Full Text (PDF)

Published OnlineFirst January 13, 2012; DOI: 10.1158/1541-7786.MCR-11-0300

Molecular Cancer Cancer Genes and Genomics Research

Role of Neuroﬁlament Light Polypeptide in Head and Neck Cancer Chemoresistance

Baishen Chen1,3, Ju Chen2, Michael G. House1, Kevin J. Cullen4, Kenneth P. Nephew1,5, and Zhongmin Guo1,5

Abstract Resistance to cisplatin-based chemotherapy is responsible for therapeutic failure of many common human cancers including cancer of head and neck (HNC). Mechanisms underlying cisplatin resistance remain unclear. In this study, we identified neurofilament light polypeptide (NEFL) as a novel hypermethylated gene associated with resistance to cisplatin-based chemotherapy in HNC. Analysis of 14 HNC cell lines revealed that downregulation of NEFL expression significantly correlated with increased resistance to cisplatin. Hypermethylation of NEFL promoter CpG islands was observed in cell lines as examined by bisulfite DNA sequencing and methylation-specific PCR (MSP) and tightly correlated with reduced NEFL mRNA and protein expression. Furthermore, in patient samples with HNC (n ¼ 51) analyzed by quantitative MSP, NEFL promoter hypermethylation was associated with resistance to cisplatin-based chemotherapy [relative risk (RR), 3.045; 95% confidence interval (CI), 1.459–6.355; P ¼ 0.007] and predicted diminished overall and disease-free survival for patients treated with cisplatin-based chemotherapy. Knockdown of NEFL by siRNA in the highly cisplatin-sensitive cell line PCI13 increased (P < 0.01) resistance to cisplatin. In cisplatin-resistant O11 and SCC25cp cells, restored expression of NEFL significantly increased sensitivity to the drug. Furthermore, NEFL physically associated with tuberous sclerosis complex 1 (TSC1), a known inhibitor of the mTOR pathway, and NEFL downregulation led to functional activation of mTOR pathway and consequentially conferred cisplatin resistance. This is the first study to show a role for NEFL in HNC chemoresistance. Our findings suggest that NEFL methylation is a novel mechanism for HNC chemoresistance and may represent a candidate biomarker predictive of chemotherapeutic response and survival in patients with HNC. Mol Cancer Res; 10(3); 305–15. 2012 AACR.

Introduction more than 70% of relapsed patients, is an important con- Although platinum-based chemotherapeutic agents are tributing cause of the more than 12,000 deaths from HNC widely used to treat human malignancies, including head reported each year in the United States (3). Alternative and neck cancer (HNC; ref. 1), intrinsic and acquired therapeutic strategies and predictive biomarkers for cisplat- resistance of cancer cells to cisplatin accounts for the majority in-resistant HNC are currently lacking. of chemotherapeutic failure in HNC (2). Resistance to Mechanisms underlying the development of resistance to cisplatin, reported in 15% to 30% of newly diagnosed and platinum drugs are not fully understood (4). Extracellular and intracellularchangesthatpromotedrugmetabolism,decreased cellulardrugaccumulation,alteredexpressionofkeymolecules Authors' Affiliations: 1Medical Sciences Program, Indiana University in the apoptotic pathway, and increased repair of DNA School of Medicine, Bloomington, Indiana; 2Department of Surgery, adducts have been reported to contribute to drug resistance Memorial Hospital, Sun Yat-Sen University School of Medicine, Guangz- hou, P. R. China; 3Department of Thoracic Surgery, the Third Affiliated (4). In HNC, alterations of key molecules mediating cisplatin- Hospital of Southern Medical University, Guangzhou, P.R.China; 4Marlene induced apoptosis have been implicated in resistance (5), and and Stewart Greenebaum Cancer Center, University of Maryland School of Medicine, Baltimore, Maryland; and 5Department of Cellular and Integrative mutated or overexpressed p53 and the Bcl-2 protein family, Physiology and Indiana University Simon Cancer Center, Indianapolis, key players in regulating apoptotic pathways, were associated Indiana with HNC resistance to cisplatin-based chemotherapy (6). In Note: Supplementary data for this article are available at Molecular Cancer addition,altered metabolism inreactive oxygen species(ROS), Research Online (http://mcr.aacrjournals.org/). an essential molecule for cisplatin-induced cell killing, was also fi B. Chen and J. Chen contributed equally to this work. observed in HNC (7). We have shown that ampli cation and overexpression of glutathione S-transferase (GST), a molecule Corresponding Author: Zhongmin Guo, Medical Sciences Program, Indi- ana University School of Medicine, Jordan Hall 104, 1001 E. Third St., important for preventing intracellular ROS accumulation, Bloomington, IN 47405. Phone: 812-856-2434; Fax: 812-855-4436; E-mail: contributes to cisplatin resistance in HNC (8, 9). [email protected] Altered DNA methylation is a hallmark of human cancer doi: 10.1158/1541-7786.MCR-11-0300 and has been shown to contribute to drug resistant disease 2012 American Association for Cancer Research. (10). Inactivation of tumor suppressor genes (TSG) by

www.aacrjournals.org 305

Chen et al.

promoter CpG island hypermethylation contributes to dendrites (21). In this study, we further examined NEFL tumor initiation and progression, and methylated DNA is methylation and expression in a panel of HNC cell lines and a potential source of cancer-specific biomarkers for clinical patient tumors with recorded responses to cisplatin-based assessment (11). Studies using DNA methylation inhibitors chemotherapy. We also investigated the functional role and [e.g., 5-aza-20-deoxycytidine (5-aza-dC)] to restore DNA potential mechanisms of NEFL in HNC cisplatin response. damage recognition, repair, and apoptosis pathways and thus Taken together, our results implicate that promoter hyper- sensitivity to cisplatin-based chemotherapy further support methylation-mediated silencing of NEFL is a novel mech- the important role of DNA methylation in cancer (12, 13). anism for HNC chemoresistance, and methylated NEFL Although epigenetic alterations in chemoresistant tumors may represent a candidate biomarker predictive of chemo- have been reported in ovarian, lung, colon, and testicular therapeutic response and survival for patients with HNC. cancers (10, 14, 15), aberrant DNA methylation in HNC chemoresistance have only been reported in in vitro cell line models (16). Materials and Methods To better understand the epigenetic mechanisms under- Cell lines lying cisplatin resistance in HNC and identify novel hyper- A total of 15 HNSCC cell lines, including 2 isogenic methylated genes as therapeutic targets and potential HNC cisplatin-sensitive/resistant lines (HN17B/HN17Bcp, biomarkers, we carried out genome-wide DNA methylation SCC25/SCC25cp) were used in this study (Table 1). In profiling of isogenic, cisplatin-sensitive, and cisplatin-resis- addition, an in vitro immortalized keratinocyte cell line tant HNC cell lines, using an approach that combines gene Hacat was used as a nonmalignant control. Cell lines expression microarray and pharmacologic unmasking (5- O11, O12, O13, O22, O28, and O29 were obtained from aza-dC treatment of cancer cells; ref. 17). This technique has the Department of Otolaryngology, Head and Neck Cancer been widely used in methylation screening for different types Research Division, Johns Hopkins University School of of cancers (16, 17). Among the identified candidates, neu- Medicine (Baltimore, MD) and have been described previ- rofilament light polypeptide (NEFL) was selected for anal- ously (22). HN17Bcp cisplatin–resistant cells were devel- ysis, due to its chromosome location of 8p21, a genetic locus oped in Dr. K. Cullen's laboratory by long-term treatment of previously reported to be frequently deleted in several cisplatin-sensitive HN17B cells with increasing concentra- common cancers, including HNC (18–20). The NEFL gene tions of cisplatin. The additional cell lines, sources, and encodes the light subunit of neurofilaments, which are culture conditions were described previously (8, 23). Cis- essential for maintaining neuronal integrity and may play platin sensitivity of these cell lines was determined by an essential role in intracellular transport to axons and treating cells with various doses of cisplatin continuously

Table 1. Correlation of expression and methylation of NEFL gene with cisplatin sensitivity in HNC cell lines

Cell lines NEFL expression NEFL methylation Cisplatin IC50 value, mmol/L Cisplatin sensitivity Hacat þ UM Not determined Unknown PCI13 þ UM 0.54 S O29 þ UM 1 S HN17B þ UM 1.4 S HN22A þ UM 1.6 S O12 þ UM 1.7 S HN38 PM 2 MR O13 þ PM 2.25 MR SCC25 þ UM 2.5 MR O11 M 2.65 MR O22 þ UM 4.7 MR PCI51 M 6.5 R Fadu M7 R SCC25cp M 9.5 R HN17Bcp PM 9.6 R O28 M 18.45 R

NOTE: Expression of NEFL was detected by real-time quantitative reverse transcriptase PCR with GAPDH as a normalization control. Abundant expression was recorded as þ and no or reduced expression as . NEFL methylation was analyzed by MSP and bisulﬁte DNA

sequencing. Cisplatin IC50 value was determined by 72-hour treatment of individual cell lines with cisplatin. Cisplatin sensitivity of the

cell lines was categorized into sensitive (IC50 < 2 mmol/L), moderately resistant (IC50 ¼ 2–5 mmol/L), and resistant (IC50 > 5 mmol/L). Abbreviations: M, methylated; MR, moderately resistant; PM, partial methylation; R, resistant; S, sensitive; UM, unmethylated.

306 Mol Cancer Res; 10(3) March 2012 Molecular Cancer Research

NEFL Methylation and Head and Neck Cancer Chemoresistance

for 72 hours and conducting MTT assays to measure collection were approved by local Institutional Review cisplatin toxicity. IC50 values were determined with Prism Boards. software (GraphPad Software). Real-time reverse transcriptase PCR: mRNA levels of Patient samples NEFL transcripts in HNC cell lines were quantitatively Frozen tumor and matched normal tissues from patients assessed by real-time reverse transcriptase PCR with HNC were also obtained from the Department of For cDNA synthesis, 2-mg RNA was used for reverse Otolaryngology (Head and Neck Cancer Research Division, transcription (First-Strand cDNA Synthesis kit; Invitrogen). Johns Hopkins University School of Medicine). Fresh tissues Twenty microliters of cDNA reaction mixture were further samples were collected and stored at 80C. Archival tumor diluted to 100 mL, and then 3 of 100 mL cDNA for each tissues from an additional 51 patients with HNC with clinical sample was mixed with SYBR Green Master mixture (Roche information, including treatment histories and survival fol- Diagnostics) and quantitatively amplified by real-time PCR low-up, were collected from Georgetown University Lom- in a LightCycler 480 instrument using a protocol suggested bardi Cancer Center (Washington, DC). The patients from by manufacturer (Roche Diagnostics). Primers for amplifi- this cohort were selected on the basis of availability of cation of NEFL transcript were: forward: 50-GAAGAG- adequate tissues for analysis. After the initial biopsy, all GAGGCAGCTGGAAGA and reverse: 50-AAGGAAAT- patients were treated with cisplatin-based chemotherapy for GGGGGTTCAATC. Quantitative amplification of glycer- organ preservation or unresectable disease. Determination of aldehyde-3-phosphate dehydrogenase (GAPDH) cDNA chemotherapy responses and patient survival in this cohort was carried out using the same protocol and used as nor- has been described previously (24). Clinicopathologic features malization control. Relative expression levels of NEFL of all patients are described in Table 2. Sample and data transcript in cell lines were calculated using the previously C C published formula: 2 ( t_NEFL t_GAPDH) (25)

NEFL 5-Aza-dC treatment and reverse transcriptase PCR Table 2. Correlation of promoter HNC cell lines (PCI13, SCC25cp, PCI51, and O28) methylation with clinical characteristics of were treated with 10 mmol/L 5-aza-dC for 3 days with patients with HNC replacement of the drug and medium every 24 hours. RNA of cell lines was purified with TRIzol reagent following the NEFL methylation analysis instruction from the manufacturer (Invitrogen). cDNA synthesis was carried out as described earlier, and 3 mLof a Clinical variables Methylated Unmethylated P the diluted cDNA for each sample was amplified by PCR No. 25 26 using a protocol described in our previous study (17). fi Age 55.0 19.37 59.0 16.23 Primers for ampli cation of NEFL transcript were the same fi Gender as for real-time PCR ampli cation of GAPDH. cDNA was m M1821 used as normalization control. PCR products (10 L) were F75NSresolved in 1% agarose gel and visualized by ethidium Locations bromide staining. Oral cavity 13 13 Methylation analysis Others 12 13 NS 0 Differentiation Methylation analysis of 5 genomic DNA sequences was fi Well and moderate 17 17 conducted on bisul te-converted DNA from cell lines and Poor 8 9 NS primary tissues by an EZ DNA Methylation-Gold Kit Node status (ZYMO Research) according to standard protocol from the N0 9 13 manufacturer. Methylation status of individual CpG island 0 fl þ þ NEFL >N1 16 13 NS in 5 anking genomic regions ( 269 to 572) of Stages gene in cell lines PCI13, SCC25cp, and PCI51 was deter- fi Early 11 13 mined by plasmid cloning and bisul te DNA sequencing. fi m Late 14 13 NS Bisul te-treated DNA (2 L) from each of above cell lines fi m Chemoresponseb were ampli ed in a 25 L PCR reactions containing 1 m Responsive 6 19 PCR buffer, 1.5 mmol/L MgCl2, 200 mol/L dNTP, 1 unit m Nonresponsive 19 7 0.007 of Platinum Taq polymerase (Invitrogen), and 0.4 mol/L of each primer. Primer sequences were: forward, GTAGT- Abbreviation: NS, not significant. TATTTTTTTAGTTTTGGAT and reverse, AAAATCT- a Fisher exact test, 2-sided. (P > 0.05). CCTCCAACCCTTC. PCR amplification was carried out b Cases with chemotherapy responses included responsive for 45 cycles consisting of 95C for 30 seconds, 56C for 30 (completely and partially responsive) and nonresponsive (no seconds, and 72C for 30 seconds, with 2 minutes at 95C response and progressive diseases). for initial denaturation and 7 minutes at 72C for final elongation. PCR products were verified by ethidium

www.aacrjournals.org Mol Cancer Res; 10(3) March 2012 307

Chen et al.

bromide staining, and 1 mL of the products was cloned into a NEFL expression construct and siRNA TA cloning vector (QIAGEN Inc.). Ten clones from each Full-length NEFL coding sequences was PCR amplified cell lines were selected for sequencing. and cloned into a Topo TA pcDNA3.1 expression vector Methylation-specific PCR (MSP) was used to analyze with an N-terminal HA tag (Invitrogen). DNA sequencing methylation of NEFL gene promoter in cell lines and frozen was used to verify the constructs, and expression efficiency primary tissues. MSP primers were designed to detect the was determined by Western blot analysis using HA tag methylation status of CpG in the promoter regions analyzed monoclonal antibody (Santa Cruz Biotechnology). NEFL by bisulfite DNA sequencing. MSP primers included meth- siRNA and scrambled control were purchased from Sigma- ylation-specific primers (forward, AACGATTTTAAGTT- Aldrich. TATTCGTACGT; reverse, CCTTATTCTACTCCAAC- TCGTA) and unmethylated primers (forward, AATGATT- MTT assay TTAAGTTTATTTGTATGT; reverse, ACCTTATTC- Cell lines with silenced expression of NEFL (O11 and TACTACTCCAACTAT). For amplification of both meth- SCC25cp) were seeded in 96-well plates at a density of 5 ylated and unmethylated alleles, 2 mL of bisulfite converted 103 cells per well. After overnight culture, cells were trans- DNA were amplified in a 25-mL PCR reaction following the fected with NEFL expression construct and control same protocol and cycling parameters as for bisulfite DNA pcDNA3.1 vector with a Lipofectamine 2000 reagent (Invi- sequencing. PCR products were resolved in agarose trogen). At 48 hours after transfection, cells were treated with gel electrophoresis and visualized by ethidium bromide various concentrations of cisplatin (0, 2.5, 5, 10, 20, 40, and staining. 80 mmol/L). MTT assays were conducted after 24 hours of NEFL promoter methylation of archival HNC tissues was cisplatin treatment. MTT assays were conducted in quantitatively analyzed by real-time MSP. Real-time MSP triplicate. was carried out according to previously published protocol The effect of NEFL knockdown was examined in the (26). Briefly, bisulfite-converted DNA (2 mL) was amplified cisplatin-sensitive cell line PCI13. Cells (5 103) were m in a 20 L reaction containing 2.5 mmol/L MgCl2,200 seeded in 96-well plates. After overnight culture, cells were mmol/L each dNTP, 1 unit Platinum Taq Polymerase, 1 transfected with NEFL siRNA and scrambled control (5 standard reaction buffer (Invitrogen), 0.5 mmol/L each prim- pmol per well) with Lipofectamine 2000 reagent (Invitro- er, 0.04 mmol/L ROX reference dye, and 0.25 mmol/L gen). After 48 hours of transfection, cells were treated with fluorescence labeled probe. Primers and probe were designed various concentrations of cisplatin (0, 2.5, 5, 10, 20, 40, and to specifically detect the methylated NEFL promoter 80 mmol/L), and MTT assay was applied to examine sequences, including forward primer, AACGATTT- cytotoxicity of cisplatin after 24 hours of treatment. TAAGTTTATTCGTACGT; reverse primer, CCTTATT- CTACTCCAACTCGTA; and probe, (6-FAM)ACGCG- Colony formation assay CTCGATAAAACTAACGAAACGAT(TAMRA). Amplifi- Effects of NEFL expression on restoration of cisplatin cation was carried out with a Roche 480 real-time thermal sensitivity were also tested by colony formation assay. The cycling analyzer (Roche) for 45 cycles consisting of 95C15 O11 and SCC25cp cell lines (5 105 cells per well) were seconds, 60C30seconds,and72C 2 minutes, with 10 placed in a 6-well plate and transfected with 1 mg of NEFL minutes at 95C for initial denaturation. Amplification of expression construct or mock vector after overnight culture. unmethylated b-actin (ACTB) gene promoter was used as a After a 24-hour transfection period, cells from each well were normalization control using the same PCR protocol as for split into three 10-cm petri dishes and cultured for an NEFL. To ensure the specificity of the real-time MSP analysis, additional 24 hours. After 48 hours of transfection, cells every 96-well plate contained wells with bisulfite-converted were treated with 20 mmol/L cisplatin and cultured for up to DNA from patient tissues and in vitro–methylated lympho- 2 weeks with replacement of fresh medium with selective cyte DNA or the methylation-positive DNA from PCI51 cell agent hygromycin in every 3 to 4 days. Colonies were stained line as positive controls and normal lymphocyte DNA or with crystal violet and counted. methylation-negative DNA from PCI13 cell line as negative controls. Relative methylation level was calculated from Western blotting and immunoprecipitation ACTB 2 (cycle number of NEFL cycle number of ),asdescribed Protein levels of NEFL in HNSCC cell lines were detected previously (27). To determine the optimal cutoff value of by Western blot analysis using monoclonal antibody against relative methylation level for discriminating methylated and N-terminal of NEFL protein (Sigma-Aldrich). Phosphory- unmethylated samples, we conducted real-time quantitative lation of ribosomal protein S6 kinase (p70S6k), which MSP (qMSP) analysis on serial mixtures of in vitro–methyl- indicates the functional status of mTOR pathway (28) was ated DNA and methylation-negative lymphocyte DNA. We detected using polyclonal antibody against phosphorylated found that a relative methylation level of value 1.5 was able to p70S6k protein (Cell Signaling Technology). Protein pre- detect 25% of methylated DNA copies in 75% of methyl- parations and Western blotting were carried out as we have ation-negative DNA background, and therefore, should be described (29). sufficiently sensitive to determine the methylation status of To investigate a physical interaction of NEFL and tuber- primary tumors containing both cancer cells and some normal ous sclerosis complex 1 (TSC1), immunoprecipitation tissues. analysis was conducted using antibodies against NEFL

308 Mol Cancer Res; 10(3) March 2012 Molecular Cancer Research

NEFL Methylation and Head and Neck Cancer Chemoresistance

(Sigma-Aldrich) and TSC1 (Cell Signaling Technology). possible association with progressive phenotypes and per- Preparation of cell lysates and procedures of immunopre- haps chemoresistance of tumors. cipitation followed the previously published protocols. To show the role of methylation-mediated silencing of Briefly, 1 107 PCI13 cells were lysed in RIPA buffer NEFL expression in HNC cisplatin resistance, we proceeded with proteinase inhibitors. Immunoprecipitation was carried to examine both expression and promoter CpG island out at 4C overnight in 1 mL reaction mixture including hypermethylation of NEFL in a panel of 15 HNC cell lines 500-mg protein, 1 RIPA buffer and 5-mg TSC1 antibody. with known cisplatin sensitivities and the immortalized Protein A agarose beads (50 mL; Upstate Biotechnology, keratinocyte cell line, Hacat. NEFL expression was greatly Inc.) were added and incubated for 1 hour at room tem- reduced in 47% of the HNC cell lines (7 of 15; Fig. 1A perature. After washing 3 times in Western lysis buffer, beads and Table 1), and expression of NEFL in Hacat was high. were boiled for 5 minutes in 30 mLof2 sample buffer Expression status of NEFL was correlated with cisplatin containing b-mercaptoethanol, incubated on ice (1 minute), sensitivity. As shown in Fig. 1A and Table 1, NEFL and then centrifuged to remove agarose beads prior to expression was observed in all cisplatin-sensitive HNC cell < m electrophoresis. lines (IC50 2 mol/L; PCI13, HN17B, HN22A, O12, and O29), and NEFL expression was absent or greatly reduced in all 5 cell lines displaying the highest level of cisplatin Statistical analysis > m Correlations of NEFL methylation with clinical charac- resistance (IC50 5 mol/L; HN17Bcp, SCC25cp, Fadu, PCI51, and O28), and in 2 of 5 moderately resistant cell lines teristics including chemotherapy responses of patients with ¼ – m HNC were tested by the Fisher 2-sided exact test. Prognostic (IC50 2 5 mol/L; HN38 and O11). The other 3 value of NEFL methylation for patient survivals was assessed moderately sensitive lines (SCC25, O13, and O22) showed by Kaplain–Meier analysis. Differences of cisplatin cytotoxic high NEFL expression. Interestingly, both highly resistant responses of cell lines (PCI13, O11, and SCC25cp) between isogenic subclones HN17Bcp and SCC25cp displayed NEFL expression or siRNA knockdown and mock transfec- markedly reduced NEFL expression as compared with their sensitive parental cells HN17B and SCC25 (Fig. 1A). tants and effects of mTOR inhibitor on cisplatin cytotoxicity fi of HNC cell lines were examined by unpaired t test. Signi cant reduction of NEFL expression in resistant cell lines was also confirmed at the protein level by Western blotting (Fig. 1B). Results To examine the role of promoter CpG island hypermethy- Methylation-mediated silencing of NEFL expression lation in downregulation of NEFL expression, bisulfite correlates with cisplatin resistance in HNC cell lines DNA sequencing was carried out on a total of 28 CpGs in Two pairs of isogenic cisplatin-sensitive parental HNC a 303-bp genomic region in exon 1 of NEFL (Fig. 1C). In cell lines and their resistant subclones (HN17B/HN17Bcp HN17Bcp, SCC25cp and PCI51 cell lines, which displayed and SCC25/SCC25cp) were used for initial screening of very low or no expression of NEFL, the interrogated CpGs novel methylated candidate genes associated with cisplatin were heavily methylated (Fig. 1A and C). In PCI13 HN17B resistance using our previously established approach (17). and SCC25 cells, which displayed abundant NEFL expres- The HN17Bcp and SCC25cp cell lines were treated with the sion, the CpGs assessed were mostly completely unmethy- demethylating agent 5-aza-dC (10 mol/L) for 3 days to lated (Fig. 1A and C). On the basis of the patterns of reactivate the genes epigenetically silenced in cisplatin-resis- methylated CpGs identified from plasmid cloning and tant clones. Gene expression profiles of cisplatin-sensitive bisulfite DNA sequencing, an MSP assay was designed to HN17B and SCC25 cell lines and gene expression changes detect NEFL promoter methylation in the other HNC cell of cisplatin-resistant HN17Bcp and SCC25cp lines by 5- lines. As shown in Fig. 1D, no methylation was detected in aza-dC treatment were analyzed using microarray chips an immortalized Hacat cell, and 7 of 14 of the HNC cell lines containing 18,400 genes (Affymetrix U133A 2.0). Forty- analyzed by MSP were positive for NEFL methylation. one genes were significantly downregulated (>3-fold) in Furthermore, DNA methylation was found to strongly both HN17Bcp and SCC25cp cisplatin-resistant subclones correlate with the gene expression of NEFL in HNC cell compared with their sensitive parental cells HN17B and lines. NEFL methylation was detected in all 6 cell lines with SCC25. After treatment with the demethylating agent (10 reduced or no expression of NEFL, and of the 8 cell lines mmol/L of 5-aza-dC), derepression of 19 genes of those 41 with NEFL expression, only one (O13) displayed partial downregulated genes was observed (upregulated >3-fold) in NEFL methylation (Fig. 1A and D and Table 1). Moreover, HN17Bcp and SCC25cp cells. This group may represent as shown in Fig. 1F, treatment with a demethylating agent frequently methylated genes of candidates in cisplatin- (5-aza-dC) dramatically restored NEFL expression in resistant HNC cell lines and tumors. Among the 19 SCC25cp, PCI51, and O28 (positive for NEFL methyla- derepressed genes, NEFL was of particular interest, due tion) and had no effect on NEFL expression in PCI13 cells to its genomic location at chromosome 8p21, a locus (negative for NEFL methylation), further suggesting a crit- frequently deleted in several common cancer types includ- ical role for promoter methylation in regulating the expres- ing HNC (18–20, 30). Allelic loss of 8p21 has been sion of NEFL in HNC cells. reported to associate with clinical outcomes of breast To determine whether NEFL methylation was present in cancer and patients with HNC (20, 31), implicating its primary tumors, we used the MSP assay on a small cohort of

www.aacrjournals.org Mol Cancer Res; 10(3) March 2012 309

Chen et al.

AC5′ partial genomic structure NEFL 20 TSS of gene 15

10 +269 +572 5 0 PCI13

O11 O12 O13O22 O28 O29 HacatPCI13 HN38 FADU HN17B HN22ASCC25 PC151 HN17B HN17Bcp SCC25cp B HN17Bcp PCi51 HacatPCI13HN17BHN17BcpHN22ASCC25SCC25cp O11 NEFL SCC25 GAPDH

SCC25cp D

PCI51 MD NL Hacat PCI13 HN17B HN22A SCC25 SCC25cp HN 38 PCI51 Fadu O11 O12 O13 O22 O28 O29 Marker M

U F E PCI13 SCC25cp PCI51 O28 2,685 2,925 2,749 2,911 2,934 325 266 5-Aza-dC ––––++ + + MNLTT T T T TNNNNN T NEFL M

U GAPDH

Figure 1. Downregulation of NEFL expression correlates with its promoter hypermethylation in HNC. A, NEFL mRNA expression was examined by real-time reverse transcriptase PCR in 14 HNC cell lines and an immortalized epidermal keratinocyte Hacat using GAPDH as a normalization control. B, protein expression of NEFL tightly correlates with mRNA expression of the gene. C, methylated CpGs at 50 flanking genomic region (þ269 to þ572 bp) of NEFL gene by plasmid cloning and bisulfite DNA sequencing in NEFL-expressing PCI13, HN17B, and SCC25 cells and NEFL-negative cells HN17Bcp, SCC25cp, and PCI51. Unmethylated CpGs are indicated as unfilled squares and methylated as black squares. D and E, NEFL promoter methylation in Hacat, HNC cell lines, and primary tumors by MSP using primers specific for methylated (M) and unmethylated (U) sequences. In vitro methylated DNA (MD) and normal human lymphocyte DNA (NL) served as positive and negative controls for the MSP assay, respectively. F, treatment with 5-aza-dC restored NEFL expression in SCC25cp, PCI51, and O28 cells with NEFL promoter methylation. T, treated; N, nontreated.

patient tumors with matched normal oral mucosa. NEFL further confirmed by direct bisulfite DNA sequencing methylation was detected in 5 of 7 HNC primary tumors, (Supplementary Fig. S1). On the basis of predetermined whereas none of 5 matched normal oral mucosal tissues were cutoff value, 49% (25 of 51) of cases were considered positive for NEFL methylation (Fig. 1E), providing the positive for NEFL hypermethylation (Table 2). Further- rationale for further analysis of an expanded cohort of clinical more, NEFL methylation was associated with responses samples. to cisplatin-based chemotherapy. For nonresponders (no response to chemotherapy and progressive disease), NEFL Promoter hypermethylation of NEFL is frequently methylation was detected in 73% (19 of 26) of cases, observed in primary HNC tumors and strongly associates whereas NEFL methylation was observed in only 24% with chemoresistance and poor prognosis for patients (6 of 25) of complete or partial responders (Table 2). The with HNC difference in NEFL methylation frequency between On the basis of the earlier results in cell lines and responsive and nonresponsive cases was highly significant tumors, we analyzed NEFL promoter methylation in 51 [relative risk (RR), 3.045; 95% confidence interval (CI), HNC cases with recorded responses to cisplatin-based 1.459–6.355; P ¼ 0.007, Fisher exact test, 2-tailed]. No chemotherapy and long-term clinical follow-up. NEFL significant correlation was found between NEFL meth- promoter CpG island methylation was quantitatively ylation and other clinical parameters, including age, assessed using qMSP. We determined that the relative gender, tumor stage, node status, and histologic differ- methylation level of 1.5 as an optimal cutoff value for entiation (Table 2). detecting NEFL methylation in primary tumor tissues NEFL methylation was also strongly predictive of both and the reliability of our established qMSP analysis was diminished overall survival and disease-free survival for

310 Mol Cancer Res; 10(3) March 2012 Molecular Cancer Research

NEFL Methylation and Head and Neck Cancer Chemoresistance

tivity, as determined by MTT assay (Fig. 3B). Mock- 1.0 1.0 transfected O11 cells displayed an cisplatin IC50 value of P P m m 0.8 = 0.001 0.8 = 0.001 18.3 mol/L compared with 2.3 mol/L cisplatin IC50 Unmethylated value for NEFL-transfected O11 cells (Fig. 3B, middle). 0.6 n 0.6 = 26 Unmethylated Similar results were obtained in SCC25cp cells (60 vs. 32 n = 26 m 0.4 0.4 mol/L cisplatin IC50 value for mock-transfected SCC25cp vs. SCC25cp exogenously expressing NEFL, Cumulative survival Cumulative survival Cumulative 0.2 Methylated 0.2 Methylated n respectively; Fig. 3B, right). Essentially similar results were n = 25 = 25 0.0 0.0 obtained by colony formation assays. NEFL-transfected 0 20 40 60 80 100 120 0 20 40 60 80 100 120 O11andSCC25cpcellsshowedsubstantiallyreduced Months Months (overall survival) (disease-free survival) colony formation as compared with mock-transfected controls (Fig. 3C).

Figure 2. NEFL promoter hypermethylation predicts reduced survival of NEFL inhibits mTOR pathway activity by physically patients with HNC who received cisplatin-based chemotherapy. A total of interacting with an mTOR suppressor TSC1 51 cases of patients with HNC that received cisplatin-based chemotherapy were included in the study. All tumor tissues were To further investigate the mechanism underlying NEFL- collected prior to clinical treatment. NEFL methylation was detected by mediated cisplatin resistance, we examined potential path- quantitative real-time MSP and predictive value of NEFL methylation in ways modulated by NEFL. NEFL is a critical component of patient survival was assessed by Kaplan–Meier survival curve. Difference the cytoskeleton, interacting with multiple protein targets of patient survival between methylation-positive (Methylated) and -negative (Unmethylated) groups was statistically analyzed by log-rank essential for cell plasticity, proliferation, migration, apopto- test. sis, and molecule transporters (32, 33). Among these, we further investigated the TSC1 tumor suppressor. TSC1 has been shown to function as a molecular inhibitor of the patients with HNC who received cisplatin-based chemo- mTOR oncogenic pathway, reported to be frequently con- therapy. Patients with NEFL methylation had median stitutively activated in cisplatin-resistant cancers (28). NEFL survival of 29 months (95% CI, 15.6–32.4) for overall has been shown to bind TSC1 and stabilize the TSC1/2 survival and 28 months (95% CI, 16.4–31.5) for disease- complex (33). Therefore, we hypothesized that inactivation free survival, whereas patients negative for NEFL methyl- of NEFL, leading to abnormal activation of mTOR path- ation had median survival of 67 months (95% CI, 47.6– way, confers cisplatin resistance. 144.3) for overall and 58 months (95% CI, 37.4–82.5) for To test this hypothesis, we examined the status of the disease-free survival, respectively (Fig. 2). Difference in mTOR pathway and the effects of altered NEFL expression both overall and disease-free survival between NEFL-methy- on mTOR activity in HNC cell lines with and without lated and -unmethylated patients was highly significant NEFL silencing. Immunoprecipitation assays revealed an (P < 0.001, log-rank test). association between NEFL and TSC1 in PCI13 cells with endogenous NEFL expression (Fig. 4A). Consistent with NEFL functionally associates with HNC cellular our hypothesis, mTOR pathway activity was correlated with response to cisplatin NEFL expression. Phosphorylation of p70S6k, a critical Having shown clinical relevance for NEFL inactivation by downstream mTOR activation substrate and known indi- promoter hypermethylation, it was next of interest to further cator of an active mTOR pathway (28), was detected in O11 investigate the functional relevance of aberrant NEFL and SCC25cp cells with methylation silencing of NEFL expression and cisplatin response in HNC. To do so, we and absent in NEFL-expressing PCI13 cells (Fig. 4B). altered the expression of NEFL in HNC cell lines and then Restoration of NEFL expression using pcDNA3.1/NEFL examined the effect on cisplatin sensitivity in vitro. The vector suppressed p70S6k phosphorylation in both O11 cisplatin-sensitive cell line PCI13 was used to examine the and SCC25cp cells (Fig. 4C, right), whereas knockdown effect of NEFL knockdown on cisplatin response. Knock- of NEFL expression resulted in increased phosphorylation of down efficiency of siRNA on NEFL expression was verified p70S6k in PCI13 cells (Fig. 4C, left). by immunoblotting using NEFL-specific antibody (Fig. 3A, NEFL modulation of mTOR pathway activation indicat- left). PCI13 cells were transfected with 10 mmol/L of NEFL ed a role for mTOR activity in cellular cisplatin responsive- siRNA or scrambled control siRNA. After 48 hours of ness. To test this possibility, we examined the effects of an siRNA incubation, PCI13 cells were treated with various mTOR inhibitor, RAD001, on cisplatin sensitivity in HNC doses of cisplatin for 24 hours and cell viability was measured cell lines with or without NEFL methylation. Cell lines were by MTT assay. As shown in Fig. 3A, downregulation of first treated with 10 nmol/L of RAD001 for 24 hours and NEFL protein level by siRNA significantly enhanced resis- then incubated with 5 and 10 mmol/L of cisplatin or mock tance of PCI13 cells to cisplatin treatment. The IC50 value controls. MTT assays were conducted 24 hours after cis- for PCI13 transfected with NEFL siRNA was 8.5 mmol/L platin treatment. Treatment with RAD001 significantly m compared with a 5.2 mol/L IC50 value in controls. In inhibited SCC25cp, O11, and O28 cell growth (Fig. contrast, restoring full expression of NEFL in O11 and 4D). No inhibitory effect of RAD001 on the NEFL-expres- SCC25cp cell lines dramatically increased cisplatin sensi- sing cell line PCI13 was observed (Fig. 4D), although

www.aacrjournals.org Mol Cancer Res; 10(3) March 2012 311

Chen et al.

A PCI13 120 PCI13 100 Control-sc 80 Control-sc NEFL-si NEFL-si NEFL 60 40 ** 20 ** GAPDH *

Cell viability (%) 0 –20 0 2.55 10204080 Cisplatin (μmol/L)

B O11 SCC25cp 120 Mock 120 ** HEK293 ** * ** Mock 100 ** ** NEFL 100 80 80 * NEFL Mock NEFL 60 60 * NEFL 40 40 20 20

GAPDH Cell viability (%) 0 Cell viability (%) 0 –20 0 2.5 5 10 20 40 80 –20 02.55 10204080 Cisplatin (μmol/L) Cisplatin (μmol/L) C O11 SCC25cp

80 80 Mock NEFL Mock NEFL 60 60 40 40

Colony no. Colony 20 ** no. Colony 20 ** 0 0 Mock NEFL Mock NEFL

Figure 3. NEFL functionally associates with cisplatin responses in HNC cells. A, knockdown of NEFL conferred cisplatin resistance in HNC cells. PCI13 cells were transfected with 10 mmol/L NEFL siRNA (NEFL-si) or scrambled control (Control-sc) for 48 hours and treatment of the transfected cells with various concentrations of cisplatin was continued for additional 24 hours. MTT assay was conducted to measure cell growth 72 hours later. NEFL knockdown by transfected siRNA in PCI13 cells was verified by Western blotting using NEFL-specific antibody. NEFL knockdown significantly increases cisplatin resistance in PCI13 cells (unpaired t test, , P < 0.05; , P < 0.01). B, enforced overexpression of NEFL resensitized NEFL-negative cells to cisplatin treatment. O11 and SCC25cp cells were transfected with pcDNA3.1/NEFL expression construct or mock (empty vector) for 48 hours and further treated with various doses of cisplatin for 24 hours. Cell growth was measured by MTT 72 hours after drug treatment. Efficacy of pcDNA3.1/NEFL expression construct was verified by transient expression of the construct in HEK293 cells and then examined by Western blotting. Restored expression of NEFL significantly enhances cisplatin sensitivity in both O11 and SCC25cp cells (, P < 0.05; , P < 0.01). C, enforced overexpression of NEFL reduced colony formation in O11 and SCC25cp cells. O11 and SCC25cp cells were transfected with pcDNA3.1/NEFL expression construct and mock vector for 48 hours and incubated with 10 mmol/L of cisplatin for an additional 24 hours. Cells were further cultured in G418 selective medium for 2 weeks to allow the formation of visible colonies. Difference in numbers of colonies between NEFL and mock-transfected cells in both O11 and SCC25cp cells was significant (, P < 0.01, unpaired t test).

increased cell growth was seen at the lower dose. For PCI51 8p21, which has been identified as one of the genetic loci cells (hypermethylated and silenced NEFL), there was a frequently affected by both heterozygous and homozygous trend for cell growth inhibition by RAD001, but this was not deletions in a variety of common human cancers, including statistically significant. cancers of prostate and breast (18, 30). In HNC, loss of heterozygosity at the NEFL locus was reported in more than 40% of cases (20); moreover, deletions of the 8p21 NEFL Discussion locus have been shown to predict poor prognosis in patients NEFL gene encodes type IV intermediate filament hetero- with HNC (20). Second, neurofilament heavy peptide polymers that functionally maintain the neuronal caliber and (NEFH), a functional partner of NEFL and a subunit of play an important role in intracellular transport of neuro- neurofilament heteropolymers, has been recently shown to transmitters to axons and dendrites (34). Germline muta- be a putative TSG frequently inactivated by promoter tions of NEFL gene are known to cause Charcot-Marie- hypermethylation in esophageal cancers (36). Third, NEFL Tooth disease types 1F, disorders of the peripheral nervous has been shown to interact with a number of functional system that are characterized by distinct neuropathies (35). molecular targets found in several critical cancer-associated Although in cancer, a role for NEFL has not been previously pathways. These NEFL-interacting targets include TSC1 shown; several lines of evidence implicate NEFL as a (33), a potent inhibitor of mTOR pathway and glutamate potential TSG. First, NEFL is located at chromosome receptor, ionotropic, N-methyl-D-aspartate 1 (NMDAR1,

312 Mol Cancer Res; 10(3) March 2012 Molecular Cancer Research

NEFL Methylation and Head and Neck Cancer Chemoresistance

A PCI13 BC

PCI13 SCC25cp O11 Input ProteinTSCI A Ab

68 kD NEFL PCI13 SCC25cp O11 Mock NEFL Mock NEFL Control-sc NEFL-si pp7056k pp7056k

GAPDH GAPDH

D 350 300

250

200

150 MTT value

100

0 RAD001 (10 nmol/L) ––++–+ –+ ––++ –+–+ ––+ +

Cisplatin 5 10 5 10 5 10 5 10 5 10 μmol/L μmol/L μmol/L μmol/L μmol/L μmol/L μmol/L μmol/L μmol/L μmol/L

PCI13 SCC25cp PCI51 O11 O28

Figure 4. NEFL physically interacts with mTOR suppressor TSC1, and functional inactivation of NEFL results in mTOR pathway activation in HNC cells. A, direct binding of NEFL and TSC1 in PCI13 cells was determined by immunoprecipitation (IP) using TSC1 antibody for IP and NEFL antibody for Western blotting (see Materials and Methods). B, activation of mTOR pathway was assessed by Western blot analysis detection of phosphorylation of p70S6k, a substrate of mTOR pathway, using a speciﬁc antibody against phosphorylated p70S6k (pp70S6k). Phosphorylated p70S6k was detected in O11 and SCC25cp cell lines with NEFL inactivation whereas absent in PCI13 cell with normal NEFL expression. C, NEFL expression, restored by transfecting an NEFL expression vector, dramatically reduced the level of phosphorylated p70S6k in O11 and SCC25cp cells as compared with mock vector (left). In contrast, downregulation of NEFL by siRNA (NEFL-si) in PCI13 cells induced p70S6k phosphorylation as compared with scrambled siRNA transfection (Control-sc). D, inhibition of mTOR activity resensitized cell lines with NEFL methylation to cisplatin. NEFL-negative cell lines (O11, SCC25cp, PCI51, and O28) and -positive cell line PCI13 were treated with mTOR inhibitor RAD001 (10 nmol/L) for 24 hours and further incubated with 5 and 10 mm of cisplatin for additional 24 hours. Cell growth was measured by MTT assay. indicates the difference in cell growth between RAD001 treatment and control groups (P < 0.05; unpaired t test).

Grin1; ref. 37), a newly identified TSG in human esophageal Drug resistance is a major limitation for the clinical carcinoma (26). Taken together, these studies strongly application of cisplatin and its analogues in HNC chemo- suggest that NEFL, and an intact neurofilament structure, therapy and contributes significantly to morbidity and may play a critical role in suppressing cancer initiation mortality from HNC (38). Despite intensive efforts for and/or progression. several decades, the mechanism(s) underlying resistance to In the current study, we show that downregulation of cisplatin in patients with HNC remains unclear (39). Pre- NEFL expression is seen in about 50% of HNC cell lines and vious studies have shown that p53 mutations, leading to correlates with NEFL hypermethylation (Fig. 1). In addi- inactivation of p53 protein and defects of apoptotic tion, NEFL CpG island promoter methylation was detected responses to chemotherapeutic agents, are strongly associ- in 49% of HNC primary tumors (Table 2) but was absent in ated with cisplatin resistance in HNC (40). In addition, we matched normal oral mucosa tissues (Fig. 1D). Our study have reported a strong association between chromosome provides the first direct evidence that NEFL is epigenetically 11q13 amplification, resulting in overexpression of GST pi, silenced in a human cancer and supports the possibility that which detoxifies cisplatin-induced ROS, and decreased NEFL may be the putative TSG candidate located at chemotherapy responses in patients with HNC (8). In this chromosome 8p21, which may have important implications study, we show that NEFL is silenced by DNA methylation for a broad range of common cancers with frequent 8p21 in cisplatin-resistant but not in cisplatin-sensitive HNC cell deletions, including prostate, breast, bladder, lung, and liver lines and our functional studies further show a requirement cancers (18, 30). for NEFL in cisplatin-induced growth inhibitory responses.

www.aacrjournals.org Mol Cancer Res; 10(3) March 2012 313

Chen et al.

Clinically, NEFL methylation was strongly associated in including HNC (44) and has been shown to associate with patient responses to cisplatin-based chemotherapy. This is resistance to several chemotherapeutic agents, including the first report to identify downregulation of NEFL expres- cisplatin (28). In cultured cortical neurons, NEFL was found sion by promoter methylation as a novel and important to physically bind to the putative tumor suppressor TSC1 mechanism for HNC chemoresistance. (33), a molecular inhibitor of mTOR activity. We conducted Predictive and prognostic markers for chemotherapy a series of functional studies to show that NEFL is required responses and patient outcomes are highly valuable for for suppression of mTOR activity in HNC cells. Further- guiding therapeutic strategies and improving survival of more, downregulation of NEFL leads to mTOR activation, patients with cancer (41). Lack of such biomarkers for conferring cisplatin resistance. We further show that NEFL clinical use has had a substantial influence on the successful physically associates with TSC1 in NEFL-expressing PCI13 management of patients with HNC, as HNC morbidity has cells and modulates mTOR activity. Altered mTOR path- remained essentially unchanged for several decades (42).We way activation also associates with NEFL-negative HNC. and others, using immunohistochemical analysis have iden- These findings provide convincing evidence that activation tified protein biomarkers that significantly correlated with of mTOR pathway, a consequence of NEFL downregulation chemotherapy responses and patient prognosis with HNC by promoter hypermethylation, is one of the mechanisms (43). In this study, patients with HNC with positive NEFL underlying cisplatin resistance in HNC. However, we rec- promoter methylation were nearly 3 times more likely to be ognize that inhibition of mTOR activity by RAD001 alone resistant to cisplatin-based chemotherapy than those with- was not able to restore cisplatin sensitivity in all cell lines with out NEFL methylation, regardless of other clinical para- NEFL methylation (e.g., PCI51), suggesting that other meters of the patients (Table 2). NEFL methylation also mechanisms may be involved in NEFL-associated responses predicted reduced overall and disease-free survival for to cisplatin. NEFL has been found to interact with a number patients with HNC who received cisplatin-based chemo- of target proteins that participate in various important therapy (Fig. 2). These findings suggest that NEFL meth- cellular processes. Among these, glutamate receptor, iono- ylation is a candidate independent predictive marker for tropic, NMDAR1, a subunit of NMDAR, has recently been response to cisplatin-based chemotherapy and that methyl- shown to be a novel TSG in esophageal carcinoma (26). A ated NEFL may predict a more malignant clinical phenotype crucial role for NMDAR in response to stress signals, in patients with HNC. We recognize that a limitation of the including cisplatin in neurologic system (47), has also been current study is the small number of patients and that our recently reported. As the interaction of NEFL and NMDAR analysis is not able to account for possible effects of other may represent an additional novel mechanism for NEFL- reported prognostic factors, such as human papilloma virus associated cisplatin responses in cancer cells, we are currently (HPV) status and the combined radiation therapy. However, investigating this potential mechanism. the strong correlation of NEFL methylation with chemo- Disclosure of Potential Conflicts of Interest therapy responses and patient outcomes, as well as our fl findings in vitro functional experiments, support further No potential con icts of interest were disclosed. analysis of a larger patient cohort. Grant Support mTOR is a key downstream protein kinase of the phos- This work was supported by awards #062545 from the Flight Attendant Medical phatidylinositol 3-kinase (PI3K)/AKT signaling pathway Research Institute (to Z. Guo) and by the National Cancer Institute Integrative Cancer Biology Program U54CA113001 (to K.P. Nephew). and has reported to play a central role in controlling cancer The costs of publication of this article were defrayed in part by the payment of page cell growth (44). AKT and mTOR function as "switch" charges. This article must therefore be hereby marked advertisement in accordance with proteins to modulate metabolism, the cell cycle, and apo- 18 U.S.C. Section 1734 solely to indicate this fact. ptosis in cancer cells (45, 46). Constitutive activation of Received June 23, 2011; revised December 13, 2011; accepted December 29, 2011; mTOR has been found in common human malignancies, published OnlineFirst January 13, 2012.

References 1. Specenier PM, Vermorken JB. Current concepts for the manage- 6. Andrews GA, Xi S, Pomerantz RG, Lin CJ, Gooding WE, Wentzel AL, ment of head and neck cancer: chemotherapy. Oral Oncol 2009; et al. Mutation of p53 in head and neck squamous cell carcinoma 45:409–15. correlates with Bcl-2 expression and increased susceptibility to cis- 2. Seve P, Dumontet C. Chemoresistance in non-small cell lung cancer. platin-induced apoptosis. Head Neck 2004;26:870–7. Curr Med Chem Anticancer Agents 2005;5:73–88. 7. Parise O Jr, Janot F, Luboinski B, Massaad L, Albin N, Toussaint C, 3. Colevas AD. Chemotherapy options for patients with metastatic or et al. Thymidylate synthase activity, folates, and glutathione system in recurrent squamous cell carcinoma of the head and neck. J Clin Oncol head and neck carcinoma and adjacent tissues. Head Neck 2006;24:2644–52. 1994;16:158–64. 4. Borst P, Rottenberg S, Jonkers J. How do real tumors become 8. Cullen KJ, Newkirk KA, Schumaker LM, Aldosari N, Rone JD, Haddad resistant to cisplatin? Cell Cycle 2008;7:1353–9. BR. Glutathione S-transferase pi ampliﬁcation is associated with 5. Higuchi E, Oridate N, Furuta Y, Suzuki S, Hatakeyama H, Sawa H, cisplatin resistance in head and neck squamous cell carcinoma cell et al. Differentially expressed genes associated with CIS-diammi- lines and primary tumors. Cancer Res 2003;63:8097–102. nedichloroplatinum (II) resistance in head and neck cancer using 9. Rudin CM, Yang Z, Schumaker LM, VanderWeele DJ, Newkirk K, differential display and CDNA microarray. Head Neck 2003;25: Egorin MJ, et al. Inhibition of glutathione synthesis reverses Bcl-2- 187–93. mediated cisplatin resistance. Cancer Res 2003;63:312–8.

314 Mol Cancer Res; 10(3) March 2012 Molecular Cancer Research

NEFL Methylation and Head and Neck Cancer Chemoresistance

10. Shen L, Kondo Y, Ahmed S, Boumber Y, Konishi K, Guo Y, et al. Drug 28. Mabuchi S, Kawase C, Altomare DA, Morishige K, Sawada K, Hayashi sensitivity prediction by CpG island methylation profile in the NCI-60 M, et al. mTOR is a promising therapeutic target both in cisplatin- cancer cell line panel. Cancer Res 2007;67:11335–43. sensitive and cisplatin-resistant clear cell carcinoma of the ovary. Clin 11. Baylin SB, Chen WY. Aberrant gene silencing in tumor progression: Cancer Res 2009;15:5404–13. implications for control of cancer. Cold Spring Harb Symp Quant Biol 29. Fan M, Bigsby RM, Nephew KP. The NEDD8 pathway is required for 2005;70:427–33. proteasome-mediated degradation of human estrogen receptor (ER)- 12. Fang F, Balch C, Schilder J, Breen T, Zhang S, Shen C, et al. A phase 1 alpha and essential for the antiproliferative activity of ICI 182,780 in and pharmacodynamic study of decitabine in combination with car- ERalpha-positive breast cancer cells. Mol Endocrinol 2003;17:356–65. boplatin in patients with recurrent, platinum-resistant, epithelial ovar- 30. Macoska JA, Trybus TM, Benson PD, Sakr WA, Grignon DJ, Wojno KD, ian cancer. Cancer 2010;116:4043–53. et al. Evidence for three tumor suppressor gene loci on chromosome 13. Fang X, Zheng C, Liu Z, Ekman P, Xu D. Enhanced sensitivity of 8p in human prostate cancer. Cancer Res 1995;55:5390–5. prostate cancer DU145 cells to cisplatinum by 5-aza-20-deoxycytidine. 31. Seitz S, Werner S, Fischer J, Nothnagel A, Schlag PM, Scherneck S. Oncol Rep 2004;12:523–6. Refined deletion mapping in sporadic breast cancer at chromosomal 14. Wei SH, Balch C, Paik HH, Kim YS, Baldwin RL, Liyanarachchi S, et al. region 8p12-p21 and association with clinicopathological parameters. Prognostic DNA methylation biomarkers in ovarian cancer. Clin Cancer Eur J Cancer 2000;36:1507–13. Res 2006;12:2788–94. 32. Goehler H, Lalowski M, Stelzl U, Waelter S, Stroedicke M, Worm U, 15. Olasz J, Mandoky L, Geczi L, Bodrogi I, Csuka O, Bak M. Influence of et al. A protein interaction network links GIT1, an enhancer of hMLH1 methylation, mismatch repair deficiency and microsatellite huntingtin aggregation, to Huntington's disease. Mol Cell 2004; instability on chemoresistance of testicular germ-cell tumors. Anti- 15:853–65. cancer Res 2005;25:4319–24. 33. Haddad LA, Smith N, Bowser M, Niida Y, Murthy V, Gonzalez-Agosti C, 16. Chang X, Monitto CL, Demokan S, Kim MS, Chang SS, Zhong X, et al. et al. The TSC1 tumor suppressor hamartin interacts with neurofila- Identification of hypermethylated genes associated with cisplatin ment-L and possibly functions as a novel integrator of the neuronal resistance in human cancers. Cancer Res 2010;70:2870–9. cytoskeleton. J Biol Chem 2002;277:44180–6. 17. Wu G, Guo Z, Chang X, Kim MS, Nagpal JK, Liu J, et al. LOXL1 and 34. Leermakers FA, Zhulina EB. How the projection domains of NF-L and LOXL4 are epigenetically silenced and can inhibit ras/extracellular alpha-internexin determine the conformations of NF-M and NF-H in signal-regulated kinase signaling pathway in human bladder cancer. neurofilaments. Eur Biophys J 2010;39:1323–34. Cancer Res 2007;67:4123–9. 35. Houlden H, Reilly MM. Molecular genetics of autosomal-dominant 18. Knowles MA, Shaw ME, Proctor AJ. Deletion mapping of chromosome demyelinating Charcot-Marie-Tooth disease. Neuromolecular Med 8 in cancers of the urinary bladder using restriction fragment length 2006;8:43–62. polymorphisms and microsatellite polymorphisms. Oncogene 36. Kim MS, Chang X, LeBron C, Nagpal JK, Lee J, Huang Y, et al. 1993;8:1357–64. Neurofilament heavy polypeptide regulates the Akt-beta-catenin path- 19. Lerebours F, Olschwang S, Thuille B, Schmitz A, Fouchet P, Buecher way in human esophageal squamous cell carcinoma. PLoS One B, et al. Fine deletion mapping of chromosome 8p in non-small-cell 2010;5:e9003. lung carcinoma. Int J Cancer 1999;81:854–8. 37. Ehlers MD, Fung ET, O'Brien RJ, Huganir RL. Splice variant-specific 20. Coon SW, Savera AT, Zarbo RJ, Benninger MS, Chase GA, Rybicki BA, interaction of the NMDA receptor subunit NR1 with neuronal interme- et al. Prognostic implications of loss of heterozygosity at 8p21 and diate filaments. J Neurosci 1998;18:720–30. 9p21 in head and neck squamous cell carcinoma. Int J Cancer 38. Kelland L. The resurgence of platinum-based cancer chemotherapy. 2004;111:206–12. Nat Rev Cancer 2007;7:573–84. 21. Previtali SC, Zerega B, Sherman DL, Brophy PJ, Dina G, King RH, et al. 39. Le Tourneau C, Chen EX. Molecularly targeted agents in the treatment Myotubularin-related 2 protein phosphatase and neurofilament light of recurrent or metastatic squamous cell carcinomas of the head and chain protein, both mutated in CMT neuropathies, interact in peripheral neck. Hematol Oncol Clin North Am 2008;22:1209–20, ix. nerve. Hum Mol Genet 2003;12:1713–23. 40. Schumaker L, Nikitakis N, Goloubeva O, Tan M, Taylor R, Cullen KJ. 22. Tokumaru Y, Yamashita K, Osada M, Nomoto S, Sun DI, Xiao Y, et al. Elevated expression of glutathione S-transferase pi and p53 confers Inverse correlation between cyclin A1 hypermethylation and p53 poor prognosis in head and neck cancer patients treated with che- mutation in head and neck cancer identified by reversal of epigenetic moradiotherapy but not radiotherapy alone. Clin Cancer Res silencing. Cancer Res 2004;64:5982–7. 2008;14:5877–83. 23. Newkirk K, Heffern J, Sloman-Moll E, Sessions RB, Rasmussen AA, 41. Zalcman G, Bergot E, Creveuil C, Levallet G, Lechapt E. Integrating Andrews PA, et al. Glutathione content but not gamma glutamyl biomarkers into clinical trials: methodological issues for a new para- cysteine synthetase mRNA expression predicts cisplatin resistance digm in nonsmall cell lung cancer. Curr Opin Oncol 2011;23:106–11. in head and neck cancer cell lines. Cancer Chemother Pharmacol 42. Brockstein BE, Vokes EE. Head and neck cancer in 2010: maximizing 1997;40:75–80. survival and minimizing toxicity. Nat Rev Clin Oncol 2011;8:72–4. 24. Shiga H, Heath EI, Rasmussen AA, Trock B, Johnston PG, Forastiere 43. Cullen KJ, Schumaker L, Nikitakis N, Goloubeva O, Tan M, Sarlis NJ, AA, et al. Prognostic value of p53, glutathione S-transferase pi, and et al. beta-Tubulin-II expression strongly predicts outcome in patients thymidylate synthase for neoadjuvant cisplatin-based chemotherapy receiving induction chemotherapy for locally advanced squamous in head and neck cancer. Clin Cancer Res 1999;5:4097–104. carcinoma of the head and neck: a companion analysis of the TAX 25. Monney L, Sabatos CA, Gaglia JL, Ryu A, Waldner H, Chernova T, 324 trial. J Clin Oncol 2009;27:6222–8. et al. Th1-specific cell surface protein Tim-3 regulates macrophage 44. Zoncu R, Efeyan A, Sabatini DM. mTOR: from growth signal integration activation and severity of an autoimmune disease. Nature 2002;415: to cancer, diabetes and ageing. Nat Rev Mol Cell Biol 2011;12:21–35. 536–41. 45. Foster DA. Phosphatidic acid signaling to mTOR: signals for the 26. Kim MS, Yamashita K, Baek JH, Park HL, Carvalho AL, Osada M, et al. survival of human cancer cells. Biochim Biophys Acta 2009;1791: N-methyl-D-aspartate receptor type 2B is epigenetically inactivated 949–55. and exhibits tumor-suppressive activity in human esophageal cancer. 46. Castedo M, Ferri KF, Kroemer G. Mammalian target of rapamycin Cancer Res 2006;66:3409–18. (mTOR): pro- and anti-apoptotic. Cell Death Differ 2002;9:99–100. 27. Yang H, Kadia T, Xiao L, Bueso-Ramos CE, Hoshino K, Thomas DA, 47. Gozdz A, Vashishta A, Kalita K, Szatmari E, Zheng JJ, Tamiya S, et al. et al. Residual DNA methylation at remission is prognostic in adult Cisplatin-mediated activation of extracellular signal-regulated kinases Philadelphia chromosome-negative acute lymphocytic leukemia. 1/2 (ERK1/2) by inhibition of ERK1/2 phosphatases. J Neurochem Blood 2009;113:1892–8. 2008;106:2056–67.

www.aacrjournals.org Mol Cancer Res; 10(3) March 2012 315

Role of Neurofilament Light Polypeptide in Head and Neck Cancer Chemoresistance

Baishen Chen, Ju Chen, Michael G. House, et al.

Mol Cancer Res 2012;10:305-315. Published OnlineFirst January 13, 2012.

Updated version Access the most recent version of this article at: doi:10.1158/1541-7786.MCR-11-0300

Supplementary Access the most recent supplemental material at: Material http://mcr.aacrjournals.org/content/suppl/2012/03/13/1541-7786.MCR-11-0300.DC1

Cited articles This article cites 47 articles, 19 of which you can access for free at: http://mcr.aacrjournals.org/content/10/3/305.full#ref-list-1

Citing articles This article has been cited by 1 HighWire-hosted articles. Access the articles at: http://mcr.aacrjournals.org/content/10/3/305.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mcr.aacrjournals.org/content/10/3/305. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.