EXPERIMENTAL CARDIOVASCULAR AND LUNG RESEARCH
From mutation to methylation – molecular markers in lung cancer

Introduction

Lung cancer is one of the most common types of malignant neoplasms in Poland and in the world, and is associated with the worst prognosis [1]. In Poland, lung cancer occupies the third place (after breast and colon cancer) with over 6 thousand female patients and the first place with 14.5 thousand male patients. Within recent years, significant progress has been made when it comes to high-throughput and high-resolution technologies in the field of molecular biology, i.e. next-generation sequencing, microarrays, and mass spectrometry, which contributed to the development of research in the fields of genomics, transcriptomics and epigenetics in lung cancer. Implementing the results achieved by these disciplines quickly allows for the analysis of somatic point mutations or translocations to be used in molecular cancer diagnostics, and, thus, for the personalization of treatment for patients with NSCLC. Unfortunately, certain markers and prognostic predictors that could facilitate the detection of lung cancer in its early stages are yet to be found.

Genomics

Genomics study the genome and deals with both the physical traits and the properties of the genome. Numerous studies are being carried out to identify the genetic changes related to the process of carcinogenesis. The results of genomic studies on lung cancer appear promising and may be applicable in clinical practice in the future. The identification of specific genetic changes in cancer tissue is not only a part of the diagnostic process, but is also an important factor in prognostication and “to tailor the treatment” [2]. For example Thomas et al. in their work on the molecular profile of lung cancer, underlined thirteen genes with translocations in ALK, ROS1, KIF5B-RET , amplifications in the MET and FGFR1 genes and other mutations in following genes EGFR, HER2, PI3K, AKT1, KRAS, BRAF, MEK, DDR2 [2].

Mutations in the gene coding the epidermal growth factor receptor are the best known mutations in non-small cell lung cancer (NSCLC), widely used in clinical practice. Deletions in exon 19 of the EGFR gene constitute one type of such mutations. These deletions most commonly affect 9, 12, 15, 18 or 24 nucleotides [3]. Some of these deletions share a common feature – they cause the removal of four amino acids in codons 747-750: leucine, arginine, glutamic acid, and alanine. The second most common mutation in the EGFR gene is substitution in exon 21, mainly L858R [3].

Identifying the above-mentioned somatic mutations provides significant clinical benefits. Patients with activating mutations in the EGFR tyrosine kinase domain respond very well to tyrosine kinase inhibitors, such as erlotinib or gefitinib. The latter ones inhibit EGFR phosphorylation, which prevents the division of neoplastic cells and leads to their death [4].

Taking into consideration the fact that mutations in exons 18, 19, 20, and 21 take place in about 10-13% of Caucasian NSCLC patients [5, 6], their therapeutic usefulness is significant. Further lung cancer research helped identify the EML4-ALK fusion gene. This gene appears in NSCLC cells as a result of inversion in the short arm of chromosome 2 [7]. Nine variants of the EML4-ALK fusion gene were identified. Each variant contains exons 20-29 of the ALK gene as well as fragments of the EML4 gene. Depending on the variant, these fragments are: exon 13, exon 20, exon 6, exon 6 with the insertion of 11 additional amino acids, exon 2, exon 15, and exon 18. There are also two variants with exon 14 of EML4. In one of them, the beginning of the fusion takes place in nucleotide 13 of exon 20 of the ALK gene, while in the other, insertion of 11 additional nucleotides occurs in exon 20 of the ALK gene [8]. The resulting fusion gene is an oncogene with the ability to promote neoplastic development, which is related to its catalytic activity [7]. This characteristic is utilized in NSCLC therapy. It has been demonstrated that the use of specific inhibitors, blocking tyrosine phosphorylation in the EML4-ALK oncogene, inhibits the activity of tyrosine kinase in this fusion gene. This leads to the death of neoplastic cells in which the expression of this gene is present [7]. Last year, it was reported that other fusion genes can also be found in lung cancer: KLC1-ALK, GOPC-ROS1 [9], TPM3-ROS1, SDC4-ROS1, SLC34A2-ROS1, CD74-ROS1, EZR-ROS1, LRIG3-ROS1 [10], and KIF5B-RET [11]. The fusion between KIF5B and the RET proto-oncogene seems to be an interesting new molecular target in the personalization of diagnostics and the treatment of patients with NSCLC, and it was found in a non-smoking patient without mutations in the EGFR, KRAS, or ALK genes (triple-negative). Another patient with identified KIF5B-RET rearrangement (10p11.22-q11.21) was in the group of double-negative patients (without mutations in the EGFR and ALK genes). Since high expression of the tyrosine kinase domain of the RET gene was observed in only those lung cancer samples which contained the KIF5B-RET fusion gene, it seems that this fusion gene may potentially be a good new therapeutic target. However, more research needs to be conducted on the amplification of the RET gene and the assessment of the fusion of the RET gene with other genes [11].

The application of genomics in lung cancer molecular diagnostics

The above-mentioned mutation analyses constitute a good example of predictive biomarkers which are currently used in oncological molecular diagnostics. However, there are still no biomarkers that would enable quick and early detection of lung cancer.

The loss of heterozygosity in the 3p region may potentially represent such a biomarker. This type of deletion occurs in over 70% of NSCLC cases and over 90% of cases of small cell lung cancer. The change takes place very early on, before the structure of the epithelium is significantly affected. Mutations of the TP53 gene are also present at the early stage of respiratory system changes. It is important to note that the diagnostic value of the molecular analysis is greatly enhanced when a combination of several mutation is evaluated [12]. Thus, the identification of mutations appearing in neoplastic tissue provides significant benefits, both in the diagnostics and the therapy of NSCLC. It is, therefore, important for the methods of assessing mutations in lung cancer to be quick, efficient, and accurate [13]. Depending on the type of mutation (point mutation, translocation), these conditions are met by the dideoxy and next-generation (NGS) sequencing methods, real-time polymerase chain reaction (PCR), fluorescence in situ hybridization (FISH), reverse transcriptase PCR, or genotyping with the use of mass spectrometry [43].

Transcriptomics

Detailed knowledge about the genomescale gene expression profile in thousands of clincal samples were obtained by high throughput microarray technology. It allows precise molecular classification of neoplasms and making more reliable disease development predictions, stratifying patients, as well as developing and implementing new therapies directed at specific targets characteristic for different types of neoplasms.

One of the first evaluation of NSCLC transcripts was performed using a microarray. This study compared the gene expression between normal and malignant lung tissue samples in various stages. This study enabled the identification of 50 genes whose differences in expression allow for the classification of patients with first stage NSCLC into groups with higher and lower chances of survival. That 50 genes signature includes genes related to apoptosis (CASP4, P63), cell structure and adhesion (KRT7, LAMB1), cell cycle and growth regulation (BMP2, GAPD, CDC6, STX1A), and cell signaling (ADM, AKAP12, ARHE, GRB7, VEGF), as well as genes coding chaperones (HSPA8), receptors (ERBB2, FXYD3, SLC20A1), enzymes (CSTB, CYP24, FUT3, MLN64, SLC1A6), and proteins related to transcription and translation (COPEB, CRK, RELA) [15].

Another study, also employing microarrays, has enabled the identification of 133 genes, whose expression profile performs the function of a predictive signature, allowing for the prediction of disease recurrence risk in patients at the early stages of NSCLC [16, 17]. The signature of these genes is especially helpful in making decisions concerning the necessity of implementing supplementary chemotherapy in patients with stage 1A tumors after resection. During the first stage of the experiment, a predictive model was created with 93% accuracy, while the accuracy of the predictive model based on clinical data was only 64%. During the second stage, two validation studies were conducted, the predictive accuracy of which was 72% and 79%, respectively [16]. Further research on the signature helped to identify 5 genes whose level of expression was correlated with the overall survival time and recurrence-free survival time of patients with NSCLC: DUSP6, MMD, STAT1, ERBB3, and LCK [18]. Signatures of high and low risk were identified based on the expression of these five genes. Patients with high risk signature exhibited median overall survival of 20 months, while patients with low risk signatures had median overall survival of 40 months. Moreover, the high risk signature was related to shorter median recurrence-free survival in comparison to the low risk signature (13 months and 29 months, respectively) [18].

The published results of research on gene expression in primary lung tumors samples (adenocarcinoma and squamous cell carcinoma), performed using microarray platforms (Affymetrix), helped identify 91 genes which may serve as survival indicators [19]. Selected genes play a rolein different biological processes, the main ones including: the communication of signals, regulation of transcription, cell cycle, adhesion and proliferation of cells. Some of these genes have previously been identified as predictive markers: DUSP6 and ERBB3 [18], SLC2A1 and MEF2C, AKAP12, CYP24A1, CSTL, SLC2A1, GAPD [15, 19].

Guo et al. proposed a 35-gene signature whose prognostic value enables the stratification of stage 1A NSCLC patients into groups of high and low risk of cancer recurrence [20]. Moreover, this signature was independent of clinical prognostic criteria. No significant relations were found between the level of expression of the 35 studied genes and the age of the patients (over 60 years), tumor differentiation or gender. What is more, it was demonstrated that the 35-gene signature had better prognostic value than other signatures previously used in lung cancer, including the 5-gene signature proposed by Chen et al. and the 133-gene signature proposed by Potti et al. In this study, the 35-gene signature was validated as an independent prognostic factor for NSCLC [20].

Another example of a molecular signature in NSCLC is the three-gene prognostic classifier. It is based on the expression of mRNA genes: STX1A, HIF1A, and CCR7. Verification of the usefulness of the classifier demonstrated that it enables the stratification of patients into groups of various prognosis, and that it is more effective in establishing prognosis than assessment based on histopathological examination [21]. The expression analysis of five genes – ERCC1, BAG-1, BRCA1, RRM1, and TUBB3 – showed that the ERCC1 and

BAG-1 signatures provide both prognostic and predictive information [22]. Patients with reduced ERCC1 expression exhibited longer overall survival by over 21 months and progression-free survival longer by over 27 months. Meanwhile, patients with lowered BAG-1 expression exhibited longer overall survival by over 25 months and progression-free survival longer by over 29 months. Moreover, patients with negative expression of both ERCC1 and BAG-1 responded better to therapy combined with cisplatin or carboplatin [22].

The latest transcriptomic trends utilizes next-generation sequencing technologies (RNA-seq) [23]. Expression profiles of mRNA acquired from bronchoscopy material were compared between healthy smokers and healthy individuals who had never smoked, as well as between healthy smokers and smokers suffering from lung cancer. Smokers exhibited increased expression of the genes related to the xenobiotic metabolic pathway including cytochrome P450, retinol metabolism and oxidoreductase activity. In smokers with lung cancer, the increased expression concerned the genes related to the chemokine signaling pathway and to the interaction of cytokine-cytokine receptors, as well as the genes coding cell adhesion molecules. Apart from RNA-seq, gene expression was additionally assessed using the microarray and qRT-PCR methods. However, only the RNA-seq technique allowed for the identification of additional transcripts whose changes in expression were correlated with smoking: S100A8 and S100A9 (in the group of healthy non-smokers and healthy smokers), as well as of those correlated with the occurrence of lung cancer: MUC5AC and SCGB3A1 (in the group of smokers with and without lung cancer) [23]. Recent studies put also attention to importance of the tumor microenvironment for cancer development, progression, and ability to metastasize. One of the conducted experiments compared the gene expression in a primary culture of fibroblasts acquired from an NSCLC tumor stroma and in normal fibroblasts collected from the same patient [24]. Thanks to microarray analysis, 46 genes involved in the TNF signaling pathway were identified; the expression of these genes was different in the normal fibroblasts than in the fibroblasts associated with the tumor. Repeated validation of the microarray technology and the correlation of expression results with patient survival helped narrow down the prognostic signature to 11 genes: ICAM-1, THBS2, MME, OXTR, B3GALT2, EVI2B, MCTP2, the expression of which increased, and PDE3B, CLU, COL14A1, GAL, the expression of which was reduced. In the next step of this experiment, a data-mining comparison of cancer-associated fibroblasts (CAFs) and normal fibroblasts (NFs) as well as NSCLC tumor’s stroma stroma and normal lung tissue allowed to create 14-gene classifier. Expression level increased in the COL11A1, MFAP5, SULF1, ITGA11, THBS2, and CTHRC1 genes, and decreased in the GPR126, TMOD1, PDE3B, CCDC102B, IGSF10, CLU, FLRT3, and A2M genes. This study confirms the prognostic function of gene expression changes in fibroblasts from the tumor microenvironment [24].

The application of transcriptomics in lung cancer molecular diagnostics

High throughput technologies (gene expression microarrays and RNA-seq) open new possibilities in biomarkers discovering. However, the high cost of research and inaccessible equipment in clinical centers [20] severely limit the use of this technology at the moment and prevent its application in routine diagnostic tests [13]. Moreover, at each stage of experiment there are many factors which may influence analysis results and cause them to be difficult to replicate [25]. In this situation, combining the real time-PCR technique with microarray analysis to analyze genome-scale gene expression seems to be an optimal solution [20].

Epigenetics

Although each cell of an organism has the same set of genes, their expression differs in individual types of tissue, which allows the cells to specialize in different specific functions. The modifications the DNA of a cell is subjected to, aimed at changing the way the DNA is decoded, include DNA methylation, covalent posttranslational modification of histones (e.g. acetylation, methylation, phosphorylation, ubiquitination), and changes in miRNA. Abnormalities in the heredity of epigenetic markers may cause abnormal activation or inhibition of various signaling pathways participating in cell growth regulation, cell differentiation, transformation, as well as apoptosis, and may lead to diseases, such as cancer [26, 27].

Cytosine methylation in CG dinucleotides has a significant influence on gene expression. DNA methyltransferases are the enzymes that create and maintain methylation patterns in cells. They transfer a methyl group onto the fifth carbon atom in a pyrimidine ring of cytosine, creating 5-methylcytosine. In the initiation and progression of a neoplasm, DNA methylation plays significant role: DNA methylation in promoter regions or in the first exons of suppressor genes, causes the silencing of their transcription [28], while hypermethylation of intergenic regions and introns of proto-oncogenes affects chromosome instability [29]. What is more, these modifications are often related to changes in DNA methyltransferase (DNMT) expression [30].

Recent studies on lung cancer points to the diversification of the occurrence frequency of increased suppressor gene promoter methylation in individual types of neoplasms, which may be applicable for diagnostics purpose. The usefulness of the research increases when specific methylation pattern is evaluated for neoplastic cells and normal lung cells. The comparison of the methylation patterns of IMP4, GATA4, SOX18, and EGFL7 in the neoplastic tissue of NSCLC patients, the tissue surrounding the tumor, and normal lung tissue showed that the first three genes exhibit excessive methylation in neoplastic cells and in the normal tissue surrounding the tumor. This may point to the early initiation of epigenetic processes, which precede the subsequent genetic and morphological changes leading to the appearance of neoplastically transformed cells [31]. In previous research, a set of genes was defined in which the increase in promoter sequence methylation was characteristic mainly of neoplastic tissue samples. Methylation of the following genes was identified: RASSF1A (44% in neoplastic tissue/0% in normal tissue), p16 (47%/14%), DLC1 (61%/25%) [32]. Moreover, an increase in RASSF1A methylation was observed in the material obtained from patients who were long-term smokers [32]. RASSF1A inactivation is associated with progression of the neoplasm and worsening of prognosis. Other suppressor genes characterized by increased methylation of promoter regions in lung cancer are: FHIT [28], CDH13 [33, 34], KLK10 [33], EFEMP1 [33], SFRP1 [33], RARβ [33, 34], APC [33], and DAPK [35]. Silencing these genes causes a decrease in the amount of synthesized protein to a level insufficient for normal cell cycle regulation [36].

Changes in the expression of DNA methyltransferases mainly consist in their increased transcription, which leads to the heightening of these enzymes’ activity in neoplastic cells and may be related to the hastening of lung cancer development and the worsening of prognosis [30]. A correlation was found between DNMT1 and DNMT3b overexpression and the increased methylation of suppressor gene promoters, especially in smokers. This may result from methyltransferases combining with protein 2, which binds methylated CpG (MeCP2). This protein prevents transcription factors from binding to a promoter [37].

Modifications of histones are also conducive to the faster development of the disease and lower probability of survival. Di-methylation of 4 lysine and acetylation of 18 lysine in histone H3 may serve as examples. An increase in the overall histone modification level points to higher aggressiveness of the neoplasm; significant similarities of these relations were found in adenocarcinomas acquired from various types of epithelial tissue, including prostate, lung, and renal cancer [38].

The application of epigenetics in lung cancer molecular diagnostics

Epigenetic changes taking place in neoplastic cells can be detected at the early stage of the disease; therefore, they may be used for screening and diagnostic examinations in the future. Moreover, these changes are not permanently inscribed in the genome and it is possible to reverse or modify them. The currently used biomarker is the assessment of MGMT gene methylation in gliomas; unfortunately, the use of epigenetics in lung cancer is still in the research phase and is not yet employed in routine molecular diagnostics. For example, the research concerning miR-29s, the expression of which is inversely correlated with DNMT3A and -3B in lung cancer, indicated that restoring the high expression of miR-29s normalizes methylation pattern deviations and inhibits in vitro and in vivo development of neoplasms [39]. This confirmed the role of miR-29s in the epigenetic normalization of NSCLC and in the development of a new potential strategy for lung cancer treatment [39].

The examination of easily obtainable samples is of great importance in cancer diagnostics. Many scientific studies use fresh neoplastic tissue, usually acquired during surgery or biopsy. The use of blood, saliva, or sputum is less invasive and more beneficial for screening purposes; therefore, many studies follow this direction in the search for biological markers in body fluids. A decrease in the expression of the p16/INK4 gene, related to the increased methylation of its promoter, can be observed in non-small cell lung cancer [40]. In the aforementioned study, the hypermethylation levels of p16/INK4a were compared in the DNA isolated from sputum and bronchial lavage fluid acquired from long-term smokers and lung cancer patients. The cancer patients exhibited an increase in methylation (51%), but this tendency was also found in smokers (28%). The examination of the methylation levels of promoter genes CDKN2A, CDH1, and MGMT in patients with chronic obstructive lung disease and patients with lung cancer, where the used material was sputum, showed an increase in the methylation of CDKN2A and MGMT in both patient groups. CDH1 methylation increased in patients with lung cancer only [41].

Future directions for research and molecular diagnostics

One of the latest trends in genomics, transcriptomics, and epigenetics is the analysis of patient molecular profiles using data mining and correlating these results with the patients’ response to particular types of treatment. This research contributed to the development of targeted therapies and personalized medicine [42]. The next step was the use of high-throughput technologies in molecular cancer diagnostics, which allows for the analysis of larger numbers of genes at lower examination costs [13]. For example, the LungCarta™ Panel (Sequenom, Inc.) enables the analysis of 214 somatic mutations in 26 genes in DNA isolated from lung cancer tissue, while OncoDEEP

(ONCODNA SA) uses next-generation sequencing, providing an initial analysis of 40 genes or a more detailed analysis of up to 406 genes. The above-mentioned analyses not only further the understanding of the molecular changes taking place in cancer tissue, but also provide key molecular information when selecting a personalized, “made-to-measure” therapy for a patient.

Acknowledgements

This work was written as part of the Homing Plus 2010-2/7 grant, awarded by the Foundation for Polish Science, co-financed from the Innovative Economy Programme 2007-2013, implemented in the F. Łukaszczyk Oncology Center in Bydgoszcz.

References

1. Wojciechowska U, Didkowska J, Zatoński W. Nowotwory złośliwe w Polsce w 2010 roku. Centrum Onkologii – Instytut, Warszawa 2012; 1-108.

2. Thomas A, Rajan A, Lopez-Chavez A, Wang Y, Giaccone G. From targets to targeted therapies and molecular profiling in non-small cell lung carcinoma. Ann Oncol 2013; 24: 577-585.

3. Kamel-Reid S, Chong G, Ionescu DN, Magliocco AM, Spatz A, Tsao M, Weng X, Young S, Zhang T, Soulieres D. EGFR tyrosine kinase mutation testing in the treatment of non-small-cell lung cancer. Curr Oncol 2012; 19: e67-e74.

4. D’Angelo SP, Pietanza MC, Johnson ML, Riely GJ, Miller VA, Sima CS, Zakowski MF, Rusch VW, Ladanyi M, Kris MG. Incidence of EGFR exon 19 deletions and L858R in tumor specimens from men and cigarette smokers with lung adenocarcinomas. J Clin Oncol 2011; 29: 2066-2070.

5. Lewandowska MA, Jóźwicki W, Starzynski J, Kowalewski J. Analysis of EGFR mutation frequency and coexistence of KRAS and EGFR mutations using RT-PCR in lung adenocarcinoma: may a clinical and pathological model of a patient’s qualification for targeted therapy have an impact on time to obtain genetic results? Kardiochir Torakochir Pol 2012; 9: 443-451.

6. Krawczyk P, Ramlau R, Powrózek T, Wojas-Krawczyk K, Sura S, Jarosz B, Walczyna B, Pankowski J, Szumiło J, Dyszkiewicz W, Woźniak A, Milanowski J. Wykrywalność mutacji w genie EGFR u chorych na niedrobnokomórkowego raka płuca w wybranych ośrodkach w Polsce zaangażowanych w diagnostykę molekularną. Kardiochir Torakochir Pol 2012; 9: 431-438.

7. Soda M, Choi YL, Enomoto M, Takada S, Yamashita Y, Ishikawa S, Fujiwara S, Watanabe H, Kurashina K, Hatanaka H, Bando M, Ohno S, Ishikawa Y, Aburatani H, Niki T, Sohara Y, Sugiyama Y, Mano H. Identification of the transforming EML4-ALK fusion gene in non-small-cell lung cancer. Nature 2007; 448: 561-566.

8. Zhang X, Zhang S, Yang X, Yang J, Zhou Q, Yin L, An S, Lin J, Chen S, Xie Z, Zhu M, Wu YL. Fusion of EML4 and ALK is associated with development of lung adenocarcinomas lacking EGFR and KRAS mutations and is correlated with ALK expression. Mol Cancer 2010; 9: 188.

9. Suehara Y, Arcila M, Wang L, Hasanovic A, Ang D, Ito T, Kimura Y, Drilon A, Guha U, Rusch V, Kris MG, Zakowski MF, Rizvi N, Khanin R, Ladanyi M. Identification of KIF5B-RET and GOPC-ROS1 fusions in lung adenocarcinomas through a comprehensive mRNA-based screen for tyrosine kinase fusions. Clin Cancer Res 2012; 18: 6599-6608.

10. Takeuchi K, Soda M, Togashi Y, Suzuki R, Sakata S, Hatano S, Asaka R, Hamanaka W, Ninomiya H, Uehara H, Lim Choi Y, Satoh Y, Okumura S, Nakagawa K, Mano H, Ishikawa Y. RET, ROS1 and ALK fusions in lung cancer. Nat Med 2012; 18: 378-381.

11. Ju YS, Lee WC, Shin JY, Lee S, Bleazard T, Won JK, Kim YT, Kim JI, Kang JH, Seo JS. A transforming KIF5B and RET gene fusion in lung adenocarcinoma revealed from whole-genome and transcriptome sequencing. Genome Res 2012; 22: 436-445.

12. Chorostowska-Wynimko J, Szpechcinski A. The impact of genetic markers on the diagnosis of lung cancer: a current perspective. J Thorac Oncol 2007; 2: 1044-1051.

13. Poznański M, Lewandowska M. Zastosowanie wysokowydajnych technologii w diagnostyce molekularnej chorób nowotworowych. Onkol Prakt Klin 2013; 9: 70-77.

14. Schena M, Shalon D, Davis RW, Brown PO. Quantitative monitoring of gene expression patterns with a complementary DNA microarray. Science 1995; 270: 467-470.

15. Beer DG, Kardia SL, Huang CC, Giordano TJ, Levin AM, Misek DE, Lin L, Chen G, Gharib TG, Thomas DG, Lizyness ML, Kuick R, Hayasaka S, Taylor JM, Iannettoni MD, Orringer MB, Hanash S. Gene-expression profiles predict survival of patients with lung adenocarcinoma. Nat Med 2002; 8: 816-824.

16. Potti A, Mukherjee S, Petersen R, Dressman HK, Bild A, Koontz J, Kratzke R, Watson MA, Kelley M, Ginsburg GS, West M, Harpole DH Jr, Nevins JR. A genomic strategy to refine prognosis in early-stage non-small-cell lung cancer. N Engl J Med 2006; 355: 570-580.

17. Potti A, Mukherjee S, Petersen R, Dressman HK, Bild A, Koontz J, Kratzke R, Watson MA, Kelley M, Ginsburg GS, West M, Harpole DH Jr, Nevins JR. Retraction: A genomic strategy to refine prognosis in early-stage non-small-cell lung cancer. N Engl J Med 2006; 355: 570-80. N Engl J Med 2011; 364: 1176.

18. Chen HY, Yu SL, Chen CH, Chang GC, Chen CY, Yuan A, Cheng CL, Wang CH, Terng HJ, Kao SF, Chan WK, Li HN, Liu CC, Singh S, Chen WJ, Chen JJ, Yang PC. A five-gene signature and clinical outcome in non-small-cell lung cancer. N Engl J Med 2007; 356: 11-20.

19. Chen G, Kim S, Taylor JM, Wang Z, Lee O, Ramnath N, Reddy RM, Lin J, Chang AC, Orringer MB, Beer DG. Development and validation of a quantitative real-time polymerase chain reaction classifier for lung cancer prognosis. J Thorac Oncol 2011; 6: 1481-1487.

20. Guo NL, Wan YW, Tosun K, Lin H, Msiska Z, Flynn DC, Remick SC, Vallyathan V, Dowlati A, Shi X, Castranova V, Beer DG, Qian Y. Confirmation of gene expression-based prediction of survival in non-small cell lung cancer. Clin Cancer Res 2008; 14: 8213-8220.

21. Lau SK, Boutros PC, Pintilie M, Blackhall FH, Zhu CQ, Strumpf D, Johnston MR, Darling G, Keshavjee S, Waddell TK, Liu N, Lau D, Penn LZ, Shepherd FA, Jurisica I, Der SD, Tsao MS. Three-gene prognostic classifier for early-stage non small-cell lung cancer. J Clin Oncol 2007; 25: 5562-5569.

22. Leng XF, Chen MW, Xian L, Dai L, Ma GY, Li MH. Combined analysis of mRNA expression of ERCC1, BAG-1, BRCA1, RRM1 and TUBB3 to predict prognosis in patients with non-small cell lung cancer who received adjuvant chemotherapy. J J Exp Clin Cancer Res 2012; 31: 25.

23. Beane J, Vick J, Schembri F, Anderlind C, Gower A, Campbell J, Luo L, Zhang XH, Xiao J, Alekseyev YO, Wang S, Levy S, Massion PP, Lenburg M, Spira A. Characterizing the impact of smoking and lung cancer on the airway transcriptome using RNA-Seq. Cancer Prev Res (Phila) 2011; 4: 803-817.

24. Navab R, Strumpf D, Bandarchi B, Zhu CQ, Pintilie M, Ramnarine VR, Ibrahimov E, Radulovich N, Leung L, Barczyk M, Panchal D, To C, Yun JJ, Der S, Shepherd FA, Jurisica I, Tsao MS. Prognostic gene-expression signature of carcinoma-associated fibroblasts in non-small cell lung cancer. Proc Natl Acad Sci U S A 2011; 108: 7160-7165.

25. Mager J, Schultz RM, Brunk BP, Bartolomei MS. Identification of candidate maternal-effect genes through comparison of multiple microarray data sets. Mamm Genome 2006; 17: 941-949.

26. Egger G, Liang G, Aparicio A, Jones PA. Epigenetics in human disease and prospects for epigenetic therapy. Nature 2004; 429: 457-463.

27. Feinberg AP, Ohlsson R, Henikoff S. The epigenetic progenitor origin of human cancer. Nat Rev Genet 2006; 7: 21-33.

28. Tomizawa Y, Iijima H, Nomoto T, Iwasaki Y, Otani Y, Tsuchiya S, Saito R, Dobashi K, Nakajima T, Mori M. Clinicopathological significance of aberrant methylation of RARbeta2 at 3p24, RASSF1A at 3p21.3, and FHIT at 3p14.2 in patients with non-small cell lung cancer. Lung Cancer 2004; 46: 305-312.

29. Eden A, Gaudet F, Waghmare A, Jaenisch R. Chromosomal instability and tumors promoted by DNA hypomethylation. Science 2003; 300: 455.

30. Vallböhmer D, Brabender J, Yang D, Schneider PM, Metzger R, Danenberg KD, Hölscher AH, Danenberg PV. DNA methyltransferases messenger RNA expression and aberrant methylation of CpG islands in non-small-cell lung cancer: association and prognostic value. Clin Lung Cancer 2006; 8: 39-44.

31. Azhikina T, Kozlova A, Skvortsov T, Sverdlov E. Heterogeneity and degree of TIMP4, GATA4, SOX18, and EGFL7 gene promoter methylation in non-small cell lung cancer and surrounding tissues. Cancer Genet 2011; 204: 492-500.

32. Dammann R, Strunnikova M, Schagdarsurengin U, Rastetter M, Papritz M, Hattenhorst UE, Hofmann HS, Silber RE, Burdach S, Hansen G. CpG island methylation and expression of tumour-associated genes in lung carcinoma. Eur J Cancer 2005; 41: 1223-1236.

33. Zhang Y, Wang R, Song H, Huang G, Yi J, Zheng Y, Wang J, Chen L. Methylation of multiple genes as a candidate biomarker in non-small cell lung cancer. Cancer Lett 2011; 303: 21-28.

34. Kubo T, Yamamoto H, Ichimura K, Jida M, Hayashi T, Otani H, Tsukuda K, Sano Y, Kiura K, Toyooka S. DNA methylation in small lung adenocarcinoma with bronchioloalveolar carcinoma components. Lung Cancer 2009; 65: 328-332.

35. Tang X, Khuri FR, Lee JJ, Kemp BL, Liu D, Hong WK, Mao L. Hypermethylation of the death-associated protein (DAP) kinase promoter and aggressiveness in stage I non-small-cell lung cancer. J Natl Cancer Inst 2000; 92: 1511-1516.

36. Bird A. DNA methylation patterns and epigenetic memory. Genes Dev 2002; 16: 6-21.

37. Lin RK, Hsu HS, Chang JW, Chen CY, Chen JT, Wang YC. Alteration of DNA methyltransferases contributes to 5’CpG methylation and poor prognosis in lung cancer. Lung Cancer 2007; 55: 205-13.

38. Seligson DB, Horvath S, McBrian MA, Mah V, Yu H, Tze S, Wang Q, Chia D, Goodglick L, Kurdistani SK. Global levels of histone modifications predict prognosis in different cancers. Am J Pathol 2009; 174: 1619-1628.

39. Fabbri M, Garzon R, Cimmino A, Liu Z, Zanesi N, Callegari E, Liu S, Alder H, Costinean S, Fernandez-Cymering C, Volinia S, Guler G, Morrison CD, Chan KK, Marcucci G, Calin GA, Huebner K, Croce CM. MicroRNA-29 family reverts aberrant methylation in lung cancer by targeting DNA methyltransferases 3A and 3B. Proc Natl Acad Sci U S A 2007; 104: 15805-15810.

40. Kersting M, Friedl C, Kraus A, Behn M, Pankow W, Schuermann M. Differential frequencies of p16(INK4a) promoter hypermethylation, p53 mutation, and K-ras mutation in exfoliative material mark the development of lung cancer in symptomatic chronic smokers. J Clin Oncol 2000; 18: 3221-3229.

41. Guzmán L, Depix MS, Salinas AM, Roldán R, Aguayo F, Silva A, Vinet R. Analysis of aberrant methylation on promoter sequences of tumor suppressor genes and total DNA in sputum samples: a promising tool for early detection of COPD and lung cancer in smokers. Diagn Pathol 2012; 7: 87.

42. Lewandowski R, Roszkowski K, Lewandowska MA. Personalized medicine in oncology: vision or realistic concept? Wspolczesna Onkol 2011; 15: 1-6.

43. Lewandowska MA, Jóźwicki W, Zurawski B. KRAS and BRAF Mutation Analysis in Colorectal Adenocarcinoma Specimens with a Low Percentage of Tumor Cells. Mol Diagn Ther 2013; 17: 193-203. doi: 10.1007/s40291-013-0025-8.