Histone deacetylase inhibitors dysregulate DNA repair proteins and antagonize metastasis-associated processes
We set out to determine whether clinically tested epigenetic drugs against class I histone deacetylases (HDACs) affect hallmarks of the metastatic process.
We treated permanent and primary renal, lung, and breast cancer cells with the class I histone deacetylase inhibitors (HDACi) entinostat (MS-275) and valproic acid (VPA), the replicative stress inducer hydroxyurea (HU), the DNA-damaging agent cis-platinum (L-OHP), and the cytokine transforming growth factor-β (TGFβ). We used proteomics, quantitative PCR, immunoblot, single cell DNA damage assays, and flow cytometry to analyze cell fate after drug exposure.
We show that HDACi interfere with DNA repair protein expression and trigger DNA damage and apoptosis alone and in combination with established chemotherapeutics. Furthermore, HDACi disrupt the balance of cell adhesion protein expression and abrogate TGFβ-induced cellular plasticity of transformed cells.
HDACi suppress the epithelial–mesenchymal transition (EMT) and compromise the DNA integrity of cancer cells. These data encourage further testing of HDACi against tumor cells.
KeywordsAdhesion EMT DNA damage HDACi MET TGFβ
The formation of metastases that originate from a primary cancer is commonly associated with increased drug resistance and patient death (Fidler and Kripke 2015). EMT and the subsequent transition of cells back to the mesenchymal state have been associated with metastasis for decades (Nieto 2013; Zeisberg and Neilson 2009). However, there is an ongoing dispute whether EMT is a prime event in the metastatic process or whether the mesenchymal phenotype of breast and pancreatic cancer cells represents predominantly an indicator of cellular resistance to DNA damage (Aiello et al. 2017; Brabletz et al. 2018; Fischer et al. 2015; Ye et al. 2017; Zheng et al. 2015). Irrespective of this conceptual conflict, it is undoubted that novel drugs are necessary to combat clinical metastasis formation to enhance patient survival. Such drugs should be analyzed for their impact on both genomic integrity and modulation of EMT.
HDACi are epigenetic drugs that enhance protein acetylation and thereby impact a large number of cellular functions (Bayat Mokhtari et al. 2017; Mrakovcic et al. 2019; Müller and Krämer 2010; Nikolova et al. 2017; Vancurova et al. 2018). Since the Food and Drug Administration has approved four HDACi for the treatment of hematological malignancies, additional research is warranted to demonstrate how a pharmacological inactivation of HDACs affects metastasis formation. To solve this issue, global assays and analyses of various tumor cell types are required. We recently revealed that HDACi did not shift renal cell carcinoma (RCC) cells to a distinct epithelial or mesenchymal phenotype, but rather disrupted functional EMT/MET protein expression signatures and triggered apoptosis of RCC cells (Kiweler et al. 2018). These data are coherent with previous reports that show beneficial effects of HDACi against RCC cells and EMT (Chun 2018; Jones et al. 2009a, b; Juengel et al. 2014; Mao et al. 2017). Furthermore, HDACi counteracted the acquired resistance of RCC cells against the mammalian target of rapamycin-inhibitor everolimus and the glucose-regulating biguanide metformin (Juengel et al. 2014; Wei et al. 2018). In light of the chemoresistance and the poor prognosis of metastatic RCC (Barbieri et al. 2017; Chang et al. 2019), these findings suggest that HDACi pose an interesting therapeutic option for this cancer type.
The analysis of drug-dependent effects on metastasis should also involve conditions that promote this process. The secreted cytokine TGFβ is a tumor suppressor in non-transformed cells through its cell cycle arresting activity. Tumor cells, including those from RCC, are insensitive to this effect of TGFβ and undergo metastasis-promoting EMT and acquire chemotherapy resistance (Hao et al. 2019; Singla et al. 2018; Tretbar et al. 2019). These processes can be abrogated with HDACi in RCC cells. The HDACi trichostatin-A (TSA) and butyrate suppressed TGFβ-induced EMT (Yoshikawa et al. 2007) and VPA prevented the TGFβ-dependent activation of the EMT-associated transcription factor SMAD4 (Mao et al. 2017). Such effects could be particularly relevant to tumors that are or become resistant to standard cancer drugs such as renal and lung tumors (Barbieri et al. 2017; Chang et al. 2019; Dietrich and Gerber 2016; Foy et al. 2017).
In light of the recent discussion on the impact of DNA repair and EMT for the metastatic process and its relation to acquired chemoresistance, and due to the influence of HDACi on the EMT of RCC cells, we set out to investigate if HDACi affect the expression of DNA repair proteins, including p53, and cell adhesion molecules. In addition, we investigated whether HDACi modulate TGFβ-induced cell plasticity and if the combination of HDACi and L-OHP or HU is effective against cancer cells.
HDACi suppress DNA repair protein expression
To assess if the observed alterations in protein expression result in DNA damage at a global level, we used the comet assay technique, which is also known as single cell electrophoresis assay. MS-275 evoked a time- and dose-dependent increase in DNA tails indicating DNA damage (Fig. 1b).
Due to these findings, we analyzed the phosphorylation of histone H2AX at serine-139 (ɣH2AX), which is a marker for replication stress, single-/double-strand breaks in DNA, and apoptosis (Nikolova et al. 2017; Rogakou et al. 1998). MS-275 and VPA, which we used as additional class I HDACi (Bradner et al. 2010; Müller and Krämer 2010), induced a time- and dose-dependent accumulation of ɣH2AX (Fig. 1c).
These data encouraged us to test for a dysregulation of DNA repair factors and DNA injury upon class I HDAC inhibition in primary human RCC cells (Mz-ccRCC2). We found that MS-275 reduced the expression of the recombinase RAD51, which is a key factor for HR-mediated DNA repair (Nikolova et al. 2017), in primary RCC cells (Fig. 1d). This was associated with a highly significantly increased level of DNA fragmentation, which is typically seen in cells undergoing apoptosis (Fig. 1e).
We conclude that the activity of class I HDACs is necessary for the expression of proteins that control genomic integrity. Accordingly, HDACi cause replicative stress and DNA damage in RCC cells.
Status of p53 and its regulation by HDACi in Renca cells
The tumor suppressor p53 determines the fate of HDACi-treated cells (Mrakovcic et al. 2019) and is relevant for DNA replication and HR (Gottifredi and Wiesmüller 2018; Klusmann et al. 2016). As we see replication stress/DNA damage in HDACi-treated cells (Fig. 1), we consequently analyzed the p53 status of and a putative regulation of p53 expression by HDACi in Renca cells. We searched major databases for the characterization of cell lines concerning the p53 status of the Renca cell line. One database provided information that Renca cells carry one R210C exchange in the DNA-binding domain of p53 (Zeitouni et al. 2017). Meta-data analysis of whole-exome sequencing data (WES data; accession no. PRJEB12925) for Renca cells from Mosely et al. (2017) revealed that these cells have one wild-type p53 allele and one allele with the R210C mutation (Supplementary Fig. S2).
These results prompted us to test for an accumulation of p53 in response to genotoxic stress, which indicates wild-type p53 (Schneider et al. 2010). When we analyzed the p53 expression of Renca cells in response to doxorubicin treatment, we found that total p53 protein expression was upregulated in response to this genotoxic stimulus (Fig. 2b).
It is known that HDACi decrease the expression of wild-type and mutant p53 through a transcriptional mechanism in pancreatic and colorectal cancer cells (Göder et al. 2018; Schäfer et al. 2017; Sonnemann et al. 2014; Stojanovic et al. 2017). When we measured the expression of p53 mRNA in Renca cells, we detected time- and dose-dependent effects of class I HDACi on p53 mRNA expression. A treatment of Renca cells with 1.5 µM MS-275 for 48 h led to a significant reduction of p53 mRNA to 46.5 ± 1.34% of control levels. This effect was more pronounced at higher doses of MS-275 (Fig. 2c).
Immunoblot analyses revealed that this reduction of the p53 mRNA translated into reduced levels of the p53 protein after 24-h incubations with MS-275 or VPA (Fig. 2d).
These data suggest that HDACi repress the expression of wild-type p53 and p53R210C in Renca cells.
HDAC inhibition does not promote chemoresistance
Since wild-type p53 is a tumor suppressor (Gottifredi and Wiesmüller 2018; Klusmann et al. 2016), its reduction by HDACi raises concerns that such drugs promote chemoresistance. Furthermore, HDACi-induced alterations in EMT factors (Kiweler et al. 2018) may promote the mesenchymal phenotype that is linked to chemoresistance (Fischer et al. 2015; Zheng et al. 2015). To address these concerns, we incubated Renca cells with combinations of HDACi, and the commonly used chemotherapeutics L-OHP, a DNA crosslinking agent that damages DNA directly, and HU, a ribonucleotide reductase inhibitor that can lead to DNA double-strand breaks secondary to a stalling of replication forks (Nikolova et al. 2017).
Moreover, the co-treatment with the HDACi and L-OHP or HU reduced the cellular adhesion factor E-cadherin, led to an accumulation of caspase-6, and induced caspase-3 cleavage (Fig. 3b). In addition, MS-275 and VPA slightly increased the accumulation of ɣH2AX that is caused by L-OHP (Fig. 3b) and this was linked to an abrogated induction of RAD51 by L-OHP and HU in Renca cells (Fig. 3c).
To further extend these findings, we analyzed cells derived from another notoriously chemoresistant tumor, non-small cell lung cancer (NSCLC). We incubated human H1299 lung cancer cells with L-OHP, because platinum compounds are a standard therapy for lung tumors (Foy et al. 2017; Gelsomino et al. 2019). In accordance with the data obtained in renal cancer cell lines, three structurally different HDACi (VPA, MS-275, LBH-589) promoted anti-proliferative effects of L-OHP against H1299 cells (Fig. 3d).
We sum up that HDACi do not attenuate drug sensitivity, but rather increase the efficacy of well-established anti-cancer drugs.
HDACi modulate the expression of adhesion factors
Proteomics also showed a reduction of integrin-β1 by HDACi (Fig. 4a), which is a key factor for cell adhesion to collagen and the colonization of tumor cells at distant tissues (Chua et al. 2010). Immunoblot analysis verified that 5 µM MS-275 and 5 mM VPA decreased the levels of integrin-β1 (Fig. 4c).
Since overexpression and somatic mutations in the tyrosine kinase ACK1 promote oncogenic signaling, cell proliferation, EMT, and chemoresistance (Chua et al. 2010; Jenkins et al. 2018; Mahajan and Mahajan 2015; Mahajan et al. 2018; Mahendrarajah et al. 2017; Yeh et al. 2012), we additionally tested whether HDACi have an impact on ACK1 in Renca cells. Western blot analyses showed that MS-275 and VPA attenuated ACK1 in a time- and dose-dependent manner (Fig. 4d).
These data demonstrate that HDACi dysregulate proteins that control the metastatic spread to and the colonization of secondary tissues by transformed cells.
HDACi suppress TGFβ-induced cellular plasticity
Cell adhesion and cellular plasticity are key steps in the progression from a local tumor to a distantly spread metastasis. TGFβ promotes EMT, metastasis, and drug resistance of cancer cells (Brabletz et al. 2018; Fidler and Kripke 2015; Hao et al. 2019; Nieto 2013; Zeisberg and Neilson 2009).
As a substitute, we analyzed N-cadherin levels in mammary epithelial NM18 cells, which represent a standard TGFβ-sensitive cell system (Deckers et al. 2006). Immunocytochemistry of NM18 cells revealed that VPA and MS-275 significantly reduced TGFβ-induced N-cadherin expression (Fig. 5b, c).
Furthermore, we show in primary Mz-ccRCC2 cells that HDACi abolish a basal expression of N-cadherin (Fig. 5d).
These results suggest that HDACi dysregulate the expression of cell adhesion factors and impair TGFβ-induced cell plasticity.
Our data illustrate that HDACi promote an induction of replicative stress and DNA damage in cultured cancer cells. This finding is coherent with previously published works, which illustrate that HDACs are required for the expression of factors mediating DNA damage, the recognition and repair of DNA lesions, and for scheduled origin firing (Conti et al. 2010; Miller et al. 2010; Nikolova et al. 2017; Noack et al. 2017; Wang et al. 2012; Wells et al. 2013). Our proteomics approach shows that the HDACi-induced accumulation of ɣH2AX in Renca cells is linked to a disturbed expression of proteins that stabilize DNA replication forks and contribute to DNA damage recognition and repair. For example, we demonstrate that VPA and MS-275 diminish the expression levels of RAD51, which is a key HR protein and a survival factor for cancer cells harboring damaged DNA. These findings are consistent with literature evidence from other tumor-derived cells (Göder et al. 2018; Krumm et al. 2016; Nikolova et al. 2017). The observed correlation between DNA damage and prolonged growth arrest of HDACi-treated Renca cells is coherent with DNA damage being a major trigger of cell cycle arrest (Kiweler et al. 2018; Lanz et al. 2019; Nikolova et al. 2017). Nevertheless, more investigations are necessary to assess the relative contribution of replication stress/DNA damage to HDACi-induced anti-proliferative effects. For instance, HDACi-induced alterations of proteins that control apoptosis and autophagy are further pathways through which HDACi might restrict tumor cell growth (Koeneke et al. 2015; Mrakovcic et al. 2019; Vancurova et al. 2018). Moreover, our proteome analyses show HDACi-induced alterations of proteins that control immune tolerance (Kiweler et al. 2018), raising the possibility that HDACi combat tumors through immune modulation.
Despite being a frequently used model system to analyze RCC in vitro and in vivo as syngeneic mouse model (Kiweler et al. 2018), the p53 status of Renca cells was undefined. While wild-type p53 is a short-lived protein, the majority of mutations in p53 are missense mutations that lead to the stable expression a p53 protein variant with a prolonged half-life (Conradt et al. 2012; Freed-Pastor and Prives 2012). We observed a strongly diminished expression of total p53 protein in Renca cells in comparison to a cell line expressing a defined mutant p53 isoform. This lower expression of p53 in Renca cells suggests its wild-type status. Our finding of an accumulation of p53 in doxorubicin treated Renca cells supports this notion. Nonetheless, one allele of p53 carries an ill-defined R210C exchange (Zeitouni et al. 2017). In general, p53 wild-type expression in Renca cells would correspond to the majority of cells derived from common renal cancers. TP53 mutation rates in this disease are exceptionally low in comparison to other cancer types, with 2.5% for renal papillary-cell carcinoma and 2.4% for renal clear-cell carcinoma (Wang et al. 2017).
Since wild-type p53 can suppress tumorigenesis (Gottifredi and Wiesmüller 2018; Klusmann et al. 2016), the reduction of p53 in HDACi-treated Renca cells appears to be counterintuitive with the anti-proliferative effects of HDACi. However, p53 might not be inactivated and its reduction by HDACi is not complete. There is, for example, an accumulation of p21, which is positively regulated by p53, and a repression of survivin, which is negatively regulated by p53 in HDACi-treated Renca cells (Kiweler et al. 2018). Apparently, the reduction of total p53 may not necessarily lead to a suppression of p53 target gene regulation, because p53 is also activated by acetylation. For example, low and very active levels of acetylated p53 can drive the expression of its target genes and apoptosis upon replication stress and DNA damage in colorectal cancer cells (Brandl et al. 2012). On the other hand, we may also detect p53-independent growth arrest and cell death induction by HDACi in Renca cells, as seen in p53-negative colorectal cancer cells (Sonnemann et al. 2014). Moreover, replication stress triggers apoptosis and mitotic catastrophe after HDACi treatment despite a reduced expression of p53 and its target genes (Göder et al. 2018). One should additionally consider that there are even cases in which p53 antagonizes apoptosis induction (Barckhausen et al. 2014) and the HDACi-evoked loss of various DNA repair proteins including p53 may trigger cytotoxic DNA damage. In conclusion, the observed loss of p53 expression in HDACi-treated Renca cells is not linked to diminished cytotoxic responses or an induction of chemoresistance.
Two recent studies point out that the mesenchymal transition of transformed cells ties in with the resistance of pancreatic and breast cancer cells against DNA-damaging agents (Fischer et al. 2015; Zheng et al. 2015). So far, our data illustrate that HDACi themselves attenuate the expression of DNA repair and promote cell death of HU-treated RCC and NSCLC cultures. Studies including small groups of patients treated with HU and HDACi suggest that such combinatorial treatment might be successful (Bug et al. 2005; Müller and Krämer 2010). Undoubtedly, additional in vivo evidence is necessary to clarify the therapeutic validity of HDACi/HU combination treatment schedules. This also applies to the doses that can be achieved without significant toxicity in RCC patients. A recent report found that up to 1.51 µM MS-275 was achieved without gross toxicity in mice, but large variation of maximal plasma concentrations from 4 to 53.1 ± 92.4 and a half-life from 33.4 to 150 h occurred in humans (Connolly et al. 2017; Kurmasheva et al. 2019), indicating unexplained large patient-to-patient variability. The maximum-tolerated dose of VPA was reported to range, for example, from 50 mg/kg daily to 140 mg/kg/day, which is within the therapeutic serum concentrations of VPA from 0.35 to 0.7 mM (Bug et al. 2005; Münster et al. 2007; Phiel et al. 2001).
In addition to the effects of HDACi on replication stress/DNA damage, their impact on EMT needs to be investigated. Preceding work demonstrated a dysregulation of various proteins that control cell adhesion and migratory properties in RCC cells following class I HDAC inhibition (Kiweler et al. 2018) and we verify an HDACi-mediated downregulation of integrin-β1 in Renca cells. This finding is coherent with the literature that reports an inhibition of integrin-α/β expression and their downstream signaling pathways in HDACi-treated RCC cells (Jones et al. 2009b). However, a plethora of additional factors determines EMT in HDACi-treated cells. For example, we see that HDACi decrease RhoA, which regulates EMT and interacts with HDACs (Mertsch and Krämer 2017). Likewise, proteomics suggests an HDACi-induced increase in the inducible transcription factor STAT1 in Renca cells and we found that such an accumulation of STAT1 contributes to apoptosis in HDACi-treated melanoma cells (Krämer et al. 2006). Another example is ACK1, for which evidence collected in various tumor types suggests a pro-tumorigenic role (Chua et al. 2010; Jenkins et al. 2018; Mahajan and Mahajan 2015; Mahajan et al. 2018). Furthermore, data collected with gastric and liver cancer cells show that ACK1 is overexpressed in primary clinical specimen and that ACK1 promotes the invasive capacity and EMT of such tumor cells through its direct activating effects on AKT kinases (Lei et al. 2015; Xu et al. 2015). AKT also contributes to the spread of RCC cells into bone tissue and a hyperstabilized, mutant ACK1 isoform promotes hallmarks of cancer in RCC cells (Chua et al. 2010). Hence, the reduction of ACK1 by HDACi could cause anti-proliferative, therapeutically relevant effects. We found that HDACi decrease ACK1 by a caspase-dependent mechanism in leukemic cells (Mahendrarajah et al. 2016) and we see that the decline in ACK1 is linked to apoptosis of Renca cells [(Kiweler et al. 2018) and this study]. Accordingly, the loss of ACK1 could occur through caspase-dependent degradation. Additional experiments are necessary to clarify whether ACK1 degradation in renal cancer presents a functional signal for cell growth reduction, cell death induction, and metastasis or poses a downstream marker of cell death.
TGFβ-induced EMT signaling promotes metastasis, chemoresistance, angiogenesis, and immune evasion of tumor cells (Hao et al. 2019). Although the treatment with TGFβ induces N-cadherin in other cell lines (Mikami et al. 2016; Zeisberg and Neilson 2009), Renca cells do not express the TGFβ receptor-II (Engel et al. 1999) and, therefore, fail to accumulate N-cadherin. Such a loss of TGFβ receptor II is also seen in 31 out of 62 RCC patients and correlates with a lower apoptotic index and statistically significant lower survival rates (Miyajima et al. 2003). The expression of E-cadherin, the cytoplasmic localization of β-catenin (Kiweler et al. 2018), and the absence of the mesenchymal marker N-cadherin (this work) verify that Renca cells remain epithelial cells independent of HDACi treatment. In contrast to this, HDACi suppress TGFβ-induced N-cadherin expression in mammary epithelial cells and HDACi decrease basal N-cadherin expression in primary human RCC cells. Our finding that both epithelial (Renca cells) and mesenchymal (Mz-ccRCC2 cells) respond with an induction of cell death to class I HDACi shows that such drugs can kill cells having either differentiation status. Further studies are underway to address if HDACi shift transformed cells to certain molecular signatures of one of the states and thereby eliminate them.
Taken together, our work illustrates that class I HDACi evoke DNA damage and suppress the metastasis-associated phenotype. These data suggest exploiting HDACi further for the treatment of cancer.
Materials and methods
Cell culture conditions and drugs
Fetal calf serum (FCS) was from Gibco Invitrogen Life Technologies, Darmstadt, Germany (catalogue numbers 102270/10270106, EU approved origin: South America, lots. 41Q8207K/42G8258K). Renca cells grow at 37 °C and 5% CO2 in RPMI medium (Sigma-Aldrich, Munich, Germany) supplemented with 10% FCS, 1% penicillin/streptomycin and 2% glutamine. NM18 and H1299-TO-p53 cells grow at 37 °C and 5% CO2 in DMEM medium (Sigma-Aldrich, Munich, Germany) supplemented with 10% FCS and 1% penicillin/streptomycin; 1% glutamine and 5 µg/mL insulin were added to NM18 cells. Renca cells were obtained from Prof. W. Wels, GSH Frankfurt/Main, Germany (derived from a spontaneously developed renal cortical adenocarcinoma in a male Balb/c mouse; ATCC® CRL-2947™). PPT-5436 were developed in the group of one of the coauthors (G.S.). These are a low passage cell line from primary pancreatic tumors of a PTF1a/p48ex1Cre/+;LSL-KRASG12D/+;LSL-p53R172H+/+;LSL-R26Tva−lacZ/− mouse. H1299-TO-p53 were given to us by Dr. G. Rohaly, HPI, Hamburg, Germany. These are a derivative of NCI-H1299 (ATCC® CRL-5803™) cells, which are epithelial cells from a metastatic site lymph node of a lung carcinoma of a male patient. NM18 cells are a subclone of NMuMG cells (ATCC® CRL-1636™), which were isolated by their strong response to TGFβ by Deckers et al. (2006). NMuMG cells are epithelial mammary gland cells from a Namru strain mouse. Mz-ccRCC2 renal tumor cells were isolated as described in Haber et al. (2015) from tumor specimens that were obtained shortly after nephrectomy. Tumor tissue was dissociated mechanically and with 1 mg/ml collagenase II, pressed through a cell strainer (70 μm) and centrifuged under sterile conditions. The obtained cells were first cultured in AmnioMAX C100 Basal Medium including AmnioMAX C100 Supplement (Gibco, Life Technologies, Darmstadt, Germany). After the first passage, they were transferred to DMEM medium supplemented with 10% FCS and 1% penicillin/streptomycin and cultured at 37 °C and 5% CO2; i.e., conditions as for Renca cells. The epithelial origin was confirmed by immunohistochemical cytokeratin staining. For the experiments, the cells were used in passages 2–8. No commonly mischaracterized cells were used and cells were tested free of mycoplasma every 4–8 weeks. TGFβ and VPA were purchased from Sigma-Aldrich, Germany. Entinostat was purchased from Selleckchem, Germany.
Database search for p53 in Renca cells
To search for information on the p53 status of Renca cells, we considered the International Agency for Research on Cancer TP53 Database (IARC, https://p53.iarc.fr/), the Catalogue Of Somatic Mutations In Cancer (COSMIC, https://cancer.sanger.ac.uk/cosmic), the Cancer Cell Line Encyclopedia (CCLE, https://portals.broadinstitute.org/ccle), the American Type Culture Collection (ATCC, https://www.atcc.org), the TP53 website cell line compendium (https://p53.fr), and Charles River (Zeitouni et al. 2017). Reanalysis of WES data is based on accession no. PRJEB12925 (Mosely et al. 2017).
Protein lysates, Western blot, densitometric analysis, and antibodies
We have recently summarized the Western blot method used to collect data shown here (Stojanovic et al. 2017). Data acquisition was performed with the Odyssey Infrared Imaging System (Licor), using IRDye® 680RD-coupled or IRDye® 800CW-coupled secondary antibodies. Immunoblots are representative of at least two independent experiments. The following antibodies were used: Cell Signaling Technology (Frankfurt/Main, Germany): cleaved caspase-3/#9661, caspase-6/#9762, E-cadherin/#3195; Enzo Life Sciences (Lörrach, Germany): HSP90/ADI-SPA-830-F; Merck Millipore (Darmstadt, Germany): AB1952; Santa Cruz Biotechnology (Heidelberg, Germany): ACK1/sc-28336, pS139-H2AX/sc-101696 (ɣH2AX), β-actin/sc-47778, P53/sc81168; Abcam: RAD51/ab63801; BD Bioscience (Heidelberg, Germany): N-cadherin/BD610921.
Cell cycle and cell death analysis
Cells were harvested with trypsin/EDTA and fixed with 80% ethanol. Samples were then stored at − 20 °C for at least 2 h. Thereafter, cells were incubated with RNAse A (Carl Roth; final concentration 20 µg/mL) for 1 h at RT and stained with propidium iodide (PI) (Sigma-Aldrich; final concentration 16.5 µg/mL) for 10 min on ice. Annexin V/PI staining was performed according to the manufacturer’s instructions with Annexin V-FITC (Becton Dickinson) and PI at RT for 15 min in the dark. Cells were subjected to flow cytometry with the FACSCanto Flow Cytometer (BD Biosciences). Data were analyzed with the FACSDIVA™ Software (BD Biosciences).
Immunofluorescence to detect N-cadherin
NM18 cells were seeded on chamber coverslips (µ-Slide 8-well, iBidi) and treated with 1.5 mM VPA or 5 µM MS-275, 5 ng/mL TGFβ, or combinations thereof for 48 h. Cells were washed thrice with phosphate-buffered saline (PBS), and fixed for 20 min in 4% paraformaldehyde. After cell permeabilization using 0.1% Triton X-100 in PBS for 10 min, cells were stained for 1 h at RT using α-N-cadherin antibody (BD Bioscience) 1:100 in PBS + 1% FCS. After incubation, slides were washed three times for 5 min in PBS and incubated for 1 h at RT with α-mouse Cy3 secondary antibody (Dianova). DNA/cell nuclei were visualized by staining with 0.5 µg/mL Hoechst-33258 (Sigma-Aldrich). Observation, image acquisition, and analysis of stained cells were performed using AxioVert 200 M, a digital AxioCam CCD camera and Axiovision software (Carl Zeiss, Jena, Germany).
Proteomics and pathway analysis
NuPAGE® LDS Sample Buffer (1×) supplemented with 100 µM DTT was added to treated cells and controls. Lysates were subjected to mass spectrometry (Dejung et al. 2016) for protein detection. Proteins were run on gels and digested with mass spectrometry-grade trypsin (Sigma-Aldrich, St. Louis, Missouri) and purified with a StageTip. Data were analyzed with Maxquant v220.127.116.11 using LFQ, ENSEMBL GRCm38 peptide database (57,751 entries), and custom R scripts. (Dejung et al. 2016). Full proteomics data are available upon scientific request. Preceding statistical analysis by R software version 3.2 using unpaired t test (two conditions) or one-way ANOVA (multiple conditions) resulted in an individual set of significantly regulated proteins. The background set was composed of all proteins successfully quantified in the experiment (5812 proteins). For each set of significantly regulated proteins, three hypergeometric tests (for biological processes, for molecular functions, and for cellular components) were performed using the R package “GOstats”. By this means, it was determined if the GO terms that were associated with significantly changing genes were over-represented over the defined background (Falcon and Gentleman 2007). For each protein listed, the Entrez gene ID was obtained using the annotation R package “BiomaRt”; see also cells (Kiweler et al. 2018).
Determination of mRNA transcripts
The primer sequences that we used to determine p53 transcript numbers were AGAGACCGCCGTACAGAAGA (forward)/CTGTAGCATGGGCATCCTTT (reverse); β-actin: GTCGAGTCGCGTCCACC (forward)/GTCATCCATGGCGAACTGGT (reverse). Isolation of total RNA was carried out by means of the RNeasy Mini-Kit (Qiagen, Hilden, Germany). To quantify the relative gene expression by ∆∆-Ct method, 20 ng cDNA and 100 nM primer mix (MWG Biotech, Ebersberg, Germany) were mixed together with the Power SybrGreen master mix and applied to the StepOnePlus real-time PCR device (Applied Biosystems/Thermo Fischer, Frankfurt/Main, Germany). Primer efficiencies were previously determined by the formula 10(−1/slope). All experiments were carried out in technical and biological triplicates.
One-way ANOVA/two-way ANOVA were used as indicated for the respective experiments (GraphPad Prism Vers.6.01) and corrected for multiple testing using Dunnett’s test or Bonferroni’s multiple comparisons test. Two-paired t test with Welch’s correction was used when not more than two conditions were compared and for proteomics as comparison with untreated samples. p values are indicated in the legends.
Open Access funding provided by Projekt DEAL. We thank Franziska Müller, Christina Brachetti, and Andrea Piée-Staffa (ITOX Mainz, Germany) for excellent technical support; Mandy Beyer (ITOX Mainz, Germany) for help with statistics; Dr. Mario Dejung (IMB Mainz, Germany) for help with proteomics. We are indebted to Prof. W. Wels, GSH Frankfurt/Main, Germany, Dr. G. Rohaly, HPI, Hamburg, Germany, and Deckers and colleagues, University Medical Center, Leiden, The Netherlands, for cell lines. This study was mainly supported by grants to OHK from the Wilhelm Sander-Stiftung (#2010.078). The group of OHK is additionally supported by the Deutsche Forschungsgemeinschaft (#KR2291/7-1/8-1/9-1), funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – Project-ID 393547839 – SFB 1361, and received intramural funding from the NMFZ Mainz and the University Medical Center Mainz.
Compliance with ethical standards
Conflict of interest
The authors declare no conflict of interest.
- Deckers M, van Dinther M, Buijs J et al (2006) The tumor suppressor Smad4 is required for transforming growth factor beta-induced epithelial to mesenchymal transition and bone metastasis of breast cancer cells. Cancer Res 66:2202–2209. https://doi.org/10.1158/0008-5472.CAN-05-3560 CrossRefPubMedGoogle Scholar
- Münster P, Marchion D, Bicaku E et al (2007) Phase I trial of histone deacetylase inhibition by valproic acid followed by the topoisomerase II inhibitor epirubicin in advanced solid tumors: a clinical and translational study. J Clin Oncol 25:1979–1985. https://doi.org/10.1200/JCO.2006.08.6165 CrossRefPubMedGoogle Scholar
- Zeitouni B, Tschuch C, Davis JM et al (2017) Abstract 1840: whole-exome somatic mutation analysis of mouse cancer models and implications for preclinical immunomodulatory drug development. Cancer Res 77(13 Suppl):Abstract nr 1840. doi: 10.1158/1538-7445.AM2017-1840Google Scholar
Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.