MYC directs transcription of MCL1 and eIF4E genes to control sensitivity of gastric cancer cells toward HDAC inhibitors

Abstract

Histone deacetylases (HDACs) control fundamental physiological processes such as proliferation and differentiation. HDAC inhibitors (HDACi) induce cell cycle arrest and apoptosis of tumor cells. Therefore, they represent promising cancer therapeutics that appear particularly useful in combination therapies. Although HDACi are tested in current clinical trials, the molecular mechanisms modulating the cellular responses toward HDACi are incompletely understood. To gain insight into pathways that limit HDACi efficacy in gastric cancer, we treated a panel of gastric cancer cells with the clinically relevant HDACi suberoylanilide hydroxamic acid (SAHA). We report that higher expression levels of the anti-apoptotic BCL2 family members MCL1 and BCL_XL were detectable in cells with high inhibitory concentration 50 (IC₅₀) values for SAHA. Using RNAi, we show that MCL1 and BCL_XL lower the efficacy of SAHA. To find strategies to interfere with MCL1 and BCL_XL expression, we investigated molecular regulation of both proteins. We show that specific siRNAs against c-MYC as well as pharmacological inhibition of this cancer-relevant transcription factor reduced MCL1 and BCL_XL expression. Subsequently, we observed an increase in SAHA efficacy. Our data furthermore demonstrate that two different molecular mechanisms are responsible for the modulation of these factors. Whereas c-MYC controls transcription of MCL1 directly, regulation of BCL_XL was due to c-MYC’s capability to regulate the eIF4E gene, which encodes a rate-limiting factor of eukaryotic translation. Our data reveal a new molecular mechanism for how c-MYC controls cell autonomous apoptosis and provide a rationale for a concerted inhibition of HDACs and c-MYC in gastric cancer.

Keywords: :

• BCL_XL
• HDAC
• MCL1
• apoptosis
• c-Myc
• eIF4E
• gastric cancer

Introduction

Histone deacetylases (HDACs) are divided into four distinct classes according to phylogenetic analysis, sequence homology and function. The Zn²⁺-dependent enzymes are grouped into class I (HDAC1, 2, 3 and 8), class II (HDAC4, 5, 6, 7, 9 and 10) and class IV (HDAC11) enzymes, while the NAD⁺-dependent sirtuins (SIRT 1–7) represent class III.¹ HDACs are involved in the control of proliferation, apoptosis, differentiation, migration and angiogenesis of cancer cells.² Thus, they are promising therapeutic targets.^3-5 Furthermore, a recent study demonstrates the sensitivity of tumor-initiating cells toward HDAC inhibitors (HDACi), pointing to a good therapeutic potential in solid tumors.⁶ The pan-HDACi suberoylanilide hydroxamic acid (SAHA; Vorinostat) and the cyclic peptide depsipeptide (Romidepsin; FK-228), which preferably inhibits class I HDACs, have been approved by the FDA for the treatment of cutaneous T-cell lymphoma (CTCL).⁷ Although HDACi are currently being tested in diverse clinical trials for solid cancers and leukemias, mechanisms controlling the efficacy of HDACi and predictive markers for these agents are still not fully identified.

Class I HDACs are often overexpressed in solid tumors, and high expression of these enzymes is also found in gastric cancer.^8-12 Of note, the concomitant nuclear expression of HDAC 1, 2 and 3 correlates strongly with reduced 3-y survival rates.¹¹ Furthermore, high HDAC1 expression is a predictor of reduced survival in a group of gastric cancer patients responding to a 5-Fluorouracil (5-FU)/platinum-based neoadjuvant chemotherapy.¹² It has been postulated that the levels of class I HDACs predicts the responsiveness of gastric cancer toward HDACi.¹¹ However, recent observations suggest that neither HDAC1 nor HDAC2 expression are convincing markers for HDACi sensitivity of gastric cancer cell lines.¹²

Results

BCL_XL and MCL1 expression levels inversely correlate with the responsiveness of gastric cancer cells toward HDACi.

To elucidate mechanisms controlling the sensitivity of gastric cancer cells exposed to HDACi, we used a panel of murine (mGC4 and mGC8) and human (ST23132, KatoIII, HSC45-M2, ST2957, NUGC4 and MKN45) gastric cancer cell lines. These were treated with increasing concentrations of SAHA, and IC₅₀ values were determined for each cell line (Table 1 and Fig. S1). SAHA IC₅₀ values ranged between 1 ∝M and 3.2 ∝M. This finding is consistent with recent observations.^12,13 One important response of tumor cells toward current cancer therapies is the induction of programmed cell death, which is regulated by the fine-tuned interplay of BCL2 family members.¹⁴ In this family, pro-survival factors, such as BCL-2, BCL_XL or MCL1, counteract the action of death inducers, which are the multidomain pro-death proteins BAX and BAK or the BH3-only proteins, such as BIM, PUMA, NOXA or BID.¹⁴ As shown in Figure 1A, BCL_XL and MCL1 protein levels correlated with SAHA IC₅₀ values. This finding provided us a first hint that both proteins modulate the response of gastric cancer cells to HDACi.

Table 1. Response of gastric cancer cell lines to SAHA

Cell line	IC50 (48 h) (μM SAHA)	CI (95%)
MGC8	1.0	0.7–1.3
ST23123	1.4	1.1–1.7
MGC4	1.5	1.2–1.8
HSC45-M2	2.3	1.9–2.8
KatoIII	2.4	1.8–3.3
ST2957	2.8	2.1–3.7
NuGC4	2.9	2.4–3.6
MKN45	3.2	2.6–3.9

Figure 1. BCL_XL restricts SAHA efficacy of gastric cancer cells. (A) western blot analysis of MCL1 and BCL_XL expression of the indicated gastric cancer cell lines. β-actin served as loading control. (B) ST23132 and MKN45 cells were treated with increasing doses SAHA for 24 h as indicated or were left as vehicle treated controls. Western blots detected expression of BCL_XL and β-actin (loading control). (C) ST2957 and MKN45 cells were transfected with the indicated siRNAs. 48 h after the transfection whole-cell lysates were prepared and western blots detected expression of BCL_XL, MCL1 and β-actin (loading control). (D) ST2957 and MKN45 were transfected with the indicated siRNAs. 48 h after the transfection cells were treated with SAHA as indicated for additional 24 h. Viability of cells was measured in MTT assays and viability vehicle treated controls were arbitrary set to 100% in order to compare the SAHA responses within each siRNA condition. Data are presented as mean and standard error of the mean (S.E.M). (Student’s t-test: * p < 0.05 vs. controls).

BCL_XL and MCL1 counteract the efficacy of HDACi.

To investigate regulation of BCL_XL by HDACi, we performed western blot analysis. As shown in Figure 1B, SAHA treatment downregulated BCL_XL protein expression in the SAHA-sensitive ST23132 cells (IC₅₀ 1.4 ∝M). No regulation of BCL_XL was detected in the more resistant MKN45 cells (IC₅₀ 3.2 ∝M), when SAHA concentrations below the IC₅₀ value were used (Fig. 1B). SAHA concentrations over the IC₅₀ value decreased BCL_XL expression also in the more resistant MKN45 and ST2957 (IC₅₀ 2.8 ∝M) cells (Fig. 2D). We used RNAi to analyze the contribution of BCL_XL for cellular responses toward SAHA. Knockdown of this protein in ST2957 cells and MKN45 cells was specific and did not influence MCL1 expression (Fig. 1C). In both cell lines, a reduction of BCL_XL decreased the viability of both cell lines upon treatment with SAHA (Fig. 1D). Thus, BCL_XL limits the pro-apoptotic potency of HDACi against gastric cancer cells.

Figure 2. MCL1 restricts SAHA efficacy of gastric cancer cells. (A) ST23132 and MKN45 cells were treated for 24 h with increasing doses SAHA as indicated or left as vehicle treated controls. Western blots detected expression of MCL1 and β-actin (loading control). (B) ST2957 and MKN45 cells were transfected with the indicated siRNAs. Forty-eight hours after transfection whole-cell lysates were prepared and western blots detected expression of MCL1, BCL_XL and β-actin (loading control). (C) ST2957 and MKN45 cells were transfected with the indicated siRNAs. 48 h after the transfection cells were treated with SAHA as indicated for additional 24 h. Viability of cells was measured in MTT assays and viability of vehicle treated controls were arbitrary set to 100% in order to compare the SAHA responses within each siRNA condition. Data are presented as mean and standard error of the mean (S.E.M). (D) ST2957 and MKN45 cells were transfected with the indicated siRNAs. For MCL1, the MCL1–1 siRNA was used. siRNA amount was kept constant using control siRNA. Forty-eight hours after transfection cells were treated with SAHA for additional 24 h as indicated or were left as untreated controls. Western blots detected expression of MCL1, BCL_XL and β-actin (loading control). (E) ST2957 and MKN45 cells were transfected with the indicated siRNAs. For MCL1, the MCL1–1 siRNA was used. siRNA amount was kept constant using control siRNA. Forty-eight hours after transfection cells were treated with SAHA for additional 24 h as indicated or were left as untreated controls. Cells were stained with Propidium iodide (PI) and FITC-labeled anti-Annexin V. Depicted is the Annexin V positive fraction (early apoptosis = Annexin V⁺/PI^- and late apoptosis = Annexin V⁺/PI⁺). (Student’s t-test: * p < 0.05 vs. controls).

HDACi have been reported to induce apoptosis of leukemic cells without decreasing MCL1 protein expression.¹⁵ We observed that SAHA treatment neither altered MCL1 expression in ST23132 nor in MKN45 gastric cancer cells (Fig. 2A). Also, SAHA concentration over the IC₅₀ value failed to downregulate MCL1 protein expression (Fig. 2D). Nevertheless, attenuation of MCL1 levels was reported to enhance HDACi-induced cytotoxicity in leukemic cells.¹⁵ This notion prompted us to test how MCL1 affects the HDACi responses of our cellular systems. Knockdown of MCL1 with RNAi in ST2957 and MKN45 cells was specific and did not influence BCL_XL expression (Fig. 2B). The reduction of MCL1 protein expression slightly increased the SAHA-induced therapeutic response of ST2957 and MKN45 cells (Fig. 2C). The fact that this observation reached statistical significance argues that MCL1 counteracts HDACi-induced death programs in gastric cancer cells. To corroborate our notion that BCL_XL and MCL1 limit SAHA efficacy, we simultaneously depleted both BCL2 family members in ST2957 and MKN45 cells (Fig. 2D). Simultaneous knockdown of BCL_XL and MCL1 distinctly increased SAHA-induced apoptosis compared with the knockdown of each anti-apoptotic protein alone (Fig. 2E).

MYC controls BCL_XL, MCL1 and the HDACi response.

Having found that MCL1 and BCL_XL counteract the HDACi response, we deciphered pathways that control their expression in gastric cancer cells. After excluding NFκB and STAT3 as main MCL1 and BCL_XL regulators in ST2957 and MKN45 gastric cancer cells (Fig. S2), we investigated c-MYC, since gene expression patterns regulated by c-MYC are associated with therapeutic resistance.^16-19 Knockdown of c-MYC distinctly decreased MCL1 protein and mRNA levels in ST2957 and MKN45 gastric cancer cells (Fig. 3A and B). While BCL_XL protein expression was distinctly reduced in c-MYC siRNA transfected cells (Fig. 3A), BCL_XL mRNA expression was not altered significantly (Fig. 3B). These observations argue for a non-genomic function of c-MYC on BCL_XL.

Figure 3. c-MYC controls MCL1 and BCL_XL expression of gastric cancer cells. (A) ST2957 and MKN45 cells were transfected with the indicated siRNAs. 48 h after the transfection whole-cell lysates were prepared and western blots detected expression of c-MYC, BCL_XL, MCL1 and β-actin (loading control). (B) ST2957 and MKN45 cells were transfected with the indicated siRNAs. 48 h after the expression of c-MYC, MCL1 and BCL_XL mRNAs were determined by quantitative RT-PCR using cyclophilinA mRNA as reference. (C) ST2957 and MKN45 cells were transfected with the indicated siRNAs. 48 h after the transfection cells were treated with SAHA as indicated for additional 24 h. Viability of cells was measured in MTT assays and viability of vehicle treated controls were arbitrary set to 100% in order to compare the SAHA responses within each siRNA condition. Data are presented as mean and standard error of the mean (S.E.M). (Student’s t-test: * p < 0.05 vs. controls).

Knockdown of c-MYC in ST2957 and MKN45 cells significantly reduces viability of cells (Fig. S3), confirming a role for c-MYC in tumor maintenance of gastric cancer cells.^20,21 Remarkably, the cellular responses toward SAHA in the rather HDACi-resistant cells (ST2957 and MKN45) was markedly increased upon attenuation of c-MYC (Fig. 3C). Hence, this factor appears to counteract the effects of HDACi on gastric cancer cells. To validate the RNAi results, we treated ST2957 and MKN45 cells with the c-MYC inhibitor (myc-I) 10058-F4, which interferes with c-MYC/MAX heterodimerization.²² Dose-dependent downregulation of MCL1 and BCL_XL was observed in both cell lines (Fig. 4A). Myc-I as well as SAHA treatment induced apoptosis of ST2957 and MKN45 cells, which was further increased by the myc-I and SAHA co-treatment (Fig. 4B).

Figure 4. MCL1 is a direct c-MYC target in gastric cancer cells. (A) ST2957 and MKN45 cells were treated with the MYC inhibitor (myc-I) for 24 h. Western blots detected the expression of BCL_XL, MCL1 and β-actin (loading control). (B) ST2957 and MKN45 cells were treated with the MYC inhibitor (myc-I), SAHA or the combination of both for 24 h as indicated or were left as an untreated control. Cells were stained with Propidium iodide (PI) and FITC-labeled anti-Annexin V. Depicted is the Annexin V positive fraction (early apoptosis = Annexin V⁺/PI^- and late apoptosis = Annexin V⁺/PI⁺). (Student’s t-test: * p < 0.05). (C) and (D) ST2957 and MKN45 cells were treated with the MYC inhibitor for 24 h or were left as an vehicle treated control. ChIP analysis revealing the binding of c-MYC (C) or the RNA Polymerase II (D) to the E-box of the MCL1 promoter or an 13 kb 3′ control.

To test the direct binding of c-MYC to the MCL1 promoter, we used quantitative chromatin immunoprecipitation (qChIP). In MKN45 as well as in ST2957 cells, c-MYC was recruited to the E-box in the promoter of the MCL1 gene (Fig. 4C). As expected, binding of c-MYC was reduced after treatment of both cell lines with the myc-I (Fig. 4C). Furthermore, c-MYC binding correlated with the binding of RNA polymerase II to the MCL1 promoter, which was reduced by the myc-I treatment in both cell lines (Fig. 4D). Together, the ChIP data determine the MCL1 gene as a c-MYC target gene in gastric cancer cells.

c-MYC-induced transcription of eIF4E regulates BCL_XL expression in gastric cancer cells.

c-MYC can regulate protein translation, e.g., via control of genes coding for translation initiation factors, such as eIF4E.^23-25 Accordingly, we observed downregulation of eIF4E expression in c-MYC siRNA-transfected ST2957 and MKN45 cells at the protein (Fig. 5A) and mRNA levels (Fig. 5B). Furthermore, reduced eIF4E expression was also detected in myc-I treated gastric cancer cells (Fig. 5C). To prove direct transcriptional regulation of eIF4E by c-MYC, we again used qChIP. Analogously to the MCL1 promoter, we detected c-MYC recruitment to the E-box in the proximal eIF4E promoter. As expected, recruitment of this transcription factor was impaired after myc-I treatment (Fig. 5D). Moreover, reduced binding of c-MYC to the eIF4E promoter was accompanied by a decline in RNA polymerase II recruitment to this promoter in ST2957 and MKN45 cells (Fig. 5E). These data show transcriptional control of eIF4E by c-MYC in gastric cancer cells.

Figure 5. MYC controls eIF4E transcription to regulate BCL_XL expression in gastric cancer cells. (A) ST2957 and MKN45 cells were transfected with the indicated siRNAs. 48 h after the transfection whole-cell lysates were prepared and western blots detected expression of c-MYC, eIF4E and β-actin (loading control). (B) ST2957 and MKN45 cells were transfected with the indicated siRNAs. 48 h after the expression of the eIF4E mRNA were determined by quantitative RT-PCR using cyclophilinA mRNA as reference. (C) ST2957 and MKN45 cells were treated with the MYC inhibitor for 24 h. Western blots detected the expression of eIF4E and β-actin (loading control). (D) and (E) ST2957 and MKN45 cells were treated with the MYC inhibitor for 24 h or were left as an vehicle treated control. ChIP analysis revealing the binding of c-MYC (D) or the RNA Polymerase II (E) to the E-box of the eIF4E promoter or an cyclophilin A 3′ control. (F) ST2957 and MKN45 cells were transfected with the indicated siRNAs. 48 h after the transfection whole-cell lysates were prepared and western blots detected expression of eIF4E, BCL_XL and β-actin (loading control).

To corroborate regulation of BCL_XL expression by eIF4E, we used two specific eIF4E siRNAs. As shown in Figure 5F, transfection of ST2957 and MKN45 cells with such siRNAs indeed blocked BCL_XL expression. From these findings, we conclude that a c-MYC-eIF4E axis controls BCL_XL translation in gastric cancer cells.

Since expression of c-MYC is downregulated by HDACi in various cell types,^26-28 we tested c-MYC expression upon SAHA treatment. Interestingly, although the c-MYC protein expression was downregulated in ST2957 and MKN45 cells upon SAHA treatment, we observed no regulation of the herein described c-MYC target genes MCL1 and eIF4E (Fig. 6A), which may argue that both genes are actively repressed upon inhibition of c-MYC by a HDAC-dependent repressor complex. This observation might also explain the rather small effect of the simultaneous treatment of both cell lines with the myc-I and SAHA toward apoptosis induction (Fig. 4B) compared with the distinct increase in the therapeutic efficacy of SAHA after c-MYC depletion via RNAi (Fig. 3C). Consequently, we concluded that drug scheduling composed out of a 24 h myc-I pretreatment, to assure downregulation of MCL1 and eIF4E, followed by a combination of myc-I and SAHA might be more potent. Indeed, we observed a distinct increase of the therapeutic SAHA efficacy in ST2957 and MKN45 cells (Fig. 6B).

Figure 6. MYC expression is regulated by HDACi in gastric cancer cells. (A) ST2957 and MKN45 cells were treated with increasing doses SAHA for 24 h as indicated or were left as vehicle treated controls. Western blots detected expression of c-MYC, MCL1, eIF4E and β-actin (loading control). (B) ST2957 and MKN45 cells were pre-treated with myc-I as indicated or were left as an untreated control. After 24 h cells were again treated with myc-I or with the combination of the myc-I and SAHA for additional 24 h. Untreated cells remained untreated. Viability of cells was measured in MTT assays and viability of vehicle treated controls were arbitrary set to 100%. Data are presented as mean and standard error of the mean (S.E.M). (Student’s t-test: * p < 0.05 vs. controls). (C) Schematic presentation of how gene expression signatures induced by c-MYC restrain the efficacy of HDACi against gastric cancer cells.

Discussion

Most patients with gastric cancer are diagnosed with an advanced tumor stage, in which systemic chemotherapy remains the key treatment option. Compared with best supportive care, chemotherapy in advanced gastric cancer improves median survival.²⁹ Nonetheless, chemotherapy remains palliative, and 5-y survival rates are below 10%. HDACi are currently tested in advanced gastric cancer (www.clinicaltrials.gov), but molecular mechanisms and markers relevant for the efficacy of HDACi in gastric cancer cells are largely unknown. Here we provide evidence that c-MYC counteracts HDACi efficacy in gastric cancer cells by regulating the expression of the anti-apoptotic BCL2 family members BCL_XL and MCL1. Our data illustrate two independent mechanisms by which c-MYC protects gastric cancer cells from HDACi. First, c-MYC directly controls transcription of MCL1 and, second, regulation of BCL_XL protein expression was due to c-MYC’s capability to control the eIF4E gene and thereby translation of the BCL_XL mRNA (Fig. 6C).

Apoptosis is one major cellular response of cancer cells treated with HDACi and was also observed in SAHA-treated gastric cancer cells.³⁰ Induction of this type of cell death predicts a beneficial therapeutic success of these agents.³¹ Consistently, we observed that the pro-survival BCL2 family members BCL_XL and MCL1 limit the efficacy of these compounds in a cell-based model of gastric cancer. Furthermore, our results are in line with in vivo data demonstrating that pro-survival BCL2 family members limit the efficacy of HDACi.³¹ High MCL1 expression has been detected in about 70% of patients with gastric cancer, and expression of MCL1 is an independent prognostic factor.³² Furthermore, lowering MCL1 expression by an antisense oligonucleotide in NCI-N87 gastric cancer cells increased sensitivity toward docetaxel and cisplatin.³³ MCL1 hence appears relevant for gastric tumorigenesis and appears as a valid chemotherapeutic marker protein. In addition to MCL1, BCL_XL is overexpressed in gastric cancers, and depletion of BCL_XL leads to an increased sensitivity of such cells toward extrinsic apoptosis.^34-36 Therefore, our observation that c-MYC controls expression of important anti-apoptotic BCL2 family members might argue that c-MYC is a common mediator of therapeutic resistance of gastric cancer cells. This conclusion is supported by recent microarray data, which point to a role of c-MYC in primary and secondary resistance of gastric cancer patients toward conventional chemotherapy.^37,38

The oncogene and basic helix-loop-helix leucine-zipper transcription factor c-MYC heterodimerizes with MAX. These bind to cis-acting elements, the so called E-boxes (CAC GTG) of cancer-promoting genes.³⁹ Inhibition of c-MYC has recently been demonstrated to be well tolerated in vivo, and inhibition of c-MYC eradicated tumors in a murine Kras-dependent lung cancer model.⁴⁰ Therefore, several approaches to inhibit c-MYC are currently under development.⁴¹ c-MYC overexpression has frequently been observed in gastric cancer and is associated with therapeutic resistance of certain tumor entities.^16-19,42 A high-affinity c-MYC binding site in the MCL1 promoter has also been found in U937 histiocytoma cells.⁴³ Consistent with the downregulation of MCL1 protein and mRNA after the knockdown of c-MYC, we detected binding of c-MYC to the MCL1 promoter in gastric cancer cell lines, demonstrating direct transcriptional regulation.

In addition to c-MYC’s direct transcriptional mode of action, this factor can regulate protein expression at the level of translation. c-MYC can control genes coding for translation initiation factors, such as eIF4E.^23-25 Dysregulated translational control can contribute to diverse human diseases, including cancer.⁴⁴ Translational initiation is regulated by the eIF4F complex, which is composed of the RNA helicase eIF4A, the scaffolding protein eIF4G and the cap-binding protein eIF4E, which is believed to be the rate-limiting protein in the eIF4F complex.^44,45 eIF4E harbors transforming characteristics in vivo and in vitro and was found to be overexpressed in various tumors, contributing to proliferation, survival, angiogenesis and metastasis of cancer cells.^44-46 Also in gastric cancer patients, high expression of eIF4E was correlated with vascular invasion and a worse outcome for patients.⁴⁷ Our data demonstrate that c-MYC controls the transcription of eIF4E in gastric cancer cells by direct binding to its proximal high affinity E-box. Since eIF4E knockdown results in a distinct downregulation of BCL_XL, our data argue that BCL_XL is regulated at the translational level in gastric cancer cells. These observations are consistent with observations from breast cancer models in which the knockdown of eIF4E reduced BCL_XL expression.⁴⁸ The observation that SAHA doses above the IC₅₀ value regulate BCL_XL expression in the cellular models investigated, without changing eIF4E expression, argues that HDACi control BCL_XL beyond the c-MYC-eIF4E axis.

The c-MYC antagonist of the MAD family of transcription factors are alternative binding partners of MAX, primarily acting as repressors.⁴⁹ For instance, MAD/MAX heterodimers repress transcription by the recruitment of an mSin3/HDAC containing co-repressor complex.⁵⁰ We detected downregulation of the c-MYC protein expression in the investigated gastric cancer cells upon HDACi treatment. In contrast, the c-MYC target genes MCL1 and eIF4E were not altered by HDACi treatment, which might argue for the involvement of an HDAC containing co-repressor complex. Such observations are important, since they argue for a specific drug regimes to assure optimal therapeutic efficacy. Consistently, inhibition of c-MYC before HDACi treatment revealed an increased efficacy. The observation that HDACi synergize with DNA damaging agents only in case of simultaneous application⁵¹ argues that each HDACi combination therapy posses its own application schedule. Although attractive, these possibilities clearly need more experimental evaluations.

In contrast to our data obtained with HDACi in gastric cancer cells, there is evidence that c-MYC can promote the efficacy of diverse chemotherapeutics and γ-irradiation in other tumor entities.⁵² While the molecular determinants that define the pro-therapeutic or therapeutic-resistance role of c-MYC are still unclear, these data clearly argue for the need to determine the molecular actions of novel therapeutic strategies in pre-clinical settings, especially considering that even novel targeted therapies are used in combinations with conventional therapeutics.

Here we have demonstrated that the anti-apoptotic BCL2 family members MCL1 and BCL_XL limit the biological effects HDACi exert against gastric cancer cells. Since MCL1 and BCL_XL are regulated by c-MYC in gastric cancer cells, c-MYC inhibition applied in combination with HDACi might be a rationally based therapeutic option for this type of cancer. Since c-MYC is included in a three gene predictor indicating poor prognosis for cisplatin/5-FU-treated gastric cancer patients and a c-MYC-controlled gene set is enriched in gastric cancer patients with acquired resistance toward cisplatin/5-FU treatment, c-MYC might limit the therapeutic response of gastric cancers beyond HDACi.^37,38

Material and Methods

Cell culture, reagents, siRNAs, transfection, viability assay and annexin V FACS analysis.

Suberoylanilide hydroxamic acid (SAHA; Vorinostat) was purchased from LC Laboratories, and the MYC inhibitor (myc-I, 10058-F4; (Z,E)-5-(4-Ethylbenzylidine)-2-thioxothiazolidin-4-one) was purchased from EMD Biosciences. Cell lines were cultivated in DMEM or RPMI medium supplemented with 10% fetal calf serum and 1% (w/v) penicillin/streptomycin (Invitrogen) as recently described in references 53–55. KatoIII were purchased from ATCC and MKN45 cells from DSMZ. The murine MGC4 and MGC8 cells were a kind gift from Dr. W. Zimmermann, HSC45-M2 cells a kind gift from Dr. Senekowitsch-Schmidtke and ST23123, ST2957 and NUGC4 cells were a kind gift from Dr. R. Mejías-Luque.⁵⁶ Double-stranded siRNAs were transfected with polyethylenimine (Sigma-Aldrich) at a final concentration of 50 nM as recently described in reference 54. siRNAs were purchased from Eurofins. Target sequences of the siRNAs used were: control siRNA 5'-CAG TCG CGT TTG CGA CTG G-3'; MCL1-1 si 5'-GAA ACG CGG TAA TCG GAC T 3'; MCL1–2 si 5'-CGC CGA ATT CAT TAA TTT A-3'; BCL_XL si 5'-GGA GAT GCA GGT ATT GGT G-3'; RelA si 5'-GAT CAA TGG CTA CAC AGG A-3'; STAT3 5'-CAT CTG CCT AGA TCG GCT A-3'; c-MYC-1 si 5'-CTT CTA CCA GCA GCA GCA G-3'; c-MYC-2 si 5'-GAA CAC ACA ACG TCT TGG A-3'; eIF4E-1 5'-si GGA CGA TGG CTA ATT ACA T-3'; eIF4E-2 si 5'-GGA TGG TAT TGT GCC TAT G-3'. Viability of the cells was measured using MTT-assays performed according to the manufacturer’s protocol (Roche Applied Science) and described recently in reference 53 and 54. To determine Annexin V, cells were stained with Propidium iodide (PI) and FITC-labeled anti-Annexin V using the FITC Annexin V Apoptosis Detection Kit I (BD Biosciences) and analyzed by FACS.⁵⁴

Total cell lysates and western blot.

Whole-cell lysates were prepared and western blots were performed as recently described in reference 53. The following antibodies were used: anti-MCL1 (Santa Cruz Biotechnology, sc-819), anti-BCL_XL (Cell signaling technology, #2762S), STAT3 (Cell signaling technology, #9132), RelA (Santa Cruz Biotechnology, sc-372), c-MYC (Cell signaling technology, #9402S), eIF4E (BD Biosciences, #610269) and β-actin (Sigma-Aldrich). Western blots were performed using Odyssey Infrared Imaging System (LI-COR Biosciences) as described in reference 53.

Quantitative reverse-transcriptase PCR.

Total RNA was isolated from pancreatic carcinoma cell lines using the RNeasy kit (Qiagen) following the manufacturer’s instructions. Quantitative mRNA analyses were performed as previously described using real-time PCR analysis (TaqMan, PE Applied Biosystems).⁵³ The following primers were used: cyclophilinA-fw 5'-ATG GTC AAC CCC ACC GTG T-3'; cyclophilinA-rv 5'-TCT GCT GTC TTT GGG ACC TTG TC-3'; c-MYC-fw 5'-AGC GAC TCT GAG GAG GAA CA-3'; c-MYC-rv 5'-CTC TGA CCT TTT GCC AGG AG 3'; MCL1-fw 5'-GCA TCG AAC CAT TAG CAG AAA G-3'; MCL1-rv 5'-AAA GCC AGC AGC ACA TTC-3'; BCL_XL-fw 5'-CCA CTT ACC TGA ATG ACC ACC TAG A-3'; BCL_XL-rv 5'-GCT GCA TTG TTC CCA TAG AGT TC-3'; eIF4E-fw 5'-ACA AGT CAG TCT GAA ACC ATC GAA C-3'; eIF4E-rv 5'-CTT CAT CCT CTT CGG CCA CTC CTC C-3'.

Quantitative chromatin immunoprecipitation.

Chromatin Immunoprecipitations were performed using magnetic SimpleChIP Enzymatic Chromatin IP Kit (Cell Signaling technology) according to the manufacturer’s protocol with the following antibodies: c-MYC (Cell signaling technology, #9402S), RNA Polymerase II (Santa Cruz Biotechnology, N-20) and rabbit IgG (Cell signaling technology) as control. c-MYC and RNA Polymerase II binding was determined by quantitative PCR.⁵⁷ ΔCt values for each sample were calculated using ΔCt = [Ct (sam ple) - Ct (input)]. ΔΔCt was calculated using ΔΔCt = [ΔCt (experimental sample) -ΔCt (negative control)]. Shown is the fold difference between the experimental sample and the negative control (IgG), which was determined as 2^-ΔΔCt. The following primers with similar amplification efficiencies were used: MCL1–587-fw 5'-TTA GGG TAG CAC GTG GAG CA-3'; MCL1–587-rv 5'-ACG GGG AAG GTT CAG TGA TGG-3'; eIF4E-75-fw 5'-TAC TCA CGC AGC CGC AGT C-3'; eIF4E-75-rv 5'-TCG CAC AAC CGC TCC AG-3'; MCL1–13 kb 3'-fw 5'-GCT GTG CTG AGA GGC CTG GG-3'; MCL1–13 kb 3'-rv AGC ACA CAA ACA TGC CGA CCC-3'; cyclophilin A 3' control-fw 5'-GCG ACC TTG AGG CCT GCG TT-3'; cyclophilin A 3'-control-rv 5'-CGG CTC TTC GGC CGT TGT CA-3'.

Statistical methods.

A two-tailed Student t-test was used to test statistical significance. p-values are indicated, and * denotes a p-value of at least < 0.05. Inhibitor concentration 50 (IC₅₀) values were calculated with GraphPad Prism4 software using a nonlinear regression model with classic equations as described in reference 53.

Acknowledgements

We thank Dr. R. Mejías-Luque, Dr. W. Zimmermann and Dr. R. Senekowitsch-Schmidtke for providing cell lines. The work was supported by KAAD, Bundesministerium für Bildung und Forschung, MoBiMed Programm (01 EZ 0802), Deutsche Krebshilfe (Grant 109265) and Deutsche Forschungsgemeinschaft (Grant SCHN 959/1–2).

Disclosure of potential Conflicts of interest

No potential conflicts of interest were disclosed.

Note

Supplemental material can be found at: www.landesbioscience.com/journals/cc/article/20008