Báo cáo khoa học: Apoptosis and autophagy: BIM as a mediator of tumour cell death in response to oncogene-targeted therapeutics pptx

Abbreviations AML, acute myeloid leukaemia; b-TrCP, b-transducin repeat containing protein; BAD, BCL-xL⁄ BCL-2-associated death promoter; BAK, BCL-2 homologous antagonist ⁄ killer; BAX,

Apoptosis and autophagy: BIM as a mediator of tumour cell death in response to oncogene-targeted therapeutics Annette S Gillings, Kathryn Balmanno, Ceri M Wiggins, Mark Johnson and Simon J Cook

Laboratory of Molecular Signalling, The Babraham Institute, Babraham Research Campus, Cambridge, UK

Introduction

The conserved, ‘cell intrinsic’ or ‘mitochondrial’

apop-tosis pathway is controlled by the interplay between

three groups of B-cell lymphoma 2 (BCL-2) proteins [1,2] The multidomain, pro-apoptotic proteins BCL-2

Keywords

B-cell lymphoma 2 (BCL-2); breakpoint

cluster region ⁄ Abelson murine leukaemia

viral oncogene (BCR ⁄ ABL);

BCL-2-interacting mediator of cell death (BIM);

v-raf murine sarcoma viral oncogene

homologue B1 (BRAF); epidermal growth

factor receptor (EGFR); extracellular

signal-regulated kinase 1 ⁄ 2 (ERK1 ⁄ 2);

mitogen-MAPK or ERK Kinase 1 ⁄ 2 (MEK1 ⁄ 2); protein

kinase B (PKB); ribosomal protein S6 kinase

(RSK)

Correspondence

Simon J Cook, Laboratory of Molecular

Signalling, The Babraham Institute,

Babraham Research Campus, Cambridge

CB22 3AT, UK

Fax: 44-1223-496023

Tel: 44-1223-496453

E-mail: simon.cook@bbsrc.ac.uk

(Received 16 March 2009, revised 23 June

2009, accepted 9 July 2009)

doi:10.1111/j.1742-4658.2009.07329.x

The BCL-2 homology domain 3 (BH3)-only protein, B-cell lymphoma

2 interacting mediator of cell death (BIM) is a potent pro-apoptotic protein belonging to the B-cell lymphoma 2 protein family In recent years, advances in basic biology have provided a clearer picture of how BIM kills cells and how BIM expression and activity are repressed by growth factor signalling pathways, especially the extracellular signal-regulated kinase 1⁄ 2 and protein kinase B pathways In tumour cells these oncogene-regulated pathways are used to counter the effects of BIM, thereby promoting tumour cell survival In parallel, a new generation of targeted therapeutics has been developed, which show remarkable specificity and efficacy in tumour cells that are addicted to particular oncogenes It is now apparent that the expression and activation of BIM is a common response to these new therapeutics Indeed, BIM has emerged from this marriage of basic and applied biology as an important mediator of tumour cell death in response to such drugs The induction of BIM alone may not be sufficient for significant tumour cell death, as BIM is more likely to act in concert with other BH3-only proteins, or other death pathways, when new targeted therapeutics are used in combination with traditional chemotherapy agents Here we discuss recent advances in understanding BIM regulation and review the role of BIM as a mediator of tumour cell death in response to novel oncogene-targeted therapeutics

Abbreviations

AML, acute myeloid leukaemia; b-TrCP, b-transducin repeat containing protein; BAD, BCL-xL⁄ BCL-2-associated death promoter; BAK, BCL-2 homologous antagonist ⁄ killer; BAX, BCL-2-associated x protein; BCL-2, B-cell lymphoma 2; BCL-x L, B-cell lymphoma-extra large;

BCR ⁄ ABL, breakpoint cluster region ⁄ Abelson murine leukaemia viral oncogene; BH3, BCL-2 homology domain 3; BIM, BCL-2-interacting mediator of cell death; BOP, BH3-only protein; BRAF, v-raf murine sarcoma viral oncogene homologue B1; CBL, Casitas B-lineage

lymphoma oncogene; CML, chronic myelogenous leukaemia; CUL2, Cullin 2; DLC1, dynein light chain 1; EGFR, epidermal growth factor receptor; ERK, extracellular signal-regulated kinase; FLT3, FMS-like tyrosine kinase 3; FOXO3A, Forkhead box 3A; KIT, oncogene of HZ4 feline sarcoma virus; MCL, myeloid cell leukaemia 1; MEK, MAPK or ERK Kinase; mTOR, mammalian target of rapamycin; NSCLC, non-small cell lung cancer; PDGFR, platelet-derived growth factor receptor; PI3K, phosphoinositide 3¢-kinase; PKB, protein kinase B (also known

as Akt); PUMA, p53-upregulated modulator of apoptosis; RACK1, receptor for activated C-kinase-1; RAS, rat sarcoma virus concogene; RNAi, RNA interference; RSK, ribosomal protein S6 kinase.

associated x protein (BAX) and bcl-2 homologous

antagonist⁄ killer (BAK) can activate

caspase-dependent cell death by promoting the release of

cytochrome c from the mitochondria; however, in

viable cells, BAX and BAK are restrained by their

interaction with the prosurvival proteins such as

BCL-2, B-cell lymphoma-extra large (BCL-xL) or

mye-loid cell leukaemia 1 (MCL-1) The third group of

BCL-2 proteins, the BCL-2 homology domain 3

(BH3)-only proteins (BOPs), includes

BCL-2-interact-ing mediator of cell death (BIM), p53-upregulated

modulator of apoptosis (PUMA), NOXA (‘damage’),

BCL-2 modifying factor (BMF) and BCL-xL⁄

BCL-2-associated death promoter (BAD); they are activated

(that is, expressed de novo, post-translationally

modi-fied and⁄ or stabilized) in response to various

pro-apoptotic stimuli (including loss of survival signals)

and promote cell death in a manner dependent on the

presence of BAX and BAK The precise mechanism by

which this is achieved remains controversial, but a

wealth of data now favours a model in which the

BOPs bind to the prosurvival BCL-2 proteins,

seques-tering them and allowing BAX and⁄ or BAK to

pro-mote cell death [3] One observation that favours this

model is that the relative toxicity of different BOPs

segregates well with the repertoire of prosurvival

BCL-2 proteins to which they can bind [4] For example,

BIM and PUMA can bind to all BCL-2 proteins with

high affinity and are potent killers, whereas NOXA

only binds to MCL-1 and BCL2-related protein A1

and is less toxic (except presumably in cells in which

MCL-1 and A1 are the predominant BCL-2 proteins?)

BIM has a number of properties that set it apart

from most other BOPs In addition to its strong

toxic-ity [4], alternative splicing [5,6] gives rise to a variety

of BIM isoforms with different intrinsic toxicities and

modes of regulation [7] Some BIM splice variants

exhibit no apparent toxicity; are these naturally

occur-ring dominant negatives or do they point to additional

functions for BIM that are unrelated to the promotion

of cell death? The most extensively studied splice

vari-ants, BIMS, BIML and BIMEL (BIM-short, BIM-long

and BIM-extra long, respectively), are all cytotoxic

and subject to different modes of regulation by various

prodeath and prosurvival signalling pathways [7]

Some, such as BIML and BIMEL, are phosphorylated

by c-Jun N-terminal kinase in response to various

stresses, and this promotes apoptosis [8,9] In addition,

particular attention has focussed recently on the

regu-lation of BIM by the prosurvival extracellular

signal-regulated kinase 1⁄ 2 (ERK1 ⁄ 2) and protein kinase B

(PKB) pathways that act downstream of oncogenic

protein kinases [10,11] It is increasingly apparent that

these pathways are utilized by oncogenes to inhibit or neutralize BIM, thereby facilitating tumour cell survival Arising directly from this is the growing appreciation that the new generation of oncogene-targeted therapeutics cause loss of ERK1⁄ 2 and ⁄ or PKB signalling and, as a consequence, promote increased expression of BIM and BIM-dependent cell death in tumour cells Here we review recent advances

in understanding BIM regulation and analyze the results of studies which suggest that BIM is an impor-tant mediator of tumour cell death in response to novel oncogene-targeted therapeutics

Regulation of BIM by cell survival signalling pathways

Transcriptional regulation of BIM Transcription of the BIM gene is normally repressed

by serum, growth factors and cytokines, and increases upon the withdrawal of such survival factors; indeed, expression of BIM is required for optimal cell death following cytokine withdrawal [12–14] Transcription

of BIM is promoted by the Forkhead box 3A (FOXO3A) transcription factor [15], which itself is inhibited by both the ERK1⁄ 2 and PKB pathways PKB phosphorylates FOXO3A directly at three serine residues and this allows binding to 14-3-3 proteins, thereby sequestering FOXO3A in the cytosol and pre-venting it from activating BIM transcription [16] In addition, direct ERK1⁄ 2-dependent phosphorylation targets FOXO3A for proteasome-dependent degrada-tion [17] These studies provide relatively simple expla-nations for the fact that inhibition of either the ERK1⁄ 2 or PKB pathways is sufficient to increase BIMmRNA in many cell types

Post-translational regulation of BIM by phosphorylation

The extra long splice variant, BIMEL, is the most abundant isoform and undergoes the most dynamic changes in expression upon withdrawal of survival factors [14] In addition, expression of BIMEL often precedes that of BIMS or of BIML [14,18], suggesting that BIMEL is subject to some unique mode of regula-tion Indeed, many studies have now shown that BIMEL is phosphorylated at multiple sites in response

to activation of the ERK1⁄ 2 pathway, and this has the effect of promoting its ubiquitination and proteasome-dependent degradation [7,19–21] BIMELis phosphory-lated on at least three Ser-Pro motifs, including Ser69 (Ser65 in mouse and rat) (Fig 1) The first effect of

this phosphorylation appears to be to promote the

dis-sociation of BIMEL from prosurvival BCL-2 proteins

[14] (Fig 1); because BIM promotes cell death by

binding to prosurvival BCL-2 proteins, this alteration

in the binding properties of BIMELserves as a

cell-sur-vival mechanism In addition, this may constitute part

of the signal for BIMELdegradation because a BIMEL

mutant that fails to bind to BCL-2 proteins is

degraded more rapidly in cells [14,22]

The nature of the E3 ubiquitin ligase responsible for

poly-ubiquitination of BIMEL remains a matter of

some debate The really interesting new gene (RING)

finger protein, Casitas B-lineage lymphoma oncogene

(CBL), was originally proposed as the relevant E3 [23];

however, this suggestion was rather controversial

because substrates of CBL are almost invariably

phosphotyrosine-containing proteins and to date the

only pathways suggested to play a role in BIMEL

degradation result in its serine phosphorylation

Subse-quently, other studies have failed to demonstrate any

role for CBL by showing that BIMELis not

phosphor-ylated on tyrosine, CBL and BIM fail to interact, and

ERK1⁄ 2-driven BIMEL turnover proceeds normally in

cells lacking CBL [24,25] These studies reveal that

CBL is not the E3-ubiquitin ligase responsible for

tar-geting BIMEL for degradation in fibroblasts and

epi-thelial cells, and that any role it may play in other cell

types is likely to be an indirect one In a separate

study, receptor for activated C-kinase-1 (RACK1) and

cytokine-inducible SH2 protein (CIS) were reported to

be members of an ElonginB⁄ C-Cullin-SOCS-Box

(ECS)–regulator of cullins (Roc) complex responsible

for the degradation of BIMELin response to treatment

with paclitaxel [26] Initially, dynein light chain 1 (DLC1, a known BIM-binding protein) was found to bind to RACK1 in a yeast two-hybrid screen Further overexpression studies suggested a large E3 ligase complex involving RACK1 in complex with DLC1, BIMEL, CIS and Cullin 2 (CUL2), with assembly of some components being enhanced by paclitaxel RNA interference (RNAi)-mediated knockdown of RACK1

or DLC1 resulted in BIMEL accumulation [26] How-ever, a recent study failed to reproduce the CUL2– BIMEL interaction but rather demonstrated co-immu-noprecipitation of BIMEL with CUL1, leading to the proposal that BIMEL degradation occurs via a classic Skp-Cullin-F-box (SCF) E3 ligase [27]

The search for the relevant E3 ligase now appears to have been resolved with the report that ribosomal pro-tein S6 kinase (RSK), activated downstream of ERK1⁄ 2, phosphorylates BIMEL, providing a binding site for the F-box proteins beta-transducin repeat con-taining protein (bTrCP)1 and bTrCP2, which promote the poly-ubiquitination of BIMEL[27] It is known that ERK1⁄ 2 can phosphorylate BIMEL at Ser55, Ser69 and Ser73 within cells, and Ser69 seems to contribute

to BIMEL turnover [20,28] The new study proposes that ERK1⁄ 2-dependent phosphorylation of BIMEL at Ser69 facilitates optimal phosphorylation by RSK at Ser93, Ser94 and Ser98, and this motif then serves as the binding site for bTrCP1⁄ 2 [27] (Fig 1) This attrac-tive model may explain why mutation of a single ERK1⁄ 2 phosphorylation site, Ser69, causes loss of at least two further phosphorylation sites in cells [28] However, it also reveals that one of the important RSK phosphorylation sites, Ser98, lies within the

UPS 26S

BIM

RAS

RAF

MEK

ERK

RSK

TrCP1/2

BIM EL BIM EL

P P

MCL-1

BIM EL

P

P P P P Ub Ub

BIM EL

BIM EL

P

P P P P

P

P P P P

Fig 1 Regulation of BIM-binding properties and stability by ERK1 ⁄ 2 The pro-apoptotic BH3-only protein BIM is expressed de novo following cytokine withdrawal and binds

to prosurvival proteins, such as MCL-1, thereby releasing BAX or BAK to promote cell death BIM EL , the most abundant BIM splice variant, is phosphorylated directly by ERK1 ⁄ 2 on up to three different sites This promotes the dissociation of BIM EL from prosurvival proteins [14,22] ERK1 ⁄ 2-cataly-sed phosphorylation may also ‘prime’ BIMEL for phosphorylation by RSK1 or RSK2, providing a binding site for the bTrCP E3 ubiquitin ligase [27]; bTrCP promotes the poly-ubiquitination of BIMEL, thereby targeting it for destruction by the 26S proteasome See the text for details.

previously mapped ERK1⁄ 2 docking domain [29]

Pre-sumably this must mean that the binding of ERK1⁄ 2,

RSK and bTrCP1⁄ 2 is subject to fine temporal

coordi-nation within the cell Does ERK1⁄ 2 dissociate rapidly

after phosphorylating BIMEL to allow binding of

RSK, which in turn dissociates to allow binding of

bTrCP1⁄ 2? Are these events coordinated by a scaffold

protein that brings the components together at the

outer surface of the mitochondria? No doubt these

details will emerge in the future

Taken together, a wealth of literature now clearly

indicates that activation of the PKB or ERK1⁄ 2

path-ways can repress BIM transcription, whilst activation

of ERK1⁄ 2 can selectively target the major BIM splice

variant, BIMEL, by reducing its binding to prosurvival

BCL-2 proteins and promoting its

proteasome-depen-dent destruction (Fig 1) As the ERK1⁄ 2 and PKB

pathways are two of the major cell-survival signalling

pathways [10,21], it follows that these pathways play a

major role in growth factor-dependent cell-survival

sig-nalling, including the repression or inhibition of BIM

Arising from this is a growing understanding of (a) the

role that these pathways play in repressing BIM and in

promoting aberrant cell survival in tumours that

harbour mutations in oncogenes which control these

pathways and (b) the role of BIM in tumour cell death

arising from targeted inhibition of such oncogenes or

pathways

Oncogene addiction and the tumour

cell response to novel targeted

therapies

Research into the molecular basis of cancer in the last

25 years has identified hundreds of genes that can

cause malignant transformation when overexpressed or

mutated, and it is known that tumours typically

accu-mulate dozens of mutations during their lifetime

How-ever, some of these mutations (so-called ‘drivers’) are

more important than others (‘passengers’) [30] Driver

mutations promote the initiation, development and

maintenance of the tumour, whereas passenger

muta-tions confer no selective advantage and probably arise

through genomic instability Tumour cells exhibit a

series of hallmarks that set them apart from normal

cells [31] and driver mutations are thought to promote

the acquisition and underpin the maintenance of these

tumour-specific traits It seems that tumours evolve to

be dependent upon certain key driver mutations and

on the signalling pathways they control, to maintain

their malignant phenotype – a concept known as

‘oncogene addiction’ [32] This evolved dependency

upon particular oncogenes often reflects a loss of

signal pathway redundancy, providing a therapeutic window for tumour-selective intervention The new, targeted therapeutics take advantage of this window

by targeting the specific driver oncoproteins, or their downstream effector pathways, to which tumours are addicted Because tumour cells typically evolve to be dependent upon their driver oncoproteins for survival signals, tumour cell death is a common and clinically desirable response to these new, targeted therapeutics Recent studies have shown that in certain tumour types pharmacological inhibition of these driver onco-proteins results in inactivation of the ERK1⁄ 2 and PKB pathways, increased expression of BIM and cell death These agents do not target BIM directly; expres-sion or activation of BIM occurs indirectly, resulting from the inactivation of signalling pathways that nor-mally repress BIM It is increasingly clear that whilst drug-induced expression of BIM alone may not be suf-ficient to kill these tumour cells, death is at least partly BIM-dependent, with the degree of BIM involvement reflecting the role of other BOPs or other pathways in different cell types Here we review the most pertinent recent examples

Tumours with BRAF mutations

The v-raf-1 murine leukemia viral oncogene (RAF)–MAPK or ERK Kinase 1⁄ 2 (MEK1 ⁄ 2)– ERK1⁄ 2 signalling cascade has received much atten-tion in terms of drug discovery because of its role in promoting cell proliferation and survival [10] and as a result of the high frequency of rat sarcoma virus con-cogene (RAS) [33] and v-raf murine sarcoma viral oncogene homolog B1 (BRAF) [34] mutations identi-fied in certain human cancers Within the ERK1⁄ 2 pathway, MEK1⁄ 2 are attractive targets for pharmaco-logical intervention because of (a) their strict selectivity for their downstream targets, ERK1 and ERK2, and (b) the presence of a unique hydrophobic inhibitor binding pocket adjacent to the Mg⁄ ATP-binding site that exhibits little homology to other kinases and explains the high degree of specificity that has been observed with the MEK1⁄ 2 inhibitors reported to date [35] The first-generation pan-MEK inhibitors PD98059 and U0126 can also inhibit MEK5 [10] and exhibit poor potency and pharmacokinetic properties PD184352 (CI-1040) is selective for the MEK1⁄ 2-ERK1⁄ 2 pathway and was the first MEK1 ⁄ 2 inhibi-tor to demonstrate oral anticancer activity in a preclin-ical model [36]; however, despite encouraging results in phase I clinical trials [37] it showed inadequate clinical activity to justify development [38] PD0325901 [39] and AZD6244 (ARRY-142886) [40] are both selective

for the MEK1⁄ 2–ERK1 ⁄ 2 pathway and show similar

oral activity, with AZD6244 undergoing clinical

evalu-ation at the time of writing AZD6244 can cause a G1

cell cycle arrest and in some cases apoptosis; mouse

xenograft studies have revealed both tumour stasis,

associated with reduced tumour proliferation, and

tumour regression, accompanied by apoptosis [40] An

understanding of how and under what circumstances

MEK inhibition can promote apoptosis may permit a

more targeted clinical use of AZD6244 and related

molecules

Several studies have recently implicated BIM as a

tumour cell executioner in response to inhibitors of the

BRAF–MEK–ERK signalling pathway (Fig 2)

Colorectal cancer cell lines harbouring a BRAF600E

mutation are relatively resistant to death arising from

serum starvation and fail to upregulate BIM; however,

this is readily overcome by treatment with AZD6244,

indicating that these cells are addicted to the ERK1⁄ 2

pathway for repression of BIM and growth

factor-independent survival RNAi-mediated knockdown

revealed a major role for BIM in AZD6244-induced

cell death [41] In a separate study, treatment of

melanoma cells harbouring BRAF600E with PD184352

or the BRAF600E-selective inhibitor PLX4720 also syn-ergized with growth factor withdrawal to increase BIM expression, and cell death was partially dependent upon BIM [42] Similar results in both colorectal can-cer and melanoma cell lines have also been reported [43], and in all cases BIMEL was the predominant iso-form expressed upon MEK inhibition Indeed, ERK1⁄ 2-dependent turnover of BIMEL via the protea-some was primarily responsible for the repression of BIMEL in all three studies [41–43] (Fig 2) BIM, alone

or in combination with other BOPs, has been impli-cated in melanoma cell death, arising from MEK1⁄ 2 inhibition, in several other studies [44–46]

It is interesting to note that in all cases MEK inhibi-tor-induced BIM expression alone was not sufficient to induce a dramatic increase in cell death; pronounced increases in apoptosis were only observed when MEK inhibition was combined with serum deprivation [41,42] This suggests that loss of other serum-dependent sur-vival pathways, such as the phosphoinositide 3¢-kinase (PI3K)–PKB pathway, may be required to cooperate with MEK1⁄ 2 inhibition for optimal tumour cell death,

FOXO3A

PI3K

PDK

PKB RAS

BIM

Mut

BIMEL

BRAF

MEK1/2

ERK1/2

BAX

AZD6244 PD184352 PD0325901

cyt-c

CASP

BAX

BCR-ABL

EGFR

BCL-2

Gefitinib / Erlotinib

Imatinib Dasatinib Nilotinib

ABT-737 Mut

Mut

PLX4720 Mut

Fig 2 BIM-dependent tumour cell death – a common response to oncogene-targeted therapeutics Tumours with mutated oncogenic kinases, such as BCR–ABL (in CML), EGFR (in NSCLC) or BRAF (in melanoma and colorectal cancer), typically evolve to be addicted to these oncoproteins for cell survival Selective oncogene-targeted therapeutics, such as imatinib (BCR–ABL), erlotinib (EGFR) or PLX4720 (BRAF), inhibit these kinases or, in the case of MEK inhibitors (such as AZD6244), inhibit one of the key signalling pathways they control; in all cases inhibition of the oncogenic kinase results in the loss of downstream survival signalling pathways and a consequent increase in the expres-sion of BIM Inactivation of the ERK1 ⁄ 2 pathway seems to be particularly important for upregulation of the most abundant BIM isoform, BIMEL Dephosphorylated BIMELthen binds to prosurvival BCL-2 proteins, such as MCL-1, to release BAX and ⁄ or BAK to promote cell death Tumour cell death arising from such drug treatments requires BIM to varying degrees; in many cases BIM acts in concert with other BOPs, such as BAD Despite this, cell death in such circumstances can be quite modest as a result of buffering by high levels of prosurvival proteins, such as BCL-2 and BCL-xLfound in some tumours The co-administration of BH3 mimetics, such as ABT-737, inhibits BCL-2 and BCL-x L and synergizes effectively with oncogene-targeted therapeutics that mobilize BIM to promote tumour cell killing and tumour regres-sion See the text for details CASP, caspase-9; cyt-c, cytochrome c.

providing a rationale for the use of combinations of

MEK1⁄ 2 inhibitors and PI3K–PKB pathway inhibitors

Indeed, rapamycin, an inhibitor of mammalian target of

rapamycin (mTOR) downstream of PKB, can synergize

with the MEK1⁄ 2 inhibitor, PD0325901, to promote

regression of established melanomas in a mouse model

in which melanoma is driven by Braf600Eand

phospha-tase and tensin homologue deleted on chromosome 10

(Pten) loss In this case a cell line established from this

mouse model exhibited increased BIM expression upon

treatment with PD0325901 [47] Furthermore, the

MEK1⁄ 2 inhibitor, AZD6244, can cooperate with PI3K

inhibitors to inhibit the growth of otherwise refractory

colorectal cancer cell lines [48]

Synergistic interactions between MEK inhibitors and

other kinase inhibitors have also been reported [49,50]

UCN-01 is a reversible and ATP-competitive inhibitor,

which targets several protein kinases such as

cyclin-dependent kinases (CDKs), checkpoint protein 1

(CHK1), 3¢-phosphoinositide-dependent kinase-1

(PDK1) and protein kinase Cs (PKCs) Treatment of

multiple myeloma cells with UCN-01 alone resulted in

the activation of ERK1⁄ 2 and in the phosphorylation

and loss of BIMEL; however, co-administration of the

MEK1⁄ 2 inhibitor PD184352 stabilized BIMEL and

effectively synergized with UCN-01 to promote tumour

cell death [49,50] These observations suggest that

whilst MEK inhibition is sufficient to cause BIMEL

accumulation and subsequent abrogation of the

anti-apoptotic properties of BCL-2⁄ BCL-xL, other

UCN-01-inducible signals are required to cooperate with

BIM to induce apoptosis

The results of clinical trials utilizing some MEK1⁄ 2

inhibitors as a monotherapy [38] highlight the need for

a greater understanding of how these compounds act

to initiate cell death, which could guide the selection

of suitable therapeutic partners for combination

treat-ments BRAF or MEK inhibition has emerged as a

pivotal mediator of synergistic effects in combination

with other therapeutics Early indications suggest that

this can be attributed in part to the role the ERK1⁄ 2

pathway plays in controlling BIM expression

Treat-ment of ERK1⁄ 2 pathway-addicted cancer cells with a

BRAF or MEK inhibitor, resulting in an accumulation

of BIM, may serve to sensitize cells to other

pro-apop-totic stimuli or therapeutics

Non-small cell lung cancers with

epidermal growth factor receptor

mutation

Over-expression of the epidermal growth factor

recep-tor (EGFR) is frequently observed in a variety of

epithelial malignancies, including non-small cell lung cancer (NSCLC) A subgroup of NSCLC patients exhibit somatic activating mutations in the EGFR tyrosine kinase domain; these primary mutations corre-late well with clinical responses to the EGFR-specific, ATP-competitive tyrosine kinase inhibitors gefitinib and erlotinib, indicating that human NSCLC cells are addicted to these mutant EGFR oncoproteins [51] Several studies have now shown that human NSCLC cell lines harbouring these primary EGFR mutants undergo apoptosis upon treatment with gefitinib or erlotinib [52–55] Acquired resistance to gefitinib and erlotinib is a very real issue clinically, and most EGFR-mutant tumours that respond well in the first instance eventually become resistant, allowing disease progression Acquired resistance is frequently associ-ated with secondary mutations in the kinase domain (the most frequent being the T790M gatekeeper muta-tion, which impairs drug binding) but may also arise

as a result of the amplification of other oncogenes Gefitinib- or erlotinib-induced NSCLC cell death proceeds via the cell-intrinsic mitochondrial pathway, and increased expression of BIM is invariably an early event following treatment with these drugs in NSCLC harbouring primary EGFR mutations [52–54] Erloti-nib treatment also blocks the formation of tumours in transgenic mice that conditionally express the L858R EGFR mutation and inhibits the growth of NSCLC cells as xenografts; in both cases this is associated with increased BIM expression [52] In NSCLC cell lines harbouring primary EGFR mutations, knockdown of BIM by RNAi significantly, but not completely, reversed the cell death induced by gefitinib or erlotinib The partial protection afforded by knockdown of BIM may reflect a role for other BOPs, such as BAD or PUMA, or may simply reflect incomplete knockdown

of BIM Furthermore, NSCLC cell lines expressing the secondary T790M mutant EGFR were resistant to gefitinib and erlotinib and failed to upregulate BIM;

in such cases, BIM induction and cell death were re-imposed by administration of the structurally dis-tinct, covalent EGFR inhibitor, CL-387785 [53] These studies demonstrate that NSCLCs are addicted to the activity of their primary EGFR mutant for repression

of BIM and cell survival, and demonstrate that BIM, whilst not acting alone, is an important key effector of gefitinib- or erlotinib-induced cell death (Fig 2) Virtually all NSCLCs harbouring primary EGFR mutations exhibit strong activation of the ERK1⁄ 2 and PKB pathways, and treatment with gefitinib or erlotinib causes inactivation of both pathways Thus, expression of BIM and cell death could reflect loss of either pathway, or both In fact drug-induced loss of

ERK1⁄ 2 signalling contributes substantially to the

increase in BIM expression In the majority of NSCLC

cell lines treated with erlotinib or gefitinib, BIMELwas

the predominant isoform induced [52–54] and was

expressed predominantly as the dephosphorylated,

sta-bilized, active form, correlating with loss of ERK1⁄ 2

activity Finally, inhibitors of PI3K or PKB did not

cause accumulation of BIM, whereas inhibitors of

MEK–ERK1⁄ 2 signalling did [54] However, despite

causing increased BIM expression, inhibition of the

ERK1⁄ 2 pathway alone caused little cell death in

com-parison to that seen with gefitinib, suggesting that loss

of other signalling pathways (and activation of other

BOPs?) must also contribute to gefitinib-induced cell

killing, as discussed elsewhere [10]

BCR-ABL inhibitors and chronic

myeloid leukaemia

Chronic myeloid leukaemia (CML) is characterized by

the presence of the t(9;22)(q34;q11) reciprocal

trans-location, giving rise to the breakpoint cluster region–

Abelson murine leukaemia viral oncogene (BCR–ABL)

fusion oncoprotein [56] The mutant BCR–ABL

tyro-sine kinase activates several signalling pathways,

includ-ing the ERK1⁄ 2 pathway, the PKB pathway and the

Janus kinase⁄ signal transducer and activator of

tran-scription (JAK-STAT) pathway, to promote

prolifera-tion, survival and transformation [56,57] The

importance of the BCR–ABL tyrosine kinase in the

sur-vival of CML cells led to the development of the tyrosine

kinase inhibitor imatinib (STI571, Gleevec), which is a

potent inhibitor of BCR–ABL, platelet-derived growth

factor receptor (PDGFR) and oncogene of HZ4 feline

sarcoma virus (KIT) and has produced impressive

results in clinical trials in CML [57,58] Resistance to

imatinib has proved to be a problem clinically, and 40%

of patients who relapse on imatinib therapy have point

mutations in the BCR–ABL kinase domain, including

the T315I gatekeeper mutation that impairs imatinib

binding [58,59] Accordingly, new therapies are being

tested and these include second-generation tyrosine

kinase inhibitors, such as dasatinib (inhibits BCR–ABL,

KIT, PDGFR and SRC family kinases), nilotinib

(a more potent inhibitor of BCR–ABL, KIT and

PDGFR), INNO406 (a dual BCR-ABL and Lyn

inhibi-tor) and PPY-A and PHA-739358 (which can inhibit the

T315I mutant of BCR–ABL) [57,60]

Several observations suggest that BIM is important

in promoting cell death following BCR–ABL

inhibi-tion Downregulation of BIM is of key importance in

cytokine-mediated survival in murine haematopoetic

progenitor cells [61] Expression of BCR–ABL in

inter-leukin-3-dependent Baf-3 cells represses BIM and allows interleukin-3–independent survival; treatment of these cells with imatinib reverses the effects of BCR– ABL, resulting in increased expression of BIM and cell death [62,63] RNAi has shown that BIM expression is required, at least in part, for imatinib-induced apopto-sis in BCR–ABL-transformed murine progenitor cells [64] and in BCR–ABL+ K562 CML cells in response

to both imatinib and nilotinib [65] INNO-406, a more potent inhibitor of BCR–ABL, also induces apoptosis

in BCR–ABL+ K562 cells, and whilst BCR–ABL-expressing myeloid progenitor cells from BIM-⁄ - mice are partially protected against INNO-406-induced apoptosis, substantially greater protection is seen in BIM-⁄ -BAD-⁄ - double knockout cells or upon BCL-2 overexpression [66]

These results reveal that a variety of first-generation and second-generation BCR–ABL inhibitors increase BIM expression and elicit BIM-dependent cell death in CML cells, with BIM acting in concert with other BOPs, such as BAD (Fig 2) A notable exception to this is the third-generation dual BCR–ABL and pan-aurora kinase inhibitor MK-0457, which can inhibit both wild-type and imatinib-resistant BCR–ABL mutants (including T315I) Despite inhibiting BCR– ABL, MK-0457 predominantly induces polyploidy, rather than apoptosis, in BCR–ABL+ CML cells, probably reflecting its activity against the aurora kinases; the lack of cell death induced by MK-0457 correlates with its inability to increase BIM expression [67] However, because other BCR–ABL inhibitors do increase BIM expression, it is surprising that MK-0457 does not; does this suggest that the additional inhibi-tory activity against aurora kinases antagonizes BIM expression, or does it suggest other targets for MK-0457? Regardless, MK-0457 can synergize with the histone deacetylase inhibitor vorinostat to promote apoptosis, and this synergy does involve BIM; vorino-stat increases BIM expression and BIM plays a signifi-cant role in the induction of apoptosis observed with the combination of the two drugs [67]

MEK inhibitors alone tend to cause only modest cell death in CML cells, but PD184352 can act synergisti-cally with imatinib or dasatinib in BCR–ABL+ cells, leading to a substantial increase in apoptosis [68,69] Treatment of BCR–ABL-expressing Baf-3 cells with either the pan-MEK1⁄ 2 ⁄ 5 inhibitor, PD98059, or the pan-PI3K inhibitor, LY294002, could induce BIM expression [62], although other studies showed that PD98059, but not LY294002, could increase BIM expression in BCR–ABL-expressing Baf-3 cells and primary CML cells [63] In summary, the repression of BIM downstream of BCR-ABL appears to be mediated

predominantly by the constitutive activity of the

ERK1⁄ 2 and perhaps by the PI3K pathway

Regulation of BIM by other oncogenes/

pathways

FMS-like tyrosine kinase 3 (FLT3), a receptor tyrosine

kinase related to PDGFR and KIT, is frequently

mutated in acute myeloid leukemia (AML) and this

correlates with poor prognosis; mutant FLT3 proteins

typically exhibit ligand-independent dimerization and

activation Treatment of primary AML cells with

either of two FLT3 inhibitors (AG1295 or PKC412)

caused a substantial cell-death response [70]

Activa-tion of the PI3K–PKB pathway downstream of FLT3

was the major pathway responsible for repressing

FOXO3A and BIM expression, and whilst both BIM

and PUMA were upregulated following FLT3

inhibi-tion, only loss of BIM was able to preserve clonogenic

survival in FDC-p1 cells expressing mutant FLT3

pro-teins Thus, AML cells expressing mutant FLT3 are

addicted to FLT3-dependent signalling via the PI3K–

PKB pathway for repression of BIM and cell survival

Whilst there is a well-defined role for the PI3K–

PKB pathway in repressing FOXO3A (see above),

studies have also suggested a role for mTOR in

regu-lating BIM expression The earliest study to suggest

this demonstrated that treatment of haematopoietic

progenitor cells with the mTOR inhibitor, rapamycin,

increased BIM expression and overcame

RAS-depen-dent survival signals to promote cell death, arguing

that mTOR was an important survival signal that

acted, in part, by repressing BIM [61] Most recently,

prominent effects of rapamycin on BIM were

demon-strated in a mouse model of androgen-independent

prostate cancer [71] In this instance, the combination

of the MEK1⁄ 2 inhibitor, PD0325901, and rapamycin

was remarkably effective at inhibiting the growth of

prostate cancer cell lines and the growth of prostate

cancer in vivo in Pten-deficient mice Interestingly,

although rapamycin alone failed to increase BIM

expression, the combination of PD0325901 and

rapa-mycin was more effective than PD0325901 alone, and

death arising from the combination therapy was at

least partially BIM-dependent [71] Both of these

stud-ies suggest that TOR activity normally represses BIM

and that therapeutic inhibition of TOR will increase

BIM levels and contribute to tumour cell death

How-ever, a recent study suggests that repression of BIM

might not be a direct effect of TOR mTOR exists in

mammalian cells as two distinct complexes: mTORC1

(composed of mTOR, mLST8 and raptor) regulates

cell growth via the effectors S6K1 and 4E-BP1, whilst

mTORC2 (composed of mTOR, mLST8 and rictor) phosphorylates PKB at Ser473, contributing to its acti-vation Rapamycin binds to FKBP12 and, in this form, inhibits preformed mTORC1 complexes but not preformed mTORC2; as a result, the effects of rapa-mycin are frequently attributed to inhibition of mTORC1 alone However, FKBP12–rapamycin can bind to free mTOR, and Sabatini and co-workers have recently shown that prolonged treatment of cells with rapamycin can actually cause disassembly of mTORC2, loss of PKB Ser473 phosphorylation and apoptosis [72] Thus, it is quite possible that the ability

of rapamycin to contribute to BIM expression during prolonged treatments with drug [61,71] may actually reflect loss of PKB phosphorylation and activation of FOXO3A, rather than loss of a direct effect of mTOR Such details do not, of course, detract from the strik-ing synergy seen between MEK1⁄ 2 inhibitors and rapamycin [47,71], but are important in understanding the mechanisms by which these drugs cooperate to kill tumour cells

In addition to promoting cell proliferation and transformation, the c-Myc proto-oncogene is renowned for its ability to promote cell death [73] and there is now good evidence to indicate that (a) BIM is impor-tant in Myc-induced cell death and (b) that this may

be an arbiter of tumour progression B-lymphoid cells from El-Myc transgenic mice exhibited increased expression of BIM and an increased propensity to undergo apoptosis, which was lost on a BIM-⁄ - back-ground Loss of even a single BIM allele accelerated Myc-induced tumour progression, giving rise to acute B-cell leukaemia These results demonstrate that Myc can promote expression of BIM and show that BIM is

a tumour suppressor in this system [74] Myc-induced BIM expression may be therapeutically relevant in other tumour models For example, in human glioma cell lines, several distinct glycogen synthase kinase 3 inhibitors cause activation of c-Myc, expression of Myc target genes (including BIM) and glioma death, although the role of BIM, as opposed to other Myc target genes, was not defined [75]

BH3 mimetics: giving BIM a helping hand

By virtue of its ability to engage with and inhibit all of the prosurvival BCL-2 proteins, BIM is one of the most potent and effective BOPs, in terms of cell kill-ing, when assayed by overexpression [4] Despite this, oncogene-targeted therapeutics alone can often cause quite significant increases in BIM expression, but rela-tively modest tumour cell death Similarly, the

response observed in the clinic may often be cytostatic

(i.e associated with tumour stasis, rather than with

cytotoxicity and tumour regression) This may be

because the level of BIM upregulation achieved is not

sufficient for apoptosis and⁄ or because tumours often

exhibit elevated expression of certain prosurvival

BCL-2 proteins, providing an effective buffer to the

drug-induced expression of BIM or other BOPs In

this context, recent studies suggest that the use of new

oncogene-targeted therapeutics, in combination with

BH3 mimetics, may prove particularly effective

BH3 mimetics are small, cell-permeant molecules that

mimic BOPs by binding and inhibiting prosurvival

BCL-2 proteins [11,76,77] The prototype, ABT-737,

binds to BCL-2, BCL-xLand BCL-2-like protein 2 with

high affinity and is thought to act by liberating BAX

and BAK from these proteins; in addition, BOPs

dis-placed by treatment with ABT-737 may bind and inhibit

MCL-1, providing further activation of BAX⁄ BAK

ABT-737 can kill certain tumour cells as a single agent

or when administered with conventional cytotoxic

chemotherapeutics; more importantly, it can cooperate

with oncogene-targeted therapeutics to provide

some-times quite striking synergistic tumour cell killing

(Fig 2)

Tumours with BRAF mutations

Even in tumour cells with BRAF600E that show strong

addiction to ERK1⁄ 2 signalling for proliferation, MEK

inhibition alone can often induce quite striking increases

in BIM expression but only modest tumour cell death

[41–43] Cragg et al [43] noted that high levels of the

anti-apoptotic BCL-2 protein correlated with low levels

of cell death in response to the first-generation

pan-MEK1⁄ 2 ⁄ 5 inhibitor, U0126, in a range of tumour-cell

lines harbouring BRAF600E and found that ABT–737

synergized with U0126 to promote extensive apoptosis

in BRAF mutant SkMel-28 melanoma and Colo205

colon cancer cells; this was associated with the

ABT-737-dependent redistribution of BIM from BCL-2

to MCL-1 Striking synergy was also observed when

ABT-737 and the second-generation MEK1⁄ 2-specific

compound, PD0325901, were combined to treat

SkMel-28 and Colo205 xenografts in nude mice, resulting in

partial tumour regression [43], providing compelling

support for the use of MEK inhibitors in combination

with BH3 mimetics in tumours with BRAF600E

NSCLC with EGFR mutations

In NSCLC cell lines that are sensitive to erlotinib or

gefitinib and exhibit drug-induced expression of BIM,

the induction of apoptosis is often modest However, two studies have now shown that erlotinib or gefitinib-induced cell death can be greatly enhanced by co-administration of ABT-737 In the case of erlotinib, synergy with ABT-737 was observed in PC-9 and H2355 cells [52], whilst in the case of gefitinib, synergy with ABT-737 was most pronounced in H358, H1975 and H1650 cells, gefitinib alone being more efficacious

in H3255 cells [54]

CML with BCR-ABL Prosurvival BCL-2 family proteins such as MCL-1, BCL-2 and BCL-xL are often expressed at high levels

in CML cells [62,63,78,79], and this prompted an investigation of the efficacy of combinations of

ABT-737 and BCR-ABL inhibitors Indeed, ABT-ABT-737 coop-erates effectively with imatinib [64] and INNO-406 [66]

to promote death of CML cells This cooperation was also seen in Baf-3 cell lines expressing two different mutant BCR-ABL proteins (E255K and H396P), but not in those expressing the T315I gatekeeper mutant These authors also demonstrated that ABT-737 could enhance apoptosis in response to 17-AAG [66], which inhibits the activity of the HSP90 chaperone required for correct BCR–ABL folding

Together these examples indicate the powerful synergy that is observed when therapies targeting oncogenic kinases are combined with BH3 mimetics, giving rise to substantially greater tumour cell killing

in vitro and tumour regression in vivo [77] (Fig 2) Obviously this is a desirable outcome in its own right, but it may have other advantages For example, the predominantly cytostatic effects of oncogenic kinase inhibitors alone mean that tumour cells stay alive and receive prolonged exposure to the drug; this may explain the frequent emergence of acquired resistance

in CML and NSCLC In contrast, the more substantial and precipitate cell-death response seen with combina-tions of kinase inhibitors and BH3-mimetics may sub-stantially shorten the window of opportunity for acquisition and⁄ or selection of secondary mutations, making acquired resistance less likely to arise Answers

to such speculation may be informed by tissue culture and animal models, but ultimately will come from clinical studies

Conclusions

A combination of basic and applied biology in the last

5 years has provided a good working model for how BIM is inhibited by survival signalling pathways, notably the ERK1⁄ 2 pathway, and has led to the

recognition that BIM plays a major response in

tumour cell death arising from inhibition of oncogenic

kinases Indeed, the increased expression of

dephos-phorylated BIMEL could even be viewed as a

biomar-ker for drug-induced inactivation of ERK1⁄ 2 and

engagement of the BCL-2 axis Whether this increase

in BIM gives rise to substantial tumour cell death will

depend on the activation of other survival pathways

and expression of other BOPs or prosurvival BCL-2

proteins In instances of high BCL-2 or BCL-xL

expression, drug-induced, BIM-dependent cell death

will be greatly enhanced by combination with BH3

mimetics, giving tumour regression and potentially less

opportunity for resistance to arise Moving forward, if

we are to take advantage of this clinically, tumours

will be sampled for oncogene mutations, biochemical

signatures of signal pathway activation (to address

pathway redundancy [10]) and expression of BCL-2

family proteins to match the treatment combination to

the tumour fingerprint of the patient

Acknowledgements

We apologise to colleagues in the field whose work

we have had to omit because of space constraints

Work in the Cook laboratory is funded by the

Asso-ciation for International Cancer Research,

Astra-Zeneca, the Babraham Institute, the Biotechnology

and Biological Sciences Research Council and Cancer

Research UK

References

1 Adams J & Cory S (2007) The Bcl-2 apoptotic switch in

cancer development and therapy Oncogene 26, 1324–

1337

2 Youle RJ & Strassser A (2009) The BCL-2 protein

family: opposing activities that mediate cell death Nat

Rev Mol Cell Biol 9, 47–59

3 Willis SN, Fletcher JI, Kaufmann T, van Delft MF,

Chen L, Czabotar PE, Lerino H, Lee EF, Fairlie WD,

Bouillet P et al (2007) Apoptosis initiated when BH3

ligands engage multiple Bcl-2 homologs, not Bax or

Bak Science 315, 856–859

4 Chen L, Willis SN, Wei A, Smith BJ, Fletcher JI, Hinds

MG, Colman PM, Day CL, Adams JM & Huang DC

(2005) Differential targeting of prosurvival Bcl-2

pro-teins by their BH3-only ligands allows complementary

apoptotic function Mol Cell 17, 393–403

5 O’Connor L, Strasser A, O’Reilly LA, Hausmann G,

Adams JM, Cory S & Huang DC (1998) Bim: a novel

member of the Bcl-2 family that promotes apoptosis

EMBO J 17, 384–395

6 U M, Miyashita T, Shikama Y, Tadokoro K & Yamada M (2001) Molecular cloning and characteriza-tion of six novel isoforms of human Bim, a member of the proapoptotic Bcl-2 family FEBS Lett 509, 135–141

7 Ley R, Ewings KE, Hadfield K & Cook SJ (2005) Reg-ulatory phosphorylation of Bim: sorting out the ERK from the JNK Cell Death Differ 12, 1008–1014

8 Lei K & Davis RJ (2003) JNK phosphorylation of Bim-related members of the Bcl2 family induces Bax-dependent apoptosis Proc Natl Acad Sci USA 100, 2432–2437

9 Hu¨bner A, Barrett T, Flavell RA & Davis RJ (2008) Multisite phosphorylation regulates Bim stability and apoptotic activity Mol Cell 30, 415–425

10 Balmanno K & Cook SJ (2009) Tumour cell survival signalling by the ERK1⁄ 2 pathway Cell Death Differ

16, 368–377

11 Fesik SW (2005) Promoting apoptosis as a strategy for cancer drug discovery Nat Rev Cancer 5, 876–885

12 Bouillet P, Metcalf D, Huang DC, Tarlinton DM, Kay

TW, Ko¨ntgen F, Adams JM & Strasser A (1999) Proa-poptotic Bcl-2 relative Bim required for certain apopto-tic responses, leukocyte homeostasis, and to preclude autoimmunity Science 286, 1735–1738

13 Whitfield J, Neame SJ, Paquet L, Bernard O & Ham J (2001) Dominant-negative c-Jun promotes neuronal sur-vival by reducing BIM expression and inhibiting mito-chondrial cytochrome c release Neuron 29, 629–643

14 Ewings KE, Hadfield-Moorhouse K, Wiggins CM, Wickenden JA, Balmanno K, Gilley R, Degenhardt K, White E & Cook SJ (2007) ERK1⁄ 2-dependent phosphorylation of BimEL promotes its rapid dissocia-tion from Mcl-1 and Bcl-xL EMBO J 26, 2856–2867

15 Gilley J, Coffer PJ & Ham J (2003) FOXO transcrip-tion factors directly activate bim gene expression and promote apoptosis in sympathetic neurons J Cell Biol

162, 613–622

16 Fu Z & Tindall DJ (2007) FOXOs, cancer and regula-tion of apoptosis Oncogene 27, 2312–2319

17 Yang JY, Zong CS, Xia W, Yamaguchi H, Ding Q, Xie X, Lang JY, Lai CC, Chang CJ, Huang WC et al (2008) ERK promotes tumorigenesis by inhibiting FOXO3a via MDM2-mediated degradation Nat Cell Biol 10, 138–148

18 Weston CR, Balmanno K, Chalmers C, Hadfield K, Molton SA, Ley R, Wagner EF & Cook SJ (2003) Activation of ERK1⁄ 2 by DRaf-1:ER* represses Bim expression independently of the JNK or PI3K pathways Oncogene 22, 1281–1293

19 Ley R, Balmanno K, Hadfield K, Weston C & Cook SJ (2003) Activation of the ERK1⁄ 2 signaling pathway promotes phosphorylation and proteasome-dependent degradation of the BH3-only protein, Bim J Biol Chem

278, 18811–18816