spatiotemporalregulationof dnp63bytgfb-regulated mirnas is ... · cancer research | molecular cell...

CANCER RESEARCH | MOLECULAR CELL BIOLOGY

Spatiotemporal Regulation ofDNp63by TGFb-RegulatedmiRNAs Is Essential for Cancer Metastasis A C

Ngoc H.B. Bui1,2, Marco Napoli1,2, Andrew John Davis1,2, Hussein A. Abbas3, Kimal Rajapakshe4,Cristian Coarfa4, and Elsa R. Flores1,2

ABSTRACT◥

DNp63 is a transcription factor of the p53 family and hascrucial functions in normal development and disease. Theexpression pattern of DNp63 in human cancer suggests dynam-ic regulation of this isoform during cancer progression andmetastasis. Many primary and metastatic tumors express highlevels of DNp63, while DNp63 loss is crucial for tumor dis-semination, indicating an oscillatory expression of DNp63during cancer progression. Here, we use genetically engineeredorthotopic mouse models of breast cancer to show that whiledepletion of DNp63 inhibits primary mammary adenocarcino-ma development, oscillatory expression of DNp63 in estab-

lished tumors is crucial for metastatic dissemination in breastcancer. A TGFb-regulated miRNA network acted as upstreamregulators of this oscillatory expression of DNp63 duringcancer progression. This work sheds light on the pleiotropicroles of DNp63 in cancer and unveils critical functions ofTGFb in the metastatic process.

Significance: This study unveils TGFb signaling and a networkof four miRNAs as upstream regulators of DNp63, providing keyinformation for the development of therapeutic strategies to treatcancers that commonly overexpress DNp63.

IntroductionThe TP63 gene encodes multiple isoforms categorized into two

groups, TAp63 and DNp63, which are expressed in different cellularcompartments and have distinct functions in many biological pro-cesses including cancer (1). DNp63, which lacks the full-length trans-activation domain, is the predominant isoform found in the basallayers of stratified epidermis and is essential for terminal differenti-ation and maintenance of basal epidermal cells. Both TAp63 andDNp63 transcripts can be transcribed into proteins with at least threedifferent C-termini termed a, b, and g via alternative splicing (2). Incancer, DNp63a is overexpressed in various primary tumors, such aslung squamous cell carcinomas (LUSC), head and neck squamous cellcarcinomas (HNSCC), basal-like bladder cancer, and ovariancancer (3–6). DNp63 exhibits oncogenic functions through its antag-onistic effects on the transcriptional activities of p53, TAp63, andTAp73 (1, 7–10). Importantly, tumor cells with high invasion capacityexpress low levels of DNp63, which induces an epithelial-to-mesenchymal transition (EMT) program in these cells (11–13). Onthe contrary, metastases at distant organs have high expression ofDNp63 (14, 15), indicating a dynamic oscillatory expression of DNp63during cancer progression.

Rare cells within a primary tumor lose contact with surroundingcells due to the loss of cell–cell adhesion, enter into the bloodcirculation, and successfully infiltrate and form metastases at distantorgans. The multistep metastatic process is orchestrated by a complexnetwork of signaling pathways, not only within primary tumor cells,but also between tumor cells and the surrounding microenviron-ment (16). Even though there are numerous studies that have focusedon the characterization of metastatic progression, definedmechanisticinsights into metastatic progression are still unknown. Interestingly,while DNp63 is found to be overexpressed in primary tumors andmetastases of multiple cancers (14, 15, 17), there is mounting evidencethat DNp63 is a suppressor of the EMT, cell migration and invasion,and potentially cancer metastasis (12, 18, 19). In particular, the loss ofDNp63 in epithelial-like bladder cancer cells promotes EMT andenhances invasion, while the induction of DNp63 in mesenchymal-like bladder cancer cells triggers the mesenchymal-to-epithelial tran-sition and suppresses invasion (19). This oscillatory expression ofDNp63 expression from primary tumor to tumor dissemination todistant metastasis resembles the spatiotemporal regulation of Twist1during metastasis (20), suggesting a potential dynamic regulation ofDNp63 in cancer metastasis. More importantly, we have recentlydemonstrated pleiotropic functions of DNp63 during tumor develop-ment and progression using a global pan-cancer approach to examinevarious types and stages of cancers in The Cancer Genome Atlas(TCGA; ref. 21). Together, these data indicate an intricate regulation ofDNp63 in tumor development and metastatic progression.

To understand the biological consequences of spatiotemporalexpression of DNp63 on cancer metastasis, we modulated DNp63expression during breast cancer metastasis in a genetically engi-neered orthotopic mouse model of breast cancer. By generating aninducible short hairpin RNA (shRNA) knockdown of DNp63 tomodulate its expression in vivo, we demonstrated that an oscillatoryexpression of DNp63 is required for efficient metastatic colonizationof breast cancer. We found that only breast cancer cells with anoscillatory expression of DNp63 (i.e., these cells initially expressDNp63 followed by a silencing of DNp63 and then a reexpression ofDNp63) efficiently metastasized and colonized the lungs of micewith mammary tumors. Furthermore, we identified a novel network

1Department of Molecular Oncology, H. Lee Moffitt Cancer Center, Tampa,Florida. 2Cancer Biology and Evolution Program, H. Lee Moffitt Cancer Center,Tampa, Florida. 3Hematology/Oncology Fellowship Program, The University ofTexas MD Anderson Cancer Center, Houston, Texas. 4Department of Molecularand Cellular Biology, Baylor College of Medicine, Houston, Texas.

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

Corresponding Author: Elsa R. Flores, Moffitt Cancer Center, 12902 USF Mag-nolia Drive, Tampa, FL 33612. Phone: 813-745-1473; E-mail:[email protected]

Cancer Res 2020;80:2833–47

doi: 10.1158/0008-5472.CAN-19-2733

�2020 American Association for Cancer Research.

AACRJournals.org | 2833

on August 5, 2020. © 2020 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst April 20, 2020; DOI: 10.1158/0008-5472.CAN-19-2733

http://crossmark.crossref.org/dialog/?doi=10.1158/0008-5472.CAN-19-2733&domain=pdf&date_stamp=2020-6-17

http://crossmark.crossref.org/dialog/?doi=10.1158/0008-5472.CAN-19-2733&domain=pdf&date_stamp=2020-6-17

http://cancerres.aacrjournals.org/

of four TGFb-Smad3–regulated miRNAs, including miR-22-3p,miR-30a-5p, miR-203a-3p, and miR-222-3p, which target DNp63and efficiently silence it. These TGFb-Smad3–dependent miRNAscooperatively regulate the expression of DNp63, not only in breastcancer, but also in other types of cancers where DNp63 is over-expressed. This novel TGFb/miRNA/DNp63 regulatory axis pro-vides new insights into the pleiotropic functions of DNp63 andTGFb in tumorigenesis and cancer metastasis.

Materials and MethodsCell lines and culture conditions

Human mammary epithelial MCF10A, and human breast cancerMCF10DCIS.com andMCF10CA1D.cl1 cell lines were obtained fromBarbara Ann Karmanos Cancer Institute, and cultured as describedpreviously (22). All the cells lineswere regularly authenticated via shorttandem repeat profiling by theMolecular Genomics Core at the H. LeeMoffitt Cancer Center (Tampa, FL).

Animal studiesMouse studies were approved by the Institutional Animal Care and

Use Committee at Moffitt Cancer Center. Female athymic nu/nu micewere purchased from Envigo and used at 6–8 weeks of age. Age-matched mice were used for all experiments.

Generation of inducible shRNA against DNp63An shRNA specifically targeting DNp63 (shDNp63: GAGGGAC-

TTGAGTTCTGTTAT) was designed and cloned into pLV-H1TetO-GFP-Puro Lentiviral Vector (SORT-CO1, Biosettia). Three micro-grams of the inducible shDNp63 lentiviral (pLV-i-shDNp63)vector were transfected into HEK 293T cells along with 3mg oflentivirus packaging vectors, pCMV-VSVG and pRSV-REV, usingX-tremeGENE HP (Roche) in accordance with the manufacturer'sprotocol. Forty-eight hours after the transfection, supernatants werecollected, filtered, and added to MCF10DCIS cells for 48 hours in thepresence of 8mg/mL Polybrene (Santa Cruz Biotechnology). Puromy-cin (2mg/mL) was added to the media 48 hours after infection for2 days to select for stable clones. Doxycycline (1 mg/mL) was added tothe stable cells for 3 days and then cell pellets were collected forWestern blot analysis to determine knockdown efficiency of DNp63protein.

In vivo mammary fat pad injectionA total of 2 � 106 cells mixed 1:1 ratio with growth factor–

reduced Matrigel (BD Biosciences) were injected into the left andright mammary fat pads of nude mice. Two weeks after theinjection, the tumors became palpable and the mice were fed witheither control diet or doxycycline diet (200 mg/kg, Bio-Serv) toinduce the expression of the shDNp63. At the end of the experi-ments, all the mice were euthanized and necropsy performed tocollect tumors and internal organs for further analysis. Tumor sizes(length and width) were measured using a caliper. Tumor volumewas calculated using the following formula: tumor volume ¼ 1/2(length � width2)

NanoString human miRNA panel to identify differentiallyexpressed miRNAs

A total of 2 � 105 MCF10DCIS cells were treated with TGFb1(10 ng/mL) for 72 hours, then collected for total RNA extraction usingmiRNaesy Kit (Qiagen) following themanufacturer's protocol. Experi-ments were done in quadruplicate. Then, RNA samples were run in the

NanoString Human microRNA Panel (NanoString Technologies)following the manufacturer's protocol.

Statistical analysisFor statistical comparison between two groups, an unpaired two-

tailed t test was used. A P value of less than 0.05 was consideredsignificant. Statistical analyses were performed in GraphPad Prism 7.

Data availabilityThe NanoString human microRNA data and the chromatin immu-

noprecipitation-sequencing (ChIP-seq) data were deposited to NCBIGene Expression Omnibus repository: GSE134681 and GSE144995.

ResultsOscillatory expression of DNp63 in breast cancer is associatedwith invasion and progression

To determine the pattern of expression of DNp63 during theprogression of breast cancer in human patients, we examined DNp63expression in a tissue microarray (TMA) including samples fromnormal breast tissue, lobular hyperplasia, ductal carcinoma in situ(DCIS), and invasive breast cancer biopsies (BR480, US Biomax). Weperformed DNp63 immunostaining and revealed a significantlydecreased expression of DNp63 in invasive breast cancer samplescompared with hyperplasia and carcinoma in situ (Fig. 1A–E),indicating that DNp63 must be lost for breast cancer invasion tooccur. We then determined the expression level of DNp63 in invasiveprimary tumors at different stages of disease progression using adistinct breast cancer TMA (BR20837a, US Biomax). In this TMA,we found a dramatic increase in the number of samples with lowDNp63 expression in stage IIB and IIIA compared with stage IIA(Fig. 1F), indicating that loss of DNp63 is associated with tumorprogression. Paradoxically, we found that the decrease in DNp63expression observed in the primary tumors progressing from stageIIA to IIIB was associated with an increase in the levels ofDNp63 in thematched lymph node metastases (Fig. 1G).

The association of DNp63 loss and the induction of cancer cellmigrationand invasionhavebeen reported inother cancers (12, 18, 19).To determine whether these features were also induced in breastcancer upon loss of DNp63, we made use of the MCF-10A breastcancer progression model (22–24). In this model, we depletedDNp63 in normal human mammary epithelial cells (MCF10A), theDCIS cells (MCF10DCIS), and metastatic cells (MCF10CA1D), andfound increased expression of EMT-associated factors, vimentinand Twist1 (Fig. 1H). E-cadherin expression did not significantlychange upon the depletion of DNp63, suggesting the induction of apartial EMT (Fig. 1H). Importantly, we also observed a significantincrease in the cell migration and invasion of MCF10DCIS cellswhen DNp63 was knocked down (Fig. 1I and J), indicating that thedownregulation of DNp63 is associated with the migratory andinvasive potential of breast cancer cells similarly to what has beenobserved in other cancers.

Oscillatory expression of DNp63 enhances metastaticdissemination

To decipher the activities of DNp63 on different stages of themetastatic cascade in breast cancer, we mimicked the oscillatoryexpression of DNp63 during breast cancer progression by generatinga doxycycline-inducible shRNA against DNp63 (i-shDNp63). In thepresence of doxycycline, protein expression of DNp63 inMCF10DCIScells harboring the inducible shDNp63 (MCF10DCIS-i-shDNp63) was

Bui et al.

Cancer Res; 80(13) July 1, 2020 CANCER RESEARCH2834




Figure 1.

Oscillatory expression DNp63 in breast cancer progression. A–D, Representative IHC staining images for DNp63 in breast cancer TMA. DNp63-positive nuclei arebrown. Hematoxylin was used as a nuclear counterstain. Magnification, �5. E, Quantification of DNp63 expression (DNp63 expression score) in breast cancer TNA.Score of each sample was determined by multiplying the number of positive nuclei by the average staining intensity. F, Percentage of primary tumors with eitherlow or high expression of DNp63 at different stages of breast cancer. G, IHC staining images for DNp63 in invasive primary human breast tumors at different stagesandmatched lymph nodemetastases. Percentage of DNp63-positive tumor cells is indicated.H, RepresentativeWestern blots of MCF10A (10A), MCF10DCIS (DCIS),and MCF10CA1D (CA1D) transfected with control siRNA (siC) or siRNA against DNp63 (siDNp63). Immunoblots were probed with the indicated antibodies. Actinwas used as a loading control. I and J,Quantification of cell migration (I) and invasion (J) assays of MCF10DCIS cells transfected with control siRNA or siDNp63. Dataare mean � SD. n ¼ 3. � , P < 0.05.

Spatiotemporal Regulation of DNp63 Enhances Metastasis

AACRJournals.org Cancer Res; 80(13) July 1, 2020 2835




Figure 2.

Oscillatory expression ofDNp63 in amousemodel ofmammary adenocarcinoma enhancesmetastatic dissemination.A,Experimental design to investigate effects ofoscillatory expression of DNp63 in breast cancermetastasis using an inducible shRNA for DNp63 and an orthotopic model of metastatic mammary adenocarcinoma.B, Primary mammary tumor volume (mm3) of mice injected with DCIS-i-shDNp63 cells and fed with different diet schedules of doxycycline (dox) as described in A.� , P < 0.05; ns, nonstatistically significant using a two-tailed t test. C, Pie graphs indicating the percentage of mice with lung metastases among the experimentalgroups as described in A. D, Working model demonstrating the effects of oscillatory expression of DNp63 in breast cancer metastasis.

Bui et al.





dramatically decreased after 3 days (Supplementary Fig. S1A). With-drawing doxycycline from the culture media for 2 days restoredexpression of DNp63 with a complete recovery of DNp63 expressionafter 4 days without doxycycline in the culture media (SupplementaryFig. S1A). Loss ofDNp63 expression inMCF10DCIS cells dramaticallyreduced primary tumor growth in vivo by more than two fold(Supplementary Fig. S1B and S1C). These data are in accordance withprevious studies demonstrating the essential oncogenic roles ofDNp63for the establishment of primary tumors in thymic lymphoma andcutaneous squamous cell carcinoma (cuSCC; refs. 9, 10).

Because p63 was previously reported to be overexpressed in bothprimary tumors and metastases of cuSCCs (14, 15), we wanted tofurther investigate how DNp63 modulates the complete process ofmetastatic dissemination, from primary tumor formation to lungcolonization using metastatic breast cancer as a model system. As

described in Fig. 2A, MCF10DCIS-i-shDNp63 cells were injected intomammary fat pads of nude mice. After 2 weeks tumors becamepalpable, and one cohort of 10 mice was fed with control diet tomaintain a “highDNp63” expression (group 1). The second cohort wason doxycycline diet for 3 weeks, then switched to control diet for 3moreweeks to oscillate the expression ofDNp63 “low thenhighDNp63expression” (group 2). The last cohort was continuously fed a doxy-cycline diet to keep a “low DNp63” expression (group 3). Consistentwith our previous results shown in Supplementary Fig. S1C, down-regulation of DNp63 in established tumors caused reduction in tumorgrowth (Fig. 2B). In addition, primary tumors from group 3 with “lowDNp63” showed decreased expression of DNp63 and Ki67 comparedwith group 1 and group 2 (Supplementary Fig. S2A and S2B),indicating that the downregulation of DNp63 resulted in a decreasein tumor cell proliferation, which in turn led to reduced tumor growth.

Figure 3.

DNp63 reexpression is required for lung extravasation and colonization of breast cancer MCF10DCIS cells. A, Experimental design to investigate roles of DNp63 inextravasation (48 hours) and colonization (8weeks) to the lungs ofmice that have been treatedwithMCF10DCIS cells expressing an inducible shRNA againstDNp63.B and C, Flow cytometry assay (B) and quantification (C) of the percentage of RFP-positive MCF10DCIS cells in the lungs. Data are mean � SD. n ¼ 6. � , P < 0.05.D andE,Representative hematoxylin and eosin (H&E) cross-sections of lungs from the indicated groups (D) and quantification ofmetastatic colonies per lung sectionper mouse (E). Data are mean � SD. n ¼ 4 (group 1 and group 2) or n ¼ 5 (group 3). � , P < 0.0005. dox, doxycycline.






Bui et al.





All the mice were also assessed for spontaneous metastasis to the lungswithout removal of the primary tumors. Strikingly, tumor cells with“oscillatory DNp63” (group 2) exhibited the highest capacity tometastasize and colonize the lungs, with 70% of mice developing lungmetastases compared with 30% from group 1 (Fig. 2C). Tumor cellswith “low DNp63” (group 3) also showed an increase in metastaticfrequency to the lungs compared with “high DNp63” group (group 1).Interestingly, lung metastases from all three groups were positive forDNp63 staining (Supplementary Fig. S2C–S2E), suggesting that sub-sets of cells with high expression of DNp63 may possess molecularfeatures that ultimately favor metastatic colonization.

In summary, these data revealed that the oscillatory expression ofDNp63 facilitates metastatic dissemination of tumor cells to distantorgans. The depletion of DNp63 in primary tumor cells, while inhibit-ing cell proliferation, also triggered a cell migration and invasionprogram that could potentially increase intravasation into the bloodcirculation (Fig. 2D). This mechanism of action sheds new lights onvarying expression of DNp63 during mammary adenocarcinomadisease progression.

DNp63 is essential for breast cancer cell extravasation and lungcolonization of circulating tumor cells

The ability of tumor cells to disseminate and proliferate in distantorgans is crucial for the establishment of micrometastatic lesions andsubsequent colonization (16, 25). Because our data show that DNp63promotes cancer cell proliferation and formation of primary tumors,we further determined whether the expression of DNp63 is alsorequired for the establishment of distant metastases to the lungs, thusexplaining why an oscillatory expression of this isoform enhancesmetastasis as shown previously in this study. There are two criticalsteps involved in the formation of overt metastatic colonies at distantsties: extravasation and colonization. To investigate contribution ofDNp63 in extravasation of circulating tumor cells into the lungs, weperformed an extravasation assay in which fluorescence-labeled tumorcells were injected into nude mice via tail vein, and extravasated cellsinto the lungs after 48 hours were quantified using flow cytometry (20).Specifically, MCF10DCIS-i-shDNp63 cells were labeled with red fluo-rescence protein (RFP) and cultured in doxycycline for 3 days tosilence DNp63 prior to tail vein injection into nude mice receivingcontrol diet or doxycycline diet as described in Fig. 3A. The untreatedMCF10DCIS-i-shDNp63 cells were injected into nudemice on controldiet as “high DNp63” (group 1). After 48 hours, the number of RFP-positive cells successfully extravasating into the lungs was quantified.The results showed that depletion of DNp63 in circulating tumor cells(“low DNp63”, group 3) resulted in a significant decrease in thenumber of RFP-positive cells in the lungs compared with controltumor cells with “high DNp63” expression (group 1). Notably, tumorcells with “oscillatory DNp63” (group 2) had similar percentage ofextravasated RFP-positive cells to “high DNp63” control cells(group 1; Fig. 3B and C), indicating that DNp63 expression is

essential for circulating cells to extravasate into the lungs at the earlystage of lung colonization.

We next determined the ability of these three groups with differentDNp63 expression patterns to colonize the lungs. To do this, werepeated the above assay by injecting MCF10DCIS-i-shDNp63 cellstreated with doxycycline for 3 days to silence DNp63 into nude micefed with either control diet or doxycycline diet as described in Fig. 3A.In this experiment, the mice were aged for 8 additional weeks afterinjection and the lungs were collected and quantified for the number ofmetastatic lesions (Fig. 3A). In-line with our previous results, “oscil-latory DNp63” (group 2) had the greatest number of visible nodules inthe lungs similar to “high DNp63” (group 1), whereas “low DNp63”(group 3) exhibited a dramatic reduction in overt metastasis formation(Supplementary Fig. S2F and S2G). The quantification data from lungcross-sections stained with hematoxylin and eosin confirmed thehighest number of metastatic lesions in “oscillatory DNp63” (group 2)and “high DNp63” (group 1) compared with “low DNp63” (group3; Fig. 3D and E). These data indicate that expression of DNp63 iscritical for efficient lung colonization of breast cancer cells.Surprisingly, metastases in all three groups showed high expressionof DNp63 regardless of our experimental manipulation of DNp63during the course of the study (Supplementary Fig. S2H), indicatingthat cancer cells expressing DNp63 exhibit favorable characteristics tocolonize distant lungs and that only cells that escape DNp63 depletionare capable of colonization. Taken together, these in vivo data dem-onstrated that DNp63 is required for both extravasation and coloni-zation of breast tumor cells at distant lungs.

To determine whether DNp63 regulates genes and pathwaysrelevant to cancer progression and dissemination in our in vivomodel,we injected MCF10DCIS-i-shDNp63 cells into the mammary fat padsof nude mice and fed them with either control or doxycycline diet.After 5 weeks, the primary tumors were collected to perform ChIP-seqanalyses with DNp63 and RNA polymerase II (Pol II) antibodies asdescribed in Supplementary Fig. S3A. Differential binding profilesof DNp63 and Pol II between the MCF10DCIS-control andMCF10DCIS-shDNp63 tumors were compared to identify DNp63-induced and -repressed gene signatures. Then, pathway analyseswere performed. We found enrichment of pathways involved indevelopment, cell metabolism, and cell motility, including someTGFb/EMT-related pathways (Supplementary Fig. S3B and S3C;Supplementary Table S1). These data are in accordance with previ-ously reported roles of DNp63 in developmental processes and cellmotility regulation (12, 18, 19, 26–28), indicating thatDNp63 regulatesa large network of downstream target genes during cancer progression.

TGFb inhibits expression of DNp63 via canonical Smad-dependent signaling in breast cancer

Given that the regulation of DNp63 plays pivotal roles in metastaticdissemination, we aimed to identify upstream regulators of DNp63during this process. TGFb signaling has been shown to have

Figure 4.TGFb inhibits expression of DNp63 via canonical Smad-dependent signaling in MCF10A progression model. A and B, Representative Western blots of MCF10A,MCF10DCIS (DCIS), andMCF10CA1D (CA1D) cells treatedwith TGFb (10 ng/mL) or with TGFb and the TGFBRI inhibitor LY2157299 (1 mmol/L) for 72 hours (A) or withTGFb and siSmad3 (B). Immunoblotswere probedwith the indicated antibodies. Actinwas used as a loading control.C andD,Quantification of cell migration (C) andinvasion (D) assay ofMCF10DCIS cells overexpressingDNp63a in the presence of TGFb (10 ng/mL). E,RepresentativeWestern blots of MCF10DCIS cells treatedwithorwithout LY2157299 (1mmol/L) for 72 hours. Immunoblotswere probedwith the indicated antibodies. Actinwas used as a loading control. F andG,Quantification ofcell migration (F) and invasion (G) assay of MCF10DCIS cells treated with siDNp63 and/or LY2157299 (1 mmol/L). H, Heatmap of differentially expressed miRNAsbetween TGFb-treated (þTGFb-1, 2, 3, and 4) and untreated (�TGFb-1, 2, 3, and 4) MCF10DCIS cells. I, Venn diagram showing 64 TGFb-induced miRNAs that arepredicted to bind to the 30 UTR of TP63. J,RepresentativeWestern blot of MCF10DCIS cells transfected with the indicatedmiRNAmimics. Immunoblotswere probedwith the indicated antibodies. Actin was used as a loading control. �, P < 0.05.






duplicitous roles in breast cancer development and metastasis. Innormal cells and early-stage tumors, TGFb exerts its tumor suppres-sive functions to inhibit tumorigenesis via inducing apoptosis and cellgrowth arrest. On the contrary, in advanced cancer, TGFb functions asa prometastatic factor by inducing EMT and promoting breast cancercell migration and invasion, which ultimately favors metastatic dis-semination (29, 30). Interestingly, a previous study revealed thatconstitutive activation of TGFb signaling promotes single-cell inva-sion and intravasation in primary tumors, yet fails to facilitate tumorgrowth of distant metastases, whereas a localized and reversible TGFbsignaling is required for sufficient metastatic dissemination of breastcancer (31). This mechanism of action resembles our observations onhow DNp63 exerts its function during the multistep process ofmetastasis, suggesting a potential regulatory network connectingTGFb signaling and DNp63.

Indeed, in the presence of TGFb, MCF10DCIS cells shared similarmorphology to DNp63-depleted cells (Supplementary Fig. S4A).Remarkably, DNp63 expression was dramatically reduced whenMCF10A, MCF10DCIS, and MCF10CA1D cells were treated withTGFb (Fig. 4A; Supplementary Fig. S4B). We also observed increasedphosphorylation of Smad2/3 in all three cell lines upon TGFb treat-ment (Supplementary Fig. S4B), demonstrating the activation of TGFbdownstream signaling. To determine whether TGFb signaling regu-lated DNp63, we treatedMCF10A,MCF10DCIS, andMCFCA1D cellswith LY2157299, an inhibitor of TGFb-receptor I (TGFbRI), toinactivate the TGFb signaling cascade. We found that the inhibitoryeffect of TGFb on DNp63 was rescued by the addition of LY2157299inhibitor (Fig. 4A). In addition, the overexpression of the active formof TGFb, but not the latent form, reduced the expression of DNp63(Supplementary Fig. S4C). MCF10DCIS cells cultured in conditionedmedia from these cells exhibited decreased level of DNp63, suggestingthat endogenous TGFb secreted by tumor cells can also reduce DNp63expression (Supplementary Fig. S4D). TGFb signaling is transducedthrough either canonical Smad-dependent pathways or noncanonicalSmad-independent pathways (32). To understand DNp63 regulationby TGFb, we knocked down Smad3 in the presence of TGFb todetermine whether the regulation of TGFb on DNp63 is Smad-dependent. The ablation of Smad3 in MCF10A andMCF10DCIS cellsrestored DNp63 expression (Fig. 4B), demonstrating that TGFbreduces DNp63 expression via Smad2/3-dependent transcriptionalactivities. Moreover, while TGFb enhanced migration and invasionin MCF10DCIS cells, the overexpression of exogenous DNp63a inTGFb-treated cells largely counteracted these effects (Fig. 4C and D;Supplementary Fig. S4E). Next, we assessed whether endogenousTGFb can modulate DNp63 expression similarly to exogenous TGFb.To do this, we inhibited the activation of the signaling cascade usingLY2157299 in MCF10DCIS cells, which had high level of p-Smad2/3compared with MCF10A andMCF10CA1D (Fig. 4A), and found thatinhibition of the endogenous TGFb signaling increasedDNp63 level inthese cells (Fig. 4E). In addition, downregulation of DNp63 in thepresence of the inhibitor partially restored the cell migration andinvasion ability of MCF10DCIS cells (Fig. 4F and G). Together, thesefindings indicate that DNp63 is regulated downstream of canonicalSmad2/3-dependent TGFb signaling and affects TGFb-dependentbiological processes.

Novel TGFb-induced miRNAs, including miR-22-3p,miR-30a-5p, miR-203a-3p, and miR-222-3p, regulateexpression of DNp63

Tounderstandwhether there is a direct transcriptional repression ofSmad2/3 complex on the expression of DNp63, we examined mRNA

level of DNp63 upon TGFb signaling activation. However, no signif-icant reduction of DNp63 mRNA was observed in the presence ofTGFb (Supplementary Fig. S4F). This data suggested a posttranscrip-tional regulation of Smad2/3 on DNp63 expression. miRNAs havebeen intensively studied for their roles in posttranscriptional regula-tions of gene expression. Cross-talk between TGFb and miRNAmachinery has been reported in multiple cancers (33). Hence, weused the NanoString nCounter Human miRNA platform to profiledifferentially expressed miRNAs in MCF10DCIS cells upon TGFbtreatment to identify TGFb-regulated miRNAs that potentially targetDNp63. Using this method, we identified differentially expressedmiRNAs upon TGFb treatment including several known targets ofTGFb signaling such as miR-181, miR-145, and miR-21 (33), furtherconfirming the activation of TGFb signaling in these cells (Fig. 4H).Our screen identified 74 upregulated miRNAs upon TGFb treatment(Fig. 4H; Supplementary Table S2). We then merged these identifiedmiRNAs with a list of miRNAs predicted to have binding sites to the 30

untranslated region (UTR) of TP63 gene using the miRWalk predic-tion algorithm because both TAp63 and DNp63 isoforms share thesame 30 UTR of TP63. This approach resulted in 64 miRNAs that canpotentially bind to the 30 UTR of TP63 and regulate DNp63 expressionupon activation of TGFb signaling (Supplementary Table S3). Amongthe 64 miRNAs, nine miRNAs, including miR-10b-5p, miR-21-5p,miR-22-3p, miR-30a-5p, miR-141-3p, miR-181a-5p, miR-200a-3p,miR-203a-3p, and miR-222-3p, were selected for further validationon the basis of predicted binding affinity to the 30 UTR of TP63 andpreviously reported connections with breast cancer and/or p63(Fig. 4I). We then overexpressed mimics of these nine miRNAs andfound that miR-22-3p, miR-30a-5p, miR-203a-3p, and miR-222-3pdownregulated DNp63 protein level in MCF10DCIS cells (Fig. 4J). Asimilar reduction in DNp63 expression was obtained in MCF10A andMCF10CA1D cells (Supplementary Fig. S5A). In addition, these fourmiRNAs caused reduced luciferase activity when transfected togetherwith the 30 UTR of DNp63 mRNA cloned into a luciferase reportervector (Fig. 5A–D).We also generatedmutated versions of the 30 UTRof DNp63 at the binding sites of miR-22-3p (Supplementary Fig. S5B),miR-30a-5p (Supplementary Fig. S5C), miR-203a-3p (SupplementaryFig. S5D), and miR-222-3p (Supplementary Fig. S5E) by replacingthese binding sites on the 30 UTR with the seed sequences of thecorresponding miRNAs. Mutations in the binding sites of miR-22-3p(Fig. 5A), miR-30a-5p (Fig. 5B), miR-203a-3p (Fig. 5C), andmiR-222-3p (Fig. 5D) on the 30 UTR abolished miRNA-mediatedrepressive activities. These results indicate that these four TGFb-regulated miRNAs control DNp63 expression in breast cancer cellsand could provide a mechanism for the spatiotemporal regulation ofDNp63.

We further determined whether these four miRNAs are regulateddownstream of canonical TGFb signaling. Our data showed that TGFbtreatment upregulated the expression of these four miRNAs inMCF10DCIS cells, whereas the addition of TGFbRI inhibitor pre-vented increased expression of the miRNAs triggered by TGFbactivation (Supplementary Fig. S5F–S5I). Likewise, we found thatmiRNA upregulation by TGFbwas rescued by knocking down Smad3in MCF10DCIS cells (Fig. 5E–H). We then performed ChIP to assessSmad2/3 binding to the promoters of the four identified miRNAs inMCF10DCIS cells. Our results revealed that Smad2/3 indeed binds tothese four promoter regions, demonstrating that these four miRNAsare direct targets of canonical Smad-dependent TGFb signaling(Fig. 5I–L). Next, we asked whether inhibition of the miRNAs canrescue inhibitory effects of TGFb on DNp63 expression. Interestingly,inhibition of all four miRNAs, but not any single miRNA, restored the

Bui et al.





Figure 5.

Novel TGFb-induced miRNAs (miR-22-3p, miR-30a-5p, miR-203a-3p, and miR-222-3p) regulate expression of DNp63. A–D, Relative luciferase expression inMCF10DCIS (DCIS) cells transfectedwith 30 UTRof either human TP63wild-typemRNAormutated versions of the 30 UTR at binding sites formiR-22-3p (A),miR-30a-5p (B),miR-203a-3 (C),miR-222-3p (D), and indicatedmiRNAmimics. Data aremean� SD. n¼ 3.E–H,qRT-PCRofmiR-22-3p (E),miR-30a-5p (F),miR-203a-3p (G),andmiR-222-3p (H) inMCF10DCIS cells treatedwith TGFb (10 ng/mL) alone or TGFb and siSmad3. Data aremean� SD. n¼ 3. J–L,ChIP to detect binding of Smad2/3to the promoter of miR-22-3p (I), miR-30a-5p (J), miR-203a-3p (K), and miR-222-3p (L) upon TGFb treatment. Data are mean � SD. n ¼ 3. M, Representativeimmunoblots of MCF10DCIS cells treated with TGFb alone or TGFb and the indicated miRNA inhibitors. Immunoblots were probed with the DNp63 antibody. Actinwas used as a loading control. N and O, Quantification of cell migration (N) and invasion (O) assay of MCF10DCIS cells treated with a combination of four miRNAinhibitors for miR-22-3p, miR-30a-5p, miR-203a-3p, and miR-222-3p in the presence of TGFb (10 ng/mL). Data are mean� SD. n¼ 3. P, qRT-PCR of the indicatedmiRNAs in MCF10DCIS cells treated with or without LY2157299 (1 mmol/L). Data are mean� SD. n ¼ 3. Q and R, Quantification of cell migration and invasion assayshowingmigration (Q) and invasion (R) capacity ofMCF10DCIS cells treatedwith a combination of fourmiRNAmimics formiR-22-3p,miR-30a-5p,miR-203a-3p, andmiR-222-3p in the presence of LY2157299 (1 mmol/L). Data are mean � SD. n ¼ 3. � , P < 0.05.






expression ofDNp63 in the presence of TGFb (Fig. 5M), indicating theimportance of all four in DNp63 silencing. The inhibition of these fourmiRNAs also reducedmigration and invasion capacity ofMCF10DCIScells treated with TGFb (Fig. 5N and O), thus indicating theirrelevance for TGFb-dependent biological activities. We also observeda decrease in Smad3 phosphorylation level in the presence of the fourmiRNA inhibitors and TGFb (Supplementary Fig. S5J), suggesting apossible feedback regulation among the four miRNAs and TGFbsignaling to modulate DNp63. In addition, the inhibition of endog-enous TGFb signaling by LY2157299 in MCF10DCIS cells alsoreduced the expression of all four miRNAs (Fig. 5P). The overexpres-sion of all four miRNAs using miRNA mimics in the presenceof LY2157299 rescued the migration and invasion capacity ofMCF10DCIS cells (Fig. 5Q and R). Together, these data demonstratethat canonical TGFb signaling modulates the expression of DNp63 viaa network of four TGFb-inducedmiRNAs, includingmiR-22-3p,miR-30a-5p, miR-203a-3p, and miR-222-3p.

In vivo TGFboverexpression reduces primary tumor growth andmetastatic colonization

To demonstrate the in vivo effects of the TGFb/DNp63 axis onprimary tumor growth and lung colonization, we injectedMCF10DCIS cells expressing either a latent or an active form ofTGFb1 into mammary fat pads and tail veins of nude mice(Fig. 6A). Because TGFb is produced in a biologically inactive (orlatent) form, then gets activated by various extracellular matrixcomponents (34), andwedid not observe the activation of the signalingcascade when overexpressing the latent TGFb in vitro (SupplementaryFig. S4C), we wanted to examine whether the overexpression of thislatent form in vivo would activate the cascade. Indeed, the over-expression of both latent and active form of TGFb1 significantlyreduced DCIS primary tumor volume (Fig. 6B). Immunostainingresults confirmed high expression of TGFb1 together with the acti-vation of the downstream signaling and downregulation of DNp63 inthese tumors (Fig. 6C and D). Similar to the depletion of DNp63(Fig. 3D andE), TGFb1 overexpression also hindered the formation ofovert metastatic colonies in the lungs (Fig. 6E and F). Taken together,these data further emphasize the contribution of TGFb in reducingprimary tumor growth and metastatic progression in vivo through themodulation of DNp63.

TGFb/miRNAs axis is required for the downregulation ofDNp63in squamous cell carcinomas and bladder cancer

BecauseDNp63 is overexpressed inmany cancers (14, 15, 35, 36), weinvestigated the importance of the TGFb/miRNAs/DNp63 axis inother types of cancers with high expression of DNp63. To do this,we treated a panel of human cancer cell lines from various sites,including cuSCC, bladder cancer, HNSCC, and LUSC, with TGFb andexamined DNp63 protein expression. We found that TGFb signalingactivation decreased expression of DNp63 in three cuSCC lines,including COLO16, RDEB2, and IC1, one bladder cancer line 5637,and one HNSCC line 22B. In accordance with our previous findings,treatment with TGFbRI inhibitor, LY2157299, also restored DNp63expression in these cells (Fig. 7A–C). Surprisingly, TGFb did notexhibit inhibitory effects on DNp63 expression of one cuSCC line,SRB12, and one LUSC line, HCC95, despite the fact that the signalingcascade was activated indicated by the phosphorylation of Smad3(Supplementary Fig. S6A and S6B). Notably, although DNp63 is aknown marker for LUSC, we found that out of five LUSC cell linesDNp63 expression was high only in the HCC95 cell line (Supplemen-tary Fig. S6C). This prompted us to ask whether the four miRNAs

identified in breast cancer were upregulated byTGFb signaling in thesecancer cell lines. Indeed, miR-22-3p, miR-30a-5p, miR-203a-3p, andmiR-222-3p expression was increased in COLO16, RDEB2, and IC1cells treated with TGFb (Fig. 7D–F). However, we did not observesignificant induction of these fourmiRNAs in both SRB12 andHCC95(Supplementary Fig. S6D and S6E). We further asked whether theLUSC cell lines with undetectable or low levels of DNp63 actually havehigh expression of the four miRNAs of our interest. Indeed, H226,H1703, and H2286 lines showed significantly higher expression ofmiR-22-3p, miR-30a-5p, and miR-222-3p compared with HCC95(Supplementary Fig. S6F–S6H), indicating that high levels of thesemiRNAs decrease expression of DNp63, leading to its low steady statelevel in these cell lines. Moreover, treatment of LY2157299 in H226,H170, and H2286 lines not only induced expression of DNp63(Supplementary Fig. S6I), but decreased level of the four miRNAs(Supplementary Fig. S6J–S6L), indicating that the high level of the fourmiRNAs and low expression of DNp63 in these cells are TGFbdependent. In summary, these data indicate that elevated expressionof these miRNAs is essential to modulate DNp63 levels downstream ofTGFb signaling cascade in various types of cancer and is critical to thespatiotemporal expression of DNp63.

Bioinformatic analyses revealed an anticorrelation betweenTGFb-driven transcriptional signature and DNp63-driventranscriptional signature in different cancermolecular subtypes

Our results showing that TGFb signaling affects DNp63 in cell linesof different origin, including breast and lung cancer, prompted us todetermine whether the cross-talk between TGFb and DNp63 is alsooccurring in tumors of patients with breast and lung cancer. Therefore,we utilized a TGFb signature (37) and a DNp63 signature (21) asreadouts of the activity of TGFb and DNp63, respectively, to analyzefour clusters of TCGA datasets (36): C1 (enriched in lung adenocar-cinomas), C2 (enriched in LUSC), C3 (enriched in luminal-type breastcancers), and C4 (enriched in basal-type breast cancers). Principalcomponent analysis (PCA)was performed using theTGFb andDNp63signatures in each of these four clusters. Notably, in C2 (LUSC) but notin C1 (lung adenocarcinomas) there is a negative correlation betweenthe two signatures (Fig. 7G and H; Supplementary Table S4). This isin-line with DNp63 being a crucial marker in LUSC (38) and with ourdata showing that DNp63 levels increase in LUSC cell lines if TGFbsignaling is inhibited (Supplementary Fig. S6I). In the case of the twobreast cancer clusters, there is a negative correlation between the TGFband DNp63 signatures in C3 (luminal subtype) but not in C4 (basalsubtype; Fig. 7I and J; Supplementary Table S4). This suggests that thecontrol of DNp63 by TGFb could be more relevant in breast cancerswith low levels of DNp63 (luminal subtype) than in those with highlevels of DNp63 (basal subtype; refs. 39–42). In summary, these datademonstrate a negative correlation between TGFb and DNp63 sig-natures in subtypes of breast and lung tumors, further providingevidence for the regulatory axis of TGFb/DNp63 in human cancers.

DiscussionGiven its pivotal roles in the maintenance and proliferation of

epithelial cells, DNp63 is highly expressed in primary tumors andmetastases across multiple epithelial cancers (35, 36, 43). Work fromour laboratory and others have unveiled tumor-promoting activities ofDNp63 in various cancer types (9, 10, 21). However, the suppressivefunction of DNp63 on the cell adhesion and motility through regu-lation of different transcription factors and miRNAs has been dem-onstrated in multiple cell lines and cancers (11–13, 18, 19, 44),

Bui et al.





Figure 6.

TGFb overexpression reduces mammary adenocarcinoma primary tumor growth and metastatic lung colonization. A, Experimental design to investigate effects ofTGFb on mammary tumor development and metastatic lung colonization using orthotopic mouse models and tail vein injections. B, Primary mammaryadenocarcinoma volume (mm3) of mice injected with MCF10DCIS (DCIS) cells expressing either empty vector (EV), latent TGFb1, or active TGFb1 as describedin A. Data are mean � SD. n ¼ 6. C, IHC to detect TGFb1, Smad2, p-Smad2, and DNp63 in cross-sections of MCF10DCIS primary tumors overexpressing TGFb1. D,Quantification of the percentage ofDNp63-positive cells inMCF10DCIS primary tumors inC.E andF,Representative hematoxylin and eosin–stained cross-sections oflung tissues (E) and quantification of metastases per lung lobe (F). Data are mean � SD. n ¼ 15. � , P < 0.05; ns, not significant.






Figure 7.

TGFb/miRNAs axis is required for the downregulation of DNp63 in other cancers including SCCs and bladder cancer. A–C, Representative immunoblots ofcuSCC lines COLO16, RDEB2, and IC1 (A), bladder cancer cells 5637 (B), and HNSCC 22B (C) treated with TGFb or with TGFb plus TGFBRI inhibitor, LY2157299.Immunoblots were probed with the indicated antibodies. Actin was used as a loading control. D–F, qRT-PCR of the indicated miRNAs in COLO16 (D), RDEB2(E), and IC1 (F) cells treated with or without TGFb. Data are mean � SD. n ¼ 3. � , P < 0.05. G–J, PCA analyses comparing a DNp63 transcriptional signature anda TGFb transcriptional signature in the indicated clusters, including C1, lung adenocarcinoma (G), C2, LUSC (H), C3, luminal breast cancer (I), and C4, basalbreast cancer (J) subtypes.

Bui et al.





indicating tumor suppressive activities of DNp63 in cell migration andinvasion. In this study, we found thatDNp63 expression is essential forbreast primary tumor development, consistent with its role as anoncogene in multiple cancers, and that this is achieved via the DNp63-dependent regulation of numerous pathways, including in develop-ment, metabolism, signal transduction, and TGFb-regulated EMT.More interestingly, our in vivo data demonstrated an indispensablerole of DNp63 for circulating tumor cells to extravasate and establishnew colonies at distant organs. This is in-line with the fact that DNp63expression is crucial for the maintenance of epithelial characteristicsand that the reversion to epithelial state of circulating tumor cellsarrested at distant organs is important for cell proliferation andestablishment of metastatic lesions (16, 45). Importantly, we showedthat an oscillatory expression of DNp63 is essential for efficientmetastatic progression of breast cancer both in genetically engineeredorthotopic mouse models and in patients with breast cancer. Thisdynamic expression of DNp63 is modulated by TGFb and a novelnetwork of four TGFb-regulated miRNAs in various DNp63-expressing cancers. Together, our findings support a model in whichspatiotemporal regulation of DNp63 and its oscillatory expression iscrucial for metastasis.

The overexpression of p63/DNp63 has been implicated as amolecular feature of various squamous cell carcinomas (SCC;refs. 35, 36). The most common mechanistic explanation for highp63 in these cancers is the amplification of chromosome 3q (35).Here, we provide a detailed mechanism through canonical TGFbsignaling and its miRNA targets for the regulation of DNp63.Previously, multiple studies have focused on characterizing molec-ular pathways downstream of DNp63 that contribute to cancerprogression and metastasis (9, 13, 18, 19, 44, 46, 47). In this study,we aimed to elucidate the regulatory network upstream of DNp63and emphasize its importance in cancers. Our work demonstratedthat TGFb signaling represses DNp63 expression at the posttran-scriptional level in both normal mammary epithelial cells and breastcancer cells via canonical Smad3-dependent pathway and down-stream miRNAs. Activated TGFb signaling has been stronglyassociated with increased metastatic dissemination via inductionof an EMT program in tumor cells (30, 32). Yet, impaired TGFbsignaling cascade also resulted in enhanced metastasis (48, 49).TGFb has also been shown to function as a molecular switchbetween collective and single-cell invasion programs in breastcancer (31), further emphasizing pleiotropic roles of TGFb inmetastatic dissemination, which is similar to our proposed modelon the contributions of DNp63 in metastasis. Of note, TGFb hasalso been previously reported to have a positive regulatory effect onDNp63 (50, 51), even though these studies were performed in othertumor types (51) or did not distinguish between the different p63isoforms (50), which have opposing effects on cell motility andcancer metastasis (19, 46, 52).

The four TGFb-induced miRNAs identified as upstream regulatorsof DNp63, including miR-22-3p, miR-30a-5p, miR-203a-3p, andmiR-222-3p, exhibit inhibitory effects on the protein expression ofDNp63. Interestingly, the inhibition of four miRNAs, but not anysingle miRNA, rescued the inhibitory effects of TGFb on DNp63expression, suggesting cooperative functions of those miRNAs inmodulating DNp63 downstream of TGFb signaling. This TGFb/miRNAs/DNp63 axis was further validated in several cancers thathighly express DNp63, such as bladder cancers, and cutaneous andlung SCCs. Among the four miRNAs, miR-203a-3p is a knownmodulator of DNp63 during terminal differentiation of the epider-

mis (53). The identification of the negative regulatory effect of thismiRNA on DNp63 downstream of TGFb signaling indicates itsconsistent role as a regulator of DNp63 across different cell types. Inbreast cancer, the roles of miR-203 are not clear. Studies have shownits function as either a suppressor or promoter of breast can-cers (54, 55), suggesting a perplexing role of this miRNA in breastcancer. Similarly, tumor-promoting and tumor-suppressive roles ofmiR-22-3p have also been documented (56, 57). miR-30a-5p, on theother hand, has been consistently demonstrated to be a suppressorof breast cancer cell growth and metastasis (58, 59). Conversely,miR-222-3p has been implicated as an oncogene in breast cancermetastasis by promoting EMT in basal-like breast cancer (60, 61).These contradictory findings indicate a more intricate regulatorynetwork involving these miRNAs in breast cancer and can be dueto context-dependent effects likely regulated by TGFb and furtherby DNp63.

Taken together, our data have unveiled oscillatory expression ofDNp63 during breast cancer progression. Moreover, we identified anovel regulatory mechanism for DNp63 downstream of TGFb signal-ing and miRNAs in modulating metastatic dissemination. Our dataprovide key information of the upstream regulation of DNp63 oscil-latory expression during breast cancer metastasis that is key todesigning DNp63-targeted therapies to treat cancer.

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

Authors’ ContributionsConception and design: N.H.B. Bui, M. Napoli, A.J. Davis, H.A. Abbas, E.R. FloresDevelopment of methodology: N.H.B. Bui, H.A. Abbas, E.R. FloresAcquisition of data (provided animals, acquired and managed patients, providedfacilities, etc.): N.H.B. Bui, M. Napoli, A.J. Davis, H.A. AbbasAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): N.H.B. Bui, M. Napoli, A.J. Davis, H.A. Abbas,K. Rajapakshe, C. Coarfa, E.R. FloresWriting, review, and/or revision of the manuscript: N.H.B. Bui, M. Napoli,H.A. Abbas, E.R. FloresAdministrative, technical, or material support (i.e., reporting or organizing data,constructing databases): N.H.B. BuiStudy supervision: E.R. Flores

AcknowledgmentsThis work was supported by R35CA197452 to E.R. Flores. E.R. Flores is a NCI

Outstanding Investigator,Moffitt Distinguished Scholar, and Scholar of the Leukemiaand Lymphoma Society, the Rita Allen Foundation, and the V Foundation for CancerResearch. N.H.B. Bui is a VietnamEducation Foundation Fellow and a recipient of theAndrew Sowell-Wade Huggins Scholarship in Cancer Research. K. Rajapakshe andC. Coarfa were partially supported by The Cancer Prevention Institute of Texas(CPRIT) RP170005, NIH P30 shared resource grant CA125123, and NIEHS grants1P30ES030285 and 1P42ES0327725. We thank X. Su, R. Checker, X. Li, andI. Grammatikakis for technical advice, J. Yao for the analysis of the NanoString dataand the ChIP-seq data, and E.A. Welsh for the PCA analysis of the TCGA datasets.This work has been supported in part by the Molecular Genomics Core, the FlowCytometry Core, the Biostatistics and Bioinformatics Core, the Vivarium servicesCore, the Analytic Microscopy Core and the Tissue Core at Moffitt Cancer Center,and an NCI-designated Comprehensive Cancer Center (P30-CA076292).

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.

Received September 1, 2019; revised March 18, 2020; accepted April 15, 2020;published first April 20, 2020.






References1. Su X, Chakravarti D, Flores ER. p63 steps into the limelight: crucial roles in

the suppression of tumorigenesis and metastasis. Nat Rev Cancer 2013;13:136–43.

2. Chakravarti D, Su X, Cho MS, Bui NH, Coarfa C, Venkatanarayan A, et al.Inducedmultipotency in adult keratinocytes throughdown-regulation ofDNp63or DGCR8. Proc Natl Acad Sci U S A 2014;111:E572–81.

3. Marchini S, Marabese M, Marrazzo E, Mariani P, Cattaneo D, Fossati R, et al.DeltaNp63 expression is associated with poor survival in ovarian cancer.Ann Oncol 2008;19:501–7.

4. Crook T, Nicholls JM, Brooks L, O'Nions J, Allday MJ. High level expression ofdeltaN-p63: a mechanism for the inactivation of p53 in undifferentiatednasopharyngeal carcinoma (NPC)? Oncogene 2000;19:3439–44.

5. Flores ER. The roles of p63 in cancer. Cell Cycle 2007;6:300–4.6. Choi W, Porten S, Kim S, Willis D, Plimack ER, Hoffman-Censits J, et al.

Identification of distinct basal and luminal subtypes of muscle-invasive bladdercancer with different sensitivities to frontline chemotherapy. Cancer Cell 2014;25:152–65.

7. Yang A, Kaghad M, Wang Y, Gillett E, Fleming MD, Dotsch V, et al. p63, a p53homolog at 3q27–29, encodes multiple products with transactivating, death-inducing, and dominant-negative activities. Mol Cell 1998;2:305–16.

8. Flores ER, Tsai KY, CrowleyD, Sengupta S, YangA,McKeon F, et al. p63 and p73are required for p53-dependent apoptosis in response to DNA damage. Nature2002;416:560–4.

9. NapoliM, VenkatanarayanA, Raulji P,Meyers BA, NortonW,Mangala LS, et al.DNp63/DGCR8-dependent microRNAs mediate therapeutic efficacy of HDACinhibitors in cancer. Cancer Cell 2016;29:874–88.

10. Venkatanarayan A, Raulji P, Norton W, Chakravarti D, Coarfa C, Su X, et al.IAPP-driven metabolic reprogramming induces regression of p53-deficienttumours in vivo. Nature 2015;517:626–30.

11. Barbieri CE, Tang LJ, Brown KA, Pietenpol JA. Loss of p63 leads to increased cellmigration and up-regulation of genes involved in invasion and metastasis.Cancer Res 2006;66:7589–97.

12. Olsen JR, Oyan AM, Rostad K, Hellem MR, Liu J, Li L, et al. p63 attenuatesepithelial to mesenchymal potential in an experimental prostate cell model.PLoS One 2013;8:e62547.

13. Zhao W, Wang H, Han X, Ma J, Zhou Y, Chen Z, et al. DNp63a attenuatestumor aggressiveness by suppressing miR-205/ZEB1-mediated epithelial-mesenchymal transition in cervical squamous cell carcinoma. Tumour Biol2016;37:10621–32.

14. Kanitakis J, Chouvet B. Expression of p63 in cutaneous metastases. Am J ClinPathol 2007;128:753–8.

15. Reis-Filho JS, Torio B, Albergaria A, Schmitt FC. p63 expression in normal skinand usual cutaneous carcinomas. J Cutan Pathol 2002;29:517–23.

16. Valastyan S, Weinberg RA. Tumor metastasis: molecular insights and evolvingparadigms. Cell 2011;147:275–92.

17. Noszczyk BH, Majewski ST. p63 expression during normal cutaneous woundhealing in humans. Plast Reconstr Surg 2001;108:1242–7.

18. Tucci P, AgostiniM, Grespi F,Markert EK, Terrinoni A, Vousden KH, et al. Lossof p63 and its microRNA-205 target results in enhanced cell migration andmetastasis in prostate cancer. Proc Natl Acad Sci U S A 2012;109:15312–7.

19. Tran MN, Choi W,Wszolek MF, Navai N, Lee IL, Nitti G, et al. The p63 proteinisoform DNp63a inhibits epithelial-mesenchymal transition in human bladdercancer cells: role of MIR-205. J Biol Chem 2013;288:3275–88.

20. Tsai JH, Donaher JL, Murphy DA, Chau S, Yang J. Spatiotemporal regulation ofepithelial-mesenchymal transition is essential for squamous cell carcinomametastasis. Cancer Cell 2012;22:725–36.

21. AbbasHA, BuiNHB, RajapaksheK,Wong J, Gunaratne P, Tsai KY, et al. DistinctTP63 isoform-driven transcriptional signatures predict tumor progression andclinical outcomes. Cancer Res 2018;78:451–62.

22. Santner SJ, Dawson PJ, Tait L, Soule HD, Eliason J, Mohamed AN, et al.Malignant MCF10CA1 cell lines derived from premalignant human breastepithelial MCF10AT cells. Breast Cancer Res Treat 2001;65:101–10.

23. Miller FR, Santner SJ, Tait L, Dawson PJ. MCF10DCIS.com xenograft model ofhuman comedo ductal carcinoma in situ. J Natl Cancer Inst 2000;92:1185–6.

24. Miller FR, Soule HD, Tait L, Pauley RJ, Wolman SR, Dawson PJ, et al. Xenograftmodel of progressive human proliferative breast disease. J Natl Cancer Inst 1993;85:1725–32.

25. Kang Y, Pantel K. Tumor cell dissemination: emerging biological insights fromanimal models and cancer patients. Cancer Cell 2013;23:573–81.

26. Mills AA, Zheng B, Wang XJ, Vogel H, Roop DR, Bradley A. p63 is a p53homologue required for limb and epidermal morphogenesis. Nature 1999;398:708–13.

27. Koster MI, Kim S, Mills AA, DeMayo FJ, Roop DR. p63 is the molecular switchfor initiation of an epithelial stratification program. Genes Dev 2004;18:126–31.

28. Carroll DK, Carroll JS, Leong CO, Cheng F, Brown M, Mills AA, et al. p63regulates an adhesion programme and cell survival in epithelial cells. Nat CellBiol 2006;8:551–61.

29. Lebrun JJ. The dual role of TGFb in human cancer: from tumor suppression tocancer metastasis. ISRN Mol Biol 2012;2012:381428.

30. Colak S, Ten Dijke P. Targeting TGF-b signaling in cancer. Trends Cancer 2017;3:56–71.

31. Giampieri S, Manning C, Hooper S, Jones L, Hill CS, Sahai E. Localized andreversible TGFbeta signalling switches breast cancer cells from cohesive to singlecell motility. Nat Cell Biol 2009;11:1287–96.

32. Ikushima H, Miyazono K. TGFbeta signalling: a complex web in cancerprogression. Nat Rev Cancer 2010;10:415–24.

33. Butz H, Racz K, Hunyady L, Patocs A. Crosstalk between TGF-b signaling andthe microRNA machinery. Trends Pharmacol Sci 2012;33:382–93.

34. Ballesteros A, Mentink-Kane MM, Warren J, Kaplan GG, Dveksler GS. Induc-tion and activation of latent transforming growth factor-beta1 are carried out bytwo distinct domains of pregnancy-specific glycoprotein 1 (PSG1). J Biol Chem2015;290:4422–31.

35. Campbell JD, Yau C, Bowlby R, Liu Y, Brennan K, Fan H, et al. Genomic,pathway network, and immunologic features distinguishing squamous carcino-mas. Cell Rep 2018;23:194–212.

36. Hoadley KA, Yau C, Wolf DM, Cherniack AD, Tamborero D, Ng S, et al.Multiplatform analysis of 12 cancer types reveals molecular classification withinand across tissues of origin. Cell 2014;158:929–44.

37. Gordian E, Welsh EA, Gimbrone N, Siegel EM, Shibata D, Creelan BC, et al.Transforming growth factor b-induced epithelial-to-mesenchymal signaturepredicts metastasis-free survival in non-small cell lung cancer. Oncotarget2019;10:810–24.

38. Cancer Genome Atlas Research Network. Comprehensive genomic character-ization of squamous cell lung cancers. Nature 2012;489:519–25.

39. Nielsen TO, Hsu FD, Jensen K, Cheang M, Karaca G, Hu Z, et al. Immunohis-tochemical and clinical characterization of the basal-like subtype of invasivebreast carcinoma. Clin Cancer Res 2004;10:5367–74.

40. Badowska-Kozakiewicz AM, Budzik MP. Immunohistochemical characteristicsof basal-like breast cancer. Contemp Oncol 2016;20:436–43.

41. Kuroda N, Ohara M, Inoue K, Mizuno K, Fujishima N, Hamaguchi N, et al. Themajority of triple-negative breast cancer may correspond to basal-like carcino-ma, but triple-negative breast cancer is not identical to basal-like carcinoma.Med Mol Morphol 2009;42:128–31.

42. Kamarlis RK, Lubis MN, Hernowo BS, Kar AS. Immunoexpression of P63and SOX2 in triple-negative breast cancers, Indonesia. F1000Res 2017;6:1780.

43. Lo Muzio L, Santarelli A, Caltabiano R, Rubini C, Pieramici T, Trevisiol L, et al.p63 overexpression associates with poor prognosis in head and neck squamouscell carcinoma. Hum Pathol 2005;36:187–94.

44. Urist MJ, Di Como CJ, Lu ML, Charytonowicz E, Verbel D, Crum CP, et al. Lossof p63 expression is associated with tumor progression in bladder cancer. Am JPathol 2002;161:1199–206.

45. Lambert AW, PattabiramanDR,WeinbergRA. Emerging biological principles ofmetastasis. Cell 2017;168:670–91.

46. Bergholz J, Zhang Y, Wu J, Meng L, Walsh EM, Rai A, et al. DNp63aregulates Erk signaling via MKP3 to inhibit cancer metastasis. Oncogene2014;33:212–24.

47. Di Franco S, Turdo A, Benfante A, Colorito ML, Gaggianesi M, Apuzzo T, et al.DNp63 drivesmetastasis in breast cancer cells via PI3K/CD44v6 axis. Oncotarget2016;7:54157–73.

48. Lu SL, Herrington H, Reh D, Weber S, Bornstein S, Wang D, et al. Loss oftransforming growth factor-beta type II receptor promotesmetastatic head-and-neck squamous cell carcinoma. Genes Dev 2006;20:1331–42.

49. Forrester E, Chytil A, Bierie B, Aakre M, Gorska AE, Sharif-Afshar AR, et al.Effect of conditional knockout of the type II TGF-beta receptor gene inmammary epithelia on mammary gland development and polyomavirusmiddle T antigen induced tumor formation and metastasis. Cancer Res2005;65:2296–302.

Bui et al.





50. Vasilaki E, Morikawa M, Koinuma D, Mizutani A, Hirano Y, Ehata S, et al. Rasand TGF-b signaling enhance cancer progression by promoting the DNp63transcriptional program. Sci Signal 2016;9:ra84.

51. Hu L, Liu J, Li Z, Wang C, Nawshad A. Transforming growth factor-b1 activatesDNp63/c-Myc to promote oral squamous cell carcinoma. Oral Surg Oral MedOral Pathol Oral Radiol 2016;122:460–82.

52. Su X, Chakravarti D, Cho MS, Liu L, Gi YJ, Lin YL, et al. TAp63 suppressesmetastasis through coordinate regulation of Dicer and miRNAs. Nature 2010;467:986–90.

53. Yi R, PoyMN, Stoffel M, Fuchs E. A skin microRNA promotes differentiation byrepressing ‘stemness’. Nature 2008;452:225–9.

54. He S, Zhang G, Dong H, Ma M, Sun Q. miR-203 facilitates tumor growth andmetastasis by targeting fibroblast growth factor 2 in breast cancer. Onco TargetsTher 2016;9:6203–10.

55. Zhao S, Han J, Zheng L, Yang Z, Zhao L, Lv Y. MicroRNA-203regulates growth and metastasis of breast cancer. Cell Physiol Biochem2015;37:35–42.

56. Xu D, Takeshita F, Hino Y, Fukunaga S, Kudo Y, Tamaki A, et al. miR-22represses cancer progression by inducing cellular senescence. J Cell Biol 2011;193:409–24.

57. Song SJ, Poliseno L, Song MS, Ala U, Webster K, Ng C, et al. MicroRNA-antagonism regulates breast cancer stemness and metastasis via TET-family-dependent chromatin remodeling. Cell 2013;154:311–24.

58. Fu J, Xu X, Kang L, Zhou L,Wang S, Lu J, et al. miR-30a suppresses breast cancercell proliferation and migration by targeting Eya2. Biochem Biophys ResCommun 2014;445:314–9.

59. Li L, Kang L, Zhao W, Feng Y, Liu W, Wang T, et al. miR-30a-5p suppressesbreast tumor growth and metastasis through inhibition of LDHA-mediatedWarburg effect. Cancer Lett 2017;400:89–98.

60. Stinson S, LacknerMR, Adai AT, YuN, KimHJ,O'Brien C, et al. TRPS1 targetingby miR-221/222 promotes the epithelial-to-mesenchymal transition in breastcancer. Sci Signal 2011;4:ra41.

61. Shah MY, Calin GA. MicroRNAs miR-221 and miR-222: a new level ofregulation in aggressive breast cancer. Genome Med 2011;3:56.






2020;80:2833-2847. Published OnlineFirst April 20, 2020.Cancer Res Ngoc H.B. Bui, Marco Napoli, Andrew John Davis, et al. Essential for Cancer Metastasis

-Regulated miRNAs IsβNp63 by TGF∆Spatiotemporal Regulation of

Updated version

10.1158/0008-5472.CAN-19-2733doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2020/04/22/0008-5472.CAN-19-2733.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/80/13/2833.full#ref-list-1

This article cites 61 articles, 13 of which you can access for free 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/80/13/2833To 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-19-2733

http://cancerres.aacrjournals.org/content/suppl/2020/04/22/0008-5472.CAN-19-2733.DC1

http://cancerres.aacrjournals.org/content/80/13/2833.full#ref-list-1

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

mailto:[email protected]

http://cancerres.aacrjournals.org/content/80/13/2833


spatiotemporalregulationof dnp63bytgfb-regulated mirnas is ... · cancer research | molecular cell...

Documents