∆ Np63 α inhibits Rac1 activation and cancer cell invasion through suppression of PREX1

Δ Np63 α , a member of the p53 family of transcription factors, plays a critical role in maintaining the proliferative potential of stem cells in the strati�ed epithelium. Although Δ Np63 α is considered an oncogene and is frequently overexpressed in carcinoma cells, loss of Δ Np63 α expression is associated with increased cancer invasion and metastasis. We recently identi�ed a Δ Np63 α /miR-320a/PKC γ signaling axis that regulates cancer cell invasion by inhibiting phosphorylation of the small GTPase Rac1, a master switch of cell motility that positively regulates cell invasion in multiple human cancers. In this study, we identi�ed a novel mechanism by which Δ Np63 α negatively regulates Rac1 activity, by inhibiting the expression of the Rac-specic Guanine Exchange Factor PREX1. Δ Np63 α silencing in multiple squamous cell carcinoma cell lines leads to increased Rac1 activation, which is abrogated by treatment with Rac1 inhibitor NSC23766. Furthermore, Δ Np63 α binds to the PREX1 promoter, leading to reduced PREX1 transcript and protein levels. Using a Rac-GEF activation assay, we also showed that Δ Np63 α reduces the levels of active PREX1. The inhibition of the PREX1-Rac1 signaling axis by Δ Np63 α leads to impaired cell invasion, thus establishing the functional relevance of this signaling axis. Our results elucidated a novel molecular mechanism by which Δ Np63 α negatively affects cancer cell invasion and identi�es the Δ Np63 α /Rac1 axis as a potential target for metastatic cancers.


Introduction
ΔNp63α, the major isoform of p63 expressed in epithelial tissues, plays a dual role in cancer, both as an oncogene and as a suppressor of tumor metastasis (1,2,3).While ΔNp63α is frequently overexpressed and promotes cell proliferation during the early stages of squamous cell carcinoma (SCC) (1,4), loss of ΔNp63α promotes cancer metastasis and is associated with poor prognosis (5,6,7,8).It is wellestablished that ΔNp63α reduces cancer cell invasiveness and prevents metastatic disease in diverse cancer cell types (5,6,8,9), however the precise mechanisms underlying these effects are poorly characterized.
The small GTPase Rac1 regulates multiple signaling pathways that control cytoskeleton organization, transcription, and cell proliferation.Deregulated Rac1 expression and/or activity is a common event in cancer, and has been associated with anchorage-independent growth, transformation, migration, and invasion (10,11).Rac1 is activated by guanine exchange factors (GEFs) that promote GDP/GTP exchange, and inactivated by GTPase activating proteins (GAPs).Rac1 hyperactivation in cancer cells is frequently caused by the oncogenic activation of upstream GEFs, or by the dysregulation of GEF expression or activity (12).The Rac-GEF family comprises more than 40 members, with most of them (> 30) belonging to the Dbl-like class, and a smaller subset corresponding to the DOCK family (13).As a large family of multidomain proteins, Rac-GEFs have distinctive regulatory modes and display cell-type speci c differences in expression.Rac-GEFs have been shown to regulate cell invasiveness and metastatic dissemination of cancer cells (13,14).Moreover, elevated expression of Rac-GEFs has been associated with poor patient outcome in several cancer types (15,16,17).Among the pro-metastatic Rac-GEFs, Phosphatidylinositol-3,4,5-Trisphosphate Dependent Rac Exchange Factor 1 (PREX1) has been found to be highly expressed in many types of tumors, including melanoma, breast, prostate cancer and others (18).Elevated PREX1expression has been linked to migratory and invasive phenotypes of cancer cells, and conversely, PREX1 knockdown suppresses cell migration and invasion (19,20,21,22,23).
PREX1 is synergistically activated by PIP3 (a lipid product of PI3K) and Gβγ subunits of heterotrimeric G proteins (18,24).The pathways that regulate PREX1 expression and function are poorly understood, and its role in SCC has not been yet identi ed.
Here, we demonstrate that ΔNp63α acts as a negative regulator of Rac1 activation by suppressing the expression of PREX1.Our data provide critical insight into the nature of ΔNp63α effectors and underscore a novel regulatory signaling pathway for Rac1-mediated cancer cell invasion.

ΔNp63α inhibits cancer cell invasion by impairing Rac1 activation
Next, we examined the effect of NSC23766, a small molecule inhibitor that inhibits the interaction between the Rac1 Switch II domain and Rac-speci c GEFs (27,28), on the Rac1 activation effect caused by ∆Np63α silencing.Notably, NSC23766 abrogated the elevation in Rac1-GTP levels resulting from ΔNp63α knockdown (Figs.2A and 2B).∆Np63α protein knockdown and the lack of an effect on total Rac1 levels was con rmed by Western blot as in Fig. 1.
In order to determine the functional consequences of ∆Np63α on Rac1 activity, we assessed cell invasion by Matrigel transwell assay in JHU-006 cells.ΔNp63α knockdown signi cantly increased cell invasion relative to non-treated control (NTC) cells.Conversely, NSC23766 treatment reduced cell invasion in both NTC and sip63 cells relative to corresponding vehicle treated controls (Figs.2C and 2D), suggesting ΔNp63α-dependent inhibition of Rac1-GTP levels leads to decreased cell invasiveness.

ΔNp63α knockdown negatively regulates PREX1 expression
Since Rac1 hyperactivation in cancer cells frequently correlates with elevated expression and/or activation of Rac1-GEFs (13), we speculated that ∆Np63α-mediated reduction of Rac1 activity could be associated with a reduction in the levels of upstream Rac-GEFs.Since the Rac-GEF family comprises > 40 members, we took advantage of a pre-designed Q-PCR array (29) to assess the effect of ΔNp63α knockdown on Rac-GEF expression in A431 cells.Two different ΔNp63α siRNA duplexes were used to minimize misinterpretation of results due to non-speci c effects of RNAi.Our analysis revealed that several DOCK GEFs and Dbl-like Rac-GEFs were up-regulated by ΔNp63α silencing with both siRNA duplexes (Supplemental Fig. 1).Among the up-regulated GEFs, PREX1 was chosen for subsequent analysis due to of its known pro-invasive role in cancer (21,22).The effect of ΔNp63α knockdown on PREX1 expression was examined in A431, JHU006, JHU029 and FaDu SCC cell lines.Silencing ΔNp63α led to a signi cant up-regulation of PREX1 mRNA (Fig. 3A, upper panel) and increased protein levels (Fig. 3A, lower panel), as determined by qRT-PCR and Western blot analysis, respectively.These results were also recapitulated in HaCaT cells, a non-tumorigenic human keratinocyte cell line that exhibits measurable levels of PREX1 mRNA (30) (Supplemental Figs.2A and B).Taken together, these results indicate that ΔNp63α negatively regulates PREX1 transcript and protein levels in SCC.
To determine whether ΔNp63α directly represses PREX1 gene transcription, a 2 kB fragment of the PREX1 promoter (-2,047 to -23 from the transcriptional start site) was cloned upstream of a luciferase reporter gene in the pGL3-Basic vector.The resulting pGL3-PREX1-Luc reporter plasmid (23) (PREX1-Luc) was cotransfected into p63 null H1299 cells with a Renilla luciferase plasmid (for normalization), along with either empty vector (control) or increasing concentrations of a ΔNp63α expression plasmid.Increasing concentrations of ΔNp63α, as con rmed by immunoblot, led to a dose-dependent reduction in luciferase reporter activity (Fig. 3B).
The p63 ChIP-seq database GSE59827 (31,32) was used to identify putative p63 binding sites in the PREX1 promoter.This analysis predicted a p63 binding site located at chr20:48828385-48828404 with the sequence 5'-GCGCAGGCTCCTGCTTGCAG-3'.Next, a 229 bp fragment of the PREX1 promoter containing the putative p63 binding site was cloned upstream of the luciferase reporter gene in pGL3 to generate the ΔPREX1-Luc reporter plasmid.ΔPREX1-Luc was co-transfected into p63 null H1299 cells together with either empty vector (control) or increasing concentrations of the expression plasmid encoding ΔNp63α.Again, as observed with the full-length PREX1 promoter, dose-dependent expression of ΔNp63α caused a signi cant decrease in luciferase reporter activity (Fig. 3C).Taken together, these results suggest that ΔNp63α likely inhibits PREX1 expression via direct transcriptional repression.

ΔNp63α knockdown increases Rac1 binding to activated PREX1
To determine whether modulation of ΔNp63α expression leads to changes in PREX1 activity, we used a pull-down assay for activated PREX1.We took advantage of the G15A-Rac1 "nucleotide-free" mutant, which binds poorly to GDP and GTP, and thus mimics the intermediate state that binds to active Rac-GEFs with high a nity (33).PREX1 was immunoprecipitated from whole cell lysates of JHU-006 cells transiently transfected with control siRNA or siRNA for ΔNp63α, and then incubated with either GST-WT-Rac1 or GST-G15A-Rac1 fusion proteins.This assay revealed that ∆Np63α knockdown increased binding of PREX1 to WT Rac1 relative to control cells (Fig. 4A, lane 2vs.lane 1, Fig. 4B), and that the effect was much stronger when pull-down was done with the G15A-Rac1 mutant (Fig. 4A, lane 4vs.lane 3, and Fig. 4B).These results indicate that ΔNp63α knockdown not only increases the abundance of PREX1 as a consequence of its inhibition of PREX1 transcription, but also results in elevated levels of activated PREX1.

Knockdown of PREX1 reduces cell invasion in ΔNp63α knockdown cells
Since PREX1 levels positively correlate with cell migration and invasion (18,19,20,21,22,34), we next determined whether ΔNp63α-dependent repression of PREX1 inhibits cell invasion.A Matrigel-based invasion assay was performed using JHU-006 cells transiently transfected with ΔNp63α and/or PREX1 siRNA.As expected, ΔNp63α knockdown increased cell invasion relative to the nontargeted control.Interestingly, while PREX1 RNAi had no signi cant effect on cell invasiveness (Fig. 5C and 5D), ΔNp63α and PREX1 double knockdown cells were signi cantly less invasive than ΔNp63α knockdown cells, indicating that knockdown of PREX1 reversed the effect of ΔNp63α knockdown on cell invasion.Altogether, these data demonstrate that ΔNp63α inhibits cell invasion by negatively regulating expression of PREX1.

Discussion
ΔNp63α plays a dual role in cancer, acting as a promoter of tumor initiation and a suppressor of tumor metastasis (1,2,3).As a potent oncogenic protein in SCC, ΔNp63α promotes cell survival and angiogenesis, suppresses apoptosis, and associates with poor prognosis (35,36,37,38).Nevertheless, loss of ΔNp63α correlates with increased invasiveness and metastasis (5,6,8).Moreover, decreased ΔNp63α expression at advanced cancer stages upregulates epithelial-to-mesenchymal transition (EMT) genes in cell culture and highly metastatic tumors in mice (6, 9, 39).The precise mechanisms by which ΔNp63α inhibits cancer cell invasion have been poorly characterized.We recently showed that ΔNp63α reduces Rac1 signaling by inhibiting protein kinase C γ (PKCγ), which in turn inhibits cancer cell invasion (26).In the present study, we show that silencing ΔNp63α PREX1 in SCC cells leads to elevated Rac1 activity and upregulation of the Rac-speci c GEF.To our knowledge, this is the rst study that implicates decreased activation of Rac1 in the anti-invasive role of ΔNp63α.This observation, together with the evidence that ΔNp63α inhibits cancer cell migration and invasion, provides critical mechanistic insight into the inhibitory role of p63 in SCC metastasis.
The small GTPase Rac1 dynamically regulates cytoskeletal organization, and therefore acts as a major regulator of cell morphology, adhesion and migration.Rac1 hyperactivation is a hallmark of a variety of cancers, contributing to enhanced cancer cell migration, invasion, and metastasis (11,12).As such, Rac1 represents a potential therapeutic target for cancer metastasis.The mechanisms leading to elevated Rac1 activity in metastatic cancer are not fully understood, and to date, efforts to exploit Rac1-GTP as a therapeutic target in cancer have not been comprehensively examined.Our ndings indicate that loss of ΔNp63α is permissive for Rac1-dependent invasion in SCC, and this ts with the loss of ΔNp63α and advanced metastatic disease (6, 9, 39).To our knowledge, this is the rst study to suggest that reduced activation of Rac1 mediates the anti-invasive effect of ΔNp63α.
Hyperactivation and/or overexpression of Rac-GEFs has been linked to aberrant activation of Rac1 in cancer (13).However, due to the large complexity and context-speci c Rac-GEF expression and regulation, there is a pressing need to dissect upstream regulatory events as well as their expression control mechanisms in speci c cancers.Our unbiased analysis identi ed PREX1 as a Rac1-GEF that is upregulated upon ΔNp63α knockdown in SCC cells.PREX1 is primarily expressed in hematopoietic cells, neurons, and endothelial cells, and its expression is low in most normal epithelial cells (18, 23,40).Previous studies revealed that PREX1 is prominently up-regulated in multiple human cancers.Moreover, PREX1 has been causally linked to increased cancer metastasis and tumorigenesis (18).PREX1 is highly expressed in oral SCC (OSCC), and associates with metastatic disease and poor prognosis.Ampli cation or epigenetic dysregulation of PREX1 may contribute to its overexpression in some human cancers (23,41,42).Our study provides evidence for an additional mechanism of PREX1 expression control that involves transcriptional regulation.We found that ΔNp63α suppresses the expression of PREX1 in cutaneous SCC cells (A431) as well as in JHU-006, JHU-029, and FaDu HNSCC cells, an effect that was observed both at mRNA and protein levels.The mechanism underlying ΔNp63α-mediated suppression of P-Rex1 was investigated using a PREX1 promoter-driven luciferase reporter, whose activity inversely correlated with ΔNp63α abundance.We identi ed a putative p63 binding site in the PREX1 promoter that is essential for its regulation by ΔNp63α.These results demonstrate that ΔNp63α plays a critical role in regulating PREX1 in SCC, likely involving a transcriptional mechanism.Whether non-transcriptional mechanisms are also involved in the control of PREX1 expression in SCC would need to be explored.
Silencing endogenous PREX1 reduces Rac1 activity and cancer cell invasion in SCC cells.In mouse models, deletion of the PREX1 gene leads to reduced Rac1 activity and metastasis in melanoma (20).It has been well established that PREX1-mediated activation of Rac1 promotes membrane ru ing and lamellipodia formation, which contributes to cell migration (22).We showed that ΔNp63α inhibits PREX1-mediated activation of Rac1 by reducing expression of PREX1, and that PREX1 silencing abrogates the elevated SCC cell invasion observed as a consequence of ΔNp63α knockdown.Thus, loss of PREX1 represents a critical event for mediating the effect of ΔNp63α on Rac1 activation.We postulate that ΔNp63α signaling is a promising therapeutic target for reducing the expression or activation of Rac-GEF/Rac1 metastatic signaling in cancer.
In summary, our study identi ed a novel mechanism by which ΔNp63α inhibits Rac1 activity and cancer cell invasion.ΔNp63α controls PREX1 promoter transcriptional activity and downregulates its expression, which in turn inhibits activation of Rac1 and reduces cancer cell invasion.These results may facilitate development of novel effective therapeutic approaches for metastatic cancer.

Declarations Con ict of Interest
The authors do not have any competing nancial interests related to the described work.

Data Availability
The data used to support the ndings of this study are available from the corresponding author upon request.

PREX1 promoter cloning and luciferase reporter assay
The full-length 2024 bp PREX1 promoter (23) was ampli ed by PCR using the forward primer 5′-CGGACTCGAGGCTCTCACAAAGACTCCCCTTTT-3′ and the reverse primer 5′-CCCAAGCTTGCTCCTTCCGTCGCGCCGAG-3′ and the PCR product was inserted into the XhoI and HindIII cloning sites in the pGL3-basic vector, which carries the luciferase reporter gene (pGL3-basic-Luc, Promega, Madison, WI, USA).A 229 bp fragment of the PREX1 promoter containing a putative p63 binding site (chr20:48828385-48828404) (44,45) was PCR-ampli ed using forward primer 5′-CGGACTCGAGCTCCGCAGCGAGCTTTCCCAGCCCC-3′ and the reverse primer 5′-CCCAAGCTTCCGGAAGGGCCCCGCGGAGCC-3 and the PCR product was inserted into the XhoI and HindIII cloning sites in pGL3-basic-Luc.H1299 cells were plated on 12-well plates and co-transfected with empty vector (EV) or expression plasmid encoding ΔNp63α and PREX1 or ΔPREX1-luciferase reporter plasmids.Cells were transfected with a Renilla luciferase expression plasmid to estimate transfection e ciency and normalize experimental data.Twenty-four hours after transfection, cells were harvested in passive lysis buffer, and cell lysates were used for dual luciferase assay, according to the manufacturer's instructions (Promega, Madison, WI).Relative luciferase unit (RLU) values were calculated from the ratio of Fire y luciferase activity to Renilla luciferase activity and normalized to luciferase activity in cells co-transfected with empty pLG3 promoter or basic vector plasmid and increasing concentrations of the ΔNp63α expression plasmid.

Cell invasion assay
Cell invasion was assessed using a two-chamber transwell system.A total of 5.0 × 10 4 transiently transfected JHU-006 cells were suspended in serum-free medium, seeded into 8 µm pore size inserts (BD Biosciences) coated with 1 mg/mL Matrigel (BD Biosciences), and placed into the well of a 24-well plate.Then, RPMI containing 10% FBS was added to the bottom of each insert and incubated for 21 h.Cells that did not invade were removed using a cotton swab.Invading cells attached to the bottom of the transwell were xed with 4% of paraformaldehyde and washed once with Dulbeco's PBS.Cells were stained with 0.1% crystal violet and imaged in 4-6 random elds at 100× magni cation using a Leica CTR 6000 Microscope (Leica Microsystems, Wetzlar, Germany) and ImagePro 6.2 software (Media Cybernetics, Bethesda, MD).Cells were counted manually and the average number of cells per eld was calculated.
isolation and TaqMan qRT-PCR Total RNA was extracted using the EZNA RNA isolation kit according to the manufacturer's instructions (Omega Bio-Tek, Norcross, GA, USA).Quantitative RT-PCR was carried out as previously described using the Applied Biosystem 7900HT or QuantStudio 7 Flex Real-Time PCR Systems and Assays on Demand™ (AOD) for GAPDH (4325792) and PREX1 (Hs01031507_m1).Data were normalized to endogenous GAPDH (Life Technologies, Carlsbad City, CA, USA) (46, 47).qRT-PCR reactions were run in triplicate.Data were analyzed using the 2 −ΔΔCT method (48) and statistical signi cance was analyzed using two-tailed Student's unpaired t-test.

Rac1-GTP pull-down assay
Rac1-GTP was quanti ed using a Rac1 pull-down activation assay kit (Cytoskeleton, BK035, Denver, CO), in which PAK-PBD is fused to GST and glutathione a nity beads are used to isolate and quantify PAK-PBD-bound proteins.Assays were performed 48 h after transfection with siRNA or 24 h after transfection with overexpression plasmid DNA.Brie y, cells were lysed in ice-cold lysis buffer (50 mM Tris pH 7.5, 10 mM MgCL2, 0.5 M NaCl, and 2% Igepal), containing 1× protease inhibitor cocktail.Cell lysates were immediately clari ed by centrifugation at 10,000 × g for 1 min at 4°C.Protein concentrations were determined by BCA assay (Thermo Fisher Scienti c Inc., Fremont, CA, USA).Equivalent concentrations of protein (300-500 µg) were added to 10 µL PAK-PBD beads and rotated at 4°C on a tube rotator for 1 h.The PAK-PBD beads were pelleted by centrifugation at 5,000 × g at 4°C for 3 min, washed twice with 500 µL wash buffer, and resuspended in 20 µL wash buffer.Rac1-GTP-bound PAK-PBD beads were loaded on a 10% SDS-PAGE gradient gel, followed by immunoblot analysis using mouse monoclonal anti-Rac1 (23A8) to detect total Rac1 and rabbit polyclonal anti-phospho-Rac1 (Ser71) to detect pRac1.
GST-Rac1 fusion beads and immunoprecipitation assay expressing recombinant (WT) and G15A were obtained from Dr. Garcia-Mata (University of Toledo).Recombinant proteins were expressed and puri ed as previously described (33).Recombinant proteins were incubated with 500 µL pre-equilibrated Glutathione-Sepharose 4B beads (Sigma-Aldrich) at 4°C for 1 h.After incubation, the samples were centrifuged, pelleted beads were washed with once with lysis buffer and twice with 20 mM HEPES pH 7.5, 150 mM NaCl (HBS) containing 5 mM MgCl 2, and 1 mM DTT. Washed beads were resuspended in an aliquot of the supernatant, 250 µL glycerol was added, and the resuspended beads were diluted in HBS containing 5 mM MgCl 2 , 1 mM DTT, and glycerol to a nal total protein concentration of 1-3 mg/mL.For immunoprecipitation, cells were lysed in NP-40 lysis buffer (50 mM Tris-HCl pH 8, 150 mM NaCl, 5 mM EDTA pH 8, 1% NP-40, 2 mM DTT, and 2 mM PMSF) containing protease inhibitor cocktail (Sigma-Aldrich).Whole-cell extracts (1 mg) were added to 20 µg WT or G15A GST-Rac1 beads and rotated at 4°C for 1 h.After centrifugation, pelleted beads were washed three times with wash buffer (0.05% in 1× PBS) and in 20 µL wash buffer.Immunoprecipitated protein complexes were resolved on a 7.5% SDS-PAGE gradient gel followed by immunoblot analysis.

Statistical analysis
Data are presented as mean ± 1 standard deviation (SD) or Standard Error of the Mean (SEM).Statistical signi cance between groups was determined using Student's unpaired t-test.Statistically signi cant values (P ≤ 0.05) are indicated with an asterisk or the # symbol.

Figures
Figure 1 ΔDNp63anegatively regulates Rac1-GTP in SCC cells.(A) JHU-006, A431 and JHU-029 cells were transfected with nontargeting control (NTC) siRNA or p63-targeted siRNA (sip63).(B) JHU-006 cells were transfected with empty vector (EV) or ΔNp63α-expression plasmid DNA.Whole-cell lysates of transfected cells were immunoprecipitated with p21-activated kinase (PAK) protein binding domain (PBD) followed by immunoblot with Rac1 antibody to detect Rac1-GTP, as described in Materials and Methods.Immunoblot analysis with the indicated antibodies is shown in the bottom panels.b-actin was used as a control to normalize for differences in total protein per lane.The relative abundance of Rac1-GTP was calculated by normalizing to total Rac1 and β-actin signal in the corresponding NTC (A) or EV control (B), and the resulting fold-changes are shown in the bar plots (top).Data are presented as mean ±1 S.E.M. Asterisks indicate P ≤ 0.05 relative to corresponding NTC.