Background
Nuclear factor κB (NF-κB) is a family of transcription factors that are implicated in many physiological and pathological processes, including immunity, inflammation, carcinogenesis and chemoresistance[
1,
2]. In mammals, the NF-κB family consists of five structurally related proteins, RelA (p65), NF-κB-1 (p105/p50), NF-κB-2 (p100/p52), RelB, and c-Rel, which form various homodimers and heterodimers. The predominant form of NF-κB consists of p50 and p65 subunits. In most cell types, NF-κB is mainly trapped in the cytoplasm in an inactive form bound to IκB proteins, the inhibitors of NF-κB. In response to stimuli like tumor necrosis factor alpha (TNFα) or interleukin-1β (IL1β), TGFβ-activated kinase-1 (TAK1) and its adaptors TAB2/3 are recruited to the receptor proximal signaling complex, leading to the activation of IκB kinase (IKK). The activated IKK phosphorylates IκB proteins and triggers the ubiquitination and degradation of IκB, which allows p50/p65 heterodimer to be released, translocate to the nucleus and act as a sequence-specific DNA-binding transcription factor. Meanwhile, IKK phosphorylates the Ser536 of p65 and thereby enhances the transactivation activity of NF-κB[
1,
3]. In addition to extracellular ligands that signal through membrane receptors, the chemotherapy-induced DNA damage also activates NF-κB in some cell contexts[
1,
4]. Many potent anti-apoptosis genes are transactivated by NF-κB[
1,
2]. Therefore, the activation of NF-κB may desensitize cells to apoptosis and thereby promote cancer progression. Unfortunately, abnormal activation of the NF-κB pathway is a common phenomenon in cancer cells[
2].
MicroRNAs (miRNAs) are evolutionarily conserved small non-coding RNAs that suppress protein expression by binding to the 3’-untranslated region (3’UTR) of target mRNA. Several miRNAs have been reported to modify cell behavior by regulating the NF-κB pathway. For example, miR-301a, miR-30e* and miR-182 promote NF-κB activity and thereby enhance tumor growth, invasiveness or angiogenesis[
5‐
7]. On the contrary, miR-15/16/195 and miR-146a/b have been shown to impair NF-κB activity, thus reducing the proliferation and metastasis of tumor cells[
8‐
10]. Very few miRNAs have been characterized to affect chemosensitivity by regulating the NF-κB pathway: miR-143 sensitizes colorectal cancer cells to 5-fluorouracil treatment by downregulating ERK5, Bcl-2 and p65 expression[
11]; miR-146a enhances the chemosensitivity of NK/T cell lymphoma to etoposide by targeting TRAF6[
12]. Clearly, identification of miRNAs that target NF-κB signaling may provide novel molecular targets for cancer therapy.
It is reported that NF-κB signaling is frequently activated in hepatocellular carcinoma (HCC)[
2,
13]. In order to uncover the HCC-associated miRNAs which may regulate the NF-κB pathway, we predicted the targets of those deregulated miRNAs that we found in HCC tissues[
14], using target prediction algorithms (TargetScan). Among the miRNAs that were predicted to target the regulators of NF-κB pathway, miR-26b stood out as a potential candidate, with TAK1 and TAB3 as its putative targets. Studies from us and other groups show that the miR-26 family (miR-26a/b) is frequently downregulated in multiple types of cancer, including HCC, breast cancer, nasopharyngeal carcinoma and melanoma[
15‐
19]. To date, there is no report disclosing the regulatory role of miR-26a/b on the NF-κB pathway and its biological significance. Because of the share of common seed sequences for target recognition, members of a miRNA family usually play similar, if not identical, roles. Therefore, we explored the impact of miR-26b on NF-κB signaling and its biological significance. We found that miR-26b suppressed the TNFα- and doxorubicin-activated NF-κB signaling in HCC cells, and sensitized cancer cells to the doxorubicin-induced apoptosis by targeting TAK1 and TAB3.
Discussion
NF-κB is frequently activated in the various types of tumors and promotes cancer development and chemoresistance. Therefore, miRNAs that possess the NF-κB inhibitory activity may provide novel targets for anti-cancer therapy.
The aberrant activation of NF-κB pathway may result from different causative mechanisms, like inactivating mutations or deletions of
IκBα in Hodgkin’s lymphoma[
23], activating mutations of
NF-κB2 gene in B- and T-cell lymphomas[
23], amplifications and/or translocations of
NIK in multiple myeloma[
2]. In addition to the genomic aberrations of protein-coding genes, deregulation of the miRNAs that regulate the NF-κB pathway also results in the abnormal activation of the NF-κB pathway. For example, the NF-κB activators miR-301a and miR-30e* are overexpressed in pancreatic cancer and glioma, respectively[
5,
6]; the NF-κB inhibitor miR-195 is downregulated in HCC[
9]. The decreased expression of the miR-26 family members has been observed in HCC tissues and the low miR-26a/b levels were associated with the short survival of patients[
15,
16]. Our data reveal a significant inhibitory role of miR-26b on NF-κB signaling, suggesting miR-26b downregulation as a novel mechanism that contributes to the abnormal activation of the NF-κB pathway in HCC cells.
Doxorubicin and its analogs, like daunorubicin, have gained broad application for the chemotherapy of various malignant tumors. They trigger the apoptosis of cancer cells by interfering with the actions of DNA topoisomerase IIα and creating DNA double-strand breaks[
21]. Upon DNA damage, ataxia telangiectasia mutated (ATM) and sumoylated NF-κB essential modulator (NEMO) are jointly exported from the nucleus and mediates the TAK1/TAB2/3-dependent IKK activation[
4]. Activated IKK then induces the phosphorylation and degradation of IκBα, which further liberates NF-κB to the nucleus and stimulates its DNA-binding activity. In lymphoma and cervical cancer cells, treatment with doxorubicin/daunorubicin enhances the transactivation activity of NF-κB[
24,
25]. On the other hand, although doxorubicin/daunorubicin also induces the phosphorylation and degradation of IκBα and increases the DNA-binding activity of NF-κB in osteosarcoma and breast cancer cells[
26,
27], yet they repress the NF-κB reporter activity and the expression of the NF-κB-regulated anti-apoptosis genes. Histone deacetylase has been shown to be recruited to p65 in the daunorubicin-treated osteosarcoma cells, which converts p65 into an active transcription repressor of anti-apoptosis genes[
26]. In breast cancer cells, NF-κB induced by doxorubicin is deficient in phosphorylation and acetylation and represses the NF-κB-dependent transcription in a histone deacetylases-independent manner. The cellular context-dependent response calls for a cell-type-specific analysis in determining the outcome of doxorubicin-stimulated NF-κB signaling. Herein, we disclosed that doxorubicin markedly increased the level of phosphorylated p65 at the serine-536 residue and obviously activated the anti-apoptosis genes in HCC cells. Furthermore, the depletion of p65 dramatically increased the apoptosis rates of doxorubicin-exposed HCC cells, suggesting that NF-κB activation impairs the chemosensitivity of tumor cells and the inhibition of NF-κB signaling represents an effective strategy to overcome the chemoresistance of HCC to doxorubicin. Consistently, it has been shown that doxorubicin treatment activates NF-κB in other types of cancer cell lines, and blocking NF-κB activation sensitizes these cells to doxorubicin-triggered apoptosis[
28]. Importantly, we found that the restoration of miR-26b expression significantly inhibited the phosphorylation of IκBα and p65, reduced the NF-κB reporter activity, blocked the nuclear translocation of NF-κB, consequently abrogated the expression of anti-apoptosis genes and sensitized HCC cells to the doxorubicin-induced apoptosis in HCC cells. These observations indicate miR-26b as a potent NF-κB inhibitor that may increase the chemosensitivity of HCC cells.
TAK1 is a member of the mitogen-activated protein kinase kinase kinase (MAP3K) family. It mediates the activation of IKK in various cell systems[
3]. Upregulation of TAK1 is found in clear cell renal cell carcinomas and aggressive esophageal squamous cell carcinomas[
29,
30]. TAB3 is markedly overexpressed in skin, testis and small intestinal cancers[
31]. Our results showed that miR-26b could decrease the protein level of TAK1 and TAB3 by directly binding to their 3’UTR. Considering that the expression of miR-26 family is frequently reduced in multiple types of cancer[
15‐
19], we suggest that miR-26b downregulation may represent one of the mechanisms responsible for the overexpression of TAK1 and TAB3 in cancers.
In previous studies, miR-26a/b have been shown to block the G1/S transition of the cell cycle by targeting CCND2, CCNE1/2, CDK6, and EZH2 in HCC and nasopharyngeal carcinoma[
15,
18,
32], and to restrain metastasis by suppressing the expression of IL6 in HCC[
33]. miR-26a/b have also been reported to induce cell apoptosis by targeting MTDH, EZH2 and SLC7A11 in breast cancer cells and by targeting SODD in melanoma cells[
17,
19,
34]. Our findings suggest that miR-26b may also suppress NF-κB signaling and enhance the chemosensitivity of hepatocellular carcinoma cells by targeting TAK1 and TAB3. It is exciting to find that a single miRNA may suppress tumor growth via multiple mechanisms, which makes miR-26b a promising anti-cancer target.
Methods
Human tissue specimens
Histologically confirmed HCC tissues were collected from 20 patients who underwent HCC surgical resection at the Cancer Center of Sun Yat-sen University in Guangzhou, P.R. China. None of the patients had received any local or systemic anticancer treatment prior to the surgery. This study was approved by the Institute Research Ethics Committee at the Cancer Center and informed consent was obtained from each patient.
Cell lines and transient transfection
Human hepatocellular carcinoma cell lines QGY-7703 and MHCC-97H were maintained in Dulbecco’s modified Eagle’s medium (DMEM, Invitrogen Corp., Buffalo, NY, USA) supplemented with 10% fetal bovine serum (FBS, Hyclone, Thermo Fisher Scientific, Victoria, Australia).
RNA oligos were reversely transfected using Lipofectamine RNAiMAX (Invitrogen, Carlsbad, CA, USA). A final concentration of 50 nM RNA duplex or 200 nM miRNA inhibitor was used. Transfection of plasmid DNA alone or together with RNA duplex was conducted using Lipofectamine 2000 (Invitrogen).
RNA oligoribonucleotides
hsa-miR-26b (Pre-miR miRNA Precursor Product, cat. AM17100) and its negative control (Pre-miR Negative Control #2, cat. AM17111) were purchased from Applied Biosystems (Foster City, CA, USA). The sequence-specific miR-26b inhibitor (anti-miR-26b, cat. miR20000083) and its control (anti-NC, cat. miR02201) were obtained from Ribobio (Guangzhou, P.R. China). All siRNA duplexes were purchased from GenePharma (Shanghai, P.R. China). siTAK1, siTAB3 and sip65 targeted the mRNAs of human
TAK1 [GenBank: NM_145333],
TAB3 [GenBank: NM_152787] and
p65 [GenBank: NM_001145138] genes, respectively. The negative control RNA duplex (siNC) for siRNA was non-homologous to any human genome sequences. The sequences of all siRNA duplexes are listed in Additional file
7: Table S1.
Vectors and luciferase reporter assay
To quantitatively examine NF-κB activity, luciferase reporter plasmid containing the minimal promoter with multiple tandem NF-κB binding sites (pNF-κB-Luc, Clontech, Palo Alto, CA, USA) and its control vector (pTAL-Luc, Clontech) were employed. Cells were first reversely transfected with 50 nM RNA duplex in a 48-well plate for 24 hours, followed by co-transfection with 10 ng pRL-TK (Promega, Madison, WI, USA) and 50 ng pNF-κB-Luc or pTAL-Luc for 32 hours, then remained untreated or treated with 20 ng/ml TNFα (cat. 210-TA-010, R&D Systems, Oxon, UK) for 4 hours or doxorubicin hydrochloride (cat. D1515, Sigma-Aldrich, St. Louis, MO, USA) for 12 hours before luciferase activity analysis.
To verify the miR-26b-targeted 3’UTR, firefly luciferase reporter plasmids pGL3cm-TAK1-3’UTR-WT, pGL3cm-TAK1-3’UTR-MUT, pGL3cm-TAB3-3’UTR-WT and pGL3cm-TAB3-3’UTR-MUT were constructed. The 3’UTR segment of human
TAK1 (529 bp) or
TAB3 (540 bp) mRNA that contained the putative wild-type (WT) or mutant (MUT) miR-26b binding site (Additional file
3: Figure S3) was PCR-amplified and inserted into the
Eco RI and
Xba I sites downstream of the stop codon of firefly luciferase in pGL3cm vector, which was created based upon the pGL3-control vector (Promega), as previously described[
14]. The sequences of all primers are listed in Additional file
7: Table S1. Cells cultured in a 48-well plate were co-transfected with 20 nM RNA duplex, 10 ng pRL-TK and 20 ng firefly luciferase reporter containing the wild-type or mutant 3’UTR of target genes for 48 hours before luciferase activity analysis.
Luciferase activity was measured using the dual-luciferase reporter assay system (Promega). pRL-TK, which expresses Renilla luciferase, was used as an internal control to adjust for discrepancies in both transfection and harvest efficiencies. The luciferase activity of miR-26b-transfectants was normalized to the mean luciferase activity of NC-transfectants.
Immunofluorescent staining for p65
Cells cultured on coverslips in a 48-well plate, were reversely transfected with 50 nM RNA duplexes for 48 hours and remained untreated or treated with 20 ng/mL TNFα for 10 minutes. Then the cells were fixed with 4% paraformaldehyde (PFA, cat. 16005, Sigma-Aldrich) and stained with rabbit monoclonal antibody (mAb) against p65 (cat. #4764, Cell Signaling Technology, CST, Beverly, MA, USA), followed by incubation with HiLyte Fluor 555-conjugated goat anti-rabbit IgG (cat. 28176-05-H555, AnaSpec, Fremont, CA, USA) and nuclear counterstaining with DAPI (cat. D9542, Sigma-Aldrich). Fluorescent pictures were photographed with Zeiss Axio Imager Z1 (Zeiss, Jena, Germany).
RNA extraction and real-time quantitative RT-PCR
Total RNA was extracted from cultured cells using TriPure Isolation Reagent (cat. 11667165001, Roche Applied Science, Germany) according to the manufacturer’s instructions.
For real-time quantitative RT-PCR (qPCR) analysis of mRNA, 2 μg of total RNA was subjected to DNaseI digestion (Fermentas, Hanover, MD, USA) at 37°C for 30 minutes and then to heat inactivation of DNaseI at 65°C for 10 minutes, followed by reverse-transcription using Moloney murine leukemia virus reverse transcriptase (Promega). mRNA level was detected using Power SYBR® Green PCR Master Mix (Applied Biosystems) and
β-actin was used as an internal control. The primers used for qPCR are listed in Additional file
7: Table S1. For qPCR analysis of miRNA, cDNA was synthesized using the Taqman miRNA reverse transcription kit (Applied Biosystems). The expression levels of miR-26b and the reference gene RNU6B were quantified using the TaqMan MicroRNA Assay Kit (Applied Biosystems).
All reactions were performed on a LightCycler® 480 (Roche Diagnostics, Germany) and were run in triplicate. The cycle threshold (Ct) values did not differ by more than 0.5 among the triplicates. The levels of target genes were normalized to the levels of the internal control genes to permit the calculation of the 2-ΔΔCt value.
Immunoblotting
Cellular proteins were separated in SDS-polyacrylamide gels, electrophoretically transferred to polyvinylidene difluoride membranes (cat. #162-0177, Bio-Rad, Cambridge, MA, USA), then detected with antibodies. The sources of antibodies were as follows: mouse mAb against IκBα (cat. #4814, CST), phospho-Ser32/Ser36 of IκBα (cat. 551818, BD, Franklin Lakes, NJ, USA) and β-actin (cat. BM0627, Boster, Wuhan, China); rabbit mAb for p65 (cat. #4764, CST), phospho-Ser536 of p65 (cat. #3033, CST) and TAK1 (cat. #5206, CST); rabbit polyclonal antibodies for TAB3 (cat. ab85655, Abcam, Cambridge, MA, USA) and caspase-3 (cat. #9662, CST). β-actin was used as an internal control. All results were reproduced in three independent experiments, and the representative immunoblots are shown.
Apoptosis analysis
For cultured cells, 24 hours after transfection, cells were treated with doxorubicin for 48 hours, then applied to morphological examination and detection of caspase-3 activity. For morphological examination, the cells were fixed with 4% PFA, stained with DAPI and those with condensed or fragmented nuclei were considered as apoptotic cells. At least 500 cells were counted for each sample. The activity of caspase-3 was detected by immunoblotting. Activated caspase-3 resulting from the cleavage of the inactive proenzyme form was indicated as 17/19 kDa bands below the full length caspase-3 (35 kDa) band.
For HCC tissues, TUNEL staining was performed using the In Situ Cell Death Detection Kit (cat. 11684817910, Roche Applied Science), according to the manufacturer’s protocol. At least 750 cells were counted for each sample.
Statistical analysis
Data were expressed as the mean ± standard error of the mean (SEM) from at least three independent experiments. Analyses on the differences between groups were performed using GraphPad Prism version 4.0 (GraphPad Software, Inc., San Diego, CA, USA). Student’s t test was performed to compare the differences between two groups and one-way ANOVA was applied to compare more than two groups. Correlation between the miR-26b level and the apoptosis rate in HCC tissues was explored using Spearman’s correlation coefficient. All statistical tests were two-sided and P < 0.05 was considered to be statistically significant.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
NZ and RW designed and performed experiments, discussed and interpreted the data. LZ and JG performed experiments. YZ gave suggestion on study design, discussed and interpreted the data. SMZ designed and supervised study, discussed and interpreted the data, wrote the manuscript. All authors read and approved the final manuscript.