Mutation-Specific RAS Oncogenicity Explains NRAS Codon 61 Selection in Melanoma

Published OnlineFirst September 24, 2014; DOI: 10.1158/2159-8290.CD-14-0729

Research Article

Christin E. Burd1,2, Wenjin Liu3,4, Minh V. Huynh5, Meriam A. Waqas1, James E. Gillahan1,2, Kelly S. Clark3,4, Kailing Fu3,4, Brit L. Martin1, William R. Jeck3,4,. George P Souroullas3,4, David B. Darr3,4, Daniel C. Zedek6, Michael J. Miley7, Bruce C. Baguley8, Sharon L. Campbell4,5, and Norman E. Sharpless3,4

Abstrt ac NRAS mutation at codons 12, 13, or 61 is associated with transformation; yet, in melanoma, such alterations are nearly exclusive to codon 61. Here, we compared the melanoma susceptibility of an NrasQ61R knock-in allele to similarly designed KrasG12D and NrasG12D alleles. With concomitant p16INK4a inactivation, KrasG12D or NrasQ61R expression efficiently promoted melanoma in vivo, whereas NrasG12D did not. In addition, NrasQ61R mutation potently cooperated with Lkb1/Stk11 loss to induce highly metastatic disease. Functional comparisons of NrasQ61R and NrasG12D revealed little difference in the ability of these proteins to engage PI3K or RAF. Instead, NrasQ61R showed enhanced nucleotide binding, decreased intrinsic GTPase activity, and increased stability when compared with NrasG12D. This work identifies a faithful model of human NRAS-mutant melanoma, and suggests that the increased melanomagenecity of NrasQ61R over NrasG12D is due to heightened abundance of the active, GTP-bound form rather than differences in the engagement of downstream effector pathways.

SIGNIFICANCE: This work explains the curious predominance in human melanoma of mutations of codon 61 of NRAS over other oncogenic NRAS mutations. Using conditional “knock-in” mouse models, we show that physiologic expression of NRASQ61R, but not NRASG12D, drives melanoma formation. Cancer Discov; 4(12); 1–10. ©2014 AACR.

1Department of Molecular Genetics, The Ohio State University, Columbus, Note: Supplementary data for this article are available at Cancer Discovery Ohio. 2Department of Molecular and Cellular Biochemistry, The Ohio State Uni- Online (http://cancerdiscovery.aacrjournals.org/). 3 versity, Columbus, Ohio. Department of Genetics, University of North Carolina Corresponding Author: Norman E. Sharpless, The Lineberger Compre- 4 School of Medicine, Chapel Hill, North Carolina. The Lineberger Comprehen- hensive Cancer Center, University of North Carolina School Medicine, CB sive Cancer Center, University of North Carolina School of Medicine, Chapel #7295, Chapel Hill, NC 27599-7295. Phone: 919-966-1185; Fax: 919- 5 Hill, North Carolina. Department of Biochemistry and Biophysics, University 966-8212; E-mail: [email protected] of North Carolina School of Medicine, Chapel Hill, North Carolina. 6Department of Dermatology, University of North Carolina School of Medicine, Chapel Hill, doi: 10.1158/2159-8290.CD-14-0729 North Carolina. 7Department of Pharmacology, University of North Carolina ©2014 American Association for Cancer Research. School of Medicine, Chapel Hill, North Carolina. 8Auckland Cancer Society Research Centre, University of Auckland, Auckland, New Zealand.

OF1 | CANCER DISCOVERY December 2014 www.aacrjournals.org

INTRODUCTION mediators of mutation-specific therapeutic response (14). Together, these results suggest that distinct, codon-specific One third of all human cancers harbor activating KRAS, properties of RAS mutations have important clinical and HRAS, or NRAS mutations, which localize predominantly to biologic implications. codons 12, 13, or 61 (1, 2). RAS proteins function as canonical Cancers display tissue-specific preferences for mutation of GTPase switches, binding to effectors in the presence of GTP the RAS homologs (Supplementary Table S1). In melanoma, and activating downstream signaling pathways to influence NRAS is by far the most frequently mutated RAS isoform, cellular proliferation, differentiation, and survival. The return and notably, 84% of these mutations localize to codon 61 of RAS to an inactive, GDP-bound state is catalyzed by GTPase- versus only 7% to glutamine 12 (Supplementary Table S1). A activating proteins (GAP), which stimulate the weak, intrinsic similar preference for codon 61 mutations is noted in thyroid GTPase activity of these proteins. Mutations at codons 12 or cancer, but is not observed in other cancer types. Codon 12 13 render RAS proteins insensitive to GAP activity, resulting and 13 mutations constitute more than 90% of KRAS muta- in constitutive, oncogenic signaling (3). Similarly, mutation of tions observed in human colon, pancreatic, lung, and ovarian Q61, a catalytic residue required for efficient GTP hydrolysis, cancers (Supplementary Table S1 and ref. 1). Likewise, glycine impedes the return of RAS to an inactive GDP-bound state (4). 12 is the most common site of NRAS mutation in acute mye- Historically, RAS proteins with codon 12, 13, or 61 altera- loid leukemia (Supplementary Table S1). The mechanistic tions have been considered oncogenic equivalents; however, basis for codon 61 selection in melanoma and thyroid cancer recent clinical observations suggest functional differences for is unclear. Some have suggested that cytosine to thymidine each RAS mutation. For example, in colorectal cancer, KRAS transversions caused by ultraviolet (UV) light may explain mutational status is used as a prognostic indicator of resist- the preference for certain mutations in melanoma, but the ance to therapy with EGFR antibodies (e.g., cetuximab; refs. majority of codon 61 mutations do not exhibit a characteris- 5–8). Retrospective analyses, however, of “all-comer” trials tic UV-damage signature (15). Alternatively, it is possible that suggest mutational specificity in this regard: patients harbor- codon mutation preferences reflect differences in oncogenic ing KRAS codon 13 mutations appear to benefit from cetuxi- signaling. mab therapy, whereas those with codon 12 mutations were Comparing the oncogenic potential of various RAS unresponsive (9–13). Moreover, progression-free survival on mutants is challenging for several reasons. RAS gene dosage targeted therapies may also be codon-specific in non–small clearly influences downstream signaling, and artifacts of RAS cell lung cancer (NSCLC; ref. 14). Here, molecular modeling overexpression are well described. Likewise, endomembrane and reverse-phase protein analysis pinpointed differential localization is critical for physiologic RAS signal transduc- effector engagement and downstream signaling as potential tion (16), and may not be adequately recapitulated using

December 2014 CANCER DISCOVERY | OF2

RESEARCH ARTICLE Burd et al.

A G12D LSL-Nras Figure 1. Activation of LSL-NrasQ61R or LSL-NRASG12D (*G12D) slows melanocyte proliferation. A, diagrammatic representa- tion of the LSL-NrasG12D (17) and LSL-NrasQ61R alleles in the 0 PUROSA 3 STOP SD 1 2 3 4 5 6 ë presence or absence of CRE recombinase. Green triangles (*G12D) denote lox P sites and unnumbered black boxes represent 0 1 2 3 4 5 6 FRT sites. Blue boxes indicate the location of a puromy- + CRE cin (PURO) resistance cassette. SA, splice acceptor; SD, splice donor; 3× STOP, transcription and translational stop LSL-NrasQ61R sequence. B, PCR genotyping of melanocytes treated for 6 days with ethanol vehicle (E; lane 7) or 1.0 µmol/L 4-OHT (*Q61R) (lane 7). DNA from mice of the indicated genotypes is 0 SA 4ë STOP 1 2 3 4 5 6 included as a control (lanes 3–5). WT, wild-type. C, melanocytes cultured as in B were incubated with EdU, a thymidine (*Q61R) analogue, for 16 hours and then analyzed by flow cytometry. EdU incorporation in vehicle-treated cells was set to 100%. + CRE 0 1 2 3 4 5 6 Each dot represents a biologic replicate with the mean and standard error of the mean indicated by black lines. Concen- trations of 4-OHT are provided in µmol/L. B O 2 MW H WT/WT LSL/WT LSL/LSL LSL/LSL LSL/LSL +CRE - WT G12D - LSL +CRE - WT Q61R - LSL 4-OHT: –– –– E 1.0 Lane: 1 2 3 4 5 6 7

C 120

100

60 12D/12D TpN 40 61R/61R TpN 20 No CRE Percentage of change in EdU 0 0 0.5 1.0 0 0.5 1.0 1.0

[4-OHT] exogenous protein expression. In addition, genetic alterations followed by a single missense mutation in the endogenous private to a given cell line or tumor sample could obscure Ras gene (G12D or Q61R, respectively; Fig. 1A). The codon distinct functions of individual RAS mutants. To circumvent 12 LSL-Kras and Nras alleles have been previously described these issues, we generated a knock-in allele (LSL-NrasQ61R) (17, 18). We generated and confirmed a related LSL-NrasQ61R allowing for the conditional, tissue-specific, and somatic allele using standard homologous recombination followed expression of NrasQ61R under control of the endogenous by Southern blot, PCR, and genomic sequencing (Supple- promoter. We used this allele in combination with similarly mentary Fig. S1A–S1D). To minimize strain-specific effects, designed knock-in NrasG12D (17) and KrasG12D (18) models to all alleles were backcrossed more than seven generations to compare the transforming potential of each mutant when C57Bl/6J in the presence of a conditional p16INK4a knockout expressed at physiologic levels in melanocytes. allele (p16L; ref. 19) and a melanocyte-specific, 4-hydrox- ytamoxifen (4-OHT)–inducible CRE recombinase (Tyr-CRE- T2 RESULTS ER ; ref. 20). Cohorts of all three alleles were born at normal Mendelian ratios (data not shown) and showed no defects in Generation and Characterization development or fertility. of LSL-Nras Alleles Using primary melanocytes derived from syngeneic Tyr-CRE- To compare the ability of Nras mutants to promote ERT2 p16L/L LSL-NrasG12D/G12D (TpN12D/12D) or LSL-NrasQ61R/Q61R melanoma formation, we used three conditional knock-in (TpN61R/61R) neonates, we verified the functionality of the LSL alleles: LSL-NrasG12D (17), LSL-KrasG12D (18), and LSL-NrasQ61R. alleles. Melanocyte purities of >99% were confirmed using tyro- Each allele contains a floxed transcriptional stop sequence sinase-related protein 1 (TRP1) as a marker for flow cytometry

OF3 | CANCER DISCOVERY December 2014 www.aacrjournals.org

Codon Specificity in RAS-Driven Melanoma RESEARCH ARTICLE

(Supplementary Fig. S2A; ref. 21). In culture, melanocytes were A 100 treated with ethanol vehicle or 4-OHT to induce CRE activity. 80 Tp (n = 24) CRE-dependent excision of the transcriptional stop element 12D/12D was verified by PCR (Fig. 1B), and resulted in the production TpN (n = 29) 60 61R/61R of mutant Nras mRNA (Supplementary Fig. S2B). Activation TpN (n = 20) of either allele did not induce changes in melanocyte morphol- 40 TpK12D/WT (n = 21) ogy or pigmentation (Supplementary Fig. S2C), but induced 20 a decrease in melanocyte proliferation as measured by EdU Melanoma-free survival incorporation (Fig. 1C). Despite codeletion of p16INK4a, physi- 0 ologic expression of either Nras mutant caused comparable 0 20 40 60 80 antiproliferative effects (Fig. 1C). In addition, both TpN12D/12D Weeks of age 61R/61R and TpN melanocytes failed to bypass senescence in the B presence of 4-OHT, and the cells invariably ceased proliferat- ing after two to three passages in culture (data not shown). Therefore, when expressed under the control of an endogenous promoter, neither NrasG12D nor NrASQ61R was capable of 61R/61R 12D/WT immortalizing p16INK4a-deficient melanocytes. 12D/12D TpK TpN NRAS–Driven Melanomagenesis Is Codon-Specific TpN We next sought to examine the phenotypic effects of melanocyte-specific Ras mutations in vivo. Toward that end, we generated contemporaneous colonies of p16INK4a-deficient 12D/12D 61R/61R 12D/WT Kras- or Nras-mutant mice in a common genetic background. C TpN TpN TpK We elected to study Nras mutations in the homozygous state for two reasons: (i) we noted no melanoma formation despite 80 weeks of monitoring in a large cohort of Tyr-CRE-ERT2 p16L/L LSL-NrasG12D/WT (TpN12D/WT) animals (Supplementary Fig. S3A), and (ii) deep sequencing of nine NRAS-mutant human melanoma cell lines failed to detect the presence of a wild-type allele, demonstrating consistent NRAS loss of heterozygosity in vivo (22). Prior work from our group has demonstrated that melanocytic expression of KrasG12D efficiently promotes melanoma in vivo in the setting of p16INK4a inactivation (19, 23). Herein, syngeneic Tyr-CRE-ERT2 p16L/L LSL-KrasG12D/WT (TpK12D/WT) mice were maintained in a heterozygous state due to the homozygous lethality of Kras deletion (24). All mice in these cohorts were Figure 2. In vivo melanomagenesis is isoform and mutation-specific. treated neonatally with 4-OHT (19) to induce melanocyte- A, Kaplan–Meier curve of melanoma-free survival. Animals were followed for a total of 80 weeks. Survival of each strain was compared pairwise specific expression of the desired Ras mutant and delete to that of TpN61R/61R animals using the log-rank (Mantel–Cox) test. The p16INK4a. following significant P values were determined: Tp and TpN12D/12D, P < Initial examination of the skin, paws, and tails of these mice 0.0001; TpK12D/WT, P < 0.0027. B, representative images of the amelanotic tumors found throughout the skin of TpN12D/12D, TpN61R/61R, and TpK12D/WT revealed the presence of nevi and hyperpigmented regions animals. White arrowheads, the tumor location. C, hematoxylin and eosin on the paws and tails (Supplementary Fig. S4A). The pene- staining of representative tumors of the indicated genotypes. Scale bars, trance and severity of these phenotypes was allele-specific with 20 µm; white arrowheads, areas of melanin deposition. TpK12D/WT animals having the most pronounced effect and the Nras codon 12 mutation producing the least (TpK12D/WT > TpK12D/WT animals developed tumors with high penetrance TpN61R/61R > TpN12D/12D). To quantify nevi frequency, nevi (Fig. 2A; Table 1; ref. 19). Tumors were very rare in TpN12D/12D presence was scored weekly for 10 weeks. As expected, nevi mice (one tumor found in 29 mice observed for 80 weeks), were rarely observed on Tyr-CRE-ERT2 p16L/L (Tp) mice, but was whereas TpN61R/61R mice readily developed melanoma with consistently found on TpK12D/WT animals (Supplementary Fig. high penetrance and a median latency of 26.3 weeks (Fig. 2A S4B). Members of both the TpN12D/12D and TpN61R/61R cohorts and B and Table 1). To explain this result, we considered had more nevi than control Tp mice (P < 0.001). However, the possibility that the LSL-NrasQ61R and LSL-NrasG12D alleles the frequency with which NrasQ61R triggered nevus forma- had different recombination efficiencies. Toward that end, tion was significantly higher than that observed in TpN12D/12D we examined allelic recombination in primary melanocyte animals (P = 0.03). These data suggest that in p16INK4a-deficient cultures and noted that, if anything, the codon 61 allele melanocytes, physiologic NrasQ61R expression more effi- recombined with lower efficiency than the codon 12 allele ciently promotes nevus formation than NrasG12D. (Supplementary Fig. S5A). We also considered the possi- These established colonies of syngeneic TpN12D/12D, bility that the recombined LSL-NrasG12D allele was poorly TpN61R/61R, and TpK12D/WT mice were aged and serially assessed expressed. Sequencing of cDNA from treated melanocytes for melanoma formation. In accordance with previous data, confirmed 4-OHT–dependent expression of NrasG12D mRNA

December 2014 CANCER DISCOVERY | OF4

RESEARCH ARTICLE Burd et al.

induce Nras61R expression. The onset of tumor formation T able 1. Comparative summary of melanomagenesis in TpN61R/61R neonates was not affected by Lkb1 loss (Fig. 3A and macrometastasis in LSL-Ras–driven models and Table 1); however, the propensity for distant metastasis was markedly affected. TpLN61R/61R animals, treated as adults Tumors/treated Median tumor Macro or neonates, exhibited enhanced nevus formation relative to Genotype mice, % latency, wks metastasis Lkb1-proficient counterparts (compare Supplementary Fig. S4A with Fig. 3B). Upon sacrifice due to increasing tumor burden, Tp 1/24 (4.1%) >80 None many of TpLN61R/61R showed signs of metastasis. We performed 12D/12D TpN 1/29 (3.4%) >80 None thorough autopsies on 14 TpLN61R/61R mice, noting lymph node TpN61R/61R 14/20 (70%) 26.3 None enlargement and other signs of macroscopic disease (Fig. 3C) in TpK12D/WT 16/21 (76%) 36.3 None the majority of animals. The presence of visceral metastatic disease to the lung, spleen, and/or liver was confirmed in 5 of these TpLN61R/61R 17/20 (85%) 22.1 5/14 (36%) animals by histologic examination (36%; Fig. 3D and Table 1). Metastases appeared similar in morphology and incidence to (Supplementary Fig. S2B). Finally, functional validation of those observed in a prior Kras–driven, Lkb1−/− melanoma the LSL-NrasG12D allele was accomplished by inducing allelic model (23). Flow cytometric analyses of TpLN61R/61R mice with recombination in the hematopoietic lineage using an inter- primary melanomas showed the presence of infiltrating tumor feron-inducible Mx1-CRE driver. Expression of NrasG12D in the cells expressing the melanocyte marker TRP1 (Fig. 3E; note that hematopoietic compartment efficiently induced a myelopro- CD45+ cells were excluded from splenic analyses). Unlike cells liferative syndrome in accord with prior findings (ref. 25; and from TpN61R/61R tumors, we were able to culture TpLN61R/61R data not shown). These data establish that when expressed at melanoma cells in vitro. These cell lines exhibited varied mor- physiologic levels in melanocytes, NrasQ61R and KrasG12D phologies, but were invariably TRP1 positive (Supplementary are inherently more transforming than NrasG12D. Fig. S6). Together, these findings establish TpLN61R/61R as a We performed additional analyses to determine whether faithful murine model of metastatic, cutaneous melanoma there were phenotypic differences in the tumors of TpN12D/12D, driven by an endogenous, oncogenic NRAS allele. TpN61R/61R, and TpK12D/WT mice. Similar growth rates were Q61R G12D observed in tumors from TpN12D/12D, TpN61R/61R, and TpK12D/WT NRAS and NRAS Similarly Bind mice, albeit only one tumor was found in the TpN12D/12D cohort Melanogenic Effector Pathways (Supplementary Fig. S5B). Moreover, melanomas from all three The distinct oncogenicity of HRAS, KRAS, and NRAS is groups were histologically similar, containing both spindle-cell often attributed to the isoform-specific, preferential engage- and desmoplastic cell types with no overt signs of macrometa- ment of downstream effector pathways (17, 32–35). We postu- static spread (Fig. 2C and Table 1). In the TpN61R/61R tumors, lated that a similar mechanism might also drive codon-specific we confirmed the presence of the recombined LSL-NrasQ61R melanomagenesis. To this end, the interaction of purified KRAS, allele (Supplementary Fig. S5C) and used quantitative real- NRAS, NRASG12D, and NRASQ61R with PI3K and RAF [i.e., time PCR to look for potential gene amplifications. Compared the BRAF RAS binding domain (RBD)] was examined in vitro. with TpK12D/WT tumors, in which endogenous Nras expression Each GTPase was first loaded with a fluorescent GTP analogue is unaltered, TpN61R/61R melanomas displayed no significant [2′-/-3′-O-(N′methylanthraniloyl)-guanosine-5′-[(β,γ)-imido]tri- change in Nras mRNA levels, suggesting that the allele was phosphate (mantGMPPNP)]. Next, purified PI3K or RAF-RBD not amplified during tumorigenesis (Supplementary Fig. S5D). was titrated into reactions containing a constant amount of These data show that NrasQ61R, NrASG12D, and KrASG12D GMPPNP-RAS protein at both 5 mmol/L and 100 µmol/L Mg2+ tumors, once established, are phenotypically similar, and sug- concentrations. Fluorescence quenching, caused by effector gest that oncogenic differences between the alleles is most binding, was monitored to determine the Kd for each protein– pronounced during tumor initiation. protein interaction. In these assays, NRASQ61R bound both PI3K and the RAF-RBD with a lower affinity than either wild-type Lkb1 Loss Promotes Melanoma Metastasis NRAS or NRASG12D (Table 2). However, the noted differences 61R/61R in TpN Mice were small (<1-fold for PI3K; ∼4-fold for the RAF-RBD) and Melanoma mortality is predominantly attributed to meta- unlikely to translate to an in vivo phenotype. Likewise, isother- static disease; however, very few RAS-driven genetically engi- mal titration calorimetry experiments did not reveal significant neered mouse models, including the TpN61R/61R model, show differences in the RAF-RBD–binding affinities of NRASQ61R, evidence of macrometastases (Table 1; ref. 26). We recently NRASG12D, and wild-type NRAS (Supplementary Fig. S8). These demonstrated that the loss of liver kinase B1 (Lkb1/Stk11) data indicate that distinct engagement of the oncogenic RAS promotes widespread macrometastases in a Kras12D–driven effector pathways frequently targeted in melanoma (i.e., RAF mouse melanoma model (23). LKB1 is mutated in approxi- and PI3K) is not the cause of codon-specific melanomagenesis. mately 10% of malignant human melanomas (15, 27, 28) and functionally inactivated by oncogenic BRAF (22, 29, 30). To test Activation of MAPK and ERK Is Codon- whether Lkb1 loss promotes metastasis in TpN61R/61R tumors, Independent in NRAS-Mutant Melanomas we crossed these animals to a previously described conditional The distinct subcellular localizations of HRAS, KRAS, and Lkb1 knockout allele (Lkb1f; ref. 31). TpLN61R/61R mice were NRAS are suggested to influence effector availability and con- treated neonatally or at adulthood with 4-OHT to stimulate tribute to isoform-specific RAS oncogenicity (17, 32–35). To melanocyte-specific deletion of both Lkb1 and p16INK4a and determine whether the availability of specific effector pools

OF5 | CANCER DISCOVERY December 2014 www.aacrjournals.org

Codon Specificity in RAS-Driven Melanoma RESEARCH ARTICLE

A D Figure 3. Lkb1 deficiency promotes metas- 100 tasis in TpN61R/61R tumors. A, Kaplan–Meier curve comparing the melanoma-free survival 80 of TpN61R/61R and TpLN61R/61R mice treated Tp (n = 24) neonatally with 4-OHT. The log-rank (Mantel–Cox) Primary test revealed no significant difference in disease 60 61R/61R TpN (n = 20) onset in these models (P > 0.05). B, representa- TpLN61R/61R (n = 20) tive photographs showing severe melanocytic 40 hyperproliferation in the ears, tails, and paws of TpLN61R/61R mice. C, images of macrometastases 20 in the lung, spleen, and lymph nodes (LN) of Liver Melanoma-free survival TpLN61R/61R mice. D, representative hematoxylin 61R/61R 0 and eosin–stained TpLN tissues showing 0 20 40 60 80 tumor invasion into the liver, spleen, and lymph Weeks of age node. E, expression of TRP1 in the spleen and liver of TpLN61R mice with primary melanomas. The aqua line designates tissue from a wild-

B Spleen type (WT) animal, and the red line shows TRP1 Tail Paw Ear expression in a TpLN61R/61R mouse with observed macrometastases. For splenic tissues, CD45+ cells were excluded before analysis. Lymph node

C E Spleen Liver

WT TpLN 61R Count

Lung Spleen LN 0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 αTRP1

in vivo contributes to codon-specific signaling, we examined Because of the confounding potential of secondary muta- MAPK and PI3K activation in a variety of human melanoma cell tions acquired in human melanoma cell lines, we also exam- lines harboring a mutation in NRAS codon 12, 13, or 61 (n = 11). ined ERK and AKT activation in primary and immortalized These cell lines exhibited variable levels of activated ERK and AKT melanocytes derived from the TpN12D/12D and TpN61R/61R mod- that did not correlate with genotype (Fig. 4A and B). NRASG12/13- els. Expression of each Nras allele was induced in vitro using CRE mutant cell lines exhibited a moderate increase (<2-fold) recombinase and verified by genomic PCR (Supplementary Fig. in total ERK and AKT expression; however, the significance S10A and S10B, top). Immortalized TpN61R/61R melanocytes of this observation is unclear (Fig. 4A and B, right) and this were particularly resistant to sustained allelic recombination finding would not explain the increased transforming activity (Supplementary Fig. S10B, top). Therefore, we analyzed these of codon 61 mutants. In addition, no difference was observed cells at an early time point following CRE induction (2 days). In in the proliferation of NRAS codon 12/13 and 61–mutant both cases, we found that phospho-ERK and AKT levels were cell lines as measured by EdU incorporation (Supplementary variable and codon-independent (Supplementary Fig. S10A Fig. S9). and S10B, bottom). These data, along with observations in human melanoma cell lines, demonstrate that melanomage- neic and nonmelanomageneic NRAS mutants similarly acti- T able 2. Binding affinities of RAS·mantGMPPNP vate the MAPK and PI3K pathways. to PI3Kg and RAF-RBD NRASQ61R Exhibits Distinct Biochemical Properties Surprised by the finding that NRAS codon 12, 13, and 61 mant GMPPNP binding mutants appear equally to engage the oncogenic PI3K and MAPK effector pathways, we further investigated the bio- RAS mutant Kd oft bRa -RBD (µmol/L) Kd ot PI3K (µmol/L) chemical properties that differentiate these two mutants. WT NRAS 0.07 ± 0.03 2.2 ± 0.4 We observed during mantGMPPNP–based binding assays that NRASG12D 0.08 ± 0.04 1.8 ± 0.5 the nucleotide exchange rate of NRASQ61R was significantly G12D NRASQ61R 0.28 ± 0.04 2.7 ± 1.2 retarded compared with WT or NRAS (data not shown). To follow up on this observation, we measured the intrinsic and KRASG12D 0.06 ± 0.03 2.5 ± 0.8 son of sevenless (SOS)–mediated nucleotide disassociation

December 2014 CANCER DISCOVERY | OF6

RESEARCH ARTICLE Burd et al.

A 0.4 3

NZM63NZM24WM3629WM3670MaMel27IIMel224SK-Mel-119VMM39SK-Mel-103SK-Mel-147WM1366kD 0.3 pERK 2 (T202/Y204) –37 0.2 ERK –37

ERK/ β -Actin 1

β-Actin pERK/total ERK –37 0.1 Codon 12/13 Codon 61 0.0 0 G12/13 Q61 G12/13 Q61

* B 0.15 2.0

1.5 NZM63NZM24WM3629WM3670MaMel27IIMel224SK-Mel-119VMM39SK-Mel-103SK-Mel-147WM1366kD 0.10 pAKT –75 (S473) –50 0.05 1.0 –75 AKT –50 AKT/ β -Actin

pAKT/total AKT pAKT/total 0.00 0.5 β-Actin –37 Codon 12/13 Codon 61 –0.05 0.0 G12/13 Q61 G12/13 Q61

Figure 4. Activation of ERK and AKT in human melanoma cell lines is not codon-specific. Shown are immunoblots for total and phosphorylated (p) ERK (A) or AKT (B) in established human melanoma cell lines harboring the indicated NRAS mutations. Protein expression levels were quantified using LI-COR ImageStudio software. Each dot represents a single cell line. The mean is indicated by a line, and the standard error is shown as whiskers. *, P < 0.05. rates of purified NRAS, NRASG12D, and NRASQ61R. To moni- NRASQ61R. Indeed, when directly assessed, the intrinsic GTP tor exchange, each NRAS variant was preloaded with either hydrolysis rate of NRASQ61R was much slower than NRASG12D or mantGMPPNP (mimicking the GTP-bound state) or 2′-/-3′- NRAS (1,150 and 2,300 times slower, respectively; see Table 3). O-(N′methylanthraniloyl)-guanosine diphosphate (mantGDP) These results, taken together, suggest that NRASQ61R has and then incubated with 1,000-fold excess unlabeled nucle- higher affinity for GTP relative to NRAS and NRASG12D. There- otide. The resulting change in fluorescence over time was used fore, we examined the relative stability of NRAS, NRASG12D, to monitor exchange rates. NRAS, NRASG12D, and NRASQ61R and NRASQ61R using thermal unfolding measurements for showed a similar ability to exchange GDP (Table 3). However, both GDP- and GMPPCP-bound proteins. NRASQ61R remained the GMPPNP exchange rate of NRASQ61R was significantly stable until reaching 80°C ± 5°C, whereas the wild-type and slower than NRAS and NRASG12D (Table 3). This distinction NRASG12D proteins were destabilized at much lower tempera- became more pronounced when SOS was added to each tures (67°C ± 3°C and 70°C ± 3°C, respectively; see Supple- reaction (Table 3). Specifically, SOS was unable to stimulate mentary Table S2). The thermostability of NRASQ61R was lower either GDP or GMPPNP exchange in the codon 61 variant in the GDP-bound versus GMPPCP-bound state (74°C ± 4°C (Table 3). These results indicate that NRASQ61R possesses a vs. 80°C ± 5°C; see Supplementary Table S2), suggesting that clear affinity for GTP that cannot be overcome by guanine NRASQ61R adopts a conformation that stabilizes nucleotide nucleotide exchange factor (GEF) interaction at the concen- binding, especially in the GTP-bound state. Together, our work trations used in our assays. reveals that NRASQ61R exhibits distinct nucleotide-binding These data, along with previous observations in other RAS capacity, stability, and GTPase resistance likely responsible isoforms (36), suggest a decreased intrinsic GTPase activity of for its exceptional melanomagenic properties.

T able 3. Nucleotide binding affinities and intrinsic GTP hydrolysis rates of relevant NRAS isoforms

GMPPNP off + 1:2 Intrinsic hydrolysis GDP off (rate, GDP off + 1:2 SOScat GMPPNP off (rate, SOScat (rate, per (rate, per second × RAS mutant per second × 10−4) (rate, per second × 10−4) per second × 10−4) second × 10−4) 10−4) NRASWT 1.5 ± 0.2 8.0 ± 0.7 1.7 ± 0.4 9.8 ± 0.9 4.6 ± 0.4 NRASG12D 1.3 ± 0.3 3.3 ± 0.4 1.6 ± 0.5 3.9 ± 0.6 2.3 ± 0.5 NRASQ61R 1.1 ± 0.2 1.3 ± 0.3 0.4 ± 0.2 0.5 ± 0.2 0.002 ± 0.03

OF7 | CANCER DISCOVERY December 2014 www.aacrjournals.org

Codon Specificity in RAS-Driven Melanoma RESEARCH ARTICLE

DISCUSSION genic mechanisms found in human melanomas. For these reasons, we believe that the TpN61R and TpLN61R models will Our data explain a long-standing mystery in the field, dem- be extremely valuable for preclinical drug testing, especially onstrating through biochemical and genetic analyses that the given the metastatic nature of the Lkb1-deficient tumors. predominance of codon 61 mutants in human melanoma can Q61R be attributed to their distinct oncogenic properties. In our NRAS Exhibits Unique Biochemical Properties novel suite of mouse models, endogenous levels of NrasQ61R, In contrast to prior studies showing that preferential effector but not NrasG12D, were able to efficiently drive in vivo usage drives isoform-specific RAS oncogenesis (17, 32–35), we melanomagenesis (Fig. 2). Although prior work has linked found that distinct biochemical properties inherent to each RAS isoform-specific RAS oncogenicity to differential effector mutant likely drive codon-specific melanomagenesis (Table 2 binding (17, 32–35), we found that NRASQ61R and NRASG12D and Supplementary Table S2). Nearly three decades ago, Der similarly engaged the PI3K and MAPK pathways (Table 2; Fig. and colleagues (48) reported that HRAS codon 61 mutants, 4; Supplementary Figs. S8 and S10). Our data suggest an alter- regardless of their transforming potential, exhibited an approx- native model for codon-specific oncogenicity, demonstrating imately 10-fold decrease in intrinsic GTP hydrolysis. Later, that the melanomagenic NRASQ61R mutant possesses distinct Donovan and colleagues (49) used predicted RAS-GTP levels biochemical properties not found in NRASG12D (Table 3). to form the hypothesis that defects in intrinsic GTPase activity Before this study, it was unclear whether the prevalence of and nucleotide exchange dictate the transforming potential of NRAS codon 61 mutations in melanoma reflected a preferen- individual RAS mutants. Our experimental results are consist- tial pattern of mutagenesis or codon-specific differences in the ent with this model, demonstrating that the melanomagenic biologic function of each allele. We now show that NRASG12D NRASQ61R variant exhibits high-affinity GTP binding, increased does not efficiently drive cutaneous melanomagenesis (Fig. 2), stability, and reduced intrinsic GTPase activity when compared and therefore codon-specific RAS biology influences tumori- with NRASG12D (Table 2 and Supplementary Table S2). genic potential. Tissue and temporal specificity are also likely These results suggest that NRASQ61R may more efficiently to affect oncogenic RAS activity. For instance, activation of activate downstream effectors in vivo. However, the activation the LSL-NRASG12D allele in hematopoietic stem cells and early of ERK and AKT was similar in both primary melanocytes and melanocyte progenitors drives leukemogenesis and melanoma human melanoma cultures harboring an NRAS codon 12 or 61 of the central nervous system (refs. 25, 37, 38; data not mutation (Fig. 4 and Supplementary Fig. S10). Moreover, even shown). Moreover, it is possible that our experimental system, through unbiased probes of the kinome using PhosphoScan which relies upon p16INK4a loss to facilitate melanomagenesis, technology (Cell Signaling Technology), we were unable to is biased to favor Nras codon 61 mutations. Patient studies detect consistent differences between NrasG12D and NrASQ61R have suggested exquisite cooperation between NRAS codon signaling in primary melanocytes (data not shown). Because 61 mutations and p16INK4a inactivation (39, 40), and NrasQ61K of feedback mechanisms within many pathways downstream transgenic mice appear to be “addicted” to the activity of of RAS, changes in signaling flux may not be readily appar- CDK4/6, which is enhanced by p16INK4a loss (41). Therefore, ent. Alternatively, enhanced cooperation with secondary onco- perhaps other cooperating oncogenic events commonly found genic events during the initial stages of tumor development in melanoma would more efficiently synergize with rN asG12D may be responsible for the selection of NRASQ61R mutants in in tumor formation. Nonetheless, p16INK4a loss is found in the melanoma. In line with this observation, RAF activation is a majority of human melanomas, occurring through a variety common event in both thyroid carcinomas and melanomas, of mechanisms (i.e., epigenetic silencing, mutation, and dele- the two major tumor types wherein NRAS codon 61 mutations tion). Given the significance of this event in the clinical set- predominate (1, 2). Mounting evidence supports the idea that ting, we believe that the choice of the p16INK4a-deficient genetic the intensity of oncogenic RAS signaling can influence transfor- context is highly relevant to the human disease. mation potential (38). In addition, RAF binding in conjunction Widely efficacious and durable treatment options are not with unidentified allosteric regulators is speculated to further available for the 17% to 30% of patients with NRAS-mutant, reduce the intrinsic GTPase activity of HRAS codon 61 mutants metastatic melanomas (1, 15, 42). With mounting data to (36). These findings, along with our work, suggest that NRAS suggest that genetically engineered mouse models of cancer codon 61 mutants exhibit unique biochemical properties that more faithfully report therapeutic efficacy (43–45), a number promote the transformation of RAF-responsive tissues. Pars- of RAS-mutant melanoma models have been developed (see ing the mechanism of this specific dependency will require Supplementary Table S3). Most of these models use trans- the structural analyses of multiple NRAS mutants as well genic technologies that can alter the expression, subcellular as comprehensive functional screens using endogenous RAS compartmentalization, intracellular signaling, and trans- expression systems and relevant cell types. However, our data forming potential of RAS (16). To this end, the activation of suggest that this work would be extremely valuable, identifying endogenous RAS oncogenes triggers minimal ERK and AKT codon-specific tumor vulnerabilities for therapeutic targeting. activation in our system and others (Supplementary Fig. S10; refs. 17, 18, 25); yet, transgenic RAS alleles robustly stimulate these pathways (46, 47). Other murine melanoma models use METHODS oncogenic RAS isoforms rarely observed in human melano- Murine Alleles and Husbandry mas (i.e., HRAS and KRAS mutants). As each RAS isoform Animal work was conducted in accordance with protocols approved can initiate a unique set of downstream signals, it is unclear by the Institutional Care and Use Committee for animal research at to what degree these models faithfully recapitulate the onco- the University of North Carolina (Chapel Hill, NC) and the Ohio

December 2014 CANCER DISCOVERY | OF8

RESEARCH ARTICLE Burd et al.

State University (Columbus, OH). The Tyr-CRE-ERT2, p16L, Lkb1f, for that time point. The percentage of time nevus positive was calcu- LSL-Kras12D, and LSL-Nras12D alleles have been previously described lated as: (sum of all 10 measurements/10) × 100. Each dot represents (17–20, 31). All animals in this study were backcrossed more than 1 animal with the mean indicated by a line. Comparisons between seven generations to C57BL/6. The Lkb1f allele was initially estab- sample pairs were conducted using the Student t test. lished on an albino C57BL/6J background and then crossed into the TpN61R colony. Therefore, some of the mice in this cohort are albino. Culture of Murine and Human Melanoma Cell Lines Analysis of tumor-free survival was conducted using GraphPad Prism The NZM24 and NZM63 cell lines were developed in the laboratory software. To determine statistical significance, the log-rank (Mantel– of Dr. Bruce C. Baguley (51) and obtained from the Cell Line Collection Cox) test was performed for each experimental pairing. maintained by the Auckland Cancer Society Research Centre, Faculty of Medical and Health Sciences, The University of Auckland (Auckland, Q61R Generation of the LSL-Nras Allele New Zealand). The MaMel27II cell line was generated by Drs. G. Finlay, Standard homologous recombination procedures were used to insert C. Posch, and S. Ortiz (Mount Zion Cancer Research Center, San Fran- a conditionally excisable transcriptional stop codon and point mutation cisco, CA; ref. 52). WM3670, WM3629, and WM1366 cells were created into the endogenous Nras locus (see Supplementary Fig. S1A). Founding in the laboratory of Meenhard Heryln and obtained from the Wistar mice were first crossed to C57Bl/6 FlpE mice (The Jackson Laboratory; Institute (Philadelphia, PA; ref. 53). VMM39 cells were the kind gift #005703) to remove the neomycin resistance cassette. Southern blot- of C. Slingluff (University of Virginia, Charlottesville, VA; ref. 54). The ting was performed on embryonic stem (ES) cells and founder DNA SK-Mel-119, SK-Mel-103, and SK-Mel-147 cell lines were generously using the standard procedures. PCR primers and cycling conditions provided by Dr. A. Houghton (Memorial Sloan Kettering, New York, used for genotyping LSL-NrasQ61R were as follows: Q61R GENO2—5′- NY; ref. 55). The Mel224 cell line was produced and supplied by GCAAGAGGCCCGGCAGTACCTA-3′ (0.15 µmol/L); Primer 1—5′-AG J. Hansson (Karolinska Institutet, Solna, Sweden). Tumor-derived ACGCGGAGACTTGGCGAGC-3′ (0.15 µmol/L); Primer 2—5′-GCTGG murine cell lines from the TpLN61R/61R and TpK12D/WT models were gen- ATCGTCAAGGCGCTTTTCC-3′ (0.15 µmol/L); cycling—95°C 15 min- erated, genotyped, and maintained as previously described (23). utes, 35 × [94°C 30 seconds, 62°C 30 seconds, 72°C 45 seconds], 72°C 5 minutes. The resulting PCR products were 487 (wild-type), 371 (LSL- Authentication of Cell Lines NrasQ16R), and 562 (LSL-NrasQ16R + CRE) base pairs in size. All cell lines were obtained from their original sources or authorized distributors and maintained as previously described (51–56). Cell lines Primary Melanocyte Culture from the Wistar Institute (WM3670, WM3629, and WM1366) are rou- Skin was isolated from newborn pups and placed dermis side down tinely validated by short-tandem repeat (STR) profiling using the Amp- in 0.25% trypsin, 0.1% EDTA for 3 hours at 37°C, and 5% CO2. Using FISTR Identifiler PCR Amplification Kit (Life Technologies) and were forceps, the epidermal layer was separated from the dermis. Using used for the experiments shown within 3 months of receipt. For pri- surgical scissors, epidermal cells were minced finely in phosphate- mary cell lines where original STR data is not currently available, NRAS buffered saline supplemented with 0.02% EDTA. To further dissoci- mutations were validated by PCR followed by Sanger sequencing. ate the cell suspension, each sample was subjected to two rounds of program A on the GentleMACs dissociator (Miltenyi Biotec). The Flow Cytometry resulting samples were spun down and plated on rat tail collagen- coated dishes in melanocyte growth medium (Ham F10, 10 µg/mL Cell-cycle analysis of primary melanocytes After 10 days in culture, 5 insulin, 0.5 ng/mL bovine serum albumin, 5% fetal bovine serum, 1 4.4 × 10 primary melanocytes were seeded onto collagen-coated µmol/L ethanolamine, 1 µmol/L phophoethanolamine, 10 nmol/L 60-mm dishes. A day later, the medium was changed and cells were sodium selenite, 20 nmol/L TPA, 50 pmol/L cholera toxin, 1× allowed to recover overnight. EdU was added to each culture at a con- penicillin/streptomycin, 100 nmol/L melanocyte stimulating hor- centration of 2.5 µg/mL for a period of 16 hours. Cells were harvested mone, and 0.05 mmol/L di-butyril cyclic AMP). Media was changed and processed as described in the Click-iT EdU Flow Cytometry Assay every other day. Kit (Invitrogen) using saponin for permeabilization.

Cell-cycle analysis of human melanoma cell lines Human mela Induction of CRE Recombinase noma cell lines grown to 50% to 70% confluence and were treated For neonatal induction, pups were painted dorsally (postnatal days with 10 µmol/L EdU for a period of 6 hours. Cells were har- 2–4) with 25 mg/mL 4-OHT dissolved in DMSO. Adult induction vested, processed, stained for EdU incorporation, and analyzed on a of CRE was performed as previously described (50). For cell culture FlowSight cytometer (Amnis). studies in primary melanocytes, 4-OHT was dissolved in ethanol and added to the growth media. Cells were treated for 6 consecutive days Trp1 staining To validate that our cultures contained a pure replacing the media and 4-OHT every other day. In the immortalized population of melanocytes and verify the identity of TpLN61R metas- 61R TpN cultures, which were resistant to recombination, an adenovi- tases, PEP1 antibody (αTrp1; ref. 21) was added at a 1:100 dilu- ral CRE recombinase was used (Ad5-MCV-Cre-GFP; Baylor College tion to saponin-permeabilized cells followed by incubation with of Medicine Vector Development Laboratory, Houston, TX). Before an Alexa Fluor 488–conjugated anti-rabbit secondary (Molecular proliferation or morphologic assessment, melanocyte cultures were Probes, 1:1,000 dilution). In the analysis of splenic tissues (Fig. 3E), returned to untreated media for 3 days. Allelic recombination in co-staining with CD45 was first used to gate out any CD45+ cells. tumors and primary cultures was confirmed by genomic PCR using the genotyping primer sets and conditions described for each allele. In Vitro Protein Purification A vector encoding human NRAS (1-172) was acquired from Scoring of Murine Nevi Addgene (vector #25256, NRAS-A). The G12D and Q61R mutations Nevi on a 4-cm2 dorsal area were counted by staff, blinded of the were each introduced into this vector using standard quick change animal’s genotype, every week for 10 weeks. So as not to interfere mutagenesis. All proteins were expressed and purified from Rosetta with melanomagenesis, we chose not to depilate or shave these ani- 2 BL21(DE3) cells (Novagen). Briefly, cells were grown at 37°C in a mals during the study. The presence of a single nevus at any time shaking culture of Terrific Broth media supplemented with kanamy- point was scored as a “1.” If no nevus was observed a “0” was entered cin. Once the cells reached an OD600 between 0.6 and 1.0, the culture

OF9 | CANCER DISCOVERY December 2014 www.aacrjournals.org

Codon Specificity in RAS-Driven Melanoma RESEARCH ARTICLE was chilled in an ice bath (to ∼20°C) and 0.1 mmol/L IPTG added. of unlabeled nucleotide, and the rate of dissociation was determined The cultures were then shaken for an additional 12 to 15 hours at by monitoring the change in fluorescence of the mantGMPPNP– 18°C. Cells were harvested by centrifugation and pellets stored at loaded protein (excitation and emission wavelengths of 365 and −80°C. NRAS proteins were purified on Ni Sepharose6 columns 435 nm, respectively) using a SpectraMax M2 plate reader (59). Each following the manufacturer’s protocol (GE Life Sciences). To further nucleotide dissociation curve was fit to a one-phase single exponen- purify the proteins from contaminants, size-exclusion chromatog- tial to determine kobs. The dissociation rates were plotted against the raphy was performed (Superdex 75 10/300 GL; GE Life Sciences). effector concentrations and fit to the equation: Purity exceeding 95% was confirmed by SDS-PAGE analysis. 2 0. 5 cat (([][]PPKPPK1+ 2 +d )(()() −1 + 2 +d ) −4 ×[[PP1 ]× [ 2 ]) The minimum catalytic domain of the human protein SOS YAAB= −() − × was expressed in the pQlinkH vector (Addgene) and purified as 2×[]P1 previously described (57). The Echerichia coli codon-optimized RAS- binding domain of BRAF (amino acids 149–232) containing an where Y is observed rate; A is maximum rate; B is minimum rate; P1 is N-terminal purification tag (MGHHHHHHSSGVDLGTENLYFQS) concentration of protein 1 (cell); P2 is concentration of protein 2 (ligand/ was synthesized by Genewiz, subcloned into pET28a and expressed effector), which was used to determine the NRAS·mantGMPPNP:effector in BL21 (DE3) cells. The resulting BRAF-RBD was purified on Ni binding constant (Kd; ref. 60). Sepharose6, cleaved overnight with TEV protease, and then subjected to subtractive Ni column purification. The tag-less BRAF-RBD was Nucleotide Exchange in the Absence and further purified using size exclusion chromatography and verified Presence of SOScat to be >95% pure by SDS-PAGE analysis. Purified PI3Kγ (amino acids Nucleotide dissociation rates were measured using a PerkinElmer 144–1102) was kindly provided by Genentech. LS50B fluorimeter as previously reported (58). NRAS was loaded with mant-labeled nucleotide and added to a final concentration Loading of NRAS with Nucleotide Derivatives of 1 µmol/L in 1 mL of exchange buffer (20 mmol/L HEPES, 50

To observe nucleotide dissociation activity in the GTP-bound mmol/L NaCl, 5 mmol/L MgCl2, and 100 µmol/L DTPA, pH 7.4). A state, NRAS was loaded with mantGMPPNP. NRAS was exchanged 1,000-fold excess of unlabeled nucleotide was added to initiate the out of excess MgCl2 into a buffer containing 20 mmol/L HEPES, 50 dissociation reaction, and the rate of nucleotide dissociation was mmol/L NaCl, 125 mmol/L (NH4)2SO4, and 1 mmol/L MgCl2 at pH measured by monitoring the change in fluorescence λ( ex = 365 nm; 7.4. The concentration of NRAS was adjusted to 100 µmol/L and a λem = 435 nm). For reactions that showed dissociation rates too slow 5-mol/L excess of mantGMPPNP was added. Alkaline phosphatase to reach completion within the time frame of the assay, 25 mmol/L conjugated to sepharose beads was added to 1:10 volume and EDTA EDTA was added to induce complete nucleotide dissociation. For was added to 1 mmol/L to increase the nucleotide exchange rate. GEF-facilitated nucleotide exchange, the minimum catalytic domain The protein mixture was incubated at 4°C until all unlabeled nucle- of SOScat was added to a 1:2 RAS-to-SOScat ratio (chosen to limit RAS otide was converted to guanosine. The alkaline phosphatase beads binding to the allosteric site in SOScat). The fluorescence data were were removed by centrifugation and 10 mmol/L MgCl2 was added normalized and fit to one-phase exponentials to determine rates. to induce nucleotide binding. The concentration of mantGMPPNP- bound RAS was determined by measuring the absorbance of the Determination of Intrinsic GTP Hydrolysis mant fluorophore [ε350 = 5,700/((mol/L)*cm)]. For assays requiring Activity of NRAS unlabeled NRAS in the GTP-bound state, NRAS·GMPPCP was pre- pared in an identical manner. To measure GTP hydrolysis rates, the phosphate sensor FLIPPi For assays observing NRAS nucleotide dissociation activity in the (Addgene) was used for the reaction, which binds to free phosphate in GDP-bound state, the protein was loaded with mantGDP as previously solution (61). Because of the nature of the assay, all trace phosphate described (58). NRAS was exchanged into a buffer containing 20 was removed from buffers before use to minimize background fluorescence. NRAS was exchanged into a buffer containing 20 mmol/L mmol/L HEPES, 125 mmol/L (NH4)2SO4, 100 µmol/L EDTA, and HEPES, 20 mmol/L (NH4)2SO4, 1 mmol/L EDTA, and 1 mmol/L 200 µmol/L DTPA at pH 8.0 to remove excess MgCl2 and nucleotide. A 5 mol/L excess of mantGDP was added to the protein and the mix- inosine, pH 8.0. NRAS was incubated with GTP and passed through 2+ ture was incubated for 30 minutes at 37°C to increase the nucleotide a PD-10 column to remove trace Mg (62). NRAS was diluted to 10 µmol/L and combined with 10 µmol/L FLIPPi, and 2 mmol/L exchange rate. MgCl2 was then added to 20 mmol/L and the protein was placed at 4°C for 2 hours to induce binding, after which NRAS was MgCl2 was added to initiate the hydrolysis reaction. A phosphate exchanged into a buffer containing 20 mmol/L HEPES, 50 mmol/L standard curve was used to convert the raw fluorescence output to a measurement of [GTP] hydrolyzed. NaCl, 5 mmol/L MgCl2, and 100 µmol/L DTPA, pH 7.4 for storage.

Quantitative Real-Time PCR Thermal Melting Analysis by Circular Dichroism Spectroscopy RNA from melanocytes or melanomas was isolated using the Qiagen AllPrep Kit. Equal amounts of RNA were subjected to reverse A Jasco J-815 CD Spectrometer was used for circular dichroism transcription using the PrimeScript First Strand cDNA Synthesis Kit (CD) measurements. NRAS was exchanged into a buffer containing (Takara). Real-time PCR was run in triplicate and a standard curve 10 mmol/L KH2PO4/K2HPO4 at pH 7.45 and diluted to 15 µmol/L. of known plasmid concentration was included on every plate (ABI MgCl2 and guanine nucleotide (GDP or GMPPCP) were added to a TaqMan primer set #Mm00477878). Molecules of NRAS/µg total final concentration of 500 and 80 µmol/L, respectively, immediately RNA were calculated assuming a cDNA conversion efficiency of 100%. before analysis. A thermal melt scan from 20° to 90°C was performed to determine the temperature (Tm) at which half of the protein is Determination of NRAS Binding to BRAF-RBD and PI3Kg unfolded. For quantitative binding to BRAF-RBD and PI3Kγ (amino acids 144–1102), NRAS·mantGMPPNP was incubated with different con- RAS Mutational Analysis in Human Cancers centrations of the desired effectors in a buffer containing 50 mmol/L Data were downloaded from COSMIC (release v64, March 26,

HEPES, 50 mmol/L NaCl, and 5 mmol/L MgCl2 at pH 7.4. Nucle- 2013; ref. 1). Cell lines and cultured cells were removed from the analy- otide dissociation was initiated by the addition of 1,000-mol/L excess sis to prevent inclusion of events secondary to culture. Remaining

December 2014 CANCER DISCOVERY | OF10

RESEARCH ARTICLE Burd et al. events were limited to observations of somatic missense mutations 3. Adari H, Lowy DR, Willumsen BM, Der CJ, McCormick F. Guanosine in codons 12, 13, or 61 of HRAS, NRAS, or KRAS. Multiple subtypes triphosphatase activating protein (GAP) interacts with the p21 ras and histologies were then summed to give information for the dis- effector binding domain. Science 1988;240:518–21. played tumor types. 4. Frech M, Darden TA, Pedersen LG, Foley CK, Charifson PS, Anderson MW, et al. Role of glutamine-61 in the hydrolysis of GTP by p21H- Disclosure of Potential Conflicts of Interest ras: an experimental and theoretical study. Biochemistry 1994;33: 3237–44. No potential conflicts of interest were disclosed. 5. Garm Spindler KL, Pallisgaard N, Rasmussen AA, Lindebjerg J, Andersen RF, Cruger D, et al. The importance of KRAS mutations Authors’ Contributions and EGF61A>G polymorphism to the effect of cetuximab and iri- Conception and design: C.E. Burd, S.L. Campbell, N.E. Sharpless notecan in metastatic colorectal cancer. Ann Oncol 2009;20:879–84. Development of methodology: C.E. Burd, M.V. Huynh, M.J. Miley, 6. Lievre A, Bachet JB, Le Corre D, Boige V, Landi B, Emile JF, et al. KRAS N.E. Sharpless mutation status is predictive of response to cetuximab therapy in Acquisition of data (provided animals, acquired and managed colorectal cancer. Cancer Res 2006;66:3992–5. patients, provided facilities, etc.): C.E. Burd, W. Liu, M.V. Huynh, 7. Karapetis CS, Khambata-Ford S, Jonker DJ, O’Callaghan CJ, Tu D, M.A. Waqas, J.E. Gillahan, K.S. Clark, K. Fu, G.P. Souroullas, D.B. Darr, Tebbutt NC, et al. K-ras mutations and benefit from cetuximab in D.C. Zedek, M.J. Miley, B.C. Baguley, S.L. Campbell, N.E. Sharpless advanced colorectal cancer. N Engl J Med 2008;359:1757–65. Analysis and interpretation of data (e.g., statistical analy- 8. Lievre A, Bachet JB, Boige V, Cayre A, Le Corre D, Buc E, et al. KRAS sis, biostatistics, computational analysis): C.E. Burd, W. Liu, mutations as an independent prognostic factor in patients with M.V. Huynh, M.A. Waqas, J.E. Gillahan, W.R. Jeck, G.P. Souroullas, advanced colorectal cancer treated with cetuximab. J Clin Oncol 2008;26:374–9. D.B. Darr, D.C. Zedek, M.J. Miley, S.L. Campbell, N.E. Sharpless 9. Tejpar S, Celik I, Schlichting M, Sartorius U, Bokemeyer C, Van Writing, review, and/or revision of the manuscript: C.E. Burd, Cutsem E. Association of KRAS G13D tumor mutations with out- W. Liu, M.V. Huynh, M.A. Waqas, J.E. Gillahan, B.L. Martin, D.B. Darr, come in patients with metastatic colorectal cancer treated with D.C. Zedek, M.J. Miley, S.L. Campbell, N.E. Sharpless first-line chemotherapy with or without cetuximab. J Clin Oncol Administrative, technical, or material support (i.e., reporting or 2012;30:3570–7. organizing data, constructing databases): C.E. Burd, J.E. Gillahan, 10. Chen J, Ye Y, Sun H, Shi G. Association between KRAS codon 13 B.L. Martin, M.J. Miley mutations and clinical response to anti-EGFR treatment in patients Study supervision: N.E. Sharpless with metastatic colorectal cancer: results from a meta-analysis. Cancer Chemother Pharmacol 2013;71:265–72. Acknowledgments 11. Mao C, Huang YF, Yang ZY, Zheng DY, Chen JZ, Tang JL. KRAS The authors thank Wayne Joseph, Drs. Graeme Finlay, Christian p.G13D mutation and codon 12 mutations are not created equal in Posch, Susana Ortiz, Vincent Hearing, Meenhard Herlyn, Altaf Wani, predicting clinical outcomes of cetuximab in metastatic colorectal Kevin Haigis, and Tyler Jacks for reagents, equipment, and/or ani- cancer: a systematic review and meta-analysis. Cancer 2013;119: mals; Drs. Ruben Carrasco, Yuri Fedoriw, and Arlin Rogers for advice; 714–21. Dr. Nabeel Bardeesy for critical reading of the article; and the UNC 12. Modest DP, Jung A, Moosmann N, Laubender RP, Giessen C, Schulz C, et al. The influence of KRAS and BRAF mutations on the efficacy Structural Bioinformatics Core Facility (Brenda Temple), Animal of cetuximab-based first-line therapy of metastatic colorectal cancer: Histopathology Core, Biostatistics Core (Dominic Moore), Protein an analysis of the AIO KRK-0104-trial. Int J Cancer 2012;131:980–6. Expression and Purification Core (Mischa Machius), Macromolecu- 13. De Roock W, Jonker DJ, Di Nicolantonio F, Sartore-Bianchi A, Tu D, lar Interactions Facility (Ashutosh Tripathy), and Mouse Phase I Unit Siena S, et al. Association of KRAS p.G13D mutation with outcome (Luke Hunter) for technical assistance. Purified PI3Kγ was generously in patients with chemotherapy-refractory metastatic colorectal can- provided by Genentech. cer treated with cetuximab. JAMA 2010;304:1812–20. 14. Ihle NT, Byers LA, Kim ES, Saintigny P, Lee JJ, Blumenschein GR, Grant Support et al. Effect of KRAS oncogene substitutions on protein behavior: This work was supported by NIH R00AG036817 (to C.E. Burd), implications for signaling and clinical outcome. J Natl Cancer Inst NIH R01CA163896 (to N.E. Sharpless), NIH U01CA141576 (to N.E. 2012;104:228–39. Sharpless), American Cancer Society IRG-67-003-50 (to C.E. Burd), 15. Hodis E, Watson IR, Kryukov GV, Arold ST, Imielinski M, Theuril- NIH T32ES007126 (to C.E. Burd), NIH T32CA009156 (to G.P. Sour- lat JP, et al. A landscape of driver mutations in melanoma. Cell oullas), and The UNC University Cancer Research Fund (UCRF). 2012;150:251–63. The costs of publication of this article were defrayed in part by 16. Omerovic J, Prior IA. Compartmentalized signalling: Ras proteins the payment of page charges. This article must therefore be hereby and signalling nanoclusters. FEBS J 2009;276:1817–25. 17. Haigis KM, Kendall KR, Wang Y, Cheung A, Haigis MC, Glickman JN, marked advertisement in accordance with 18 U.S.C. Section 1734 et al. Differential effects of oncogenic K-Ras and N-Ras on prolifera- solely to indicate this fact. tion, differentiation and tumor progression in the colon. Nat Genet 2008;40:600–8. Received July 11, 2014; revised September 11, 2014; accepted Sep- 18. Tuveson DA, Shaw AT, Willis NA, Silver DP, Jackson EL, Chang S, tember 11, 2014; published OnlineFirst September 24, 2014. et al. Endogenous oncogenic K-ras(G12D) stimulates proliferation and widespread neoplastic and developmental defects. Cancer Cell 2004;5:375–87. 19. Monahan KB, Rozenberg GI, Krishnamurthy J, Johnson SM, Liu W, References Bradford MK, et al. Somatic p16(INK4a) loss accelerates melanom- . 1 Forbes SA, Bindal N, Bamford S, Cole C, Kok CY, Beare D, et al. agenesis. Oncogene 2010;29:5809–17. COSMIC: mining complete cancer genomes in the Catalogue of 20. Bosenberg M, Muthusamy V, Curley DP, Wang Z, Hobbs C, Nelson B, Somatic Mutations in Cancer. Nucleic Acids Res 2011;39:D945–50. et al. Characterization of melanocyte-specific inducible Cre recombi- 2. Cerami E, Gao J, Dogrusoz U, Gross BE, Sumer SO, Aksoy BA, nase transgenic mice. Genesis 2006;44:262–7. et al. The cBio cancer genomics portal: an open platform for 21. Hearing VJ. Antibodies specific to melanocyte-specific proteins avail- exploring multidimensional cancer genomics data. Cancer Discov able from the Hearing Laboratory. Pigment Cell Melanoma Res 2012;2:401–4. 2009;22:1.

OF11 | CANCER DISCOVERY December 2014 www.aacrjournals.org

Codon Specificity in RAS-Driven Melanoma RESEARCH ARTICLE

22. Jeck WR, Parker J, Carson CC, Shields JM, Sambade MJ, Peters EC, 44. Combest AJ, Roberts PJ, Dillon PM, Sandison K, Hanna SK, Ross C, et al. Targeted next generation sequencing identifies clinically action- et al. Genetically engineered cancer models, but not xenografts, faith- able mutations in patients with melanoma. Pigment Cell Melanoma fully predict anticancer drug exposure in melanoma tumors. Oncolo- Res 2014;27:653–63. gist 2012;17:1303–16. 23. Liu W, Monahan KB, Pfefferle AD, Shimamura T, Sorrentino J, Chan 45. Usary J, Zhao W, Darr D, Roberts PJ, Liu M, Balletta L, et al. Predict- KT, et al. LKB1/STK11 inactivation leads to expansion of a prometa- ing drug responsiveness in human cancers using genetically engi- static tumor subpopulation in melanoma. Cancer Cell 2012;21:751–64. neered mice. Clin Cancer Res 2013;19:4889–99. 24. Johnson L, Greenbaum D, Cichowski K, Mercer K, Murphy E, Sch- 46. Ackermann J, Frutschi M, Kaloulis K, McKee T, Trumpp A, Beermann mitt E, et al. K-ras is an essential gene in the mouse with partial F. Metastasizing melanoma formation caused by expression of acti- functional overlap with N-ras. Genes Dev 1997;11:2468–81. vated N-RasQ61K on an INK4a-deficient background. Cancer Res 25. Li Q, Haigis KM, McDaniel A, Harding-Theobald E, Kogan SC, Akagi K, 2005;65:4005–11. et al. Hematopoiesis and leukemogenesis in mice expressing oncogenic 47. VanBrocklin MW, Robinson JP, Lastwika KJ, Khoury JD, Holmen SL. NrasG12D from the endogenous locus. Blood 2011;117:2022–32. Targeted delivery of NRASQ61R and Cre-recombinase to post-natal 26. Damsky WE Jr, Bosenberg M. Mouse melanoma models and cell lines. melanocytes induces melanoma in Ink4a/Arflox/lox mice. Pigment Pigment Cell Melanoma Res 2010;23:853–9. Cell Melanoma Res 2010;23:531–41. 27. Rowan A, Bataille V, MacKie R, Healy E, Bicknell D, Bodmer W, et al. 48. Der CJ, Finkel T, Cooper GM. Biological and biochemical properties Somatic mutations in the Peutz-Jeghers (LKB1/STKII) gene in spo- of human rasH genes mutated at codon 61. Cell 1986;44:167–76. radic malignant melanomas. J Invest Dermatol 1999;112:509–11. 49. Donovan S, Shannon KM, Bollag G. GTPase activating proteins: 28. Guldberg P, thor Straten P, Ahrenkiel V, Seremet T, Kirkin AF, Zeu- critical regulators of intracellular signaling. Biochim Biophys Acta then J. Somatic mutation of the Peutz-Jeghers syndrome gene, LKB1/ 2002;1602:23–45. STK11, in malignant melanoma. Oncogene 1999;18:1777–80. 50. Hanna SC, Krishnan B, Bailey ST, Moschos SJ, Kuan PF, Shimamura 29. Esteve-Puig R, Canals F, Colome N, Merlino G, Recio JA. Uncoupling T, et al. HIF1alpha and HIF2alpha independently activate SRC to of the LKB1-AMPKalpha energy sensor pathway by growth factors promote melanoma metastases. J Clin Invest 2013;123:2078–93. and oncogenic BRAF. PLoS ONE 2009;4:e4771. 51. Stones CJ, Kim JE, Joseph WR, Leung E, Marshall ES, Finlay GJ, et al. 30. Zheng B, Jeong JH, Asara JM, Yuan YY, Granter SR, Chin L, et al. Comparison of responses of human melanoma cell lines to MEK and Oncogenic B-RAF negatively regulates the tumor suppressor LKB1 to BRAF inhibitors. Front Genet 2013;4:66. promote melanoma cell proliferation. Mol Cell 2009;33:237–47. 52. Posch C, Moslehi H, Feeney L, Green GA, Ebaee A, Feichtenschlager 31. Bardeesy N, Sinha M, Hezel AF, Signoretti S, Hathaway NA, Sharpless V, et al. Combined targeting of MEK and PI3K/mTOR effector path- NE, et al. Loss of the Lkb1 tumour suppressor provokes intestinal ways is necessary to effectively inhibit NRAS mutant melanoma in polyposis but resistance to transformation. Nature 2002;419:162–7. vitro and in vivo. Proc Natl Acad Sci U S A 2013;110:4015–20. 32. Yan J, Roy S, Apolloni A, Lane A, Hancock JF. Ras isoforms vary in 53. Smalley KS, Xiao M, Villanueva J, Nguyen TK, Flaherty KT, Letrero their ability to activate Raf-1 and phosphoinositide 3-kinase. J Biol R, et al. CRAF inhibition induces apoptosis in melanoma cells with Chem 1998;273:24052–6. non-V600E BRAF mutations. Oncogene 2009;28:85–94. 33. Voice JK, Klemke RL, Le A, Jackson JH. Four human ras homologs 54. Molhoek KR, Brautigan DL, Slingluff CL. Synergistic inhibition differ in their abilities to activate Raf-1, induce transformation, and of human melanoma proliferation by combination treatment with stimulate cell motility. J Biol Chem 1999;274:17164–70. B-Raf inhibitor BAY43-9006 and mTOR inhibitor Rapamycin. J 34. Walsh AB, Bar-Sagi D. Differential activation of the Rac pathway by Transl Med 2005;3:39. Ha-Ras and K-Ras. J Biol Chem 2001;276:15609–15. 55. Houghton AN, Albino AP, Cordon-Cardo C, Davis LJ, Eisinger M. 35. Quinlan MP, Quatela SE, Philips MR, Settleman J. Activated Kras, but Cell surface antigens of human melanocytes and melanoma. Expres- not Hras or Nras, may initiate tumors of endodermal origin via stem sion of adenosine deaminase binding protein is extinguished with cell expansion. Mol Cell Biol 2008;28:2659–74. melanocyte transformation. J Exp Med 1988;167:197–212. 36. Buhrman G, Wink G, Mattos C. Transformation efficiency of RasQ61 56. Shields JM, Thomas NE, Cregger M, Berger AJ, Leslie M, Torrice C, mutants linked to structural features of the switch regions in the et al. Lack of extracellular signal-regulated kinase mitogen-activated presence of Raf. Structure 2007;15:1618–29. protein kinase signaling shows a new type of melanoma. Cancer Res 37. Pedersen M, Kusters-Vandevelde HV, Viros A, Groenen PJ, Sanchez- 2007;67:1502–12. Laorden B, Gilhuis JH, et al. Primary melanoma of the CNS in 57. Sondermann H, Soisson SM, Boykevisch S, Yang SS, Bar-Sagi D, children is driven by congenital expression of oncogenic NRAS in Kuriyan J. Structural analysis of autoinhibition in the Ras activator melanocytes. Cancer Discov 2013;3:458–69. Son of sevenless. Cell 2004;119:393–405. 38. Xu J, Haigis KM, Firestone AJ, McNerney ME, Li Q, Davis E, et al. Dominant 58. Lenzen C, Cool RH, Wittinghofer A. Analysis of intrinsic and role of oncogene dosage and absence of tumor suppressor activity in Nras- CDC25-stimulated guanine nucleotide exchange of p21ras-nucle- driven hematopoietic transformation. Cancer Discov 2013;3:993–1001. otide complexes by fluorescence measurements. Methods Enzymol 39. Jonsson A, Tuominen R, Grafstrom E, Hansson J, Egyhazi S. High 1995;255:95–109. frequency of p16(INK4A) promoter methylation in NRAS-mutated 59. Herrmann C, Martin GA, Wittinghofer A. Quantitative analysis of the cutaneous melanoma. J Invest Dermatol 2010;130:2809–17. complex between p21ras and the Ras-binding domain of the human 40. Eskandarpour M, Hashemi J, Kanter L, Ringborg U, Platz A, Hansson J. Raf-1 protein kinase. J Biol Chem 1995;270:2901–5. Frequency of UV-inducible NRAS mutations in melanomas of patients 60. Herrmann C, Horn G, Spaargaren M, Wittinghofer A. Differential with germline CDKN2A mutations. J Natl Cancer Inst 2003;95:790–8. interaction of the ras family GTP-binding proteins H-Ras, Rap1A, 41. Kwong LN, Costello JC, Liu H, Jiang S, Helms TL, Langsdorf AE, and R-Ras with the putative effector molecules Raf kinase and et al. Oncogenic NRAS signaling differentially regulates survival and Ral-guanine nucleotide exchange factor. J Biol Chem 1996;271: proliferation in melanoma. Nat Med 2012;18:1503–10. 6794–800. 42. Gao J, Aksoy BA, Dogrusoz U, Dresdner G, Gross B, Sumer SO, et al. 61. Gu H, Lalonde S, Okumoto S, Looger LL, Scharff-Poulsen AM, Gross- Integrative analysis of complex cancer genomics and clinical profiles man AR, et al. A novel analytical method for in vivo phosphate track- using the cBioPortal. Sci Signal 2013;6:pl1. ing. FEBS Lett 2006;580:5885–93. 43. Kim WY, Sharpless NE. Drug efficacy testing in mice. Current Top 62. Shutes A, Der CJ. Real-time in vitro measurement of intrinsic and Ras Microbiol Immunol 2012;355:19–38. GAP-mediated GTP hydrolysis. Methods Enzymol 2006;407:9–22.

December 2014 CANCER DISCOVERY | OF12

Mutation-Specific RAS Oncogenicity Explains NRAS Codon 61 Selection in Melanoma

Christin E. Burd, Wenjin Liu, Minh V. Huynh, et al.

Cancer Discovery Published OnlineFirst September 24, 2014.

Updated version Access the most recent version of this article at: doi:10.1158/2159-8290.CD-14-0729

Supplementary Access the most recent supplemental material at: Material http://cancerdiscovery.aacrjournals.org/content/suppl/2014/09/24/2159-8290.CD-14-0729.DC1

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

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

Permissions To request permission to re-use all or part of this article, use this link http://cancerdiscovery.aacrjournals.org/content/early/2014/11/13/2159-8290.CD-14-0729. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.