marker genes in five cases of PB cells. Seven of the 14 genes, CLDN5, RAFTLIN, SFTPA2, SCGB1A1, DKFZp434B227, DNAI2 and FABP6, were expressed at adequate levels in PB cells. However, five other genes, RARRES2, CLST11240, MUC4, LOC123872 and DKFZp586M1120, were not expressed, and the remaining two genes, SFTPB and SFTPC, were expressed at extremely low levels in PB cells (Table 4).
Genes highly and specifically expressed in cancer cells are useful for molecular detection of cancer cells. Genes that are expressed in a tissue specific manner in precursor cells and whose expression is retained in cancer cells are useful for differential diagnosis to predict the origin of cancer cells. To identify these genes, we isolated lung AdC cells and noncancerous lung epithelial cells of three different anatomic regions by the LCM method. These cells were then used for oligo nucleotide microarray analysis of 14 500 genes. Hierarchical clustering analysis revealed that lung AdC cells and noncancerous lung epithelial cells were significantly different in their gene expression profiles, suggesting the presence of genes up or downregulated specifically and commonly in lung AdC cells in comparison with noncancerous lung epithelial cells. Subsequently, nine genes were identified as being upregulated in AdC cells in vivo in comparison with noncancerous lung epithelial cells, and thus, as possible lung AdC specific tumor markers. Moreover real-time RT–PCR analysis against 45 AdCs and 14 noncancerous lung tissues revealed that expression of four of the nine genes, SGNE1, COL11A1, GAGED2 and TOP2A, was significantly high in lung AdCs in comparison with noncancerous lung tissues. The SGNE1 gene encodes a neuroendocrine protein called 7B2, and its serum levels are elevated in several types of neuroendocrine tumors, including small cell lung cancer (Vieau et al., 1991; Mbikay et al., 2001). The COL11A1 gene encodes one of the two alpha chains of type XI collagen, and it was previously reported that COL11A1 is expressed in colorectal cancers but not expressed in normal colon epithelia (Fischer et al., 2001). GAGED2 encodes a protein of a family of CT antigens and is known to be expressed in a variety of cancers (Egland et al., 2002). TOP2 is also known to be highly expressed in various types of cancers. Since these four genes are expressed in a variety of cancers, they are not likely to be expressed specifically in lung AdC cells. Therefore, molecular diagnosis with a set of these markers, rather than with any of a single marker, will be useful for the detection and characterization of lung AdC.
Hierarchical clustering analysis also revealed that there were considerable numbers of genes differentially expressed among three different types of noncancerous lung epithelial cells. Subsequently, we identified 16 genes that were expressed specifically in either alveolar/bronchiolar or bronchiolar/bronchial epithelial cells. As expected, SFTPA2, SFTPB and SFTPC were identified as genes that were expressed in alveolar/bronchiolar epithelial cells but not in bronchial epithelial cells (Otto, 2002). Five genes were identified as being expressed only in alveolar epithelial cells but not in bronchiolar/bronchial epithelial cells. In addition to CC10 (SCGB1A1), seven genes were identified as being expressed in bronchiolar/ bronchial epithelial cells, but not in alveolar epithelial cells (Broers et al., 1992). Thus, this is the first report showing the difference in gene expression profiles among three different regions of lung epithelial cells, and identifying various differentiation marker genes for respective epithelial cells. Moreover, hierarchical clustering analysis with these 16 genes indicated that the gene expression profile of lung AdC is more similar to that of type II alveolar epithelial cells than that of bronchiolar/bronchial epithelial cells. This finding supports the previous reports showing that lung AdCs are derived from type II alveolar epithelial cells and not from Clara cells (Ten Have-Opbroek et al., 1997; Otto, 2002). In particular, expression of 10 genes, SFTPA2, SFTPB, SFTPC, RARRES2, RAFTLIN, MUC4, LOC123872, DFKZp434B227, DNAI2 and FABP6, was retained in a considerable fraction of lung AdC cases. SFTPA2, SFTPB and SFTPC are typical examples of differentiation markers isolated in this study. However, MUC4 is known to be expressed in various normal epithelial cells as well as several types of carcinomas, including lung carcinoma (Hanaoka et al., 2001; Moniaux et al., 2004). At present, tissue specificities for the expression of the remaining six genes are unknown. Thus, further studies should focus on the characterization of these genes as lung epithelial cell-specific markers.
Finally, we found that five tumor markers, COL11A1, GAGED2, TM4SF4, SGNE1 and TFPI2, and seven differentiation markers, SFTPB, SFTPC, RARRES2, CLST11240, MUC4, LOC123872 and DKFZp586M1120 were not expressed or expressed at extremely low levels in PB cells. Up to the present, several genes have been used for detection of lung AdC cells in PB, such as CK19 and CEA (Matsunaga et al., 2002; Mitas et al., 2003). However, both the sensitivity and specificity of these genes for detection of lung AdC cells were not high enough to use as a single molecular marker in clinics. Thus, it will be important to further evaluate the specificity and sensitivity for the combinational use of the 12 markers identified in this study with several known markers for lung AdC cell detection in PB.
Materials and methods
Preparation of tissue samples
Ten primary lung AdCs and nine normal lung tissues used for oligonucleotide microarray analysis were obtained from 15 patients with lung AdC at surgery at the National Cancer Center Hospital, Tokyo, Japan (Table S1). AdCs were histologically classified according to the Histological Typing of Lung and Pleural Tumors (Travis et al., 1999) and staged according to the TNM Classification of Malignant Tumors (Sobin, 2002). These tissues were embedded in OCT medium and frozen at −80°C as described previously (Kobayashi et al., 2004).
Another 45 lung AdC samples and 14 noncancerous lung tissues were obtained from 46 patients at surgery and kept at −80°C until RNA extraction. Total RNA was extracted from these macrodissected samples using an RNeasy Mini Kit (QIAGEN, Valencia, CA, USA). Peripheral blood cells were obtained from five individuals with no past histories of cancer, and total RNA from these samples was extracted using a QIAamp RNA Blood Mini Kit (QIAGEN).
Laser capture microdissection and total RNA extraction
The OCT-embedded tissues were cryostat sectioned at 8 μm thickness, mounted on glass slides, and then cells were microdissected using the PixCell II LCM (Arcturus Engineering, Mountain View, CA, USA) as described previously (Kobayashi et al.,. 2004). Approximately 100 type II alveolar epithelial cells, 200 bronchiolar epithelial cells, 200 bronchial epithelial cells and 500 AdC cells were captured onto a single transfer film from a single section. Total RNA was extracted from 2000 microdissected cells on the film using the Micro RNA Isolation Kit (Stratagene, San Diego, CA, USA) as described previously (Kobayashi et al., 2004).
RT–PCR analysis for SFTPB, SCGB1A1 and GAPDH were performed using RNA from approximately 1000 cells as described previously (Kobayashi et al., 2004).
cRNA preparation and GeneChip hybridization
Total RNA (10 μl) solutions, containing RNA from 1000 cells, were subjected to PCR-mediated global mRNA amplification using the TALPAT (T7 RNA polymerase promoter-attached, adaptor ligation-mediated, and PCR amplification followed by in vitro T7-transcription) method (Aoyagi et al., 2003). Biotin-labeled cRNA was prepared as described in the Affymetrix GeneChip Expression Analysis Manual (Affymetrix Santa Clara, CA, USA) and hybridized to the human U133 GeneChip.
Microarray data analysis
Array chip slides were scanned with a GMS418 Array Scanner (Genetic Microsystems, Woburn, MA, USA), and image analysis of the array was performed according to the manufacturer's protocol. Raw data were analysed with an Affymetrix Microarray Analysis Suite (MAS) v4.0 and were globally normalized and scaled to a target intensity of 1000 to facilitate inter-array comparison. The MAS software generated absolute calls (present, marginal or absent), and the calls were used to determine whether a gene was expressed or not. After being filtered through ‘absent call’ for all samples, lists of genes expressed in AdC cells or noncancerous lung epithelial cells were ranked based on absolute call. Hierarchical clustering analysis was performed using the GeneSpring 6.1 (Silicon Genetics, Redwood City, CA, USA) software with unsupervised and supervised clustering of normalized data.
Real-time RT–PCR analysis
Real-time RT–PCR analysis was performed using the ABI Prism 7900HT sequence detection system (Applied Biosystems, Lincoln Centre Drive Foster City, CA, USA). The expression level of each gene was normalized to RNA content for each sample by using GAPDH as an internal control. A primer pair and a TaqMan probe for each gene are listed in Table S3. PCR for each target gene and the reference gene GAPDH was performed in a single tube in duplicate, and the results were expressed as the average of the two independent tests.
Aoyagi K, Tatsuta T, Nishigaki M, Akimoto S, Tanabe C, Omoto Y et al. (2003). Biochem Biophys Res Commun 300: 915–920.
Barth PJ, Koch S, Muller B, Unterstab F, von Wichert P, Moll R . (2000). Virchows Arch 437: 648–655.
Bhattacharjee A, Richards WG, Staunton J, Li C, Monti S, Vasa P et al. (2001). Proc Natl Acad Sci USA 98: 13790–13795.
Bonner RF, Emmert-Buck M, Cole K, Pohida T, Chuaqui R, Goldstein S et al. (1997). Science 278: 1481, 1483.
Borczuk AC, Gorenstein L, Walter KL, Assaad AA, Wang L, Powell CA . (2003). Am J Pathol 163: 1949–1960.
Broers JL, Jensen SM, Travis WD, Pass H, Whitsett JA, Singh G et al. (1992). Lab Invest 66: 337–346.
Chhieng DC, Cangiarella JF, Zakowski MF, Goswami S, Cohen JM, Yee HT . (2001). Cancer 93: 330–336.
D’Cunha J, Corfits AL, Herndon II JE, Kern JA, Kohman LJ, Patterson GA et al. (2002). J Thorac Cardiovasc Surg 123: 484–491; discussion 491.
Egland KA, Kumar V, Duray P, Pastan I . (2002). Mol Cancer Ther 1: 441–450.
Fischer H, Stenling R, Rubio C, Lindblom A . (2001). Carcinogenesis 22: 875–878.
Garber ME, Troyanskaya OG, Schluens K, Petersen S, Thaesler Z, Pacyna-Gengelbach M et al. (2001). Proc Natl Acad Sci USA 98: 13784–13789.
Hanaoka J, Kontani K, Sawai S, Ichinose M, Tezuka N, Inoue S et al. (2001). Cancer 92: 2148–2157.
Hosch SB, Scheunemann P, Izbicki JR . (2001). Semin Surg Oncol 20: 278–281.
Kobayashi K, Nishioka M, Kohno T, Nakamoto M, Maeshima A, Aoyagi K et al. (2004). Oncogene 23: 3089–3096.
Kufer P, Zippelius A, Lutterbuse R, Mecklenburg I, Enzmann T, Montag A et al. (2002). Cancer Res 62: 251–261.
Matsunaga H, Hangai N, Aso Y, Okano K, Kawamura M, Kobayashi K et al. (2002). Int J Cancer 100: 592–599.
Mbikay M, Seidah NG, Chretien M . (2001). Biochem J 357: 329–342.
Mitas M, Hoover L, Silvestri G, Reed C, Green M, Turrisi AT et al. (2003). J Mol Diagn 5: 237–242.
Miura K, Bowman ED, Simon R, Peng AC, Robles AI, Jones RT et al. (2002). Cancer Res 62: 3244–3250.
Moniaux N, Andrianifahanana M, Brand RE, Batra SK . (2004). Br J Cancer 91: 1633–1638.
Naruke T, Tsuchiya R, Kondo H, Asamura H . (2001). Ann Thorac Surg 71: 1759–1764.
Otto WR . (2002). J Pathol 197: 527–535.
Player A, Barrett JC, Kawasaki ES . (2004). Expert Rev Mol Diagn 4: 831–840.
Segal E, Friedman N, Koller D, Regev A . (2004). Nat Genet 36: 1090–1098.
Sobin Ch (Christian) LHW, International Union against Cancer. (2002). In: LH Sobin and Ch Wittekind (eds). TNM classification of malignant tumours. 6th edn, Wiley-Liss: New York, pp 97–103.
Takamochi K, Yoshida J, Nishimura M, Yokose T, Sasaki S, Nishiwaki Y et al. (2004). Eur J Cardiothorac Surg 25: 877–883.
Ten Have-Opbroek AA, Benfield JR, van Krieken JH, Dijkman JH . (1997). Histol Histopathol 12: 319–336.
Tomida S, Koshikawa K, Yatabe Y, Harano T, Ogura N, Mitsudomi T et al. (2004). Oncogene 23: 5360–5370.
Travis WD, Colby TV, Corrin B, Shimosato Y, Brambilla E . (1999). Histological typing of lung and pleural tumours/WD Travis… [et al.]; in collaboration with LH Sobin and pathologists from 14 countries, 3rd edn. Springer-Verlag: Berlin, New York, pp.
Vieau D, Rojas-Miranda A, Verley JM, Lenne F, Bertagna X . (1991). Clin Endocrinol (Oxf) 35: 319–325.
Virtanen C, Ishikawa Y, Honjoh D, Kimura M, Shimane M, Miyoshi T et al. (2002). Proc Natl Acad Sci USA 99: 12357–12362.
Wang KK, Liu N, Radulovich N, Wigle DA, Johnston MR, Shepherd FA et al. (2002). Oncogene 21: 7598–7604.
Xu P, Hashimoto S, Miyazaki H, Asabe K, Shiraishi S, Sueishi K . (1998). Virchows Arch 432: 17–25.
Zamecnik J, Kodet R . (2002). Virchows Arch 440: 353–361.
This work was supported in part by Grants-in-Aid from the Ministry of Health, Labor and Welfare of Japan for the 2nd- and 3rd-term Comprehensive 10-year Strategy for Cancer Control, for Cancer Research (10-4 and 16-1), and for the Program for Promotion of Fundamental Studies in Health Sciences of the Organization for Pharmaceutical Safety and Research. KK and NN were awardees of the Research Resident Fellowship from the Foundation for Promotion of Cancer Research during this study.
About this article
Cite this article
Nakamura, N., Kobayashi, K., Nakamoto, M. et al. Identification of tumor markers and differentiation markers for molecular diagnosis of lung adenocarcinoma. Oncogene 25, 4245–4255 (2006). https://doi.org/10.1038/sj.onc.1209442
- lung adenocarcinoma
- lung epithelial cells
- expression profile
- differential diagnosis
Nature Medicine (2020)
Retinoic Acid affects Lung Adenocarcinoma growth by inducing differentiation via GATA6 activation and EGFR and Wnt inhibition
Scientific Reports (2017)
Journal of Cancer Research and Clinical Oncology (2017)
Tumor Biology (2014)
A robust tool for discriminative analysis and feature selection in paired samples impacts the identification of the genes essential for reprogramming lung tissue to adenocarcinoma
BMC Genomics (2011)