A Novel AKT1 Mutant Amplifies an Adaptive Melanoma Response to BRAF Inhibition

Published OnlineFirst November 21, 2013; DOI: 10.1158/2159-8290.CD-13-0279

RESEARCH BRIEF

A Novel AKT1 Mutant Ampliﬁ es an Adaptive Melanoma Response to BRAF Inhibition

Hubing Shi 1 , 6, Aayoung Hong 1 , 5 , 6, Xiangju Kong 1 , 6, Richard C. Koya 2 , 6, Chunying Song 1 , 6, Gatien Moriceau 1 , 6, Willy Hugo 1 , 6, Clarissa C. Yu 1 , 6, Charles Ng 3 , 6, Thinle Chodon 3 , 6, Richard A. Scolyer 7 , 8 , 10, Richard F. Kefford 7 , 9 , 10, Antoni Ribas 2 , 3 , 4 , 5 , 6, Georgina V. Long 7 , 9 , 10, and Roger S. Lo 1 , 4 , 5 , 6

ABSTRACT BRAF inhibitor (BRAFi) therapy leads to remarkable anti melanoma responses, but the initial tumor shrinkage is commonly incomplete, providing a nidus for subsequent disease progression. Adaptive signaling may underlie early BRAFi resistance and inﬂ uence the selection pattern for genetic variants, causing late, acquired resistance. We show here that BRAFi (or BRAFi + MEKi) therapy in patients frequently led to rebound phosphorylated AKT (p-AKT) levels in their melanomas early on-treatment. In cell lines, BRAFi treatment led to rebound levels of receptor tyrosine kinases (RTK; β including PDGFR ), phosphatidyl (3,4,5)-triphosphate (PIP3 ), pleckstrin homology domain recruitment, and p-AKT. PTEN expression limited this BRAFi-elicited PI3K–AKT signaling, which could be rescued by the introduction of a mutant AKT1 (Q79K) known to confer acquired BRAFi resistance. Functionally, AKT1Q79K conferred BRAFi resistance via ampliﬁ cation of BRAFi-elicited PI3K–AKT signaling. In addition, mitogen- activated protein kinase pathway inhibition enhanced clonogenic growth dependency on PI3K or AKT. Thus, adaptive or genetic upregulation of AKT critically participates in melanoma survival during BRAFi therapy.

SIGNIFICANCE: This study provides a mechanistic link between early, adaptive and late, acquired BRAF inhibitor resistance in melanoma, with early BRAFi-induced signaling alterations shaping the subsequent evolutionary selective pressure. These ﬁ ndings argue for upfront, combined targeting of the mutant BRAF genotype and a pervasive drug-adaptive, AKT-dependent tumor response. Cancer Discov; 4(1); 69–79. ©2013 AACR.

See related commentary by Solit and Rosen, p. 27.

INTRODUCTION kinase (MAPK) pathway and result in oncogene addiction. Therapy with a BRAF inhibitor (BRAFi) or its combination About 50% of metastatic melanomas harbor BRAF V600 with a MAP-ERK kinase (MEK) inhibitor (MEKi) leads to rapid mutations, most commonly a V600E substitution ( 1 ), which and high rates of clinical responses ( 2, 3 ), but the initial tumor constitutively hyperactivate the mitogen-activated protein shrinkage is typically partial, providing niduses for eventual

Authors’ Afﬁ liations: 1 The Division of Dermatology, Department of Medicine, H. Shi and A. Hong contributed equally to this work. 2 3 Division of Surgical Oncology, Department of Surgery, Division of Hematology Corresponding Author: Roger S. Lo, 52-121 CHS Department of Medicine/ 4 and Oncology, Department of Medicine, Jonsson Comprehensive Cancer Center, Dermatology, 10833 Le Conte Avenue, Los Angeles, CA 90095-1750. 5 6 Department of Molecular and Medical Pharmacology, David Geffen School of Phone: 310-825-5420; Fax: 310-206-9878; E-mail: [email protected] Medicine, University of California, Los Angeles, California; 7 Melanoma Institute of Australia, 8 Royal Prince Alfred Hospital, 9 Westmead Millennium Institute, and doi: 10.1158/2159-8290.CD-13-0279 10Westmead Hospital, University of Sydney, New South Wales, Australia ©2013 American Association for Cancer Research. Note: Supplementary data for this article are available at Cancer Discovery Online (http://www.cancerdiscovery.aacrjournals.org).

JANUARY 2014CANCER DISCOVERY | 69

RESEARCH BRIEF Shi et al.

disease progression caused by acquired drug resistance ( 4 ). It Supplementary Table S1) are critical for this analysis because is unclear what adaptive signal(s) underlies initial incomplete baseline p-AKT levels, assessed in parallel technically, var- tumor responses and whether a defi ciency in such adaptive ied over a considerable range among tumors from distinct signaling might favor genetic selection for its upregulation. patients. Importantly, in the majority of tumors biopsied We have recently shown that a majority of melanomas with early on-treatment (7 of 9), p-AKT increased upon treatment. late or acquired resistance to a BRAFi display molecular or We also examined the PDGFRβ levels by IHC, as this receptor genetic alterations that lead to MAPK pathway reactivation tyrosine kinase (RTK) has previously been correlated with (5–9 ). However, a signifi cant minority of disease-progressing BRAFi resistance (5 , 12 ). Induction of PDGFRβ expression melanomas harbor genetic lesions that upregulate the phos- upon treatment was observed in 5 of 7 tumors (Supplemen- phoinositide 3-kinase (PI3K)–PTEN–AKT pathway ( 9 ). These tary Fig. S3) that displayed a relative increase in p-AKT levels, genetic alterations include nonsynonymous, gain-of- function but was not observed in the two tumors that did not display mutations in positive regulators (PIK3CA , AKT1 , and AKT3 ) a relative p-AKT increase, consistent with our observations in and loss-of-function mutations (PIK3R2 , PTEN , and PHLPP1) cell lines (see below). Interestingly, as Patient #1 developed dis- and gene deletion (PTEN ) in negative regulators. While some ease progression (via mutant NRAS) on BRAFi, he was treated somatic variants (e.g., PIK3CAE545G, AKT1/3E17K) harbor with BRAFi + MEKi (BRIM7 clinical trial). This combinatorial known mechanisms of action, the structure-functional con- treatment (day 15) was also associated with PDGFRβ induc- sequences of others (e.g., PIK3R2N561D, AKT1Q79K, PTENM134del) tion (Supplementary Fig. S3). In preliminary work with fi ve are novel. sets of transcriptomes for patient-matched pre- and early on- Given that the landscape of mechanisms of acquired BRAFi treatment (days 6–22; BRAFi or BRAFi + MEKi) melanomas, resistance refl ect the essential survival functions of MAPK we observed PDGFRB mRNA induction by MAPK pathway and PI3K–PTEN–AKT signaling pathways in BRAF -mutant inhibition in 5 of 5 patients (fold increase from 1.4 to 2.9, FDR melanoma, we explored the roles of these pathways in acute adjusted P ≤ 0.05; unpublished data). Thus, inhibition of the or adaptive resistance. Recent works in cancer cell lines (10, MAPK pathway in BRAF V600 -mutant melanoma was associ- 11) have demonstrated how targeted inhibition of the MAPK ated with a tumor-associated rebound PDGFRβ and p-AKT pathway can lead to adaptive alterations in signal transduc- induction in the majority of patients. tion crosstalk/feedback and transcriptional output. Here, As in BRAF -mutant melanoma tumors, BRAF -mutant we show that in clinical melanoma biopsy specimens, MAPK human melanoma cell lines displayed a wide range of basal pathway inhibition elicited a pervasive rebound in AKT acti- p-AKT levels at exponential growth densities (Supplemen- vation. In cell line work, this rebound in AKT activation could tary Fig. S4). We have previously shown that certain BRAF - be traced to enhanced phosphatidyl (3,4,5)-triphosphate mutant human melanoma sublines chronically treated with

(PIP3 ) generation and was antagonized by wild-type (WT) a BRAFi (or some cases of short-term cultures derived from PTEN. Moreover, the pleckstrin homology domain (PHD) BRAFi-resistant melanoma tissues in patients) upregulated class of AKT mutants, including the novel AKT1Q79K mutant, the expression and tyrosine phosphorylation levels of two β displayed enhanced affi nity for PIP3 , dramatically amplifi ed key RTKs, namely PDGFR and EGF receptor (EGFR), as the BRAFi-induced upregulation of PI3K–AKT signaling in well as the levels of p-AKT. More recent work on triple- the presence of PTEN, and as such conferred BRAFi resist- negative breast cancer cell lines uncovered a negative feed- ance in a cell context–dependent manner. These data thus back of the MAPK pathway (via MYC-mediated transcrip- reveal key insights into the dynamic evolutionary continuum tional repression) on RTK expression (10 ). To assess the of BRAFi resistance. Early adaptive events in response to timing of p-AKT induction in response to BRAF (or MAPK MAPK inhibitors, which limit the initial effi cacy, may shape pathway) inhibition, we treated BRAF -mutant melanoma cell subsequent mechanisms of acquired resistance, which fur- lines with a BRAFi (vemurafenib), a MEKi (AZD6244), or an ther augment adaptive responses and lead to tumor regrowth. ERKi (FR180204). Inhibitor treatment for 48 hours consistently led to an induction of p-AKT (Thr308) levels, suggesting crosstalk at a level below ERK signaling ( Fig. 1B ). MAPK RESULTS pathway inhibition (by BRAFi, MEKi, or ERKi) for 48 hours Although AKT has been implicated in BRAFi resistance, led to concomitant c-MYC downregulation, RTK/PDGFRβ the timing of its upregulation during BRAFi therapy is not (but not EGFR, see below) overexpression, and p-AKT and clear. To determine whether MAPK pathway targeting results p-CRAF upregulation, correlating with an induction of in a rebound increase in phosphorylated AKT (p-AKT) in PDGFRB but not EGFR mRNA levels (data not shown). the tumors of patients treated with a BRAFi (or BRAFi + While EGFR tyrosine phosphorylation can transiently MEKi), we assessed the p-AKT levels in melanoma tumors increase within hours of BRAFi treatment (13 ), its protein biopsied early on-treatment (days 4 to 25) relative to their expression level was not induced until after prolonged (weeks) patient-matched baseline (pretreatment) tumors by immu- BRAFi treatment coinciding with drug-tolerant persisting nohistochemistry (IHC; Fig. 1A ) using a validated antibody (but slowly cycling; unpublished data) cell subpopulations (Supplementary Fig. S1) and quantifying signals via image entering a proliferative phase (Supplementary Fig. S5A and analysis (Supplementary Fig. S2). These early on-treatment S5B). Interestingly, although BRAFi induced p-AKT levels in biopsies took place during a period of time to objective M229 cells at 48 and 72 hours, cotreatment with a PDGFRβ tumor responses, before or at maximal tumor responses. inhibitor, sunitinib, or an EGFR inhibitor, gefi tinib, reduced Such patient-matched (and in most cases also tumor site– (but did not completely abolish) this BRAFi-induced p-AKT matched) melanoma samples (n = 9 patients; clinical data in level, with the strongest effect observed with the cotreatment

70 | CANCER DISCOVERYJANUARY 2014 www.aacrjournals.org

MAPK Inhibition Elicits a Widespread, AKT-Dependent Adaptive Response in Melanoma RESEARCH BRIEF

AB Baseline Early on-Rx M229 M238 WM2664 Intensity Intensity p-AKT p-AKT BRAFi MEKi ERKi BRAFi MEKi ERKi BRAFi MEKi ERKi overlay overlay 048048 048 0 48 0 48 0 48 0 48 0 48 0 48 Time (48 h) p-AKT Pt 1 Thr308

p-AKT Ser473

AKT Pt 2

p-ERK1/2

Pt 3 ERK1/2

c-MYC

Pt 4 PDGFR-β

p-CRAF Pt 5

CRAF

Tubulin Pt 6

0.7 μ Pt 7 Vemu (1 mol/L) 0.6 0 h 0.5 6 h 0.4 48 h ( μ mol/L)

Pt 8 3 0.3 PIP 0.2 0.1 Pt 9 0 M229 M238 WM2664

D Vemu 48 h (1 μmol/L) 0 h 6 h 24 h 48 h (zoom) M229 M238 WM2664

Figure 1. MAPK pathway inhibition leads to rebound upregulation of AKT signaling in melanoma tumors and cell lines. A, BRAFi (vemurafenib or dabrafenib) or BRAFi + MEKi treatment led to increased p-AKT Ser473 levels in early on-treatment tumor biopsies (days #4 to 25) relative to the patient-matched, baseline (pretreatment) biopsies. p-AKT IHC of melanin-bleached tumor sections (×400; quantiﬁ cation shown as heat intensity). Rx, symbol for treatment. B, BRAF -mutant melanoma cell lines were treated with the indicated MAPK inhibitor (1 μmol/L) or dimethyl sulfoxide (DMSO) for 48 hours. Phospho- and total protein levels were then probed by Western blotting. Tubulin, loading control. C, lipids were extracted from the indicated cell lines (n = 3 per group) μ treated with vemurafenib (1 mol/L) for increasing durations (h), and the PIP3 levels were detected by ELISA (average of biologic triplicates; error bar, SD). D, localization of WT PHD–GFP cells treated with vemurafenib (1 μmol/L) for increasing durations (h; scale bar, 20 μm). Note the cellular morphologic response to vemurafenib treatment in the surviving subpopulations. Photomicrographs representative of two independent experiments. Vemu, vemurafenib.

JANUARY 2014CANCER DISCOVERY | 71

RESEARCH BRIEF Shi et al.

of BRAFi with both sunitinib and gefi tinib (Supplementary We fi rst confi rmed that both AKT1 mutants are associated Fig. S5C). This suggests that these RTKs either contribute with upregulated levels of activation-associated phosphoryla- independently to p-AKT induction or can compensate for tion (Fig. 2B ). We then tested the structure prediction (9 ) that each other’s kinase inhibition. After chronic vemurafenib the mechanism of AKT1Q79K gain-of-function is based on its

treatment, drug-resistant, BRAF -mutant melanoma sublines mutant PHD displaying enhanced PIP3 binding and mem- (M229 R5, M238 R1), which overexpress both wild-type brane recruitment, as has been shown for the mutant PHD of PDGFRβ and EGFR, could be resensitized to vemurafenib AKT1E17K ﬁ rst described in breast cancer ( 14 ). In agreement by cotreatment with sunitinib and geﬁ tinib, suggesting a with our structure-based prediction, an AKT1 PHD contain- BRAFi-induced dependency on RTK activities (Supplemen- ing the Q79K mutation and fused to GFP localized to the tary Fig. S5D). Accordingly, BRAFi treatment induced apop- cell surface independently of serum stimulation, in contrast tosis in M229 and M238 cell lines (Supplementary Fig. S6A) to WT PHD–GFP and similar to a PHD–GFP fusion pro- and the cell-surface expression of PDGFRβ (Supplementary tein containing a known AKT1-activating mutation, E17K Fig. S6B). In fact, cell-surface PDGFRβ expression was higher (Fig. 2C ; ref. 14 ). Moreover, the increased recruitment of in the live subpopulations compared with the dead sub- Q79K PHD–GFP to the cell surface was also less sensitive to populations (Supplementary Fig. S6C), and the live sub- PI3K inhibition by LY294002 when compared with the WT populations over time were also more strongly enriched PHD–GFP in the presence of serum stimulation, suggesting with PDGFRβ-positive cells (Supplementary Fig. S6D). These that AKT1Q79K may be hyper-responsive to cell surface mem-

data suggest that BRAF inhibition speciﬁ cally (and MAPK brane recruitment by low levels of PIP3 . pathway inhibition in general) upregulates RTK expression, To assess the baseline (no BRAFi treatment) activation which contributes to adaptive AKT signaling. We then traced status of AKT1Q79K in the context of BRAF -mutant human the BRAFi-induced p-AKT levels to an induction of lipid/ melanoma cell lines with varying PTEN expression (Supple-

membrane-associated PIP3 levels ( Fig. 1C ). This increase in mentary Fig. S4), we derived stable cell lines expressing either Q79K PIP 3 levels correlated with an increase in cell-surface recruit- the vector or AKT1 in a doxycycline-repressible manner ment of the PHD (Fig. 1D ), as visualized by PHD–GFP fusion (Fig. 2C ). Upon induction of expression for 2 days, similar protein expressed at similar levels (Supplementary Fig. S7). FLAG-tagged AKT1Q79K expression levels were achieved in Thus, BRAFi treatment led to rebound RTK upregulation, all cell lines ( Fig. 2C ), resulting in an expression level of the

PIP3 membrane accumulation, PHD membrane recruitment, FLAG-tagged AKT1 (top band) comparable with that of the and AKT activation. endogenous AKT (bottom band). Notably, in the absence These studies in melanoma tissues and cell lines (Fig. 1 ) of BRAF inhibition, activation-associated phosphorylation support the notion that BRAF inhibition leads to early, adap- of FLAG-AKT1Q79K (top band) was greater than that of the tive AKT signaling. In a subset of melanoma tissues with late, endogenous WT AKT1 (bottom band) in each cell line. This acquired BRAFi resistance, we recently uncovered mutations, was most evident in the PTEN WT-expressing line M229, including gain-of-function AKT1 and AKT3 mutations, which where the endogenous p-AKT level was very low, despite upregulate the PI3K–AKT pathway and confer BRAFi resist- the total AKT levels being similar across all cell lines. The ance ( 9 ), identifying the PI3K–AKT pathway as another core p-AKT1Q79K level was lowest in the cell line (M229) with drug escape pathway in addition to MAPK reactivation. In the lowest basal endogenous p-AKT level, indicating that, this setting of late, acquired BRAFi resistance, the detection although the AKT1Q79K mutant is more readily phospho- frequency of MAPK reactivating molecular alterations was rylated and activated compared with the WT, its maximal greater than that of PI3K–PTEN–AKT-upregulating altera- phosphorylation still requires upstream signal activation (i.e.,

tions, raising the possibility that this pattern might be shaped PIP3 generation). Introduction of WT PTEN into the PTEN by early pathway responses to BRAFi therapy (Fig. 2A ). In other nonexpressing cell lines WM2664 and M249 (Supplementary words, lack of a robust early adaptive p-AKT tumor response Fig. S4) suppressed the endogenous p-AKT level (Fig. 2D ; to BRAFi treatment would eventually provide a selective data not shown). Upon restoration of PTEN expression and

growth advantage for (potentially preexisting) genetic sub- under a lowered PIP3 environment, the AKT1 PHD mutants clones that harbor PI3K–AKT-upregulating alterations. Thus, (E17K, Q79K) were much more activated than AKT1 WT, i.e., the prevalence of adaptive AKT upregulation early during AKT1 PHD mutants were much more sensitive to limiting

BRAFi therapy would be inversely correlated with the preva- PIP 3 levels (Fig. 2E ). lence of genetic variants that upregulate AKT later during We then hypothesized that BRAFi treatment, by increasing

the acquisition of late resistance. To further understand this PIP3 levels, would provide the necessary upstream signal to potential mechanistic link between early and late resistance, maximally activate AKT1Q79K, enabling the mutant to rescue we hypothesized that a preexisting determinant of AKT acti- the negative effect of PTEN and to amplify the BRAFi-induced vation (e.g., PTEN mutation/expression status) would restrict rebound p-AKT (Thr308) level. Consistent with prior experi- the magnitude of BRAFi-elicited rebound p-AKT levels, but ments (Fig. 1B ), treatment of M229 (low basal p-AKT), M238 the presence of a gain-of-function AKT mutant, for instance, (intermediate basal p-AKT), and WM2664 (high basal p-AKT) would lift this restriction via signal amplifi cation. cells with the BRAFi vemurafenib induced endogenous p-AKT To test this hypothesis, we took advantage of two gain-of- Thr308 in a time-dependent manner (followed up to 48 hours; function AKT mutants (AKT1E17K and AKT1Q79K) identifi ed Fig. 2F ). The strength of this BRAFi-induced p-AKT rebound specifi cally in melanomas with acquired BRAFi resistance level was weakest in BRAF -mutant melanoma cell lines dis- (but not in their patient-matched pretreatment melanomas) playing the lowest basal level of p-AKT and WT PTEN expres- and shown to be capable of conferring BRAFi resistance (9 ). sion, such as M229 cells. In contrast, BRAFi induced strong

72 | CANCER DISCOVERYJANUARY 2014 www.aacrjournals.org

MAPK Inhibition Elicits a Widespread, AKT-Dependent Adaptive Response in Melanoma RESEARCH BRIEF

WM A B D M229 M238 M263 2664 M249 Majority of melanomas Frequencies of core Minority of melanomas pathways of acquired resistance BRAF or Vector E17K Q79K WT

MAPK inhibition Vector F-AKT1 Q79K Vector F-AKT1 Q79K Vector F-AKT1 Q79K Vector F-AKT1 Q79K Vector F-AKT1 Q79K

Flag FLAG Tumor volume p-AKT p-AKT Thr308 Time Thr308 p-AKT p-AKT Ser473 MAPK- Ser473 reactivation

p-ERK AKT AKT Time p-ERK1/2 PI3K–PTEN–AKT- Tubulin activation ERK1/2 p-AKT

Time Tubulin

Adaptive (early) Acquired (late) M249 resistance resistance E Vector PTEN C F-AKT1 F-AKT1 Ly294002 – – ++ Serum – + – + Vector WT E17K Q79K Vector WT E17K Q79K

PTEN WT PHD–GFP p-AKT Thr308

p-AKT Ser473 E17K PHD–GFP AKT

p-ERK1/2

Q79K ERK1/2 PHD–GFP

Tubulin

F M229 M238 WM2664 Vector AKT1Q79K Vector AKT1Q79K Vector AKT1Q79K

0 6 12 24 48 0 6 12 24 48 0 6 12 24 48 0 6 12 24 48 0 6 12 24 48 0 6 12 24 48 Vemu (h) p-AKT Thr308

p-AKT Ser473

AKT

p-ERK1/2

ERK1/2

Tubulin

Figure 2. AKT1Q79K mutant ampliﬁ es BRAFi-induced PI3K–AKT signaling in the presence of PTEN. A, a hypothetical mechanistic link between early, adaptive and late, acquired resistance to BRAF or MAPK pathway inhibition in BRAF -mutant melanomas. Model depicting temporal response patterns of tumor volume and pathway status (p-ERK and p-AKT) to BRAF or MAPK targeting as distinct subsets (majority vs. minority). The frequencies of ERK and AKT activation status early on-treatment are inversely correlated with (and hence may inﬂ uence) the relative frequencies of acquired resistance mechanisms in the two core pathways (shown as pie charts). B, AKT1E17K and AKT1Q79K displayed upregulation of activation-associated phosphorylation. Indicated constructs were transfected into human HEK293T cells, and levels of phospho- and total proteins were probed by Western blotting. Tubulin, loading control. C, AKT1 PHD containing the E17K or Q79K mutations localized to the cell surface independently of serum stimulation (10% FBS, 1 hour), and serum-induced cell-surface localization of mutant PHD (vs. WT PHD) was less sensitive to a PI3K inhibitor (20 μmol/L LY294002, 1 hour). PHD of AKT1 was fused to GFP, expressed stably in M229 cells and visualized (scale bar, 20 μm). D, melanoma cell line stably expressing vector or doxycycline-repressible FLAG-tagged AKT1Q79K (doxycycline withdrawal, 48 hours) were probed for endogenous and exogenous p-AKT and indicated total protein levels. E, stable expression

of PTEN WT and FLAG-AKT1 WT and mutants (vs. vector) in M249 revealed stronger PIP3 signal-amplifying effects of AKT1 PHD mutants with PTEN WT reexpression. F, indicated stable melanoma cell lines were treated with vemurafenib (1 μmol/L) for increasing durations (h) without change of media. Protein lysates were probed for levels of endogenous p-AKT (vector) versus exogenous p-AKT (AKT1Q79K) and indicated phospho- and total protein levels.

JANUARY 2014CANCER DISCOVERY | 73

RESEARCH BRIEF Shi et al.

phosphorylation of exogenous AKT1Q79K in M229 cells, to the lack of PTEN expression correlated with cancer cell line an extent similar to the endogenous p-AKT levels induced by sensitivity to AKT inhibition (15 ). Consistently, PTEN non- BRAFi treatment in WM2664 cells. Thus, consistent with our expressing, BRAF -mutant melanoma cell lines (e.g., WM2664, hypothesis, AKT1Q79K ampliﬁ ed a weak BRAFi-induced signal M249, M397) were exquisitely sensitive to AKT inhibition upstream of AKT. Taken together, BRAFi treatment leads to alone (vs. BRAF or MEK inhibition alone; Fig. 4 and Sup-

rebound RTK upregulation, PIP 3 membrane accumulation, plementary Fig. S4 and S10). Thus, the intrinsic sensitivity PHD membrane recruitment, and AKT activation. In addi- of a subset of melanoma cell lines (those with no PTEN tion, the AKT1Q79K mutant, which can confer acquired BRAFi activity) to AKT inhibition and the BRAFi-induced sensitivity resistance (9 ), produces a signaling phenotype dependent on of another subset of melanoma cell lines (those with PTEN the cell context, which is related to the basal p-AKT level or activity) to AKT inhibition rationalize the combination of PTEN status, and BRAF inhibitor treatment. PI3K–AKT and MAPK pathway inhibition to maximally sup- It is likely then that the AKT1Q79K mutant would confer press melanoma survival signaling. a BRAFi resistance phenotype in a cell context–dependent We further tested the effi cacy of combinatorial PI3K– manner such that the greatest impact would be observed AKT and MAPK pathway targeting using small-molecule where the mutant AKT1 most robustly amplifi es an adaptive inhibitors (either in clinical use or in development) to sup- response. To test this prediction, we treated multiple cell lines press adaptive BRAFi resistance relative to the effi cacies of of varying PTEN genetic and protein expression status (Sup- single target/pathway targeting and dual targeting of the plementary Fig. S4) with either dimethyl sulfoxide (DMSO) MAPK pathway (i.e., BRAFi + MEKi). We measured adap- or increasing vemurafenib concentrations for 3 days ( Fig. 3A , tive escape from drug-induced cell death by long-term clo- left) or 10 days (Fig. 3A , right). From both short- and long- nogenic assays, visualizing and quantifying the growth of term drug treatment regimens, AKT1Q79K expression conferred survival fractions after repeated drug treatments. For drug vemurafenib resistance robustly in M229 cells but weakly in treatments, we selected doses suffi cient to induce complete M238 and WM2664 cells (as well as M263 and M249 cells; on-target inhibition (p-ERK for the MAPK pathway inhibi- Supplementary Fig. S8). In M229 cells, where AKT1Q79K and tors; p-AKT for the PI3K inhibitor BKM120; p-AKT and AKT1E17K expression conferred a two-log increase in BRAFi p-GSK3β for the AKT inhibitors MK2206 and GSK2141795) resistance, the expression of AKT1 WT led to no signifi cant at 1 hour after treatment (Fig. 3D and Supplementary Figs. change in BRAFi sensitivity. PTEN knockdown in M229 cells S11 and S12; data not shown for the AKTi GSK2141795). The increased the basal p-AKT (Thr308) level and further boosted MAPK pathway inhibitors studied include the ATP-compet- the BRAFi-elicited p-AKT level rebound (Fig. 3B ). PTEN itive BRAFi vemurafenib (PLX4032) and the allosteric MEK knockdown in M229 cells conferred resistance to vemurafenib inhibitors AZD6244 and GSK1120212. In these long-term but rendered the AKT1 PHD mutants incapable of conferring survival assays, BRAFi treatment alone (vs. DMSO) was used further vemurafenib resistance (Fig. 3C ). These observations as the “reference” (Fig. 4 ). Dual-target MAPK pathway inhi- are consistent with the notion that WT PTEN activity limits bition (BRAFi + MEKi) was more potent than single-target the BRAFi-elicited adaptive response and that PHD, gain-of- MAPK pathway inhibition, consistent with the clinical data function AKT1/3 mutants counteract this negative effect of demonstrating an improvement in response rate with the WT PTEN to provide survival benefi ts under BRAF- or MAPK- combination of dabrafenib and trametinib compared with inhibited conditions. dabrafenib alone (16 ). Notably, single-target MAPK pathway Given that AKT1Q79K expression in M229 cells amplifi ed inhibition combined with PI3K or AKT inhibition ( Fig. 4A–C BRAFi-induced p-AKT (Thr308) and conferred vemurafenib and Supplementary Fig. S10) led to similarly effi cacious or resistance, we tested whether AKT1Q79K expression would even more profound suppression of clonal growth escape render cell survival more dependent on AKT signaling in than dual MAPK pathway inhibition. This is consistent with the presence of a BRAFi. First, we showed that vemurafenib- the notion that PI3K–AKT signaling provides either mutant induced p-AKT (and the downstream substrate p-GSK3β) BRAF-independent survival or BRAFi-induced compensatory could be downregulated by cotreatment with the AKTi survival. In addition, triple treatment with BRAFi + MEKi + MK2206 in a dose-dependent manner (using 1 hour treatment AKTi led to the most profound suppression of clonogenic to gauge on-target inhibition; Fig. 3D ). M229 cells expressing growth regardless of PTEN expression status (Supplementary AKT1 Q79K exhibited reduced sensitivity to vemurafenib alone Fig. S10). It is important to note that the drug treatment (as compared with M229 vector) but only slightly increased duration of these clonogenic assays mostly measured the slow sensitivity to MK2206 alone (Supplementary Fig. S9). Impor- growth of drug-tolerant persisters (Supplementary Fig. S5A). tantly, the combination of BRAFi + AKTi was more effective In summary, we have shown that melanoma therapy based than either agent alone in reducing the clonogenic growth on MAPK pathway suppression can frequently unleash a of M229 cells expressing AKT1Q79K ( Fig. 3E ), consistent with rebound increase in PI3K–AKT pathway signaling in the BRAF inhibition leading to an increase in MAPK-redundant, tumors of treated patients. In cell lines, this signal crosstalk AKT-dependent survival. In additional PTEN-expressing, varied in strength, and one factor limiting the extent of a BRAF-mutant melanoma cell lines (e.g., M238, M395; Fig. 4 BRAFi-induced rebound in PI3K–AKT signaling was PTEN and Supplementary Fig. S4 and S10) that are relatively insen- expression or activity. Interestingly, a novel PHD mutant sitive to an AKTi alone (compared with a BRAFi or MEKi of AKT1 (Q79K), which was detected in melanoma with alone), the combination of BRAFi + AKTi (or MEKi + AKTi) acquired (late) BRAFi resistance, could counteract PTEN’s was also more effective than either agent alone in reducing negative effect, amplify this BRAFi-elicited rebound level of the clonogenic growth. Previous studies have suggested that AKT activation (by enhancing its cell-surface recruitment

74 | CANCER DISCOVERYJANUARY 2014 www.aacrjournals.org

MAPK Inhibition Elicits a Widespread, AKT-Dependent Adaptive Response in Melanoma RESEARCH BRIEF

A Vemu (μmol/L) 0 0.01 0.1 1 10

Vector

M229 Vector 100 AKT1 WT 80 AKT1Q79K 60 WT 40

% Survival 20 M229 0 1 0.1 10 E17K 0.01 0.001 [vemu]

Q79K Vector 100 M238 AKT1Q79K 80 60 40 Vector

% Survival 20 0 M238 1 0.1 10 0.01 0.001 [vemu] Q79K

WM2664 Vector 100 AKT1Q79K 80 60 Vector 40

% Survival 20 WM2664 0 1 0.1 10 Q79K 0.01 0.001 [vemu]

M229 B D Q79K E M229 M229 Vector shPTEN M229 AKT1 Vector Q79K 0612 24 48 0612 24 48 Vemu (h) 0 0 0.01 0.11 10 MK2206 (μmol/L) – +++++ Vemu (48 h) PTEN DMSO p-AKT Thr308 p-AKT 100% 100% Thr308 p-AKT Ser473 AKT Vemu

AKT 62% 86% p-ERK1/2 p-GSK3-β Ser9 MK2206 ERK1/2

GSK3-β 66% 49% Tubulin Vemu Tubulin + MK2206

C 49% 28% M229 100 shVector shPTEN 80 Vector Vector 60 AKT1 WT AKT1 WT Q79K Q79K 40 AKT1 AKT1

% Survival E17K E17K 20 AKT1 AKT1 0 1 0 0.1 1 0.01 0.001 [vemu]

Figure 3. AKT PHD mutants confer vemurafenib resistance in a PTEN context-dependent manner. A, indicated stable melanoma cell lines were withdrawn from doxycycline (48 hours), treated with indicated concentrations of vemurafenib (vemu) for 72 hours, and survival (relative to DMSO-treated controls; mean ± SEM, n = 5) measured by the MTT (left; dashed line, 50% inhibition) or clonogenic assays (right). Media and vemurafenib at indicated concentration were replenished every 2 days for 10 days, starting 2 days after doxycycline withdrawal. B, the levels of vemurafenib-induced p-AKT in M229 cells without or with PTEN knockdown as shown by Western blotting. C, the effects of AKT1 PHD–mutant expression on vemurafenib sensitivity (MTT) in the context of PTEN WT expression or its knockdown. D, M229 cells stably expressing AKT1Q79K were treated with DMSO or vemurafenib (1 μmol/L, 48 hours) with and without increasing concentrations of the AKT1/2/3 inhibitor MK2206 (1 hour), and the lysates probed for the indicated phospho- and total protein levels. E, M229 cells stably expressing vector or AKT1Q79K were withdrawn from doxycycline (48 hours), seeded, treated with at vemurafenib (1 μmol/L), and/or MK2206 (5 μmol/L) every 2 days (four treatments, 9 day drug exposure), ﬁ xed/stained, and quantiﬁ ed (% survival relative to DMSO).

JANUARY 2014CANCER DISCOVERY | 75

RESEARCH BRIEF Shi et al.

A Single inhibitor Dual inhibitors Dual MAPK inhibiton PI3Ki and MAPK inhibition Figure 4. Suppression of clonogenic BRAFi BRAFi PI3Ki PI3Ki PI3Ki melanoma growth via cotargeting of the + + + + + DMSO BRAFi MEKi-AZD MEKi-GSK PI3Ki MEKi-AZD MEKi-GSK BRAFi MEKi-AZD MEKi-GSK PI3K–AKT and MAPK pathways regardless of PTEN status. A, M229 vector or M229 AKT1Q79K stable cell lines (2 days M229 Vector after doxycycline washout), M238, and WM2664 were treated with DMSO or indicated inhibitor(s) every day. BRAFi M229 Q79K (vemurafenib), MEKi-AZD6244, and PI3Ki were used at 1 μmol/L (except for WM2664, 0.4 μmol/L). MEKi-GSK was μ

M238 used at 0.01 mol/L (except for WM2664, 0.004 μmol/L). All cultures were ﬁ xed and stained with crystal violet (results shown are representative of two experiments).

WM2664 B, BRAFi (vemurafenib), MEKi-GSK (GSK1120212), and AKTi (MK2206) Single inhibitor Dual inhibitors μ B D RTKs were used at 1, 0.01, and 5 mol/L for Dual MAPK AKT + MAPK M229 vector and M229 AKT1Q79K cells and inhibiton inhibiton PTEN PI3K BRAF CRAF at 0.5, 0.005, and 2.5 μmol/L for M238, BRAFi AKTi AKTi and WM2664 cells. C, quantiﬁ cation of + + + DMSO BRAFi MEKi-GSK AKTi MEKi-GSK BRAFi MEKi-GSK AKT MEK clonogenic growth in A and B expressed as percentage growth inhibition (relative ERK to DMSO). MAPK inhibition (at BRAF M229 Vector or MEK) indicated in gray; PI3K or AKT inhibition in red. D, model showing that

RTKs BRAFi treatment of BRAF -mutant M229 Q79K melanoma can lead to a context-dependent, PTEN PI3K BRAF CRAF adaptive RTK–PI3K–AKT upregulation. BRAFi (or MAPKi) treatment derepresses AKT MEK

M238 RTK upregulation, resulting in activation of PI3K–AKT, and CRAF. With WT PTEN ERK activity, this BRAFi-induced, rebound AKT upregulation is weak. In this context,

WM2664 AKT-activating mutants can counteract

RTKs the effect of WT PTEN and amplify PIP3 C signaling. Brown, mutated; blue, wild-type. PTEN PI3K BRAF CRAF BRAF PI3K MEK AKT BRAF PI3K AKT MEK 80 M229 vector MEK AKT

70 60 M229 vector 50 ERK 60 40 30 50 20 90 10 RTKs M229 Q79K 65 70 55 M229 Q79K PTEN PI3K BRAF CRAF 45 50 35 25 AKT MEK 30 15 5 95 ERK M238 90 M238 80 80 % Growth inhibition % Growth % Growth inhibition % Growth 70 65 60 50 50 85 95 WM2664 WM2664 75 70 55 55 35 40 15

despite limiting PIP3 levels), and confer BRAFi resistance DISCUSSION (Fig. 4D ). This BRAFi-induced, PI3K–AKT-dependent adaptive response, along with the intrinsic sensitivity of PTEN Strategies to deter acquired BRAFi resistance in melanoma nonexpressing, BRAF-mutant melanoma cells to AKT inhibi- have been based on the frequent occurrence of mechanisms tion, helped to explain the efﬁ cacy of combined PI3K–AKT that reactivate the MAPK pathway or the phenomenon of and MAPK pathway targeting. Taken together, these ﬁ ndings drug addiction displayed by the tumor cells with acquired provide key insights into the common clinical observation of resistance. These strategies have been either validated clini- partial initial tumor responses to BRAFi or BRAFi + MEKi cally or proposed to be tested clinically by the combination therapy and a strong rationale for combinatorial, upfront of BRAFi + MEKi or its intermittent dosing, respectively (16, targeting of both the PI3K–AKT and MAPK melanoma sur- 17). However, although the combination of BRAFi + MEKi vival pathways. likely achieves more profound pathway inhibition (compared

76 | CANCER DISCOVERYJANUARY 2014 www.aacrjournals.org

MAPK Inhibition Elicits a Widespread, AKT-Dependent Adaptive Response in Melanoma RESEARCH BRIEF

with BRAFi or MEKi alone) and hence a higher rate of initial reprogramming of growth and survival signaling. Our data clinical responses, the prolonged durability of response is indicate that this AKT upregulation is likely accompanied by still cut short by the late acquisition of resistance, strongly a certain degree of MEK–ERK reactivation driven by CRAF, suggesting that a strategy based solely on MAPK pathway which is relatively much weaker compared with that driven inhibition would be missing another essential melanoma by, for instance, NRAS mutations detected later during late or survival pathway. This is even more evident when patients acquired resistance. This weaker form of MAPK reactivation who progressed on vemurafenib were then treated with the occurring during the adaptive phase likely also constrains the combination of BRAFi + MEKi; the secondary response rates initial effi cacy of BRAFi and results from BRAFi-induced loss were low and, if seen, the durability short ( 18 ). This work on of ERK-dependent negative feedback, which normally sup- adaptive BRAFi resistance and our recent study of the land- presses ligand/RTK-driven RAS/MAPK signaling via BRAFi- scape of acquired BRAFi resistance mechanisms ( 9 ) strongly insensitive RAF dimers ( 11 ). Data presented here and our support upregulation of the PI3K–AKT pathway as a critical unpublished work show that certain RTKs (e.g., PDGFRβ, event during the early and late evolution of resistance to EGFR) are overexpressed and ligand-stimulated in a specifi c MAPK pathway inhibition in patients. temporal order along the evolutionary continuum of adap- Early and late resistance could potentially be mechanisti- tive resistance, spanning from an early period of maximal cell cally linked. In other words, the effects of BRAF inhibition on death induction, to an (overlapping) phase characterized by melanoma signaling may infl uence and shape the evolution- a surviving but slow-cycling subpopulation of drug-tolerant ary selective pressure on the residual tumor cells. We propose persisters (DTP), and then another transition marked by that, just as potent mutant BRAF inhibition provides strong renewed proliferative clonal escape (i.e., drug-tolerant pro- selective pressure for MAPK reactivation, a potent BRAFi- liferating persisters or DTPP). These transitions could be induced AKT upregulation would attenuate selective pres- marked by signifi cant cellular morphologic and gene expres- sure for gain-of-function lesions in the PI3K–AKT pathway. sion alterations (5 ). Subsequent to the DTPP stage, further On the basis of our data, we surmise that BRAFi-induced clonal outgrowth can occur due to enhanced growth and pro- AKT upregulation during the fi rst month of treatment may liferative fi tness driven by specifi c genetic variants (e.g., NRAS be widespread as an adaptive response. In those tumors in mutations, mutant BRAF amplifi cation, AKT1/3 mutations; which this adaptive response was attenuated (e.g., by wild- ref. 9 ). Thus, late acquired resistance mechanisms not only type PTEN activity), genetic alterations upregulating the mirror but also augment adaptive resistance mechanisms. PI3K–PTEN–AKT pathway would distinctly confer growth Melanoma therapeutics has entered the era of combinato- and/or survival advantage. In fact, in the setting of acquired rial approaches. It has become evident recently that about or late BRAFi resistance, our data support the notion that 20% of BRAF -mutant metastatic melanomas harbor readily BRAF inhibition can select for amplifi cation of the early screenable genetic alterations that upregulate the PI3K–PTEN– adaptive response driven by stable, genetic alterations, lead- AKT pathway. Our studies show that melanomas can adap- ing to dramatically enhanced survival dependency on the tively upregulate the PI3K–PTEN–AKT pathway early during PI3K–PTEN–AKT pathway. MAPK–targeted therapy to compensate for MAPK pathway In this context, recent studies ( 19-21 ) and our own unpub- inhibition and, with further evolutionary selection, acquire lished data place partial or complete genetic inactivation of genetic lesions to further enhance PI3K–AKT signaling for PTEN at 10%–30% of BRAF -mutant melanoma tumors or growth and survival. Thus, upfront, combinatorial targeting of short-term cultures. It is not yet clear how frequently loss of both the PI3K–PTEN–AKT pathway and the MAPK pathway PTEN function occurs as a result of epigenetic (DNA meth- would be expected to curtail innate (lack of initial responses; ylation, small non-coding RNAs, etc.), transcriptional, and ref. 23 ), adaptive (limited initial responses), and acquired (ces- posttranslational modifi cation mechanisms in BRAF -mutant sation of responses) resistance to MAPK-targeted therapies. melanomas. Presumably, the subgroup of BRAF -mutant melanomas with wild-type and functional PTEN expression (i.e., the low p-AKT cohort) would mount the weakest AKT- METHODS dependent adaptive response when treated with a MAPK Cell Culture, Infections, and Drug Treatments pathway inhibitor. Interestingly, recent clinical data suggest Cells were maintained in Dulbecco’s Modifi ed Eagle Medium with that BRAF -mutant, p-AKT–low melanomas are more likely to 10% FBS and glutamine. The melanoma cell lines were established regress in response to treatment with the MEKi selumetinib at UCLA with Institutional Review Board approval and routinely (AZD6244) than BRAF -mutant, p-AKT–high melanomas authenticated by mitochondrial DNA sequencing. A375, SK, and ( 22 ). The relative lack of responses of the BRAF- mutant, WM cell lines were obtained from MSKCC and the Wistar Institute p-AKT–high melanomas may be explained by redundant via Material Transfer Agreements and were not further authenticated survival pathways and/or a robust, AKT-dependent adaptive except for verifi cation of the BRAF -mutant status. WT PHD, E17K response. Another prediction from the current study is that PHD, and Q79K PHD of AKT1 fused to GFP as well as FLAG-tagged E17K Q79K the subgroup of BRAF -mutant, wild-type PTEN-expressing full-length AKT1, AKT1 , and AKT1 were subcloned into the doxycycline-repressible lentiviral vector pLVX-Tight-Puro (Clontech, melanomas, when treated with a MAPK-targeted therapy, Inc.); viral supernatants generated by cotransfection with three pack- would be susceptible to the development of acquired resist- aging plasmids into HEK293T cells; and infections carried out with ance driven by genetic amplifi ers of the PI3K–PTEN–AKT protamine sulfate. Stocks and concentrations of small-molecule pathway. kinase inhibitors were made in DMSO. Cells were quantifi ed using The AKT-dependent adaptive melanoma response is likely CellTiter-GLO Luminescence (Promega) or crystal violet staining fol- an early event in a set or series of coordinated, stereotypic lowed by NIH Image J quantifi cation.

JANUARY 2014CANCER DISCOVERY | 77

RESEARCH BRIEF Shi et al.

Drug Sensitivity, Protein, and PIP3 Detection Bethesda, MD). Stained immunohistochemical slides were scanned × Cell proliferation experiments were performed in a 96-well format with Scan Scope CS section scanner (Aperio) with 40 magnifi cation, (5 replicates) and drug treatments initiated at 24 hours after seeding and analyzed by Tissue Studio 2.0 image analysis software (Defi n- for 72 hours. Stocks and dilutions of PLX4032 (Plexxikon), LY294002, iens). Briefl y, tumor-rich regions were selected as regions of interest MK2206, AZD6244, BKM120, GSK1120212, GSK2141795, and (ROI), and the software was trained to recognize tumor cells by GSK2118436 (Selleck Chemicals) were made in DMSO. In these hematoxylin and DAB intensity, nuclear size, and nuclear morphol- short-term assays, surviving cells were quantifi ed using CellTiter- ogy. Cell morphology was simulated per parameter 5. GLO Luminescence (Promega) following the manufacturer’s rec- Disclosure of Potential Confl icts of Interest ommendations. Clonogenic assays were performed by plating cells at single-cell density in 6-well plates and providing fresh media, R.A. Scolyer has received honoraria from the speakers’ bureau of doxycycline (if applicable), and vemurafenib (vs. DMSO) either every Roche and is a consultant/advisory board member of GlaxoSmith day or every other day. In these long-term assays, surviving clo- Kline. R.F. Kefford has received honoraria from the speakers’ nogenic colonies were fi xed by 4% paraformaldehyde and stained bureaus of Roche and Novartis and is a consultant/advisory board with 0.05% crystal violet. Cell lysates for Western blotting were member of Roche, GlaxoSmithKline, and Novartis. A. Ribas has made in radioimmunoprecipitation assay buffer (Sigma) with pro- ownership interest (including patents) in Entrogen. G.V. Long is a tease (Roche) and phosphatase (Santa Cruz Biotechnology) inhibitor consultant/advisory board member of GlaxoSmithKline and Roche. cocktails. Western blots were probed with antibodies against p-AKT R.S. Lo is an inventor on a patent. No potential confl icts of interest (Ser473), p-AKT (Thr308), AKT, p-GSK3β (Ser9), GSK3β, p-ERK1/2 were disclosed by the other authors. (T202/Y204), ERK1/2, PTEN, c-MYC, PDGFRβ, p-CRAF (Ser338), CRAF, EGFR, GFP (Cell Signaling Technology), and FLAG, Tubulin Authors’ Contributions (Sigma). Samples in the same fi gure subpanels were run, transferred, Conception and design: H. Shi, A. Hong, R.F. Kefford, R.S. Lo blotted, and developed (with the same exposure time) together. Development of methodology: H. Shi, X. Kong, R.C. Koya, T. Chodon, In IHC experiments, after deparaffi nization and rehydration, all R.F. Kefford, R.S. Lo sections from tissue series with extensive microscopic or macroscopic Acquisition of data (provided animals, acquired and man- melanin deposits in any tumor were fi rst subjected to a bleaching aged patients, provided facilities, etc.): H. Shi, A. Hong, X. Kong, ° step using 3% H 2O 2 for 2 hours at 55 C. Tissue sections were then R.C. Koya, C. Song, G. Moriceau, C.C. Yu, C. Ng, T. Chodon, antigen-retrieved with a vegetable steamer at 95°C for 30 minutes, R.A. Scolyer, R.F. Kefford, A. Ribas, G.V. Long, R.S. Lo and immunostaining with anti–p-AKT (Ser473) or anti-PDGFRβ Analysis and interpretation of data (e.g., statistical analysis, (Cell Signaling Technology) was then performed followed by a stand- biostatistics, computational analysis): H. Shi, A. Hong, X. Kong, ard streptavidin–biotin complex technique with horseradish per- C. Song, G. Moriceau, W. Hugo, R.F. Kefford, G.V. Long, R.S. Lo oxidase (HRP) and DAB chromogen (Vector labs). After mounting, Writing, review, and/or revision of the manuscript: H. Shi, R.C. Koya, stained slides were scanned in their entirety with Scan Scope CS sec- R.A. Scolyer, R.F. Kefford, G.V. Long, R.S. Lo tion scanner (Aperio) at ×40 magnifi cation, and the images analyzed Administrative, technical, or material support (i.e., reporting

by Tissue Studio 2.0 (Defi niens). PIP3 measurement was performed or organizing data, constructing databases): H. Shi, X. Kong, by competitive PIP3 Mass ELISA (Echelon K-2500). Briefl y, the acidic W. Hugo, C.C. Yu, G.V. Long, R.S. Lo lipids of each sample were extracted by the TCA/chloroform/metha- Study supervision: R.S. Lo nol method, and samples were normalized on the basis of protein concentration. Cellular PI(3,4,5)P3 quantities were calculated by com- Acknowledgments paring the values from the wells containing PI(3,4,5)P extraction 3 The authors thank P. Lin and G. Bollag (Plexxikon Inc.) for provid- products to the values in the standard curve. ing PLX4032, A. Villanueva (UCLA) for clinical data management, and all patient volunteers. Fluorescent Microscopy Melanoma cell lines expressing GFP–AKT PHD or PHD mutants Grant Support were cultured in 6-well plates over cover slides. Serum-starved cells were treated with serum or a small-molecule inhibitor for indicated R. S. Lo is supported by a Stand Up To Cancer Innovative Research durations. Cells were fi xed with 4% paraformaldehyde and mounted Grant, a Program of the Entertainment Industry Foundation by VECTASHIELD Mounting Media (Vector Lab). GFP signal was (SU2C-AACR-IRG0409). Additional funding for R.S. Lo came from photographed with a Zeiss microscope (AXIO Imager A1) mounted the National Cancer Institute (K22CA151638, 1R01CA176111), with a charge-coupled device camera (Retiga EXi QImaging), and the Burroughs Wellcome Fund, Melanoma Research Alliance, American images captured by Image-pro plus 6.0. Skin Association, Sidney Kimmel Foundation for Cancer Research, Eli & Edythe Broad Center of Regenerative Medicine & Stem Cell Apoptosis Analysis Research, Harry J. Lloyd Charitable Trust, National Center for Advancing Translational Sciences UCLA CTSI Grant UL1TR000124, Cultures of indicated melanoma cell lines were treated with and the Ian Copeland Melanoma Fund, and from the Wesley Coyle PLX4032 (at different time points) for increasing durations of time, Memorial Fund, The Seaver Institute, and National Cancer Institute fi xed (at the same time point), permeablized, and treated with (1P01CA168585; to R.S. Lo and A. Ribas). Postdoctoral fellowship RNase (QIAGEN). Cells were then stained with Annexin V-V450 and funding came from the American Association of Cancer Institutes, β anti-PDGFR FITC (BD Pharmingen) for 15 minutes at room tem- the American Skin Association (to H. Shi), and the California Insti- perature and mixed with 7-AAD before sample loading (LSR II Flow tute for Regenerative Medicine (G. Moriceau). R.A. Scolyer, R.F. Cytometry, BD Bioscience). Flow cytometry data were analyzed by the Kefford, and G.V. Long are supported by the National Health and FACS Express V2 software. Medical Research Council of Australia, Translational Research Pro- gram of the Cancer Institute New South Wales. Data Quantifi cation Clonogenic assays were stained with 0.05% crystal violet and Received June 11, 2013; revised October 21, 2013; accepted Octo- photographed, and colonies were quantifi ed by NIH Image J (NIH, ber 29, 2013; published OnlineFirst November 21, 2013.

78 | CANCER DISCOVERYJANUARY 2014 www.aacrjournals.org

MAPK Inhibition Elicits a Widespread, AKT-Dependent Adaptive Response in Melanoma RESEARCH BRIEF

12. Shi H , Kong X , Ribas A , Lo RS . Combinatorial treatments that over- REFERENCES come PDGFRβ-driven resistance of melanoma cells to B-RAF(V600E) 1. Davies H , Bignell GR , Cox C , Stephens P , Edkins S , Clegg S , et al. inhibition. Cancer Res 2011 ; 71 : 5067 – 74 . Mutations of the BRAF gene in human cancer . Nature 2002 ; 417 : 13. Lo RS . Receptor tyrosine kinases in cancer escape from BRAF inhibi- 949 – 54 . tors. Cell Res 2012 ; 22 : 945 – 7 . 2. Chapman PB , Hauschild A , Robert C , Haanen JB , Ascierto P , Lark in 14. Carpten JD , Faber AL , Horn C , Donoho GP , Briggs SL , Robbins CM , J , et al. Improved survival with vemurafenib in melanoma with BRAF et al. A transforming mutation in the pleckstrin homology domain of V600E mutation . N Engl J Med 2011 ; 364 : 2507 – 16 . AKT1 in cancer. Nature 2007 ; 448 : 439 – 44 . 3. Sosman JA , Kim KB , Schuchter L , Gonzalez R , Pavlick AC , Weber J S, 15. Vasudevan KM , Barbie DA , Davies MA , Rabinovsky R , McNear CJ , et al. Survival in BRAF V600-mutant advanced melanoma treated Kim JJ , et al. AKT-independent signaling downstream of oncogenic with vemurafenib. N Engl J Med 2012 ; 366 : 707 – 14 . PIK3CA mutations in human cancer. Cancer Cell 2009 ; 16 : 21 – 32 . 4. Lo RS . Combinatorial therapies to overcome B-RAF inhibitor resist- 16. Flaherty KT , Infante JR , Daud A , Gonzalez R , Kefford RF , Sosman J , ance in melanomas. Pharmacogenomics 2012 ; 13 : 125 – 8 . et al. Combined BRAF and MEK inhibition in melanoma with BRAF 5. Nazarian R , Shi H , Wang Q , Kong X , Koya RC , Lee H , et al. Melano- V600 mutations. N Engl J Med 2012 ; 367 : 1694 – 703 . mas acquire resistance to B-RAF(V600E) inhibition by RTK or N-RAS 17. Das Thakur M , Salangsang F , Landman AS , Sellers WR , Pryer NK , upregulation. Nature 2010 ; 468 : 973 – 7 . Levesque MP , et al. Modelling vemurafenib resistance in melanoma 6. Poulikakos PI , Persaud Y , Janakiraman M , Kong X , Ng C , Moriceau reveals a strategy to forestall drug resistance. Nature 2013 ; 494 : 251 – 5 . G , et al. RAF inhibitor resistance is mediated by dimerization of aber- 1 8. Lo RS , Ribas A , Long GV , Ballotti R , Berger M , Hugo W , et al. Meet- rantly spliced BRAF(V600E). Nature 2011 ; 480 : 387 – 90 . ing report from the Society for Melanoma Research 2012 Congress, 7. Shi H , Moriceau G , Kong X , Koya RC , Nazarian R , Pupo GM , et Hollywood, California. Pigment Cell Melanoma Res 2013 ;26:E1-7. al. Preexisting MEK1 exon 3 mutations in V600E/KBRAF melano- 19. Hodis E , Watson IR , Kryukov GV , Arold ST , Imielinski M , Theurillat JP , mas do not confer resistance to BRAF inhibitors . Cancer Discov et al. A landscape of driver mutations in melanoma. Cell 2012 ; 150 : 2012 ; 2 : 414 – 24 . 251 – 63 . 8. Shi H , Moriceau G , Kong X , Lee MK , Lee H , Koya RC , et al. Melanoma 20. Krauthammer M , Kong Y , Ha BH , Evans P , Bacchiocchi A , McCusker whole-exome sequencing identifi es (V600E)B-RAF amplifi cation-medi- JP , et al. Exome sequencing identifi es recurrent somatic RAC1 muta- ated acquired B-RAF inhibitor resistance. Nat Commun 2012 ; 3 : 724 . tions in melanoma. Nat Genet 2012 ; 44 : 1006 – 14 . 9. Shi H , Hugo W , Kong X , Hong A , Koya RC , Moriceau G , e t al. 21. Shull AY , Latham-Schwark A , Ramasamy P , Leskoske K , Oroian D , Acquired resistance and clonal evolution in melanoma during BRAF Birtwistle MR , et al. Novel somatic mutations to PI3K pathway genes inhibitor therapy. Cancer Discov 2014 ;4:80–93 . in metastatic melanoma. PLoS ONE 2012 ; 7 : e43369 . 10. Duncan JS , Whittle MC , Nakamura K , Abelll AN , Midland AA , 22. Catalanotti F , Solit DB , Pulitzer MP , Berger MF , Scott SN , Iyriboz T , Zawistowski JS , et al. Dynamic reprogramming of the kinome in et al. Phase II trial of MEK inhibitor selumetinib (AZD6244, ARRY- response to targeted MEK inhibition in triple-negative breast cancer. 142886) in patients with BRAFV600E/K-mutated melanoma. Clin Cell 2012 ; 149 : 307 – 21 . Cancer Res 2013 ; 19 : 2257 – 64 . 11. Lito P , Pratilas CA , Joseph EW , Tadi M , Halilovic E , Zubrowski M , 23. Trunzer K , Pavlick AC , Schuchter L , Gonzalez R , McArthur GA , et al. Relief of profound feedback inhibition of mitogenic signaling Hutson TE , et al. Pharmacodynamic effects and mechanisms of by RAF inhibitors attenuates their activity in BRAFV600E melano- resistance to vemurafenib in patients with metastatic melanoma . mas. Cancer Cell 2012 ; 22 : 668 – 82 . J Clin Oncol 2013 ; 31 : 1767 – 74 .

JANUARY 2014CANCER DISCOVERY | 79

A Novel AKT1 Mutant Amplifies an Adaptive Melanoma Response to BRAF Inhibition

Hubing Shi, Aayoung Hong, Xiangju Kong, et al.

Cancer Discovery 2014;4:69-79. Published OnlineFirst November 21, 2013.

Updated version Access the most recent version of this article at: doi:10.1158/2159-8290.CD-13-0279

Supplementary Access the most recent supplemental material at: Material http://cancerdiscovery.aacrjournals.org/content/suppl/2013/11/21/2159-8290.CD-13-0279.DC1 http://cancerdiscovery.aacrjournals.org/content/suppl/2021/03/02/2159-8290.CD-13-0279.DC2

Cited articles This article cites 23 articles, 5 of which you can access for free at: http://cancerdiscovery.aacrjournals.org/content/4/1/69.full#ref-list-1

Citing articles This article has been cited by 26 HighWire-hosted articles. Access the articles at: http://cancerdiscovery.aacrjournals.org/content/4/1/69.full#related-urls

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

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

Permissions To request permission to re-use all or part of this article, use this link http://cancerdiscovery.aacrjournals.org/content/4/1/69. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.