Lung cancer

Abstract

Epigenetic is the study of heritable changes in gene expression that occur without changes in DNA sequence. This process is important for gene expression and genome stability and its disruption is now thought to play a key role in the onset and progression of numerous tumor types. The most studied epigenetic phenomena includes post-translational modifications in DNA and histone proteins as well as microRNAs expression. As epigenetic aberrations are potentially reversible, their correction has emerged as a potential strategy for the treatment of cancer. This review highlights the roles of chromatin epigenetic modifications and of microRNAs expression in lung tumorigenesis and discusses the emerging epigenetic therapies which are being developed for the treatment of lung cancer.

Cancers are traditionally viewed as genetic disorder. However it has become increasingly evident that aberrant epigenetic modifications also play major roles in the tumorigenic process. Epigenetic defines heritable and reversible modifications of gene expression without changes in the nucleotidic sequence. The three most studied epigenetic phenomena are (1) DNA methylation, (2) chromatin modifications (including post-translational modifications of histones and chromatin modifying complexes) and (3) non-coding RNAs mediated regulations. Epigenetic processes are finely tuned, undergo many regulations in response to environment and involve all the signaling pathways described so far. Epigenetic plays a crucial role in the control of nuclear architecture and gene activity and constitute one of the basis of the biological diversity. In this review, we focus on the growing number of publications describing the huge aberration of the epigenetic landscape in lung cancer cells (Table 1). Additionally, we outline advances in the potential use of these epigenetic events for cancer diagnostic, prognostic and targeted epigenetic therapy.

Table 1.

Major epigenetic changes in lung cancer

Epigenetic changes	Targets	Cancer type	Ref	Proposed or effective therapy
DNA methylation
CpG islands hypermethylation	p16INK4a	NSCLC	5, 6, 36, 37–39
	CDH13	NSCLC	7, 39
	14-3-3s	SCLC, NSCLC	8	Demethylating agents (41)
	DAPK1	SCLC, NSCLC	5, 9, 38
	RASSF1A	SCLC, NSCLC	10, 39	DNA methyltranferase inhibitors
	Caspase8	SCLC, NSCLC	11
	RAR-b	NSCLC	12	Demethylated agents in combination with HDACi
	TIMP3	NSCLC	12
	MGMT	NSCLC	12, 36, 37
	ECAD	NSCLC	12	Demethylated agents in combination with chemotherapy
	GSTP1	NSCLC	12
	APC	NSCLC	38, 39
	FHIT	NSCLC	38
Genome wide hypomethylation		NSCLC	21–24
Chromatin modifications
Histone modification
Acetylation	H2AK5ac	NSCLC	54	HDACi (56, 60)
	H3K9ac	NSCLC	54
	H3K18ac	NSCLC	42	HDACi in combination with ionizing radiation (57)
	H4K5ac	NSCLC	44
	H4K8ac	NSCLC	44
	H4K12ac	NSCLC	44	HDACi in combination with Chemotherapy or radiotherapy (56, 58–60)
	K4K16ac	NSCLC	44
Methylation	H3K4me2	NSCLC	42, 54
	H4K20me3	NCSLC	44
Chromatin remodelling complexes
	BRG1	NSCLC	63–67
Micro-RNAs
Reduced expression	Let7	NSCLC	72	Chromatin modifying drugs (HDACi.. …)
	miR-128b	NSCLC	75
Overexpression	miR-17-92	NSCLC	70, 76	synthetic oligonucleotides
	miR-155	NSCLC	71	anti-miRNRNA oligonucleotides (antagomirs)
	miR-21	NSCLC	92

DNA Methylation and Lung Cancer

DNA methylation is the best known and the most widely studied epigenetic modification. Methylation is the only covalent DNA modification described in mammals and is restricted to cytosines which are followed by guanines residues, a motif called CpG dinucleotide.¹ While CpG dinucleotides seem depleted from the vast majority of the genome, they are found enriched in regions known as “CpG islands.” Typically, CpG islands are genomic regions of at least 200 bp and up to several Kb in length characterized by a high GC percentage. CpG islands are mainly found near or at the transcription start site within the promoter of ∼40% of mammalian genes. CpG islands play a major role in the process of transcriptional regulation, the unmethylated status of a CpG-island correlating with the ability of a gene to be transcribed in the presence of the required co-regulators. Most CpG islands are usually unmethylated in normal cells allowing gene transcription. In contrast, CpG sites are usually methylated throughout the genome. Methylation is a normal physiological function in the cell, involved in embryonic development,² genomic imprinting³ and chromosome-X inactivation.⁴ In cancer, frequent alterations in DNA methylation are observed and include: (1) locus-specific hypermethylation (at CpG island) which often occurs at tumor suppressor gene loci and leads to the loss of their expression, (2) genome-wide hypomethylation mainly found in the body of genes and in DNA repetitive sequences leading to genomic instability, and (3) altered DNA methyltransferases (DNMTs) expression.

In human lung cancer, promoter DNA hypermethylation is involved in the silencing of various tumor suppressor genes. The best studied example is the case of the CDKi p16^INK4a (CDKN2A), which promoter hypermethylation prevents the negative control exerted by p16^INK4a on RB phosphorylation thereby promoting cell cycle progression. p16^INK4a hypermethylation is considered as one of the earliest event in lung tumorigenesis and increases constantly with disease progression.^5,6 Other examples include H-CADHERIN (CDH13),⁷ 14-3-3σ,⁸ DEATH ASSOCIATED PROTEIN KINASE 1 (DAPK1),⁹ RAS ASSOCIATION DOMAIN FAMILY 1 gene (RASSF1A),¹⁰ CASPASE-8,¹¹ RETINOIC ACID RECEPTOR β-2 (RAR-β), TISSUE INHIBITOR of METALLOPROTEINASE 3 (TIMP3), O⁶-METHYLGUANINE DNA METHYLTRANSFERASE (MGMT), E-CADHERIN (ECAD) and GLUTHATIONE S-TRANSFERASE P1 (GSTP1).¹² As those genes are involved in a broad range of biological processes, promoter DNA hypermethylation appears as a key event in lung carcinogenesis. Consistently, genome wide analyses have pointed to the fact that the extend of promoter DNA hypermethylation is probably under-appreciated.^13–16 By using a high-throughput global expression profiling approach, Shames et al. recently identified 132 genes that are methylated with high penetrance in lung cancer cells.¹⁶ More strikingly, the analysis performed by Brena et al. supports the notion that 4.8% of all CpG island promoters might be aberrantly methylated, suggesting that the expression of about 1,400 genes might be disturbed in lung cancer.¹³ Overexpression of DNA methyltranferases DNMT1, DNMT3A and DNMT3B has been described in NSCLC especially among smoker patients, and correlates with hypermethylation of tumor suppressor gene such as p16^INK4a, FHIT and RARβ.^17,18 Furthermore, polymorphisms that influence expression of the DNMT3B gene have been connected with risk of lung cancer.^19,20 Genome wide hypomethylation may also have a role in the onset and development of lung cancer. Global DNA hypomethylation is highly associated with the progression of lung tumors and is already detected in the normal part of the lung from cancer patient as compared to a normal individual.²¹ Hypomethylation status within exons 5–8 of p53 from peripheral blood DNA has been reported as a relevant predictor of lung cancer among male smokers.²² Furthermore, extensive DNA hypomethylation at repetitive sequences has also been observed in lung tumors.^23,24

Accumulating evidence argues that epigenetic gene silencing through DNA hypermethylation can actually predispose to mutational events.²⁵ Because of its spontaneous hydrolytic deamination under physiological conditions, methylated cytosine is considered as a potent endogenous mutagen. Whereas methylated cytosine represents only 1% of the bases in the mammalian genome, it is estimated that it might be at the origin of as much as 30% of all transition mutations found in human disease such as cancers.^26,27 Methylation of CpG sites has been reported to facilitate the binding for benzo[a]pyrene found in cigarette smoke, leading to the formation of major DNA damage hotspots in human lung cancer.^28,29 This has been particularly well-illustrated for the occurrence of some hotspot mutations of the p53 tumor suppressor gene in lung tumors.^30,31 Other studies have shown that silencing through promoter hypermethylation of DNA repair genes such as MGMT may predispose to mutation of key genes such as p53 and K-RAS.^32–34 Although this has not been demonstrated in lung cancer it can be envisaged that such mechanism could exist as silencing of MGMT has been associated with p53 mutation in these tumors.³⁵

The development of squamous cell carcinoma can be predicted by p16^INK4a and MGMT promoter methylation up to three years before clinical diagnosis^36,37 and DNA methylation may serve as a marker for the early detection of lung cancer when detected in the sputum of the patient.^5,38 In a multivariate model, Brock et al. have recently shown that promoter methylation of CDKN2A, CDH13, RASSF1A and APC is associated with early recurrence in patients with stage I NSCLC.³⁹ Altogether, theses studies highlight promoter methylation as a promisive epigenetic approach for early detection and prognosis of NSCLC. In addition, demethylating drugs have a great and promising clinical potential as their use restores the expression of epigenetic silenced genes and inhibits tumor cell growth.⁴⁰ They induce manageable short-term side effects at doses showing therapeutic efficacy⁴⁰ although their long-term effects remain to be fully evaluated. The 5-aza-2′-deoxycytidine demethylating agent has been reported to increase the survival of NSCLC patients in the absence of prior chemotherapy, up to 6 years in some cases.⁴¹ Current investigations are aimed at combining epigenetic therapies (i.e., hypomethylating agents with histone deacetylase inhibitors) and at attempting to integrate epigenetic therapy with more standart therapy.

Chromatin Modifications in Lung Cancer

The fundamental unit of chromatin is called the nucleosome. This specialized regulatory structure consists of 147 bp of genomic DNA wrapped around an histone octamere composed of two copies of each of the core histones H2A, H2B, H3 and H4. Each of these core histone possesses a globular domain as well as an amino-terminal tail which protrudes outside of the nucleosome and in that way is accessible to numerous nuclear regulators. Amino-terminal tails of histone proteins are subjected to wide variety of post-translational covalent modifications such as acetylation, methylation, phosphorylation and ubiquitinylation. These modifications can either favor (e.g., acetylation, methylation) or inhibit (e.g., methylation) the access to the chromatin and are involved in all DNA-based processes described so far. Clearly, global alterations of histone modification patterns have the potential to affect the structure and integrity of the genome and to disrupt normal patterns of gene expression. Several protein complexes involved in transcription regulation function by modifying histones or altering chromatin structure. They are mainly represented by Histones AcetylTransferases (HATs)/Histones Deacetylases (HDACs) and Histones methyltransferases (HMTs)/Histones Demethylases (DHMTs) complexes that determine the level of acetylation and methylation respectively of the amino-terminal domains of nucleosomal histones associated with them, and by ATP-dependent complexes such as SWI/SNF which use the energy of ATP hydrolysis to locally disrupt or alter the association of histones with DNA. The last years have supported the notion that histones modifications may contribute to tumorigenesis^42,43 and have shown an altered expression pattern of histone and chromatin modifying enzymes in human tumors.

Histones modifications.

Recently, we demonstrated the existence of a global post-translational modifications profile of histone H4 in a comprehensive panel of normal lung tissue and primary lung tumors. In this study, cancer cells exhibited a gain of H4K5ac and H4K8ac and a loss of H4K12ac, H4K16ac and H4K20me3.⁴⁴ Interestingly, loss of H4K20me3 already occurs in early precursors lesions of squamous cell carcinoma and increases with disease progression. The finding that these changes occurs so soon in the course of lung tumorigenesis indicate that they might be relevant steps in the transformation process. Some clues as to how these cancer specific histone modifications arise are emerging. The demonstration that loss of H4K20me3 correlates with decreased expression of the histone methyltransferase SUV4-20H2,⁴⁴ fits well with a recent study showing that Suv4-20 h-double-null mice have lost nearly all H4K20me3 and H4K20me2 states.⁴⁵ Similarly, reduced mRNA levels of the MYST Tip60 HAT⁴⁶ could explain the hypoacetylation of Histone H4K16, as this enzyme mediates the H4K16 acetylation.^47,48 Aberrant expression of other histone modifying enzymes has also been reported to discriminate lung tumor samples from their normal counterparts. For instance, mutations and deletions of the CBP HAT gene,⁴⁹ variable levels of HDAC1-10^50,51 or overexpression of the MAPJD and JMJD2C HDMs have been described.^52,53 Whether such aberrant expression patterns could correlate with specific histone modifications remains to be determined.

Analysis of histone modifications is able to cluster the tumor samples according to their histological type suggesting that post-translational modifications of histones could be a nice alternative for the diagnosis of lung cancer.⁴⁴ Changes in global level of individual histone modifications also influence prognosis. Loss of H4K20me3 correlates with reduced survival in patients with stage I adenocarcinoma.⁴⁴ The epigenetic pattern of H3K4me2, H2AK5ac and H3K9ac influences the clinical outcome of NSCLC patients, especially in early-stage tumors.⁵⁴ The cellular levels of H3K4me2 and H3K18ac have also been reported as independent predictors of clinical outcome in lung adenocarcinoma, and may be general predictors of clinical outcome in adenocarcinomas of different tissue origins.^42,55 Further studies are needed to analyze whether distinct combinations of histone modifications might define “tumor signature profiles” that could be used as complementary diagnostic tools, prognostic factors and predictors of responses to treatment. Inhibitors of HDAC (HDACi) have emerged as novel and promising anticancer agents⁵⁶ and several molecules have been shown to increase the cytotoxic effects of radiation in NSCLC by decreasing DNA repair efficiency and promoting cell death.⁵⁷ HDACi also display benefits when used in combination with standard NSCLC chemotherapeutic agents and are likely to be a novel approach for the treatment of NSCLC because of an anti-growth activity against NSCLC cells.^58,59 Phases I and II clinical trials with HDACi in the treatment of advanced NSCLC have been completed⁵⁶ and ongoing clinical trials are exploring the use of many new HDACi singly or as part of a combination with others therapeutic modalities such as chemotherapy or radiotherapy.⁶⁰ Clearly, HDACi have a specific antitumor effect and thorough studies analyzing the full potential and mechanism of these drugs with regards to optimal dose, schedule, patient selection and combination strategies would allow the development of molecules with more effective therapeutic effect.

Alteration of chromatin-remodeling complexes.

BRG1 is a catalytic component of the SWI/SNF chromatin-remodeling complex and regulates gene expression by disrupting DNA-histone interactions at the nucleosomes in an ATP-dependent manner.⁶¹ This protein has been proposed to be a tumor suppressor and mice studies support a role for BRG1 loss in lung cancer development. Indeed, inactivation of BRG1 enhances the tumorigenic effect of carcinogens and induces the development of lung adenomas.⁶² In human lung cancer, loss of heterozygosity surrounding the BRG1 loci and somatic point mutations of BRG1 have been described.^63,64 As a result, loss of BRG1 protein is observed in some NSCLC⁶⁵ and correlates with reduced survival.^66,67

Micro-RNAs for Major Biological Insights in Lung Cancer

Micro-RNAs are small non-coding RNAs initially transcribed as large RNA precursor (called pri-miRNA) which are processed into a ∼70 nucleotide pre-miRNA and exported to the cytoplasm to undergo final processing steps to obtain a mature miRNA of ∼22 nucleotides length.⁶⁸ Depending on the degree of homology to their 3′UTR target sequence, miRNAs induce translational repression or degradation of mRNAs. It is estimated that 1,000 miRNAs are transcribed and that 30% of the human genome is under miRNA regulation, one miRNA being able to modulate post-transcriptionally hundreds of downstream genes. In this regard, miRNAs control a wide range of biological processes including apoptosis, development, proliferation and differentiation.⁶⁹ High-throughput analyses have highlighted aberrant miRNAs expression profiles in an increasing range of human cancer types^70–72 and all these studies support the same view: the alterations seen in cancer cells that express miRNAs consist of both overexpressed and downregulated miRNAs. Therefore, miRNAs may function either as tumor suppressors or oncogenes and the genomic abnormalities found to influence their activity are the same as those described for protein-coding genes. To date, both “in vivo” and “in vitro” studies demonstrate a deregulation of miRNA expression in lung cancer and highlight them as useful diagnostic, pronostic and therapeutic tools.

A growing number of miRNAs has been found aberrantly expressed in lung cancer and our understanding of miRNAs expression patterns and function in normal and lung cancer cells is just starting to emerge. One of the first miRNAs identified is Let7 which appears to be important in lung cancer. Indeed, over-expression of let-7 inhibits Ras protein expression⁷³ and represses proliferation of lung cancer cells both “in vitro” and “in vivo,”⁷⁴ identifying let-7 as a tumor suppressor. This is confirmed by clinical data as reduced expression of Let-7 miRNA is observed in primary lung tumors.⁷² Other miRNAs with tumor suppressor function include miR-128b which is a direct negative regulator of the EGFR oncogene and which expression is lost in lung tumors.⁷⁵ As an example of oncogenic miRNAs is the miR-17-92 cluster which upregulation is observed in lung cancer cells.^70,76 Overexpression of miR-17-92 positively stimulates cell proliferation^76,77 inhibits differentiation of lung epithelial progenitor cells in transgenic mice.⁷⁸ Expression of the E2F1 transcription factor is negatively regulated by miR-17-92,⁷⁹ suggesting that its differential pattern in lung tumors⁸⁰ could rely on aberrant expression of miR-17-92. Predicted regulatory targets of the miR-17-92 cluster also include the PTEN and RB2 tumor suppressors⁸¹ that are known to play important roles in lung cancer. More fundamentally, abrogation of global miRNA processing through targeted silencing of components of the miRNA machinery promotes lung tumorigenesis⁸² suggesting that global decrease of miRNA expression causally contributes to the transformed phenotype. Mice exposed to cigarette smoke exhibit variations in miRNA profiles expression (including let-7 and the p53 tumor suppressor responsive miRNA miR-34) especially during the weaning period.^83,84 These results demonstrate that miRNA alterations occur as an early response to environmental carcinogens “in vivo”, before the onset of cancer. The precise mechanisms regulating miRNA expression in lung tumors is largely unknown but the few existing studies suggest that genetic and epigenetic alteration might affect miRNA status. MicroRNA-128b is located on chromosome 3p, which allelic loss is the most frequent and earliest genetic event in lung carcinogenesis. The cluster miR-17-92 is located at chromosome 13q31 in a region amplified in lung cancer. Expression of miRNA-124a is epigenetically silenced by DNA hypermethylation⁸⁵ and DNMTi restores miRNA-124a expression. Abnormalities in miRNA-processing genes might also be involved in aberrant miRNA pattern as decreased levels of Dicer expression are observed in lung tumors with a significant prognostic impact on the survival of surgically treated patients.⁸⁶

It is becoming apparent that miRNA expression profiles confer important clues for clinical diagnosis and prognosis of human lung cancer. MicroRNA microarray analyses have identified statistical profiles which could discriminate lung cancers from noncancerous lung tissues, as well as molecular signatures that differ according to tumor histology.^71,87 Also interesting is the recent identification of Has-miR-205 as a highly specific marker for squamous carcinoma⁸⁸ suggesting that a clinical diagnostic assay based on miR-205 expression levels could aid for differential diagnosis of NSCLCs. Aberrant miRNA expression can be used as a marker for the diagnosis of NSCLC in sputum specimen⁸⁹ and detection of miRNA expression in peripheral blood or in serum is a good indicator of its expression in the tumor sample.^90,91 As miRNAs are more stable than mRNA and more tissue specific than DNA, their measurement could provide a novel promising non invasive approach to discriminate between normal and cancer patient samples. Several miRNAs are reported to be associated with the clinical outcome of lung cancer. Reduced Let-7 expression correlates with shorter survival in both univariate an multivariate analyses^71,72 and is an independent prognostic factor for the stage of the disease.⁷² Overexpression of miR-155 correlates with a poor prognosis when all clinical variables are considered together.⁷¹ Overexpression of miR-21 is an independent negative prognostic factor for overall survival in NSCLC patients.⁹² Importantly, a five-miRNA signature (let-7a, miR-221, miR-137, miR-372 and miR-182) has been recently associated with survival and cancer recurrence in NSCLC patients.⁹³ Remarkably, this signature is valuable even after patients stratification by stage or histology.

The potential for using miRNAs in lung cancer therapy is now being explored. Let-7 overexpression confers radio-sensitivity to lung cancer cells “in vitro.”⁹⁴ miR-128b LOH, a direct regulator of EGFR, correlates with clinical response and survival following gefitinib treatment.⁷⁵ miR-221, miR-222 and miR-17-92 alter the phenotype of lung cancer cells and sensitize them to cytotoxic agents.^76,77,95 Such results offer the experimental bases for the use of miRNAs as therapeutic targets. Further experiments are needed to uncover the emerging power of small non-coding RNAs to improve lung cancer therapeutics, and would have significant consequences for cancer patients in the clinical area.

Concluding Remarks

It is becoming increasingly clear that disruption of epigenetic processes promotes lung tumor development and growth, and evidence is accumulating that epigenetic changes may account for some of the heritable effects of cigarette smoking. The management of aberrant epigenetic states as a way to target early tumor development as well as tumor progression is therefore a logical therapeutic approach. In the future, developing new strategies to avoid the pleiotropic properties of anti-cancer drugs, such as DNMTi or HDACi, will be of particular interest. Indeed, the side effects of these both compounds could have unscheduled consequences in term of genes expression, in that they may display growth-promoting effects on tumor cells. Otherwise, as the number of miRNAs increases constantly and the expression of many of them is reduced in lung cancer cells, targeting miRNA is also a promising strategy in term of cancer treatment and further studies are required to uncover the potential usefulness of chromatin modifying drugs in restoring the loss of expression of tumor suppressor miRNAs. In this context, administration of synthetic oligonucleotides that mimic endogenous miRNAs, might also be used to treat specific tumor types. Conversely, targeting oncogenic miRNAs through administration of anti-sense oligonucleotides, called AMO (anti-miRNA oligonucleotides) will become into focus, given that the use of antagomirs, which are AMOs conjugated with cholesterol, has recently emerged as an efficient approach to inhibit miRNA activity.⁹⁶ In the coming years, a more complete dissection of the cellular and molecular pathways controlled by epigenetic process will undoubtedly provide novel insights into tumor related mechanisms and will highlight promising fields for the development of novel therapies to fight lung cancer.

Abbreviations

NSCLC

non small cell lung cancer

CDKi

cyclin-dependent-kinase inhibitor

HAT

histone acetyltransferase

HDAC

histone deacetylase

HDACi

histone deacetylase inhibitor

DNMT

DNA methyltransferase

miRNA

microRNA