TAK1 signaling regulates p53 through a mechanism involving ribosomal stress

Triple-negative breast cancer (TNBC) is among the most aggressive forms of breast cancer with limited therapeutic options. TAK1 is implicated in aggressive behavior of TNBC, while means are not fully understood. Here, we report that pharmacological blockade of TAK1 signaling hampered ribosome biogenesis (RBG) by reducing expression of RBG regulators such as RRS1, while not changing expression of ribosomal core proteins. Notably, TAK1 blockade upregulated expression of p53 target genes in cell lines carrying wild type (wt) TP53 but not in p53-mutant cells, suggesting involvement of ribosomal stress in the response. Accordingly, p53 activation by blockade of TAK1 was prevented by depletion of ribosomal protein RPL11. Further, siRNA-mediated depletion of TAK1 or RELA resulted in RPL11-dependent activation of p53 signaling. Knockdown of RRS1 was sufficient to disrupt nucleolar structures and resulted in activation of p53. TCGA data showed that TNBCs express high levels of RBG regulators, and elevated RRS1 levels correlate with unfavorable prognosis. Cytotoxicity data showed that TNBC cell lines are more sensitive to TAK1 inhibitor compared to luminal and HER2+ cell lines. These results show that TAK1 regulates p53 activation by controlling RBG factors, and the TAK1-ribosome axis is a potential therapeutic target in TNBC.


Results
Inhibition of TAK1 up-regulates the p53-p21 signaling axis. TAK1 signaling in breast cancer cells contributes to tumor invasion, angiogenesis, and metastases, and TAK1 inactivation effectively blocks these responses 8,11,22,33 . To define the molecular targets of TAK1 in tumor cells, we thought to compare gene expression profiles in breast carcinoma (MDA-MB-231 and MCF7) and non-tumor MCF10A cell lines upon treatment with a selective TAK1 inhibitor 5Z-7-oxozeaenol 23 . We found that TAK1-i effectively blocked TNF-induced activation of NFκB signaling measured by phosphorylation of RELA and protein levels of a negative regulator IκB-alpha Immunoblotting of whole-cell lysates from breast cancer MDA-MB-231 and MCF7 cell lines treated with 10 ng/ml TNFα -/ + 5 µM 5Z-7-oxozeaenol. Graphs (panels A, E, F) show fold changes relative to control in the GAPDH-normalized band density. (B) Treatment with TAK1-i decreases mRNA levels of genes responsive to TAK1 signaling. Expression profiles were measured using microarrays and two independent preparations of total RNA from cells treated with TAK1-i for 6 hours. (C) TAK1-i increases expression of p53 target genes in MCF7 and MCF10A cell lines carrying p53 wt but not in p53mutant MDA-MB-231. Microrray data for two independent preparations of total RNA from cells treated with TAK1-i for 6 hours. (D) Quantitative RT-PCR of p21 and MDM2 relative to GAPDH. Experiments were done in triplicates and repeated at least two times, **, P < 0.01. (E,F) Immunoblots of whole-cell lysates from cells treated with TAK1-i or TNFα for 24 hours, unless indicated otherwise. (Fig. 1A). Expression profiles were measured at 6 hours of treatment with TAK1-i in order to limit the effects associated with cell cycle arrest in G1-phase induced by 24-hour incubation ( Supplementary Fig. 1). As expected, TAK1-i markedly reduced mRNA levels of known target genes for TAK1-NFκB signaling such as IL8, BCL3, JUN, PLAU, and Cyclin D1 in all tested cell lines (Fig. 1B, Supplementary Table 1). In two ER-negative cell lines (MCF10A and MDA-MB-231), TAK1-i decreased expression of about 15% of genes involved in TAK1-NFκB responses, including chemokines CXCL1, CXCL2 and IL8/CXCL8, while their levels were lower in ER-positive MCF7 cell line (Supplementary Table 1). Notably, about 15% of down-regulated genes in all three cell lines were associated with ribosomal RNA (rRNA) processing and ribosome assembly while ribosomal proteins were not regulated, suggesting involvement of TAK1 in ribosome biogenesis (Supplementary Table 1). The gene compositions in TAK1-i-induced groups were markedly dissimilar in cells carrying mutant and wt p53. Known p53 target genes 34 constituted roughly 25% of genes commonly induced by TAK1-i in MCF7 and MCF10A cell lines carrying p53-wt, whereas these genes were not regulated by TAK1-i in MDA-MB-231 carrying mutant p53-R280K ( Fig. 1C; Supplementary Table 1). TAK1 blockade induced well-known p53 targets such as p21/CDKN1A, MDM2, SESN1, and BTG2 in p53-wt cell lines without affecting p53 mRNA levels, while their basal levels were lower in p53-mutant cells (Fig. 1C). These findings were further validated by quantitative (q) RT-PCR for MDM2 and p21/CDKN1A (Fig. 1D).
Next, we assessed protein levels of p53 and p21 in cell lines with wt (MCF7, MCF10A and WI-38), and mutant p53, MDA-MB-231 (Fig. 1E,F). These cell lines exhibit p53 and TAK signaling responses 8,35 . We found that TAK1-i increased p53 protein levels in all tested cell lines, whereas up-regulation of p21 was observed only in p53-wt cells (MCF7, MCF10A, and WI-38), indicating activation of p53-p21 signaling by TAK1-i (Fig. 1E,F, and their quantification shown in bar graphs). Together these data demonstrate that TAK1 inhibition decreases TAK1-NF-κB signaling while activating p53. Notably, our data indicate that blockade of TAK1 decreases expression of ribosome biogenesis regulators.
The contribution of signaling pathways in the TAK1-p53 link. TAK1 activates MAP kinases (ERK, p38) and NFκB signaling while cytokine-mediated activation of TAK1 depends on cIAP-mediated K63-ubiquitination, see above and 17 . To define a signaling mechanism leading to activation of p53 in response to TAK1 blockade, we examined responses to selective inhibitors of the aforementioned signaling pathways (see Supplementary Fig. 2) in cell line models with wild-type p53: epithelial MCF7 and A549 cell line, and mesenchymal WI-38 line. We found that increasing concentrations of TAK1-i 5Z-7-Oxozeaenol decreased phosphorylation of p65 while inducing p53 and p21 protein levels in A549 and MCF7 cell lines ( Supplementary Fig. 3A). Then, we examined the response to a structurally-unrelated TAK1-IKK inhibitor, CAY10657 36 ( Supplementary Fig. 2). Treatment with TAK1-i CAY10657 increased p53 and p21 levels while reducing phospho-p65 ( Supplementary  Fig. 3B). Next, we evaluated the response to agents blocking the activity of TAK1 downstream signaling mediators. Treatment with a selective IKKβ inhibitor BMS-345541 37 ( Supplementary Fig. 2) induced p53 levels comparable to TAK1 inhibitors in all tested p53-wt lines ( Fig. 2A-C). All three tested inhibitors of TAK1-IKK signaling reduced phosphorylation of p65 (Supplementary Figs. 3 and 4A,B). In contrast, a selective MEK inhibitor (U0126) or cIAP inhibitor/Smac-mimetic TL-32711 38 did not induce p53 protein in MCF7 and A549 cell lines ( Fig. 2A,B) although both inhibitors effectively suppressed the cognate targets ( Supplementary Fig. 4C,D). In addition, we found that treatment with TNFα had no significant effect on the p53-p21 axis (Fig. 1F). Notably, induction of p53 by TAK1 blockade in MCF7 and A549 cell lines was temporal and reduced by 24 hours. This effect may associate with a negative feedback mechanism by the p53-MDM2 axis 39 . To address this possibility, we examined the response to TAK1 inhibitor alone and in combination with Nutlin 3A that disrupts the p53-MDM2 interaction 40 . We found that TAK1 blockade induced the p53-MDM2 axis at 6 hours and this response declined by 24 hours while p53, MDM2 and p21 protein levels remained high for 24 hours in cells treated with Nutlin 3A ( Supplementary Fig. 5). Co-treatment with TAK1-i and Nutlin 3A resulted in stabilization of p53 and p21 protein levels at 24 hours in two tested cell lines ( Supplementary Fig. 5B,C).
Then, we explored regulation of p53 signaling by TAK1-IKK inhibitors in human mesenchymal WI-38 cell line carrying wild-type p53 (Fig. 2C). Treatment with TAK1 or IKK inhibitors induced levels of p53 protein and Ser15 phosphorylation (Fig. 2C). Phosphorylation of p53 at Ser15 stimulates transcriptional activity of p53 target genes such as p21 41 . Accordingly, TAK1-IKK blockade increased p21 protein levels (Fig. 2C). To verify the role of p53 in regulation of p21 by TAK1-i, we tested the response in isogenic colon carcinoma HCT116 cell lines with p53-wt and p53-deletion. The data showed that TAK1-IKK inhibitors up-regulated p53 and p21 protein levels, while the latter response was strictly dependent on p53 (Fig. 2D). Thus, these findings showed that inhibition of TAK1-IKK signaling stabilizes p53 protein and increases expression of p53-target genes, whereas MEK-ERK signaling and activity of cIAPs are not required for this response. TAK1 signaling inhibitors induce nucleolar stress. Stabilization and activation of p53 can be stimulated by various cellular stresses including DNA damage and ribosome biogenesis stress 31 . Examination of phospho-Ser139-H2AX (γH2AX), a DNA damage marker, showed no induction of DNA damage within 6 hours of TAK1-i treatment. Given down-regulation of ribosome biogenesis genes by TAK1-i, we examined whether TAK1 blockade affects the structure of nucleolus, a site of ribosome biogenesis. Disruption of the nucleolar architecture indicates an interruption in ribosome biogenesis that leads to stabilization of p53 42 . Phase microscopy revealed alterations in the nucleolar structures upon treatment with TAK1-i in both wt-p53 and mut-p53 cell lines (Fig. 3A). To validate disruption of the nucleolus, the cells were stained for fibrillarin, a nucleolar protein required for the organization and processing of rRNA 43 . Fluorescence microscopy showed that the nucleus of control-untreated cells contain 3-5 fibrillarin-stained speckles, whereas the organization of the fibrillarin speckles was altered in TAK1-i treated cells (Fig. 3B). The observed changes in the fibrillarin speckles were similar to www.nature.com/scientificreports www.nature.com/scientificreports/ those produced by 5-fluorouracil (5-FU) which is known to induce the nucleolar stress and ribosome biogenesis interruption by blocking rRNA processing 42 . Thus, these data suggest that inhibition of TAK1 signaling disrupts the nucleolar architecture and this may activate a p53-dependent response.
Next, we examined whether TAK1 blockade disrupts ribosomal RNA synthesis in comparison to a potent RNA polymerase I inhibitor CX5461 44 , and Nutlin 3A, a known activator of p53 that stabilizes p53 through a direct disruption of the p53-MDM2 interactions 40 . First, testing the effects on the p53-p21 axis, we found that treatment with increasing amounts of CX5461 resulted in upregulation of p53 and p21 protein levels (Fig. 3C). These effects were comparable with those of TAK1-i 5Z-7-oxozeaenol obtained from two different sources, Ox1 and Ox2 (Fig. 3C). As expected, Nutlin 3A induced robust up-regulation of the p53-p21 axis compared to CX5461 and TAK1-i (Fig. 3C). Similar results were obtained for MCF10A cells ( Supplementary Fig. 6). Next, rRNA synthesis was measured by assessing levels of unprocessed pre-rRNA that contains internal transcribed spacer 1 (ITS1), which is rapidly removed following pre-rRNA synthesis and its presence indicates newly synthesized rRNA [45][46][47] . Quantitative RT-PCR (qRT-PCR) data revealed that TAK1-i moderately reduces pre-rRNA ITS1 levels in MCF7 cells, while CX5461 showed a more pronounced reduction in pre-RNA levels (Fig. 3D). Note, CX5461 at the concentrations ≥1 μM may induce DNA damage in some cell lines 44 .
To validate the qRT-PCR data, we evaluated de novo rRNA synthesis by measuring incorporation of 5-ethynyl uridine (EU) into newly synthesized RNA 48 . RNA-incorporated EU was labeled with Cy3 fluorophore using click-it chemistry 49 . Fluorescence microscopy showed the presence of bright EU-containing speckles (nucleoli) within the nucleus of control cells (Fig. 3E). As expected, CX-5461 markedly reduced the number and intensity of these speckles (Fig. 3E,F). TAK1 inhibitor also reduced the EU speckle intensities (Fig. 3E,F). Nutlin 3A did not affect the EU speckle parameters, suggesting that activation of p53 alone is not sufficient to impact rRNA synthesis. Together, these data indicate that TAK1 blockade leads to disruption of ribosome biogenesis, as indicated by changes in the nucleolus architecture and rRNA synthesis. Nucleolar stress and induction of p53 in response to TAK1-IKK inhibition. Next, we examined whether activation of p53 by TAK1 blockade is associated with disruption of ribosome biogenesis. Prior studies reveal that interruption in ribosome biogenesis leads to accumulation in the nucleoplasm of ribosomal protein RPL11 as a part of the 5 S rRNP complex, which inhibits MDM2 and stabilizes p53 35,50 . Further, depletion of RPL11 is sufficient to suppress up-regulation of p53 in response to the ribosome biogenesis (nucleolar) stress 35,50 . To assess the role of this mechanism in the TAK1-dependent activation of p53, cells were transduced with siRNA to RPL11 followed by treatment with TAK1 inhibitors. Blockade of TAK1 or IKK increased p53 and p21 protein levels in MCF7 cells transfected with scramble-control siRNA, whereas this response was greatly reduced in RPL11-depleted cells (Fig. 4A). As expected, Nutlin 3A, a MDM2 inhibitor, induced p53 and p21 proteins independently of RPL11 (Fig. 4A). Depletion of RPL11 in MCF10A or A549 cells also diminished activation of p53 by TAK1 or IKK inhibitors (Fig. 4B,C). These data implicate ribosome stress in the activation of p53 by TAK1 blockade.
To validate the results obtained with pharmacological probes, we assessed regulation of p53 in response to siRNA-mediated depletion of TAK1 and RELA. First, depletion of TAK1 in A549 cells ( Supplementary Fig. 7A) resulted in up-regulation of p53 and p21 compared to scramble-control while depletion of RPL11 blocked this response (Fig. 4D). Next, we found that depletion of RELA/p65 by increasing amounts of siRNA resulted in upregulation of p53 and p21 protein levels (Fig. 4E), while depletion of RPL11 reduced siRELA-induced activation of p53 (Fig. 4F). Similar results were obtained in MCF7 and HCT116 cell lines (Fig. 4G). These findings were also verified in non-tumor MCF10A cell line ( Supplementary Fig. 7B). The data show that blockade of the TAK1-RELA axis activates p53 and ribosome stress contributes to this response.

TAK1 signaling regulates expression of ribosome biogenesis genes.
To better define the mechanism of TAK1-dependent ribosomal stress, we analyzed changes in gene expression after treatment with TAK1-i for 6 hours in three cell lines described in Fig. 1. We found that mRNA levels of cytosolic and mitochondrial ribosomal proteins, including RPL11, RPS7, and MRPL4, were not affected by TAK1 blockade (Fig. 5A). In contrast, mRNA levels of ribosome biogenesis factors (RBFs) regulating rRNA processing and ribosome assembly such as RRS1, PPAN, URB2, and PUS1 were greatly reduced (Fig. 5A). In comparison, levels of p53 targets p21/CDKN1A and BTG2 were up-regulated (Fig. 5A, bottom rows). Further, RT-PCR assays confirmed down-regulation of RRS1 and URB2 by TAK1-i (Fig. 5B). In MCF7 cells, inhibitors of TAK1-i (Oxo) or IKKβ (BMS) reduced RRS1 www.nature.com/scientificreports www.nature.com/scientificreports/ and up-regulated MDM2 protein levels (Fig. 5C). Next, we examined whether depletion of RRS1, which controls 60 S ribosome assembly and trafficking 51,52 , affects the p53-p21 axis. Depletion of RRS1 in A549 or MCF7 cells reduced RRS1 protein levels and upregulated expression of p21 (Fig. 5D). Phase contrast microscopy showed disruption of the nucleoli in RRS1-depleted cells similar to those observed in cells after depletion of RELA (Fig. 5E). Together, these data indicate that blockade of TAK1 reduces expression of RBFs leading to disruption of ribosomal biogenesis and activation of p53.

TAK1 signaling and ribosome biogenesis genes in breast cancer. To gain insights in clinical signif-
icance of our findings, we assessed expression of TAK1 targets and ribosome biogenesis genes in breast cancers using TCGA datasets. Analysis of the TCGA data for breast cancers 53 revealed that basal-like breast cancers (BLBCs) express high mRNA levels of TAK1 targets (Fig. 6A, gene group B). Majority of TNBC cases exhibit BLBC molecular phenotype 2,3 with elevated levels of EGFR receptor and deficiency in tumor suppressors p53, RB1, and PTEN (gene group A). Importantly, basal cancers express high mRNA levels of genes encoding regulators of ribosome biogenesis such as BYSL, RRS1, RIOK1, POLR1C, and POLR1E (Fig. 6A, group C). In contrast, mRNA levels of ribosomal proteins were not elevated in basal cancers (Fig. 6A, group E). Similar results were obtained using MetaBric dataset 54 for 1,951 breast cancer patients (Supplementary Fig. 8). Evaluation of relapse-free survival (RFS) using the Kaplan-Meier plotter (www.kmplot.com) 55 showed that high mRNA level of RRS1 is associated with poor prognosis for TNBC patients (Fig. 6B). This association was further confirmed in two TCGA datasets (Supplementary Fig. 9). Finally, we analyzed a public database for the sensitivity to TAK1-inhibitor of breast cancer cell lines 56 . The data showed that cell lines derived from TNBCs exhibited higher sensitivity (lower IC50) compared to cell lines derived from HER2 + or luminal ER-positive cancers (Fig. 6C, and Supplementary Fig. 10). Whole-cell lysates were probed with antibodies to RELA, p21, and GAPDH. Graphs show fold changes relative to control and normalized to GAPDH.

Discussion
Accumulating evidence suggest that TAK1 is a potential therapeutic target for treatment of aggressive cancers such as metastatic breast cancer 11,22 , pancreatic cancer 9 and colon cancer 12 . However the contribution of TAK1 to cancer remains underexplored. Here, we identified a link of TAK1-RELA signaling to the regulation of ribosome biogenesis and p53 signaling (Fig. 6D). The data suggest that blockade of the TAK1-RELA signaling axis disrupts ribosome biogenesis leading to activation of p53. In support of this conclusion, we found that blockade of TAK1-RELA signaling by pharmacological or genetic means results in up-regulation of p53 protein and expression of p53 target genes such as p21 and MDM2. The p53-p21 link was validated in isogenic p53-wt and p53-deficient HCT116 cells (Fig. 2D). Second, we found that TAK1 blockade interfered with ribosome biogenesis, affecting the nucleolar structure and reducing expression of regulators of ribosome biogenesis including RRS1 and URB2. Further, depletion of RPL11, a major mediator of p53 activation through ribosomal stress, reduced p53 activation in response to the TAK1 blockade. Third, depletion of RRS1 was sufficient to impair ribosome biogenesis and activate p53. Clinical metadata showed that TNBC/basal-like breast cancers express elevated levels of ribosome biogenesis regulators and high RRS1 levels are associated with poor prognosis. Thus, the TAK1-ribosome axis may represent a novel therapeutic target for treatment of TNBC and advance disease.
Disruption of the TAK1-IKK-RELA signaling module leads to activation of p53. Blockade of TAK1 increased expression of p53-target genes in cells with wild-type p53 (MCF10A, MCF7, WI-38, A549 and HCT116) but not in p53-mutant (MDA-MB-231) or deficient (HCT116-p53ko) cells (Figs. 1-2). Assessment of TAK1 signaling modules revealed a major role of the TAK1-IKK-RELA axis in this response, while inhibition of MEK or cIAPs did not activate p53 (Figs. 2-3). These data are in agreement with prior reports of Verma's group, who found elevated levels of p53 and p21 in Ikk1/2-deficient murine embryonic fibroblasts (MEFs) 57,58 . Consistent with our data, Gilmore's group found inactivation of p53 in immortalized RelA-deficient MEFs 59 . Together, these data suggest that the TAK1-RELA axis may function through ribosome biogenesis to prevent activation of p53 in non-tumor and cancer cells.

TAK1-p53 link and ribosome biogenesis.
The inverse relationship between p53 and NFκB activities has been reported in several prior studies 57,58,60,61 and reviewed in 62,63 . Early studies suggest that p53 and NFκB may compete for limiting pools of transcriptional transactivator p300/CBP and IKK-α-mediated phosphorylation of CBP favors a complex from p53 to NFκB 60,61 . Studies with IKK1/2 knockout fibroblasts implicated IKKβ in NFκB-mediated up-regulation of E3 ubiquitin-protein ligase MDM2 levels leading to ubiquitin-mediated degradation of p53 protein 57 . Further, IKKβ can phosphorylate p53 leading to recruitment of β-TrCP1/BTRC E3 www.nature.com/scientificreports www.nature.com/scientificreports/ ubuquitin ligase that ubiquitinate p53 and facilitates its degradation 58 . Besides IKKs, NFκB transcriptional target BCL3 can induce p53 degradation by up-regulating MDM2 64 , whereas p53 may repress BCL3 expression 65 . On the other hand, induction of NF-κB by oxidative stress can also stabilize and activate p53 66 . Here, we found that inhibition of TAK1 or IKKs resulted in stabilization of p53 in both wt and mutant p53 cell lines. While this observation is consistent with IKKβ-dependent mechanisms mentioned 57,58 , our data on RELA depletion suggest involvement of NFκB transcription in the regulation of p53, and this is compatible with a CBP-competing model 61 . Furthermore, our finding that depletion of ribosomal protein RPL11 reduces activation of p53 by TAK1 blockade suggests a distinct mechanism underlying activation of p53 through the ribosomal stress response. This finding was verified by pharmacological and genetic approaches in several cell lines, and separates our model from the above-mentioned mechanisms. Our data also show that p53 activation is temporal and is negatively regulated via a mechanism involving the p53-MDM2 interaction 39 , which is blocked by Nutlin 3A ( Supplementary  Fig. 5).
Ribosome biogenesis stress mediates activation of p53 in response to blockade of TAK1. The data argue that ribosomal (nucleolar) stress mediates activation of p53 by TAK1-NFκB blockade. We found that TAK1 or RELA blockade disrupted nucleolar structures and attenuated rRNA synthesis (Figs. 3 and 5), and reduced expression of ribosome biogenesis factors (RRS1, URB2) (Fig. 5). Notably, TAK1 blockade did not alter mRNA levels of the core ribosomal proteins (e.g. RPL5, RPL11, RPS5, and RPS6) in the tested cell lines. Accordingly, depletion of RRS1 was sufficient to induce the p53-p21 axis. RRS1 is a key regulator of the 5 S RNP complex entry into the large ribosome subunit 51,67 . Interruption in ribosome biogenesis leads to accumulation in the nucleoplasm of ribosomal 5 S RNP complex containing RPL11 and other ribosomal proteins that inhibit E3-ubiquitin ligase MDM2 thereby stabilizing p53 protein 35,39,50 . Here, the data revealed that depletion of RPL11 prevented activation of p53 in response to blockade of TAK1 or RELA (Fig. 4). Together, these findings support a critical role of ribosomal stress in activation of p53 in response to TAK1 blockade. This new TAK1-ribosome axis may protect cells from abrupt interruption in ribosome production in cancer and normal cells.
The TCGA data showed elevated levels of ribosome biogenesis factors in TNBC cancers suggesting that intervention in ribosome biogenesis may selectively impede TNBCs (Fig. 6). Ribosome biogenesis consumes significant cellular energy and material resources, and defects in ribosome production may contribute to various human diseases including cancer 32 . Thus, the significance of our findings may extend beyond cancer and can be important for better understanding human diseases associated with ribosomal defects such as Treacher Collins syndrome 68 . www.nature.com/scientificreports www.nature.com/scientificreports/ In summary, this work suggests that ribosome biogenesis stress mediates activation of p53 in response to TAK1 blockade. TAK1-RELA signaling may protect cells from unsolicited interruption in ribosome biogenesis. This mechanism might be particularly important for cancers exhibiting increased ribosome biogenesis such as TNBCs. In addition, this mechanism may contribute to maintenance of homeostasis by restraining activation of p53 that may lead to cell death. Thus, the TAK1-ribosome axis provides a novel strategy for therapeutic intervention in TNBC and advanced breast cancer. Immunoblot analysis. A detailed description of immunoblotting has been reported elsewhere 10,11 . Briefly, whole-cell lysates were collected using NP40 Lysis Buffer (0.88% NP-40, 132 mM NaCl, 44 mM Hepes, 8.8 mM NaF) supplemented with 2 mM sodium orthovanadate, 1 mM PMSF and 1X Protease Inhibitor Cocktail (Cat# 11836153001; Roche; Basel, Switzerland). Prior to lysis, cells were grown to 70-80% confluency and, if necessary, treated with 10 ng/mL TNFα. Inhibitors were added 1 hour prior to cytokine treatment. Protein concentrations were measured using the Bio-Rad DC Protein Assay according to the manufacturer's instructions. Proteins were resolved using SDS-PAGE and transferred to nitrocellulose membranes (Cat# 162-0112; BIO-RAD). Transfer was validated by Ponceau S staining. Protein bands were visualized using ECL chemiluminescent reagent (Cat# 32209; Pierce). Relative changes in protein levels were quantified using ImageJ software version 1.52a and presented as fold change in the GAPDH-normalized band density relative to control. q-RT-PCR. Procedures were performed as described in 69 . Briefly, cells were seeded in media containing 10% serum which was changed to 5% serum or serum-free media the following day. RNA extraction was performed using the TRIzol Reagent (Cat# 15596-026; Invitrogen) according to the manufacturer's instructions. cDNA samples were prepared from equal amounts of RNA using M-MLV RT (Cat# M1701; Promega; Madison, WI), and then amplified using 5X HOT FIREPol EvaGreen qPCR Mix Plus (ROX) (Cat# 08-24-00001; Solis Biodyne, Tartu, Estonia) in the Applied Biosystems StepOnePlus Real-Time PCR System (Thermo Fisher Scientific; Waltham, MA). Samples were run in triplicate. Results were analyzed as follows: threshold cycle (Ct) values were normalized using the mean Ct for the reference gene, 5 SrRNA, defined as ΔCt = Ct (test gene) -Ct (mean for the reference gene). The final data were presented as the fold change (FC) between the test and control samples, defined as FC = 2^-(ΔCt (test gene) -ΔCt (mean for control)). Human primer sequences are as follows: CDKN1A/p21,

RT-PCR.
Total RNA was prepared as for q-RT-PCR and RT-PCR reaction were performed as described Flow cytometry. For cell cycle analysis, cells were seeded at 750,000 in 10 cm 2 dishes and left to attach overnight. Cells were treated with 5 µM of 5Z-7-Oxozeaenol for 6 and 24 hours. Collected cells were fixed in ice cold 70% ethanol for 1 hour and stained with Krishan DNA Buffer for 2 hours at RT. Krishan DNA Buffer consisted of (2020) 10:2517 | https://doi.org/10.1038/s41598-020-59340-5 www.nature.com/scientificreports www.nature.com/scientificreports/ the following reagents: Propidium Iodide, Sodium Citrate, RNAase A, NP40, and HCl. Samples were sorted using a BD LSRFortessa cytometer running Diva version 6.1.1 and cell cycle analysis was done using Modfit LT software version 4.1.7. Cell cycle analysis experiments were repeated three times with representative histograms shown.

Microscopy.
Microscopy experiments were done as described in 69 . Briefly, cells were grown on glass coverslips (22 × 22 mm) and treated with inhibitors for 24 hours, then fixed with 4% PFA and permeabilized with 0.05% Triton X-100. For fibrillarin staining samples were blocked with 3% milk in PBS for 30 min at room temperature. The cells were incubated for 1 h with antibodies to fibrillarin (1:400) in 1% milk/PBS followed by incubation for 30 min with Texas red-conjugated secondary antibody (1:500) at room temperature. Fluorescence images were taken with a Plan Apochromat 60×/1.40 NA oil objective at ambient temperature using Nikon TE2000-E inverted microscope equipped with a charge-coupled device camera (CoolSNAP HQ; Photometrics). The images were acquired using MetaVue imaging software (v7.7.3, Molecular Devices). rRNA synthesis. rRNA synthesis was evaluated by pulse-labeling of de novo synthesized RNA with 5-ethynyl uridine followed by coupling of fluorophore using click-it chemistry. Briefly, cells were grown on glass cover slips (22 × 22 mm) and treated with inhibitors for 6 hours, followed by a 1 hour pulse with 5-ethynyl uridine (Cat# PY7563, Berry & Associates, Dexter, MI). Cells were fixed in 4% PFA and permeabilized with 0.05% Triton X-100. To label RNA with Cy3 fluorophore, cells were subjected to 'click-it' reaction with Cu(II)SO 4 , Tris-pH 8.5, Cy3-azide, THTPA, and ascorbic acid for 1 hour, followed by wash and staining with Hoechst. Cover slips were mounted onto glass-slides and fluorescence images were acquired with a 60× oil lens using Nikon TE2000-E microscope as described above. Fluorescence intensity of the nucleus was evaluated for >30 cells/case from at least 3 randomized fields in each case by using MetaVue imaging software. Experiments were repeated at least 2 times.
Metadata analysis. The Kaplan-Meier curves of recurrence-free survival in BC patients were generated using Kaplan-Meier Plotter 55 (http://kmplot.com/analysis/index.php?p=service). Metastasis-free survival data were obtained using a published dataset of 295 BC patients 70 and a prognostic PROGgeneV2 tool 71 , http://genomics.jefferson.edu/proggene/. Heat-map of gene expression profiles was generated with the TCGA BC dataset using the cBioPortal for Cancer Genomics online tool https://www.cbioportal.org/.
Statistical analysis. Data in each experiment was compared using the Student's t test. Statistical significance was achieved when P < 0.05.