The Epigenetic Regulator I-BET151 Induces BIM-Dependent Apoptosis and Cell Cycle Arrest of Human Melanoma Cells

The Epigenetic Regulator I-BET151 InducesBIM-Dependent Apoptosis and Cell Cycle Arrest ofHuman Melanoma CellsStuart J. Gallagher1, Branka Mijatov1, Dilini Gunatilake1, Jessamy C. Tiffen1, Kavitha Gowrishankar1, Lei Jin1,Gulietta M. Pupo2, Carleen Cullinane3,4, Rab K. Prinjha5, Nicholas Smithers5, Grant A. McArthur3,4,Helen Rizos2 and Peter Hersey1,6

Epigenetic changes are widespread in melanoma and contribute to the pathogenic biology of this disease. In thepresent study, we show that I-BET151, which belongs to a new class of drugs that target the BET family ofepigenetic ‘‘reader’’ proteins, inhibits melanoma growth in vivo and induced variable degrees of apoptosis in apanel of melanoma cells. Apoptosis was caspase dependent and associated with G1 cell cycle arrest. Allmelanoma cells tested had increased levels of the BH3 proapoptotic protein BIM, which appeared to be regulatedby the BRD2 BET protein and to some extent by BRD3. In contrast, knockdown experiments indicated thatinhibition of BRD4 was associated with decreased levels of BIM. Apoptosis was dependent on BIM in some butnot all cell lines, indicating that other factors were determinants of apoptosis, such as downregulation ofantiapoptotic proteins revealed in gene expression arrays. G1 cell cycle arrest appeared to be mediated by p21and resulted from inhibition of the BRD4 protein. The activity of BET protein inhibitors appears independent ofthe BRAF and NRAS mutational status of melanoma, and further studies to assess their therapeutic role inmelanoma are warranted.

Journal of Investigative Dermatology advance online publication, 3 July 2014; doi:10.1038/jid.2014.243

INTRODUCTIONAlthough there have been major advances in the treatment ofmelanoma based on the use of targeted therapies (Jang andAtkins, 2013; Sullivan et al., 2013) and immunotherapy withmonoclonal antibodies against checkpoint inhibitors (Riley,2013; Robert et al., 2013), the disease remains incurable inthe majority of patients. We have reviewed elsewhere some ofthe evidence that epigenetic changes underlie many aspects ofmelanoma biology (Lai et al., 2012b), including resistance totargeted therapies, melanoma invasiveness (Saladi et al.,2010), and regulation of microphthalmia-associatedtranscription factor (Keenen et al., 2010). These and studies

by others indicate that dysregulation of chromatin is a frequentevent in melanoma (Becker et al., 2009; Hodis et al., 2012;van den Hurk et al., 2012) and other cancers (Fullgrabe et al.,2011; Blair and Yan, 2012; Dawson and Kouzarides, 2012);yet, the role of new therapies targeting the epigenetic changesin melanoma remains under explored.

Chromatin regulation is largely due to the methylationof DNA or due to modification of histones. Histones aremodified by several protein groups that act either as ‘‘writers’’by adding acetyl, methyl, or other groups to histones or as‘‘erasers’’ that remove such groups and included histonedeacetylases (HDACs) and demethylases. The writers anderasers establish what is referred to as a ‘‘histone code’’ thatis ‘‘read’’ by a third family of proteins that recognize thehistone code and act to focus large protein complexes to thosesites (Arrowsmith et al., 2012; Dawson et al., 2012). Theseprotein complexes or modules appear characteristic of thegene expression status, such as repression or activation, andthe particular tissues involved (Ram et al., 2011; Ernst andKellis, 2012). For example, mapping of just nine acetyl andmethyl marks was found to define 15 chromatin states relatingto transcriptional activity of surrounding genes (Arrowsmithet al., 2012).

Particular interest has focused on the development ofinhibitors against a highly conserved class of ‘‘reader’’proteins, referred to as bromodomain and extraterminal (BET)proteins. The BET family, which consists of BRD2, BRD3,

ORIGINAL ARTICLE

1Melanoma Research Group, Kolling Institute of Medical Research, Universityof Sydney, St Leonards, New South Wales, Australia; 2Westmead Institute forCancer Research, The University of Sydney at Westmead Millennium Institute,Westmead Hospital, Westmead, New South Wales, Australia; 3TranslationalResearch Laboratory, Peter MacCallum Cancer Centre, Melbourne, Victoria,Australia; 4Oncogenic Signalling and Growth Control Program, PeterMacCallum Cancer Centre, Melbourne, Victoria, Australia; 5Epinova DiscoveryPerformance Unit, GlaxoSmithKline, Stevenage, UK and 6Melanoma Instituteof Australia, North Sydney, New South Wales, Australia

Correspondence: Peter Hersey, Melanoma Research Group, Kolling Institute ofMedical Research, University of Sydney, Pacific Hwy/RNSH, St Leonards, NewSouth Wales 2065, Australia. E-mail: [email protected]

Received 16 January 2014; revised 14 March 2014; accepted 8 April 2014;accepted article preview online 6 June 2014

Abbreviations: BET, bromodomain and extraterminal proteins; HDAC, histonedeacetylases

& 2014 The Society for Investigative Dermatology www.jidonline.org 1

BRD4, and the testis-specific BRDT, has two bromodomains inthe N-terminal region that bind to acetylated lysines in histonesand a C-terminal region that binds to transcription elongationfactors (Filippakopoulos et al., 2012). The bromodomains actto target protein complexes to particular chromosomal regionsinvolved in gene transcription, and act as coregulators oftranscription (Belkina and Denis, 2012; Prinjha et al., 2012).The first BET inhibitor JQ1 inhibited the growth of rareNUT (nuclear protein of testis) midline carcinomas that arecaused by fusion of the BRD4 BET protein with theNUT(Filippakopoulos et al., 2010). Related BET protein inhibitorsI-BET151 and I-BET762 have been described by Nicodemeet al. (2010) (Nicodeme et al., 2010) and shown to inhibit BETproteins BRD2, BRD3, and BRD4. Subsequent studies haveshown that BRD4 appears to be the target of these drugs inAML (Zuber et al., 2011), neuroblastoma (Puissant et al.,2013a), and melanoma (Segura et al., 2013b). The effects onAML and of other c-Myc-dependent tumors were believed tobe largely due to the downregulation of c-Myc (Delmore et al.,2011; Mertz et al., 2011).

We and others have shown that pan-HDAC inhibitors caninduce apoptosis in melanoma that is associated with upre-gulation of BIM, BAX, and BIK and downregulation of Bcl-XLand XIAP (Zhang et al., 2004; Gillespie et al., 2006; Lai et al.,2012b). Such inhibitors were also strongly synergistic withselective BRAF inhibitors in induction of apoptosis ofmelanoma (Hersey et al., 2010). In the present study, weexamined the influence of the BET protein inhibitor I-BET151on the survival of a series of human melanoma cells includingshort-term melanoma cultures established from patients whorelapsed while on treatment with the BRAF inhibitor vemu-rafenib. We report that I-BET 151 induces cell cycle arrest andapoptosis in melanoma and inhibits melanoma cell survival.These effects are at least in part due to the inhibition of BET-mediated suppression of the cyclin-dependent kinase inhibitorp21 and BET-mediated activation of the proapoptotic BH3-only protein BIM.

RESULTSI-BET151 induces melanoma cell death and cell cycle arrest andinhibits melanoma growth in vivoTo determine the effect of I-BET151 on the growth ofmelanoma cells, we examined the sensitivity of Me1007melanoma cells over a range of drug doses and cultureperiods. As shown in Figure 1a, growth inhibition was seenafter 48–72 hours at both 10mM and 100mM of drug, and alsoafter 24 hours at the higher concentration. In view of this andpharmokinetic studies showing attainable in vivo blood con-centrations of at least 10mM for 12 hours following dosing(Dawson et al., 2011), subsequent studies were carried out atthe lower 10mM I-BET151 concentration after 48 hourstreatment, unless shown otherwise. I-BET151 also inhibitedthe growth of other melanoma cell lines to various degrees(Supplementary Figure S1a).

To assess whether I-BET151 also had activity in vivo, wetested the effects of I-BET151 in NOD/SCID mice with mela-noma xenografts. Xenografts were established by injectingPatient-1-post cells into the flank of NOD/SCID mice, and

after tumors had established the mice were given 25 mg kg� 1

I-BET151 or vehicle once daily by oral gavage for 7 daysfollowed by a 2-day break before treatment was recom-menced at 20 mg kg�1. This schedule was well tolerated withmean body weight not exceeding 10% (Supplementary FigureS1b). I-BET151 treatment significantly inhibited Patient-1-posttumor growth by 56% on day 15 (Po0.001, t-test; Figure 1b).

We then assessed whether I-BET151 could induce apoptosisor inhibit the cell cycle in a panel of melanoma cell lines.Little apoptosis was observed after 24 hours of I-BET151treatment (data not shown); however, apoptosis was observedin some lines after 48 hours (Figure 1c), suggesting relativelyslow kinetics of apoptosis induction. Of the continuousmelanoma cell lines, the NRAS mutant cell line (Mel-RM)and the NRAS/BRAF wild-type line (Me1007) were sensitive toI-BET151, whereas the NRAS mutant/BRAF wild-type (Mel-JD)and NRAS wild-type/BRAF mutant cell lines (SK-Mel-28, Mel-RMU) were relatively insensitive to this drug (Figure 1c andSupplementary Figure S2). Apoptosis was induced at highlevels in the pretreatment lines from patient 3 and less so inthe pretreatment line from patient 1. Both post-relapse linesfrom these patients were relatively resistant in apoptosisassays. Normal neonatal human epidermal melanocytes anddermal fibroblasts did not undergo cell death with I-BET151treatment (Figure 1c).

G1 arrest was induced by I-BET151 and was more evidentin the pretreatment primary cultures from patients 1 and 3than in the corresponding post-relapse lines (Figure 1dand Supplementary Figure S3). A reduction in cells enteringS-phase and replicating DNA was confirmed usingBrdU/7-AAD dual staining in Patient-1-pre and Patient-3-pre cells, confirming G1 arrest (supplementary Figure S4A).Varied levels of G1 arrest were documented in the continuouscell lines, ranging from a prominent arrest (Mel-RM) tonegligible levels in Mel-RMU. There was relatively little G1arrest in the sensitive Me1007, although S-phase inhibitionand an increase in G2/M population were observed. Adecrease in S-phase was observed in all cell lines (P-valueo0.05, Welch’s t-test).

I-BET151 upregulates the proapoptotic BH3 protein BIM but notother BH3 proteins

To investigate the mechanism of I-BET151-induced apoptosis,we carried out western blots on a range of pro- andantiapoptotic proteins in five of the melanoma cell lines. Asshown in Figure 2a, after 48 hours of I-BET151 treatment, therewas consistent upregulation of BIM-EL and its isoforms anddownregulation of Puma, Bik, tBID, and BMF to varyingdegrees in the different cell lines. SK-Mel-28 had high levelsof BIM in the untreated and treated samples. There wereminimal or variable changes in BAX, BAK, and NOXA.Among the antiapoptotic proteins, there were consistentdecreases in the XIAP inhibitor of apoptosis protein, smallincreases in MCL-1 levels, and small decreases in Bcl-2 andBcl-XL. The increases in BIM (particularly the short isoform)and decreases in the antiapoptotic proteins were most evidentin the two relatively sensitive lines Me1007 and Mel-RM, andthis coincided with PARP cleavage and the presence of

SJ Gallagher et al.I-BET151 Induces Apoptosis and Arrest in Melanoma

2 Journal of Investigative Dermatology (2014), Volume 00

cleaved caspase 3 (Figure 2a). Similar changes were seenin the western blots of cell lines from patients 1 and 3(Supplementary Figure S4B).

Apoptosis was caspase dependent and mediated in part by BIM

We investigated whether apoptosis was caspase dependent bypretreating cells with the pan-caspase inhibitor Q-VD-OPh(Kuzelova et al., 2011). As shown in Figure 2b, caspaseinhibition almost completely prevented I-BET151-mediatedapoptosis in the three most sensitive cell lines. Tumor necrosisfactor–related apoptosis-inducing ligand-induced apoptosiswas used as a positive control and is known to be caspasedriven. Previous studies have shown that Me1007 cellslack caspase 8 and are resistant to tumor necrosis factor–related apoptosis-inducing ligand (Hersey and Zhang, 2001).Caspase-dependent apoptosis can occur via the intrinsic

apoptotic pathway, which involves depolarization ofmitochondria, controlled by changes in the Bcl-2 familymembers, including BIM. Measurement of mitochondrialouter membrane potential by JC-1 showed that I-BET151treatment leads to depolarized mitochondria in Me1007,Mel-RM, and Patient-3-pre cells (Figure 2c).

BIM is a proapoptotic member of the Bcl-2 family thatcontrols depolarization of mitochondria, leading to apoptosisvia caspase-dependent and -independent mechanisms. Theincrease in BIM protein levels (Figure 2a) appeared to be dueto transcriptional upregulation as shown by the increase inmRNA levels of BIM 6 hours after I-BET151 treatment(Figure 2d). This was also evident in the melanocyte andfibroblast lines.

To further investigate the role of BIM in I-BET151-inducedapoptosis, BIM was knocked down by siRNA in the sensitive

% P

opul

atio

n

Primary lines

0

50

100

Patien

t-3-p

re

Patien

t-3-p

ost

Patien

t-1-p

re

Patien

t-1-p

ost

HDF

HEMn

I-BET151 – –– –– +++++I-BET151 – –– –– +++++

Continuous lines

0

50

100

G1SG2/M

% P

opul

atio

n

Me1

007

SK-Mel-

28

Mel-

RMU

Mel-

JD

Mel-

RM

Continuous lines

Me1

007

SK-Mel-

28

Mel-

RMU

Mel-

JD

Mel-

RM

0

10

20

30

40ControlI-BET151

Cel

l dea

th %

Patien

t-3-p

re

Patien

t-3-p

ost

Patien

t-1-p

re

Patien

t-1-p

ost

HDF

HEMn

0

20

40

60ControlI-BET 151

Cel

l dea

th %

Primary lines

Me1007

0 h

24 h

48 h

72 h

0

1

2

3

4

5

0 μM

1 μM

10 μM

100 μM

Rel

ativ

e ce

ll nu

mbe

r

[I-BET151] Mean tumor volume

0 2 4 6 8 10 12 14 16

40

80

120VehicleI-BET151

Day

Tum

or v

olum

e (m

m3 )

Figure 1. I-BET151 inhibits the growth of melanoma cells in vitro and in vivo and induces cell death. (a) Me1007 melanoma cells were treated with 1, 10, 100mM

I-BET151, or vehicle control and cell growth was assessed over 72 hours using CellTiter Glo assay, with data normalized to a 0-h vehicle control. Error bars: SEM,

n¼ 8. (b) Average tumor volume of xenografted Patient-1-post cells in NOD/SCID mice treated with I-BET151 or vehicle control. (c) Cell death of matched pre/post

primary lines, human dermal fibroblasts (HDF), human melanocytes (HEM), and continuous melanoma cell lines after 48 hours of 10mM I-BET151 treatment was

quantified by annexin-V staining. (d) The cell cycle distribution was assessed using flow cytometry. Results are from at least three individual experiments.

SJ Gallagher et al.I-BET151 Induces Apoptosis and Arrest in Melanoma

www.jidonline.org 3

Me1007 cells and patient-3-pre cells. As shown in Figure 2e,there was a substantial (480%) reduction in I-BET151-inducedapoptosis in the Me1007 cell line and a smaller reduction inthe Patient-3-pre line. Effective reduction of BIM protein wasconfirmed by western blotting (Figure 2f). These results areborne out by the photomicrographs shown in SupplementaryFigure S5A, showing less cell death following I-BET151treatment in Me1007 cells with BIM ablated by siRNA.

I-BET151 induces cell cycle arrest by upregulation of p21

We examined the known cell cycle regulatory genes afterculture of the melanoma lines in 10mM I-BET151 for 48 hours.As shown in Figure 3a and b, there was a consistent increasein p21 and Cyclin D in the continuous lines and p21 in theprimary patient–derived lines. There was also an increase inp27 and p53 in the latter lines. p53 increased in all the linesexcept in Mel-RM and patient-3-post line. Bromodomaintarget c-myc (Dawson et al., 2011; Puissant et al., 2013a)was not detectably altered in our panel of melanoma cells.

To further investigate the role of p21 we knocked downits expression by siRNA (Figure 3d). This diminished the G1

arrest induced by I-BET151 in the pretreatment cell lines frompatients 1 and 3 (Figure 3c), consistent with the dominant roleof this CDK inhibitor in regulating G1–S progression.

BRD2 proteins appear responsible for suppression of BIM,whereas BRD4 suppresses CDKN1A (p21)To determine which BRD proteins were involved in regulationof BIM and p21, we knocked down the individual BRDproteins using two separate siRNAs for each gene. As shownin Figure 4a and b, BRD3 and BRD4 transcript levels wereeffectively suppressed (60–80%) by the targeted siRNAs in theMe1007 and Patient-3-pre cell lines. BRD2 message was alsoknocked down by more than 50%, except by BRD2 siRNA#1in the patient-3-pre treatment line.

As shown in Figure 4c and d, knockdown of BRD2 resultedin increased expression of BIM (BCL2L11) in both Me1007and Patient-3-pre cells, although the second BRD2 silenceronly caused a small, but statistically significant, increase inBIM in Me1007 (Figure 4c). In contrast, knockdown of BRD4by both silencing molecules resulted in a marked decrease inthe expression of the BIM gene in both melanoma lines.

Caspase inhibition

% C

ell d

eath

Me1007 Mel-RM Patient-3-pre

Con

trol

I-B

ET

151

TR

AIL

Con

trol

Con

trol

I-B

ET

151

I-B

ET

151

TR

AIL

TR

AIL

0

20

40

60

80

ControlQ-VD-OPh

si-C

ont

si-C

ont

si-B

IM#1

si-B

IM#1

si-B

IM#2

si-B

IM#2

Me1007 Patient-3-pre

0

20

40

60

80 ControlI-BET151

% C

ell d

eath

BIM knockdownBIM mRNA

Pat

ient

-1-p

re

Pat

ient

-1-p

ost

Pat

ient

-3-p

re

Pat

ient

-3-p

ost

HE

M

HD

F

0

5

10

15

20ControlI-BET151

mR

NA

rel

ativ

eex

pres

sion

Mitochondrial depolarization

0

20

40

60

80

% C

ells

with

ΔM

OM

P ControlI-BET151

Pat

ient

-3-p

re

Mel

-RM

Me1

007

Me1007 Patient-3-pre

I-BET151

a b c

d e

f

BIMLBIMS

BIMEL

β-Actin

– –+ +

siC

ont

siB

IM #

1

– –+ +

siC

ont

siB

IM #

2I-BET151

β-Actin

I-BET151 – –+ +

siC

ont

siB

IM #

1

– –+ +

siC

ont

siB

IM #

2

BIMLBIMS

BIMEL

β-Actin

PARP

cl. caspase 3

XIAP

MCL1

BCL-XL

BCL2

Bik

BMF

NOXA

PUMA

tBID

BID

BIMS

BIML

BIMEL

BAK

BAX

– – – – –+ + + + +

Me1

007

SK

-Mel

-28

Mel

-RM

U

Mel

-JD

Mel

-RM

25 kDa

20 kDa

15 kDa10 kDa

19 kDa17 kDa

Figure 2. I-BET151 increases BIM and induces caspase-dependent apoptosis. (a) The levels of apoptosis-related proteins were measured by western blotting

melanoma cell lines treated with 10mM I-BET151 for 48 hours. (b) Apoptosis was measured in cells pretreated with 10mM of the pan-caspase inhibitor Q-VD-OPh

and then treated for 48 hours with 10mM I-BET151. TRAIL treatment was used as a positive control to activate caspase-dependent apoptosis in Mel-RM and Patient-

3-pre cells. (c) Mitochondrial depolarization of cell lines treated with 10mM I-BET151 for 48 hours was measured with JC-1 staining. (d) BIM mRNA was measured

in cells by real-time RT-PCR after 6 hours of 10mM I-BET151 treatment. (e) Apoptosis in cells after BIM was ablated using two separate siRNAs and cells were

treated with 10mM I-BET151 for 48 hours and (f) the efficiency of BIM knockdown was shown by western blotting. The panels showing BIML and BIMS are darker

exposures of the same blot as BIMEL, to allow easier visualization of these bands. RT-PCR, reverse transcriptase–PCR; siRNA, small interfering RNA; TRAIL, tumor

necrosis factor–related apoptosis-inducing ligand.

SJ Gallagher et al.I-BET151 Induces Apoptosis and Arrest in Melanoma

4 Journal of Investigative Dermatology (2014), Volume 00

Ablation of BRD3 did not affect BIM mRNA levels. Theseresults were reflected at the protein level as determined bywestern blot (Supplementary Figure S5B).

Knockdown of BRD proteins increases apoptosis and cell cyclearrest

Although there was a clear difference in the effects of siRNAfor BRD2 and 3 compared with BRD4 on mRNA for BIM,knockdown of each of the three genes increased apoptosis inthe patient-3-pre line, whereas only BRD2 siRNA increasedapoptosis in Me1007 (Figure 4c and d). The fact that BRD4knockdown can cause apoptosis in patient-3-pre cells whilereducing BIM levels suggests that BRD4 normally suppressesother mediators of apoptosis or upregulates antiapoptoticproteins like Bcl-2, which was slightly reduced followingBRD4 ablation (Supplementary Figure S5B).

Both siRNAs against BRD4 produced G1 cell cycle arrest tovarying extent in both Me1007 and the line from patient 3(Figure 4e and f), whereas only one each of the siRNA pairsagainst BRD2 and BRD3 produced G1 arrest in the Me1007line (Figure 4e).

I-BET151 upregulates proapoptotic and cell cycle arrest genes inmelanoma cells

We used gene expression microarrays to examine changes inglobal gene expression 6 and 24 hours after exposure of thecell lines from patient 1 and 3 to I-BET151. Using these data,we first examined genes relating specifically to induction ofapoptosis. Figure 5 shows genes with expression that varied bymore than 50% following I-BET151 treatment. Genes aregrouped according to whether they were pro (red) or

antiapoptotic (green), according to GO gene ontology (GOterms 0043066 or 0010942), or had been associated witheither role (blue bars). All four cell lines showed increasedgene expression of the proapoptotic proteins Noxa, Death-associated protein kinase, RhoB, and Foxo3, 6 hours afterI-BET151 treatment. PMAIP1 (Noxa) is well described as aBH3 protein induced by p53. STK17B (Death-associatedprotein kinase ) was reported previously as a hypermethylatedgene in melanoma (Hoon et al., 2004) and to be involved inUV-induced apoptosis (Kuwahara et al., 2006). RhoB isconsidered to have a tumor suppressive role and to bedownregulated by the RAS/Akt pathway (Jiang et al., 2004).The Foxo3 transcription factor upregulates BIM (Gillings et al.,2009). There were no major differences between thepretreatment line from patient 3 that underwent apoptosisand the resistant line in this patient, which may suggest thatthe induction of cell death is dependent on the balancebetween pro- and antiapoptotic gene expression in eachline. Full data are given in Supplementary Table 1f–g.

We used the LimmaGP tool in GenePattern to identify genesdifferentially expressed after I-BET151 treatment. Gene setenrichment analysis identified a variety of downregulatedgene sets, including those relating to ribosomes, hormonereceptor binding, mitosis, and NF-kB, whereas there was amixture of upregulated gene sets (Supplementary Table 1c).After 24 hours, the overwhelming number of downregulatedgene sets were related to mitosis and cell cycle(Supplementary Table 1d). Downregulated genes within themost significantly decreased gene set—Cell_Cycle_Process—included cyclin A2, CDK4, CDK2, and aurora kinase(Supplementary Table 1e).

β-Actin

p27

p21

p53

Pat

ient

-3-p

ost

Pat

ient

-1-p

ost

G1S

p21siRNA #1 #2 #1 #2

Patient-3-pre Patient-1-pre

0

Patient-1-pre

p21

β-Actin

I-BET151siRNA Cont p21#1 p21#2

p21

β-Actin

I-BET151siRNA Cont p21#1 p21#2

Patient-3-pre

β-Actin

c-myc

cyclin D1

p27

p21

p53

Me1

007

Mel

-JD

Mel

-RM

I-BET151 – – – – –+ + + +

I-BET151 – – – –+ + + +

+

I-BET151 ––

– – ––

– –+ + + + + +

– + – + – +

– + – + – +

SK

-Mel

-28

Mel

-RM

U

Pat

ient

-3-p

re

Pat

ient

-1-p

re

100

50

% P

opul

atio

n

G2/M

Figure 3. I-BET151 induces p21-dependent cell cycle arrest. (a) The level of cell cycle regulators in continuous melanoma cells treated with 10mM I-BET151

for 48 hours was measured by western blotting, and (b) in primary melanoma cell lines. (c) Cell cycle was measured in cells after p21 silencing and cells were

treated with 10mM I-BET151 for 48 hours. (d) The efficiency of the p21 knockdown was assessed by western blotting. Top panel images are from nonadjacent lanes

of the same blot.

SJ Gallagher et al.I-BET151 Induces Apoptosis and Arrest in Melanoma

www.jidonline.org 5

We also noted that many genes transcribing histone geneswere upregulated by I-BET151 treatment (SupplementaryTable 1). Histone gene transcription is normally tightlyregulated, and alterations in histone levels can perturb thecell cycle and lead to apoptosis. These results require valida-tion and further investigation.

DISCUSSIONThe above studies have identified a number of potentiallyimportant anticancer effects of I-BET151 against melanomacells that may be exploited in future preclinical studies. Allmelanoma lines tested responded to I-BET151 with an alteredcell cycle or survival profile. This involved induction of G1cell cycle arrest, mediated in part by the upregulation of p21as shown in western blots and by inhibition of G1 arrest whenp21 was downregulated by siRNA. A subset of melanomacells also underwent caspase-dependent apoptosis in responseto I-BET151, and this was associated with the loss of

mitochondrial outer membrane potential. Apoptosis wasassociated with upregulation of the BH3 proapoptotic BIMprotein in the majority of cell lines, and siRNA knockdown ofBIM confirmed it had a role in inducing apoptosis. A numberof other changes in pro- and antiapoptotic proteins weredetected in western blots, and further studies are needed togive a more complete understanding of the changes respon-sible for induction of apoptosis.

Successful development of treatments based on epigeneticmodifiers has largely been focused on hematologic malignan-cies and with relatively nonspecific pan-HDAC inhibitors suchas vorinostat and panobinostat. Our preclinical studies havealso shown strong synergy between HDAC inhibitors andselective BRAF inhibitors (Hersey et al., 2010; Lai et al.,2012b), and recent reports suggest the benefit of targetingepigenetic modulators in BRAFi-resistant melanoma(Johannessen et al., 2013). The BET protein inhibitors likeI-BET151 and JQ1 are more restricted in their action and

Rel

ativ

e ge

ne e

xpre

ssio

n%

Of c

ontr

ol le

vel

% O

f con

trol

leve

l

BRD2

BRD3

BRD4

0

100

200

300

Cell deathBIM mRNA

0

50

100

150p21 mRNA

Rel

ativ

e ge

ne e

xpre

ssio

n%

Of c

ontr

ol le

vel

% O

f con

trol

leve

l

BRD2

BRD3

BRD4

0

100

200

300BIM mRNACell death

0

50

100

150

200p21 mRNA

*

* *

*

**

** *

*

**

* *

* **

* *

*

* *

* *

1.5

1.0

0.5

0.0

1.5

1.0

0.5

0.0

G1 phase G1phase

Contro

l siR

NA

si-BRD2

#1

si-BRD2

#2

si-BRD3

#1

si-BRD3

#2

si-BRD4

#1

si-BRD4

#2

Contro

l siR

NA

si-BRD2

#1

si-BRD2

#2

si-BRD3

#1

si-BRD3

#2

si-BRD4

#1

si-BRD4

#2

Contro

l siR

NA

si-BRD2

#1

si-BRD2

#2

si-BRD3

#1

si-BRD3

#2

si-BRD4

#1

si-BRD4

#2

Contro

l siR

NA

si-BRD2

#1

si-BRD2

#2

si-BRD3

#1

si-BRD3

#2

si-BRD4

#1

si-BRD4

#2

Contro

l siR

NA

si-BRD2

#1

si-BRD2

#2

si-BRD3

#1

si-BRD3

#2

si-BRD4

#1

si-BRD4

#2

Contro

l siR

NA

si-BRD2

#1

si-BRD2

#2

si-BRD3

#1

si-BRD3

#2

si-BRD4

#1

si-BRD4

#2

Patient-3-pre

Patient-3-preMe1007

Me1007

Me1007 Patient-3-pre

Figure 4. Individual targeting of BET proteins causes cell cycle arrest and apoptosis. (a, b) BRD2, 3, and 4 were individually knocked down with siRNA in

Me1007 and patient-3-pre cells and their levels assessed by real-time RT-PCR after 72 hours. (c, d) Levels of BIM and cell death were assessed by real-time RT-PCR

and annexin-V staining, respectively. Levels are normalized to control siRNA-treated cells. (e, f) Levels of p21 and the percentage of cells in G1 phase were

assessed by real-time RT-PCR and cell cycle analysis, respectively. Levels are normalized to control cells. *Po0.05. BET, bromodomain and extraterminal proteins;

RT-PCR, reverse transcriptase–PCR; siRNA, small interfering RNA.

SJ Gallagher et al.I-BET151 Induces Apoptosis and Arrest in Melanoma

6 Journal of Investigative Dermatology (2014), Volume 00

competitively inhibit BET protein binding to acetylated histo-nes. They have shown selectivity for hematologic malignanciesdriven by c-Myc and the rare NUT midline carcinoma. Thepresent study adds to information about nonhematologiccancers that may be responsive to the BET protein inhibitors,such as N-Myc-dependent neuroblastoma (Puissant et al.,2013b), glioma (Cheng et al., 2013), lung carcinoma

(Lockwood et al., 2012), and melanoma (Segura et al.,2013b). In the case of lung carcinoma, sensitivity to the BETprotein inhibitor appeared to be based on downregulation ofthe oncogenic transcription factor FOSL1, which could bereproduced by knockdown of BRD4 (Lockwood et al., 2012).A previous study on human melanoma lines exposed to theBET protein inhibitor MS417 did, however, find downregula-

6-h apoptosis

PM

AIP

1S

TK

17B

RH

OB

FO

XO

3E

CT

2K

LF11

MA

P3K

5P

HLD

A3

HO

XA

5S

IRT

1T

HB

S1

PT

RH

2IL

6P

IM2

DH

RS

2N

OT

CH

1G

CLC

PIM

1A

SN

SH

SP

A1A

ZN

F16

NU

P62

BC

L2L2

PLK

1B

IRC

2LE

F1

BC

L2L1

BC

L3C

LCF

1W

NT

5AF

MN

2

–2

–1

0

1

2

Pat-1-pre 6 h

–2

–1

0

1

2

Pat-1-post 6 h

–2

–1

0

1

2

Pat-3-pre 6 h

–2

–1

0

1

2

Pat-3-post 6 h

24-h apoptosis

AN

KR

D1

RH

OB

BN

IP3

FA

M16

2AB

CL6

MLL

T11

BID

NE

T1

EC

T2

KLF

11A

ES

SIA

H1

MA

P3K

5T

OP

2AH

OX

A5

PP

P2R

4K

CN

MA

1T

FA

P2A

SIR

T1

IL6

DH

RS

2G

CLC

PIM

2T

HO

C6

WF

S1

SO

D2

CA

TH

SP

A1A

CC

L2A

UR

KB

SC

G2

TM

BIM

4V

EG

FB

BIR

C2

PR

KC

HF

MN

2F

HL2

HS

PA

5P

LK1

BIR

C5

NU

P62

CR

YA

BM

AD

2L1

ITG

A5

SN

AI2

SR

CB

CL3

VE

GF

AC

AV

1LE

F1

CLC

F1

WN

T5A

–2

–1

0

1

2

Pat-1-pre 24 h

–2

–1

0

1

2

Pat-1-post 24 h

–2

–1

0

1

2

Pat-3-pre 24 h

–2

–1

0

1

2

Pat-3-post 24 h

*

*

Figure 5. Changes in expression of apoptosis-related genes. Expression level of apoptosis-related genes was determined by microarray after 6 or 24 hour I-BET151

treatment. Genes that were significantly changed collectively and with a greater than 50% median fold change (log2 FC40.58) are presented. Red bars,

proapoptotic genes. Green, antiapoptotic genes. Blue, both pro- and antiapoptotic functions described. Bars are log2 fold change compared with a DMSO-treated

control at each time point. *Value exceeds axis maximum (see Supplementary Material online for actual value).

SJ Gallagher et al.I-BET151 Induces Apoptosis and Arrest in Melanoma

www.jidonline.org 7

tion of c-Myc as well as ERK and SKP2 (Segura et al., 2013b).In studies on gliomas, there was no association betweensensitivity to JQ1 and core signaling pathways such as Akt,p53, or Rb pathways resulting from mutations in key pathwayproteins. This also appeared to be the case in the present studyon melanoma, but further analysis is needed to confirm this.

One of the most consistent effects of I-BET 151 wasupregulation of the BH3 proapoptotic Bcl-2 family proteinBIM, even though a range of other BH3 proteins such as Noxaand Puma were unchanged or downregulated in the westernblot studies. The predominant isoforms of BIM (EL, L, and S)can bind all antiapoptotic members of the Bcl-2 family andhave been shown to be potent inducers of apoptosis inmelanoma (Plotz et al., 2013; Berger et al., 2014). In ourstudies, there did not appear to be differential expression ofthe different BIM isoforms, with all being changed to the sameextent. The increase in BIM was seen at the mRNA level,suggesting that upregulation was mediated transcriptionallyrather than by inhibition of degradation through theproteasome that is known to be important in the biology ofBIM (Gillings et al., 2009; Faber et al., 2012). BIM can betranscriptionally regulated by a number of transcription factorssuch as E2F1, FOXO3a, CHOP-C during endoplasmic stress,c-Myc, Rel-A, and c-Jun (reviewed in (Pinon et al. (2008)).Phosphorylation of BIM by JNK was reported to result in itsrelease from microtubules, but this would not be expected tochange the overall protein levels (Gillings et al., 2009). Asreported elsewhere (Gallagher et al. submitted), I-BET151 is astrong inhibitor of NF-kB in melanoma. It was also reportedthat NF-kB was able to suppress BIM in response to a numberof stimuli (Banerjee et al., 2008; Wang et al., 2008; Busuttilet al., 2010), but inhibition of NF-kB by BMS345541 (Yanget al., 2006) in our studies did not upregulate BIM and did notaffect the induction of BIM by I-BET151 (data not shown). Thissuggests that supression of NF-kB alone is not responsible forthe observed increase in BIM following I-BET151 treatment.Gene expression array studies identified strong upregulation ofFOXO3 and downregulation of E2F2 by I-BET151, andongoing studies suggest these may be the main transcriptionfactors involved in regulation of BIM. There were negliblechanges in CHOP-C/EBP and c-Jun in our cells.

Although BIM was increased in all the melanoma linestested, apoptosis was cell line dependent. This may reflectinhibition of BIM by oncogenic pathways such as the MAPK/ERK kinase/extracellular signal–regulated kinase and phos-phoinositide-3-kinase pathways, which phosphorylate andinhibit the activity of BIM-EL (Gillings et al., 2009).Similarly, variation in NF-kB activity may have contributedto variation in sensitivity to apoptosis by reduction in XIAP andBCL-XL levels, as reported elsewhere (Gallagher et al.submitted). Induction of p53-dependent BH3 proteins Pumaand Noxa did not appear to be involved, as their protein levelswere unchanged or decreased by treatment with I-BET151,even though mRNA transcripts of Noxa were stronglyupregulated at 6 hours in the gene expression arrays. Wepresume that this is a result of rapid degradation of Noxa asdescribed by others (Brinkmann et al., 2013). The relativelyslow kinetics of apoptosis induction (48 hours) are consistent

with a model in which apoptosis mediated by BET inhibitionrequires a number or cascade of transcriptional changes.These changes could include both upregulation of BIM andthe disabling of pro-survival signaling provided by NF-kB orother as-yet unidentified factors. The slow kinetics could alsosuggest that the cell must progress through certain stages of thecell cycle to undergo apoptosis; however, reduction of G1arrest by ablating p21 did not increase cell death, providinglittle support for this hypothesis.

Our analysis of the mechanisms underlying cell cycle arrestwas largely consistent with previous studies showing that thisappeared to result from upregulation of p21 (Cheng et al.,2013) and in some lines p27. In melanoma lines withupregulation of both p21 and p27, knockdown of p21reduced G1 arrest, suggesting it was the main cell cycleinhibitor. The basis for upregulation of p21 is, however, notclear. Although p53 was upregulated by I-BET151,knockdown of p53 did not alter G1 arrest (data not shown).This, together with the studies on the p53-null Me4405,suggests that p53 was not a key transcription factor in cellcycle arrest mediated by I-BET151. C-Myc involvement in thecell cycle arrest appeared unlikely, in that there was no orminimal downregulation of c-Myc by I-BET151 at the proteinor mRNA level.

Knockdown studies of individual BRD proteins suggestedthat BIM was regulated principally by BRD2. Inexplicably, theknockdown of BRD4 was associated with reduced levels ofBIM in both cell lines even though apoptosis was increased inboth lines. We interpret these findings as indicating that BRD2may normally suppress BIM, whereas BRD4 is part of com-plexes upregulating BIM. It is possible that, following theablation of BRD4, the suppression by BRD2 is unopposed,resulting in lower levels of BIM, whereas I-BET151 treatmentalso relieves BRD2 suppression and, possibly, activates BIMtranscription via other indirect methods. Further studies areneeded to see whether a possible feedback mechanism byBRD2 may be involved. As reported elsewhere, BRD2 alsoappeared to be the main protein involved in regulation of NF-kB (Gallagher et al. submitted). In contrast, BRD4 appeared tobe the major protein involved in cell cycle regulation inmelanoma as shown by the siRNA studies, in agreement withreports by others (Segura et al., 2013a).

These initial studies show that induction of apoptosis maybe an important additional mechanism that can be added toeffects of BET inhibitors on cell cycle and NF-kB in treatmentof cancer. Further studies are needed to define whichtranscription factors are involved in the upregulation of Bim.BRD2 and BRD4 appeared to have opposing effects onregulation of BIM, and the basis for this requires further study.It also raises questions as to whether selective inhibitorsagainst BRD2 may have merit. In addition, should the markedheterogeneity in apoptotic responses between the cell linestranslate to heterogeneity of clinical responses, identificationof biomarkers may be critical in patient selection. Interest-ingly, I-BET151 reduced the growth of Patient-1-post (vemur-afenib resistant) cells more strongly in vivo than in vitro. Thismight be explained by the previously mentioned effects ofI-BET151 on NF-kB activity and cytokine production, as

SJ Gallagher et al.I-BET151 Induces Apoptosis and Arrest in Melanoma

8 Journal of Investigative Dermatology (2014), Volume 00

autocrine signaling can be more important to support growthin vivo than in culture.

Our results highlight the potential of using BET inhibitors incombination with other drugs, with BET inhibition primingcells to undergo apoptosis via increased levels of BIM anddecreased BCL-2, BCL-XL, and XIAP. Taken together, thepresent results provide evidence that I-BET151 has severalpotential effects against melanoma cells that are independentof known mutational and signal pathway activation and justifyfurther investigation for their use in treatment for melanoma.

MATERIALS AND METHODSCell lines

Human melanoma cell lines Mel-RMU, Sk-Mel-28, Mel-RM, Mel-JD,

and Me1007 have been described previously (Zhang et al., 1999).

Cells were cultured in Dulbecco’s modified Eagle’s medium

containing 10% fetal calf serum (AusGeneX, Brisbane, Australia). In

addition, primary melanoma cell cultures were established, as

described previously (Nguyen et al., 2001), from two patients

entered into the Roche ‘‘BRIM2’’ phase II study of vemurafenib in

patients who had failed previous treatment. The patient lines were

established before treatment and during relapse from treatment with

vemurafenib, labeled ‘‘pre’’ and ‘‘post’’, respectively, as described

elsewhere (Lai et al., 2012a). These studies were approved by the

Hunter and New England Research Ethics Committee.

Chemicals and transfections

I-BET151 was supplied by GlaxoSmithKline (Brentford, UK). The

caspases were inhibited by the addition of 10mM Q-VD-OPH (SM

Biochemicals, Anaheim, CA) to medium 3 hours before I-BET151

treatment. For gene knockdown studies, siRNA was purchased from

QIAGEN (Limburg, The Netherlands) (non-silencing control

(1027281), BIM (SI02655359; SI04951968), p21 (SI00604898;

SI00299810), BRD2 (SI04256966, SI05015150), and BRD3

(SI03125150, SI04140178)) or from Shanghai Gene Pharma

(BRD4 (5’-GGAGAUGACAUAGUCUUAATT-3’, 5’-GCACAAUCAA

GUCUAAACUTT-3’)) and transfected with Lipofectamine RNAiMax

(Invitrogen) into cells 24 hours before I-BET151 treatment.

Quantitative reverse transcriptase–PCR

RNA was extracted from cell lines using the RNeasy Plus mini prep

kit (QIAGEN), quantitated using a Nanodrop (Thermo Scientific,

Wilmington, DE), and 1mg RNA reverse transcribed with Super-

ScriptIII (Invitrogen, Carlsbad, CA). cDNA was amplified on AB7900

(Applied Biosystems, Carlsbad, CA) using Universal PCR Master Mix

and Taqman probes (Applied Biosystems) specific for BIM

(Hs00708019_s1), p21 (Hs00355782_m1), BRD2 (Hs01121986_g1),

BRD3 (Hs00201284_m1), and BRD4 (Hs04188087_m1), and nor-

malized to levels of 18S (Hs99999901_s1).

Analysis of cell viability, cell death, cell cycle, and mitochondrialmembrane potentialCell growth assays were performed using CellTiter-Glo (Promega,

Madison, WI) as described by the manufacturer. Apoptotic cells were

quantified using Annexin-V/propidium iodide staining as described by

the manufacturer (Becton Dickinson, Franklin Lakes, NJ), and mea-

sured using a Becton Dickinson FACSCalibur flow cytometer. Cell

cycle was analyzed by measuring propidium iodide-stained cells

using a FACSCalibur flow cytometer and cell cycle fitted to viable

cells using ModFit LT software (Verity Software House). For BrdU

experiments, cells were pulsed with 10mM BrdU for 40 minutes and

processed using the BD Pharmingen BrdU Flow Kit (Cat No. 552598).

Changes in mitochondrial membrane potential (DCm) were studied

by staining the cells with the cationic dye JC-1, according to the

manufacturer’s instructions (Molecular Probes, Eugene, OR).

Western blotting

Western blot analysis was carried out as described previously (Irvine

et al., 2010). Antibodies used were Bak (G-23, Santa Cruz, Dallas,

TX); Bax (Cell Signaling, Danvers, MA); Bcl-2 (C-2, Santa Cruz); Bcl-

xL (H-5, Santa Cruz); Beta Actin (AC-74, Sigma); Bid (Cell Signaling);

Bik (C33-1, BD Biosciences); BIM (C34C5, Cell Signaling); BMF

(G81, Cell Signaling); cleaved caspase 3 (Cell Signaling); c-Myc

(9E10, Santa Cruz); cyclin D1 (G124-326 BD Biosciences); Mcl-1 (22/

Mcl-1, BD); Noxa (sc-26917, Santa Cruz); p21 (Millipore, Billerica,

MA); p27 (C-19, Santa Cruz); p53 (DO-1, Santa Cruz); Puma (Cell

Signaling); and XIAP (20/hILP/X, BD).

In vivo experiments

All animal experiments were performed in accordance with the

Australian Code of Practice for the Care and Use of Animals for

Scientific Purposes, 7th Edition, and with approval from the Peter

MacCallum Cancer Centre Animal Experimentation Ethics Commit-

tee. Female NOD/SCID mice were injected subcutaneously into the

flank with 4� 106 Patient-1-post cells in 50% Matrigel. Once tumors

had grown to approximately 130 mm2, mice were randomized into

two groups of nine mice each and each group was administered

either 25 mg kg� 1 I-BET151 or 1% methylcellulose vehicle daily by

oral gavage on days 1–7 and 20 mg kg� 1 on days 10–15.

Transcriptome analysis

Total RNA was extracted in duplicate from 75 cm2 culture flasks of

both ‘‘pre’’ and ‘‘post’’ cell lines from Patient 1 and Patient 3 at 6 and

24 hours after cells had been treated with 10mM I-BET151 or DMSO

control. The RNA was isolated using TRIZOL and purified with an

RNeasy purification kit (Qiagen) with DNAse I digestion on the

column, and the RNA quality was verified using the Agilent 2100

Bioanalyser (Agilent Technologies, Palo Alto, CA). Gene expression

analysis was performed using the HumanHT-12 v4 Expression

BeadChip (Illumina, San Diego, CA) and BeadStation system from

Illumina. Quality control was performed on all chips using Geno-

meStudio (Illumina). Data were log2 transformed and quantile normal-

ized following filtering out of unexpressed probe sets. Probe sets were

collapsed into single genes, and paired LimmaGP analysis was

performed on all cell lines, comparing DMSO- vs. I-BET151-treated

cells at either 6- or 24-hour treatment lengths. These results were used

to perform Gene Set Enrichment Analysis, which identifies whether

predetermined sets of genes tend to be upregulated or downregulated,

in GenePattern using the Broad Institute database MSigDB.

CONFLICT OF INTERESTRK Prinjha and N Smithers are employees and shareholders of GlaxoSmith-Kline, which is carrying out the clinical development of BET inhibitors.GA McArthur receives research support from Pfizer, Millennium, Novartis,and performs uncompensated consulting for GlaxoSmithKline. The remainingauthors state no conflict of interest.

SJ Gallagher et al.I-BET151 Induces Apoptosis and Arrest in Melanoma

www.jidonline.org 9

ACKNOWLEDGMENTSWe thank Warren Kaplan (Garvan Institute of Medical Research) for hisassistance in microarray analysis and Kerry Ardley for technical assistance inthe conduct of the in vivo study. This work was supported by program grant633004 from the Australian National Health and Medical Research Council(NHMRC) and the Cancer Institute NSW and the Melanoma Foundation of theUniversity of Sydney.

SUPPLEMENTARY MATERIAL

Supplementary material is linked to the online version of the paper at http://www.nature.com/jid

REFERENCES

Arrowsmith CH, Bountra C, Fish PV et al. (2012) Epigenetic protein families: anew frontier for drug discovery. Nat Rev Drug Discov 11:384–400

Banerjee A, Grumont R, Gugasyan R et al. (2008) NF-kappaB1 and c-Relcooperate to promote the survival of TLR4-activated B cells by neutraliz-ing Bim via distinct mechanisms. Blood 112:5063–73

Becker TM, Haferkamp S, Dijkstra MK et al. (2009) The chromatin remodellingfactor BRG1 is a novel binding partner of the tumor suppressor p16INK4a.Mol Cancer 8:4

Belkina AC, Denis GV (2012) BET domain co-regulators in obesity, inflamma-tion and cancer. Nat Rev Cancer 12:465–77

Berger A, Quast SA, Plotz M et al. (2014) RAF inhibition overcomes resistanceto TRAIL-induced apoptosis in melanoma cells. J Invest Dermatol134:430–40

Blair LP, Yan Q (2012) Epigenetic mechanisms in commonly occurringcancers. DNA Cell Biol 31(Suppl 1):S-49–61

Brinkmann K, Zigrino P, Witt A et al. (2013) Ubiquitin C-terminal hydrolase-L1potentiates cancer chemosensitivity by stabilizing NOXA. Cell Rep 3:881–91

Busuttil V, Droin N, McCormick L et al. (2010) NF-kappaB inhibits T-cellactivation-induced, p73-dependent cell death by induction of MDM2.Proc Natl Acad Sci USA 107:18061–6

Cheng Z, Gong Y, Ma Y et al. (2013) Inhibition of BET Bromodomain TargetsGenetically Diverse Glioblastoma. Clin Cancer Res 12:12

Dawson MA, Kouzarides T (2012) Cancer epigenetics: from mechanism totherapy. Cell 150:12–27

Dawson MA, Kouzarides T, Huntly BJP (2012) Targeting epigenetic readers incancer. N Engl J Med 367:647–57

Dawson MA, Prinjha RK, Dittmann A et al. (2011) Inhibition of BETrecruitment to chromatin as an effective treatment for MLL-fusionleukaemia. Nature 478:529–33

Delmore JE, Issa GC, Lemieux ME et al. (2011) BET bromodomain inhibition asa therapeutic strategy to target c-Myc. Cell 146:904–17

Ernst J, Kellis M (2012) ChromHMM: automating chromatin-state discoveryand characterization. Nat Methods 9:215–6

Faber AC, Ebi H, Costa C et al. (2012) Apoptosis in targeted therapy responses:the role of BIM. Adv Pharmacol 65:519–42

Filippakopoulos P, Picaud S, Fedorov O et al. (2012) Benzodiazepines andbenzotriazepines as protein interaction inhibitors targeting bromodomainsof the BET family. Bioorg Med Chem 20:1878–86

Filippakopoulos P, Qi J, Picaud S et al. (2010) Selective inhibition of BETbromodomains. Nature 468:1067–73

Fullgrabe J, Kavanagh E, Joseph B (2011) Histone onco-modifications.Oncogene 30:3391–403

Gillespie S, Borrow J, Zhang XD et al. (2006) Bim has a crucial role insynergistic induction of apoptosis by the histone deacetylase inhibitorSBHA and TRAIL in melanoma cells. Apoptosis 11:2251–65

Gillings AS, Balmanno K, Wiggins CM et al. (2009) Apoptosis and autophagy:BIM as a mediator of tumour cell death in response to oncogene-targetedtherapeutics. FEBS J 276:6050–62

Hersey P, Zhang X, Jiang C (2010) Induction of apoptosis in human melanomaby the BRAF inhibitor PLX4720: the key to therapeutic success? J ClinOncol 28, Abstract No. 8559

Hersey P, Zhang XD (2001) How melanoma cells evade trail-inducedapoptosis. Nat Rev Cancer 1:142–50

Hodis E, Watson IR, Kryukov GV et al. (2012) A landscape of driver mutationsin melanoma. Cell 150:251–63

Hoon DS, Spugnardi M, Kuo C et al. (2004) Profiling epigenetic inactivation oftumor suppressor genes in tumors and plasma from cutaneous melanomapatients. Oncogene 23:4014–22

Irvine M, Philipsz S, Frausto M et al. (2010) Amino terminal hydrophobicimport signals target the p14(ARF) tumor suppressor to the mitochondria.Cell Cycle 9:829–39

Jang S, Atkins MB (2013) Treatment of BRAF mutant melanoma: the role ofvemurafenib and other therapies. Clin Pharmacol Ther 95:24–31

Jiang K, Sun J, Cheng J et al. (2004) Akt mediates Ras downregulation of RhoB,a suppressor of transformation, invasion, and metastasis. Mol Cell Biol24:5565–76

Johannessen CM, Johnson LA, Piccioni F et al. (2013) A melanocyte lineageprogram confers resistance to MAP kinase pathway inhibition. Nature504:138–42

Keenen B, Qi H, Saladi SV et al. (2010) Heterogeneous SWI/SNFchromatin remodeling complexes promote expression of microphthal-mia-associated transcription factor target genes in melanoma. Oncogene29:81–92

Kuwahara H, Nakamura N, Kanazawa H (2006) Nuclear localization of theserine/threonine kinase DRAK2 is involved in UV-induced apoptosis. BiolPharm Bull 29:225–33

Kuzelova K, Grebenova D, Brodska B (2011) Dose-dependent effects of thecaspase inhibitor Q-VD-OPh on different apoptosis-related processes.J Cell Biochem 112:3334–42

Lai F, Jiang CC, Farrelly ML et al. (2012a) Evidence for upregulation of Bim andthe splicing factor SRp55 in melanoma cells from patients treated withselective BRAF inhibitors. Melanoma Res 22:244–51

Lai F, Jin L, Gallagher S et al. (2012b) Histone deacetylases (HDACs) asmediators of resistance to apoptosis in melanoma and as targets forcombination therapy with selective BRAF inhibitors. Adv Pharmacol65:27–43

Lockwood WW, Zejnullahu K, Bradner JE et al. (2012) Sensitivity of humanlung adenocarcinoma cell lines to targeted inhibition of BET epigeneticsignaling proteins. Proc Natl Acad Sci USA 109:19408–13

Mertz JA, Conery AR, Bryant BM et al. (2011) Targeting MYC dependence incancer by inhibiting BET bromodomains. Proc Natl Acad Sci USA108:16669–74

Nguyen T, Zhang XD, Hersey P (2001) Relative resistance of fresh isolates ofmelanoma to tumor necrosis factor-related apoptosis-inducing ligand(TRAIL)-induced apoptosis. Clin Cancer Res 7:966s–73s

Nicodeme E, Jeffrey KL, Schaefer U et al. (2010) Suppression of inflammationby a synthetic histone mimic. Nature 468:1119–23

Pinon JD, Labi V, Egle A et al. (2008) Bim and Bmf in tissue homeostasis andmalignant disease. Oncogene 27:42

Plotz M, Gillissen B, Quast SA et al. (2013) The BH3-only protein Bim(L)overrides Bcl-2-mediated apoptosis resistance in melanoma cells. CancerLett 335:100–8

Prinjha RK, Witherington J, Lee K (2012) Place your BETs: the therapeuticpotential of bromodomains. Trends Pharmacol Sci 33:146–53

Puissant A, Frumm SM, Alexe G et al. (2013a) Targeting MYCN in neuro-blastoma by BET bromodomain inhibition. Cancer Discov 3:308–23

Puissant A, Frumm SM, Alexe G et al. (2013b) Targeting MYCN in Neuro-blastoma by BET Bromodomain Inhibition. Cancer Discov 21:21

Ram O, Goren A, Amit I et al. (2011) Combinatorial patterning of chromatinregulators uncovered by genome-wide location analysis in human cells.Cell 147:1628–39

Riley JL (2013) Combination checkpoint blockade—taking melanoma immu-notherapy to the next level. N Engl J Med 369:187–9

Robert C, Soria JC, Eggermont AM (2013) Drug of the year: programmed death-1 receptor/programmed death-1 ligand-1 receptor monoclonal antibodies.Eur J Cancer 49:2968–71

SJ Gallagher et al.I-BET151 Induces Apoptosis and Arrest in Melanoma

10 Journal of Investigative Dermatology (2014), Volume 00

Saladi SV, Keenen B, Marathe HG et al. (2010) Modulation of extracellularmatrix/adhesion molecule expression by BRG1 is associated withincreased melanoma invasiveness. Mol Cancer 9:280

Segura MF, Fontanals-Cirera B, Gaziel-Sovran A et al. (2013a) BRD4 sustainsmelanoma proliferation and represents a new target for epigenetictherapy. Cancer Res 73:6264–76

Segura MF, Fontanals-Cirera B, Gaziel-Sovran A et al. (2013b) BRD4 sustainsproliferation and represents a new target for epigenetic therapy inmelanoma. Cancer Res 15:15

Sullivan RJ, Lorusso PM, Flaherty KT (2013) The intersection of immune-directed and molecularly targeted therapy in advanced melanoma:where we have been, are, and will be. Clinical Cancer Res 19:5283–91

van den Hurk K, Niessen HEC, Veeck J et al. (2012) Genetics and epigeneticsof cutaneous malignant melanoma: a concert out of tune. BiochimBiophys Acta 1826:89–102

Wang Z, Zhang B, Yang L et al. (2008) Constitutive production of NF-kappaB2p52 is not tumorigenic however predisposes mice to inflammatoryautoimmune disease by repressing Bim expression. J Biol Chem283:10698–706

Yang J, Amiri KI, Burke JR et al. (2006) BMS-345541 targets inhibitor of kappaBkinase and induces apoptosis in melanoma: involvement of nuclear factorkappaB and mitochondria pathways. Clin Cancer Res 12:950–60

Zhang XD, Franco A, Myers K et al. (1999) Relation of TNF-related apoptosis-inducing ligand (TRAIL) receptor and FLICE-inhibitory protein expressionto TRAIL-induced apoptosis of melanoma. Cancer Res 59:2747–53

Zhang XD, Gillespie SK, Borrow JM et al. (2004) The histone deacetylaseinhibitor suberic bishydroxamate regulates the expression of multipleapoptotic mediators and induces mitochondria-dependent apoptosis ofmelanoma cells. Mol Cancer Ther 3:425–35

Zuber J, Shi J, Wang E et al. (2011) RNAi screen identifies Brd4 as a therapeutictarget in acute myeloid leukaemia. Nature 478:524–8

SJ Gallagher et al.I-BET151 Induces Apoptosis and Arrest in Melanoma

www.jidonline.org 11