trib3 stabilizes high twist1 expression to promote rapid ... · translational cancer mechanisms and...

Translational Cancer Mechanisms and Therapy

TRIB3 Stabilizes High TWIST1 Expression toPromote Rapid APL Progression and ATRAResistanceJian Lin1,Wu Zhang1, Li-Ting Niu1, Yong-Mei Zhu1, Xiang-Qin Weng1, Yan Sheng1,Jiang Zhu1, and Jie Xu1,2

Abstract

Purpose: The resistance to differentiation therapy andearly death caused by fatal bleeding endangers the healthof a significant proportion of patients with acute promye-locytic leukemia (APL). This study aims to investigate themolecular mechanisms of all-trans retinoic acid (ATRA)resistance and uncover new potential therapeutic strategiesto block the rapid progression of early death.

Experimental Design: The important role of TWIST1 inAPL leukemogenesis was first determined by gain- and loss-of-function assays. We then performed in vivo and in vitroexperiments to explore the interaction of TWIST1 and TRIB3and develop a potential peptide-initiated therapeutic oppor-tunity to protect against early death and induction therapyresistance in patients with APL.

Results:We found that the epithelial–mesenchymal tran-sition (EMT)-inducing transcription factor TWIST1 is highlyexpressed in APL cells and is critical for leukemic cell

survival. TWIST1 and TRIB3 were highly coexpressed in APLcells compared with other subtypes of acute myeloidleukemia cells. We subsequently demonstrated that TRIB3could bind to the WR domain of TWIST1 and contribute toits stabilization by inhibiting its ubiquitination. TRIB3depletion promoting TWIST1 degradation reverses resis-tance to induction therapy and improves sensitivityto ATRA. On the basis of a detailed functional screenof synthetic peptides, we discovered a peptide analogous tothe TWIST1 WR domain that specifically represses APL cellsurvival by disrupting the TRIB3/TWIST1 interaction.

Conclusions:Our data not only define the essential role ofTWIST1as anEMT-TF inpatientswithAPLbut also suggest thatdisrupting the TRIB3/TWIST1 interaction reverses inductiontherapy resistance and blocks rapid progression of APL earlydeath.

See related commentary by Peeke and Gritsman, p. 6018

IntroductionAcute promyelocytic leukemia (APL), which accounts for

10%–15% of acute myeloid leukemia (AML) cases, is character-ized by the t(15; 17) chromosomal translocation and is nowhighly curable by the combination of granulocytic differentiationinduction and the PML-RARa oncoprotein–targeted agents all-trans retinoic acid (ATRA) and arsenic trioxide (ATO; refs. 1, 2).Despite the striking molecular complete remission (CR) and thevery few cases of relapse associated with ATRA/ATO–based regi-

mens, mortality events typically result from early fatal bleeding,which remains the most important challenge and the largestobstacle to curing all patients with APL (3). For instance, severalstudies have reported that the risk of early hemorrhagic death(HD) reaches an incidence of 10%–20%during the firstmonth ofinduction (4–6). Importantly, patientswithAPLwith ahighwhiteblood cell count face an increased risk of early HD (7). Further-more, resistance to ATRA/ATO treatment with PML-RARa muta-tions still remains a therapeutic challenge for a significant pro-portion of patients with APL. Thus, we need to better understandthe molecular mechanism of APL pathogenesis and design moreeffective therapeutic strategies to block the rapid progression ofearly death and overcome resistance.

Oncogenic transcription factors play an important role in thedevelopment of hematologic malignancy (8, 9). The dysregula-tion of epithelial–mesenchymal transition-inducing transcrip-tion factors (EMT-TF), including SNAI1/SNAI2, ZEB1/ZEB2, andTWIST1/TWIST2, has also been explored within the context of theaggressive invasion, chemoresistance, and poor prognosis ofAML (10–12). TWIST1, a highly conserved basic helix-loop-helix (bHLH) protein, is a well-characterized EMT-TF that playsa critical role in embryonic development and cancer metasta-sis (13, 14). Recent studies have shown that TWIST1 is over-expressed in primary AML samples (15). Our previous studyreported that high expression of TWIST1 in AML contributes toextramedullary infiltration and promotes leukemic aggres-siveness (16). These findings strongly suggest that disruption ofTWIST1 is involved in leukemogenesis. However, the prognosis of

1State Key Laboratory for Medical Genomics, Shanghai Institute of Hematology,National Research Center for Translational Medicine, Rui-Jin Hospital, ShanghaiJiao-Tong University School of Medicine and School of Life Sciences andBiotechnology, Shanghai Jiao-Tong University, Shanghai, China. 2TranslationalMedicine Ward, Department of Hematology, Rui-Jin Hospital, Shanghai Jiao-Tong University School of Medicine, Shanghai, China.

Note: Supplementary data for this article are available at Clinical CancerResearch Online (http://clincancerres.aacrjournals.org/).

J. Lin and W. Zhang contributed equally to this article.

Corresponding Authors: Jie Xu, Rm 706, Science and Technology Plaza, No.197Rui Jin Er Road, Shanghai, China 200025. Phone: 86-21-64370045; Fax: 86-21-64743206; E-mail: [email protected]; and Wu Zhang, Rm 706, Science andTechnology Plaza, No.197 Rui Jin Er Road, Shanghai, China 200025. Phone: 86-21-64370045; Fax: 86-21-64743206; E-mail: [email protected]

Clin Cancer Res 2019;25:6228–42

doi: 10.1158/1078-0432.CCR-19-0510

�2019 American Association for Cancer Research.

ClinicalCancerResearch

Clin Cancer Res; 25(20) October 15, 20196228

on July 2, 2020. © 2019 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst June 24, 2019; DOI: 10.1158/1078-0432.CCR-19-0510

http://crossmark.crossref.org/dialog/?doi=10.1158/1078-0432.CCR-19-0510&domain=pdf&date_stamp=2019-9-19

http://crossmark.crossref.org/dialog/?doi=10.1158/1078-0432.CCR-19-0510&domain=pdf&date_stamp=2019-9-19

http://clincancerres.aacrjournals.org/

patients with AML with high TWIST1 expression is controversialbased on several reports (15, 17). These conflicting reports likelyoccurred because TWIST1 expression in AML is heterogeneous (18).TWIST1 expression is higher in APL, an intriguing AML subtype inwhich successful clinical tumor differentiation-induction therapyhas been applied and for which the clinical phenotype is inconsis-tent with the expression level of TWIST1 (15, 18).

In this study, we demonstrate that the EMT-TF TWIST1 issignificantly highly expressed in APL and governs the survival ofAPL cells. We show that TRIB3 interacts with TWIST1 and stabi-lizes high TWIST1 expression by repressing TWIST1 ubiquitina-tion. Our data also suggest that a peptide similar to the WRdomain disturbs the TRIB3/TWIST1 interaction, impairs rapidprogression during the early death of APL, and reverses resistanceto ATRA therapy. These results contribute to a better understand-ing of the molecular mechanism of APL pathogenesis and willallow for designing more effective therapeutic strategies to blockthe rapid progression of early death and overcome resistance.

Materials and MethodsCells and mice

Leukemic cell lines were cultured in RPMI1640 medium(Invitrogen) supplemented with 10% FBS (Invitrogen), andHEK293T cells were grown in DMEM (Invitrogen) supplemen-ted with 10% FBS. All cell lines, including NB4-R1 and PR9 cellline, were obtained from the Shanghai Institute of Hematology(Shanghai, China) as described previously (19–21). NB4-R1, ade novo ATRA-resistant cell line isolated from parental NB4cells, was obtained from Dr. Michel Lanotte (Hospital SaintLouis, Paris, France; ref. 22). HMRP8-PML-RARa transgenicmice were generated on an FVB/NJ background using thehuman MRP8 expression cassette (23). The reproducible APLtransplantation model can be propagated by injecting APLblasts, isolated from hMRP8-PML-RARa transgenic mice, intothe tail vein of syngeneic recipients (24). All transplanted APLcells were highly purified by flow-sorting to exclude the dying

cells, and then injected into the immunodeficient mice orsyngeneic recipients. NB4-luc or R1-luc cells were injected intosublethally irradiated 8-week-old NOD/SCID mice through tailveins. All animal experiments were conducted in accordancewith the approved guidelines provided by the LaboratoryAnimal Resource Center of Shanghai Jiao-Tong UniversitySchool of Medicine (Shanghai, China).

Patient samplesPrimary AML samples were obtained from the bone marrow

of patients with diagnosed AML. Leukemic blasts were purifiedand harvested in the mononuclear layer via density gradientcentrifugation. Human primary AML samples were mainlyobtained from Shanghai Rui-Jin Hospital (Shanghai, China)with written informed consent from each patient and researchethics board approval in accordance with the Declaration ofHelsinki.

Flow cytometry analysisCells were suspended in FACS buffer containing 1% FBS and

0.1%NaN3. The data were collected on an LSR-Fortessa X20 FlowCytometer (BDBiosciences). Antibodies were purchased fromBDBiosciences, including anti-CD11b and Annexin V/PI apoptosisflow kit.

Morphologic stainingMurine PB or bone marrow (BM) cells were cytospun onto

slides and stained with Wright-Giemsa staining solution byfollowing manufacturer's manual. The samples were evaluatedunder a light microscope (BX61, Olympus).

b-Gal stainingSenescence-associated b-galactosidase (SA-b-Gal) staining was

detected with the Senescence Detection Kit (Abcam). The senes-cent cells were quantitated from at least six random fields accord-ing to the manufacturers' protocols.

Colony formation unit assayTo evaluatemethylcellulose colony-forming unit (CFU) colony

numbers in human or mouse leukemic cells, highly purifiedsorted cells were plated in duplicate and cultured in MethoCultmedium (Stem Cell Technologies) in 12-plate dishes. On day 11,CFUs were counted from three independent experiments usingthe manufacturers' protocols.

shRNA viral vector construction and deliveryThe TWIST1 or TRIB3 shRNA sequence was converted from a

pair of previously reported shRNA oligos (16, 25). To generatecells stably expressing TWIST1-shRNA, TRIB3-shRNA, and thenegative control-shRNA (NC), the expression cassettes were trans-duced into leukemic cells with lentiviral vectors. siRNA oligonu-cleotides against TWIST1 were transfected using Lipofectamine2000 (Invitrogen).

qRT-PCR analysisTotal cellular RNA was extracted with RNeasy micro kit or

RNeasy Mini Kit (Qiagen) by following the manufacturer's man-ual. For qRT-PCR, reactions were performed by using SYBRPremix Ex Taq (Applied Takara Bio Inc.) on an ABI 7500Real-Time PCR system. The primer sequences of reference geneGAPDH were described previously (26).

Translational Relevance

This study was designed to improve the treatment forpatients with high-risk acute promyelocytic leukemia (APL).This proportion of patients with APL continued to suffer fromearly fatal bleeding and leukemic extramedullary infiltration,which remains the most important challenge and the largestobstacle to curing all patients with APL. We demonstrate thatthe unique EMT-inducing transcription factor TWIST1 is sig-nificantly highly expressed in APL and governs the survival ofAPL cells. We show that TRIB3 interacts with TWIST1 andstabilizes high TWIST1 expression by repressing TWIST1 ubi-quitination. Our data also suggest that a peptide similar to theWR domain disturbs the TRIB3/TWIST1 interaction, impairsrapid progression during the early death of APL, and reversesresistance to all-trans retinoic acid therapy. These results revealthe important role of a specific oncogenic transcriptionalfactor, TWIST1, in APL leukemogenesis and suggest a potentialpeptide-initiated therapeutic opportunity to protect againstearly death and induction therapy resistance in patients withAPL.

TRIB3 Stabilizes TWIST1 to Promote APL Progression

www.aacrjournals.org Clin Cancer Res; 25(20) October 15, 2019 6229




Figure 1.

High TWIST1 expression in APL. A, Relative TWIST1mRNA expression levels in 199 human AML cases with six different AML cytogenetic aberrations from thegene expression microarray of the E-MTAB-3444 database, including t (15;17; n¼ 26), t (8;21; n¼ 46), inv (16; n¼ 48), MLL (11q23; n¼ 43), t (9;11; n¼ 21), and inv(3)/t (3;3; n¼ 15; see also Supplementary Fig. S1A, �� , P < 0.01). B, In silico analysis of TWIST1mRNA expression levels in different FAB subtypes of patients withAML (n¼ 173), including M0 (n¼ 16), M1 (n¼ 44), M2 (n¼ 38), M3 (n¼ 16), M4 (n¼ 34), M5 (n¼ 18), M6 (n¼ 2), M7 (n¼ 3), and not classified (NC, n¼ 2). Theraw RNA-seq data were obtained from the TCGA database (� , P < 0.05; �� , P < 0.01, N/A: not applicable). C, qRT-PCR assay of the mRNA expression levels ofTWIST1 in primary blasts from newly diagnosed APL M3 patients (n¼ 22) versus CD34þ BM cells from healthy donors (n¼ 6) and primary blasts from other AMLsubtypes, including M1 (n ¼ 7), M2 (n ¼ 14), M4 (n ¼ 23), M5 (n ¼ 18), and M6 (n ¼ 6; Data represent the mean � SEM of three assays, � , P < 0.05;�� , P < 0.01). (Continued on the following page.)

Lin et al.

Clin Cancer Res; 25(20) October 15, 2019 Clinical Cancer Research6230




Western blotting, IHC staining, and immunofluorescenceThe resulting cell pellets were lysed with the whole-cell lysis

buffer in the boiling water for at least 10 minutes. The anti-bodies used included anti–b-Actin (Sigma), anti-PML-RARa(Abcam), anti-TRIB3 (Proteintech), and anti-TWIST1 (SantaCruz Biotechnology). Protein signals were visualized usingthe Immobilon Western Kit (Millipore). Murine brain sectionsand spinal cord slices were prepared for HE staining or humanCD45 (Cell Signaling Technology) IHC staining. Cells orfrozen tissue samples were fixed in 4% PFA and then permea-bilized in 0.2% Triton X-100. Murine brain sections and spinalcord slices were prepared for murine c-Kit (Santa CruzBiotechnology) IHC staining. Optical sections of the cells wereobserved under a Leica TCS SP8 Confocal Microscope (LeicaMicrosystems).

Treatment of APL miceTwenty days after inoculation of APL cells into syngeneic mice

or NOD/SCIDmice, peptidomimetics (10mg/kg, twice) or ATRA(10 mg/kg, once) and ATO (10 mg/kg, once; Sigma-Aldrich)treatment was started by daily intraperitoneal injection for 6 to10 successive days.

Bioluminescence imaging in vivoNB4 cells or APL murine blasts carrying a luciferase reporter

were transplanted into mice. The luciferase substrate wasinjected into living animals before imaging. Then biolumines-cence imaging (BLI) was performed by according to the man-ufacturers' protocols (Xenogen IVIS Spectrum, PerkinElmer).

Co-IP assayCell extracts were prepared with lysis buffer (50 mmol/L

pH 7.5 Tris, 150 mmol/L NaCl, 0.5% Triton X-100, 10%glycerol, 2 mmol/L EDTA, 1 mmol/L PMSF, 20 mmol/L,Protease Inhibitor Cocktail, Phosphatase Inhibitor Cocktail,and 2 mmol/L DTT). Supernatants were then incubated withpreconjugated anti-FLAG M2 (Sigma), anti-MYC (Biotool),anti-HA (Biotool), anti-GFP (MBL), or anti-IgG (CST) beadsat 4�C overnight. The beads were sequentially washed 5 timeswith co-IP lysis buffer. The bound proteins were eluted with2% SDS lysis buffer and boiled at 100�C for 15 minutes. Theproteins were analyzed by Western blotting according to thestandard protocol.

In vivo ubiquitination assaysFor the in vivo ubiquitination assays, 293T cells were tran-

siently transfected with plasmids for HA-tagged ubiquitin,Flag-PML-RARa, GFP-TWIST1, Myc-TRIB3, and other indicatedconstructs. Twenty-four hours after transfection, cells weretreated with 10 mmol/L MG-132 or DMSO for 24 hours, and

then washed with PBS and collected by centrifugation. Cellswere lysed in co-IP lysis buffer. The lysate was subjected to co-IPor IP using anti-HA–conjugated agarose beads for overnight at4�C. The samples were loaded, separated by SDS-PAGE, andimmunoblotted with the indicated antibodies as describedabove.

Statistical analysisData are presented as arithmetic means � SEM. Kaplan–Meier

survival analysis, Student t tests, orx2 testswere used to calculate Pvalues where appropriate. P < 0.05 was considered to besignificant.

ResultsHigh TWIST1 expression in patients with APL

To evaluate the expression levels of EMT-TFs in AML, weanalyzed microarray data for different cytogenetic codes from199 AML samples from the E-MTAB-3444 database (Supple-mentary Fig. S1A). Detailed quantitative assessment showedthat TWIST1 expression, but not TWIST2, SNAI1/2, and ZEB1/2expression, was significantly higher in t (15;17) APL samplesthan in other cytogenetic types of AML, including t (8;21), inv(16), MLL-(11q23), t (9;11), and inv (3)/t (3;3; Fig. 1A; Sup-plementary Fig. S1B). We also examined EMT-TF mRNA expres-sion using the RNA-seq database of The Cancer Genome Atlas(TCGA) and verified the higher quantity of TWIST1 mRNA inthe M3 subtype of AML according to FAB classification (Fig. 1B;Supplementary Fig. S1C). To further confirm the aberrant highexpression of TWIST1 in APL, we quantified the mRNA andprotein levels of APL blasts from patients with primary AML(Fig. 1C and D; Supplementary Fig. S1D; Supplementary TableS1). Indeed, higher levels of the TWIST1 mRNA and proteinwere found in APL samples than in non-APL samples. Consis-tent with the data from human leukemic cells, higher TWIST1expression was found in leukemic cells from PML-RARa trans-genic mice than in blasts from other AML mouse models(Fig. 1E). In a series of AML cell lines, the typical APL cell lineNB4 presented higher TWIST1 expression than the other celllines (Supplementary Fig. S1E). To assess whether TWIST1expression was correlated with the expression of the PML-RARaoncoprotein, we performed ZnSO4-induced PML-RARa expres-sion in PR9 cells. As expected, ZnSO4-induced PML-RARaexpression upregulated the expression of TWIST1 in a time-dependent manner; although, TWIST1 and PML-RARawere notcompletely colocalized (Fig. 1F). Consistent with this finding,TWIST1 was also correlated with PML-RARa expression in NB4cells, with partial colocalization (Supplementary Fig. S1F).Thus, these results demonstrate that TWIST1, as an EMT-TF,is highly expressed in APL cells.

(Continued.) D, The semiquantitative analysis of Western blot data showing TWIST1 and PML-RARa expression in primary leukemic blasts obtained from newlydiagnosed APL M3 patients (n¼ 12) versus CD34þ BM cells from healthy donors (n¼ 4) and other AML subtypes, including M1 (n¼ 7), M2 (n¼ 12), M4 (n¼ 14),M5 (n¼ 12), and M6 (n¼ 5). Data represent the mean� SEM of three assays (�, P < 0.05; �� , P < 0.01). E, qRT-PCR analysis of TWIST1mRNA expression in blastsfrom different AMLmouse models after BM transplantation of murine hematopoietic stem/progenitor cells freshly transduced with the indicated oncogenicfusion genes (top). The data represent the mean� SEM of three assays (�� , P < 0.001). In these murine leukemic blasts, the protein levels of TWIST1 and PML-RARawere detected byWestern blotting. Three independentWestern blotting replicates were performed (bottom). F, PR9 cells were incubated with 200 mmol/L ZnSO4 for the indicated times, and the cell lysates were blotted with an anti-TWIST1 or anti–PML-RARa antibody (top). The data represent immunoblots ofthree independent assays. Immunofluorescence microscopic inspection of the expression of TWIST1 or PML-RARa in control and ZnSO4-induced PR9 cells forindicated time. Representative images were obtained in six random fields from three independent biological replicates. Scale bar, 2 mm.






Figure 2.

High TWIST1 expression promotes APL progression. A, Lentiviruses carrying sh-Negative Control (NC) or sh-TWIST1-1 were used to transduce NB4 cells. Westernblot showing the protein levels of TWIST1, PML-RARa, p21, p-p53, cleaved PARP, and cleaved caspase-3 in the transduced cells. Three independentWesternblotting replicates were performed. B, The flow cytometric scatter plots present differentiated cells (CD11bþ, top) and apoptotic cells (Annexin Vþ, bottom) inNB4 cells with or without TWIST1mRNA knockdown. Column diagram showing the percentage of CD11bþ cells and Annexin Vþ cells in transduced NB4 cells. Thevalues are presented as the mean� SEM (n¼ 6/group; �� , P < 0.001). (Continued on the following page.)

Lin et al.





High TWIST1 expression promotes APL progressionTo assess whether high TWIST1 expression plays an important

role in APL progression, we introduced negative control(NC)-shRNA (shRNA) or sh-TWIST1-1/2 into two types of APLcells (NB4 cells and APL murine blasts) and found that sh-TWIST1-1 caused significant suppression of TWIST1 expressionin both types of cells (Supplementary Fig. S2A). We observed thatsh-TWIST1-1 decreased PML-RARa expression and induced NB4cell apoptosis and differentiation but not senescence (Fig. 2A andB; Supplementary Fig. S2B). The efficacy and specificity of TWIST1shRNAs were confirmed by rescue via TWIST1 overexpression(Supplementary Fig. S2C). Consistent with the findings obtainedin the NB4 cell line, sh-TWIST1-1 expression induced apoptosisand differentiation in leukemic blasts of APL transgenic murine,whichwere generated on an FVB/NJ background using the humanMRP8 expression cassette (Supplementary Fig. S2D–S2F). Toevaluate the role of TWIST1 in APL progression in vivo, weinoculated TWIST1-knockdown NB4 cells into sublethally irra-diated NOD/SCID mice by tail vein injection. At 3 weeks aftertransplantation, TWIST1 knockdown significantly suppressedAPL cell invasion and prolonged the survival of recipient mice(Fig. 2C and D). Similarly, APL transgenic murine blasts expres-sing sh-TWIST1-1 generated fewer leukemic cells and had amuchlonger survival rate than the NC mice (Supplementary Fig. S2G–S2I). Interestingly, in contrast to the NC-shRNA groups, in whichthe mice developed a spinning in a circle syndrome or evenparalysis, we did not observe significant central nervous system(CNS) infiltration of APL cells in the TWIST1-knockdown groups,indicating that TWIST1 as an EMT-TF may contribute to APLextramedullary infiltration (Fig. 2C and E; Supplementary Fig.S2G and S2J). As reported previously, the serious situation asso-ciated with APL relapse and poor prognosis largely predominatedin the CNS infiltration (27–29), we showed that downregulationof TWIST1 in APL cells prevented CNS invasion inmice, althoughwe also observed skin and eye involvement in some NC-shRNAcases of our APL mouse models (Supplementary Fig. S2K). Wealso examined the effect of TWIST1 depletion on blasts frompatients with primay APL and confirmed the importance ofTWIST1 in maintaining APL cell survival (Fig. 2F and G; Supple-mentary Fig. S2L). Collectively, these data suggest that TWIST1plays an important role in APL cell survival and is critical for APLdisease progression.

TRIB3 interacts with TWIST1 in APL cellsAs mentioned previously, TWIST1 correlated with the expres-

sion of the PML-RARa oncoprotein but was not completelycolocalized with PML-RARa. We then investigated whether there

was evidence linking the molecular mechanisms of PML-RARa–driven APL to an indirect influence of TWIST1. Notably, theactivity of TWIST1 as an EMT-TF seemed to be maintainedthrough a TRIB3/p62–dependent interaction (30, 31). TheTribbles proteins (TRIB1, TRIB2, and TRIB3) have been shownto play critical roles in leukemogenesis via different mechanisms(32–38). TRIB3 has been reported to suppress the degradation ofPML-RARa through interacting with PML-RARa and PML (39).Given the importance of TWIST1 and TRIB3 in APL pathogenesis,we assessed whether TWIST1 and TRIB3 were highly coexpressedin APL cells compared with other subtypes of AML. We analyzedthe profiling data from the E-MTAB-3444 and TCGA databasesand verified high TRIB3 expression in previously studied APLsamples (Supplementary Fig. S3A and S3B). As expected, TWIST1expression was significantly correlated with the expression ofTRIB3 in APL patient samples (Fig. 3A). Consistent with thisfinding, TWIST1 and TRIB3 were also highly coexpressed in NB4cells or ZnSO4-induced PR9 cells (Fig. 3B; Supplementary Fig.S3C). As high expression of TWIST1 and TRIB3 in APL cells, wesuspected that TWIST1might interact with TRIB3 to regulate PML-RARa function. We found that TWIST1 coimmunoprecipitatedwith TRIB3 in NB4 cells and APL murine blasts (Fig. 3C; Supple-mentary Fig. S3D), and this finding was confirmed in HEK293Tcells by coexpressing various tagged plasmids or performingglutathione S-transferase (GST) pull-down (Fig. 3D and E). Inaddition, using a coimmunostaining assay, we observed thatTWIST1 was strongly colocalized with TRIB3 (Fig. 3F and G;Supplementary Fig. S3E). Taking into account the previouslyconfirmed interaction of TRIB3 and PML-RARa (39), we furtherinvestigated the relationship among these three proteins. Weshowed the colocalization of TWIST1, TRIB3, and PML-RARa inbothNB4 and plasmid-coexpressing HEK293T cells (Supplemen-tary Fig. S3F). According to Flag-PML-RARa immunoprecipita-tion of APL murine blasts and plasmid-coexpressing HEK293Tcells, TRIB3, but not TWIST1, bound directly to PML-RARa(Supplementary Fig. S3G and S3H). These results suggest thatTWIST1 interacts with TRIB3 to indirectly modulate PML-RARa.

TRIB3 protects TWIST1 fromubiquitination and stabilizes highTWIST1 expression

Human TWIST1 is an 18-kDa protein that contains 202 aminoacids, which is much smaller than the PML-RARa fusion onco-protein. To determine whether TWIST1 protein has a short half-life compared with PML-RARa oncoprotein, we performed thecycloheximide (CHX) chase assay with APL cells. Using thetranslational inhibitor CHX, we observed that APL cells accumu-lated endogenous TWIST1 and TRIB3 but not the corresponding

(Continued.) C,A representative bioluminescence image of mice transplanted with NB4-luc cells stably transduced with NC or sh-TWIST1-1. Quantitativeluciferase bioluminescence was monitored at week 3 postxenografting. Representative BLI images and quantitation data were from six independentexperiments; n¼ 6 for each group (��, P < 0.01). D, Kaplan–Meier analysis shows the survival rates of mice receiving 2� 106 NB4-luc cells stably expressing anontargeting NC or an shRNA targeting TWIST1 (sh-TWIST1-1; n¼ 6 for each model). E, Hematoxylin and eosin (H&E) staining of brain biopsies and spinal cordbiopsies collected frommice transplanted with NB4-luc cells stably transduced with NC or sh-TWIST1-1 at day 20 postxenografting (left). RepresentativehCD45þ IHC staining of the murine brain and spinal cord (right). Carmine arrows indicate the CNS-infiltrating NB4-luc cells in transplanted mice. Black trianglesand squares denote the cerebral parenchyma and spinal cavities, respectively. Dashed lines indicate the meninges. Images are representative of six independentexperiments. Scale bar, 100 mm. F, TWIST1, PML-RARa, cleaved PARP, cleaved caspase-3, and p21 expression in blasts from three patients with APL that weretransduced with a nontargeting siRNA (NC) or a siRNA targeting TWIST1 (si-TWIST1) for 48 hours in vitro. Three independentWestern blotting replicates wereperformed.G, Representative scatter plots of differentiated cells (CD11bþ, top) and apoptotic cells (Annexin Vþ, bottom) in blasts from three patients with APLwith or without TWIST1mRNA knockdown. The quantitative measurement presents the percentages of CD11bþ cells and Annexin Vþ cells in transduced APLpatients' blasts. Three independent assays were performed for each group. The values are presented as the mean� SEM (n¼ 6/group; ��, P < 0.01;�� , P < 0.001).






Lin et al.





mRNA after treatment with the proteasome inhibitor MG-132.The CHX chase assay indicated that TWIST1 was degraded morerapidly than PML-RARa (Fig. 3H; Supplementary Fig. S3I andS3J). In vivo ubiquitination assays also showed increased levels ofubiquitinated TWIST1 after MG132 treatment (SupplementaryFig. S3K and S3L). To check the possibility that TWIST1 is a directtranscriptional target of PML-RARa, we screened the identifiedputative PML-RARa oncoprotein–binding sites provided byWang and colleagues (20), and we did not find the bindingcharacteristic of TWIST1 as a direct transcription target ofPML-RARa. These data demonstrate that APL cells degrade endog-enous TWIST1 in a proteasome-dependentmanner. Interestingly,we observed the half-life of TWIST1 is longer than half-life ofTRIB3 (Fig. 3H). Considering that TWIST1 interacted with TRIB3in APL cells, we hypothesized that TRIB3 might inhibit ubiqui-tination and degradation of TWIST1 in a TRIB3-dependent man-ner. TRIB3 knockdown decreased the protein level of endogenousTWIST1 and shortened the protein half-life in NB4 cells, whichwas reversed by MG-132 treatment (Fig. 3I; Supplementary Fig.S3M and S3N). In vivo ubiquitination assays also showedincreased levels of ubiquitinated TWIST1 after TRIB3 knock-down (Fig. 3J). Together, these data indicate that TRIB3 stabi-lizes high TWIST1 expression in APL cells through preventingubiquitination.

TRIB3 inhibition promotes TWIST1 degradation and reversesresistance to ATRA therapy

Although APL has been highly curable with ATRA-based dif-ferentiation therapy, a fractionof patients still relapse andbecomeresistant to ATRA (40, 41). We found that ATRA treatmentdecreased the protein levels of TWIST1 and TRIB3 in ATRA-sensitive NB4 cells but not in ATRA-resistant NB4-R1 (R1) cells(Fig. 4A). This finding led us to consider whether TRIB3 coop-erated with TWIST1 to contribute to ATRA resistance. To test thishypothesis, we introduced NC-shRNA or sh-TRIB3-1/2 into R1cells and observed significant suppression of TRIB3 expressionwith sh-TRIB3-2 (Supplementary Fig. S4A). Furthermore, sh-TRIB3-2 slightly decreased TWIST1 protein levels and promotedR1 cell differentiation but not cell apoptosis (Supplementary Fig.S4B–S4D). Silencing of TRIB3 also reduced TWIST1 expression in

response to ATRA and greatly reversed ATRA resistance in R1 cells(Fig. 4B and C; Supplementary Fig. S4E). In vivo ubiquitinationassays revealed increased levels of ubiquitinated TWIST1 afterTRIB3 knockdown in response to ATRA treatment (Fig. 4D).Moreover, TWIST1 overexpression rescued TWIST1 protein levelsand impaired the differentiation induced by TRIB3 knockdown(Fig. 4E and F; Supplementary Fig. S4F–S4J). To evaluate the roleof TRIB3/TWIST1 in APL ATRA resistance in vivo, we inoculatedTRIB3-knockdown and/or TWIST1-overexpressing R1 cells intosublethally irradiated NOD/SCID mice by tail vein injection. At25 days after transplantation with 6-day ATRA treatment, TRIB3knockdown significantly reversed ATRA resistance in R1 cells andprolonged the survival of recipient mice (Fig. 4G and H). Sim-ilarly, TWIST1 overexpression rescued the suppression of R1 cellsand shortened the survival of recipient mice caused by TRIB3knockdown in vivo (Fig. 4G and H). Interestingly, compared withthe NC-shRNA groups, reduced CNS infiltration of R1 cells wasobserved in the TRIB3-knockdown groups, and modest CNSinfiltration of R1 cells was observed in the TWIST1-overexpressinggroup, indicating that TWIST1 may contribute to APL extrame-dullary infiltration (Fig. 4I). These results show that loss of TRIB3promotes TWIST1 degradation and reverses resistance to ATRAtherapy.

The WR domain of TWIST1 is required for TWIST1/TRIB3binding and TWIST1 stabilization in APL

Because the TRIB3/TWIST1 interaction is involved in theprotein stability and functional maintenance of the TWIST1protein, we further analyzed the binding interface residues indetail. We constructed mutants of TWIST1 and TRIB3 deletedfor various domains (Supplementary Fig. S5A and S5B). Wefound that both the bHLH domain and the WR domain ofTWIST1 interacted with the C terminus of TRIB3, although thebHLH domain only weakly contributes to this interaction(Fig. 5A). We have thus unveiled the critical binding domainsof TWIST1 with TRIB3. As observed before, the binding ofTRIB3 and TWIST1 stabilized high TWIST1 expression in APLcells (Fig. 3J; Supplementary Fig. S3K and S3L). Consideringthat TRIB3 interacts mostly with the WR domain of TWIST1, wehypothesized that the TRIB3-C terminus: TWIST1-WR domain

Figure 3.TRIB3 interacts with TWIST1 in APL cells. A, A correlation analysis of the relative TWIST1 and TRIB3 mRNA expression in TCGA AML patients witheither APL (n ¼ 16) or non-APL disease (n ¼ 157). B, Total lysates of the indicated AML cell lines were extracted, and TWIST1 and TRIB3 proteinlevels were detected by Western blotting. TWIST1 and TRIB3 protein level were highly correlated in AML cell lines. The data are representativeimmunoblots of three independent assays. C, NB4 cell extracts were subjected to IP with immunoglobulin G (IgG) and anti-TWIST1 or anti-TRIB3 Absand blotted with the indicated Abs. The interaction of TWIST1 and TRIB3 was evaluated by Co-IP in NB4 cells. The data are representativeimmunoblots of three independent assays. D, HEK293T cells were cotransfected with TWIST1-Myc and TRIB3-Flag expression plasmids. Cell extractswere subjected to IP with anti-Myc or anti-Flag Abs and blotted with the indicated Abs. Ectopically expressed TWIST1 and TRIB3 interact in HEK293Tcells. The data are representative immunoblots of three independent assays. E, Retrieved proteins were evaluated by Western blotting. The GST-onlyprotein was used as the negative control. In vitro interaction of TWIST1 and TRIB3 was detected with a GST pull-down assay. The data arerepresentative immunoblots of three independent assays. F and G, Colocalization of TWIST1 and TRIB3 was detected in NB4 and HEK293T cells withimmunostaining. For F, anti-TWIST1 and anti-TRIB3 Abs were used for immunostaining. For G, TWIST1-Myc and TRIB3-Flag were ectopically expressedin HEK293T cells, and anti-Myc or anti-Flag Abs were used. Images are representative of at least six random fields. Scale bar, 2 mm. H, NB4 cellswere incubated with CHX (10 mg/mL) or CHX plus MG132 (10 mmol/L) for the indicated times. TWIST1, TRIB3, and PML-RARa protein level weredetected by immunoblotting (left). The semiquantitative analysis of TWIST1 and TRIB3 protein expression in NB4 cells subjected to the indicatedtreatment were shown (right). The data represent the mean � SEM of three assays. I, The statistics and quantification of relative TWIST1 expressionlevel were performed by densitometry of protein expression levels presented relative to GAPDH in the same lane, and were compared with theWestern blotting assay of 0–min control groups (see also Supplementary Fig. S3N). The data represent the mean � SEM of three assays. J, Theeffect of TRIB3 depletion on TWIST1 ubiquitination in vivo. The cell extracts of Control (NC) or TRIB3-silenced (sh-TRIB3-1) NB4 treated with MG132(10 mmol/L) in vitro for 12 hours were subjected to IP with immunoglobulin G (IgG) and anti-TWIST1 Abs and blotted with an ubiquitin Ab. The datarepresent immunoblots of three independent assays.






Figure 4.

TRIB3 inhibition promotes TWIST1 degradation and reverses resistance to ATRA therapy. A, TWIST1, TRIB3, and PML-RARa expression levels in NB4 and R1 cellstreated with ATRA (1 mmol/L) and harvested at the indicated times. Three independent Western blotting replicates were performed. B, TWIST1, TRIB3, and PML-RARa protein expression in transduced R1 cells after treatment with ATRA (1 mmol/L) for the indicated times. R1 cells stably transduced with a nontargetingshRNA (NC) or an shRNA targeting TRIB3 (sh-TRIB3-2). The measurements of protein levels were obtained from three independent Western blotting replicates.C, The flow cytometric scatter plots present differentiated cells (CD11bþ) in R1 cells with or without TRIB3 mRNA knockdown after treatment with ATRA (1 mmol/L)for the indicated times (top). Column diagram showing the percentage of CD11bþ cells in transduced R1 cells (bottom). (Continued on the following page.)

Lin et al.





interaction might stabilize high TWIST1 expression. Asexpected, HEK293T IP assays showed that TRIB3 was able toinhibit the ubiquitination of TWIST1, and this inhibitionmainly occurred between the C-terminus of TRIB3 and the WRdomain of TWIST1 (Fig. 5B; Supplementary Fig. S5C and S5D).

To directly examine the functional effect of interface residuescritical for binding between TRIB3 and TWIST1,we generated foursynthetic peptides containing the interface residues and variouslysine sites from the TWIST1 C terminus (Fig. 5C, top). We fusedthese peptides to a cationic cell–penetrating peptide (42). Thepeptideswere designed in theorientation containingK residues orthe WR domain and were termed peptides 1, 2, and 3. We alsodesigned an inactive peptide (termed peptide con) as a negativecontrol. To determine the ability of peptidomimetics to dissociatethe TRIB3/TWIST1 complex in APL cells, we purified theTRIB3/TWIST1 complex by IP using anti-TWIST1 antibodies inthe presence of 1 mmol/L peptidomimetics and determined itscomposition by Western blotting analysis. Compared with pep-tide con treatment, competition with peptide 3 led to significantdissociation of the cellular TRIB3/TWIST1 complex (Fig. 5C,bottom).We then treatedNB4 and R1 cells with peptidomimeticsandobserved that peptide 3, but not the peptide con, peptide 1, orpeptide 2, induced significant differentiation and apoptosis inAPL cells rather than other leukemic cells or normal hematopoi-etic cells (Fig. 5D, Supplementary Fig. S5E and S5F). Consistentwith the significant changes in cell surface CD11b and Annexin V,we observed a significant decrease in the protein abundance ofTRIB3/TWIST1 and a robust increase in intracellular caspase-3and PARP cleavage after peptide 3 treatment for only 8 hours(Supplementary Fig. S5G). Moreover, in vivo ubiquitinationassays showed increased levels of ubiquitinated TWIST1 withpeptide 3 treatment (Supplementary Fig. S5H). Thus, peptide 3is a specific peptide mimetic inhibitor of TRIB3/TWIST1 complexassembly inAPL cells and induces apoptosis anddifferentiationofAPL cells in vitro.

WR domain peptidomimetic inhibition of the TRIB3/TWIST1interaction impairs rapid progression during the early death ofAPL

Despite the striking long-term leukemia-free survival rateafter the ATRA/ATO–based regimen, the progression of APL,including early death (ED) and differentiation therapy resis-tance, still affects the health of a significant proportion ofpatients with APL (3, 7). To mimic the progressive clinicalpattern of APL ED, we began to treat APL mouse models withATRA/ATO or peptidomimetics at 20 days posttransplantation

(Supplementary Fig. S6A). We noted that APL mice had veryhigh peripheral blasts at this stage and died within 5 days if nottreated, which was very similar to the clinical pattern of APL ED(Supplementary Fig. S6B). We transplanted NB4 cells intoNOD/SCID mice, and at 20 days posttransplantation, wetreated the engrafted mice with ATRA/ATO, peptide 3, orcontrol peptide for 30 days. Peptide 3 significantly delayedleukemia progression, extended survival, and reduced CNSinfiltration of APL cells (Fig. 6A–C; Supplementary Fig.S6C). Consistent with the results obtained in APL cell–engrafted NOD/SCID mice, APL transgenic mice treated withpeptide 3 exhibited significantly delayed disease latency anddeath compared with the mice in the control or ATRA/ATOgroup (Fig. 6D and E). We also noted that peptide 3, the WRdomain peptide mimetic, inhibited APL cell proliferation andimpeded CNS infiltration in vivo (Supplementary Fig. S6D andS6E). We then examined whether peptide 3 interfered withTWIST1/TRIB3 interaction to reverse resistance to ATRA treat-ment. As expected, the use of peptide 3 in vitro greatly impairedATRA resistance in R1 cells (Supplementary Fig. S6F). Toevaluate the role of peptidomimetics in APL ATRA resistancein vivo, we inoculated R1 cells into sublethally irradiated NOD/SCID mice by tail vein injection. At 20 days after transplanta-tion, the mice were treated with ATRA and peptidomimetics,and then monitored for response to ATRA treatment. Consis-tent with the use of peptide 3 in vitro, the WR domain pepti-domimetic significantly impeded leukemia growth and sensi-tivity to ATRA and prolonged the survival of R1 cell recipientmice (Fig. 6F and G). We further confirmed the therapeuticeffect of peptide 3 in preventing rapid progression of primaryAPL ED patient blasts in vitro. We screened 5 samples of EDfrom the collected APL bone marrow specimens and found thatthey all involved cases of high WBC and peripheral blast counts(Supplementary Fig. S6G). We treated leukemia cells withpeptide 3 in vitro and confirmed that it can promote leukemiacell differentiation and apoptosis in a short time comparedwith ATRA/ATO or peptide con (Fig. 6H; Supplementary Fig.S6H). Therefore, the WR domain peptidomimetic can rapidlypromote APL cell differentiation and apoptosis, and can pre-vent early death of APL and reverse induction therapy resistance(Fig. 6I).

DiscussionAlthough the importance of EMT-TFs, including the TWIST,

SNAIL, and ZEB families, has been well documented in the

(Continued.) The values are presented as the mean � SEM (n ¼ 6/group; �� , P < 0.001). D, Control (NC) or TRIB3-silenced (sh-TRIB3-2) R1 cells after treatmentwith ATRA (1 mmol/L) for 24 hours were subjected to IP with immunoglobulin G (IgG) and anti-TWIST1 Abs and blotted with an ubiquitin Ab. The data representimmunoblots of three independent assays. E, R1 cells were stably transduced with NC or an shRNA targeting TRIB3 mRNA (sh-TRIB3-2), followed by transductionwith a lentiviral vector carrying a control or TWIST1 construct. In these cells, after treatment with ATRA (1 mmol/L) for the indicated times, the expression levels ofTWIST1 and TRIB3 were detected by Western blotting. Three independent Western blotting replicates were performed. F, Column diagram showing thepercentage of CD11bþ cells in transduced R1 cells. The values are presented as the mean � SEM (n ¼ 6/group; �� , P < 0.01; �� , P < 0.001). G, A representativebioluminescence image of mice transplanted with R1-luc cells stably transduced with NC, sh-TRIB3-2, or sh-TRIB3-2-TWIST1 (sh-TRIB3-2, followed by transductionwith a lentiviral vector carrying a control or TWIST1 construct). Quantitative luciferase bioluminescence was monitored at day 25 postxenografting after treatmentwith ATRA (10 mg/kg). Representative BLI images and quantitation data were from six independent experiments; n ¼ 6 for each group (�� , P < 0.01; �� , P <0.001). H, Kaplan–Meier analysis shows the survival rates of mice receiving 2 � 106 R1-luc cells stably expressing a nontargeting NC, an shRNA targeting TRIB3(sh-TRIB3-2) or sh-TRIB3-2-TWIST1 after 6-day treatment with ATRA (10 mg/kg, once daily; n ¼ 6 for each model). I, H&E staining of brain biopsies collected frommice transplanted with R1-luc cells stably transduced with NC, sh-TRIB3, or sh-TRIB3-2-TWIST1 at day 25 postxenografting after treatment with ATRA (10 mg/kg;left). Representative hCD45þ IHC staining of the murine brain (right). Carmine arrows indicate the CNS-infiltrating R1-luc cells in transplanted mice. Images arerepresentative of at least six random fields. Scale bar, 100 mm.






Figure 5.

TheWR domain of TWIST1 is required for the binding of TWIST1/TRIB3 complex and TWIST1 stabilization.A,Mapping of TWIST1 regions involved in C-terminalbinding to TRIB3. Top, Diagram of TWIST1 deletion mutants. Bottom, HEK293T cells were cotransfected with the indicated TWIST1 and TRIB3-C-Myc constructs.The cell extracts were subjected to IP with an anti-Myc Ab. The data are representative immunoblots of three independent assays. B, HEK293T cells werecotransfectedwith the indicated TWIST1-DbHLH, TRIB3-C, andHA-Ub constructs. Cell extracts were subjected to IP with an anti-GFP Ab and blotted with an HAAb. The data represent immunoblots of three independent assays. C, Schematic illustration showing that different peptides target relevant amino acidsequences, which include the key lysine sites and theWRdomain of the TWIST1 protein. Mapping aa106-aa202 sequences of human TWIST1 protein regions involvedin C-terminal peptide binding (top). Three peptides containing common sequences of penetrating peptides were designed to competitively inhibit the binding ofTWIST1 and TRIB3 (middle). The cell extracts of NB4 or R1 treated with different peptides (1 mmol/L) for 12 hours were subjected to IP with immunoglobulin G (IgG)and anti-TWIST1 Abs and blotted with the indicated Abs (bottom left). Quantitative and statistical analysis of TRIB3/TWIST1 gray values from IP (bottom right). Thedata are representative immunoblots of three independent assays.D, Representative scatter plots of differentiated cells (CD11bþ, top) and apoptotic cells (AnnexinVþ, bottom) in NB4 or R1 cells treated with different peptides (1 mmol/L) for 12 hours (top). The quantitative measurements present the percentages of CD11bþ cellsand Annexin Vþ cells in NB4 or R1 cells treatedwith different peptides (bottom). Three independent assayswere performed for each group. The values are presentedas the mean � SEM (n ¼ 6/group; �� , P < 0.001).

Lin et al.





Figure 6.

WR domain peptidomimetic inhibition of the TRIB3/TWIST1 interaction impairs rapid progression during the early death of APL. A, A representativebioluminescence image of mice transplanted with NB4-luc cells after peptide con (10 mg/kg twice daily), ATRA (10 mg/kg once daily)/ATO (4 mg/kg oncedaily), or peptide 3 (10 mg/kg twice daily) treatment in vivo. Representative BLI imageswere from six independent experiments. B,Quantitative luciferasebioluminescence was monitored at day 25 postxenografting related to (A; n¼ 6 for each group, �� , P < 0.001). C, Kaplan–Meier analysis shows the survival ratesof mice receiving 2� 106 NB4-luc cells after peptide 3 or ATO/ATRA treatment in vivo (n¼ 6 for each model). The black arrow indicates that administrationbegins at day 20. D, A representative bioluminescence image of mice transplanted with APLmurine-luc cells after peptide con (10 mg/kg twice daily), ATRA(10 mg/kg once daily)/ATO (4mg/kg once daily), or peptide 3 (10 mg/kg twice daily) treatment in vivo. Quantitative luciferase bioluminescence was monitoredat day 23 postxenografting. The values are presented as the mean� SEM (n¼ 6/group; �� , P < 0.001). E, Kaplan–Meier analysis shows the survival rates of micereceiving 1� 106 APLmurine-luc cells after peptide con, peptide 3, or ATO/ATRA treatment in vivo (n¼ 6 for each model). The black arrow indicates thatadministration begins at day 20. F, A representative bioluminescence image of mice transplanted with APL-luc cells after peptide control (10 mg/kg twice daily)/ATRA (10 mg/kg once daily) or peptide 3 (10 mg/kg twice daily)/ATRA treatment (10 mg/kg once daily) in vivo. Quantitative luciferase bioluminescence wasmonitored at day 25 postxenografting. Representative BLI images and quantitation data were from six independent experiments; n¼ 6 for each group. Scale bar,1 cm. ��, P < 0.001. G, Kaplan–Meier analysis shows the survival rates of mice receiving APL-luc cells after peptide control/ATRA or peptide 3/ATRA treatment invivo (n¼ 6 for each model). The black arrow indicates that administration begins at day 20. H, Representative scatter plots of differentiated cells (CD11bþ) andapoptotic cells (Annexin Vþ) in blasts from five APL early death patients with peptide con (2 mmol/L), peptide 3 (2 mmol/L), or ATRA (1 mmol/L)/ATO (1 mmol/L)treatment in vitro for 8 hours. The quantitative measurements present the percentages of CD11bþ cells and Annexin Vþ cells in cells treated with peptide con,peptide 3, or ATRA/ATO. Four independent assays were performed for each group. The values are presented as the mean� SEM. I, An illustration of TRIB3stabilizing high TWIST1 expression to promote APL rapid progression and ATRA resistance.






tumorigenesis of epithelial cancers, the role of EMT-TFs inhematologic malignancies is still unknown. In this study, wefound that EMT-TF TWIST1 is highly expressed in patients withAPL and is critical for leukemic cell survival. In light of a recentreport indicating that the stress protein TRIB3 inhibits thedegradation of PML-RARa and promotes APL progression, wefound that TWIST1 and TRIB3 were highly coexpressed in APLcells compared with other subtypes of AML. We subsequentlydemonstrated that TRIB3 could strongly bind to the WRdomains of TWIST1 to stabilize TWIST1 by inhibiting itsubiquitination.

Currently, APL is highly curative. With differentiation ther-apy, over 80% of patients with APL achieve long-term leuke-mia-free survival. However, a proportion of patients with APLsuffer from early fatal bleeding and leukemic extramedullaryinfiltration, and the underlying mechanisms of this differenceare largely unknown (27–29). On the basis of a detailedanalysis and a functional screening of synthetic peptides, wediscovered a peptide analogous to the TWIST1 WR domainthat rapidly and specifically represses APL cell survival bydisrupting the TRIB3/TWIST1 interaction. Targeting rapidTWIST1 degradation could protect against early death in APLand improved sensitivity to ATRA. Furthermore, our data alsoshowed that TWIST1 governed CNS infiltration during pro-gression in NB4-xenograft mice and APL transplantable mice.Cell-penetrating peptides may competitively inhibit theTWIST1/TRIB3 interaction and repress CNS progression byinitiating APL cell differentiation and apoptosis. Our previousstudy also reported that high expression of TWIST1 in AMLcontributes to extramedullary infiltration and promotes leuke-mic aggressiveness (16). A pioneering study using the MLL-AF9mouse model revealed that the EMT-inducer ZEB1 contributedto leukemic blast invasion and was associated with poorsurvival in patients with AML. These observations suggest thatEMT-TFs not only play important roles in solid tumors but alsopromote leukemic progression and aggressive extramedullaryinfiltration. Thus, EMT-TFs may be novel therapeutic targets fordisease progression in patients with relapsed and refractoryAML.

TWIST1 contains two highly conserved and functionallydifferent domains: the bHLH domain for DNA binding andthe WR domain for heterodimer formation (43). Recently, theWR domain was also reported to exhibit transactivation activityand to interact with RUNX2, SOX9, p53, RELA, andp62 (31, 44–47). Notably, through a p62-dependent interac-tion, the WR domain is necessary for the proteolytic activity ofTWIST1 (31). TRIB3 inhibited autophagic substrate clearance byinteracting with p62 and led to UPS-dependent accumulation ofEMT-TFs, such as TWIST1, ultimately promoting tumor growthand metastasis (30). These results strongly implicated that TRIB3was highly implicated in TWIST1 degradation. Here, we providedclear evidence that the binding of TRIB3 to the WR domaininhibited TWIST1 ubiquitination and stabilized high TWIST1expression to promote APL progression and ATRA resistance.Intriguingly, despite the absence of lysine sites in theWR domain,the TWIST1 WR domain was required for TWIST1 ubiquitylation,which appears to be caused by recruitment of potential ubiquitinE3 ligases.

APL is commonly driven by the t (15;17) chromosomaltranslocation, which yields the PML–RARa fusion oncoprotein

as a transcriptional repressor. ATRA- and/or ATO-triggeredPML-RARa proteolysis are required for the elimination of APLcells. Recent studies have shown that PML-RARa has a half-lifeof over 8 hours and requires approximately 12 hours to bepartially degraded in response toATRAand/or ATO(39, 48, 49).If this is the case, a significant number of patients with APLwho experience rapid progression and a high-risk state, includ-ing high white blood cell (WBC) counts, fatal bleeding, andsevere infection, would not rapidly repress APL cell survival viaATRA- and/or ATO-initiated PML-RARa degradation. Similar-ly, the proposal that ATRA- and/or ATO-triggered PML-RARadegradation in vitro and in vivo exerts any effect has beencontroversial. Therefore, we hypothesized that PML-RARainduces a gain-of-function transcriptional activation to upre-gulate modulators of APL pathogenesis. We provided clearevidence that high TWIST1 expression promotes APL progres-sion and that TWIST1 proteolysis initiated by a novel peptidemight help reshape the therapeutic design for rapid APLeradication. Although a previous study indicated that TRIB3suppresses the degradation of PML-RARa by interacting withPML-RARa and PML, it is difficult to determine how TRIB3instability might protect the half-life of PML-RARa. Our resultsstrongly suggest that the binding of TRIB3 to TWIST1 inhibitsTWIST1 ubiquitination and stabilizes high TWIST1 expressionto promote APL progression and ATRA resistance. This char-acteristic is a very effective target for preventing APL early deathand ATRA resistance.

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

Authors' ContributionsConception and design: W. Zhang, J. XuDevelopment of methodology: J. Lin, W. ZhangAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): J. Lin, W. Zhang, L.-T. NiuAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): J. Lin, W. Zhang, L.-T. NiuWriting, review, and/or revision of the manuscript: J. Lin, W. Zhang, J. Zhu,J. XuAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): W. Zhang, Y.-M. Zhu, X.-Q. Weng, Y. Sheng,J. ZhuStudy supervision: J. Xu

AcknowledgmentsWe thank Dr. Chun-Lin Shen for her assistance with NOD/SCID in vivo

bioluminescence imaging. This study was founded by the National NaturalScience Foundation of China (81800099, 81400106, 81430002, 81770206),the Shanghai Rising-Star Program (17QA1402200, 19QA1407800), the Shang-hai Excellent Youth Medical Talents Training Program (2018YQ09), andNational Science and TechnologyMajor Project (2018ZX09101001). W. Zhanggratefully acknowledges to be supported by Global Scholar-in-Training Award(GSITA) from AACR.

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received February 11, 2019; revised May 13, 2019; accepted June 20, 2019;published first June 24, 2019.


Lin et al.




References1. de The H, Pandolfi PP, Chen Z. Acute promyelocytic leukemia:

a paradigm for oncoprotein-targeted cure. Cancer Cell 2017;32:552–60.

2. Wang ZY, Chen Z. Acute promyelocytic leukemia: from highly fatal tohighly curable. Blood 2008;111:2505–15.

3. Lehmann S, Deneberg S, Antunovic P, Rangert-Derolf A, Garelius H,Lazarevic V, et al. Early death rates remain high in high-risk APL: updatefrom the Swedish Acute Leukemia Registry 1997–2013. Leukemia 2017;31:1457–9.

4. Lehmann S, Ravn A, Carlsson L, Antunovic P, Deneberg S,Mollgard L, et al.Continuing high early death rate in acute promyelocytic leukemia: apopulation-based report from the Swedish Adult Acute Leukemia Registry.Leukemia 2011;25:1128–34.

5. Park JH, Qiao B, Panageas KS, Schymura MJ, Jurcic JG, Rosenblat TL, et al.Early death rate in acute promyelocytic leukemia remains high despite all-trans retinoic acid. Blood 2011;118:1248–54.

6. Rahme R, Thomas X, Recher C, Vey N, Delaunay J, Deconinck E, et al.Early death in acute promyelocytic leukemia (APL) in French centers: amulticenter study in 399 patients. Leukemia 2014;28:2422–4.

7. Mantha S, Goldman DA, Devlin SM, Lee JW, Zannino D, Collins M,et al. Determinants of fatal bleeding during induction therapy foracute promyelocytic leukemia in the ATRA era. Blood 2017;129:1763–7.

8. Cheung N, So CW. Transcriptional and epigenetic networks in haemato-logical malignancy. FEBS Lett 2011;585:2100–11.

9. Dohner H,Weisdorf DJ, Bloomfield CD. Acute myeloid leukemia. N Engl JMed 2015;373:1136–52.

10. Puisieux A, Brabletz T, Caramel J. Oncogenic roles of EMT-inducingtranscription factors. Nat Cell Biol 2014;16:488–94.

11. Nieto MA, Huang RY, Jackson RA, Thiery JP. Emt: 2016. Cell 2016;166:21–45.

12. Stavropoulou V, Kaspar S, Brault L, Sanders MA, Juge S, Morettini S, et al.MLL-AF9 expression in hematopoietic stem cells drives a highly invasiveAML expressing EMT-related genes linked to poor outcome. Cancer Cell2016;30:43–58.

13. Peinado H, Olmeda D, Cano A. Snail, Zeb and bHLH factors in tumourprogression: an alliance against the epithelial phenotype? Nat Rev Cancer2007;7:415–28.

14. Yang J, Mani SA, Donaher JL, Ramaswamy S, Itzykson RA, Come C, et al.Twist, amaster regulator ofmorphogenesis, plays an essential role in tumormetastasis. Cell 2004;117:927–39.

15. Wang N, Guo D, Zhao YY, Dong CY, Liu XY, Yang BX, et al. TWIST-1promotes cell growth, drug resistance and progenitor clonogenic capacitiesinmyeloid leukemia and is a novel poor prognostic factor in acutemyeloidleukemia. Oncotarget 2015;6:20977–92.

16. Xu J, ZhangW, Yan XJ, Lin XQ, Li W,Mi JQ, et al. DNMT3Amutation leadsto leukemic extramedullary infiltration mediated by TWIST1. J HematolOncol 2016;9:106.

17. Chen CC, You JY, Gau JP, Huang CE, Chen YY, Tsai YH, et al. Favorableclinical outcome and unique characteristics in association with Twist1overexpression inde novo acutemyeloid leukemia. BloodCancer J 2015;5:e339.

18. Percio S, Coltella N, Grisanti S, Bernardi R, Pattini L. A HIF-1 networkreveals characteristics of epithelial-mesenchymal transition in acute pro-myelocytic leukemia. Genome Med 2014;6:84.

19. Ge DZ, Sheng Y, Cai X. Combined staurosporine and retinoic acid inducesdifferentiation in retinoic acid resistant acute promyelocytic leukemia celllines. Sci Rep 2014;4:4821.

20. Wang K, Wang P, Shi J, Zhu X, He M, Jia X, et al. PML/RARalpha targetspromoter regions containing PU.1 consensus and RARE half sites in acutepromyelocytic leukemia. Cancer Cell 2010;17:186–97.

21. Zhang W, Zhao C, Zhao J, Zhu Y, Weng X, Chen Q, et al. Inactivation ofPBX3 and HOXA9 by down-regulating H3K79 methylationrepresses NPM1-mutated leukemic cell survival. Theranostics 2018;8:4359–71.

22. Nason-Burchenal K, Maerz W, Albanell J, Allopenna J, Martin P, MooreMA, et al. Common defects of different retinoic acid resistantpromyelocytic leukemia cells are persistent telomerase activity andnuclear body disorganization. Differentiation 1997;61:321–31.

23. QuignonF, ChenZ, de TheH.Retinoic acid and arsenic: towards oncogene-targeted treatments of acute promyelocytic leukaemia. Biochim BiophysActa 1997;1333:M53–61.

24. Liu X, Chen J, Yu S, Yan L, Guo H, Dai J, et al. All-trans retinoic acid andarsenic trioxide fail to derepress the monocytic differentiation driverIrf8 in acute promyelocytic leukemia cells. Cell Death Dis 2017;8:e2782.

25. Du K, Herzig S, Kulkarni RN, Montminy M. TRB3: a tribbles homologthat inhibits Akt/PKB activation by insulin in liver. Science 2003;300:1574–7.

26. Zhang W, Zhao JM, Lin J, Hu CZ, Zhang WB, Yang WL, et al. Adaptivefibrogenic reprogramming of osteosarcoma stem cells promotesmetastaticgrowth. Cell Rep 2018;24:1266–77.

27. de Botton S, Sanz MA, Chevret S, Dombret H, Martin G, Thomas X, et al.Extramedullary relapse in acute promyelocytic leukemia treated with all-trans retinoic acid and chemotherapy. Leukemia 2006;20:35–41.

28. Housman E, Chang P, Lane SW, Blinder R, Galinsky I, Kesari S, et al.CNS relapse in acute promyeloctyic leukemia. J Clin Oncol 2010;28:e409–11.

29. Ganzel C, Douer D. Extramedullary disease in APL: a real phenomenon tocontend with or not? Best Pract Res Clin Haematol 2014;27:63–8.

30. Hua F, Li K, Yu JJ, Lv XX, Yan J, Zhang XW, et al. TRB3 links insulin/IGF totumour promotion by interacting with p62 and impeding autophagic/proteasomal degradations. Nat Commun 2015;6:7951.

31. Qiang L, ZhaoB,MingM,WangN,He TC,Hwang S, et al. Regulation of cellproliferation and migration by p62 through stabilization of Twist1.Proc Natl Acad Sci U S A 2014;111:9241–6.

32. Keeshan K, He Y, Wouters BJ, Shestova O, Xu L, Sai H, et al. Tribbleshomolog 2 inactivates C/EBPalpha and causes acute myelogenous leuke-mia. Cancer Cell 2006;10:401–11.

33. Qi L, Heredia JE, Altarejos JY, Screaton R, Goebel N, Niessen S, et al. TRB3links the E3 ubiquitin ligase COP1 to lipidmetabolism. Science 2006;312:1763–6.

34. Dedhia PH, Keeshan K, Uljon S, Xu L, Vega ME, Shestova O, et al.Differential ability of Tribbles family members to promote degradationof C/EBPalpha and induce acutemyelogenous leukemia. Blood 2010;116:1321–8.

35. Keeshan K, Bailis W, Dedhia PH, Vega ME, Shestova O, Xu L, et al.Transformation by Tribbles homolog 2 (Trib2) requires both the Trib2kinase domain and COP1 binding. Blood 2010;116:4948–57.

36. Rishi L, Hannon M, Salome M, Hasemann M, Frank AK, Campos J, et al.Regulation of Trib2 by an E2F1-C/EBPalpha feedback loop in AML cellproliferation. Blood 2014;123:2389–400.

37. Stein SJ, Mack EA, Rome KS, Pear WS. Tribbles in normal and malignanthaematopoiesis. Biochem Soc Trans 2015;43:1112–5.

38. Stein SJ, Mack EA, Rome KS, Pajcini KV, Ohtani T, Xu L, et al. Trib2suppresses tumor initiation in Notch-driven T-ALL. PLoS One 2016;11:e0155408.

39. Li K, Wang F, Cao WB, Lv XX, Hua F, Cui B, et al. TRIB3 promotes APLprogression through stabilization of the oncoprotein PML-RARalphaand inhibition of p53-mediated senescence. Cancer Cell 2017;31:697–710e7.

40. Gallagher RE. Retinoic acid resistance in acute promyelocytic leukemia.Leukemia 2002;16:1940–58.

41. Sanz MA, Lo-Coco F. Modern approaches to treating acute promyelocyticleukemia. J Clin Oncol 2011;29:495–503.

42. Fawell S, Seery J, Daikh Y, Moore C, Chen LL, Pepinsky B, et al. Tat-mediated delivery of heterologous proteins into cells. Proc Natl Acad SciU S A 1994;91:664–8.

43. Chang AT, Liu Y, Ayyanathan K, Benner C, Jiang Y, Prokop JW, et al. Anevolutionarily conserved DNA architecture determines target specificityof the TWIST family bHLH transcription factors. Genes Dev 2015;29:603–16.

44. Bialek P, Kern B, Yang X, Schrock M, Sosic D, Hong N, et al. A twist codedetermines the onset of osteoblast differentiation. Dev Cell 2004;6:423–35.

45. Gu S, Boyer TG, Naski MC. Basic helix-loop-helix transcription factorTwist1 inhibits transactivator function of master chondrogenic regulatorSox9. J Biol Chem 2012;287:21082–92.






46. Piccinin S, Tonin E, Sessa S, Demontis S, Rossi S, Pecciarini L, et al. A "twistbox" code of p53 inactivation: twist box: p53 interaction promotes p53degradation. Cancer Cell 2012;22:404–15.

47. Li S, Kendall SE, Raices R, Finlay J, Covarrubias M, Liu Z, et al. TWIST1associates with NF-kappaB subunit RELA via carboxyl-terminal WRdomain to promote cell autonomous invasion through IL8 production.BMC Biol 2012;10:73.

48. Zhang XW, YanXJ, ZhouZR, Yang FF,WuZY, SunHB, et al. Arsenic trioxidecontrols the fate of the PML-RARalpha oncoprotein by directly bindingPML. Science 2010;328:240–3.

49. Jeanne M, Lallemand-Breitenbach V, Ferhi O, Koken M, Le Bras M,Duffort S, et al. PML/RARA oxidation and arsenic bindinginitiate the antileukemia response of As2O3. Cancer Cell 2010;18:88–98.


Lin et al.




2019;25:6228-6242. Published OnlineFirst June 24, 2019.Clin Cancer Res Jian Lin, Wu Zhang, Li-Ting Niu, et al. Progression and ATRA ResistanceTRIB3 Stabilizes High TWIST1 Expression to Promote Rapid APL

Updated version

10.1158/1078-0432.CCR-19-0510doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2019/06/22/1078-0432.CCR-19-0510.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/25/20/6228.full#ref-list-1

This article cites 49 articles, 16 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/25/20/6228.full#related-urls

This article has been cited by 1 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://clincancerres.aacrjournals.org/content/25/20/6228To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/lookup/doi/10.1158/1078-0432.CCR-19-0510

http://clincancerres.aacrjournals.org/content/suppl/2019/06/22/1078-0432.CCR-19-0510.DC1

http://clincancerres.aacrjournals.org/content/25/20/6228.full#ref-list-1

http://clincancerres.aacrjournals.org/content/25/20/6228.full#related-urls

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

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/25/20/6228


trib3 stabilizes high twist1 expression to promote rapid ... · translational cancer mechanisms and...

Documents