Enhanced Histone Deacetylase Activity in Malignant Melanoma Provokes RAD51 And

Home , FANCD2, RAD51

Author Manuscript Published OnlineFirst on March 15, 2016; DOI: 10.1158/0008-5472.CAN-15-2680 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. 1

Enhanced histone deacetylase activity in malignant melanoma provokes RAD51 and

FANCD2 triggered drug resistance

Andrea Krumm1, Christina Barckhausen1, Pelin Kücük1, Karl-Heinz Tomaszowski1, Carmen

Loquai2, Jörg Fahrer1, Oliver Holger Krämer1, Bernd Kaina1 and Wynand Paul Roos1*

1Institute of Toxicology, Medical Center of the University Mainz, Obere Zahlbacher Str. 67,

D-55131 Mainz, Germany, 2Department of Dermatology, Medical University Center,

Langenbeckstr. 1, D-55131 Mainz, Germany

Running title: HDACs and homologous recombination in melanoma

Keywords: Melanoma, temozolomide, homologous recombination, RAD51, FANCD2,

HDAC

Financial support: Work was supported by the German Research Foundation RO3617

*Corresponding author: Wynand P. Roos, Institute of Toxicology, Medical Center of the

University Mainz, Obere Zahlbacher Str. 67, D-55131 Mainz, Germany, E-mail:

[email protected] or [email protected], Tel: 0049-6131-17-9257, Fax:

0049-6131-17-8499.

Conflicts of interest: Carmen Loquai: Speakers honorar from Roche/BMS/MSD/Leo,

advisory honorar from Roche/BMS/MSD/Amgen/Novartis.

Notes about manuscript: Word count - 5436, number of figures – 7.

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 2016 American Association for Cancer Research. Author Manuscript Published OnlineFirst on March 15, 2016; DOI: 10.1158/0008-5472.CAN-15-2680 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. 2

Abstract

DNA damaging anticancer drugs remain part of metastatic melanoma therapy. Epigenetic

reprogramming caused by increased histone deacetylase (HDAC) activity arising during

tumor formation may contribute to resistance of melanomas to the alkylating drugs

temozolomide, dacarbazine and fotemustine. Here we report on the impact of class I HDACs

on the response of malignant melanoma cells treated with alkylating agents. The data show

that malignant melanomas in situ contain a high level of HDAC1/2 and malignant melanoma

cells overexpress HDAC1/2/3 compared to non-cancer cells. Furthermore, pharmacological

inhibition of class I HDACs sensitizes malignant melanoma cells to apoptosis following

exposure to alkylating agents, while not affecting primary melanocytes. Inhibition of

HDAC1/2/3 caused sensitization of melanoma cells to temozolomide in vitro and in

melanoma xenografts in vivo. HDAC1/2/3 inhibition resulted in suppression of DNA double-

strand break (DSB) repair by homologous recombination due to down-regulation of RAD51

and FANCD2. This sensitized cells to the cytotoxic DNA lesion O6-methylguanine and

caused a synthetic lethal interaction with the poly(ADP-ribose)polymerase-1 inhibitor

olaparib. Furthermore, knockdown experiments identified HDAC2 as being responsible for

the regulation of RAD51. The influence of class I HDACs on DSB repair by homologous

recombination and the possible clinical implication on malignant melanoma therapy with

temozolomide and other alkylating drugs suggests a combination approach where class I

HDAC inhibitors such as valproic acid or MS-275 (entinostat) appear to counteract HDAC-

and RAD51/FANCD2-mediated melanoma cell resistance.

Introduction

Patients suffering from melanoma in its metastatic stage have a dismal prognosis. Although

newer therapies targeting mutant B-Raf or immunotherapies with ipilimumab or anti-PD-1-

antibody have shown promise (1-3), the cancer invariably becomes resistant, which

necessitates the switch back to genotoxic-based therapies. These therapies include

methylating agents such as dacarbazine (DTIC) and temozolomide (TMZ). Methylating

agents either require metabolic activation by cytochrome P450, in the case of DTIC (4), or

spontaneously decompose, in the case of TMZ (5), to the DNA reactive species 5-(3-

methyltriazen-1-yl)imidazole-4-carboxamide (MTIC). The clinically relevant cytotoxic DNA

lesion for these drugs is O6-methylguanine (O6MeG). O6MeG is repaired by O6-

methylguanine-DNA methyltransferase (MGMT) in a stoichiometric reaction, which restores

guanine and inactivates MGMT (6). The resistance of melanoma cells to methylating agents

is therefore proportional to the cells’ MGMT activity level (7). MGMT expression is governed

by the methylation of CpG islands in its promoter (8) and the methylation status of MGMT

has been shown to be predictive of TMZ-based melanoma therapy (9).

As melanomas express low levels of MGMT (10, 11), O6MeG often persists in the DNA.

Unrepaired O6MeG mispairs with thymine during replication and O6MeG/T mismatches are

detected and bound by the mismatch repair (MMR) machinery. Futile MMR repair cause

persistent stretches of single-stranded DNA (ssDNA) (12). If these stretches of ssDNA are

not bypassed by homologous recombination (HR) during DNA synthesis, DNA double-strand

breaks (DSB) arise (13, 14) that trigger apoptosis in melanoma cell lines (7). The protective

role of RAD51 mediated HR in methylating agent treated cells has been demonstrated in

glioma and melanoma lines (15, 16). DNA repair by MGMT and HR is therefore central in

protecting metastatic melanoma against methylating agent-based therapies.

Class I histone deacetylases (HDAC) have been described to be elevated in gastric,

breast, colorectal, lung and liver cancers (17, 18) and may contribute to those cancers’

response to therapy. In melanoma, pan-HDAC inhibitors in combination with TMZ have

shown promise (19, 20). For these reasons, we addressed the role of class I HDACs in

melanoma and determined the molecular mechanism whereby these HDACs influence the

response of melanomas to TMZ. The class I HDACs are HDAC1, HDAC2, HDAC3 and

HDAC8 (17, 18). They deacetylate histones and non-histone proteins, thereby contributing

to gene and protein regulation. In cancer cells, the inhibition of HDACs has been shown to

influence the execution of the extrinsic and intrinsic apoptosis pathways as well as affecting

DNA repair (21).

Our data reveal that melanoma specimens obtained from patients following surgery

contain nuclear localized HDAC1/2 and that melanoma cell lines overexpress HDAC1/2/3

compared to primary melanocytes, peripheral blood lymphocytes (PBLCs) and fibroblasts.

Pharmacological inhibition of class I HDACs sensitizes melanoma cells to TMZ, fotemustine

(FM) and ionizing radiation (IR) and inhibits tumor growth in a xenograft melanoma tumor

model more effectively than TMZ alone. HDAC1/2/3 inhibition sensitizes melanoma cells to

the TMZ-induced DNA lesion O6MeG due to suppression of HR-mediated DNA repair

caused by down-regulation of RAD51 and FANCD2 on gene and protein level. Finally, we

identified HDAC2 as essential for RAD51 regulation. These findings are a sound basis for

further exploring the targeting of class I HDACs during melanoma therapy with genotoxic

agents.

Materials and Methods

Cell lines and culture conditions

D05 and A2058, SK-Mel187 and Mel537 were generous gifts from C.W. Schmidt

(Queensland Institute of Medical Research, Queensland, Australia) in 2005 and W.K.

Kaufmann (Department of Pathology and Laboratory Medicine, University North Carolina at

Chapel Hill, USA) in 2012, respectively. VH10tert was a gift from L.H.F. Mullenders

(Department of Toxicogenetics, Leiden University Medical Centre, Netherlands) in 2008.

A375 and G361 were purchased from the American Type Culture Collection in 2008. PBLCs

were isolated from buffy-coat blood as described (22). Primary human epidermal

melanocytes light pigmented (Hema-LP) were purchased from Gibco in 2012 and 2013.

Upon receipt of the cell lines they were cryopreserved in liquid-N2 and fresh cell stocks were

used for every battery of tests. All lines were characterized in the laboratory of origin,

displayed the expected phenotype, but were not re-authenticated in our laboratory. D05,

Mel537 and PBLCs were cultured in RPMI-1640 while SK-Mel187, A2058, A375 and

VH10tert were cultured in DMEM. RPMI-1640 and DMEM were supplemented with 10%

fetal bovine serum and penicillin/streptomycin (PAA). Hema-LP was cultured in Medium254

(Gibco) supplemented with human melanocyte growth supplement-2 (Gibco). Cells were

maintained at 37°C in a humidified 5% CO2 atmosphere.

Drugs, treatment and irradiation

The handling and preparation of TMZ (Schering-Plough, Kenilworth, USA), FM (Muphoran,

Servier Research International), O6-benzylguanine (O6BG, Sigma-Aldrich) and Olaparib

(AZD2281, Selleck Chemicals) has been described previously (15, 23). In order to inhibit

MGMT, O6BG was added to cells (final concentration 10 µM) one hour before alkylating

agent treatment. Irradiation was performed using a Cs-137 source. Valproic acid sodium salt

(VPA, Sigma-Aldrich) was dissolved to 100 mM in dH2O and stored at -80 °C after sterile

filtration. Cells were pretreated with VPA (1 mM) for 168 h with refreshment of VPA-

containing medium every 48 h and VPA-removal before further treatment or irradiation. MS-

275 (Entinostat, Selleck Chemicals) was dissolved in DMSO to a stock solution of 5 mM and

stored at -80oC, before use it was diluted in PBS to 1 mM. Cells were pretreated with MS-

275 for 72 h and then removing the compound by medium change.

Tumor biopsies and immunohistochemistry

Samples of malignant melanoma were obtained from patients following surgery. Patient

material was obtained with informed consent and approval from the institutional ethics

committee of the University Medical Center Mainz. Slices were obtained from paraffinized

sections, which were labeled as to the tumor area. Immunohistochemical analysis of

HDAC1/2 levels were performed as described (23). For antibodies see Supplementary Table

S1. Nuclei were stained with TO-PRO-3. Microphotographs were acquired by laser scanning

microscopy (LSM710, Carl Zeiss MicroImaging).Three different tumors were analyzed.

Preparation of protein extracts and Western blot analysis

Preparation of whole cell (24) and nuclear (25) protein extracts were described previously.

Protein concentrations were determined using the Bradford method (26). For antibodies see

Supplementary Table S1.

Determination of apoptosis and cell cycle

Analysis of cell cycle, sub-diploid DNA content (sub-G1) (27) and annexinV-FITC/propidium

iodide double-stained cells have been described (7). Flow cytometry was performed using

FACSCanto II (BD Biosciences) and the data were analyzed with WinMDI (sub-G1,

annexinV/PI) or ModFit (cell cycle).

Determination of MGMT activity

MGMT activity assay was performed as described (28).

Colony survival assay.

G361, SK-Mel187, Mel537, A2058 and A375 melanoma cells were pretreated with VPA for

168 h. Following VPA removal, logarithmic growing cells were seeded allowed to attach for 6

h before O6BG addition 1 h before TMZ treatment. Colonies were fixed in acetic

acid:methanol:H2O (1:1:8), and stained with 1.25% Giemsa and 0.125% crystal violet.

RNA isolation, cDNA synthesis and real-time PCR

Total RNA isolation, cDNA synthesis and real-time PCR were performed as previously

described (29). RNA-levels were normalized to GAPDH and ACTB.

Homologous recombination repair assay

HR repair capacity was assessed with a stably genomic integrated reporter plasmid as

previously described (30). The pDR-GFP reporter plasmid (Addgene plasmid 26475) (30)

and the SceI expressing plasmid pCBASceI (Addgene plasmid 26477) was used (31). The

cells were analyzed 48 h (A375) or 72 h (D05) after pCBASceI transfection, using

Effectene™ (Qiagen), by flow cytometry.

Immunofluorescence

Immunofluorescent labeling of γH2AX-foci was performed as described (15). For antibodies

see Supplementary Table S1.

siRNA, plasmids and stable knockdown

siRNA targeting the mRNA of HDAC1 (Dharmacon), HDAC2 (Santa Cruz Biotechnologies

Inc.) and HDAC 3 (Ambion, Life Technologies) was transfected into A375 and D05 cells

using Lipofectamine RNAimax (Invitrogen). A pSuper (OligoEngine) construct was

generated to express shRNA targeting RAD51 mRNA using the previously described

sequence (5′-GAAGAAAUUGGAAGAAGCU-3′) (32). Plasmid DNA was transfected using

Effectene™ (Qiagen). Transfected cells were selected with 0.75 mg/ml G418 (Invitrogen).

Animal experiments

Immuno-deficient mice (NOD.CB17-Prkdcscid/J) were housed in a sterile environment and

allowed free access to food and water. The animal experiments were approved by the

government of Rhineland-Palatinate and the Animal Care and Use Committee of the

University Medical Center Mainz and performed according to the German federal law and

the guidelines for the protection of animals. A375 human melanoma xenografts were

initiated by injecting 8 x 106 cells in the right and left flank and the treatments began when

tumor volumes [(LxWxHxπ)/6] reached a suitable size (50-100 mm3). Four animals each

were assigned to the different groups and tumor volume was standardized across the

groups. The groups that received VPA pretreatment were treated as follows. Six days before

O6BG and TMZ injection mice were injected intraperitoneal once daily with 500 mg/kg VPA.

The mice received one dose intraperitoneal of TMZ (150 mg/kg). One hour prior to TMZ

injection, all mice received O6BG 30 mg/kg intraperitoneal. Animal weights and tumor

volumes were monitored following TMZ injection. The in vivo experiments were performed

twice, once with female mice (four per group) and once with male mice (four per group). The

presented data show the results from both groups.

Results

Melanoma cells overexpress HDAC1/2/3

Class I HDACs are often overexpressed in cancer. We addressed whether this is the case

for melanomas. Making use of the public accessible microarray data set E-GEOD-3189 from

ArrayExpress and published in (33), the expression of HDAC1, HDAC2 and HDAC3 in 45

primary melanoma samples were compared to 18 benign skin nevi and 7 normal skin

samples. The primary melanoma samples showed a significant overexpression of HDAC1

and HDAC2 compared to the benign nevi and the normal skin tissue (Supplementary Fig.

S1A). Furthermore, melanoma tumor samples analyzed by immunohistochemistry showed

strong HDAC1/2 nuclear staining compared to the cytoplasmic staining (Fig. 1A), while

melanoma cell lines express significantly more HDAC1/2/3 proteins compared to primary

human epidermal melanocytes (Hema-LP), peripheral blood lymphocytes (PBLCs), and

diploid human fibroblasts (VH10tert) (Fig. 1B and C). PBLCs do express class I HDACs

(Supplementary Fig. S1B), but was excluded from further analysis due to its very low

expression, but the significance remained when including them in the analysis

(Supplementary Fig. S1C). As class I HDACs can deacetylate histones 3 (H3) and 4 (H4)

(34), we determined whether inhibition of class I HDACs with valproic acid (VPA) (35)

influences the acetylation level of H3 and H4 in melanoma lines (Fig. 1D). Upon VPA

treatment all the melanoma cell lines showed increased H3 and H4 acetylation levels,

decreased HDAC2 levels in A375, G361, SK-Mel187, Mel537 and D05, which is consistent

with what has been reported previously (36, 37), while HDAC1/3 levels were not influenced

significantly. For the non-cancer cells Hema-LP, PBLCs and VH10tert VPA pretreatment

also caused an increase in acetylated-H3 levels (Supplementary Fig. S1D). Collectively, the

data show that malignant melanomas express HDACs and that the overexpressed class I

HDACs in melanoma cell lines can be inhibited by VPA.

HDAC1/2/3, protect malignant melanoma cells against genotoxin-induced cell death

Having shown that HDAC1/2/3 are overexpressed in melanoma cells, the influence of these

HDACs on genotoxin-induced cell death was determined. Two melanoma cell lines, D05 and

A375, as well as primary human melanocytes, Hema-LP, were pretreated with the HDAC

inhibitor (HDACi) VPA, a fatty acid inhibiting HDAC1/2/3/8 and clinically used for the

treatment of epilepsy (35, 38), and exposed to TMZ, FM or IR (Fig. 2A). It is important to

note that at the administered concentration of TMZ and FM, O6MeG and O6-

chloroethylguanine (O6ClEtG) are the clinically relevant apoptosis-inducing DNA lesions in

melanoma cells (7). For this reason, we performed all experiments, unless stated otherwise,

by inhibiting MGMT with O6BG prior to TMZ or FM treatment. As shown in Fig. 2A, VPA

significantly sensitized the melanoma cells to TMZ, FM and IR induced apoptosis while not

influencing the response of primary human melanocytes.

Furthermore, VPA sensitized melanoma cells to TMZ in a concentration (Fig. 2B) and

pretreatment time-dependent (Fig. 2C) manner. Since TMZ exerts its cytotoxicity through

O6MeG and this lesion is repaired by MGMT (7, 9), the sensitization of melanoma cells by

HDAC inhibition was determined in the absence and presence of MGMT activity using O6BG

(Fig. 2D). MGMT protects not only against TMZ-induced apoptosis, but also against the

sensitization caused by VPA (Fig. 2D), indicating that VPA sensitizes melanoma cells to

O6MeG DNA lesions. As HDAC8 was not overexpressed in melanomas (Fig. 1C), we

determined whether it contributes to the sensitization of melanoma cells to TMZ. This was

accomplished by using the benzamide HDACi MS-275, which specifically inhibits

HDAC1/2/3, but not HDAC8 (38). MS-275 sensitized D05 and A375 melanoma cells to TMZ

in a concentration-dependent manner (Fig. 2E), showing that the catalytic activity of

HDAC1/2/3, but not HDAC8, contributes to the resistance of melanoma cells to TMZ.

As MGMT protects cells against alkylating agents, we determined whether HDAC

inhibition affects this enzyme. HDAC inhibition only marginally influenced MGMT activity in

one cell line tested (A375), while not effecting the other (D05) (Fig. 2F). Seeing that HDAC

inhibition did not influence MGMT activity in D05 cells (Fig. 2F) and HDAC inhibition

sensitizes cells to IR-induced DNA damage (Fig. 2A), which are not subject to repair by

MGMT, we conclude that the sensitization of melanoma cells to TMZ is not due to an effect

on MGMT. Furthermore, cellular replication rate significantly influences the response of cells

to methylating agents (22). To exclude this as a reason for HDACi mediated sensitization to

TMZ, analysis of cell cycle distribution (Supplementary Fig. S2A) and cellular growth rates

(Supplementary Fig. S2B) were performed following VPA pretreatment. VPA neither

changed the cell cycle distribution nor the growth rate of melanoma cells. Collectively, the

data indicate that overexpression of HDAC1/2/3 protects melanoma cell lines against DNA

damage triggered cell death.

Class I HDAC inhibitors promote TMZ-induced cell growth inhibition in melanoma cell

lines and a melanoma xenograft model

To clarify that the sensitization to TMZ of melanoma cells by HDAC inhibition is a general

effect, a panel of melanoma cell lines was subjected to the colony survival assay in the

presence and absence of class I HDAC inhibition (Fig. 3A). All tested melanoma cell lines

were sensitized to TMZ following VPA. When combining the survival data of Fig. 3A, the

sensitization caused by the HDACi in TMZ-treated cells was still apparent (Fig. 3B), showing

that this sensitization effect would very likely also be observed in heterogeneous melanoma

tumors. In order to establish whether the findings obtained in cell lines in vitro can be

translated to melanoma cell growth in vivo, we determined the effect of A375 cells grown

subcutaneously in immunodeficient mice. HDAC inhibition with VPA pretreatment was able

to increase the acetylated-H3 levels in melanoma xenografts (Fig. 3C), showing that VPA

was able to inhibit class I HDACs in this model system. A single treatment with TMZ (150

mg/kg) delayed tumor growth, HDAC inhibition by VPA pretreatment on its own had no effect

on tumor growth, while TMZ (150 mg/kg) in combination with VPA showed significant more

tumor growth inhibition compared to TMZ alone (Fig. 3D and Supplementary Fig. S3A). For

the influence of this treatment schedule on the weight of the mice see supplementary figure

S3B. The melanoma xenograft model results reflect the results obtained in vitro with cell

lines and show that inhibition of class I HDACs enhances melanoma tumor growth inhibition

by TMZ.

Inhibition of HDAC1/2/3 down-regulates RAD51 and FANCD2

To identify how class I HDAC inhibition causes sensitization to TMZ, a real-time PCR array

was performed. Class I HDACs were inhibited in D05 melanoma cells using VPA and

changes in mRNA were monitored by real-time PCR (Fig. 4A and Supplementary Table S2).

Genes involved in apoptosis, BCL2, NOXA, PUMA and BAX, were either up- or down-

regulated while genes involved in DNA repair, RAD51, RAD51D, FANCD2 and KU80, were

down-regulated (Fig. 4A). To validate these results, specific real-time PCR experiments

were performed with two melanoma cell lines, D05 and A375, following VPA pretreatment

(Fig. 4B). Again, up- and down-regulation of these genes was observed. Next, we tested

whether changes in gene expression reflect changes in protein level. To this end, the

influence of VPA on RAD51, FANCD2, RAD51D, ATR, KU80, BAX, NOXA and BCL-2

protein levels was analyzed (Fig. 4C). Following HDAC inhibition only RAD51 and FANCD2

protein levels decreased, suggesting that decreased DNA repair capacity, and not increased

apoptosis fidelity, is the cause for melanoma cell sensitization towards TMZ upon HDACi.

Using RAD51 as a marker for down-regulation of DNA repair following HDAC inhibition,

we addressed the question of time and concentration dependence. The time of class I

HDAC inhibition and the concentration of class I HDACi used for pretreatment showed an

inverse relationship with the RAD51 protein level (Fig. 5A). This inverse relationship affects

the sensitization to TMZ, as TMZ alone induced approximately 6% apoptosis in D05 cells

while increasing VPA pretreatment time and concentration significantly increased TMZ-

induced apoptosis (Fig. 5B). Therefore, the decrease in RAD51 protein (Fig. 5A) mirrored

the increase in melanoma cell sensitivity to TMZ (Fig. 5B). Moreover, following class I

HDACi removal, RAD51 protein levels stayed suppressed for 48 h. TMZ caused a decrease

in RAD51 protein level 72 h after TMZ addition while pretreatment with VPA followed by

TMZ addition suppressed RAD51 protein levels for the duration of the experiment (up to 72

h) (Fig. 5C).

Having demonstrated that the inhibition of HDAC1/2/3/8 causes down-regulation of

RAD51 and FANCD2 in the melanoma cell lines D05 and A375, we expanded our

examination to additional melanoma cell lines: G361, SK-Mel187, Mel537 and A2058.

Following HDAC1/2/3/8 inhibition, all these cell lines showed a decrease in RAD51 and

FANCD2 protein (Fig. 5D). As VPA inhibits HDAC1/2/3/8, the contribution of the individual

HDACs, HDAC1/2/3/8, to the expression of RAD51 and FANCD2 was addressed. Firstly

HDAC8 was assessed by using the HDACi MS-275, which only inhibits HDAC1/2/3, in D05

and A375 melanoma cells. Upon MS-275, the down-regulation of RAD51 and FANCD2 was

still observed (Fig. 6A), showing that down-regulation of these DNA repair proteins are not

caused by the inhibition of HDAC8. To address the contribution of HDAC1, HDAC2 and

HDAC3 to the expression of RAD51, the individual HDACs were knocked-down using siRNA

(Fig. 6B). Only the down-regulation of HDAC2 corresponded to a simultaneous down-

regulation of RAD51, showing that of these three HDACs, only HDAC2 plays a role in the

expression of RAD51.

Inhibition of class I HDACs decreases homologous recombination DNA repair,

thereby sensitizing melanoma cells to TMZ

As the inhibition of HDAC1/2/3 down-regulates RAD51 and FANCD2, and both protect

cancer cells against TMZ (15, 39) due to their role in HR (13, 40), the effect of HDACi on HR

was assayed. Inhibition of class I HDACs significantly decreased HR activity in both D05 and

A375 melanoma cell lines upon DSB induction (Fig. 7A and 7B). Dysregulation of HR has

been implicated in synthetic lethality caused by poly(ADP-ribose)polymerase-1 (PARP-1)

inhibition in cells and cancers defective for this repair pathway (41). Therefore, if down-

regulation of RAD51 mediated by HDACi influences HR repair then class I HDAC inhibition

should sensitize melanoma cells to the PARP-1 inhibitor olaparib. This was in fact the case,

as VPA significantly sensitized D05 cells to olaparib (Fig. 7C). A decrease in DSB repair

caused by class I HDAC inhibition should increase the amount of unrepaired DSBs following

TMZ treatment. To this end the amount of γH2AX foci, a marker for DSBs, was determined

in TMZ treated melanoma cells following VPA. Pretreatment with VPA significantly increased

the amount of unrepaired DSBs following TMZ treatment (Fig. 7D and 7E), showing that

class I HDAC inhibition reduces tolerance of melanoma cells to DSBs that arise during the

processing of O6MeG DNA lesions induced by TMZ. Next, the question of whether RAD51,

and consequently HR, protect melanoma cells against TMZ was addressed. Following

shRNA knockdown of RAD51, A375 cells were significantly sensitized to TMZ-induced

apoptosis (Fig. 7F). The data show that inhibition of class I HDACs suppresses DSB repair

by HR in melanoma cells and that this results in their sensitization to TMZ.

Discussion

Despite our growing understanding of how DNA lesions induced by chemotherapeutics such

as DTIC or TMZ trigger death (42), our grasp of the cellular context in melanomas that

cause intrinsic resistance to these agents is still incomplete and may explain why melanoma

patients face such a bleak prognosis. Examples of why metastatic melanomas are resistant

to therapy are: they bypass cell cycle checkpoints (43), they express low levels of critical

apoptosis proteins (27, 44), they retain p53 wild-type status (45) allowing them to up-

regulate DNA repair genes (DDB2, XPC) (29) and they express an oncogenic form of B-Raf

giving them a growth advantage (46). Here we addressed the contribution of class I HDACs

in malignant melanoma protection against methylating chemotherapeutics, used, despite

tyrosine kinase inhibitors and immune-stimulating biologicals (1, 3), as genotoxic “gold

standard” in the therapy of this cancer.

First, we show that melanoma tumors obtained from pre-treatment patients express

HDAC1/2, and that melanoma cell lines overexpress HDAC1/2/3. These findings point to

possible roles of these HDACs in melanoma tumor formation and in the resistance of this

cancer to therapy. Before we addressed whether HDAC1/2/3 contribute to the resistance of

melanomas to methylating agents we first had to show that these HDACs can be inhibited.

To this end, an inhibitor of class I HDACs (35), namely VPA, was used and confirmed that

this inhibitor is active in melanoma cells.

Having established the experimental system, we demonstrate that inhibition of class I

HDACs sensitized melanoma cells to TMZ-, FM- and IR-induced apoptosis, while not

sensitizing primary melanocytes. Therefore, we conclude that class I HDAC activity protects

melanoma cells against genotoxic interventions. TMZ, FM and IR used in the therapy of

metastatic melanoma induce different spectrums of DNA lesions, repaired by different DNA

repair pathways. Repair of TMZ-induced DNA lesions is mediated by MGMT, base excision

repair (BER) and HR, repair of FM-induced DNA lesions is mediated by MGMT, nucleotide

excision repair (NER) and interstrand crosslink (ICL) repair in which HR plays a role and IR-

induced DNA lesions are repaired by BER, non-homologous end joining (NHEJ) and HR (6,

42). For TMZ, our data show that inhibition of HDAC1/2/3 causes sensitization to the DNA

lesion O6MeG and that this sensitization is likely not due to modulation of MGMT.

Furthermore, the protective role of class I HDAC activity in melanomas was confirmed in a

panel of cell lines in vitro and a melanoma xenograft model in vivo. Collectively the data

revealed that class I HDAC activity protects melanoma cells against methylating agents and

that this class of HDACs is a promising target for intervention during therapy with TMZ.

HDACs are centrally involved in protein regulation and gene expression. Therefore,

HDACs can theoretically play a role in the expression of any gene, which demands

screening methodologies for identifying the factor causing the observed phenotype. To this

end, we employed a real-time PCR array. The array data showed that VPA affects factors

involved in apoptosis and DNA repair on gene expression level. The conclusions drawn from

this are that VPA sensitizes melanoma cells to TMZ either due to better apoptosis fidelity or

less effective DNA repair. In order to test these hypotheses the changes in these two groups

of genes had to be validated by real-time PCR and Western blot analysis. RAD51 and

FANCD2 passed validation. By employing different HDAC inhibitors and siRNA targeting the

specific HDACs, we were able to identify HDAC2 as the class I HDAC responsible for the

regulation of RAD51. Further support for the role of DNA repair in the observed sensitization

to TMZ is found in the results obtained with FM- and IR-treated melanoma cells upon class I

HDAC inhibition. Although TMZ, FM and IR induce different spectrums of DNA lesions, only

HR is involved in the repair of all these genotoxins (47). As RAD51 and FANCD2 play

central roles in HR (48), the down-regulation of these HR proteins may therefore be

responsible for the observed sensitization.

Having identified decreased DNA repair by as the possible mechanism for sensitization to

TMZ upon HDAC1/2/3 inhibition, we addressed the questions of whether class I HDAC

inhibition effects HR activity, whether class I HDAC inhibition results in more unrepaired

DSBs following TMZ treatment and whether RAD51 protects melanoma cells against TMZ-

induced apoptosis. Inhibition of class I HDACs decreased HR-mediated DSB repair activity

by almost 50%. This decreased HR repair activity resulted in an increase in unrepaired

DSBs showing that class I HDAC inhibition prevents melanoma cells from tolerating DSBs

arising due to the processing of the TMZ-induced O6MeG adducts. Furthermore, RAD51

protects melanoma cells treated with TMZ as knockdown of RAD51 significantly sensitized

the cells to TMZ. We should state that this protective role of RAD51 was observed once

MGMT was inhibited and that the DSBs arising from O6MeG is dependent on MMR.

Attenuated HR could sensitize melanoma cells to PARP1 inhibition as synthetic lethality

caused by PARP1 inhibition is not influenced by MGMT or MMR. We show that HDAC

inhibition by VPA exacerbates the killing effect of PARP1 inhibition by olaparib in the

absence of TMZ, and that TMZ could possibly increase this cytotoxicity.

The findings reported here suggest a model where HDAC1/2/3 are key nodes in the

resistance of malignant melanoma cells to methylating agent-based therapies by

ameliorating the HR dependent repair of DSBs arising during the processing of O6MeG DNA

lesions. It forms a firm basis for targeting these HDACs during genotoxic agent based

therapies in malignant melanoma and possibly also in other tumor types.

Acknowledgements

We thank Georg Nagel, Rebekka Kitzinger and Vanessa Steinmetz for technical assistance.

References

1. Tumeh PC, Harview CL, Yearley JH, Shintaku IP, Taylor EJ, Robert L, et al. PD-1

blockade induces responses by inhibiting adaptive immune resistance. Nature.

2014;515:568-71.

2. Snyder A, Makarov V, Merghoub T, Yuan J, Zaretsky JM, Desrichard A, et al.

Genetic Basis for Clinical Response to CTLA-4 Blockade in Melanoma. N Engl J Med. 2014.

3. Chapman PB, Hauschild A, Robert C, Haanen JB, Ascierto P, Larkin J, et al.

Improved survival with vemurafenib in melanoma with BRAF V600E mutation. N Engl J Med.

2011;364:2507-16.

4. Reid JM, Kuffel MJ, Miller JK, Rios R, Ames MM. Metabolic activation of dacarbazine

by human cytochromes P450: the role of CYP1A1, CYP1A2, and CYP2E1. Clin Cancer Res.

1999;5:2192-7.

5. Newlands ES, Stevens MF, Wedge SR, Wheelhouse RT, Brock C. Temozolomide: a

review of its discovery, chemical properties, pre-clinical development and clinical trials.

Cancer Treat Rev. 1997;23:35-61.

6. Kaina B, Christmann M, Naumann S, Roos WP. MGMT: key node in the battle

against genotoxicity, carcinogenicity and apoptosis induced by alkylating agents. DNA

Repair (Amst). 2007;6:1079-99.

7. Naumann SC, Roos WP, Jost E, Belohlavek C, Lennerz V, Schmidt CW, et al.

Temozolomide- and fotemustine-induced apoptosis in human malignant melanoma cells:

response related to MGMT, MMR, DSBs, and p53. Br J Cancer. 2009;100:322-33.

8. Qian XC, Brent TP. Methylation hot spots in the 5' flanking region denote silencing of

the O6-methylguanine-DNA methyltransferase gene. Cancer Res. 1997;57:3672-7.

9. Schraml P, von Teichman A, Mihic-Probst D, Simcock M, Ochsenbein A, Dummer R,

et al. Predictive value of the MGMT promoter methylation status in metastatic melanoma

patients receiving first-line temozolomide plus bevacizumab in the trial SAKK 50/07. Oncol

Rep. 2012;28:654-8.

10. Chen JM, Zhang YP, Wang C, Sun Y, Fujimoto J, Ikenaga M. O6-methylguanine-

DNA methyltransferase activity in human tumors. Carcinogenesis. 1992;13:1503-7.

11. Christmann M, Verbeek B, Roos WP, Kaina B. O(6)-Methylguanine-DNA

methyltransferase (MGMT) in normal tissues and tumors: enzyme activity, promoter

methylation and immunohistochemistry. Biochim Biophys Acta. 2011;1816:179-90.

12. Mojas N, Lopes M, Jiricny J. Mismatch repair-dependent processing of methylation

damage gives rise to persistent single-stranded gaps in newly replicated DNA. Genes Dev.

2007;21:3342-55.

13. Roos WP, Nikolova T, Quiros S, Naumann SC, Kiedron O, Zdzienicka MZ, et al.

Brca2/Xrcc2 dependent HR, but not NHEJ, is required for protection against O(6)-

methylguanine triggered apoptosis, DSBs and chromosomal aberrations by a process

leading to SCEs. DNA Repair (Amst). 2009;8:72-86.

14. Ochs K, Kaina B. Apoptosis induced by DNA damage O6-methylguanine is Bcl-2 and

caspase-9/3 regulated and Fas/caspase-8 independent. Cancer Res. 2000;60:5815-24.

15. Quiros S, Roos WP, Kaina B. Rad51 and BRCA2--New molecular targets for

sensitizing glioma cells to alkylating anticancer drugs. PLoS One. 2011;6:e27183.

16. Kiyohara E, Tamai K, Katayama I, Kaneda Y. The combination of chemotherapy with

HVJ-E containing Rad51 siRNA elicited diverse anti-tumor effects and synergistically

suppressed melanoma. Gene Ther. 2012;19:734-41.

17. West AC, Johnstone RW. New and emerging HDAC inhibitors for cancer treatment. J

Clin Invest. 2014;124:30-9.

18. Wagner T, Brand P, Heinzel T, Kramer OH. Histone deacetylase 2 controls p53 and

is a critical factor in tumorigenesis. Biochim Biophys Acta. 2014.

19. Gatti L, Sevko A, De Cesare M, Arrighetti N, Manenti G, Ciusani E, et al. Histone

deacetylase inhibitor-temozolomide co-treatment inhibits melanoma growth through

suppression of Chemokine (C-C motif) ligand 2-driven signals. Oncotarget. 2014;5:4516-28.

20. Xia C, Leon-Ferre R, Laux D, Deutsch J, Smith BJ, Frees M, et al. Treatment of

resistant metastatic melanoma using sequential epigenetic therapy (decitabine and

panobinostat) combined with chemotherapy (temozolomide). Cancer Chemother Pharmacol.

2014;74:691-7.

21. Falkenberg KJ, Johnstone RW. Histone deacetylases and their inhibitors in cancer,

neurological diseases and immune disorders. Nat Rev Drug Discov. 2014;13:673-91.

22. Roos W, Baumgartner M, Kaina B. Apoptosis triggered by DNA damage O6-

methylguanine in human lymphocytes requires DNA replication and is mediated by p53 and

Fas/CD95/Apo-1. Oncogene. 2004;23:359-67.

23. Roos WP, Quiros S, Krumm A, Merz S, Switzeny OJ, Christmann M, et al. B-Raf

inhibitor vemurafenib in combination with temozolomide and fotemustine in the killing

response of malignant melanoma cells. Oncotarget. 2014;5:12607-20.

24. Bauer M, Goldstein M, Christmann M, Becker H, Heylmann D, Kaina B. Human

monocytes are severely impaired in base and DNA double-strand break repair that renders

them vulnerable to oxidative stress. Proc Natl Acad Sci U S A. 2011;108: 21105-10.

25. Roos WP, Batista LFZ, Naumann SC, Wick W, Weller M, Menck CFM, et al.

Apoptosis in malignant glioma cells triggered by the temozolomide-induced DNA lesion O6-

methylguanine. Oncogene. 2007;26:186-97.

26. Bradford MM. A rapid and sensitive method for the quantitation of microgram

quantities of protein utilizing the principle of protein-dye binding. Analytical Biochemistry.

1976;72:248-54.

27. Roos WP, Jöst E, Belohlavek C, Nagel G, Fritz G, Kaina B. Intrinsic anticancer drug

resistance of malignant melanoma cells is abrogated by IFN-Î² and valproic acid. Cancer

research. 2011;71:4150-60.

28. Preuss I, Eberhagen I, Haas S, Eibl RH, Kaufmann M, von Minckwitz G, et al. O6-

methylguanine-DNA methyltransferase activity in breast and brain tumors. Int J Cancer.

1995;61:321-6.

29. Barckhausen C, Roos WP, Naumann SC, Kaina B. Malignant melanoma cells

acquire resistance to DNA interstrand cross-linking chemotherapeutics by p53-triggered

upregulation of DDB2/XPC-mediated DNA repair. Oncogene. 2014;33:1964-74.

30. Pierce AJ, Johnson RD, Thompson LH, Jasin M. XRCC3 promotes homology-

directed repair of DNA damage in mammalian cells service XRCC3 promotes homology-

directed repair of DNA damage in mammalian cells. Genes & development. 1999:2633-8.

31. Richardson C, Moynahan ME, Jasin M. Double-strand break repair by

interchromosomal recombination: suppression of chromosomal translocations. Genes &

Development. 1998;12:3831-42.

32. Biard DS. Untangling the relationships between DNA repair pathways by silencing

more than 20 DNA repair genes in human stable clones. Nucleic Acids Res. 2007;35:3535-

50.

33. Talantov D, Mazumder A, Yu JX, Briggs T, Jiang Y, Backus J, et al. Novel genes

associated with malignant melanoma but not benign melanocytic lesions. Clin Cancer Res.

2005;11:7234-42.

34. Secrist JP, Zhou X, Richon VM. HDAC inhibitors for the treatment of cancer. Current

opinion in investigational drugs. 2003;4:1422-7.

35. Göttlicher M, Minucci S, Zhu P, Krämer OH, Schimpf a, Giavara S, et al. Valproic

acid defines a novel class of HDAC inhibitors inducing differentiation of transformed cells.

The EMBO journal. 2001;20:6969-78.

36. Kramer OH, Zhu P, Ostendorff HP, Golebiewski M, Tiefenbach J, Peters MA, et al.

The histone deacetylase inhibitor valproic acid selectively induces proteasomal degradation

of HDAC2. EMBO J. 2003;22:3411-20.

37. Kramer OH, Baus D, Knauer SK, Stein S, Jager E, Stauber RH, et al. Acetylation of

Stat1 modulates NF-kappaB activity. Genes Dev. 2006;20:473-85.

38. Bradner JE, Mak R, Tanguturi SK, Mazitschek R, Haggarty SJ, Ross K, et al.

Chemical genetic strategy identifies histone deacetylase 1 (HDAC1) and HDAC2 as

therapeutic targets in sickle cell disease. Proc Natl Acad Sci U S A. 2010;107:12617-22.

39. Patil AA, Sayal P, Depondt ML, Beveridge RD, Roylance A, Kriplani DH, et al.

FANCD2 re-expression is associated with glioma grade and chemical inhibition of the

Fanconi Anaemia pathway sensitises gliomas to chemotherapeutic agents. Oncotarget.

2014;5:6414-24.

40. Arnaudeau C, Helleday T, Jenssen D. The RAD51 protein supports homologous

recombination by an exchange mechanism in mammalian cells. J Mol Biol. 1999;289:1231-

41. Gottipati P, Vischioni B, Schultz N, Solomons J, Bryant HE, Djureinovic T, et al.

Poly(ADP-ribose) polymerase is hyperactivated in homologous recombination-defective

cells. Cancer Res. 2010;70:5389-98.

42. Roos WP, Kaina B. DNA damage-induced cell death: From specific DNA lesions to

the DNA damage response and apoptosis. Cancer Lett. 2013;332:237-48.

43. Omolo B, Carson C, Chu H, Zhou Y, Simpson DA, Hesse JE, et al. A prognostic

signature of G(2) checkpoint function in melanoma cell lines. Cell Cycle. 2013;12:1071-82.

44. Soengas MS, Capodieci P, Polsky D, Mora J, Esteller M, Opitz-Araya X, et al.

Inactivation of the apoptosis effector Apaf-1 in malignant melanoma. Nature. 2001;409:207-

11.

45. Castresana JS, Rubio MP, Vazquez JJ, Idoate M, Sober AJ, Seizinger BR, et al.

Lack of allelic deletion and point mutation as mechanisms of p53 activation in human

malignant melanoma. Int J Cancer. 1993;55:562-5.

46. Davies H, Bignell GR, Cox C, Stephens P, Edkins S, Clegg S, et al. Mutations of the

BRAF gene in human cancer. Nature. 2002;417:949-54.

47. Helleday T, Petermann E, Lundin C, Hodgson B, Sharma RA. DNA repair pathways

as targets for cancer therapy. Nat Rev Cancer. 2008;8:193-204.

48. San Filippo J, Sung P, Klein H. Mechanism of eukaryotic homologous recombination.

Annu Rev Biochem. 2008;77:229-57.

Figure legends

Figure 1: HDAC1/2 in melanoma tumors and class I HDAC protein levels in melanoma cell

lines. (A) Representative immunohistochemistry microphotographs of melanoma tumors.

HDAC1/2 (green) and TO-PRO-3 nuclei-stained are shown. (B) Western blot analysis of

HDAC1/2/3/8 protein levels in D05, G361, SK-Mel187, Mel537, A2058 and A375 melanoma

cells compared to normal (non-cancerous) cells PBLCs, VH10tert and Hema-LP. β-actin

served as loading control. Relative expression (R.E.) compared to the lowest protein level is

shown. PBLC protein level was not used (n/u) for further analysis. (C) Relative HDAC

protein level in melanoma cells compared to normal cells as quantified by Western blot. Data

are represented as mean ± SEM. **p<0.005. (D) Influence of class I HDAC inhibition by VPA

(1 mM for 168 h) on acetylated-H3, acetylated-H4 and HDAC1/2/3. H3 and ERK2 served as

loading controls. R.E. of acetylated-H3 and acetylated-H4 before and after VPA treatment is

shown.

Figure 2: Inhibition of HDAC1/2/3 sensitizes melanoma cells to genotoxin based

therapeutics. (A) Melanoma cells (D05 and A375) and primary human melanocytes (Hema-

LP) were pretreated with VPA, then cells were exposed to 50 µM TMZ, 32 µM FM or 5 Gy IR

and the apoptosis response was determined 120 h and 72 h after alkylating agents and IR

exposure, respectively. (B) VPA pretreatment sensitizes to TMZ concentration dependently.

Apoptosis was assayed 120 h after TMZ addition. (C) VPA pretreatment sensitizes to TMZ

time-dependently. Apoptosis was assayed following 50 µM TMZ exposure at the indicated

time points. (D) HDAC inhibition sensitizes to O6MeG. VPA pretreatment followed by 50 µM

TMZ in the presence and absence of MGMT inhibition by O6BG (10 µM). Apoptosis was

assayed 120 h after TMZ exposure. (E) MS-275 pretreatment for 72 h at indicated

concentrations sensitizes D05 and A375 cells to TMZ (50 µM). Apoptosis was determined

120 h after TMZ addition. (F) MGMT repair activity in D05 and A375 cells with and without

VPA pretreatment. Data are presented as mean ±SEM. *p<0.05, **p<0.005 and

***p<0.0001. For A, B, D and E the apoptotic response was determined by flow cytometric

analysis of annexin V-FITC/propidium iodide double-stained cells, and for C by sub-G1

analysis.

Figure 3: Class I HDACs protect melanoma cells in vitro and in vivo. (A) Colony survival in

G361, SK-Mel187, Mel537, A2058 and A375 melanoma cells in the presence and absence

of VPA pretreatment exposed to the indicated concentrations of TMZ. (B) Compilation of the

survival data obtained in A for G361, SK-Mel187, Mel537, A2058 and A375. (C) Western

blot analysis of acetylated-H3 in A375 xenografts of mice treated with VPA. H3 served as

loading control. (D) Effect of HDAC inhibition on growth inhibition of A375 xenografts

following TMZ. Relative tumor volume is shown as a function of time. See materials and

methods for detailed treatment conditions. Data are presented as mean ±SEM. *p<0.05,

**p<0.005 and ***p<0.0001.

Figure 4: Inhibition of class I HDACs down-regulates of RAD51 and FANCD2 on RNA and

protein level. (A) Real-time PCR array results showing the down- and up-regulation of the

listed genes. HDACs were inhibited by VPA in D05 cells and fold change in gene expression

was calculated compared to untreated controls. (B) Validation of gene regulation using real-

time PCR. HDACs were inhibited by VPA pretreatment in D05 and A375 cells and fold

change in gene expression was calculated compared to untreated controls. (C) Western blot

analysis of RAD51, FANCD2, RAD51D, ATR, KU80, BAX, NOXA and BCL-2 protein levels.

β-actin served as loading control and relative expression (R.E.) levels are indicated.

Figure 5: Inhibition of HDAC1/2/3/8 down-regulates RAD51 and sensitizes in a

concentration and time dependent manner. (A) Time and VPA concentration dependence of

RAD51 down-regulation as determined by Western blot analysis. RPA served as loading

control. (B) Time and VPA concentration dependence of sensitization of melanoma cells to

50 µM TMZ. The dashed line denotes the amount of apoptosis induced by TMZ in the

absence of HDAC inhibition. (C) Long term RAD51 down-regulation following VPA removal

as determined by Western blot analysis. D05 cells were treated with VPA for 168 h and then

the VPA was removed. RAD51 protein levels are shown at indicated time points following

VPA removal, TMZ addition or combination treatment. (D) RAD51 and FANCD2 protein

levels following VPA in a panel of melanoma cell lines, G361, SK-Mel187, Mel537 and

A2058. ERK2 and HSP90 served as loading controls. Data are represented as mean ±

SEM. *p<0.05, **p<0.005 and ***p<0.0001.

Figure 6: Inhibition of HDAC1/2/3 and knockdown of HDAC2 down-regulates RAD51. (A)

RAD51 and FANCD2 protein levels following HDAC1/2/3 inhibition by MS-275. (B) Western

blot analysis of siRNA knockdown of HDAC1, HDAC2 and HDAC3. Protein levels of HDAC1,

HDAC2, HDAC3 and RAD51 are shown. ß-actin served as loading control. siSC depicts

non-targeting siRNA. R.E. relative expression.

Figure 7: Inhibition of class I HDACs in melanoma cells suppresses DSB repair by HR that

causes the sensitization to TMZ. (A) Representative flow cytometry dot plots showing GFP

and auto-fluorescence following HDAC inhibition by VPA, DSB induction by Sce1 expression

or combination of both. (B) Relative HR activity following HDAC inhibition. D05 and A375

cells containing a stable integrated substrate for HR were transiently transfected with a

plasmid that expresses Sce1 in the presence or absence of VPA pretreatment. Flow

cytometric analysis was performed 72 h and 48 h after transient transfection for D05 and

A375 cells, respectively. (C) Effect of HDAC inhibition on synthetic lethality induced by

PARP-1 inhibition. D05 cells were treated with the indicated concentrations of the PARP-1

inhibitor olaparib in the presence or absence of VPA pretreatment. Apoptosis was assayed

72 h after olaparib addition. (D) Representative microphotographs showing γH2AX foci

(white) and TO-PRO-3 nuclear staining (blue). TMZ (10 µM) treated D05 cells in the

presence and absence of VPA pretreatment. (E) Quantification of TMZ-induced (10 µM)

γH2AX foci in the presence and absence of VPA pretreatment 48 h after TMZ addition in

D05 cells. (F) Knockdown of RAD51 sensitizes melanoma cells to TMZ. Left: Western blot

analysis of RAD51 protein level in A375 cells following RAD51 knockdown. PCNA was used

as loading control. Right: Apoptotic response of A375, A375 empty vector control and two

RAD51 knockdown clones, A375shRAD51c2 and A375shRAD51c26 following exposure to

indicated concentrations of TMZ. Apoptosis was assayed 120 h after TMZ addition. Data are

presented as mean ±SEM. *p<0.05, **p<0.005 and ***p<0.0001. For C and F apoptosis was

assayed by flow cytometric analysis of annexin V-FITC/propidium iodide double-stained

cells.

Enhanced histone deacetylase activity in malignant melanoma provokes RAD51 and FANCD2 triggered drug resistance

Andrea Krumm, Christina Barckhausen, Pelin Kücük, et al.

Cancer Res Published OnlineFirst March 15, 2016.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-15-2680

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2016/03/15/0008-5472.CAN-15-2680.DC1

Author Author manuscripts have been peer reviewed and accepted for publication but have not yet been Manuscript edited.

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 Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/early/2016/03/15/0008-5472.CAN-15-2680. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.