The linear ANRIL transcript P14AS regulates the NF-κB signaling to promote colon cancer progression

Abstract

Background

The linear long non-coding RNA P14AS has previously been reported to be dysregulated in colon cancer, but the mechanistic role that P14AS plays in colon cancer progression has yet to be clarified. Accordingly, this study was developed to explore the regulatory functions of ANRIL linear transcript-P14AS in cancer.

Methods

The expression of P14AS, ANRIL, miR-23a-5p and their target genes were detected by quantitative real-time polymerase chain reaction (qRT-PCR) and western blot. Cell supernatants of IL6 and IL8 were measured by Enzyme linked immunosorbent (ELISA) assay. Dual-luciferase reporter assays, RNA immunoprecipitation, or pull-down assays were used to confirm the target association between miR-23a-5p and P14AS or UBE2D3. Cell proliferation and chemosensitivity of NF-κB inhibitor BAY 11-7085 were evaluated by cell counting kit 8 (CCK8).

Results

When P14AS was overexpressed in colon cancer cell lines, enhanced TNF-NF-κB signaling pathway activity was observed together with increases in IL6 and IL8 expression. The Pita, miRanda, and RNA hybrid databases revealed the ability of miR-23a-5p to interact with P14AS, while UBE2D3 was further identified as a miR-23a-5p target gene. The results of dual-luciferase reporter, RNA pull-down, and RNA immunoprecipitation experiments confirmed these direct interactions among P14AS/miR-23a-5p/UBE2D3. The degradation of IκBa mediated by UBE2D3 may contribute to enhanced NF-κB signaling in these cells. Moreover, the beneficial impact of P14AS on colon cancer cell growth was eliminated when cells were treated with miR-23a-5p inhibitors or UBE2D3 was silenced. As such, these findings strongly supported a role for the UBE2D3/IκBa/NF-κB signaling axis as a mediator of the ability of P14AS to promote colon cancer progression.

Conclusions

These data suggested a mechanism through which the linear ANRIL transcript P14AS can promote inflammation and colon cancer progression through the sequestration of miR-23a-5p and the modulation of NF-κB signaling activity, thus highlighting P14AS as a promising target for therapeutic intervention efforts.

Background

The well-known CDKN2A/B locus is associated with human tumors and metabolic diseases. Although this locus has three famous tumor suppressor genes (P16^INK4A, P14^ARF, and P15^INK4B), the antisense strand of this genetic locus encodes the long non-coding RNAs (lncRNAs) ANRIL, as well as ANRIL’s linear transcript, P14AS, the former of which is capable of binding the polycomb repressive complex-1/2 (PRC-1/2) to inhibit the expression of P15^INK4B and P16^INK4. P14AS promotes ANRIL upregulation through binding to AUF1(Ma et al. 2020; Li et al. 2022). Colon cancer (CC) cells exhibit the overexpression of both ANRIL and P14AS, which promote tumor development and malignant cell proliferation in vitro and in vivo(Ma et al. 2020). MicroRNAs (miRNAs) are small 19–25 nucleotide transcripts that can regulate gene expression at both the transcriptional and post-transcriptional levels(Michlewski and Cáceres 2019; Ha and Kim 2014). Many miRNAs have been established as valuable biomarkers associated with various cancer types. Regulatory interactions between ANRIL and specific miRNAs have also been observed in many cancers, such as the interactions between ANRIL and both miR-125a and let-7a in CC and nasopharyngeal cancer(Hu et al. 2017; Wang et al. 2017; Zhang et al. 2018), or the ANRIL-miR-99a/miR-449a regulatory axis in gastric cancer(Zhang et al. 2014). However, whether miRNAs similarly interact with P14AS and how these interactions influence the pathogenesis of CC have yet to be established. Single nucleotide polymorphisms (SNPs) in the ANRIL sequence have been linked to various diseases including type 2 diabetes and coronary atherosclerosis(Kong et al. 2018). Since atherosclerosis develops due to chronic and progressive inflammation, ANRIL expression may be related to inflammation. In endothelial cells, NF-κB and TNFα can function as pro-inflammatory mediators that promote ANRIL upregulation and the expression of cytokines, including IL6 and IL8 through interactions with the transcription factor YY1 via the TNFα-NF-κB- ANRIL/YY1-IL6/8 signaling axis, thereby altering the local endothelial microenvironment(Zhou et al. 2016; Gupta et al. 2020). ANRIL can also drive tumor progression by regulating various pathways including the mTOR, MAPK, PI3K/AKT, and TGF-β(Yu et al. 2018; Wang et al. 2022a, b; Liu et al. 2019; Dong et al. 2018). An analysis of HCT116 cells in which P14AS was stably overexpressed revealed enrichment in KEGG pathways, including the TNF signaling pathway (Corrected P = 0.012). Further research is required to determine whether and how P14AS modulates the TNF-NF-κB signaling axis.

Here, P14AS was demonstrated to function as a molecular sponge that sequesters miR-23a-5p and promotes the upregulation of the miR-23a-5p target protein Ubiquitin Conjugating Enzyme E2 D3 (UBE2D3). Through interactions with the E2 CDC34 and the SCF(FBXW11) E3 ligase complex, UBE2D3 can promote NFκBIA (IκBa) polyubiquitination, thus driving its proteasome-mediated degradation. These results support a potential hypothesis wherein P14AS can promote tumorigenesis and disease progression via the miR-23a-5p/UBE2D3/IκBa pathway-mediated activation of NF-κB signaling activity.

Materials and methods

Cell culture

HCT116, SW480, LOVO, and HEK293T were provided as a kind gift from Professor Dajun Deng at Peking Union Medical College Hospital, and were cultured in DMEM (Corning, VA, USA) containing 10% fetal bovine serum (Gibco, Australia) and 1% penicillin/streptomycin (Gibco, NY, USA) in a 5% CO₂ incubator at 37 °C.

Cell transfection

The PCDH-CMV-EF1a-copGFP-T2A-Puro lentiviral vector was used to generate a P14AS expression construct by Syngentech Co., Ltd. (Beijing, China). The psiCHECK2 vector was obtained from Sangon Biotech Co., Ltd (Shanghai, China). Lipofectamine 3000 (Invitrogen, CA, USA) was used to transfect cells with appropriate miRNAs, siRNAs, or other constructs based on provided instructions. Colon cancer cells were transfected with miRNA mimics (final concentration 100nM) or miRNA inhibitors (final concentration 100nM) or siRNA (final concentration 100nM) for 48-72 h. Post-transfection, cells were collected for CCK8 assay or WB assay or qRT-PCR. Stably transfected HCT116, SW480, and LOVO cells were obtained through culture in media containing puromycin (1 µg/mL, InvivoGen, CA, USA).

CCK-8 assays

A TransDetect Cell Counting Kit-8 (CCK-8, TransGen Biotech, Beijing, China) was used based on the provided directions to assess transfected cell viability. Briefly, 100 µL of media containing either 20,000 or 80,000 cells/mL (for proliferation and cytotoxicity assays, respectively) was added to each well of a 96-well plate. In cytotoxicity assays, after allowing cells to adhere to the plate, media was exchanged for media containing a range of BAY 11-7085 (Selleck, Shanghai, China) concentrations (0, 2, 4, 8, 12, 16, 32µM) or DMSO. At appropriate time points, CCK-8 solution (10 µL) was added per well, followed by an additional 3 h incubation at 37 °C. Absorbance was then measured at 450 and 630 nm with a BioRad microplate reader. Proliferation was assessed once daily on 5 sequential days, and average absorbance values were measured.

RNA extract and qPCR

Trizol (TransGen Biotech, Beijing, China) was used to extract cellular RNA, after which a TransScript First-Strand cDNA Synthesis SuperMix (Roche, IN, USA) was used to prepare cDNA. Then, FastStart Universal SYBR Green Master (ROX) (TransGen Biotech, Beijing, China) was used to conduct qPCR analyses with an ABI-7500 Fast system (Applied Biosystems) with the following settings: 94 °C for 30 s; 40 cycles of 94 °C for 5 s, 60 °C for 15 s, 72 °C for 34 s. GAPDH was used as a normalization control. Primers used for this study are listed in Supplementary Table S1.

Western blotting

NP-40 buffer (Solarbio, Beijing, China) supplemented with protease inhibitors (LabLead, Beijing, China) was used to extract protein from cells, and these samples were then separated via SDS-PAGE and transferred onto PVDF membranes (Merck Millipore). Following a 1 h room temperature blocking step using 5% skim milk, the membranes were probed with appropriate primary antibodies overnight at 4 °C followed by probing for 1 h at room temperature with secondary anti-mouse or anti-rabbit IgG (ZSGB Biotech, Beijing, China). A chemiluminescence detection system (Cytiva, Amersham ImageQuant 800, Japan) was then used for protein detection. The primary antibodies used were specific for AUF1 (Abcam, ab61193, UK), CDCP1 (Abcam, ab252947), ADAM10 (Abcam, ab124695), UBE2D3 (Abcam, ab176568), IκBa (Abcam, ab32518), Ago2 (Abcam, ab186733), YY1 (Santa Cruz, sc-7341, TX, USA), HNF3a (Santa Cruz, sc-514,695), GAPDH (Protein Tech, 50430-2-AP; China), β-tubulin (Abcam, ab6046). Antibodies were used at dilutions ranging from 1:1000-1:5000.

RNA pull-down assay

Biotinylated P14AS probes (#1-#6) and corresponding control probes (#1-#2) that had been synthesized in vitro by RiboBio (Guangzhou, China) were incubated with SW480 cell lysates, after which an RNA pull-down assay was performed based on protocols published previously(Ma et al. 2020).

RIP assay

An RNA-Binding Protein Immunoprecipitation Kit (Cat# 17–701, EZ Magna, Millipore, USA) was used based on the directions provided. Total RNA bound to AUF1 or Ago2 was precipitated and isolated with anti-AUF1 (Abcam, ab61193, Cambridge, UK) or anti-Ago2 (Abcam, ab186733).

Luciferase reporter assay

HindIII was used to insert the full-length P14AS sequence containing wild-type or mutated binding sites for target miRNAs (miR-23a-5p, miR-6855-5p, and miR-6842-5p) into the pGL3-control vector (Promega, WI, USA). The 3’-UTR sequences for target genes or mutant isoforms were inserted into the psiCHECK2 vector using XhoIand NotI(Sangon Biotech, Shanghai, China). For further details regarding the binding sites used to generate these plasmids, see Fig. 3E. After adhering overnight to 12-well plates, cells were co-transfected with WT or mutated luciferase reporter vectors together with miRNA mimics (final concentration 100nM), inhibitors (final concentration 100nM), or control vectors and incubated for 48 h. A dual luciferase reporter assay kit (Promega, WI, USA) was used based on the directions provided.

ELISA

Supernatants were collected from cells that had been stably transfected, and the concentrations IL-6 (EH004-96) and IL-8 (EH005-96) therein were measured using appropriate ELISA kits (ExCell, Shanghai, China) based on the manufacturer’s instructions.

ChIP assay

A ChIP kit (Beyotime Biotech, Beijing, China) was used based on the provided directions, after which the fragments of DNA precipitated using YY1 (Santa Cruz, sc-7341, TX, USA), HNF3a (Santa Cruz, sc-514,695), and control IgG were analyzed via qPCR.

RNA-seq and data analyses

The transcriptomes of the stably P14AS-overexpressed HCT116 cells were sequenced by RiboBio Co., Ltd. (Guangzhou, China). The data sets were deposited in the Gene Expression Omnibus (GEO) database with accession number GSE127905. KEGG pathway analysis for differentially expression genes was performed using KOBAS3.0 software (http://www.genome.jp/kegg). P14AS-binding miRNAs and miRNA-target proteins were analyzed using the miRanda, Pita, RNAhybrid, and TarPmiR databases.

Statistical analyses

Data were compared using two-sample Student’s t-tests and presented using GraphPad Prism 7.0 software (Dotmaticus, USA). Results are reported as means ± SD. Pearson correlation analyses were used to assess relationships between variables. Two-sided statistical tests were used for all analyses. *P < 0.05, **P < 0.01, N.S: not significant.

Results

P14AS positively regulates TNF signaling in HCT116 cells

In a prior study exploring putative ncRNAs encoded in the CDKN2A/B locus of the human genome, we screened several RNAs through a CDKN2A-specific RNA capture deep-sequencing approach, ultimately leading to the PCR and RNA FISH-based validation of the lncRNA P14AS. High levels of P14AS expression were detected in CC tissues, and it was able to drive the P16-independent proliferation of tumor cells (Ma et al. 2020). To further explore the mechanisms through which P14AS can regulate growth-related signaling activity, an RNA sequencing (RNA-seq) analysis was conducted comparing P14AS-overexpressing (P14AS OE) and control (P14AS Ctrl) cells to detect changes in gene expression as a function of P14AS levels. A high degree of correlation in gene expression patterns was observed when comparing two biological replicate samples, emphasizing the consistent effects of P14AS on patterns of gene expression (Fig. 1A). In KEGG pathway enrichment analyses exploring the functional mechanisms associated with P14AS, significant TNF and PI3K-Akt signaling pathway enrichment was detected among P14AS-related genes (FC > 1.2, P < 0.05) (Fig. 1B). Signaling through the TNFα-NF-κB-IL6/IL8 axis has reportedly been linked to ANRIL(Zhou et al. 2016). We detected TNF/IL6/IL8 expression in P14AS OE and Ctrl cells by RT-PCR and the results showed that the expression of TNF/IL6/IL8 was elevated in the P14AS OE cells compared to the Ctrl cells (Fig. S1A). Western blotting analysis revealed that the P14AS OE group exhibited higher P65 protein expression than the Ctrl group (Fig. S1B). When TNF factor (100nM) was used to stimulate P14AS OE and P14AS Ctrl cells for 6 h, this resulted in the continuous upregulation of P14AS (Fig. 1C). Moreover, qRT-PCR and ELISA analyses demonstrated that P14AS overexpression was associated with higher IL6 and IL8 expression in CC cells, supporting an association between P14AS and NF-κB-IL6/IL8 pathway regulation (Fig. 1D-E).

Fig. 1

TNF-NF-κB-IL-6/IL-8 signaling is regulated by P14AS in CC cells. (A) High correlation of gene expression between two biological repeats for RNA-seq. (B) TNF signaling regulated by P14AS using KEGG pathway analysis. (C) TNFa factor stimulated P14AS OE cells. (D, E) Detection of IL6 and IL8 expression in P14AS OE cells by qRT-PCR and ELISA

P14AS serves as a molecular sponge for miR-23a-5p and thereby regulates UBE2D3 expression

To examine the mechanisms whereby P14AS can control the TNF-NF-κB signaling axis, the miRanda, RNA hybrid, and Pita tools were leveraged to detect overlap between P14AS and miRNA target genes, identifying several genes that have previously been annotated in RNA-sequencing datasets assessing the differential expression of ncRNAs. In total, 127 miRNAs were screened for interaction with P14AS using miRanda, RNA hybrid, and Pita databases (Table 1). Among them, only three miRNAs miR-6855-5p, miR-6842-5p, and miR-23a-5p were found to be associated with P14AS-associated proteins (Fig. 2A). To test the ability of these miRNAs to directly interact with P14AS in HCT116 cells, a dual-luciferase reporter assay was next conducted. While a significant reduction in luciferase activity was evident for cells co-transfected with a wild-type P14AS vector and a miR-23a-5p mimic, the same was not true following miR-6855-5p or miR-6842-5p mimic transfection. In addition, the mutation of the putative binding sequences in the P14AS reporter vector abrogated the ability of miR-23a-5p mimic transfection to suppress luciferase activity in both HCT116 and SW480 cells (Fig. 2B-C). A group of biotin-conjugated P14AS probes was also able to pull down miR-23a-5p in P14AS OE cells (Fig. 2D). Argonaute2 (AGO2) is a key protein member of the RNA-induced silencing complex (RISC) that controls miRNA function(Sheu-Gruttadauria et al. 2019). The use of anti-AGO2 to conduct RNA immunoprecipitation (RIP) using lysates prepared from SW480 cells resulted in P14AS and miR-23a-5p enrichment, with this enrichment being stronger for P14AS relative to miR-23a-5p (Fig. 2E). Overexpression of P14AS led to increased levels of the miR-23a precursor transcript (Table 2) and qPCR confirmed the upregulation of miR-23a-5p (Fig. 2F).

Table 1 P14AS-binding miRNAs predicted by miRanda/RNA hybrid/Pita databases

Fig. 2

P14AS functions as miR-23a-5p sponge in CC cells. (A) miRanda\Pita\RNAhybrid databases predicted P14AS binding of miRNAs. (B) HCT116 cells that were co-transfected with miR-6855-5p or miR-6842-5p or control mimics and P14AS wild type luciferase report vector were measured for luciferase activity. (C) Dual-luciferase reporter assays were conducted with wild type and mutated (putative binding sites for miR-23a-5p was mutated) luciferase report vectors to validate P14AS/miR-23a-5p axis. (D)miR-23a-5p was pulled down by biotin labelled P14AS probes. (E) RIP assay using AGO2 antibody found that the interaction with P14AS or miR-23a-5p and AGO2 antibody in SW480 cell. (F) The cells transfected with P14AS vector were analyzed by qRT-PCR for miR-23a-5p expression

Table 2 The expression profiles of P14AS OE cells in RNA-seq

RNA-seq data led to the identification of four target genes, including CDCP1, ADAM10, CMTM6, and UBE2D3 (Fig. 3A). Of these targets, CDCP1 is reportedly associated with PI3K/AKT signaling(Khan et al. 2021), CMTM6 is a critical regulator of PD-L1 in a broad range of cancer cells(Burr et al. 2017), while UBE2D3 is linked to NF-κB signaling(Chen et al. 2017) and ADAM10 plays a role in TNF signaling(Arduise et al. 2008). Based on the results of qRT-PCR analyses, P14AS overexpression was found to promote both CDCP1 downregulation and UBE2D3 upregulation in HCT116 and SW480 cells (Fig. 3B), while CMTM6 expression was presented in Fig. S1C. As expected, knocking out the AU-rich element (ARE) region recognized with AUF1 found in the first P14AS exon (P14AS KO)(Ma et al. 2020) suppressed UBE2D3 in HCT116 cells (Fig. 3C), as further confirmed via Western blotting (Fig. 3D). The miR-23a-5p binding sites in the 3’-UTR regions of CDCP1 and UBE2D3 were next analyzed and used to construct WT and mutant CDCP1-3’UTR and UBE2D3-3’UTR luciferase reporter vectors. Co-transfection of CDCP1-3’UTR and UBE2D3-3’UTR vectors (Luc-CDCP1-3’UTR WT and Luc-UBE2D3-3’UTR WT) with miR-23a-5p mimics led to a reduction in luciferase activity. However, this reduction was not observed with the mutated Luc-CDCP1-3’UTR MUT and Luc-UBE2D3-3’UTR MUT constructs (Fig. 3E). Given the consistent changes in UBE2D3 expression observed with changes in P14AS expression levels, it was selected as a target for further analysis. The overexpression of miR-23a-5p resulted in UBE2D3 upregulation at the mRNA and protein levels in CC cells, whereas silencing miR-23a-5p had the opposite effect (Fig. 3F-G). Overexpressing P14AS promoted UBE2D3 upregulation in these cells, while miR-23a-5p inhibitor transfection reversed this effect (Fig. 3H). These results thus indicate that P14AS can function as a miR-23a-5p sponge or compete with mir-23a-5p for AGO2 binding, thereby indirectly regulating UBE2D3 expression in CC cells.

Fig. 3

UBE2D3 is a direct target of miR-23a-5p in CC cells. (A) RNA network of P14AS/miR-23a-5p/target genes overlapped with TNF-NF-κB signaling pathway annotated in RNA-seq data. (B) P14AS OE transcriptionally changed the expression levels of miR-23a-5p target genes. (C) P14AS KO transcriptionally changed the expression levels of miR-23a-5p target genes. (D) P14AS OE or KO changed the protein levels of miR-23a-5p target genes in CC cells. (E)miR-23a-5p overexpression reduced the fluorescence intensity of vector containing wild type (wt) 3’UTR of UBE2D3 and CDCP1, but did not have impact on the fluorescence intensity of mutated (mut) 3’UTR of UBE2D3 and CDCP1. (F, G)miR-23a-5p overexpression increased the expression of UBE2D3 and downregulation of miR-23a-5p reduced the expression of UBE2D3. (H) P14AS OE increased the protein level of UBE2D3, which was reduced by miR-23a-5p inhibitors in CC cells

The oncogenic activity of P14AS is associated with the UBE2D3-mediated degradation of IκBa and consequent NF-κB signaling activity

In prior reports, UbcH5c/UBE2D3 has been shown to facilitate the polyubiquitination of the inhibitory IκBa protein and its consequent proteasomal degradation, thereby enhancing NF-κB-dependent inflammation(Qi et al. 2022). UbcH5c/UBE2D3 can further mediate PCNA and histone H2A ubiquitination, thus influencing general transcriptional activity, DNA damage responses, and the replication of genomic DNA(Sakasai et al. 2023). Given these prior findings, further experiments were conducted exploring the role of UBE2D3 in CC and its associations with P14AS and miR-23a-5p. Initially, the impact of P14AS on signaling via a UBE2D3/IκBa/NF-κB pathway was examined through analyses of IκBa levels in P14AS OE and KO cells, with β-tubulin as a reference control. Overexpressing P14AS was associated with significant decreases in intracellular IκBa levels, whereas the opposite effect was observed in P14AS KO HCT116 cells (Fig. 4A). Notably, the treatment of HCT116 cells overexpressing P14AS with MG132 (a proteasome inhibitor) to inhibit proteasomal activity for 24 h resulted in the restoration of IκBa protein levels (Fig. 4B).

Fig. 4

The regulatory effect of UBE2D3 on NF-κB signaling activation. (A) P14AS OE reduced the IκBa protein level in CC cells and P14AS KO increased the IκBa protein level in HCT116 cells. (B) Restoration of IκBa protein levels after treatment of P14AS OE in HCT116 cells with proteasome inhibitor MG132. (C) UBE2D3 protein level was knock down by siRNAs in CC cells. (D) P14AS OE reduced the protein level of IκBa, which was improved by miR-23a-5p inhibitors in CC cells. (E) P14AS OE reduced the IκBa protein level, which was improved by UBE2D3 siRNA in CC cells. (F) IC50 was measured for P14AS OE and Ctrl cells after BAY 11-7085 treatment

Subsequently, P14AS OE of HCT116 and SW480 cells were transfected using miR-23a-5p inhibitor or UBE2D3 siRNA constructs. As expected, UBE2D3 protein levels were significantly reduced at 48 h after UBE2D3 siRNA transfection (Fig. 4C). Western blotting additionally demonstrated the restoration of IκBa protein levels in these cells following the silencing of miR-23a-5p or UBE2D3 (Fig. 4D-E). These data were thus consistent with a model wherein P14AS can promote the proteasome-mediated destruction of IκBa following UBE2D3 upregulation, thus activating NF-κB signaling activity. To further validate these findings, the established NF-κB inhibitor BAY 11-7085 (7085) was used to treat these cells, revealing lower IC50 values for this compound in P14AS OE cells together with a greater sensitivity to 7085-induced cytotoxic cell death (Fig. 4F). BAY 11-7085 inhibitor (4µM) in P14AS OE cells was used and was found to significantly reduced cell proliferation activity (Fig. S2). Further analyses thus warrant exploring the potential clinical benefits of using 7085 or other NF-κB inhibitors in CC.

P14AS has previously been demonstrated to promote tumorigenesis and tumor cell proliferation in NOD/SCID mice (Ma et al. 2020). To explore whether this oncogenic activity is at least partially dependent on miR-23a-5p and UBE2D3 in CC cells, CCK-8 assays were performed to assess cellular proliferation. This approach revealed a significant increase in proliferative activity following the overexpression of miR-23a-5p, while the opposite was observed when miR-23a-5p was knocked down (Fig. S3). Consistently, the silencing of miR-23a-5p eliminated the P14AS-associated enhancement of CC cell growth (Fig. 5A). When a siRNA construct was used to transiently knock down UBE2D3 in P14AS OE cells (Fig. S4), this UBE2D3 silencing was found to be sufficient to reverse P14AS-associated proliferative activity (Fig. 5B). These findings underscore that downregulation of miR-23a-5p or UBE2D3 can counteract the pro-proliferative effects of P14AS overexpression in CC cells. This supports the role of this lncRNA in promoting tumor growth and development through the UBE2D3/IκBa-mediated NF-κB regulatory axis.

Fig. 5

P14AS-miR-23a-5p-UBE2D3 axis is required for the CC cell growth. (A, B) Co-transfection with miR-23a-5p inhibitors or siUBE2D3 mitigated the carcinogenic effect of P14AS on cell proliferation in CC cells

YY1 and HNF3a suppress ANRIL and P14AS transcriptional activity

To establish which factors are responsible for regulating ANRIL and P14AS, transcription factors predicted to bind to the 2.5 kb region upstream of these transcripts were predicted using the PROMO database (http://alggen.lsi.upc.es/cgi-bin/promo_v3/promo/promoinit.cgi?dirDB=TF_8.3). In total, 20 factors were identified as potential regulators of P14AS promoter activity. In general, transcription factors tend to positively regulate target gene expression. GEPIA database analyses (http://gepia.cancer-pku.cn/index.html) indicated that of these factors, YY1, GATA-1, NFYC, CEBPB, and HNF3a were correlated with the expression of CDKN2B-AS (ANRIL) (Fig. 6A). In COAD samples, GATA-1 and NFYC downregulation was observed, whereas high levels of YY1 and HNF3a expression were detected (Fig. 6B), in line with the high levels of intratumoral P14AS and ANRIL expression that was detected. We down-regulated the expression of YY1 and HNF3a in CC cells and assessed cell proliferation using the CCK8 assay. The results showed that the knockdown of the YY1 and HNF3a group significantly down-regulated the proliferative ability of the cells compared with the control group (Fig. S5). A ChIP-PCR experiment confirmed the ability of YY1 and HNF3a to bind the P14AS and ANRIL promoter region (Fig. 6C). When siRNA constructs were used to knock down the expression of YY1 and HNF3a (Fig. 6D), reductions in both P14AS and ANRIL levels were observed (Fig. 6E), confirming the ability of YY1 and HNF3a to serve as positive regulators of both ANRIL and P14AS.

Fig. 6

The characterization of ANRIL and P14AS promoter. (A, B) Analysis of transcription factors by PROMO and GEPIA databases. The orange box group represents colon adenocarcinoma (COAD) tissue and the grey box group represents normal tissue in Fig. 6B. (C) CHIP-PCR of the promoter region of P14AS by YY1 and HNF3a antibody. (D) YY1 and HNF3a protein level were knock down by siRNAs in HCT116 cell. (E) The transcript levels of P14AS and ANRIL were detected by qRT-PCR

Analyses of ANRIL and UBE2D3 target genes in COAD

The GEPIA and ENCORI databases were next leveraged to assess correlations between the expression of ANRIL (CDKN2B-AS1) and target genes. In COAD samples, ANRIL expression was closely associated with IL6, CXCL8 (IL8), and UBE2D3 levels (Fig. 7A). A positive correlation was also detected between UBE2D3 and the expression of both IL6 and IL8 at the mRNA level (P < 0.05), while no corresponding correlation was detected with NFKBIA (IκBa) expression (Fig. 7B). This suggests that UBE2D3 post-transcriptionally regulates IκBa protein levels. Our analysis of COAD tissues and normal tissues in the ENCORI database showed that NFKBIA (IκBa) was lowly expressed in colon cancers, and IL6 and IL8 were highly expressed in colon cancers (Fig. 7C), which suggests that the downstream regulatory molecules of UBE2D3 play an important role in the development of tumorigenesis. The results indicated that while UBE2D3 expression did not significantly differ between tumor and normal tissues, it influenced the expression of IL6 and IL8, downstream factors of the NF-κB signaling pathway by regulating IκBa expression.

Fig. 7

Databases analysis of ANRIL and its target genes in COAD. (A) The correlations between CDKN2B-AS1 and IL6, IL8 and UBE2D3 by using ENCORI tool. (B) The correlations between UBE2D3 and NFKBIA, IL6 and IL8 by using ENCORI tool. (C) Analysis of NFKBIA, IL6 and IL8 by ENCORI tool. The orange box group represents colon adenocarcinoma (COAD) tissue and the grey box group represents normal tissue

Discussion

ANRIL is encoded in the same locus of chromosome 9 that encodes the key tumor suppressor genes CDKN2A and CDKN2B. Prior research has highlighted associations between ANRIL and a range of pathological conditions such as atherosclerosis, type 2 diabetes, and obesity(Razeghian-Jahromi et al. 2022). The dysregulation of ANRIL has also been reported in a range of cancers including pancreatic cancer(Wang et al. 2022a, b), cervical cancer(Zhao et al. 2018), breast cancer(Xu et al. 2017; Ma et al. 2022), gastric cancer(Kangarlouei et al. 2019) and multiple myeloma(Wang et al. 2020). P14AS was first identified as a novel hypothetical linear ANRIL transcript, which was previously found to promote ANRIL upregulation and enhance tumor cell proliferation(Ma et al. 2020). While P14AS is related to CC progression, the specific mechanisms linking it to oncogenic activity are not yet fully established. As such, this study was designed to clarify the molecular pathways through which P14AS can influence CC cell proliferation.

NF-κB signaling plays an integral role in various physiological and pathological settings, and is mediated by five transcription factor subunits that include RelA (p65), RelB, c-Rel, p105/p50, and p100/p52. Heterodimers of RelA-p50 and RelB-p52 can promote canonical and non-canonical NF-κB pathway activation. Upon phosphorylation, members of the inhibitory IκB protein family (IκBα, IκBbβ, IκBε) undergo ubiquitination and proteasomal degradation. This process frees the heterodimerized RelA-p50, allowing it to translocate into the nucleus and promote target gene upregulation(Yu et al. 2020). A growing body of research has revealed that NF-κB pathway dysregulation can contribute to the incidence of cancer and various other forms of inflammation-related disease. In liver cancer, canonical NF-κB signaling can promote inflammation and the survival of hepatocytes(He and Karin 2011), whereas this same signaling axis regulates the invasion, proliferation and migration of ovarian cancer cells(Yang et al. 2018). This pro-inflammatory canonical NF-κB signaling also acts in host defenses against injury or infection. Still, chronic infections or inflammatory activity can contribute to the incidence of certain cancers, with hepatitis B virus infection having been linked to the development of hepatocellular carcinoma (HCC). In contrast, infection with Helicobacter pylori is associated with an increased risk of gastric cancer, while inflammatory bowel disease is linked to an elevated risk of colorectal cancer (CRC)(Yu et al. 2020). Beyond its role in the initiation of tumorigenesis, the activation of NF-κB signaling also influences hormone-dependent breast cancer progression such that the use of an IKK inhibitor as a co-treatment can promote Akt upregulation in MCF7 cells, thereby overcoming endocrine resistance(Zhou et al. 2005). Through interactions with specific miRNAs, ANRIL can regulate the expression and activity of NF-κB pathway components, thereby influencing inflammatory activity and other cancer-related processes. In an ischemic stroke model system, for example, the knockdown of ANRIL reportedly leads to the alleviation of neuroinflammation through the miR-671-5p/NF-κB axis(Deng et al. 2022), whereas circANRIL silencing can suppress NF-κB and JNK/p38 pathway activity via promoting miR-9 upregulation, shielding renal tubular epithelial cells against damage induced by lipopolysaccharide exposure(Deng et al. 2019). Here, miR-23a-5p was found to interact with the novel linear ANRIL transcript designated P14AS.

Prior evidence has suggested that miR-23a-5p exhibits oncogenic activity in bladder cancer(Li et al. 2018) and renal cell carcinoma(Quan et al. 2017), and it can additionally influence tumorigenesis and disease progression through interactions with specific lncRNAs or mRNAs. In glioblastoma, for example, the lncRNATPT1-AS1 can promote growth activity through the sequestration of miR-23a-5p(Gao et al. 2021), whereas the lncRNA FLVCR1-AS1 can enhance cervical cancer cell malignancy via the miR-23a-5p/SLC7A11 axis(Zhou et al. 2022) and hepatoblastoma progression is at least partially driven by the SNHG9/miR-23a-5p/Wnt3a signaling pathway(Feng et al. 2021). Besides affecting tumor progression, miR-23a-5p could modulate the innate host defense by promoting mycobacteria survival and inhibiting the activation of autophagy against Mycobacterium tuberculosis (M.tb.) through TLR2/MyD88/ NF-κB pathway by targeting TLR2(Gu et al. 2017), which suggests that mir-23a-5p may affect the NF-κB pathway. Furthermore, we demonstrated that UBE2D3 was a direct target of miR-23a-5p. In contrast to the traditional model wherein miRNAs inhibited the expression of target proteins via composing RNA-induced silencing complex, miR-23a-5p mimics elevated the protein level of UBE2D3 in the current study. Previous studies have reported that P14AS upregulates the transcription of ANRIL by interacting with AUF1 through the ARE region of the first exon(Ma et al. 2020). The results of RT-PCR from AUF1 antibody-RIP assays showed an enrichment of UBE2D3 mRNA in the AUF1 antibody group compared to the IgG group, potentially explaining the observed increase in UBE2D3 protein level (Fig. S6). Through cooperation with the E2 ligase CDC34 and the SCF E3 ligase complex, UBE2D3 can promote IκBa polyubiquitination such that it is degraded by the proteasome, in turn activating inflammatory signaling through an NF-κB dependent pathway. The present data thus provide strong support for the existence of the P14AS/miR-23a-5p/UBE2D3/IκBa regulatory axis. By serving as a molecular sponge for miR-23a-5p, P14AS can promote the enhanced expression of UBE2D3 and the consequent degradation of IκBa, thereby augmenting NF-κB signaling activity at UBE2D3 in CC cells and driving their ongoing proliferation. This enhanced NF-κB pathway activation may also contribute to a degree of chemosensitivity in these tumor cells, as levels of ANRIL expression have been found to correlate with tumors’ sensitivity to chemotherapeutic treatment(Zhou et al. 2021). For example, high levels of ANRIL expression were negatively correlated with chemotherapeutic responses in those patients with CRC undergoing treatment with a 5-FU-based regimen such that inhibiting ANRIL may represent a viable approach to chemosensitization(Zhang et al. 2018). Given that NF-κB can promote tumor progression in CC in the context of high P14AS expression levels, selectively inhibiting signaling via the canonical NF-κB pathway may represent an effective means of treating CC in the clinic. To confirm that this P14AS/miR-23a-5p/UBE2D3/IκBa axis specifically mediates NF-κB pathway activation, BAY 11-7085 was used as a selective inhibitor of IκBa phosphorylation, and IC50 values for this compound were measured in both P14AS OE and Ctrl cells. The P14AS OE cells exhibited a lower IC50 value relative to corresponding control cells, indicating that CC patients expressing higher levels of P14AS may be more susceptible to the effects of NF-κB signaling inhibitor treatment. Further studies are warranted to determine whether the NF-κB pathway is the sole downstream target of P14AS or if P14AS exerts its oncogenic effects through other molecular mechanisms.

While these results offer new mechanistic insights into the ability of P14AS to regulate NF-κB signaling, they are nonetheless subject to some limitations. The P14AS/UBE2D3/IκBa regulatory axis was validated only in cells overexpressing P14AS. Cells with the ARE region knocked out in the first P14AS exon (P14AS KO) were not used. This is because previous studies have shown that P14AS KO cells compensate by upregulating a mutated form of P14AS. Since the target sites for miR-23a-5p are present in the mutated P14AS sequence, P14AS KO cells could not be used in the experiments. In addition, the in vivo roles of UBE2D3 as a regulator of oncogenesis were explored by subcutaneously implanting HCT116 cells in which UBE2D3 was or was not silenced in immunodeficient mice, ultimately revealing no significant differences in tumor size or intratumoral UBE2D3 expression (Fig. S7). Further research on the oncogenic roles of UBE2D3 is warranted. These findings suggest that UBE2D3 may primarily drive inflammatory activity and NF-κB pathway activation related to P14AS’s carcinogenic effects.

Conclusions

The study highlights a novel association between the linear ANRIL-P14AS transcript, miR-23a-5p, and the UBE2D3/IκBa signaling axis in CC cells. Specifically, P14AS regulated NF-κB signaling activity and cellular proliferation by ultimately influencing the expression of UBE2D3 and IκBa as a direct or indirect factor (Fig. 8). Overall, these findings offer novel insight into the mechanisms that link ANRIL isoforms and the NF-κB pathway, potentially aiding in developing ANRIL-related biomarkers and therapeutic targets in patients suffering from CC.

Fig. 8

P14AS/mir-23a-5p/UBE2D3/ IκBa regulates the NF-κB signaling