Abnormal genome-wide DNA methylation induced by cisplatin may contribute to the chemo-resistance of human small cell lung cancer


 Background: So far, the platinum-based chemotherapy (e.g. cisplatin-etoposide doublet) is still the backbone for SCLC management due to its high respond rate both in LS-SCLC and ES-SCLC. However, cisplatin treatment often results in the development of chemo-resistance, leading to therapeutic failure and becoming the main obstacle to improve the therapeutic efficacy. Currently, little has been known about the genome-wide abnormal methylation of SCLC induced by cisplatin, which might provide prospective layouts to discover the potential genes and the signal pathways related with chemo-resistance of SCLC. 
Results: A total of 58,401 sites was identified to be differentially methylated (|Δβ| ≧ 0.20) in H446/DDP cells compared with that of H446 cells, of which 25,991 genes were found to be hypomethylated and 32,410 genes were shown to be hypermethylated. KEGG enrichment displayed that the differentially hypomethylated genes were mainly gathered in MAPK signaling pathway, ECM-receptor interaction, and Focal adhesion, while the differentially hypermethylated genes were clustered in Neuroactive ligand-receptor interaction, Type I diabetes mellitus, Focal adhesion, Allograft rejection, ECM-receptor interaction, CAMs, Graft-versus-host disease, Intestinal immune network for IgA production, ARVC, and Viral myocarditis (KEGG enrichment, qvalue < 0.05). Among the 152 genes which were selected as the MDR-related candidate genes for qRT-PCR to testify whether the abnormal methylation regulated the expression of related genes at the mRNA level, 69 hypomethylated genes were revealed to be significantly increased and the other 54 hypermethylated genes evidently decreased in H446/DDP cells compared with that of H446 cells. Moreover, the upregulated genes with the hypomethylated sites were found to be mainly clustered in Pathways in cancer, MAPK signaling pathway, Cytokine-cytokine receptor interaction, and Cell adhesion molecules (CAMs), while the downregulated genes with the hypermethylated sites were mainly clustered in MAPK signaling pathway, Pathways in cancer, Melanoma, Osteoclast differentiation, and Prostate cancer. 
Conclusions: Cisplatin could induce a large-scale abnormal methylation in the whole genome of SCLC. Pathways in cancer, MAPK signaling pathway, Cytokine-cytokine receptor interaction, and Cell adhesion molecules (CAMs) were most likely to be affected by cisplatin via. methylation to contribute to the development of chemo-resistance and other malignant biological behavior of SCLC cells.

biological behavior of SCLC cells.

Background
Global Cancer Statistics 2018 demonstrates that lung cancer remains the most commonly diagnosed cancer with a proportion of 11.6% in the total cases and the leading cause of cancer death with 18.4% of the total cancer deaths in both sexes combined worldwide [1]. The morbidity rate will rise swiftly to reach as many as one million in 2025 if no effective measures were taken [2][3][4]. Globally, non-small cell lung cancer (NSCLC) comprises approximately 80-85% of all lung cancers with adenocarcinoma and squamous cell carcinoma comprising the predominant histological subtypes of NSCLC and small cell lung cancer (SCLC, also known as oat-cell carcinoma) derived from bronchial epithelial cells accounts for 13-15% of all diagnosed lung cancers [5]. Compared with the major subtypes of NSCLC, the prognosis of SCLC is extremely poor in that SCLC is an aggressive high-grade neuroendocrine tumor asso ciated with a rapid doubling time and a high growth fraction combined with the early development of widespread metastases (most commonly to the brain, liver, or bone) resulting in a 95% mortality rate, which makes SCLC the most lethal lung cancer subtype [6]. Additionally, SCLC has the strongest association with smoking, with only 2% of cases occur ring in never-smokers [7], leading to a high load of somatic mutations induced by tobacco carcinogens [8,9].
SCLCs were classified into limited stage (LS) and extensive stage (ES) according to the Veterans Administration Lung Study Group (VALG) staging system. Current standard of treatment is concurrent chemoradiation for LS-SCLC and chemotherapy alone for ES-SCLC. In recent years, advances in the understanding of the high mutational burden of SCLC and SCLC biology have provided opportunities for therapeutic inter vention and led to the development of novel experimental therapies including targeted agents and immunotherapies [10]. For instance, Poly (ADP-ribose) polymerase (PARP) inhibitor talazoparib are under clinical investigation in combination with cytotoxic therapies and inhibitors of cell-cycle checkpoints. The reported Objective Response Rate (ORR) was 9% and the clinical benefit rate at ≥16 weeks was 26% [11]. Targeting of histone-lysine N-methyltransferase enhancer of zeste 2 polycomb repressive complex 2 subunit (EZH2) was found to maintain the sensitivity of SCLC xenografts to chemotherapy by preventing schlafen family member 11 (SLFN11) silenc ing [12]. High expression of the inhibitory Notch ligand Delta-like protein 3 (DLL3) in most SCLCs encouraged an anti-DLL3-antibody-drug conjugate for preclinical and clinical activity [13].
Additionally, though distinct from that of other solid tumors, few tumor-infiltrating lymphocytes and low levels of the immune-checkpoint protein programmed cell death 1 ligand 1 (PD-L1) were detected in SCLC, a number of clinical trials of this promising immuno therapeutic approaches, such as targeting the inhibitory immune-checkpoint proteins cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) [14] and PD-1 or its ligand PD-L1 [15,16], are underway. However, generally, in contrast to the rapidly changing status of NSCLC, which has notched success after success with a spate of targeted agents and immunotherapies, SCLC has been notorious for its lack of progress as drug after drug. In fact, inhibitors of VEGF, IGFR, mTOR, EGFR, and HGF has failed and fallen by the wayside due to little or no impact on progression-free survival (PFS) or overall survival (OS) [10,17]. The median OS duration of patients with ES-SCLC is stalled, frustratingly, at < 10 months, with a discouraging 5-year OS of 1-5% [18]. The platinum-based chemotherapy (e.g. cisplatin-etoposide doublet) is still the backbone for SCLC management due to its high respond rate both in LS-SCLC and ES-SCLC [19].
However, the platinum-based chemotherapy is a double-edged sword for SCLC. Without treatment, ES-SCLC is rapidly and invariably fatal within 2 to 4 months [20]. With care of chemotherapy, responses are dramatic (approximate 90% cases are responsive to chemotherapy primitively) but sadly short-lived: SCLC inevitably relapses and the disease recurrence characterized by drug resistance is associated with a median OS often < 6 months [21,22] (the median OS for SCLC patients in the third line setting is 4.7 months [23], a survival rate which has scarcely improved over the last 40 years). Compared to the primary disease, the recurrent SCLC is more aggressive with less response to therapy (e.g. topotecan, a topoisomerase I inhibitor) [24]. So far, no effective treatment regimens have been developed for patients whose disease has progressed after first-and second-line therapy.
Cisplatin, a platinum-derivative agent, exerts anticancer effects via multiple mechanisms, of which the most prominent mode of action involves the generation of DNA lesions followed by the activation of the DNA damage response and the induction of mitochondrial apoptosis [25]. Despite a consistent rate of initial responses, cisplatin treatment often results in the development of chemo resistance, leading to therapeutic failure and becoming the main obstacle to improve the therapeutic efficacy.
Over the past three decades, an intense research has been conducted and several mechanisms that account for the cisplatin-resistant phenotype of tumor cells were explored and classified as pretarget, on-target, post-target, and off-target resistance [26]. The known mechanisms explain the cisplatin resistance at the molecular level to a certain extent, however, regretfully, the therapeutic regimens developed from these reported mechanisms have failed to achieve improved outcomes in SCLC patients [27,28]. Therefore, to explore the other potential chemo-resistant mechanisms of SCLC is of great importance to discover novel chemotherapy agents and improve the efficacy of chemotherapy treatment.
The previous findings that DNA methylation is far more vulnerable than DNA sequence to external factors gave us clue that the epigenetic modification might play a pivotal role in the development of the acquired chemo-resistance of SCLC [29,30]. Actually, DNA methylation status changes have been reported to be the propelling factor in the acquired multidrug resistance (MDR) in glioma cell line SGH-44/ADM [31], chronic myeloid leukemia cells [32], human epithelial ovarian cancer cells [33], and NSCLC [34,35]. Additionally, histone deacetylation of ATP binding cassette subfamily B member 1 (ABCB1) promoter was found to be a potential routine for MDR induction in SCLC [36]. Currently, little has been known about the genome-wide methylation frameworks of the chemo-resistant cells of SCLC, which might provide prospective layouts to discover the potential genes and the signal pathways related with chemo-resistance of SCLC. Thus, this research reported for the first time the genome-wide abnormal methylation pattern of chemo-resistant H446/DDP cells of human SCLC induced by the cisplatin. The analysis revealed that Pathways in cancer, MAPK signaling pathway, Cytokine-cytokine receptor interaction, and Cell adhesion molecules (CAMs) might most likely be

Cell lines and culture
The human SCLC cell line H446 was purchased from the Institute of Biochemistry and Biology, Chinese Academy of Sciences (Shanghai, China). The cisplatin resistant H446/DDP cells was established as previously described [37]. In brief, the progenitor H446 cells were treated with first shock of high-dose cisplatin and then maintaining in lower dose cisplatin. The induced cells were verified to be cross-resistant to hydroxycamptothecin, vincristine, and 5-fluorouracil. Both H446 and H446/DDP cells were cultured in RPMI-1640 medium (Hyclone) containing 10% FBS and 1% streptomycin/penicillin at 37℃ in 5% CO 2 . Cells were passaged using 0.25% trypsin with 0.1% EDTA (Hyclone) when attaining to 90% confluence or harvested in logarithmic phase of growth for all experiments described below.
Then the cells of each well were incubated with 100 μl fresh culture medium containing 10 μl CCK-8 solution for 1.5 hrs (Cat: C0038, Beyotime, Shanghai, China). The absorbance value at the wavelength of 450 nm was measured. Cells incubated without chemotherapeutic agents were treated as negative controls. IC 50 was calculated using GraphPad Prism 5.0 software.

Detection of MMP, ATP, and ROS
To further validate the chemo-resistance of H446/DDP cells to cisplatin, the mitochondrial membrane potential (MMP), the intracellular ATP levels, and levels of intracellular reactive oxygen species (ROS) were analyzed between H446/DDP and H446 cells after treated with cisplatin. To detect MMP, cells were plated in 6-well plates and allowed to adhere overnight before treatment with cisplatin at 0, 0.5, 1.0, and 2.0 μg/ml, respectively. After treated by cisplatin for 24 hrs, the cells was incubated with 5 μmol/L fluorescent dye JC-1 dimly for 30 min at 37℃ (Cat: C2006, Beyotime, Shanghai, China), washed with PBS to remove the excess dye, and then observed using fluorescence microscopy (Olympus BX51, Japan).
To assay the intracellular ATP level, after seeded into 96-well plates at the density of 5×10 3 cells/well for 24 hrs, the cells were treated with cisplatin (0, 0.5, 1.0, and 2.0 μg/ml, respectively) for 48 hrs.
Then the intracellular ATP levels were detected using a commercial ATP assay kit in accordance with the manual (Cat: S0026, Beyotime, Shanghai, China). In brief, the assay buffer was gently mixed with the substrate at room temperature. The mixed reagent (100 μl) was added into each well and incubated with shaking for 15 min at room temperature. Then the luminescence was measured using a microplate reader (Beckman Coulter SP-Max2300A2).
Levels of intracellular ROS were detected with an oxidation sensitive fluorescent dye DCFH-DA (Cat: S0033, Beyotime, Shanghai, China). In brief, 5×10 5 cells were plated in 12-well plates and allowed to adhere overnight before incubation with different concentrations of cisplatin (0, 1.0, and 2.0 μg/ml, respectively) for 24 hrs. Cells were washed twice with ice-cold PBS to remove medium. Serum free medium of 1 ml with 1 μl DCFH-DA (10 mM) was added to each tube and incubated at 37℃ for 20 min. Subsequently, the DCF fluorescence picture was captured every five other minutes by fluorescence microscopy (Olympus BX51, Japan). and construct the interaction network, respectively.

Gene Expression determined by Quantitative real-time PCR (qRT-PCR)
Total RNA of the cells was isolated using TRIzol in compliance with the manufacturer's  Table S1.

Statistical Analysis
The statistical differences of between-groups were estimated by two-tailed Student's t-test. All statistical analyses were performed using the Statistical Package for Social Science 15 for Windows (SPSS Inc., Chicago, IL, USA). A statistical difference was accepted as significant if p < 0.05. Each experiment was repeated at least three times.

Genome-wide methylation patterns of H446/DDP cells relative to H446 cells
The comparison of the genome-wide methylation data between H446/DDP and H446 cells revealed that a total of 58,401 sites was identified to be differentially methylated (|Δβ| ≧ 0.20) in H446/DDP cells compared with that of H446 cells, of which 25,991 genes were found to be hypomethylated and 32,410 genes were shown to be hypermethylated, indicating that cisplatin could induce a large-scale abnormal methylation in the whole genome of H446/DDP cells. As shown in Fig.2A, most of the methylated sites were distributed in the intervals with lower |Δβ|, especially |Δβ| ≦ 0.4. The less methylated sites were detected with the |Δβ| increased. There is no distribution bias of the hypomethylated and hypermethylated sites in each intervals stratified by the different |Δβ| ( Fig.2A).
To uncover whether the genes with the differentially methylated sites were preferent or clustered on certain chromosomes, the differentially methylated genes were localized on each chromosome of H446/DDP and H446 cells. The localization displayed that there was no distribution preference or cluster of the hypomethylated or hypermethylated genes on each chromosome (Fig.2B), displaying that the methylated sites distributed evenly across the genome.

GO and KEGG enrichment analysis
To investigate whether the genes with the differentially methylated sites were clustered functionally, the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment of these genes with |Δβ| ≧ 0.4 were analyzed according to the latest GO (http://geneontology.org/) or KEGG database (http://www.genome.jp/kegg/), which disclosed that the differentially methylated genes were involved in many important biological functions. When enrich factor ≧ 5.0 was utilized as cut-point for GO enrichment, the genes with hypomethylated sites were found to be mainly clustered  Table 2).
Glomerulus morphogenesis was the only one common term of GO enrichment between the hypomethylated genes and the hypermethylated genes ( Fig.2C, 2D, Table 1, and Table 2).
When enrich factor ≧ 2.0 was utilized as cut-point for KEGG enrichment analysis, the hypomethylated genes were revealed to be enriched in Glycosaminoglycan biosynthesis, African trypanosomiasis, Maturity onset diabetes of the young, Dorso-ventral axis formation, Biosynthesis of unsaturated fatty acids, Hedgehog signaling pathway, Basal cell carcinoma, and Hypertrophic cardiomyopathy (HCM)  Table 4). Aldosterone-regulated sodium reabsorption, Amoebiasis, Arrhythmogenic right ventricular cardiomyopathy (ARVC), ECM-receptor interaction, Focal adhesion, Glycosaminoglycan biosynthesis, Hypertrophic cardiomyopathy (HCM), and Sulfur relay system were the common terms of KEGG enrichment between the hypomethylated genes and the hypermethylated genes. The GO and KEGG enrichment indicated that many important signal pathways were affected by cisplatin, causing dramatically changes in morphology, structure, function, physiology, and others.
To identify which signal pathway was significantly influenced by cisplatin through methylation, qvalue < 0.05 was used as the cut-point for KEGG enrichment. As shown in Table 5, the differentially hypomethylated genes was mainly clustered in ECM-receptor interaction, MAPK signaling pathway, and Focal adhesion. Differently, the differentially hypermethylated genes was clustered in Neuroactive ligand-receptor interaction, Type I diabetes mellitus, Focal adhesion, Allograft rejection, ECM-receptor interaction, CAMs, Graft-versus-host disease, Intestinal immune network for IgA production, Arrhythmogenic right ventricular cardiomyopathy (ARVC), and Viral myocarditis ( Table 6).
ECM-receptor interaction and Focal adhesion were the two common KEGG enrichment terms between the hypomethylated genes and the hypermethylated genes with qvalue < 0.05.

Network diagram of the differentially methylated genes
The relationship of each gene with the other genes with the differential methylation was analyzed by constructing the network using String 10.5 and Cytoscape 3.3.0. The purple and red circle represented the hypomethylated and hypermethylated genes, respectively ( Fig.3 and Fig.4). The deeper color displayed the greater difference. The line represented the relationship between genes.
Degree indicated the number of genes associated with the other genes. For example, degree = 10 represents that the gene interacts with the other 10 genes. The larger the degree, the more genes that interact with it. As shown in Fig.5A, among the hypomethylated genes, POTEF, ESR1, RAC2, PRKCA, NOTCH1, ARPM1, EFCAB3, BCL2, CACNA1C, and HDAC4 were revealed to be associated with more than the other 10 genes. In contrast, of the hypermethylated genes, HACE1, LRGUK, FYN, ACACB, AR, PIK3CG, ACTBL2, SMAD3, ERBB4, and RUNX1 were revealed to be associated with more than the other 10 genes (Fig.5B).

Expression of 152 genes with the differential methylated sites by qRT-PCR
To testify whether the gene expression was regulated by the methylation at the mRNA level, 152 genes were selected for qRT-PCR based on the beta difference (|Δβ| ≧ 0.7) and the numbers of the methylation sites on each gene. Among the 152 genes, 69 hypomethylated genes significantly upregulated and the other 54 hypermethylated genes evidently downregulated in H446/DDP cells compared with that of H446 cells (Fig.6). It is worth noting that three hypomethylated genes (ESR1, RAC2, and ABCB1) downregulated in H446/DDP cells relative to that of H446 cells (all p < 0.05). The loci of the differentially methylated sites on the three genes were list in Table 7

Discussion
Currently, little has been known about the genome-wide methylation frameworks of the chemo-resistant cells of SCLC, which might provide prospective layouts to discover the potential genes and the signal pathways related with chemo-resistance of SCLC. Hence, we compared the genome-wide methylation profiles between chemo-resistant H446/DDP cells of human SCLC with its progenitor H446 cells in the study. The comparison displayed that a total of 58,401 sites (i.e. promoters and CGIs) was identified to be differentially methylated in H446/DDP cells compared with that of H446 cells, of which 25,991 genes were found to be hypomethylated and 32,410 genes were shown to be hypermethylated and there was no distribution preference or cluster of the hypomethylated or hypermethylated genes on each chromosome, strongly suggesting that cisplatin could cause a genome-wide DNA methylation and the abnormal DNA methylation, one of the most frequent epigenetic alteration, might contribute to the development of chemo-resistance of SCLC.
GO and KEGG enrichment of the differentially methylated genes (|Δβ| ≧ 0.40) revealed that many important biological process, cellular components, molecular function of cells, and signal pathways were affected by the epigenetic alteration (Table 1-4, Fig.2), displaying that chemo-resistance phenotype of SCLC was determined by a very complicated cellular and molecular network. Once the tumor cells developed a chemo-resistant phenotype, the morphology, components, metabolism, and biological process of cells altered accordingly. Notably, among the signal pathways enriched on the basis of enrich factor ≧ 2.0 for KEGG enrichment, Aldosterone-regulated sodium reabsorption, Arrhythmogenic right ventricular cardiomyopathy (ARVC), and ECM-receptor interaction were disclosed to be the common KEGG enrichment terms between the hypomethylated genes and the hypermethylated genes (Fig.2E, 2F, Table 3, and Table 4). Aldosterone, a steroid hormone, regulates renal Na + reabsorption and, therefore plays an important role in the maintenance of salt and water balance [38]. Arrhythmic right ventricular cardiomyopathy (ARVC), also known as arrhythmic right ventricular dysplasia, is an inherited disease characterized by progressive replacement of the myocardium by adipose and fibrous tissue that predisposes to development of ventricular tachycardia (VT) and to sudden cardiac death (SCD) [39]. The abnormal methylation of genes related with Aldosterone-regulated sodium reabsorption and ARVC induced by cisplatin might explain some of the clinical manifestations that the cancer patients often suffer the edema (excess water accumulated in the body), palpitation and shortness of breath (cardiac dysfunction) after treated with cisplatin. Extracellular matrix (ECM) is a three-dimensional, non-cellular structure that constitutes a complex network to regulate the occurrence of tissue, support and connect tissue, and the physiological activities of cells, especially the abscission, adhesion, degradation, migration, and proliferation, the whole process of erosion and metastasis of malignant tumors [40]. The findings that ECM-receptor interaction was the common KEGG enrichment terms between the hypomethylated genes and the hypermethylated genes suggested that the abnormally methylated genes related with ECM-receptor interaction might contribute to the development of cisplatin-resistance of SCLC.
Additionally, Glycosaminoglycan (GAGs) biosynthesis presented the highest enrichment factor among the KEGG enrichment terms of the hypomethylated genes ( Fig.2E and Table 3). GAGs are charged, unbranched polysaccharides consisting of repeating disaccharide units and play roles in various biological events, including cell growth, cytokinesis, and differentiation via. binding to and coordinating the activity of proteins involved in cell attachment, migration and differentiation, neuronal plasticity, blood coagulation, lipid metabolism, and pathogen infectivity. In addition, GAGs carry out mechanical and rheological functions in synovial tissues and fluid [41].
Appreciably different from the hypomethylated genes, Fatty acid biosynthesis presented the highest enrichment factor among the KEGG enrichment terms of the hypermethylated genes ( Fig.2F and Table 4). Fatty acids (FAs), a diverse class of molecules consisting of hydrocarbon chains of different lengths and degrees of desaturation, are used to synthe size many lipids, which are used in energy metabolism and storage and have important roles as signaling molecules. FAs form the hydrophobic tails of phospholipids and glycolipids, which, together with cholesterol, represent major components of bio logical membranes. Additionally, FAs are assembled into triacylglycerides (TAGs), nonpolar lipids that are synthesized and stored during high nutrient availability and that release ample energy when broken down [42]. Tumors have a high rate of glucose uptake and perform glucose fermentation independently of oxygen availability. Tumor cells generate almost all their cellular FAs through de novo synthesis which almost accounted for more than 93% of the FAs biosynthesis and fatty acid synthase (FASN) was identified as the tumor antigen OA-519 in aggressive breast cancer [43,44].
Since then, numerous studies have confirmed the importance of FA biosynthesis for cancer cell growth and survival [42,45]. The GAGs with the highest enrichment factor among the KEGG enrichment terms of the hypomethylated genes and Fatty acid biosynthesis with the highest enrichment factor among the KEGG enrichment terms of the hypermethylated genes indicated that both GAGs and Fatty acid biosynthesis were regulated by the abnormal DNA methylation induced by cisplatin. How GAGS and Fatty acid biosynthesis were involved in the chemo resistance of SCLC is worth exploring.
To further identify which signal pathway was dominantly influenced by cisplatin through methylation, qvalue < 0.05 was used as the cut-point for KEGG enrichment. Among the enriched signal pathways of genes with the hypomethylated loci, MAPK signaling pathway, ECM-receptor interaction, and Focal adhesion were highlighted with qvalue < 0.05 (Table 5), showing that the three pathways were most likely to be affected by DNA hypomethylation induced by cisplatin, which were confirmed by qRT-PCR analysis (Fig.5). The findings gave us clue that the three signal pathway might play vital roles in the chemo-resistance of SCLC. In contrast, Neuroactive ligand-receptor interaction, Focal adhesion, Type I diabetes mellitus, Allograft rejection, ECM-receptor interaction, Cell adhesion molecules (CAMs), Graftversus-host disease, Intestinal immune network for IgA production, Arrhythmogenic right ventricular cardiomyopathy (ARVC), and Viral myocarditis were dominant with qvalue < 0.05 among the enriched signal pathways of genes with the hypermethylated loci (Table 6). It's worth noting that Neuroactive ligand-receptor interaction was highlighted with the most counts of hypermethylated genes (54/614) and the smallest qvalue among the enriched signal pathways of genes with hypermethylated sites (qvalue = 0.004), indicating that Neuroactive ligand-receptor interaction might be inhibited in the chemo-resistant cells, which might interpret the clinical phenomenon that the cancer patients often suffer the headache, numbness and pain of hands and feet (peripheral nerve disorder), and neurology and depression, after treated with cisplatin.
Moreover, the genes with the hypomethylated sites and the genes with the hypermethylated sites formed two complicated networks (Fig.3 and Fig.4). POTE ankyrin domain family member F (POTEF) and HECT domain and ankyrin repeat containing E3 ubiquitin protein ligase 1 (HACE1) were relatively located at the center of the network constructed by the hypomethylated genes and the hypermethylated genes, respectively ( Fig.3 and Fig.4). POTEF belongs to the POTE membrane protein family, which is primate specific and includes 13 paralogs dispersed among eight chromosomes. The POTE proteins were considered to be cancer-testis antigens because they were expressed in many cancers but restricted to only a few normal tissues in the reproductive system [46,47]. Recently, POTEF was found among the top driver oncogenic genes of breast cancer, with a mutation prevalence of over 5% [48]. Additionally, POTEF was identified as a binding partner of Ricinus communis agglutinin I, which may play a critical role in triple-negative breast cancer metastasis [49]. Moreover, POTEF-AS1 was revealed to promote cell growth, repress genes related to the Toll-like receptor signaling and apoptosis pathways, and inhibited apoptosis in docetaxel-treated LNCaP cells, suggesting that POTEF-AS1 would play a key role in the progression of prostate cancer by repressing Toll-like receptor signaling [50]. HACE1 belongs to the HECT family of ubiquitin ligases (HECT E3), which have intrinsic catalytic activity and specificity for substrates involved in the regulation of growth and apoptosis [51]. HACE1 was identified as a tumor suppressor gene involved in the spontaneous tumorigenesis of several cancers in vivo, including lymphoma [52]. The downregulated HACE1 was found to be associated with neuroblastoma progression and poor patient OS [53].
Furthermore, HACE1 is downregulated in Wilm's tumor patients, and the alteration is mediated through hypermethylation of the cytosine phosphate guanine (CpG) island 177 (CpG-177), which is located upstream of the transcription startsite (TSS) [51]. Hypermethylation of CpG-177 in the HACE1 promoter is frequently observed in colorectal and gastric carcinomas, and hypermethylation of HACE1 is associated with the severity of clinic pathological findings, especially lymph node metastasis, in colorectal carcinomas [54][55][56]. Thus, HACE1 was demonstrated to be a tumor suppressor gene in natural killer cell malignancies and tobe down-regulated through a combination of deletion and cytosine phosphate guanine island hypermethylation [57]. Therefor, the prior reports on POTEF and HACE1 indicated that the two abnormally methylated genes might be involved in the chemoresistance of SCLC. The underlying mechanisms are worthy investigating.
The verification of relationship between gene expression and methylation by qRT-PCR revealed that the expression of 69 hypomethylated genes significantly increased and the other 54 hypermethylated genes evidently decreased in H446/DDP cells compared with that of H446 cells, providing evidences that the expression of the most of genes were regulated by DNA methylation. Interestingly, the upregulated genes with the hypomethylated sites were found to be mainly clustered in Pathways in cancer, MAPK signaling pathway, Cytokine-cytokine receptor interaction, and Cell adhesion molecules (CAMs), while the downregulated genes with the hypermethylated sites were mainly clustered in MAPK signaling pathway, Pathways in cancer, Melanoma, Osteoclast differentiation, and Prostate cancer, suggesting that these pathways were most likely to be affected by cisplatin via. methylation.
Additionally, Pathways in cancer was the commonly enriched term between the upregulated genes and the downregulated genes, strongly hinting that cisplatin might cause the remodeling of cancer-

Availability of data and materials
The original data from Illumina Infinium Human Methylation 450K bead chip were available in NCBI (GEO accession number: GSE140743). The data sets supporting the results of this study are included in the manuscript and its additional files.

Ethics approval and consent to participate
Not applicable.

Consent for publication
Not applicable.      Table 7 Distribution of the methylated sites on the abnormally expressed genes