novel cancer therapeutics with allosteric …precision medicine. besides the canonical...

Therapeutics, Targets, and Chemical Biology

Novel Cancer Therapeutics with AllostericModulation of the Mitochondrial C-Raf–DAPKComplex by Raf Inhibitor Combination TherapyYi-Ta Tsai1,2, Mei-Jen Chuang2, Shou-Hung Tang2, Sheng-Tang Wu2, Yu-Chi Chen3,Guang-Huan Sun1,2,4, Pei-Wen Hsiao5, Shih-Ming Huang6, Hwei-Jen Lee6,Cheng-Ping Yu4,7, Jar-Yi Ho4,7, Hui-Kuan Lin8,9, Ming-Rong Chen2,4, Chung-Chih Lin2,Sun-Yran Chang2,10, Victor C. Lin3, Dah-Shyong Yu1,2,4, and Tai-Lung Cha1,2,4,6

Abstract

Mitochondria are the powerhouses of cells. Mitochondrial C-Raf is a potential cancer therapeutic target, as it regulates mito-chondrial function and is localized to the mitochondria by its N-terminal domain. However, Raf inhibitor monotherapy caninduce S338 phosphorylation of C-Raf (pC-RafS338) and impedetherapy. This study identified the interaction of C-Raf with S308phosphorylated DAPK (pDAPKS308), which together becamecolocalized in the mitochondria to facilitate mitochondrial re-modeling. Combined use of the Raf inhibitors sorafenib andGW5074 had synergistic anticancer effects in vitro and in vivo, buttargeted mitochondrial function, rather than the canonical Rafsignaling pathway. C-Raf depletion in knockout MEFC-Raf�/� orsiRNA knockdown ACHN renal cancer cells abrogated the cyto-toxicity of combination therapy. Crystal structure simulationshowed that GW5074 bound to C-Raf and induced a C-Rafconformational change that enhanced sorafenib-binding affinity.

In the presence of pDAPKS308, this drug–target interaction com-promised the mitochondrial targeting effect of the N-terminaldomain of C-Raf, which induced two-hit damages to cancercells. First, combination therapy facilitated pC-RafS338 andpDAPKS308 translocation frommitochondria to cytoplasm, lead-ing to mitochondrial dysfunction and reactive oxygen species(ROS) generation. Second, ROS facilitated PP2A-mediateddephosphorylation of pDAPKS308 to DAPK. PP2A then dissoci-ated from the C-Raf–DAPK complex and induced profoundcancer cell death. Increased pDAPKS308 modification was alsoobserved in renal cancer tissues, which correlated with poordisease-free survival and poor overall survival in renal cancerpatients. Besidesmediating the anticancer effect, pDAPKS308mayserve as a predictive biomarker for Raf inhibitors combinationtherapy, suggesting an ideal preclinical model that is worthy ofclinical translation. Cancer Res; 75(17); 3568–82. �2015 AACR.

IntroductionA better understanding of the molecular mechanisms of cancer

causation and progression has led to the development of targeted

therapeutics, which is a step toward personalized clinical andprecision medicine. Besides the canonical Ras–Raf–MEK–ERKsignaling pathway, Raf proteins also participate in the regulationof mitochondrial function (1). In vascular endothelial cells,fibroblast growth factor (FGF) induces C-Raf mitochondrialtranslocation through the phosphorylation of Ser338 (S338) ofC-Raf (pC-RafS338; ref. 2). C-Raf binds directly to the mitochon-dria through its N-terminal domain to regulate mitochondrialshape and cellular distribution (3). Suchmorphologic alterationsin the mitochondrial network are involved in cell migration andCa2þmetabolism, which contribute to the oncogenic processes incancer cells (4–6). The overexpression of Bcl-2 results in therecruitment of C-Raf into the mitochondria for cell survival(1). Mitochondrial targeting of C-Raf confers prosurvival prop-erties that prevent dimerization andphosphorylation ofMST2 (7)and induce the phosphorylation and inactivation of BAD (1),respectively. Furthermore, both C-Raf and B-Raf can induce someof the metabolic features mediated by the mitochondria, irre-spective of the canonical Raf signaling pathway (7–9).

Mitochondria are the powerhouses of cells and play crucialroles in regulating the cell signal transduction that contributes tovarious anabolic reactions. Thus, they are more important tocancer cells exhibiting profound modifications in their energymetabolism. The unique role of Raf proteins in the mitochondriaprovides a strong rationale to targetmitochondrial Raf proteins incancer therapeutics (1, 10, 11). Although Raf inhibitors have been

1Graduate Institute of Medical Sciences, National Defense MedicalCenter,Taipei,Taiwan, Republic of China. 2Division of Urology, Depart-ment of Surgery, Tri-Service General Hospital, National Defense Med-ical Center, Taipei, Taiwan, Republic of China. 3Division of Urology,Department of Surgery, E-DaHospital, Kaohsiung,Taiwan, Republic ofChina. 4Graduate Institutes of Life Sciences, National DefenseMedicalCenter, Taipei,Taiwan, Republic of China. 5Agricultural BiotechnologyResearch Center, Academia Sinica, Taipei, Taiwan, Republic of China.6Department of Biochemistry, National Defense Medical Center, Tai-pei, Taiwan, Republic of China. 7Graduate Institute of Pathology andParasitology, Tri-Service General Hospital, National Defense MedicalCenter, Taipei, Taiwan, Republic of China. 8Department of Molecularand Cellular Oncology, The University of Texas MD Anderson CancerCenter, Houston, Texas. 9The University of Texas Graduate School ofBiomedical Sciences at Houston, Houston, Texas. 10Taipei City Hospi-tal, Taipei, Taiwan, Republic of China.

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

Corresponding Author: Tai-Lung Cha, Division of Urology, Department ofSurgery, Tri-Service General Hospital, National Defence Medical Center, N0.325, Section 2, Cheng-Gong Road, Taipei, Taiwan, Republic of China. Phone:886-2-87927170; Fax: 886-2-87927172; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-14-3264

�2015 American Association for Cancer Research.

CancerResearch

Cancer Res; 75(17) September 1, 20153568

on June 29, 2020. © 2015 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst June 22, 2015; DOI: 10.1158/0008-5472.CAN-14-3264

http://cancerres.aacrjournals.org/

shown to have survival benefits for a variety of cancers, emergingproblems include the lack of long-term therapeutic efficacy andrapid development of evasive-resistance in the clinics (12–15). Rafinhibitor monotherapy promotes S338 phosphorylation of C-Rafand tumor progression (13–15). These facts suggest that Rafinhibitor monotherapy may paradoxically activate C-Raf in themitochondria and impede cancer treatment. Currently, no selec-tive cancer therapeutics that targetmitochondrial Raf are available.

Death-associated protein kinase (DAPK), a serine/threonineprotein kinase, was identified by a genetic screen for positivemediators of cell death (16). It participates in a number ofapoptosis-inducing pathways downstream of interferon-g ,CD95 (Fas), TNFa, and TGFb (17, 18). Besides regulating JNKsignaling, DAPK serves as a mitochondrial sensor to inducecaspase-independent necrotic cell death under oxidative stress(19). However, the detailed molecular mechanisms remainunknown.

The DAPK kinase activity is important for its biologic functionsthat mediate cell death processes. The auto-phosphorylation ofthe Ser308 (S308) residue of DAPK inhibits its kinase catalyticactivity. Dephosphorylation of S308 of DAPK by PP2A enhancesits kinase activity and promotes cell death processes (20, 21).Besides S308 phosphorylation of DAPK, other phosphorylationsites also affect its kinase activity. For instance, ERK andSrc kinasesphosphorylate S735 and Y491/Y492, to activate and inhibit theDAPK kinase activity, respectively (22–24).

DAPK functions as a tumor-suppressor gene, and its expressionis attenuated inmany cancer types by promoter hypermethylation(25). Loss of DAPK kinase function, rather than lack of proteinexpression, is also found in renal cell carcinoma (RCC), lungcancer, and other types of cancer (26–28). Posttranslationalmodification, such as S308 phosphorylation of DAPK, may beone of the reasons for the loss of tumor-suppressor function incancers with DAPK expression.

This study determined that C-Raf interacted with pDAPKS308

and directed it to become colocalized to the mitochondria forregulating mitochondrial remodeling. A novel combination ther-apy, which used the Raf inhibitors sorafenib and GW5074,targeted mitochondrial C-Raf and DAPK complexes instead ofthe canonical Raf signaling pathway. GW5074 bound to C-Rafand induced a C-Raf conformational change that enhanced sor-afenib-binding affinity. In the presence of pDAPKS308, GW5074and sorafenib bound to C-Raf and further induced a conforma-tional change of the N-terminal domain, which compromised itsmitochondrial-targeting effect. This facilitated pDAPKS308 and C-Raf translocation from the mitochondria to the cytoplasm, caus-ingmitochondrial dysfunction and reactive oxygen species (ROS)generation. ROS-facilitated PP2A-mediated dephosphorylationof DAPK at S308 activated DAPK, which then dissociated fromC-Raf–DAPK in the cytoplasm, and induced profound cancer cellnecroptosis. An animal model of orthotopic spontaneous metas-tases of RCC was established to mimic clinical situations. Thismodel demonstrated the synergistic anticancer effect of the com-bination therapy in vivo. This study also identified pDAPKS308 as amediator and predictive biomarker of sorafenib and GW5074combination therapy efficacy for cancer treatment.

Materials and MethodsThe procedural details are in the Supplementary Experimental

Procedures.

Chemicals and antibodiesThe GW5074, PLX7420, L779450, andNAC (N-acetylcysteine)

were purchased from Sigma-Aldrich. Sorafenib and zVAD-fmk(carbobenzoxy-valyl-alanyl-aspartyl-[O-methyl]-fluoromethylk-etone) were purchased from Santa Cruz Biotechnology. Canthar-idin and okadaic acid were purchased from Merck Millipore(KGaA).

The following antibodies were purchased from Cell SignalingTechnology: Akt (#9227), phospho-Akt (Ser473, #9271), phos-pho-ERK1/2 (Thr202/Tyr204, #4377), MEK1 (#9124), phospho-MEK1 (#9121), phospho-B-Raf (Ser445, #269), phospho-C-Raf(Ser259, #9421), phospho-C-Raf (Ser338, #9427), phospho-JNK(Thr183/Tyr185, #9255).GAPDH(#G9545) andphospho-DAPK(Ser308, #D4941) were from Sigma-Aldrich. ERK1/2 (SC-94), B-Raf (SC-5284), VDAC1 (SC-58649), and TOM20 (SC-11415)were from Santa Cruz Biotechnology. DAPK (#3798-1), phos-phor-PP2A (Y307, #1155-1), and a-tubulin (#1878-1) were fromEpitomics. The PP2A C subunit (#05-421) was obtained fromMillipore (KGaA).

Cell cultureTheACHN,A498,A2058,DU-145,HeLa,MDA-MB-453,MDA-

MB-231, MCF-7 293T, A459, NIH-3T3, H1299, A253, HepG2,U87, GBM8401, LN229, 22Rv1, SV-HUC-1, 786-O, Colo-205,HT-29, LNCaP, PC-3, RWPE-1, and WPMY-1 cells were obtainedfrom the Bioresource Collection and Research Centre (BCRC,Taiwan). Mouse embryonic fibroblasts (MEF) immortalized withthe SV40 large T antigen [clonesK2 (C-Raf�/�) andK9 (C-Rafþ/þ)]were obtained fromC. Pritchard (University of Leicester, Leicester,UnitedKingdom; ref. 29).HumanmtDNA-less cells 143B and rho0 143B cell lines were a gift from Y.H. Wei (Institute of Biochem-istry and Molecular Biology, National Yang Ming University,Taipei, Taiwan and Department of Medicine, Mackay MedicalCollege, New Taipei City, Taiwan; ref. 30). The cells were main-tained in culture media as recommended by the ATCC andmaintained at 37�C in a 5% CO2, 95% humidity incubator.

Cell treatmentsPreliminary experiments were performed to determine the

optimum time and dose effect of different C-Raf inhibitor treat-ments. In this study, all inhibitors of C-Raf (10 mmol/L GW5074,10 mmol/L PLX7420, and 10 mmol/L L779450), caspase (zVAD),PP2A, or the ROS scavenger NAC were added 30 minutes beforesorafenib (5 mmol/L) treatment.

Generation of TKI-resistant RCC cell linesApproval to use the cells for culture was obtained from the

patients and the local ethics committee before the study. Tumorsamples from patients were obtained as previously described(31). The cells were grown in RPMI-1640medium supplementedwith 10% fetal bovine serum, 50 U/mL penicillin, and 10 mg/mLstreptomycin (Invitrogen). The cells underwent 30 passages. Thecell lines were identified as RCC cells by the BCRC, Taiwan.

Annexin V/propidium iodide stainingFor necroptosis assays, the cells were washed with phosphate-

buffered saline (PBS), resuspended in a binding buffer from EnzoLife Sciences, and stained with Annexin V–FITC and propidiumiodide (PI) for 15 minutes. Annexin V fluorescence was measuredusing the FACScan flow cytometer (BD Biosciences) and the mem-brane integrity of cells was simultaneously assessed by PI exclusion.

C-Raf Colocalizes with DAPK in the Mitochondria

www.aacrjournals.org Cancer Res; 75(17) September 1, 2015 3569




Figure 1.Combination therapy of sorafenib and GW574 showed synergistically anticancer effect in vitro and in vivo. A, cell death in ACHN (right) and A498 (left) wasmeasured by sulforhodamine B assay at 24, 48, and 72 hours after treatment with the indicated Raf inhibitors (10 mmol/L GW5074, 5 mmol/L sorafenib, 10 mmol/LPLX4720, and 10 mmol/L L779450; mean � SEM; �� , P < 0.01, each experiment triplicated; n ¼ 4). The percentage of cell death after treatment wascompared with DMSO control (DMSO treatment defined as 100% viable). B, tumor volume measurements of ACHN xenografts treated with vehicle, sorafenib at5 mg/kg (by oral lavage), GW5074 at 25 mg/kg (intraperitoneal injection, i.p.), or drug combination once a day (mean � SEM; �� , P < 0.01; n ¼ 8). C, schematicrepresentation shows how the spontaneous metastatic Luc-ACHN-LL cells were established from parental ACHN cells. (Continued on the following page.)

Tsai et al.

Cancer Res; 75(17) September 1, 2015 Cancer Research3570




Wright-Giemsa stainingCells grown on 6-cm dishes were washed in PBS twice and

stained with Giemsa solution (5% Giemsa stain solution, 10%buffered formalin) for 15 minutes. After decanting the stainingsolution, the cells were thoroughly washed in distilled water threetimes and air-dried before being photographed.

Sulforhodamine B growth inhibition assaysThe cells were plated in 96-well plates at a plating density

of 3,000 cells per well. After 24 to 72 hours with differentdrug treatments, the cells were fixed with 10% trichloroaceticacid (Sigma-Aldrich) and stained with 0.4% sulforhodamine B(Sigma-Aldrich). The plates were read in a PerkinElmer Victor3micro-plate reader at 570 nm, as described previously (32).Triplicate wells per condition were evaluated and the data pre-sented are representative of three independent experiments.

Measurement of ROSThe cellswere under single or combined treatment for 24hours.

Intracellular ROS levels were measured by FACScan flow cyto-metry usingCM-H2DCFDA (10mmol/L). Todetectmitochondrialsuperoxide,MitoSOXRed (Invitrogen) staining was following themanufacturer's protocol. The histograms presented are represen-tative of three separate experiments.

ResultsSorafenib and GW5074 combination therapy induced cancercell death in vitro and suppressed tumor growth in vivo

Weexamined the effects of different Raf inhibitors with dosagesfrom 2.5 to 10 mmol/L, including sorafenib, GW5074, PLX7420,and L779450, on the cell viability of ACHN and A498 RCC lines.Raf inhibitor monotherapy showed low cytotoxicity in ACHNand A498 RCC cell lines (Supplementary Fig. S1A). To identifypotential combination therapies with high therapeutic efficacy, 5mmol/L sorafenib, 10 mmol/L GW5074, 10 mmol/L PLX7420, and10 mmol/L L779450were combined in pairs. Sorafenib combinedwith either PLX7420 or L779450, and the other combinationtherapies, only modestly induced cell death (Fig. 1A). However, acombination of sorafenib and GW5074 significantly inducedACHN (95% � 1.8%) and A498 (78% � 2.1%) cell death, ascompared with all other combination therapies (Fig. 1A andSupplementary Fig. S1B).

The antitumor effect of sorafenib and GW7054 combinationtherapy (SG-combination therapy) was further tested in vivo. Weused immunodeficientmice xenograftedwithACHNRCC tumorsto test the in vivo antitumor activity and safety of SG-combinationtherapy (Fig. 1B and Supplementary Fig. S1C). In addition, wegenerated an orthotopic spontaneous metastatic RCC animalmodel in order to mimic clinical situations (Fig. 1C and D andSupplementary Fig. S1D and S1E). Luc-ACHN-LL cells with highmetastatic capability and lethality were orthotopically injectedinto a renal capsule. Within 2 weeks after the injection, metastatic

lesionswere observed under the IVIS Xenogene system. Treatmentwith the sameprotocol as the xenografted experimentswas started(Fig. 1E). SG-combination therapy significantly suppressed theprimary and metastatic lesions, and prolonged the survival ofmice (Fig. 1F–H), which remained healthy and active (Supple-mentary Fig. S1F). Tumors from the mice treated with SG-com-bination therapy showed severe necrosis compared with tumorsfrom mice treated with monotherapy (Supplementary Fig. S1G).

Sorafenib and GW5074 combination therapy induced C-Raf–dependent mitochondrial dysfunction and ROS generation

Because sorafenib and GW5074 might have cellular targets,other than C-Raf, that mediate the anticancer effects of SG-combination therapy, we examined whether C-Raf was essentialfor the effectiveness of SG-combination therapy. Wild-type MEF(MEFC-Raf-WT) and C-Raf knockout MEF (MEFC-Raf�/�) were trea-ted with mono- or SG-combination therapy. SG-combinationtherapy synergistically induced cell death of MEFC-Raf-WT (Fig.2A, right and Supplementary Fig. S2A). However, <32% cell deathof MEFC-Raf�/� was observed under SG-combination therapy,which was similar to sorafenib monotherapy (Fig. 2A, left).Knockdown of C-Raf, but not B-Raf, in ACHN cells attenuatedcell death induced by SG-combination therapy for 24 hoursfrom 42.2%� 2.5% to 14%� 3.2% (Fig. 2B and SupplementaryFig. S2B).

Surprisingly, SG-combination therapy did not prevent S338phosphorylation of C-Raf (pC-RafS338). In addition, the phos-phorylation of the major C-Raf downstream pathway compo-nents MEK, ERK, and Akt (33–35) was not inhibited by SG-combination therapy in MEFC-Raf-WT, MEFC-Raf�/�, ACHN, andA498 cells (Fig. 2C and Supplementary Fig. S2C). Because thecombination therapy still increased pC-RafS338, the impact of pC-RafS338 on the therapeutic efficacy of combination therapy waschecked (Supplementary Fig. S2D and S2E). Cell death inducedby the combination therapy was most prominent in ACHN cellstransfected with C-RafS338D (mimicking S338 phosphorylation;Supplementary Fig. S2F and S2G). In contrast, C-RafS338A (mim-icking S338 nonphosphorylation) attenuated the cell death ofACHN cells under combination therapy (Supplementary Fig. S2Fand S2G). These results indicated that C-Raf was the major targetresponsible for the cytotoxic effect of SG-combination therapy.However, other novel mechanisms mediated by C-Raf, instead ofthe canonical Raf signaling pathways, were responsible for thecytotoxic effects of SG-combination therapy.

Because C-Raf plays an important role in mitochondria, mito-chondrial function and ROS production were examined undermono- or SG-combination therapy. In MEFC-Raf-WT, but notin MEFC-Raf�/�, SG-combination therapy resulted in a 6-foldincrease in ROS production, and decreased mitochondrialmembrane potential (Fig. 2D and E and Supplementary Fig.S3A–S3D). The immunofluorescence assay clearly demon-strated that mitochondrial fragmentation with donut- andblob-shape formation was induced by SG-combination therapy

(Continued.) D, the Kaplan–Meier survival curves show high mortality of mice orthotopically implanted with Luc-ACHN-LL cells compared with mice with parentalLuc-ACHN cells. E, drawings of the experimental set-up. F, images represent the entire time course for selectedmice in each group. Signals are displayed as pseudo-color images (blue, least intense; red, most intense) overlaid on the gray-scale body image of the mouse. G, the Kaplan–Meier analysis of survival for miceorthotopically implanted with Luc-ACHN-LL cells under indicated treatments. H, the course of whole-body tumor burden expressed in terms of the photonicsignal resulting from orthotopic Luc-ACHN-LL cell implantation under different treatments. The total photon emission from the entire body of the mouse wasmeasured and quantified with Xenogen's living imaging software (n ¼ 10; �� , P < 0.01).






Figure 2.SG-combination therapy induced cell death via C-Raf dependent mitochondrial dysfunction and ROS production. A, cell death of MEFC-Raf-WT (right) andMEFC-Raf�/� (left) at 24, 48, and 72 hours after treatment with sorafenib and GW5074 mono- or SG-combination therapy normalized to DMSO-treated cells(mean � SEM; �� , P < 0.01; n ¼ 4). Clonogenic assay (Wright-Giemsa staining) of MEFC-Raf-WT and MEFC-Raf�/� cells was measured after a 7-day treatmentwith sorafenib and GW5074 mono- or SG-combination therapy (representative photographs; n ¼ 3 for each condition). B, ACHN cells transfected withnontargeting siRNA, B-Raf siRNA, or C-Raf siRNA for 24 hours were treated with or without SG-combination therapy for 24 hours. Cell death was assayed bysulforhodamine B (mean � SEM; �� , P < 0.01; n ¼ 3). (Continued on the following page.)

Tsai et al.





in MEFC-Raf-WT, but not in MEFC-Raf�/� (Fig. 2F and Supple-mentary Fig. S3E). The same phenomena were also observedin ACHN and A498 cells treated with SG-combination therapy(Fig. 2G and H). Consistent with the aforementioned resultsfrom the cytotoxic experiments, flow cytometry analysis show-ed an increase of ACHN (53.95% � 3.1%) and A498(64.66% � 3.9%) cells with positive staining for both AnnexinV and PI when cells were exposed to SG-combination therapyfor 24 hours (Fig. 2I and Supplementary Fig. S3F). These resultsdemonstrated that SG-combination therapy induced C-Raf–dependent mitochondrial dysfunction, which likely contributedto the profound cell death.

Sorafenib and GW5074 combination therapy induced two-hitdamage of cancer cells throughROS generation and pDAPKS308

dephosphorylationRecent studies have suggested that DAPK serves as a mito-

chondrial sensor, which activates and triggers cell death afterdephosphorylation of its S308 by PP2A (20, 36). We observedthat S308 phosphorylation of DAPK (pDAPKS308) was signif-icantly reduced in ACHN, A498, and MEFC-Raf-WT cells aftertreatment with SG-combination therapy (Fig. 3A and Supple-mentary Fig. S4A). This reduction in phosphorylation appearedto be dependent upon C-Raf, as SG-combination therapyneither induced profound cell death nor dephosphoryltedpDAPKS308 in MEFC-Raf�/� (Fig. 3A). We also observed thatPP2A was required for the dephosphorylation of pDAPKS308

induced by SG-combination therapy, as PP2A inhibitors effi-ciently suppressed pDAPKS308 dephosphorylation (Fig. 3B).Meanwhile, PP2A inhibitors also attenuated the cell deathinduced by SG-combination therapy (Fig. 3C and Supplemen-tary Fig. S4B). The ROS production caused by SG-combinationtherapy was not rescued by PP2A inhibitors (Fig. 3D), suggest-ing that the dephosphorylation of pDAPKS308 by PP2A was adownstream event to the ROS production.

The ROS scavenger NAC and PEG-SOD erased ROS and atten-uated the cell death induced by SG-combination therapy for 24hours from 46.5% � 5.2% to 13.8% � 2.6%, and prevented thedephosphorylation of pDAPKS308 (Fig. 3E and F and Supplemen-tary Fig. S4Cand S4D).However, PP2A inhibitors only attenuatedthe cell death induced by SG-combination therapy from 45.03%� 3.6% to 20.32% � 4.1% in ACHN cancer cells, which was lessthan an attenuation of cell death that caused by theNACandPEG-SOD ROS scavengers (Fig. 3E).

These results indicated that SG-combination therapy causedtwo-hit damage in cancer cells. Thefirst hit of the SG-combinationtherapy inducedmitochondrial dysfunction and ROS generation.The second hit was the action of ROS, which facilitated the PP2A-mediated dephosphorylation of pDAPKS308 and activated its

death-causing activities. These sequential events caused by SG-combination therapy contributed to the profound cell death.

pDAPKS308 was essential for the anticancer effect of sorafeniband GW5074 combination therapy

We next wanted to determine whether DAPK or pDAPKS308

levels were essential for the cytotoxicity caused by SG-combina-tion therapy. We assayed the therapeutic efficacy of SG-combi-nation therapy and compared it with the total DAPK andpDAPKS308 levels of various cell lines. Normal fibroblasts andepithelial cells had high total DAPK protein, but very low levels ofpDAPKS308, and SG-combination therapy caused low toxicity tothese cells (Fig. 3G, blue bars). Consistent with these results, celldeath induced by SG-combination therapy in different cancercells was positively correlatedwith pDAPKS308 levels, but notwithtotal DAPK protein levels (Fig. 3G and H).

We further investigated the therapeutic efficacy of SG-combi-nation therapy on cancer cells derived from clinical patient (RCC-Sor-001; Supplementary Table S1) and animal models (786-O-T4, ACHN-T2R; Supplementary Fig. S5A–S5C) with evasive resis-tance to sorafenib monotherapy. SG-combination therapy for 24hours induced significant cell death in these cancer cells (RCC-Sor-001: 34.6% � 2.1%, 786-O-T4: 31.2% � 3.2%, ACHN-T2R:42.3%� 5.2%). SG-combination therapy also induced cell deathin HT29 (34.9% � 3.7%; ref. 37) and A2058 (44.7% � 2.8%;ref. 38), which have inherited resistance to Raf inhibitors (Fig. 3G,red bar). All of the resistant cancer cells with sensitivity to SG-combination therapy showed high levels of pDAPKS308. Impor-tantly, because normal cells have low levels of pDAPKS308, SG-combination therapy has the advantage of selective toxicitytoward cancer cells.

DU-145 prostate cancer cells expressed nearly undetectabletotal DAPK and pDAPKS308 levels, and did not respond to SG-combination therapy (Fig. 3G). We took advantage of DU-145cells to examine whether pDAPKS308 (but not DAPK protein)mediated the cell death induced by SG-combination therapy.DU-145 cells overexpressing either DAPKWT (wild-type), DAPKS308D

(mimicking S308 phosphorylation), DAPKS308A (mimickingS308 nonphosphorylation), and DAPKK42A (kinase-dead), weretreatedwith SG-combination therapy.OnlyDAPKS308D enhancedcell death induced by SG-combination therapy in DU-145 cancercells (Fig. 3I and Supplementary Fig. S6A). Furthermore, knock-down of DAPK by siRNA attenuated the cell death induced by SG-combination therapy in ACHN, 786-O, and RCC-Sut-002 cellsthat have DAPK expression and pDAPKS308 modification (Fig. 3Jand Supplementary Fig. S6B). These results indicated thatpDAPKS308 was essential for the cytotoxicity of SG-combinationtherapy, because only cancer cells with pDAPKS308 modificationresponded to the therapy.

(Continued.) The clonogenic assay was performed under aforementioned treatments for 7 days (representative photographs; n ¼ 3 for each condition).Immunoblotting was assayed with indicated antibodies. C, MEF and RCC cells treated with sorafenib, GW5074, or their combination for 24 hours. Cell lysates wereprepared for immunoblotting with indicated antibodies. D, MEF cells were treated with mono- or SG-combination therapy for 24 hours and assayed for ROSproduction by flow cytometry using the fluorescent dye CM-H2DCFDA andMitoSOX (mean� SEM; � , P < 0.05; �� , P <0.001; n¼ 4). E, MEF cells treated withmono-or SG-combination therapy for 24 hours. The decrease of mitochondrial membrane potential (Dcm), represented as JC-1 Red/Green ratio, was determined bystaining with JC-1 and FACS analysis. JC-1 Red, the signal of high Dcm; green, the signal of low Dcm (mean� SEM; � , P < 0.05, n¼ 3). F, fluorescencemicroscopy ofMEF cells treated with DMSO or SG-combination therapy and stained with mitochondria antibody and DAPI (nuclei). Scale bar, 20 mm. G and H, FACS analysis ofincreased ROS levels (G) and mitochondrial membrane potential (Dcm; H) in ACHN and A498 cells treated with 5 mmol/L sorafenib, 10 mmol/L GW5074, or theircombination compared with DMSO-treated levels (mean � SEM; � , P < 0.05; �� , P < 0.01; n ¼ 4). I, RCC Cells treated with SG-combination therapy. Cell deathwas assayed by flow cytometry staining with Annexin V and PI. Each bar represents the mean of three separate experiments � SEM. The clonogenicassay was performed under aforementioned treatment for 7 days (n ¼ 3).






Figure 3.pDAPKS308 expression positively correlated with the therapeutic efficacy of SG-combination therapy. A, ACHN and MEF cells treated with sorafenib, GW5074, ortheir combination for 24 hours. Phosphorylation (pDAPKS308) and expression of DAPK were assayed by immunoblotting. B, ACHN cells treated withSG-combination therapy with or without NAC pretreatment. Total cell lysates were immunoblotted with antibodies against DAPK and pDAPKS308. C, ACHN cellstreated with SG-combination therapy concomitant with or without the PP2A inhibitors cantharidin (C.A.) or okadaic acid (O.A.) for 24 hours; the percentageof cell death was normalized to DMSO-treated cells (mean � SEM, � , P < 0.05; �� , P < 0.01, n ¼ 4). (Continued on the following page.)

Tsai et al.





Combined sorafenib and GW5074 induced pC-RafS338 andpDAPKS308 cytosolic translocation, mitochondrialdysfunction, and ROS generation

Because SG-combination therapy was mitochondria-depen-dent, and C-Raf and pDAPKS308 were essential for the therapeuticefficacy, we hypothesized that C-Raf and DAPK might be colo-calized in the mitochondria. C-Raf and DAPK were localized toboth the cytoplasm and the mitochondria in MEFC-Raf-WT. How-ever, DAPK only localized to the cytoplasm in MEFC-Raf�/� (Fig.4A). SG-combination therapy caused a redistribution of DAPKandC-Raf from themitochondria to the cytoplasm inMEFC-Raf-WT,but not in MEFC-Raf�/� (Fig. 4A). In addition, dephosphorylationof pDAPKS308 was observed after its SG-combination therapy-induced cytoplasmic translocation from themitochondria. TheC-Raf protein and its S338phosphorylationwere not affected by SG-combination therapy. The aforementioned resultswere alsonotedin cancer cells with pDAPKS308 modification, but not in thosewithout (Fig. 4B and Supplementary Fig. S7).

To further investigate the role of pDAPKS308 in SG-combinationtherapy-inducedmitochondrial dysfunction, wild-type ormutantforms of DAPK were overexpressed in DU-145 cells that hadundetectable endogenous pDAPKS308 modification. OnlyDAPKS308D was induced to translocate from the mitochondriato the cytoplasm, concomitant with mitochondrial dysfunctionand ROS generation in DU-145 cells under SG-combinationtherapy [Fig. 4C and D (red rectangle), E and Supplementary Fig.S8]. Knockdown of DAPK also attenuated mitochondrial dys-function and ROS generation induced by SG-combination ther-apy (Fig. 4F). These results indicated that C-Raf directed DAPK tobecome colocalized to themitochondria, and that pDAPKS308wasessential for the selective toxicity of SG-combination therapy-mediated mitochondrial dysfunction and ROS generation incancer cells.

Sorafenib and GW5074 combination therapy disassembled theC-Raf, DAPK, and PP2A complex in the cytoplasm

Because pDAPKS308 played an important role in mediating SG-combination therapy-induced mitochondrial dysfunction, therole of pDAPKS308 in C-Raf–mediated mitochondrial functionwas investigated. Phorbol 12-myristate 13-acetate (PMA) wasshown to activate C-Raf for the regulation of mitochondrialremodeling (3). PMA treatment induced striking mitochondrialremodeling with fragmented mitochondria and perinuclear clus-tering in MEFC-Raf-WT, but not in MEFC-Raf�/�. DU-145 prostatecancer cells overexpressing wild-type or DAPK mutants weretreated with PMA. PMA induced significantly higher mitochon-drial remodeling in DU-145 prostate cancer cells with DAPKS308D

expression, as compared with the other DAPK versions (Fig. 5Aand B and Supplementary Fig. S5A and S5B). These results

suggested that pDAPKS308 cooperated with C-Raf to regulatephysiologic cellular function by enhancing mitochondrialremodeling.

Previous studies showed that PP2A interacts with C-Raf andDAPK (21, 39). Immunoprecipitation (IP) assays showed that C-Raf interacted with DAPK, both in the cytoplasm and in themitochondria (Fig. 5C and Supplementary Fig. S5C). C-Raf andDAPK also interacted with PP2A to form a complex in thecytoplasm. SG-combination therapy reduced the interaction ofPP2A with DAPK and C-Raf, concomitant with the reduction ofpDAPKS308 (Fig. 5D). The IP assay also revealed that DAPKS308D

had a weaker interaction with PP2A than did DAPKWT,DAPKS308A, and DAPKK42A. SG-combination therapy furtherenhanced PP2A dissociation from DAPKS308D (Fig. 5E and Sup-plementary Fig. S5D).

pDAPKS308 served as a good prognostic biomarker for RCC inthe clinics

Because the pDAPKS308 levels were correlated with the thera-peutic efficacy of SG-combination therapy, we investigated thelevels of pDAPKS308 in human cancer tissues. Twenty pairs of RCCtumor samples and their normal matched counterparts wereexamined. By Western blot analysis, 16 out of 20 tumor tissuesshowed high pDAPKS308 levels compared with their normalmatched counterparts (Fig. 6A and B). Immunohistochemical(IHC) analysis of 181 human RCC specimens with tissue arraysshowed that pDAPKS308 level was remarkably higher in cancertissues than in normal renal parenchyma (Fig. 6C). There was nodifference in pDAPKS308 level related to tumor grade or clinicalstage (Fig. 6D).

The Kaplan–Meier analysis of patients with survival datarevealed that the high-level of pDAPKS308 positively correlatedwith poor disease-free survival and poor overall survival (Fig. 6Eand F and Supplementary Table S2). These results demonstratedthat pDAPKS308 could serve as a good prognostic biomarker forRCC. It is worthy to investigate whether pDAPKS308 can serve as apredictive biomarker in the treatment of RCC with SG-combina-tion therapy in clinics.

The three-dimensional structure of sorafenib and GW5074binding to C-Raf predicted how they induced C-Raf/pDAPKS308

cytosolic translocationComputerized simulation was performed to evaluate the struc-

ture of C-Raf bound with different Raf inhibitors. The predictedmaximal binding energy of inhibitors in complexes with a C-Rafkinase domain was most prominent in the combination ofGW5074 and sorafenib (�182 kcal/mol), compared with otherRaf inhibitors. Prior docking of GW5074 led to a deep binding ofthis compound into a hydrophobic pocket formed by Ile355,

(Continued.) D, ACHN cells treated with SG-combination therapy with or without cantharidin and okadaic acid. ROS production was assayed by FACS. E, ACHNcells treated with SG-combination therapy for 24 hours in the presence or absence of 4 hours pretreatment with 10, 20 mmol/L NAC or 100, 150 U PEG-SOS.The percentage of cell death was normalized to DMSO-treated cells (mean � SEM; � , P < 0.05; ��, P < 0.01; n ¼ 4). F, ACHN cells treated with SG-combinationtherapy for 24 hours in the presence or absence of 4 hours pretreatment with 10, 20 mmol/L NAC. Cell lysates were subjected to immunoblotting. G, The celldeath of different cancer cell lines treated with SG-combination therapy for 24 hours (n ¼ 3; top). Data representing the relative intensity of pDAPKS308 areshown in the bar graph (middle). pDAPKS308 and DAPK expressions were assayed by immunoblotting. H, the correlation between levels of pDAPKS308 andgrowth inhibition in cancer cells treated with SG-combination therapy is shown in a regression plot. I, cell death of DU-145 cells stably expressing wild-typeor DAPK mutants was assayed by sulforhodamine B after treatment with SG-combination therapy for 48 hours. (mean � SEM; �� , P < 0.01; n ¼ 3). Theclonogenic assay was performed under aforementioned treatments for 7 days (representative photographs; n ¼ 3 for each condition). J, cell death wasmeasured in ACHN, 786-O, and Rcc-Sut-002 cells transfected with nontargeting siRNA, siDAPK-1, or siDAPK-2 following SG-combination therapy for 24 hours(left). The lysates were Western blotted with indicated antibodies (right).






Val363, Ala373, Leu406, Trp423, and Phe475. More of theextended area was occupied by GW5074 and sorafenib in thebinding pocket (Fig. 7A and Table 1).

The conformation for DAPKWT and DAPKS308D in complexwith C-Raf was derived from the poses with the best E_Rdock of�18.24 and�2.41 kcal/mol, respectively (Fig. 7B). In this bindingmanner, there were a total of nine hydrogen bonds and six salt

bridges formed between DAPKWT and C-Raf, but only sevenhydrogen bonds and four salt bridges were formed betweenDAPKS308D and C-Raf. Interaction of DAPKWT with C-Raf showedan antiparallel sheet (strand 3 and 4) of C-Raf protruding into thebinding pocket of DAPK, leaving the N-terminal and C-terminalends on the protein surface. The –OH group in S308 of DAPKinteracted with the OE2 group in Glu125 of C-Raf at the interface

Figure 4.SG-combination therapy induced mitochondrial dysfunction through pDAPKS308 cytosolic translocation and dephosphorylation. A, MEF cells treated with orwithout SG-combination therapy for 24 hours were fractionated into cytosolic and mitochondrial fractions and then immunoblotted with indicated antibodies.Anti-VDAC1 and anti-a-tubulin antibodies were run to distinguish between mitochondrial and cytosolic fractions, respectively. B, immunoblotting with theindicated antibodies for cytoplasmic and mitochondrial fractions of ACHN and DU-145 cells under vehicle, mono-, or SG-combination treatments for 24 hours,respectively. C, DU-145 cells were stably transfected with wild-type and DAPK mutants, followed by treatment with SG-combination therapy for 24 hours.The cells were stained with CM-H2DCFDA (left) or JC-1 (right) for 30 minutes, respectively. The ROS generation (left) and mitochondrial potential decrease (right)was analyzed and quantified by flow cytometry (� , P < 0.05; ��, P < 0.01). D, DU-145 cells transfected with V5-tagged-vector, V5-tagged-DAPK, or the indicatedmutants were treated with DMSO (control) or SG-combination therapy. The cells were fractionated into cytosolic and mitochondrial fractions, and proteintranslocationwas analyzed by immunoblottingwith the indicated antibodies. E, DU-145 cells stably transfectedwith vector, wild-type, or DAPKmutants and treatedwith DMSO or SG-combination therapy for 24 hours. Cells were stained with mitochondria antibody and nuclear marker DAPI and imaged with fluorescencemicroscopy. Scale bar, 20 mm. F, ACHN cells transfected with nontargeting or DAPK siRNAs for 24 hours were treated with SG-combination therapy for 24 hours.The cells were stained with CM-H2DCFDA or JC-1 for 30 minutes, and ROS generation (left) or mitochondrial potential decrease (right) was analyzed by flowcytometry (� , P < 0.05; �� , P < 0.01).

Tsai et al.





of the protein complex. In contrast, the bindingmanner for C-Rafin the DAPKS308D

–C-Raf complex model was different. Asp sub-stitution in residue 308 ofDAPK changed the interaction betweenthe two protein interfaces, which led C-Raf to adopt an alternativestable conformation inwhich thepositionof the antiparallel sheetin the N-terminus of C-Raf was switched to interact with thebinding pocket of DAPK.

These results demonstrated that the combination of GW5074and sorafenib only induced the conformational change in C-Rafin the presence of pDAPKS308. This change compromised themitochondrial targeting effect of theN-terminal domain of C-Raf,which triggered the translocation of pC-RafS338 and pDAPKS308

from the mitochondria to the cytoplasm. This translocationresulted in mitochondrial dysfunction and ROS generation. ROSfacilitated PP2A-mediated dephosphorylation of pDAPKS308 anddisassociation of PP2A from the C-Raf–DAPK complex in thecystoplasm. Consequently, DAPK activation in the cytoplasm ledto profound cell death (Fig. 7C).

DiscussionNonselective off-target mitochondrial toxicity is a major con-

tributor to the failure of chemotherapeutic agents in clinical

practice (40, 41). Recent studies of mitochondrial involvementin cancer have uncovered a plethora of differences in the structure,genome, and function of these organelles, by comparing meta-static mitochondria with those from nontransformed cells. How-ever, there are limited therapeutic strategies selectively targetingcancer cell mitochondria (42, 43).

This study revealed that pC-RafS338 interacted with pDAPKS308

and directed it to become colocalized in the mitochondria.GW5074 and sorafenib combination therapy (SG-combinationtherapy) resulted in the translocation of pC-RafS338 andpDAPKS308 from themitochondria to the cytoplasm, concomitantwith a decrease in mitochondrial membrane potential andincrease in ROS generation. These results suggested that pC-RafS338 cooperated with pDAPKS308 in the regulation of mito-chondrial function. DU-145 prostate cancer cells with ectopicDAPKS308D expressionweremuchmore sensitive to PMA inducedC-Raf–mediated mitochondrial remodeling, as compared withcells ectopically overexpressingDAPKWT,DAPKK42A orDAPKS308A

. This conclusion is supported by the observation that onlyDAPKS308D, but not DAPKWT, DAPKK42A, and DAPKS308A, couldbe induced to translocate from themitochondria to the cytoplasmby SG-combination therapy, subsequent to the induction ofmitochondrial dysfunction and ROS generation.

Figure 5.SG-Combination therapy induced disassembly of the C-Raf–DAPK–PP2A complex. A, DU-145 cells stably transfected with vector or DAPKS308D were treated with200 nmol/L PMA for 30 minutes. The dynamic mitochondrial remodeling was recorded as donut- or blob-shape mitochondria by staining with mitochondrialantibody and nuclearmarker DAPI underfluorescencemicroscopy. B,DU-145 overexpressing vector, wild-type, orDAPKmutantswere treatedwith 200nmol/L PMAfor 2 hours. The percentage of cells with dynamic mitochondrial remodeling was quantified by counting cells with donut- or blob-shape mitochondria undermicroscopy at each time point (n¼ 200; mean� SEM; � , P < 0.05; �� , P < 0.01; n¼ 3). C, immunoprecipitation of C-Raf from cytosolic and mitochondrial fractions ofACHN cells treated with DMSO or SG-combination therapy for 24 hours were immunoblotted with the indicated antibodies. D, ACHN cells treated with mono-or SG-combination therapy. Endogenous DAPK (top) or C-Raf (bottom) were immunoprecipitated and then immunoblotted with the indicated antibodies.E, ACHN cells transfectedwithV5-tagged-vector, V5-tagged-DAPK, or the indicatedmutantswere treatedwith DMSO (control) or SG-combination therapy. The celllysates were immunoprecipitated with anti-V5 antibody and immunoblotted with the indicated antibodies.






The N-terminal domain of C-Raf is required for its mitochon-drial localization (3). Our computer simulation results demon-strated that only in the presence of pDAPKS308, but not wild-typeor other DAPK mutants, sorafenib and GW5074 bound to C-Rafand induced a conformational change of the N-terminal domain,which compromised its mitochondrial targeting capability. Thisindicated that sorafenib and GW5074 served as C-Raf allostericmodulators, instead of Raf inhibitors, and provides a platform forthe development of new allosteric modulators in the future.Therefore, the detailed mechanisms involved in the regulationof mitochondrial function by pDAPKS308 and C-Raf are biolog-ically important and worthy of further investigation.

The generation of ROS by SG-combination therapy promotedthe PP2A-mediated dephosphorylation of S308 in DAPK. It hasbeen shown that the synthetic retinoid, N-(4-hydroxyphenyl)retinamide (4-HPR), induces ROS production in human leuke-mia cells (44). The 4-HPR–mediated ROS evokes Akt conforma-tion change by forming an intramolecular disulfide bond. Akt issubsequently dephosphorylated at Thr308 and Ser473 by PP2A(45). Whether this mechanism occurs in this study warrantsfurther evaluation.

Other important issues regarding DAPK translocation and itsphosphorylation-based regulation need to be explored. BecauseHSP90 and 14-3-3 are two chaperone proteins that interact withand regulate C-Raf phosphorylation status and function (46, 47),it would be interesting to determine whether these two proteins

are also responsible for the regulation of DAPK translocation andphosphorylation. In addition to S308 phosphorylation of DAPK,other posttranslational phosphorylation sites also affect its kinaseactivity (22). This provides a potential explanation for whynormal cells and some cancer cells expressing very low levels ofpDAPKS308 did not display cell death. The DAPK kinase inacti-vation in those cells might be attributed to other inhibitoryphosphorylation sites. However, why the S308 residue of DAPKis phosphorylated in certain types of cancer and whether S308phosphorylation ofDAPK is due to autophosphorylation or otherkinases warrant further studies.

The expression levels of pDAPKS308 in cancer cells or tumortissues are crucial for SG-combination therapy efficacy in termsof clinical cancer treatment. The detection of pDAPKS308 fromcancer tissues would be a prerequisite before the application ofSG-combination therapy. There are higher expression levels ofpDAPKS308 in RCC tumor tissues than in normal renal paren-chyma, regardless of tumor grade or clinical stage. These find-ings demonstrate that pDAPKS308 is an ideal predictive bio-marker for SG-combination therapy in clinical practice, becauseonly a single factor needs to be assayed in tumor tissuesobtained from either the primary or metastatic lesions. More-over, because pDAPKS308 also mediated the anticancer effectsof SG-combination therapy by targeting mitochondrial func-tion instead of the C-Raf signaling pathway, it is not necessaryto check upstream Ras or Raf mutation status, or other

Figure 6.The clinical significance of pDAPKS308

in RCC. A, immunoblotting analysis ofpDAPKS308 and DAPK expression intumor tissues (T) and adjacent normalcounterparts (N) from RCC patients.B, a representative graph shows therelative intensities of the pDAPKS308

normalized to GAPDH expression(pDAPKS308/GAPDH; �� , P < 0.01).C, the IHC detection of pDAPKS308

expression in human normal and renalcancer tissues; original magnification,�40 (top); �100 (bottom). D,expression and clinical relevance ofpDAPKS308 were assessed in ahuman RCC tissue array. ThepDAPKS308 staining indexes in cohortsof normal renal parenchyma (n ¼ 181)and renal cancer tissues (n ¼ 181)are shown as graphic bars.Immunoreactivity score wasdetermined according to a semi-quantitative method. E, the Kaplan–Meier plots of disease-free (left) andoverall survivals (right) after radicalnephrectomy based on thepDAPKS308 expression index inpatients.


Tsai et al.




Figure 7.Computer simulation and molecular mechanism of C-Raf and DAPK with SG combination therapy. A, comparison of structures simulated for C-Rafkinase domain in complex with sorafenib and GW5074. The structure of the protein is presented as the surface model. Sorafenib and GW5074 areshown as stick models and colored by magenta and green, respectively. B, the wild-type and S308D-mutant DAPK in the docked protein complex arepresented as a surface model colored by interpolated charge. C-Raf is shown as a ribbon model and colored by cyan. The N-terminal antiparallel b-sheet(residues 55–75) is highlighted by magenta color. The conformation for wild-type and mutant DAPK in complex with C-Raf were derived from theposes with the best E_Rdock of �18.24 and �2.41 kcal/mol, respectively. C, schematic view of combined sorafenib and GW5074 inducing two-hit celldamage in sequential events. Left, in the presence of pDAPKS308, sorafenib and GW5074 binding induces C-Raf conformational change, compromisingits N-terminal domain mitochondrial targeting effect, and leading to C-Raf/pDAPKS308 cytosolic translocation. Right, cytosolic translocation of C-Raf/pDAPKS308 induces mitochondrial dysfunction and ROS generation (first hit), and triggers PP2A-mediated dephosphorylation and activation ofpDAPKS308 (second hit).






compensatory pathways responsible for the inherited or evasiveresistance to Raf inhibitor monotherapy.

The animal model of orthotopic spontaneous RCC metastasesshows high metastatic capability of the kidneys, liver, and brainwithin 4 weeks, and lethality within 6 to 8 weeks. The animalmodel not only provides an ideal preclinical model to demon-strate the synergistic anticancer effect of SG-combination therapyin vivo, but can also be used to elucidate the underlying mechan-isms responsible for the high metastases and lethality of pheno-types. Resistant cancer cell lines derived from clinical patients oranimal models under sorafenib or sunitinib monotherapy mayalso serve as important research tools for illustrating the under-lying mechanisms involved in evasive resistance to anti-VEGFRtyrosine kinase inhibitors.

Raf inhibitor monotherapy induced the S338 phosphoryla-tion of C-Raf and promoted tumor progression (13). There-fore, allosteric inhibitors were developed to prevent S338phosphorylation of C-Raf for cancer treatment in a preclinicalstudy (48). SG-combination therapy might also induce C-Rafconformational changes by allosteric regulation mechanisms,while still inducing the S338 phosphorylation of C-Raf. Oursimulation program predicted two important issues: first,GW5074 bound to C-Raf and induced the conformationalchange of C-Raf, which enhanced the binding affinity ofsorafenib to C-Raf; and second, only in the presence ofpDAPKS308 did sorafenib and GW5074 bound to C-Raf alterits N-terminal conformation, which likely compromised the C-

Raf mitochondrial targeting effect. The biologic function of themitochondrial C-Raf/DAPK complex and its allosteric regula-tion by Raf inhibitors and other potential chemicals needfurther investigation.

Our animal model showed that only 20% of regular sorafenibdosage combined with GW5074 efficiently suppressed primaryandmetastatic tumors in vivo. This could provide an advantage toreduce the dose-related side effects of sorafenib in clinical practice(49). This finding also raises an important question as to whetherthe sequence of drug application of GW5074 and sorafenib hasany impact on the therapeutic efficacy of SG-combination therapybefore its clinical translation. The in vitro data show that theaddition of GW5074 followed by sorafenib has the advantageof stronger growth inhibition of cancer cells within 24 hours.However, there is no difference in growth inhibition in terms ofthe sequence of drug application of GW5074 and sorafenib whentreating cancer cells over 24 hours (data not shown). The impacton the sequence of drug application of sorafenib and GW5074 inclinical patients needs further investigation in thoughtfullydesigned clinic trials.

This study demonstrated the effectiveness of a novel combi-nation therapy with sorafenib and GW5074 for cancer therapy.These mechanisms by which these two inhibitors interacted withtheir target protein, C-Raf, provided an idealmodel for a newdrugdesignation. This study also illuminated a unique interplayamong C-Raf, DAPK, and PP2A in regulating mitochondrialfunction in cell death cascades. The safety issue of SG-combina-tion therapy is highly promising due to its selective toxicity towardcancer cells over normal cells. This study also identifiedpDAPKS308 as a biomarker for predicting the therapeutic efficacyof SG-combination therapy to prevent unnecessary treatments. Aspontaneous metastatic animal model used to mimic humancancer disease demonstrated the anticancer efficacy of combina-tion therapy in vivo. This study shows away to overcome obstaclesencountered in the current cancer therapeutics, and fulfills thecriteria of an ideal preclinical therapeutic model worthy of furtherclinical translation.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: Y.-T. Tsai, M.-J. Chuang, G.-H. Sun, C.-C. Lin,S.-Y. ChangDevelopment of methodology: Y.-T. Tsai, M.-R. Chen, C.-C. LinAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): Y.-T. Tsai, S.-T. Wu, Y.-C. Chen, C.-P. Yu, J.-Y. Ho,V.C. LinAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): Y.-T. Tsai, S.-H. Tang, G.-H. Sun, S.-M. Huang,H.-J. Lee, C.-C. LinWriting, review, and/or revision of the manuscript: Y.-T. Tsai, S.-H. Tang,G.-H. Sun, P.-W. Hsiao, S.-M. Huang, D.-S. Yu, T.-L. ChaAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): Y.-T. Tsai, M.-J. Chuang, S.-M. Huang,M.-R. ChenStudy supervision: H.-K. Lin, S.-Y. ChangOther (performed the experiments): C.-C. Lin

AcknowledgmentsThe authors thank Dr. Ruey-Hwa Chen for providing the plasmids of

pcDNA3-flag-DAPK and pcDNA3-flag-DAPK-K42A, Dr. Catrin Pritchard forthe C-Raf wild-type and knockout MEF, Dr. Yau-Huei Wei for mtDNA-lesscells 143B and rho 0 143B cells. The authors also thank Drs. Hsing-Jien Kung,

Table 1. Binding energy of Raf inhibitors in complex with C-Raf

InhibitorsName Structure

Binding energy,kcal/mol

SM52

NN

N

NH

N OH

�76

Sorafenib

Cl

FF

F

NH

NHO

O

N

NH

O

�82

þ L-779450 �134þ GW5074 �125þ PLX44720 �113Gw5074

N

I O

Br

Br

O�128

þ Sorafenib �182PLX44720

N N

Cl

OF

F N SO

O

�77

þ Sorafenib �140L-779450

Cl

O

N

N

N �60

þ Sorafenib �125


Tsai et al.




Jer-Tsong Hsieh, and Huey-Kang Sytwu for critically reading and editing thisarticle.

Grant SupportThis study was supported by grants from the National Science Council (NSC

101-2314-B-016-006,NSC96-2826-B-016-001-MY3, andNSC99-2826-B-016-012-MY3) and grants from the Tri-Service General Hospital Research Founda-tion (TSGH-C101-009-S02 and TSGH-C102-007-009-S02) and Ministry of

National Defence-Medical Affairs Bureau (DOD98-08-02), Taipei, Taiwan,Republic of China.

The costs of publication of this article were defrayed in part by the paymentof page charges. This article must therefore be hereby marked advertisementin accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received November 10, 2014; revised April 15, 2015; accepted May 4, 2015;published OnlineFirst June 22, 2015.

References1. Wang HG, Rapp UR, Reed JC. Bcl-2 targets the protein kinase Raf-1 to

mitochondria. Cell 1996;87:629–38.2. AlaviA,Hood JD, FraustoR, StupackDG,ChereshDA.RoleofRaf in vascular

protection from distinct apoptotic stimuli. Science 2003;301:94–6.3. Galmiche A, Fueller J, Santel A, Krohne G, Wittig I, Doye A, et al. Isoform-

specific interaction of C-RAF with mitochondria. J Biol Chem 2008;283:14857–66.

4. Campello S, Lacalle RA, Bettella M, Manes S, Scorrano L, Viola A. Orches-tration of lymphocyte chemotaxis by mitochondrial dynamics. J Exp Med2006;203:2879–86.

5. Szabadkai G, Simoni AM, Bianchi K, De Stefani D, Leo S, Wieckowski MR,et al. Mitochondrial dynamics and Ca2þ signaling. Biochim Biophys Acta2006;1763:442–9.

6. Alirol E, Martinou JC. Mitochondria and cancer: is there a morphologicalconnection? Oncogene 2006;25:4706–16.

7. O'Neill E, Rushworth L, Baccarini M, Kolch W. Role of the kinase MST2 insuppression of apoptosis by the proto-oncogene product Raf-1. Science2004;306:2267–70.

8. LeMellay V,Houben R, Troppmair J, HagemannC,Mazurek S, FreyU, et al.Regulation of glycolysis by Raf protein serine/threonine kinases. AdvEnzyme Regul 2002;42:317–32.

9. Haq R, Shoag J, Andreu-Perez P, Yokoyama S, Edelman H, Rowe GC, et al.Oncogenic BRAF regulates oxidativemetabolism via PGC1alpha andMITF.Cancer Cell 2013;23:302–15.

10. Zebisch A, Troppmair J. Back to the roots: the remarkable RAF oncogenestory. Cell Mol Life Sci 2006;63:1314–30.

11. Troppmair J, Bruder JT, Munoz H, Lloyd PA, Kyriakis J, Banerjee P, et al.Mitogen-activated protein kinase/extracellular signal-regulated proteinkinase activation by oncogenes, serum, and 12-O-tetradecanoylphorbol-13-acetate requires Raf and is necessary for transformation. J Biol Chem1994;269:7030–5.

12. Solit DB, Rosen N. Resistance to BRAF inhibition in melanomas. N Engl JMed 2011;364:772–4.

13. Hatzivassiliou G, Song K, Yen I, Brandhuber BJ, Anderson DJ, Alvarado R,et al. RAF inhibitors prime wild-type RAF to activate the MAPK pathwayand enhance growth. Nature 2010;464:431–5.

14. Heidorn SJ, Milagre C,Whittaker S, Nourry A, Niculescu-Duvas I, DhomenN, et al. Kinase-dead BRAF and oncogenic RAS cooperate to drive tumorprogression through CRAF. Cell 2010;140:209–21.

15. Poulikakos PI, Zhang C, Bollag G, Shokat KM, Rosen N. RAF inhibitorstransactivate RAF dimers and ERK signalling in cells with wild-type BRAF.Nature 2010;464:427–30.

16. Cohen O, Feinstein E, Kimchi A. DAP-kinase is a Ca2þ/calmodulin-depen-dent, cytoskeletal-associated protein kinase, with cell death-inducing func-tions that depend on its catalytic activity. EMBO J 1997;16:998–1008.

17. Jang CW, Chen CH, Chen CC, Chen JY, Su YH, Chen RH. TGF-beta inducesapoptosis through Smad-mediated expression ofDAP-kinase. NatCell Biol2002;4:51–8.

18. Inbal B, CohenO, Polak-Charcon S, Kopolovic J, Vadai E, Eisenbach L, et al.DAP kinase links the control of apoptosis to metastasis. Nature 1997;390:180–4.

19. Eisenberg-Lerner A, Kimchi A. DAP kinase regulates JNK signaling bybinding and activating protein kinase D under oxidative stress. Cell DeathDiffer 2007;14:1908–15.

20. Widau RC, Jin Y, Dixon SA, Wadzinski BE, Gallagher PJ. Protein phospha-tase 2A (PP2A) holoenzymes regulate death-associated protein kinase(DAPK) in ceramide-induced anoikis. J Biol Chem 2010;285:13827–38.

21. Guenebeaud C, Goldschneider D, Castets M, Guix C, Chazot G, Delloye-Bourgeois C, et al. The dependence receptor UNC5H2/B triggers apoptosis

via PP2A-mediated dephosphorylation of DAP kinase. Mol Cell 2010;40:863–76.

22. Lin Y, Hupp TR, Stevens C. Death-associated protein kinase (DAPK) andsignal transduction: additional roles beyond cell death. FEBS J 2010;277:48–57.

23. Wang WJ, Kuo JC, Ku W, Lee YR, Lin FC, Chang YL, et al. The tumorsuppressor DAPK is reciprocally regulated by tyrosine kinase Src andphosphatase LAR. Mol Cell 2007;27:701–16.

24. Chen CH, Wang WJ, Kuo JC, Tsai HC, Lin JR, Chang ZF, et al. Bidirectionalsignals transduced by DAPK–ERK interaction promote the apoptotic effectof DAPK. EMBO J 2005;24:294–304.

25. MichieAM,McCaigAM,NakagawaR,VukovicM.Death-associated proteinkinase (DAPK) and signal transduction: regulation in cancer. FEBS J 2010;277:74–80.

26. Liu Y, Gao W, Siegfried JM, Weissfeld JL, Luketich JD, Keohavong P.Promoter methylation of RASSF1A and DAPK and mutations of K-ras,p53, and EGFR in lung tumors from smokers and never-smokers. BMCCancer 2007;7:74.

27. Niklinska W, Naumnik W, Sulewska A, Kozlowski M, Pankiewicz W,Milewski R. Prognostic significance of DAPK and RASSF1A promoterhypermethylation in non–small cell lung cancer (NSCLC). Folia Histo-chem Cytobiol 2009;47:275–80.

28. Christoph F, Kempkensteffen C, Weikert S, Kollermann J, Krause H, MillerK, et al. Methylation of tumour suppressor genes APAF-1 and DAPK-1 andin vitro effects of demethylating agents in bladder and kidney cancer. Br JCancer 2006;95:1701–7.

29. Huser M, Luckett J, Chiloeches A, Mercer K, Iwobi M, Giblett S, et al. MEKkinase activity is not necessary for Raf-1 function. EMBO J 2001;20:1940–51.

30. Liu CY, Lee CF, Hong CH, Wei YH. Mitochondrial DNA mutation anddepletion increase the susceptibility of human cells to apoptosis. Ann N YAcad Sci 2004;1011:133–45.

31. Dangles-Marie V, Pocard M, Richon S, Weiswald LB, Assayag F, Saulnier P,et al. Establishment of human colon cancer cell lines from fresh tumorsversus xenografts: comparison of success rate and cell line features. CancerRes 2007;67:398–407.

32. Zeitlin BD, Joo E, Dong Z, Warner K, Wang G, Nikolovska-Coleska Z, et al.Antiangiogenic effect of TW37, a small-molecule inhibitor of Bcl-2. CancerRes 2006;66:8698–706.

33. Brummer T, Naegele H, Reth M, Misawa Y. Identification of novel ERK-mediated feedback phosphorylation sites at the C-terminus of B-Raf.Oncogene 2003;22:8823–34.

34. ParaisoKH, Fedorenko IV, Cantini LP,MunkoAC,HallM, SondakVK, et al.Recovery of phospho-ERK activity allows melanoma cells to escape fromBRAF inhibitor therapy. Br J Cancer 2010;102:1724–30.

35. Ritt DA, Monson DM, Specht SI, Morrison DK. Impact of feedbackphosphorylation and Raf heterodimerization on normal and mutantB-Raf signaling. Mol Cell Biol 2010;30:806–19.

36. Shang T, Joseph J, Hillard CJ, Kalyanaraman B. Death-associated proteinkinase as a sensor of mitochondrial membrane potential: role of lysosomein mitochondrial toxin-induced cell death. J Biol Chem 2005;280:34644–53.

37. Prahallad A, SunC,Huang S, Di Nicolantonio F, Salazar R, ZecchinD, et al.Unresponsiveness of colon cancer to BRAF(V600E) inhibition throughfeedback activation of EGFR. Nature 2012;483:100–3.

38. Xing F, Persaud Y, Pratilas CA, Taylor BS, Janakiraman M, She QB, et al.Concurrent loss of the PTEN and RB1 tumor suppressors attenuates RAFdependence in melanomas harboring (V600E)BRAF. Oncogene 2012;31:446–57.






39. AbrahamD, Podar K, PacherM, KubicekM,Welzel N, Hemmings BA, et al.Raf-1-associated protein phosphatase 2A as a positive regulator of kinaseactivation. J Biol Chem 2000;275:22300–4.

40. Verschoor ML, Ungard R, Harbottle A, Jakupciak JP, Parr RL, Singh G.Mitochondria and cancer: past, present, and future. Biomed Res Int2013;2013:612369.

41. Dykens JA, Marroquin LD, Will Y. Strategies to reduce late-stage drugattrition due to mitochondrial toxicity. Expert Rev Mol Diagn 2007;7:161–75.

42. Fulda S, Galluzzi L, Kroemer G. Targetingmitochondria for cancer therapy.Nat Rev Drug Discov 2010;9:447–64.

43. Smith RA, Murphy MP. Mitochondria-targeted antioxidants as therapies.Discov Med 2011;11:106–14.

44. Goto H, Takahashi H, Fujii H, Ikuta K, Yokota S. N-(4-Hydroxyphenyl)retinamide (4-HPR) induces leukemia cell death via generation of reactiveoxygen species. Int J Hematol 2003;78:219–25.

45. Cao J, Xu D, Wang D, Wu R, Zhang L, Zhu H, et al. ROS-driven Aktdephosphorylation at Ser-473 is involved in 4-HPR-mediated apoptosis inNB4 cells. Free Radic Biol Med 2009;47:536–47.

46. McCubrey JA, Steelman LS, Chappell WH, Abrams SL, Wong EW, Chang F,et al. Roles of the Raf/MEK/ERK pathway in cell growth, malignanttransformation and drug resistance. Biochim Biophys Acta 2007;1773:1263–84.

47. TzivionG, Luo Z, Avruch J. A dimeric 14-3-3 protein is an essential cofactorfor Raf kinase activity. Nature 1998;394:88–92.

48. Mielgo A, Seguin L, Huang M, Camargo MF, Anand S, Franovic A, et al. AMEK-independent role for CRAF in mitosis and tumor progression. NatMed 2011;17:1641–5.

49. Wilhelm SM, Adnane L, Newell P, Villanueva A, Llovet JM, Lynch M.Preclinical overview of sorafenib, a multikinase inhibitor that targets bothRaf and VEGF and PDGF receptor tyrosine kinase signaling. Mol CancerTher 2008;7:3129–40.


Tsai et al.




2015;75:3568-3582. Published OnlineFirst June 22, 2015.Cancer Res Yi-Ta Tsai, Mei-Jen Chuang, Shou-Hung Tang, et al. Therapy

DAPK Complex by Raf Inhibitor Combination−Mitochondrial C-Raf Novel Cancer Therapeutics with Allosteric Modulation of the

Updated version

10.1158/0008-5472.CAN-14-3264doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2015/06/23/0008-5472.CAN-14-3264.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/75/17/3568.full#ref-list-1

This article cites 49 articles, 15 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/75/17/3568.full#related-urls

This article has been cited by 2 HighWire-hosted articles. Access the articles at:

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

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/75/17/3568To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-14-3264

http://cancerres.aacrjournals.org/content/suppl/2015/06/23/0008-5472.CAN-14-3264.DC1

http://cancerres.aacrjournals.org/content/75/17/3568.full#ref-list-1

http://cancerres.aacrjournals.org/content/75/17/3568.full#related-urls

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://cancerres.aacrjournals.org/content/75/17/3568


novel cancer therapeutics with allosteric …precision medicine. besides the canonical...

Documents