Serum Response Factor (SRF) Drives the Transcriptional Upregulation of the MDM4 Oncogene in HCC.

cancers Article Serum Response Factor (SRF) Drives the Transcriptional Upregulation of the MDM4 Oncogene in HCC Rossella Pellegrino 1,* , Abhishek Thavamani 2, Diego F. Calvisi 3, Jan Budczies 1, Ariane Neumann 1, Robert Geffers 4, Jasmin Kroemer 1, Damaris Greule 1, Peter Schirmacher 1, Alfred Nordheim 2and Thomas Longerich 1   Citation: Pellegrino, R.; Thavamani, A.; Calvisi, D.F.; Budczies, J.; Neumann, A.; Geffers, R.; Kroemer, J.; Greule, D.; Schirmacher, P.; Nordheim, A.; et al. Serum Response Factor (SRF) Drives the Transcriptional Upregulation of the MDM4 Oncogene in HCC. Cancers 2021,13, 199. https://doi.org/10.3390/ cancers13020199 Received: 3 October 2020 Accepted: 4 January 2021 Published: 8 January 2021 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. Copyright: © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). 1Institute of Pathology, University Hospital Heidelberg, 69120 Heidelberg, Germany; [email protected] (J.B.); [email protected] (A.N.); [email protected]g.de (J.K.); [email protected]g.de (D.G.); Peter[email protected] (P.S.); [email protected] (T.L.) 2Department for Molecular Biology, Interfaculty Institute of Cell Biology, University of Tuebingen, 72074 Tuebingen, Germany; [email protected] (A.T.); alfred.nor[email protected] (A.N.) 3Institute of Pathology, University Hospital Regensburg, 93053 Regensburg, Germany; [email protected]g.de 4Genome Analytics, Helmholtz Centre for Infection Research, 38124 Braunschweig, Germany; robert.gef[email protected] *Correspondence: r[email protected]; Tel.: +49-(0)6221-56-34094 Simple Summary: Hepatocellular carcinoma (HCC) represents the most common type of liver cancer and has a poor prognosis. Therefore, there is an urgent need for the identification of new therapeutic options. The mouse double minute homolog 4 (MDM4) gene, a known p53 inhibitor, is upregulated in most HCCs. Here, we aimed to investigate the mechanisms leading to MDM4 transcriptional upregulation and to evaluate whether therapeutic targeting of these mechanisms might represent a suitable approach for future therapy. Using human HCC cell lines, a mouse model, and human HCC cohorts, we have identified serum response factor (SRF), ETS transcription factors ELK1 and ELK4 as transcription factors (TFs) driving MDM4 expression. Treatment of HCC cell lines with XI-011, a pharmaceutical inhibitor of MDM4 transcription, reduced the expression of both the TFs and MDM4 and impaired tumor growth, suggesting that targeting the MDM4 transcription may provide a rationale for future targeted therapy of HCC. Abstract: Different molecular mechanisms support the overexpression of the mouse double minute homolog 4 (MDM4), a functional p53 inhibitor, in human hepatocellular carcinoma (HCC). However, the transcription factors (TFs) leading to its transcriptional upregulation remain unknown. Following promoter and gene expression analyses, putative TFs were investigated using gene-specific siRNAs, cDNAs, luciferase reporter assays, chromatin immunoprecipitation, and XI-011 drug treatment in vitro . Additionally, MDM4 expression was investigated in SRF-VP16 iHep transgenic mice. We observed a copy-number-independent upregulation of MDM4 in human HCCs. Serum response factor (SRF), ELK1 and ELK4 were identified as TFs activating MDM4 transcription. While SRF was constitutively detected in TF complexes at the MDM4 promoter, presence of ELK1 and ELK4 was cell-type dependent. Furthermore, MDM4 was upregulated in SRF-VP16-driven murine liver tumors. The pharmacological inhibitor XI-011 exhibited anti-MDM4 activity by downregulating the TFs driving MDM4 transcription, which decreased HCC cell viability and increased apoptosis. In conclusion, SRF drives transcriptional MDM4 upregulation in HCC, acting in concert with either ELK1 or ELK4. The transcriptional regulation of MDM4 may be a promising target for precision oncology of human HCC, as XI-011 treatment exerts anti-MDM4 activity independent from the MDM4 copy number and the p53 status. Keywords: HCC; MDM4; ELK1; ELK4; ETS transcription factors; ERK; tumor protein p53; SRF; MDM4 transcriptional regulation; XI-011 Cancers 2021,13, 199. https://doi.org/10.3390/cancers13020199 https://www.mdpi.com/journal/cancers Cancers 2021,13, 199 2 of 17 1. Introduction Liver cancer is the fifth most common type of cancer in the world and the second most frequent cause of cancer-related death, with limited therapeutic options for patients with advanced stages of the disease [ 1 , 2 ]. Hepatocellular carcinoma (HCC) accounts for the majority (80%) of liver cancer cases, and the underlying etiological factors are well known [ 1 ]. During the last few years, the molecular landscape of human HCC has been comprehensively characterized at genomic, transcriptomic, epigenomic, and proteomic levels [ 3 – 7 ]. The tumor suppressor p53 represents the second most frequently mutated gene in HCC. Its mutation frequency shows a strong geographical variation, which parallels with the exposure to aflatoxin B1 and the prevalence of chronic HBV infection. In endemic regions, such as sub-Saharan Africa or East Asia, a mutation rate of up to 50% has been reported [ 8 ], while the p53 mutation frequency is much lower in western countries, ranging from <10% to 25% [ 5 , 9 , 10 ]. P53 is regulated by a network of interacting factors, most prominently by genes of the mouse double minute (MDM) family [ 11 ]. Both MDM2 and MDM4 are inhibitors of p53. They are able to bind to the N-terminal transactivation domain of p53. While MDM4 is a potent inhibitor of p53 transcriptional activity [ 12 , 13 ], MDM2 mainly functions as a negative feedback regulator of p53 signaling. Transcriptionally activated by p53, MDM2 acts as an E3 ligase targeting p53 for proteasomal degradation [ 13 , 14 ]; of note, the formation of MDM2-MDM4 heterodimer complexes is essential for p53 polyubiquitination, while MDM2 alone marks p53 for monoubiquitination and thus does not promote complete p53 degradation [ 15 ]. In vitro and in vivo experiments highlighted oncogenic properties of both MDM family genes, but mouse models suggested MDM4 to be a more potent p53 inhibitor than MDM2 [ 16 , 17 ]. Dysregulation of MDM4 has been detected in different cancer types and various mechanisms may promote its upregulation [ 18 , 19 ]. We previously reported recurrent amplification of the MDM4 gene locus in human HCC [ 3 ], which was recently validated by a large international consortium [ 7 ]. Furthermore, we proposed a post-transcriptional mechanism, by which activation of the EEF1A2/PI3K/AKT/mTOR signaling axis fosters the protumorigenic function of MDM4 in human HCC and showed that the MDM4 protein level is associated with the survival probability of HCC patients following liver resection [20]. The observation that increased MDM4 expression levels can also be detected in human HCC samples with balanced MDM4 gene locus, led us to hypothesize that transcriptional dysregulation may lead to an upregulation of MDM4 in these cases. As the transcriptional regulation of MDM4 remains largely elusive, we screened the putative basal MDM4 promoter for transcription factor (TF) binding sites in silico, performed validation of the candidate TFs in vitro , and explored whether transcriptional dysregulation of MDM4 might be a drug target for future translational studies. Moreover, we used an SRF transgenic mouse model that spontaneously develops HCC [ 21 ] to further validate our hypothesis in vivo . The data presented here highlight the role of SRF in driving MDM4 transcription, which requires interaction with ELK1 or ELK4 in a cell context-dependent manner. 2. Results 2.1. Transcriptional Activation Contributes to MDM4 Upregulation in Human HCC We have previously reported that genomic gains occur at the MDM4 gene locus (1q32.1) and MDM4 mRNA and protein levels are upregulated in human HCCs compared to normal livers (NLs) [ 3 , 4 ]. However, microarray expression profiling data from 37 human HCCs revealed no significant difference in MDM4 mRNA levels between HCCs with balanced (n= 13) or gained (n= 24) MDM4 gene locus, respectively (balanced: 2.3 ±0.4 vs. gained: 3.7 ± 0.4, p> 0.05; Figure 1A), suggesting that aberrant transcriptional activation may contribute to MDM4 overexpression in HCC. Putative TFs involved in the regulation of MDM4 transcription were identified by an in silico analysis of the MDM4 promoter region using the MAPPER [ 22 ] and Jaspar [ 23 ] databases. As shown in Figure 1B, a high affinity SRF binding site (Jaspar score: 10.1), as well as binding sites for ELK1 and ELK4 (Jaspar score: 9.8 and 8.6, respectively), were detected in the 5 0 -UTR of the MDM4 gene Cancers 2021,13, 199 3 of 17 predicted to contain the basal promoter. Of note, the sequence identified as an ELK1 binding site (CCGGAAG) differed from the complementary reverse ELK4 recognition sequence (TTTCCGG) by only one nucleotide (G); therefore, this sequence was considered a putative ELK1/ELK4 binding site. ELK1 and ELK4 are ETS family proteins of the ternary complex factor (TCF) subfamily, which form ternary complexes with DNA-bound SRF [ 24 ]. Cancers 2021, 13, x 3 of 17 may contribute to MDM4 overexpression in HCC. Putative TFs involved in the regulation of MDM4 transcription were identified by an in silico analysis of the MDM4 promoter region using the MAPPER [22] and Jaspar [23] databases. As shown in Figure 1B, a high affinity SRF binding site (Jaspar score: 10.1), as well as binding sites for ELK1 and ELK4 (Jaspar score: 9.8 and 8.6, respectively), were detected in the 5′-UTR of the MDM4 gene predicted to contain the basal promoter. Of note, the sequence identified as an ELK1 binding site (CCGGAAG) differed from the complementary reverse ELK4 recognition sequence (TTTCCGG) by only one nucleotide (G); therefore, this sequence was considered a putative ELK1/ELK4 binding site. ELK1 and ELK4 are ETS family proteins of the ternary complex factor (TCF) subfamily, which form ternary complexes with DNA-bound SRF [24]. Figure 1. Aberrant transcriptional activation may be involved in upregulation of mouse double minute homolog 4 (MDM4) in human HCC. (A) Relative MDM4 mRNA expression in human hepatocellular carcinomas (HCCs) with balanced (n = 13) and gained MDM4 (n = 24) gene loci, respectively. Mann–Whitney U test: p > 0.05. (B) An in silico analysis of the basal MDM4 promoter region identified putative transcription factor binding sites for serum response factor (SRF), ELK1, and ELK4. (C) Expression profiling revealed a positive association between MDM4 mRNA and the expression level of the putative transcription factors SRF, ELK1, and ELK4 in human HCC samples (n = 37). (D) MDM4, SRF, and ELK4 mRNA levels were associated with the survival probability of HCC patients in a second cohort (n = 32). Each median expression level was used for stratification. Abbreviation: n.s., not statistically significant. Gene expression profiling revealed a significant positive association between the mRNA expression of all three putative TFs and MDM4 in a human HCC cohort (n = 37; Figure 1C). These findings were corroborated in a second series of human HCC (n = 32; (Figure S1), which revealed significantly increased MDM4, SRF, ELK1, and ELK4 mRNA Figure 1. Aberrant transcriptional activation may be involved in upregulation of mouse double minute homolog 4 (MDM4) in human HCC. ( A ) Relative MDM4 mRNA expression in human hepatocellular carcinomas (HCCs) with balanced (n= 13) and gained MDM4 (n= 24) gene loci, respectively. Mann–Whitney U test: p> 0.05. ( B ) An in silico analysis of the basal MDM4 promoter region identified putative transcription factor binding sites for serum response factor (SRF), ELK1, and ELK4. ( C ) Expression profiling revealed a positive association between MDM4 mRNA and the expression level of the putative transcription factors SRF,ELK1, and ELK4 in human HCC samples (n= 37). ( D )MDM4,SRF, and ELK4 mRNA levels were associated with the survival probability of HCC patients in a second cohort (n= 32). Each median expression level was used for stratification. Abbreviation: n.s., not statistically significant. Gene expression profiling revealed a significant positive association between the mRNA expression of all three putative TFs and MDM4 in a human HCC cohort ( n= 37 ; Figure 1C). These findings were corroborated in a second series of human HCC (n= 32; (Figure S1), which revealed significantly increased MDM4,SRF,ELK1, and ELK4 mRNA levels in human HCCs compared to paired surrounding non-neoplastic liver tissues (SL) (Figure S2A–D). Furthermore, survival analysis of the latter cohort showed that high mRNA levels of MDM4,SRF, and ELK4 were associated with a lower survival probability of HCC patients following liver resection (p= 0.0003, p= 0.0386, and p= 0.0151, respectively; Figure 1D), while ELK1 expression levels were of no prognostic value (p> 0.05; personal observation, 2020). Additionally, survival was analyzed using the TCGA data set (LIHC) Cancers 2021,13, 199 4 of 17 following stratification of patients based on the p53 gene status [ 7 ]. As shown in Figure S2E, the overall survival of patients whose HCC showed a wild-type p53 gene sequence was lower, but not significantly different when compared to p53-mutated cases. In addition, further stratification of HCC cases regarding the gene expression level of MDM4,SRF, and ELK4 did not reveal significant differences in terms of survival probability between the individual groups (Figure S2F). Notably, the positive association of the MDM4 mRNA level with the gene expression of SRF and ELK4 could be again validated in the large LIHC cohort (p< 0.001, personal observation, 2020). 2.2. SRF, ELK1, and ELK4 Regulate MDM4 Expression in HCC Cell Lines As SRF is considered the central mediator of the immediate cellular serum response [ 25 ], fetal calf serum (FCS) was used to stimulate HCC cells. Indeed, both HepG2 and HLE cells that had been starved overnight showed significantly higher MDM4 mRNA and protein levels upon stimulation with FCS compared to the nonstimulated control cells (Figure 2A and Figure S3A). This effect correlated with an increased phosphorylation of extracellular signal-related kinase 1/2 (ERK 1/2), indicating an active serum response (Figure S3A). To further corroborate our findings, MDM4 gene expression was assessed in serumstarved HepG2 cells stimulated with FCS in combination with the PI3K inhibitor LY294002 or the ERK inhibitor LY3214996, respectively. The efficacy of PI3K and ERK inhibition was confirmed by detection of decreased AKT and ERK phosphorylation in treated cells compared to controls, respectively (Figure S3C,F). As expected, the inhibition of the individual pathways prevented FCS-induced upregulation of MDM4 expression (Figure S3B–E). Also, knockdown of SRF expression by two different gene-specific siRNAs (siSRF_1/2) significantly reduced MDM4 mRNA and protein levels in HepG2 and HLE cells compared to transfection of a scrambled, nonsense siRNA (siNS) (Figure 2B,C). Similar results were obtained when ELK1 or ELK4 were targeted by gene-specific siRNAs in HCC cell lines (Figure 2D–G). To validate the essential role of SRF in the upregulation of MDM4 following FCS stimulation, MDM4 protein levels were analyzed upon FCS stimulation in overnight serum-starved cells previously transfected with an SRF-specific siRNA and compared to siNS-transfected cells. As for ERK and PI3K pathway inhibition, FCS could not induce MDM4 expression in SRF-depleted HCC cells, while control cells were still FCS responsive (Figure 2H). Furthermore, MDM4 mRNA and protein levels were significantly increased in HuH7 cells, which showed low basal SRF levels (personal observation, 2020) when transiently transfected with an SRF-VP16 plasmid, which encodes for a full-length SRF cDNA fused to the transcriptional activation domain of the herpes simplex virus protein VP16 (Figure 2I). SRF-VP16 is able to bind to SRF target sites in promoter regions and activate transcription without additional co-factors that are physiologically required for SRF-mediated transcriptional activation (see below) [26]. Cancers 2021,13, 199 5 of 17 Cancers 2021, 13, x 5 of 17 Figure 2. SRF, ELK1, and ELK4 regulate MDM4 expression in HCC cell lines. (A) Increased MDM4 mRNA levels in HepG2 and HLE cell lines following fetal calf serum (FCS) stimulation compared to starved control cells. (B) MDM4 mRNA and (C) protein levels following siRNA-mediated knockdown of SRF compared to control cells transfected with a scrambled, nonsense siRNA (siNS) in HepG2 and HLE cells. (D) MDM4 mRNA and (E) protein levels following siRNA-mediated knockdown of ELK1 compared to control cells transfected with a scrambled, nonsense siRNA (siNS) in HepG2 and HLE cells. (F) MDM4 mRNA and (G) protein levels following siRNA-mediated knockdown of ELK4 compared to control cells transfected with a scrambled, nonsense siRNA (siNS) in HepG2 and HLE cells. (H) siRNA-mediated knockdown of SRF (siSRF_2) prevents FCS-stimulated MDM4 protein upregulation. (I) MDM4 mRNA and protein expression 48 h following transfection of HuH7 cells with an SRF-VP16 expression vector compared to mock transfected control cells. Original western blots are shown in Figures S7 and S8. Data are presented as mean ± SEM. Mann–Whitney U test: * p < 0.05, ** p < 0.01, *** p < 0.001. Abbreviations: siNS—scrambled, nonsense siRNA; siSRF_1/_2, siELK1_1/_2, siELK4_1/_2—siRNA 1 and 2 specifically targeting SRF, ELK1, and ELK4, respectively. Figure 2. SRF, ELK1, and ELK4 regulate MDM4 expression in HCC cell lines. ( A ) Increased MDM4 mRNA levels in HepG2 and HLE cell lines following fetal calf serum (FCS) stimulation compared to starved control cells. ( B ) MDM4 mRNA and ( C ) protein levels following siRNA-mediated knockdown of SRF compared to control cells transfected with a scrambled, nonsense siRNA (siNS) in HepG2 and HLE cells. ( D ) MDM4 mRNA and ( E ) protein levels following siRNA-mediated knockdown of ELK1 compared to control cells transfected with a scrambled, nonsense siRNA (siNS) in HepG2 and HLE cells. ( F ) MDM4 mRNA and ( G ) protein levels following siRNA-mediated knockdown of ELK4 compared to control cells transfected with a scrambled, nonsense siRNA (siNS) in HepG2 and HLE cells. ( H ) siRNA-mediated knockdown of SRF (siSRF_2) prevents FCS-stimulated MDM4 protein upregulation. ( I ) MDM4 mRNA and protein expression 48 h following transfection of HuH7 cells with an SRF-VP16 expression vector compared to mock transfected control cells. Original western blots are shown in Figures S7 and S8. Data are presented as mean ± SEM. Mann–Whitney U test: * p< 0.05, ** p< 0.01, *** p< 0.001. Abbreviations: siNS—scrambled, nonsense siRNA; siSRF_1/_2, siELK1_1/_2, siELK4_1/_2—siRNA 1 and 2 specifically targeting SRF, ELK1, and ELK4, respectively. Cancers 2021,13, 199 6 of 17 2.3. ELK1 and ELK4 Are Co-Factors for SRF-Mediated Transcriptional Regulation of MDM4 in HCC Cells To demonstrate that SRF, ELK1, and ELK4 indeed effectively induce the transcription of the MDM4 gene, the specific siRNAs targeting SRF, ELK1, and ELK4, together with an MDM4 promoter construct carrying a Gaussia luciferase as reporter, were transiently transfected in HCC cells. Efficient depletion of SRF, ELK1, and ELK4 by gene-specific siRNAs (Figure S3F) resulted in a significant decrease of luciferase activity compared to controls (Figure 3A–C). Cancers 2021, 13, x 6 of 17 2.3. ELK1 and ELK4 Are Co-Factors for SRF-Mediated Transcriptional Regulation of MDM4 in HCC Cells To demonstrate that SRF, ELK1, and ELK4 indeed effectively induce the transcription of the MDM4 gene, the specific siRNAs targeting SRF, ELK1, and ELK4, together with an MDM4 promoter construct carrying a Gaussia luciferase as reporter, were transiently transfected in HCC cells. Efficient depletion of SRF, ELK1, and ELK4 by gene-specific siRNAs (Figure S3F) resulted in a significant decrease of luciferase activity compared to controls (Figure 3A–C). Figure 3. ELK1 and ELK4 are essential co-factors for SRF-mediated transcriptional regulation of MDM4 in HCC. Luciferase activity of a MDM4 promoter reporter upon siRNA-mediated knockdown of (A) SRF, (B) ELK1, and (C) ELK4 in HepG2 and HLE cells compared to controls. (D) MDM4 mRNA levels after co-transfection of an ELK1 cDNA with siNS or siSRF. Transfection efficacy was confirmed by detection of ELK1 and SRF mRNA levels. (E) MDM4 mRNA levels following transfection of the indicated cDNA plasmids. ELK1 S383A represents an inactive variant, which cannot be activated by phosphorylation of S383 and is thus unable to initiate target gene transcription. Transfection efficacy was confirmed by detection of ELK1 and SRF mRNA levels. Data are presented as mean ± SEM. Mann-Whitney U test: * p < 0.05, ** p < 0.01, *** p < 0.001. Abbreviations: siNS, scrambled, nonsense; siSRF_1/_2, siELK1_1/_2 siELK4_1/_2, siRNA 1 and 2 specifically targeting SRF, ELK1 and ELK4, respectively; ELK1, ELK1 cDNA; ELK1 S383A, ELK1 S383A cDNA; GLuc, Gaussia luciferase; SEAP, Secreted Alkaline Phosphatase; norm., normalized against control. While ELK1 is capable of initiating target gene transcription on its own [27], SRF requires additional cofactors to activate the transcription of its target genes. These include either members of the ternary complex factor (TCF) family of ETS domain proteins (ELK1, ELK4, NET) or the myocardin-related transcription factor (MRTF) family (MKL1 and MKL2) [24,28]. Modulation of MRTF gene expression did not affect MDM4 mRNA expression (personal observation, 2020). In contrast, transient ELK1 transfection significantly increased MDM4 and SRF mRNA levels in HLE cells, which were blocked by previous siRNA-mediated knockdown of SRF expression (Figure 3D). Additionally, co-transfection Figure 3. ELK1 and ELK4 are essential co-factors for SRF-mediated transcriptional regulation of MDM4 in HCC. Luciferase activity of a MDM4 promoter reporter upon siRNA-mediated knockdown of ( A ) SRF, ( B ) ELK1, and ( C ) ELK4 in HepG2 and HLE cells compared to controls. ( D )MDM4 mRNA levels after co-transfection of an ELK1 cDNA with siNS or siSRF. Transfection efficacy was confirmed by detection of ELK1 and SRF mRNA levels. ( E )MDM4 mRNA levels following transfection of the indicated cDNA plasmids. ELK1 S383A represents an inactive variant, which cannot be activated by phosphorylation of S383 and is thus unable to initiate target gene transcription. Transfection efficacy was confirmed by detection of ELK1 and SRF mRNA levels. Data are presented as mean ± SEM. Mann-Whitney U test: * p< 0.05, ** p< 0.01, *** p< 0.001. Abbreviations: siNS, scrambled, nonsense; siSRF_1/_2, siELK1_1/_2 siELK4_1/_2, siRNA 1 and 2 specifically targeting SRF, ELK1 and ELK4, respectively; ELK1, ELK1 cDNA; ELK1 S383A, ELK1 S383A cDNA; GLuc, Gaussia luciferase; SEAP, Secreted Alkaline Phosphatase; norm., normalized against control. While ELK1 is capable of initiating target gene transcription on its own [ 27 ], SRF requires additional cofactors to activate the transcription of its target genes. These include either members of the ternary complex factor (TCF) family of ETS domain proteins (ELK1, ELK4, NET) or the myocardin-related transcription factor (MRTF) family (MKL1 and MKL2) [ 24 , 28 ]. Modulation of MRTF gene expression did not affect MDM4 mRNA expression (personal observation, 2020). In contrast, transient ELK1 transfection significantly increased MDM4 and SRF mRNA levels in HLE cells, which were blocked by previous Cancers 2021,13, 199 7 of 17 siRNA-mediated knockdown of SRF expression (Figure 3D). Additionally, co-transfection of SRF and an inactive ELK1 mutant (ELK1 S383A) significantly lowered the induction of MDM4 mRNA compared to a wildtype ELK1 cDNA in HuH7 cells (Figure 3E). Taken together, these data demonstrate that SRF drives the upregulation of MDM4 in HCC cells, most likely in combination with either ELK1 or ELK4. 2.4. SRF and TCF Family Members Control the Activity of the MDM4 Promoter in HCC Cell Lines To test for the physical interaction between the three TF candidates and the MDM4 promoter sequence in HCC cell lines, ChIP experiments were performed as outlined in Figure 4A. Specific binding of SRF, ELK1, and ELK4 were detected at their respective TF binding sites in the MDM4 promoter in both HepG2 and HLE cells (Figure 4B). Comparing the enrichment of ELK1 and ELK4 at the MDM4 promoter, ELK1 was more enriched in HLE cells, while ELK4 was the most prevalent TCF family member at the MDM4 promoter in HepG2 cells (ELK1: 17.0 ± 1.93 (HLE) vs. 1.5 ± 0.04 (HepG2), p< 0.05; ELK4: 5.1 ±0.92 (HLE) vs. 5.6 ± 0.27 (HepG2), p> 0.05). The specificity of TF binding to their cognate sequences was verified using control primers upstream (2.1 Kb) and downstream (1.3 Kb) from the basal promoter region (p< 0.05 for all the TFs analyzed). Cancers 2021, 13, x 7 of 17 of SRF and an inactive ELK1 mutant (ELK1 S383A) significantly lowered the induction of MDM4 mRNA compared to a wildtype ELK1 cDNA in HuH7 cells (Figure 3E). Taken together, these data demonstrate that SRF drives the upregulation of MDM4 in HCC cells, most likely in combination with either ELK1 or ELK4. 2.4. SRF and TCF Family Members Control the Activity of the MDM4 Promoter in HCC Cell Lines To test for the physical interaction between the three TF candidates and the MDM4 promoter sequence in HCC cell lines, ChIP experiments were performed as outlined in Figure 4A. Specific binding of SRF, ELK1, and ELK4 were detected at their respective TF binding sites in the MDM4 promoter in both HepG2 and HLE cells (Figure 4B). Comparing the enrichment of ELK1 and ELK4 at the MDM4 promoter, ELK1 was more enriched in HLE cells, while ELK4 was the most prevalent TCF family member at the MDM4 promoter in HepG2 cells (ELK1: 17.0 ± 1.93 (HLE) vs. 1.5 ± 0.04 (HepG2), p < 0.05; ELK4: 5.1 ± 0.92 (HLE) vs. 5.6 ± 0.27 (HepG2), p > 0.05). The specificity of TF binding to their cognate sequences was verified using control primers upstream (2.1 Kb) and downstream (1.3 Kb) from the basal promoter region (p < 0.05 for all the TFs analyzed). Figure 4. MDM4 gene promoter is activated by an SRF-ETS family transcription factor complex in HCC cell lines. (A) Schematic representation of the positioning of primers used for ChIP analyses of the MDM4 promoter region. (B) Specific binding of SRF, ELK1, and ELK4 at their cognate binding sites in the MDM4 promoter compared to control primers located either downor upstream of the predicted basal MDM4 promoter as detected by quantitative real-time PCR of immunoprecipitated chromatin. (C) Relative enrichment of SRF-immunoprecipitated DNA in HuH7 cells transfected with SRFVP16 expression plasmid compared to mock transfected control cells. Data are presented as mean ± SEM. Mann–Whitney U test: ** p < 0.01, *** p < 0.001. Abbreviations: upstr. control, control primer amplifying a region upstream of the MDM4 promoter; downstr. control, control primer amplifying a region downstream of the MDM4 promoter. Figure 4. MDM4 gene promoter is activated by an SRF-ETS family transcription factor complex in HCC cell lines. ( A ) Schematic representation of the positioning of primers used for ChIP analyses of the MDM4 promoter region. ( B ) Specific binding of SRF, ELK1, and ELK4 at their cognate binding sites in the MDM4 promoter compared to control primers located either downor upstream of the predicted basal MDM4 promoter as detected by quantitative real-time PCR of immunoprecipitated chromatin. ( C ) Relative enrichment of SRF-immunoprecipitated DNA in HuH7 cells transfected with SRF-VP16 expression plasmid compared to mock transfected control cells. Data are presented as mean ± SEM. Mann–Whitney U test: ** p< 0.01, *** p< 0.001. Abbreviations: upstr. control, control primer amplifying a region upstream of the MDM4 promoter; downstr. control, control primer amplifying a region downstream of the MDM4 promoter. Cancers 2021,13, 199 8 of 17 Of note, ELK1 expression was lower in HepG2 compared to HLE cells, while both cell lines expressed similar ELK4 levels, likely explaining the differential pattern observed in the immunoprecipitation experiments (Figure S4). Furthermore, the binding of SRF to the MDM4 core promoter was validated by overexpression of SRF-VP16, which resulted in a significant enrichment of SRF at the MDM4 promoter in HuH7 cells compared to control cells (Figure 4C). To further investigate the relevance of SRF for MDM4 gene expression in vivo , SRF-VP16 transgenic mice (SRF-VP16 iHep ), which express a constitutively active SRF-VP16 fusion protein in hepatocytes, were analyzed. These mice develop HCC via a premalignant nodular stage [ 21 ]. Immunohistochemistry revealed upregulation of the MDM4 protein in HCCs of SRF-VP16 iHep mice compared to control mice (Figure 5). In line with the in vitro data, increased nuclear pELK1 and ELK4 protein expression were detected in SRF-VP16-induced HCCs compared to wildtype littermates (Figure 5). Cancers 2021, 13, x 8 of 17 Of note, ELK1 expression was lower in HepG2 compared to HLE cells, while both cell lines expressed similar ELK4 levels, likely explaining the differential pattern observed in the immunoprecipitation experiments (Figure S4). Furthermore, the binding of SRF to the MDM4 core promoter was validated by overexpression of SRF-VP16, which resulted in a significant enrichment of SRF at the MDM4 promoter in HuH7 cells compared to control cells (Figure 4C). To further investigate the relevance of SRF for MDM4 gene expression in vivo, SRF-VP16 transgenic mice (SRF-VP16iHep), which express a constitutively active SRF-VP16 fusion protein in hepatocytes, were analyzed. These mice develop HCC via a premalignant nodular stage [21]. Immunohistochemistry revealed upregulation of the MDM4 protein in HCCs of SRF-VP16iHep mice compared to control mice (Figure 5). In line with the in vitro data, increased nuclear pELK1 and ELK4 protein expression were detected in SRF-VP16-induced HCCs compared to wildtype littermates (Figure 5). Figure 5. MDM4 is upregulated in SRF-VP16iHeptransgenic mice. (A) Normal liver parenchyma in control mice (HE staining). (B) Well-differentiated HCC in a 30-week-old SRF-VP16iHep mouse showing trabecular disarray and pseudogland formation. MDM4 immunostaining is negative in control mice (C), while a diffuse, predominantly nuclear staining is seen in SRF-VP16iHep mice (D). Individual hepatocyte nuclei (arrow) are positive for phosphorylated-ELK1 in the control liver (E), whereas the number of p-ELK1 positive nuclei is significantly increased in SRF-VP16iHep mice (F). Figure 5. MDM4 is upregulated in SRF-VP16 iHep transgenic mice. ( A ) Normal liver parenchyma in control mice (HE staining). ( B ) Well-differentiated HCC in a 30-week-old SRF-VP16 iHep mouse showing trabecular disarray and pseudogland formation. MDM4 immunostaining is negative in control mice ( C ), while a diffuse, predominantly nuclear staining is seen in SRF-VP16 iHep mice ( D ). Individual hepatocyte nuclei (arrow) are positive for phosphorylated-ELK1 in the control liver ( E ), whereas the number of p-ELK1 positive nuclei is significantly increased in SRF-VP16 iHep mice ( F ). There is no ELK4 immunosignal in control mice ( G ). In contrast, the SRF-VP16-induced HCC reveals weak to moderate nuclear ELK4 staining (H). Scale bar: 20 µM. Cancers 2021,13, 199 9 of 17 Furthermore, MDM4 mRNA expression was significantly upregulated in neoplastic lesions from SRF-VP16 iHep mice compared to their corresponding controls (normal liver 0.13 ± 0.07 vs. HCC 1.51 ± 0.66 (mean ± SEM), Wilcoxon test p< 0.01), confirming that SRF activates MDM4 transcription and results in an upregulation of MDM4 in vivo. 2.5. XI-011 Inhibits MDM4 Transcription by TF Downregulation Recently, a drug screening study showed that XI-011, a pseudourea derivative, is a potent p53 activator and reduces MDM4 protein levels in breast as well as head and neck cancer cells [ 29 ]. Similarly, XI-011 treatment significantly reduced MDM4 mRNA levels in a dose-dependent manner in HepG2 and HLE cells (Figure 6A). The strongest effect was recorded at concentrations of 0.5 and 1 µ M XI-011 in both cell lines, respectively. Downregulation of MDM4 protein was consistently detected after 16 h of XI-011 treatment in these cell lines (Figure 6B) and similar results were obtained at every time point tested (Figure S5A). Of note, MDM4 mRNA and protein levels were also decreased following XI-011 treatment in Hep3B cells with deleted p53 alleles (Figure S5B,C). In line with this, XI-011 significantly reduced MDM4 promoter activity in HepG2 and HLE cells (Figure 6C). Of note, MDM2 protein levels were not affected by XI-011 treatment in HCC cell lines (Figure S5D). XI-011-mediated reduction of MDM4 expression restored the p53 function in HepG2 cells, as indicated by the upregulation of p53 protein and induction of apoptosis (PARP cleavage) (Figure 6D). Additionally, reactivation of the p53-mediated transcription was confirmed by upregulation of p21 mRNA in HepG2 cells (Figure 6E). Of note, XI011-induced PARP cleavage was decreased following siRNA-mediated MDM4 depletion in HepG2 cells compared with siNS-transfected control cells, indicating that apoptosis induction by XI-011 requires MDM4 expression (Figure S5E). Although HLE cells harbor a mutant p53 gene (p.R249S), XI-011 treatment also resulted in upregulation of p21 mRNA and induction of PARP cleavage in these cells (Figure 6D,E). After inhibition of protein biosynthesis by cycloheximide (CHX) treatment, the half-life time of the wild-type p53 protein was increased in HepG2 cells following XI-011 treatment compared to dimethylsulfoxide (DMSO)-treated control cells (Figure 6F). In contrast, the half-life time of mutant p53 in HLE cells was not affected by XI-011 (Figure 6F), confirming that mutant p53 may escape from proteasomal degradation induced by MDM2-MDM4 heterodimers, as previously reported from other cancer entities [ 30 ]. However, p53 mRNA levels were not affected by XI-011 treatment in HepG2 cells, whereas increased p53 gene expression was observed in HLE cells treated with 0.2 and 0.5 µ M XI-011, which contrasts with the unaltered p53 protein levels of this cell line (Figure S5I), suggesting the possibility that the p53 variant R249S has not completely lost its function to transcriptionally activate (some) p53 target genes. Additionally, XI-011 treatment significantly decreased the viability of HepG2 and HLE cells compared to DMSO-treated controls in a doseand time-dependent manner (Figure 6G), similarly to reduced cell growth observed following MDM4 depletion in vitro and in vivo [ 3 , 20 ]. Of note, siRNA-mediated p53 inhibition in combination with XI-011 treatment lowered the MDM4 gene expression and consequently reduced the viability of HepG2 and HLE cells compared to the corresponding controls, excluding the possibility that the observed XI-011-driven biological effects were mediated by p53 ( Figure S5F–H ). Since we observed that XI-011 reduced MDM4 transcription, we hypothesized that XI-011 may affect the machinery driving MDM4 transcription. Indeed, SRF, ELK1, and ELK4 protein levels were significantly reduced upon XI-011 treatment (0.5 and 1 µ M) in both HCC cell lines (Figure 6H). Thus, XI-011 reduced MDM4 expression by targeting the central TFs required to activate MDM4 transcription. Importantly, XI-011 did not affect the protein half-life time of the TFs (Figure S6A), suggesting that the observed downregulation of transcription factors was not due to increased proteasomal degradation. In line with this, the expression of a selection of canonical SRF targets (VCL1,VIM,BCL2) [ 31 ] was reduced following XI-011 treatment in HLE cells compared to DMSO-treated control cells (Figure S6B), while ACTB mRNA levels were not significantly affected by the same treatment (Figure S6B). Additionally, c-MYC expression, which was found upregulated Cancers 2021,13, 199 16 of 17 Acknowledgments: The authors would like to thank Veronika Geissler and the Center for Model System and Comparative Pathology (CMCP) at the Institute of Pathology in Heidelberg for their excellent technical support in performing the immunohistochemistry staining and the NCT Tissue Bank Heidelberg for support with tissue banking. Conflicts of Interest: The authors declare no conflict of interest. References 1. El-Serag, H.B. Hepatocellular carcinoma. N. Engl. J. Med. 2011,365, 1118–1127. [CrossRef] [PubMed] 2. Ghouri, Y.A.; Mian, I.; Rowe, J.H. Review of hepatocellular carcinoma: Epidemiology, etiology, and carcinogenesis. J. Carcinog. 2017,16, 1. [PubMed] 3. Schlaeger, C.; Longerich, T.; Schiller, C.; Bewerunge, P.; Mehrabi, A.; Toedt, G.; Kleeff, J.; Ehemann, V.; Eils, R.; Lichter, P.; et al. Etiology-dependent molecular mechanisms in human hepatocarcinogenesis. Hepatology 2008 ,47, 511–520. [CrossRef] [PubMed] 4. Neumann, O.; Kesselmeier, M.; Geffers, R.; Pellegrino, R.; Radlwimmer, B.; Hoffmann, K.; Ehemann, V.; Schemmer, P.; Schirmacher, P.; Lorenzo Bermejo, J.; et al. Methylome analysis and integrative profiling of human HCCs identify novel protumorigenic factors. Hepatology 2012,56, 1817–1827. [CrossRef] [PubMed] 5. Guichard, C.; Amaddeo, G.; Imbeaud, S.; Ladeiro, Y.; Pelletier, L.; Maad, I.B.; Calderaro, J.; Bioulac-Sage, P.; Letexier, M.; Degos, F.; et al. Integrated analysis of somatic mutations and focal copy-number changes identifies key genes and pathways in hepatocellular carcinoma. Nat. Genet. 2012,44, 694–698. [CrossRef] [PubMed] 6. Fujimoto, A.; Furuta, M.; Totoki, Y.; Tsunoda, T.; Kato, M.; Shiraishi, Y.; Tanaka, H.; Taniguchi, H.; Kawakami, Y.; Ueno, M.; et al. Whole-genome mutational landscape and characterization of noncoding and structural mutations in liver cancer. Nat. Genet. 2016,48, 500–509. [CrossRef] 7. Ally, A.; Balasundaram, M.; Carlsen, R.; Chuah, E.; Clarke, A.; Dhalla, N.; Holt, R.A.; Jones, S.J.; Lee, D.; Ma, Y.; et al. Comprehensive and integrative genomic characterization of hepatocellular carcinoma. Cell 2017,169, 1327–1341. [CrossRef] 8. Schulze, K.; Nault, J.C.; Villanueva, A. Genetic profiling of hepatocellular carcinoma using next-generation sequencing. J. Hepatol. 2016,65, 1031–1042. [CrossRef] 9. Zucman-Rossi, J.; Villanueva, A.; Nault, J.C.; Llovet, J.M. Genetic landscape and biomarkers of hepatocellular carcinoma. Gastroenterology 2015,149, 1226–1239. [CrossRef] 10. Kubicka, S.; Trautwein, C.; Schrem, H.; Tillmann, H.; Manns, M. Low incidence of p53 mutations in European hepatocellular carcinomas with heterogeneous mutation as a rare event. J. Hepatol. 1995,23, 412–419. [CrossRef] 11. Marine, J.C.; Francoz, S.; Maetens, M.; Wahl, G.; Toledo, F.; Lozano, G. Keeping p53 in check: Essential and synergistic functions of Mdm2 and Mdm4. Cell Death Differ. 2006,13, 927–934. [CrossRef] 12. Shvarts, A.; Bazuine, M.; Dekker, P.; Ramos, Y.F.; Steegenga, W.T.; Merckx, G.; van Ham, R.C.; van der Houven van Oordt, W.; van der Eb, A.J .; Jochemsen, A.G. Isolation and identification of the human homolog of a new p53-binding protein, Mdmx. Genomics 1997,43, 34–42. [CrossRef] [PubMed] 13. Momand, J.; Zambetti, G.P.; Olson, D.C.; George, D.; Levine, A.J. The mdm-2 oncogene product forms a complex with the p53 protein and inhibits p53-mediated transactivation. Cell 1992,69, 1237–1245. [CrossRef] 14. Momand, J.; Zambetti, G.P. Mdm-2: “Big brother” of p53. J. Cell Biochem. 1997,64, 343–352. [CrossRef] 15. Wang, X.; Wang, J.; Jiang, X. MdmX protein is essential for Mdm2 protein-mediated p53 polyubiquitination. J. Biol. Chem. 2011 , 286, 23725–23734. [CrossRef] [PubMed] 16. Toledo, F.; Krummel, K.A.; Lee, C.J.; Liu, C.W.; Rodewald, L.W.; Tang, M.; Wahl, G.M. A mouse p53 mutant lacking the proline-rich domain rescues Mdm4 deficiency and provides insight into the Mdm2-Mdm4-p53 regulatory network. Cancer Cell 2006 ,9, 273–285. [CrossRef] 17. Danovi, D.; Meulmeester, E.; Pasini, D.; Migliorini, D.; Capra, M.; Frenk, R.; de Graaf, P.; Francoz, S.; Gasparini, P.; Gobbi, A.; et al. Amplification of Mdmx (or Mdm4) directly contributes to tumor formation by inhibiting p53 tumor suppressor activity. Mol. Cell Biol. 2004,24, 5835–5843. [CrossRef] 18. Wade, M.; Li, Y.C.; Wahl, G.M. MDM2, MDMX and p53 in oncogenesis and cancer therapy. Nat. Rev. Cancer 2013 ,13, 83–96. [CrossRef] 19. Markey, M.P. Regulation of MDM4. Front. Biosci. 2011,16, 1144–1156. [CrossRef] 20. Pellegrino, R.; Calvisi, D.F.; Neumann, O.; Kolluru, V.; Wesely, J.; Chen, X.; Wang, C.; Wuestefeld, T.; Ladu, S.; Elgohary, N.; et al. EEF1A2 inactivates p53 by way of PI3K/AKT/mTOR-dependent stabilization of MDM4 in hepatocellular carcinoma. Hepatology 2014,59, 1886–1899. [CrossRef] 21. Ohrnberger, S.; Thavamani, A.; Braeuning, A.; Lipka, D.B.; Kirilov, M.; Geffers, R.; Autenrieth, S.E.; Romer, M.; Zell, A.; Bonin, M.; et al . Dysregulated serum response factor triggers formation of hepatocellular carcinoma. Hepatology 2015 ,61, 979–989. [CrossRef] 22. Marinescu, V.D.; Kohane, I.S.; Riva, A. The MAPPER database: A multi-genome catalog of putative transcription factor binding sites. Nucleic Acids Res. 2005,33, D91–D97. [CrossRef] 23. Khan, A.; Fornes, O.; Stigliani, A.; Gheorghe, M.; Castro-Mondragon, J.A.; van der Lee, R.; Bessy, A.; Cheneby, J.; Kulkarni, S.R.; Tan, G.; et al. JASPAR 2018: Update of the open-access database of transcription factor binding profiles and its web framework. Nucleic Acids Res. 2018,46, D1284. [CrossRef] Cancers 2021,13, 199 17 of 17 24. Olson, E.N.; Nordheim, A. Linking actin dynamics and gene transcription to drive cellular motile functions. Nat. Rev. Mol. Cell Biol. 2010,11, 353–365. [CrossRef] [PubMed] 25. Norman, C.; Runswick, M.; Pollock, R.; Treisman, R. Isolation and properties of cDNA clones encoding SRF, a transcription factor that binds to the c-fos serum response element. Cell 1988,55, 989–1003. [CrossRef] 26. Li, L.; Liu, Z.; Mercer, B.; Overbeek, P.; Olson, E.N. Evidence for serum response factor-mediated regulatory networks governing SM22alpha transcription in smooth, skeletal, and cardiac muscle cells. Dev. Biol. 1997,187, 311–321. [CrossRef] [PubMed] 27. Sharrocks, A.D. Complexities in ETS-domain transcription factor function and regulation: Lessons from the TCF (ternary complex factor) subfamily. The Colworth Medal Lecture. Biochem. Soc. Trans. 2002,30, 1–9. [CrossRef] 28. Posern, G.; Treisman, R. Actin’ together: Serum response factor, its cofactors and the link to signal transduction. Trends Cell Biol. 2006,16, 588–596. [CrossRef] 29. Berkson, R.G.; Hollick, J.J.; Westwood, N.J.; Woods, J.A.; Lane, D.P.; Lain, S. Pilot screening programme for small molecule activators of p53. Int. J. Cancer 2005,115, 701–710. [CrossRef] 30. Li, D.; Marchenko, N.D.; Schulz, R.; Fischer, V.; Velasco-Hernandez, T.; Talos, F.; Moll, U.M. Functional inactivation of endogenous MDM2 and CHIP by HSP90 causes aberrant stabilization of mutant p53 in human cancer cells. Mol. Cancer Res. 2011 ,9, 577–588. [CrossRef] 31. Esnault, C.; Stewart, A.; Gualdrini, F.; East, P.; Horswell, S.; Matthews, N.; Treisman, R. Rho-actin signaling to the MRTF coactivators dominates the immediate transcriptional response to serum in fibroblasts. Genes Dev. 2014,28, 943–958. [CrossRef] [PubMed] 32. Gilkes, D.M.; Pan, Y.; Coppola, D.; Yeatman, T.; Reuther, G.W.; Chen, J. Regulation of MDMX expression by mitogenic signaling. Mol. Cell Biol. 2008,28, 1999–2010. [CrossRef] [PubMed] 33. Price, M.A.; Rogers, A.E.; Treisman, R. Comparative analysis of the ternary complex factors Elk-1, SAP-1a and SAP-2 (ERP/NET). EMBO J. 1995,14, 2589–2601. [CrossRef] [PubMed] 34. Sun, K.; Battle, M.A.; Misra, R.P.; Duncan, S.A. Hepatocyte expression of serum response factor is essential for liver function, hepatocyte proliferation and survival, and postnatal body growth in mice. Hepatology 2009 ,49, 1645–1654. [CrossRef] [PubMed] 35. Chai, J.; Tarnawski, A.S. Serum response factor: Discovery, biochemistry, biological roles and implications for tissue injury healing. J. Physiol. Pharm. 2002,53, 147–157. 36. Boyault, S.; Rickman, D.S.; de Reynies, A.; Balabaud, C.; Rebouissou, S.; Jeannot, E.; Herault, A.; Saric, J.; Belghiti, J.; Franco, D.; et al . Transcriptome classification of HCC is related to gene alterations and to new therapeutic targets. Hepatology 2007,45, 42–52. [CrossRef] 37. Longerich, T.; Mueller, M.M.; Breuhahn, K.; Schirmacher, P.; Benner, A.; Heiss, C. Oncogenetic tree modeling of human hepatocarcinogenesis. Int. J. Cancer 2012,130, 575–583. [CrossRef] 38. Wang, H.; Ma, X.; Ren, S.; Buolamwini, J.K.; Yan, C. A small-molecule inhibitor of MDMX activates p53 and induces apoptosis. Mol. Cancer. 2011,10, 69–79. [CrossRef] 39. Miranda, P.J.; Buckley, D.; Raghu, D.; Pang, J.B.; Takano, E.A.; Vijayakumaran, R.; Teunisse, A.F.; Posner, A.; Procter, T.; Herold, M.J.; et al. MDM4 is a rational target for treating breast cancers with mutant p53. J. Pathol. 2017 ,241, 661–670. [CrossRef] 40. Wang, H.; Yan, C. A small-molecule p53 activator induces apoptosis through inhibiting MDMX expression in breast cancer cells. Neoplasia 2011,13, 611–619. [CrossRef] 41. Calvisi, D.F.; Donninger, H.; Vos, M.D.; Birrer, M.J.; Gordon, L.; Leaner, V.; Clark, G.J. NORE1A tumor suppressor candidate modulates p21CIP1 via p53. Cancer Res. 2009,69, 4629–4637. [CrossRef] [PubMed] 42. Lee, S.M.; Vasishtha, M.; Prywes, R. Activation and repression of cellular immediate early genes by serum response factor cofactors. J. Biol. Chem. 2010,285, 22036–22049. [CrossRef] [PubMed] 43. Pellegrino, R.; Calvisi, D.F.; Ladu, S.; Ehemann, V.; Staniscia, T.; Evert, M.; Dombrowski, F.; Schirmacher, P.; Longerich, T. Oncogenic and tumor suppressive roles of polo-like kinases in human hepatocellular carcinoma. Hepatology 2010 ,51, 857–868. [CrossRef] [PubMed] 44. Liu, J.; Lichtenberg, T.; Hoadley, K.A.; Poisson, L.M.; Lazar, A.J.; Cherniack, A.D.; Kovatich, A.J.; Benz, C.C.; Levine, D.A.; Lee, A.V.; et al . Cancer genome atlas research. An integrated TCGA pan-cancer clinical data resource to drive high-quality survival outcome analytics. Cell 2018,173, 400–416. [CrossRef] [PubMed]