The human cervical cancer cell lines, HeLa (CCL-2) and C33A (HTB-31), were originally purchased from the American Type Culture Collection (ATCC) and cultured in the tumor, virus and ncRNA group. The (HeLa and C33A) cells were cultured in DMEM (VivaCell Biosciences), supplemented with 10% fetal bovine serum (FBS; Biological Industries), 100 µg/ml streptomycin and 100 IU/ml penicillin (cat. no. P1400, Beijing Solarbio Science & Technology Co., Ltd.). All cells were cultured at 37°C with 5% CO₂.

All transfection experiments were carried out using Lipofectamine 2000^® reagent (Thermo Fisher Scientific, Inc.) following the manufacturer's instructions. Plasmids (2 µg per well in a six-well plate or 500 ng per well in 24-well plate) or antisense oligonucleotides (ASO-NC/ASO-miR-27a-3p, 50 nM) were transfected into cervical cancer cells. After adding the complex (mixture of transfection reagent and DNA or ASO) to the wells, the cells were incubated at 37°C for 24–48 h for further analyses.

The strand-specific miR-27a-3p expression vector was constructed following a previously described strategy (34). In brief, the mature miR-27a-3p sequence with its first 19 nucleotides complementary sequence and the loop sequence were synthesized and inserted into the pcDNA3-U6M2 vector (modified from the pcDNA3 vector, replacing the CMV promoter with U6 promoter, abbreviated as U6M2; a gift from Dr Per Johansson, Karolinska Institutet, Sweden) between the BglII/KpnI sites. For miR-27a-3p inhibition, the synthesized 2′-O-methyl-modified antisense oligonucleotides of miR-27a-3p (ASO-miR-27a-3p) and the scramble control oligonucleotides (ASO-NC) were purchased from Shanghai GenePharma Co., Ltd.

To construct the overexpression vector for TAB3, the full coding sequence of TAB3 from the cDNA of HeLa cells was amplified and cloned into the pcDNA3-3×Flag vector (a modified vector, named as pcDNA3/Flag-KBE, abbreviated as KBE, since the 3×Flag sequences inserted between the NcoI/KpnI sites, the cloned fragment can only be inserted downstream of the KpnI site) between the KpnI/EcoRI sites, respectively. The overexpression vectors for p65 (KBE-p65), p50 (KBE-p50) were provided by Dr Weiying Liu (Tianjin Medical University), the shRNA vector targeting NFKB1 and the empty vector control pSilencer2.1-U6 neo were provided by Dr Qi Sun (Tianjin Medical University). The shRNA vectors targeting TAB3, NF-κB/p65 and Agonaute (AGO)2 (shR-TAB3, shR-p65, shR-AGO2) were constructed by annealing the synthesized oligonucleotides into pcDNA3-U6M2 vector between the BglII/KpnI sites.

The luciferase reporter vectors for the miR-27a promoter were constructed by amplifying the respected sequences from the genomic DNA of HeLa cells and cloned into the pGL3-enhancer vector (Promega Corporation), between the XhoI/HindIII sites (pGL3-P1 and pGL3-P2). The series truncated promoter reporter vectors (R175, R259, R509, F259 and F509) were constructed using the same strategy by amplifying the related fragments from the pGL3-P2 plasmid. The mutated reporter vector with the p65 binding site deletion (F509-Del546) was synthesized by Sangon Biotech Co., Ltd.

The wild-type and the mutated form of the 3′UTR EGFP reporter vectors for TAB3, mitogen-activated protein kinase kinase kinase 14 (MAP3K14) and TNF receptor associated factor 3 (TRAF3) were constructed by annealing the synthesized oligonucleotides (containing the predicted miR-27a-3p binding sites, wild-type and mutated form) and cloning into the pcDNA3/EGFP vector between the BamHI/EcoRI sites. All constructed vectors were verified using Sanger sequencing at GENEWIZ. All primers used for PCR amplification and all oligonucleotides for preparing constructs were synthesized by Synbio Technologies or General Biol, and their sequences are listed in Supplementary Table SI.

The online miRNA target prediction tool, TargetScan (https://www.targetscan.org/), was used to predict the potential targets of miR-27a-3p. An online analyses tool (Assistant for Clinical Bioinformatics, http://www.aclbi.com) based on The Cancer Genome Atlas (TCGA)-Cervical Squamous Cell Carcinoma and Endocervical Adenocarcinoma (CESC) dataset (http://portal.gdc.cancer.gov) was used to analyze the expression levels of miRNAs and the selected potential targets. The NCBI GEO datasets (GSE86100 and GSE20592) were also selected to analyze the expression levels of miR-27a-3p and the miR-23a/27a/24-2 cluster members. GEPIA2 (http://gepia2.cancer-pku.cn/#index) was used to analyze the overall and disease-free survival based on the expression levels of TAB3, MAP3K14 and TRAF3 from the TCGA-CESC dataset, and the online tool HIPLOT (https://hiplot.com.cn/cloud-tool/drawing-tool/detail/122) was applied to draw the survival curves based on miR-27a-3p expression from the TCGA-CESC dataset.

For promoter prediction, with the help of FANTOM (https://fantom.gsc.riken.jp/) and Promoter 2.0 (https://services.healthtech.dtu.dk/service.php?Promoter-2.0) predictions, combining with CpG island analysis, the upstream (~8 kb) sequences of miR-27a-3p were analyzed.

Total RNA was extracted from the cultured cells using TRIzol^® reagent (Invitrogen; Thermo Fisher Scientific, Inc.) following the manufacturer's recommendations. The concentration of the RNA was measured using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Inc.) and stored at −80°C. cDNA was synthesized using the HiScript II Q Select RT SuperMix for qPCR kit (cat. no. R233-01, Nanjing Vazyme Biotech Co., Ltd.). A total of 1 µg RNA was first treated by genome DNA wiper mix at 42°C for 2 min and followed by 50°C/15 min plus 85°C/5 sec for cDNA synthesis. The expression levels of miRNAs and mRNAs were quantified by qPCR using 2X Universal SYBR-Green fast qPCR Mix (ABclonal Biotech Co., Ltd.). β-actin and U6 snRNA were used as the endogenous controls for mRNA and miRNA quantification, respectively, reported as 2^−∆CCq (35) and further normalized to their respective controls. All primers used for RT-qPCR were synthesized by Synbio Technologies and are listed in Supplementary Table SI.

Cells were collected after 48 h post-transfection and lysed in RIPA buffer (cat. no. R0010, Beijing Solarbio Science & Technology Co., Ltd.). Proteins (50–100 µg per lane, determined using the BCA protein assay kit, Beyotime Institute of Biotechnology, Inc.), were separated by 10% SDS-PAGE and transferred to nitrocellulose membranes. After blocking with 5% non-fat milk in 1×TBST (cat. no. BF-0113, Beijing Dingguo Biotechnology Co. Ltd.) at room temperature for 2 h, followed by incubation with the indicated dilutions of the primary antibodies overnight at 4°C. GAPDH served as the endogenous loading control for the whole cell lysates and the cytoplasmic sub-fractions, and Lamin B1 was selected for the internal reference index for the nuclear fractions. After washing three times (5 min for each) with TBST, the secondary antibodies (anti-mouse or anti-rabbit IgG-HRP) were added and incubated for a further 1 h at room temperature. Detection was performed using ECL HRP chemiluminescent substrate reagent (cat. no. BL520B, Biosharp Life Sciences). The protein expression levels were quantified on the blots using ImageJ software (version 1.52v, National Institutes of Health). All antibodies used in the present study are listed in Table I.

For promoter luciferase reporter assays, the HeLa cells were seeded in 96-well plates and transfected with the indicated series reporter vectors (100 ng/well) with/without the modulation of NF-κB/p65 or NFKB1/p50. Each well was co-transfected with the pRL-CMV vector (10 ng/well) for normalization. The activities of luciferases (Firefly and Renilla, FL and RL) were measured at 30 h post-transfection using the Dual-Luciferase Assay kit (cat. no. E1910, Promega Corporation) following the recommendations of the manufacturer. The relative Firefly activities were normalized to the Renilla luciferase activities in each well.

For 3′UTR EGFP fluorescence reporter assays, the HeLa cells were seeded in 24-well plates and transfected with the wild-type or mutated form of the reporter vectors (500 ng/well) with/without the overexpression of miR-27a-3p. pDsRed2-N1 (100 ng/well, Clontech; Takara Bio USA) co-transfected into each well for normalization. At 48 h following transfection, the cells were lysed with RIPA buffer and the fluorescence intensities of EGFP and RFP were determined using a microplate reader (Tecan Spark, Tecan Group, Ltd.).

Following 24 h of transfection, the HeLa and C33A cells were re-suspended and seeded in a 24-well plate (300 cells/well) or 12-well plate (500 cells/well) and incubated for a further 10–14 days a 37°C in a 5% CO₂ incubator. During the incubation period, fresh culture medium was changed every 3–4 days. The cells were fixed with 4% paraformaldehyde (cat. no. P1110; Beijing Solarbio Science & Technology Co., Ltd.) for 30 min and stained with 0.5% crystal violet (cat. no. C0121, Beyotime Institute of Biotechnology) for 10 min at room temperature, and the colonies including >50 cells were counted using a light microscope (ECLIPSE TS100, Nikon Corporation).

Cell migration and invasion were analyzed using 24-well Boyden chambers with an 8-µm pore size polycarbonate membrane. The insert chambers were placed in a 24-well plate, containing 800 µl culture medium with 30% FBS. For the invasion assays, the Matrigel (cat. no. 356234, Corning Inc.) was 1:7 diluted with cold serum-free medium, placed onto the chamber membrane in advance and incubated at 37°C for at least 2 h. In brief, following 24 h of transfection, 6×10⁵ cells (for migration) or 8×10⁵ cells (for invasion) were re-suspended in serum-free culture medium and seeded in the upper chamber and incubated at 37°C for a further 36 h. At the end of the incubation period, the non-migrated or non-invaded cells on the top surface of membrane were removed using cotton swabs, and the migrated or invaded cells on the bottom surface of membrane were fixed with 4% paraformaldehyde for 30 min and stained with crystal violet for 10 min at room temperature. For quantification, the stained migrated or invaded cells were dissolved using 33% acetic acid reagent and the optical densities at 570 nm were determined using an MQX200 microplate reader (BioTek Instruments, Inc.).

The HeLa cells were seeded in a 12-well plate with coverslips inside. Following transfection with 1 µg plasmids (miR-27a-3p, TAB3 and their respective vector controls) for 48 h, the cells were fixed with 4% paraformaldehyde for 10 min at room temperature, permeabilized with 0.01% Triton X-100 (v/v, in 1X PBS) for 2 min, and blocked in 10% donkey serum (cat. no. S9100, Beijing Solarbio Science & Technology Co., Ltd.) at room temperature for 30 min. Following overnight incubation with anti-p65 antibody (1:200; cat no. AF1234, Beyotime Institute of Biotechnology) at 4°C, the FITC-labeled secondary antibody (1:200; cat. no. AJ1027, Tianjin Sungene Biotech Co., Ltd.) were added and incubated for 2 h at 4°C in the dark. The cells were mounted with DAPI Fluoromount-G and images were captured using a TS2 fluorescence microscope (Nikon Corporation). ImageJ software (version 1.52v, National Institutes of Health) was used to analyze the mean fluorescence intensities from at least three views, including ~15-25 individual cells, and the quantification was normalized to their respective control groups.

At 48 h post-transfection, the cells (in a six-well plate) were trypsinized, harvested with cold PBS and lysed with 150 µl Buffer I [0.4% NP-40 (cat. no. DH218 (Guangzhou Dingguo Biology) in PBS with 1X Protease Inhibitor Cocktail, cat. no. P6730, Beijing Solarbio Science & Technology Co., Ltd.] by vortexing. Following incubation for 5 min at 4°C, the lysates were centrifuged at 5,000 × g for 10 min at 4°C. The supernatant (i.e., the cytosolic extracts) was carefully collected and stored at −20°C for further analyses. The precipitates were re-suspended in 100 µl Buffer II (0.1% NP-40 in PBS with 1X Protease Inhibitor Cocktail) by gently pipetting up and down, followed by centrifugation at 5,000 × g for 10 min at 4°C. The supernatants were discarded and the precipitates were dissolved in 100 µl RIPA buffer (cat. no. R0010, Beijing Solarbio Science & Technology Co., Ltd.) and incubated at −80°C for 30 min; these were the nuclear extracts used for further analyses.

All experiments were performed at least in triplicate independently. All data were expressed as the mean ± SEM and were analyzed using GraphPad Prism 7 (GraphPad Software, Inc.). The significance of the differences between two or multiple groups was evaluated using an unpaired Student's t-test or one-way ANOVA followed by a Tukey's post hoc test. The paired Student's t-test was only applied with the GSE20592 dataset. P≤0.05 was considered to indicate a statistically significant difference. Pearson's correlation test was applied to analyze the correlation between the miR-27a-3p and TAB3 expression levels.

To investigate the expression levels of miR-27a-3p and the other members from miR-23a/27a/24-2 cluster in cervical cancer tissues, expression analyses were performed based on the GSE86100, GSE20592 and TCGA-CESC datasets. As demonstrated in Figs. 1A and S1, the expression of miR-27a-3p was significantly higher in cervical cancer tissues compared with normal tissues, while the expression levels of the other miRNAs were not consistent. Further RT-qPCR analyses were first applied to investigate the effectiveness of the constructed miR-27a-3p strand specific expression vector and the synthesized ASO-miR-27a-3p. As shown in Fig. 1B, the stand-specific miR-27a-3p construct enhanced and ASO-miR-27a-3p decreased the expression of miR-27a-3p in HeLa cells. Subsequently, gain- and loss-of-function experiments were performed to investigate its multiple effects on cell malignancies. The overexpression of miR-27a-3p significantly enhanced the colony formation rates of cervical cancer cells, while the inhibition of miR-27a-3p resulted in decreased colony formation rates in both the HeLa and C33A cells (Fig. 1C). Moreover, miR-27a-3p overexpression inhibited the apoptosis of cervical cancer cells, since the expression of cleaved PARP (89 kDa) increased upon miR-27a-3p inhibition and decreased with the overexpression of miR-27a-3p (Fig. 1D).

Furthermore, the expression levels of different biomarkers in the epithelial-mesenchymal transition (EMT) processes were examined using western blot analyses. As shown in Fig. 1E, the expression of the epithelial cell marker, EpCAM, decreased and that of the mesenchymal cell marker, Vimentin, increased with the overexpression of miR-27a-3p, and vice versa. Transwell migration and invasion assays were used to evaluate the effects of miR-27a-3p in cervical cancer cells. The ectopic expression of miR-27a-3p significantly enhanced the migratory (upper panel) and invasive (lower panel) abilities of the HeLa and C33A cells, and the inhibition of miR-27a-3p significantly decreased the migration and invasion of cervical cancer cells (Fig. 1F). Taken together, these results indicate the cancer-promoting effects of miR-27a-3p in cervical cancer cells.

Since miR-27a-3p was identified as one of the differentially expressed miRNAs associated with NF-κB silencing in previous unpublished screening data by the authors, in the present study, western blot and immunofluorescence analyses were performed to examine the effects of miR-27a-3p on nuclear p65 expression and the potential activation of NF-κB signaling. The sub-fraction extracts examined using western blot analysis revealed the increased nuclear p65 and decreased cytoplasmic p65 expression levels upon the overexpression of miR-27a-3p in the HeLa and C33A cells, and the opposite results were obtained following the inhibition of miR-27a-3p expression (Fig. 1G). The enhanced nuclear p65 signals upon miR-27a-3p overexpression were also observed in the HeLa cells, as evaluated using immunofluorescence analyses (Fig. 1H and I). Further RT-qPCR assays were performed to evaluate the expression levels of NF-κB-dependent genes, such as VEGFB, c-Myc and cyclin D1 upon miR-27a-3p modulation. As shown in Fig. 1J, the mRNA expression levels of VEGFB, c-Myc and cyclin D1 were significantly enhanced/decreased by miR-27a-3p overexpression/inhibition in HeLa cells. Taken together, these data suggest that the enhanced p65 nuclear translocation and the activation of NF-κB signaling are induced by miR-27a-3p overexpression.

To further investigate the regulatory mechanisms of miR-27a-3p and its possible connections with the NF-κB signaling pathway, bioinformatics analyses were first performed to search for its potential targets and focused on the molecules involved in NF-κB signaling. TAB3, MAP3K14 and TRAF3 were first selected for verification using the 3′UTR EGFP fluorescence reporter systems. As shown in Fig. 2A, there was a putative miR-27a-3p binding site in the position 368–390 of 3′UTR of TAB3. When the miR-27a-3p expression vector and the wild-type 3′UTR reporter vector were co-transfected, the normalized EGFP intensities were significantly increased, while the mutant 3′UTR reporter vector co-transfection abolished the enhancement of EGFP intensities induced by miR-27a-3p, suggesting the direct binding of miR-27a-3p with the 3′UTR of TAB3. Further western blot and RT-qPCR analyses revealed the increased TAB3 expression at both the protein and mRNA level upon miR-27a-3p overexpression, and the decreased TAB3 expression upon miR-27a-3p inhibition using ASO-miR-27a-3p in HeLa and C33A cells (Fig. 2B). Further correlation analyses based on the available TCGA-CESC dataset including 303 cervical cancer tissues revealed a significant positive correlation between miR-27a-3p and TAB3 expression levels (Fig. 2C).

Similarly, bioinformatics analyses also revealed the putative miR-27a-3p binding sites present in the 3′UTRs of MAP3K14 (position:1245-1267) and TRAF3 (position:1613-1635), as presented in Fig. S2A. The 3′UTR EGFP fluorescence reporter analyses obtained the similar results as TAB3, i.e., the increased EGFP intensities were induced by miR-27a-3p overexpression, and this enhancement was abolished by co-transfection with the mutated 3′UTR reporter vectors (Fig. S2B). The MAP3K14 and TRAF3 expression levels were also analyzed upon the modulation of miR-27a-3p expression in cervical cancer cells. As shown in Fig. S2C, the mRNA expression levels of MAP3K14 and TRAF3 were significantly increased following miR-27a-3p overexpression and decreased upon miR-27a-3p inhibition in both cervical cancer cells, and enhanced protein expression levels were also observed upon the ectopic expression of miR-27a-3p in HeLa cells (Fig. S2D). Further survival analyses based on the expression levels of TAB3, MAP3K14, TRAF3 and miR-27a-3p from the TCGA-CESC dataset were conducted using online tools. As illustrated in Fig. S3, only MAP3K14 expression was associated with significant changes in the overall survival of patients; i.e., patients with a higher MAP3K14 expression exhibited an improved survival rate; however, no significant associations were found between TAB3, TRAF3 and miR-27a-3p expression and patient survival.

To further investigate the mechanisms responsible for the promoting effects on TAB3 expression induced by miR-27a-3p, the present study first determined whether AGO2 protein participated in this process. By constructing shRNA vectors for AGO2 (shR1-AGO2 and shR2-AGO2) and determining their efficiencies (Fig. S4A), miR-27a-3p was overexpressed in C33A cells with/without shR-AGO2 and the protein expression levels of TAB3 were re-evaluated. As demonstrated in Fig. S4B, a similar upregulated TAB3 expression was observed upon miR-27a-3p overexpression with the inhibition of AGO2, indicating that AGO2 was involved in the upregulation of TAB3 induced by miR-27a-3p. Notably, the globally decreased TAB3 protein expression was also observed upon the silencing of AGO2 compared with the control group, suggesting the potential regulation of TAB3 by other AGO2-associated mechanisms or other AGO2-dependent miRNA regulations.

Overexpression and knockdown vectors were constructed for TAB3 (KBE-TAB3, shR1-TAB3 and shR2-TAB3) for the further investigation of its biological roles. Firstly, their efficiencies were evaluated using western blot analyses with anti-TAB3 and anti-Flag. As shown in Fig. 3A, the KBE-TAB3 vector efficiently enhanced TAB3 expression, and both the shR1-TAB3 and shR2-TAB3 vectors inhibited TAB3 expression. The shR2-TAB3 vector was selected for performing the further loss-of-function experiments, since its silencing effect was more prominent than that of shR1-TAB3. By gain- and loss-of-function transit transfection, the series functional effects of TAB3 on the colony formation, apoptosis, migration and invasion of cervical cancer cells, as well as the expression levels of EMT-associated biomarkers were determined. The results revealed increased colony formation rates upon TAB3 overexpression in cervical cancer cells, and TAB3 inhibition using shRNA significantly decreased the colony formation abilities (Fig. 3B). The decreased expression of cleaved PARP upon TAB3 overexpression and the enhanced expression of cleaved PARP following the silencing of TAB3 expression were observed using western blot analyses in the cervical cancer cells (Fig. 3C).

Transwell migration and invasion assays revealed the elevated migration (upper panel) and invasion (lower panel) properties following the ectopic expression of TAB3, while the significantly decreased migratory and invasive capacities were observed in the TAB3 inhibition groups (Fig. 3D). Moreover, the increased expression of Vimentin, one of the mesenchymal cell markers, and the reduced expression of the epithelial cell marker, EpCAM, were observed following TAB3 overexpression in the cervical cancer cells. The opposite results, i.e., the decreased expression of Vimentin and the increased expression of EpCAM were observed upon the inhibition of TAB3, suggesting that TAB3 promoted the EMT processes (Fig. 3E). Furthermore, the analyses of p65 expression based on the sub-fraction extracts revealed the increased nuclear p65 and decreased cytoplasmic p65 expression following TAB3 overexpression; opposite results were obtained by the silencing of TAB3 expression in both HeLa and C33A cells (Fig. 3F). The immunofluorescence assays also revealed the enhanced p65 nuclear signals upon TAB3 overexpression in HeLa cells (Fig. 3G and H). RT-qPCR analyses to determine the expression of NF-κB-dependent genes (VEGFB, c-Myc and cyclin D1) revealed their significantly increased expression levels upon TAB3 overexpression, while the silencing of TAB3 decreased their expression levels in HeLa cells (Fig. 3I), suggesting the enhanced nuclear p65 expression and the activation of NF-κB signaling by TAB3 overexpression in cervical cancer cells.

Since it was observed that miR-27a-3p and TAB3 exerted similar effects on the aggressive phenotypes of cervical cancer cells and that miR-27a-3p promoted the expression of TAB3, the present study therefore determined whether the promoting effects of miR-27a-3p on the malignant potential of cervical cancer cells were achieved due to its promoting effects on TAB3 expression. A series of functional rescue experiments was conducted by co-transfection with miR-27a-3p + shR-TAB3 and ASO-miR-27a-3p + TAB3. As was expected, the elevated colony formation rates induced by miR-27a-3p overexpression were counteracted by the silencing of TAB3 expression, and the suppressed colony formation properties induced by ASO-miR-27a-3p were rescued by TAB3 overexpression in both HeLa and C33A cells (Fig. 4A). The expression levels of EMT-associated specific molecules, i.e., the decreased EpCAM and increased Vimentin expression induced by miR-27a-3p were restored to their normal levels by co-transfection with shR2-TAB3, and vice versa (Fig. 4B). Furthermore, the knockdown of TAB3 mostly abolished the enhanced migratory (Fig. 4C) and invasive (Fig. 4D) abilities of the cervical cancer cells induced by miR-27a-3p overexpression; in addition, the suppressed migratory and invasive properties induced by ASO-miR-27a-3p transfection were counteracted by the ectopic expression of TAB3. The decreased/increased expression of cleaved PARP induced by miR-27a-3p/ASO-miR-27a-3p was also neutralized by the silencing/overexpressing TAB3 expression (Fig. 4E).

Taken together, the findings demonstrated that the suppression of TAB3 at least partially eliminated the effects generated by overexpression miR-27a-3p, and the ectopic expression of TAB3 abolished the effects induced by the inhibition of miR-27a-3p expression, indicating that TAB3 served as a direct functional target gene of miR-27a-3p, and that the enhanced malignant effects induced by miR-27a-3p overexpression were mediated by its promoting effects on TAB3 expression in cervical cancer cells.

Previous studies have reported the dysregulated expression of miR-27a-3p in cervical cancer samples (25,26). In an unpublished screening study by the authors, miR-27a-3p was identified as one of the dysregulated miRNAs by silencing NFKB1 in human cancer cells. The present study thus further focused on the transcriptional regulation of miR-27a-3p, as well as the miR-23a/27a/24-2 cluster. Using bioinformatics analyses, two putative transcription start sites (TSSs) were identified in the analyzed fragment and two promoter reporter vectors containing the two predicted TSSs were constructed, referred to as pGL3-P1 and pGL3-P2, respectively (Fig. 5A). As demonstrated in Fig. S5A, the relative luciferase activity of the fragment P2 was ~4-fold higher than that of the fragment P1, suggesting that the fragment P2 contained the functional promoter sequences. Further analyses of the fragment P2 sequence identified two potential p65 binding sites presented (Fig. 5A, positions: 214–223 and 546–555 in this 773 bp full length) and a series of truncated promoter reporter vectors with none or either one p65 binding site and flanking sequence were constructed, respectively. Of note, the truncated fragments F259 and F509 had higher promoter activities compared with the full-length fragment P2, and both of these only contained the predicted second p65 binding site (position 546–555), suggesting a critical positive regulatory role for this site, while the other one (position 214–223) appeared to have negative effects (Fig. 5B). Following the verification of the efficiencies of the shR-p65 and KBE-p65 vectors (Fig. S5B), the enhanced/decreased promoter activities of fragment P2 and its truncated form F509 were observed upon the modulation of p65 expression (Fig. 5C). Furthermore, this p65 binding site was deleted in fragment F509 and the promoter activities were re-evaluated using dual luciferase reporter systems. As illustrated in Figs. 5C and S5C and D, the significantly decreased promoter activities induced by the silencing of p65 and the silencing of NFKB1 were abolished, and the increased promoter activities induced by the ectopic expression of p65 or NFKB1/p50 were partially decreased, although these were still significantly higher compared with the wild-type sequence of F509 which included the p65 binding site, suggesting the importance of this binding site.

Furthermore, RT-qPCR assays revealed the increased or decreased primary transcript for this cluster (miR-23a/27a/24-2) and miR-27a-3p expression by p65 overexpression or inhibition, as well as the similar expression patterns for miR-23a-3p and miR-24-2-3p (Fig. 5D), the other two abundant members from the same miRNA cluster, suggesting the members from the same miRNA cluster transcript together. Taken together, these data suggested that the direct binding of p65 to this specific site contributed to the upregulated expression of miR-27a-3p by enhancing its transcription.