Background
Human esophageal squamous cell carcinoma (ESCC) is one of the most frequently diagnosed carcinomas, ranked as the sixth leading cause of death from cancers worldwide. ESCC remains the most common histology and occurs at a very high frequency in China, South Africa, France and Italy [
1]. Although modest advances have been made in chemotherapy for esophageal cancer, ESCC is still one of the most aggressive types of cancer with a 5-year survival rate less than 15%. The underlying reasons for this disappointingly low survival rate remains to be greatly elucidated. Therefore, a better understanding of the molecular mechanisms of ESCC pathogenesis is expected to facilitate the development of novel therapies for this disease.
The
Mcl-1 is an antiapoptotic gene of the Bcl-2 family members. Mcl-1 is overexpressed in many human tumor specimens, including hepatocellular carcinoma [
2], pancreatic cancer [
3], prostate cancer [
4] and others [
5]. Overexpression of Mcl-1 was found in malignant melanoma compared to benign nevi and increased expression of Mcl-1 was also observed by comparing primary and metastatic melanoma samples utilizing a tissue microarray [
6]. In addition, frequent
Mcl-1 gene amplification was identified in lung, breast, neural and gastrointestinal cancers, through which cancer cells depend on the expression of this gene for survival [
7]. A survey of antiapoptotic Bcl-2 family member expression in breast, brain, colon, lung, ovarian, renal and melanoma cell lines revealed that
Mcl-1 mRNA is more abundant than Bcl-2 or Bcl-xL [
8]. These studies demonstrated that Mcl-1 plays a critical role in carcinogenesis and malignancy development in a broad range of human tumors, making it an attractive therapeutic target. However, the underlying mechanisms causing its elevation are not fully understood.
Expression of
Mcl-1 gene can be regulated at transcriptional level. Analysis of human
Mcl-1 gene 5′-flanking promoter regions for potential transcription factor binding sites revealed consensus sequences including STAT, SRE, Ets, Sp1, CRE-BP [
9]. Multiple intracellular signaling pathways and transcription factors have been confirmed to influence Mcl-1 expression, including PI3K/Akt [
10], Stat3 [
11,
12], CREB [
10], Ets family members Elk-1 [
13] and PU.1 [
14]. In addition, putative binding sites for NF-κB were identified in the
Mcl-1 promoter region [
9]. Previous studies demonstrated that inhibition of NF-κB activation by a novel NF-κB inhibitor V1810 [
15] or Thiocolchicoside [
16] accompanied by the downregulation of Mcl-1 expression. However, the underlying mechanistic link between NF-κB and Mcl-1 expression has not been clearly established in these studies. Moreover, although reports [
17,
18] have revealed that p65 subunit of NF-κB involves in TRAIL induced expression of Mcl-1 in HCT-116 colon carcinoma cells [
17] and the interaction of p65 with N-a-Acetyltransferase 10 protein regulates Mcl-1 expression [
18], the precise mechanism of
Mcl-1 transcriptionally controlled by NF-κB family members is not fully elucidated. Therefore, a better understanding the role of this regulatory molecule in Mcl-1 expression in cancers may allow for the development of rational therapeutics that control Mcl-1 levels.
Transcripition factor NF-κB comprised of homo- and heterodimers of the RelA (p65), RelB, c-Rel, p50/p105 (NF-κB1) and p52/p100 (NF-κB2) polypeptides can both induce and repress gene expression by binding to discrete κB elements in promoters and enhancers. The genes regulated by NF-κB include those controlling apoptosis, cell adhesion, proliferation, and inflammation. In most untransformed cell types, NF-κB complexes are largely cytoplasmic by a family of inhibitory proteins known as inhibitors of NF-κB (IκBs) and therefore remain transcriptionally inactive [
19]. Activation of NF-κB typically involves the phosphorylation of IκB by the IκB kinase (IKK) complex, which results in IκB degradation. This liberates NF-κB and allows it to translocate freely to the nucleus and binds to the κB elements in the relevant downstream genes to activate a series of transcriptional events [
19]. It has become apparent that aberrant activation of NF-κB in human cancers are common [
20]. Activation of NF-κB has been detected in tumor samples from patients, such as breast, colorectal, ovarian, pancreatic, prostate cancers and so forth [
21,
22]. Constitutive NF-κB activation has also reported in esophageal carcinoma tissues [
22,
23] and cell lines [
24], implying NF-κB activation plays an important role in the tumorigenesis and development of human ESCC. Expression of Mcl-1 has been shown in human esophageal carcinoma cell lines CE81T/VGH [
25] and KYSE450 [
26]. We thus speculated that a direct link might exist between NF-κB and Mcl-1 expression in human ESCC.
The present study was performed to determine whether Mcl-1 expression is modulated by NF-κB signal pathway in human ESCC. Using human ESCC cell lines as models, reporter gene assays demonstrate that human Mcl-1 promoter activity is decreased by mutation of κB site, specific NF-κB inhibitor Bay11-7082 or dominant inhibitory molecule DNMIκBα in TE-1 and KYSE150 cells. Mcl-1 level is attenuated by Bay11-7082 treatment or co-transfection of DNMIκBα in TE-1 and KYSE150 cells. NF-κB subunits p50 and p65 are further confirmed bound to Mcl-1-κB probe in vitro by EMSA assay and directly bound to human Mcl-1 promoter in intact cells by ChIP assay, respectively. Our data provided evidence that one of the regulatory mechanisms by which Mcl-1 expression in human ESCC is by binding of p50 and p65 to κB site within human Mcl-1 promoter. This NF-κB mediating Mcl-1 expression also contributes to the viability of TE-1 cells. In conclusion, the newly identified mechanism might provide a scientific basis for developing effective approaches to treatment human ESCC.
Methods
Cell lines and culture
Human esophageal carcinoma cell lines TE-1 and Eca109 were purchased from Cell Bank of Chinese Academy of Sciences, Shanghai, China. Human esophageal carcinoma cell lines KYSE150 and KYSE510 were kindly provided by Dr. Qian Tao from The Chinese University of Hong Kong, HongKong, China. Immortalized human keratinocyte cell line HaCaT derived from human adult trunk skin was previous described [
27,
28]. TE-1, Eca109, KYSE150 and KYSE510 cells were cultured in RPMI 1640 medium (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum, 100 units/ml penicillin and 100 mg/ml streptomycin. HaCaT was cultured in DMEM medium (Invitrogen, Carlsbad, CA) containing 10% fetal bovine serum and antibiotics as described above. All cell lines were incubated at 37˚C in a humidified atmosphere containing 5% CO2.
Chemicals and cell treatments
The specific NF-κB inhibitor Bay11-7082 (Calbiochem, Darmstadt, Germany) was prepared as a stock solution of 20 mM in DMSO (Sigma, St. Louis, MO). Subconfluent cells were treated with the compound at indicated concentrations for an indicated time. Detailed treatment procedures were described in figure legends. The final concentration of DMSO in the culture media was kept less than 0.1% which had no significant effect on the cell growth. Vehicle controls were prepared for all treatments.
Plasmids
The pGL2-Mcl-1-κBwt (Addgene plasmid 19132) which contains a 325 bp long human
Mcl-1 promoter fragment including NF-κB binding-site (GGGGTCTTCC) and the pGL2-Mcl-1-κBmt (Addgene plasmid 19133) in which the κB site sequence GGGGTCTTCC being changed to GTTGTCTTCC were constructed by Dr. El-Deiry [
17] and obtained through Addgene (Cambridge, MA). The pGL2-Basic vector was purchased from Promega (Madison, WI). The pGL3-Basic vector and pGL3-NF-κB-Luc were the same as described previously [
29,
30]. Expression plasmid of dominant negative mutant of IκBα (pcDNA3-DNMIκBα) [
30] and the pcDNA3.1 empty vector [
31] were identical to those used previously. The human full-length Mcl-1 expression vector pCMV6-A-Puro-Mcl and pCMV6-A-Puro empty vector were kindly provided by Dr. Chengchao Shou [
18].
Transfection and luciferase reporter assays
Cells were cultured in 24-well plates at a density of 1 × 105 per well overnight and transfected with Lipofectamine™ 2000 (Invitrogen, Carlsbad, CA) according to manufacturer’s instructions. In luciferase assay for NF-κB transactivation, each transfection contained 800 ng/well of pGL3-Basic or pGL3-NF-κB-Luc together with 40 ng/well of internal control pRL-SV40 (Promega, Madison, WI) (Total DNA 840 ng/well). 24 h after transfection, cells were either left untreated (DMSO) or treated with 20 μM Bay11-7082 for 12 h. Cells were harvested at 36 h after transfection and lysates were analyzed for luciferase activity using the Dual Luciferase Reporter assay (Promega, Madison, WI) with a GloMax™ Microplate Luminometer (Promega, Madison, WI). In luciferase assay for the Mcl-1 promoter, each transfection contained 400 ng/well of pGL2-Basic, pGL2-Mcl-1-κBwt or pGL2-Mcl-1-κBmt together with 400 ng/well of pcDNA3.1 or pcDNA3-DNMIκBα expression plasmid. Each transfection contained 40 ng/well of pRL-SV40 as internal control (Total DNA 840 ng/well). 24 h after transfection, cells were either left untreated (DMSO) or treated with 20 μM Bay11-7082 for 12 h. Cells were harvested at 36 h after transfection and lysates were analyzed as described above. The pRL-SV40 was co-transfected in all experiments to correct the variations in transfection efficiency. The data represent the mean ± S.D. of at least two independent experiments performed in triplicate.
RNA interference
TE-1 cells were grown in 6-well plates at a density of 3 × 105 cells per well overnight. Cells reached 60-70% confluency on the day of transfection and were transfected with a p50 (sc-29407; 100 pmol), a p65 (sc-29410; 100 pmol) or a scrambled control (sc-37007; 100 pmol) siRNA (all from Santa Cruz Biotechnology) using HiPerFect transfection reagent (Cat no: 301705, Qiagen) for 72 h according to the manufacturer’s instructions. Cells were harvested for protein extraction and immunoblotting to confirm p50 or p65 knockdown.
Cell viability assay
Cell viability assays were performed using the 4-[3-(4-iodophenyl)-2-(4-nitrophenyl)-2H-5-tetrazolio]-1,3-benzene disulfonate (WST-1) assay kit (Roche, Indianapolis, IN) according to the manufacturer’s instructions. The assay is based on the cleavage of WST-1 to formazan dye by cellular mitochondrial dehydrogenases. Because cleavage of WST-1 to formazan dye occurs only in viable cells, the amount of dye produced, measured in OD values, directly corresponds with the number of viable cells present in the culture. Briefly, TE-1 cells were firstly transfected with the control, p50 or p65 siRNA in six-well plates as described above. To investigate whether reintroduction of Mcl-1 restored cell viability, 24 h following the first transfection, a second transient transfection was carried out to ectopically express Mcl-1. Each transfection contained 2 μg pCMV6-A-Puro empty vector or pCMV6-A-Puro-Mcl construct using SuperFect transfection reagent (Cat no: 301305, Qiagen) according to the manufacturer’s instructions. At 24 h post-transfection, cells were trypsinized, an aliquot of cells was maintained in six-well plate, harvested at 120 h after NF-κB subunit siRNA transfection and analyzed the Mcl-1 levels by Western blotting. The remainder was transferred as six replicates to 96-well plates at a concentration of 2.5 × 103 cells per well in 100 μl of complete RPMI 1640. After culturing for another 24, 48, 72 h (i.e. 72 h, 96 h, 120 h after each siRNA transfection, respectively), 10 μl of WST-1 was added to each well and cells incubated for 2 h at 37°C. The cellular reduction of WST-1 to formazan and its absorbance were measured at 450 nm.
Protein preparation and western blotting
Cultured cells were harvested and whole cell lysates were prepared according to the method previously described [
30]. Nuclear extracts were prepared using a Nuclear Extract kit (Cat. no. 40010, Active Motif, Carlsbad, CA) following the manufacturer’s instructions. Protein concentration was determined using the BCA Assay Reagent (Cat. no. 23228, Pierce, Rockford, IL). Western blotting was performed as previously described [
30]. The following antibodies were used for immunodetection with appropriate dilutions: Mcl-1 (sc-819, 1:1000), p50 (sc-114, 1:1000), p52 (sc-298, 1:1000), p65 (sc-8008, 1:1000), c-Rel (sc-272, 1:1000), RelB (sc-226, 1:1000) and GAPDH (sc-47724, 1:2000) (all from Santa Cruz, CA); Histone H3 (#9715, 1:1000) were purchased from Cell Signaling Technology (Beverly, MA); β-actin (A5316, 1:5000) was purchased from Sigma (St. Louis, MO).
mRNA extraction and reverse transcription-polymerase chain reaction (RT-PCR)
Total RNA was extracted using Trizol reagent (Invitrogen, Carlsbad, CA). First-strand cDNA was synthesized from 2 μg of total RNA using the Reverse Transcription System Kit (Cat. No. A3500, Promega, Madison, WI). The resulted cDNA was subjected to PCR (94°C for 5 min followed by 34 cycles of 94°C for 30 s, 58°C for 30 s, 72°C for 40 s, and an extension for 10 min at 72°C) using primers designed for human Mcl-1 [
11]: sense, 5′-cggcagtcgctggagattat-3′ and antisense, 5′-gtggtggtggttggtta-3′, yield a 573-bp product; or for GAPDH: sense, 5′-caaagttgtcatggatgacc-3′ and antisense, 5′-ccatggagaaggctgggg-3′, yield a 195-bp product. Real-time RT-PCR experiments were done in triplicate as described previously [
32] and the primers used were as following [
33]: forward 5′-gggcaggattgtgactctcatt-3′; reverse 5′-gatgcagctttcttggtttatgg-3′. The relative Mcl-1 mRNA expression levels were calculated according to the comparative CT (∆∆CT) method after normalizing to GAPDH expression. Semiquantitive RT-PCR products were separated on 1.5% agarose gels and visualized with ethidium bromide. The identity of Mcl-1 PCR product was confirmed by direct sequencing after purification.
Electrophoretic mobility shift assays
Nuclear proteins from cultured cells were prepared and protein concentration was determined as described above. EMSA was performed using the LightShift™ Chemiluminescent EMSA Kit (Cat. No. 20148, Pierce, Rockford, IL) following the manufacturer’s instructions. The reaction mixtures (20 μl) containing 8 μg nuclear extracts were incubated with 2 nM of biotin-labeled double-stranded oligonucleotide probes in reaction buffer for 20 min at room temperature. Samples were subjected to electrophoresis in 5% nondenaturing polyacrylamide gel and transferred to Biodyne™ BNylon membrane (Cat. No. 77016, Pierce, Rockford, IL). For competition analyses, 100-fold excess of unlabeled probes were included in the binding reaction. For antibody supershift experiments, the reaction mixtures were preincubated with 2 μg of p50 (sc-8414X), p52 (sc-298X), p65 (sc-8008X), c-Rel (sc-272X), RelB (sc-226X) or rabbit IgG (sc-2027) antibody (all from Santa Cruz, CA) for 30 min at room temperature. Biotin-labeled double-stranded oligonucleotides were used as probes listed below: wild-type NF-κB consensus binding sequence: 5′-agttgag
gggactttcccaggc-3′ [
34]; wild-type Mcl-1-κB binding sequence: 5′-ggagtc
ggggtcttccccagtttt-3′, corresponding to the nucleotides of the human
Mcl-1 promoter. Unlabeled double-stranded oligonucleotides used for competition analyses were: wild-type NF-κB consensus binding sequence: 5′-agttgag
gggactttcccaggc-3′; mutated NF-κB consensus binding sequence: 5′-agttgag
gagatctggccaggc-3′ [
34]; mutant Mcl-1-κB binding sequence: 5′-ggagtc
g
tt
gtcttccccagtttt-3′; The AP-1 consensus probe was used as a nonspecific competitor for NF-κB: 5′-cgcttga
tgagtcagccggaa-3′ [
35]. The probes were commercially synthesized by TaKaRa Bio Inc. (Dalian, China). Binding sites were indicated in italics type and mutations were shown in bold type. The mutated nucleotides for NF-κB binding site of human
Mcl-1 promoter in EMSA were identical to those of the mutated sequences in the reporter construct.
Chromatin immunoprecipitation (ChIP) assay
ChIP was performed using the ChIP assay kit (Upstate Biotechnology, Lake Placid, NY) as previously described [
30]. Antibodies used for immunoprecipitation were: p50 (sc-8414X), p52 (sc-298X), p65 (sc-8008X), c-Rel (sc-272X), RelB (sc-226X) and rabbit IgG (sc-2027) (all from Santa Cruz, CA). 2 μg of each antibody was used for each immunoprecipitation. The following primers were used in the ChIP assays: human
Mcl-1 promoter including the NF-κB binding region, 5′-cacttctcacttccgcttcc-3′ and 5′-ttctccgtagccaaaagtcg-3′ (200 bp).
Statistical analysis
Statistical analysis was done with the statistical software program SPSS ver.12.0. Results expressed as mean ± S.D. were analyzed using the Student’s t test. Differences were considered significant when P value was <0.05.
Discussion
Expression of Mcl-1 is frequently increased in various human tumors, so the mechanisms that increase Mcl-1 levels are of paramount importance. In addition to being modulated at transcriptional level by various transcription factors that bind and activate the
Mcl-1 promoter aforementioned, Mcl-1 could be regulated on multiple levels, such as translational and post-translational. For instance, E3 ubiquitin ligase Mule has been identified to required and sufficient for the polyubiquitination of Mcl-1. Elimination of Mule expression by RNA interference stabilizes Mcl-1 protein, resulting in an increase of Mcl-1 protein level [
41]. Another E3 ligase β-TrCP facilitates the ubiquitination and degradation of GSK-3β-phosphorylated Mcl-1, which contributes to GSK-3β-induced apoptosis [
42]. Mutational inactivation of E3 ligase FBW7 was found to occur in several neoplastic diseases, which can decrease Mcl-1 degradation, resulting in increased Mcl-1 protein levels and resistance to chemotherapeutic agents [
43]. In contrast, deubiquitinase USP9X, which is overexpressed in some malignancies, stabilizes Mcl-1 and promotes tumor cell survival. Knockdown of USP9X decreased Mcl-1 levels [
5]. Moreover, phosphorylation of Mcl-1 at Thr 163 by ERK [
43] prolongs the Mcl-1 half-life while phosphorylation at Thr 163 by GSK-3β [
42] or Thr 92 by CDK1 [
43] enhances Mcl-1 degradation. In addition, Mcl-1 transcripts can be influenced by microRNAs (miRs). For example, miR29b has been demonstrated to downregulate Mcl-1 protein and sensitize cells to apoptosis [
44]. Future studies need to explore whether these mechanisms contribute to the elevated Mcl-1 protein in human ESCC.
Increased Mcl-1 protein level has been reported to compromise the apoptotic effects of chemotherapeutic agents, resulting in therapeutic resistance [
43]. Thus, the pathways that are critical for regulating Mcl-1 expression have been employed to target Mcl-1 for cancer therapy. For instance, in large granular lymphocyte leukemia, targeting Stat3 with its upstream kinase JAK-selective inhibitor AG490 transcriptionally suppresses Mcl-1 and promotes apoptosis [
12]. PI3K/Akt signaling is involved in Mcl-1 induction [
10], targeting this pathway by newly developed PI3K inhibitor PI103 is showed to suppress Mcl-1 and induced apoptosis and restore sensitivity to TRAIL-induced apoptosis in neuroblastoma [
45]. Treatment with MEK/ERK inhibitor U0126 resulted in Mcl-1 downregulation and induced marked apoptosis in Mel-RM melanoma cells [
46]. Therefore, identification of pathways that regulate Mcl-1 may help to improve the therapeutic effect of chemotherapy. Our data indicated that inhibition of NF-κB pathway by Bay11-7082 (Figure
4A, B), DNMIκBα (Figure
4C, D) or NF-κB subunit siRNA (Figure
6) attenuates Mcl-1 expression in human ESCC cells. We also found that the survival of TE-1 cells is impaired when NF-κB is blocked by expression of p50 siRNA or p65 siRNA and reintroduction of Mcl-1 to the siRNA-transfected TE-1 cells significantly restores cell viability (Figure
6E). These data that decrease Mcl-1 expression and inhibits cell viability by inhibition of NF-κB pathway support the use of selective NF-κB inhibitors in the treatment of Mcl-1-overexpressing human ESCC.
By gel shift analysis, nuclear extracts of TE-1 cells were preincubated with antisera directed against individual NF-κB family members p50, p52, p65, c-Rel, RelB or with a nonspecific antisera prior to interaction with the Mcl-1-κB site probe. We found that NF-κB family members p50, p52 and p65 were able to bind to the same probe
in vitro. The result was in agreement with the earlier findings that most κB sites show no or little selectivity for a given NF-κB species and different dimers have broad sequence recognition specificities although relatively small differences in the relative affinity of NF-κB dimers for a given site can be found [
47‐
49]. However, p50 and p65 but not p52 were revealed directly binding to the κB site of human
Mcl-1 promoter in intact cells by ChIP assays. The discrepancy between the measured
in vitro affinity of NF-κB for the κB probe and the real
in vivo occupancy at κB site of the natural promoter is not without precedent. For instance, ChIP result showed that, in LPS-stimulated DCs, the κB site of
IL-8 promoter is a highly selective p65 recruiter [
50], while in
in vitro experiments, it is bound and activated by both p65 and c-Rel homodimers [
51]. The ability of a specific gene to selectively recruit various NF-κB dimers in vivo cannot be predicted on the basis of in vitro results [
50]. The context of κB site physiological promoter rather than the κB site itself is the major determinant of which NF-κB dimmer will ultimately be loaded onto a certain promoter.
Although putative binding sites for NF-κB were identified in the
Mcl-1 promoter region [
9] and two recent reports have shown that NF-κB is directly involved in Mcl-1 regulation [
17,
18]. In the first article, by using ChIP assay, the authors show that p65 subunit of NF-κB following TRAIL treatment binds to the
Mcl-1 promoter, which suggested that TRAIL induced expression of Mcl-1 through activation of NF-κB in HCT-116 colon carcinoma cells [
17]. In the second study, the authors show that transcriptional activation of
Mcl-1 gene required the recruitment of N-a-Acetyltransferase 10 protein/p65 complex to the p65-binding site of the
Mcl-1 promoter region [
18]. However, both studies focused only on the role of NF-κB p65 subunit in Mcl-1 expression and the report of other NF-κB subunits involved in Mcl-1 expression is relatively limited. Since dimerization is required for NF-κB binding to DNA and more than 12 homo- and heterodimers have been described [
50]. The analysis of other members of the NF-κB family to bind to κB site and regulate Mcl-1 expression would allow for a better understanding of the precise mechanism of
Mcl-1 transcriptional control by NF-κB. Our results indicate that effect of NF-κB on Mcl-1 expression in TE-1 cells is due to activation of NF-κB subtypes p65 and p50, without activation of other subtypes (Figure
5B) and reveal that activations of p65 and p50 are involved in Mcl-1 expression thus affecting cell viability (Figure
6E). Notably, we did not observe the involvement of NF-κB pathway in human
Mcl-1 promoter activity in Eca109 cells (Figure
3A). In addition to NF-κB binding site, the 325 bp long
Mcl-1 promoter fragment contains CRE-BP, Ets, Sp1, SRE, STAT binding sites [
17,
52]. We speculated that, in Eca109 cells, other transcription factor(s) rather than NF-κB might play a leading role in Mcl-1 expression. Our results suggested that the existence of other regulatory cascades that modulate Mcl-1 expression in different ESCC cells.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
HDL conceived the study, analyzed data and drafted the manuscript. JFY YCY acquired and analyzed data. ZKX MJC JW LX XLM acquired data. SFO QW provided material support. XMZ YFY FLY YC reviewed the manuscript. BLY JGH supervised the study, analyzed data and finalized the manuscript. All authors read and approved the final manuscript.