RNA helicase p68 deploys β-catenin in regulating RelA/p65 gene expression: implications in colon cancer

Formalin-fixed paraffin-embedded (FFPE) post-surgical human normal (n = 22) colon and colon carcinoma (n = 45) samples were collected from Indian patients. All clinical and ethical regulations were followed for collecting the samples, including patient consent and with formal approval from the ethical committee of both CSIR-IICB and Park Clinic (source) [19].

Human colorectal cancer (CRC) cell lines [HCT 116, SW-480, HT-29, HCT-15]; human embryonic kidney cell line [HEK-293] and murine adenocarcinoma cells [CT26] were procured from ATCC (Manassas, VA, USA). The cell lines were cultured in McCoy’s 5A, DMEM or RPMI-1640, as required, supplemented with 10% heat-inactivated fetal bovine serum (FBS) and maintained at 5% CO2, 37 °C in a humid incubator. Penicillin/Streptomycin cocktail and Gentamycin (Invitrogen, Life Technologies) were added as prescribed antibiotic at the recommended dose. Transfections of different DNA constructs were performed using Lipofectamine 2000 (Invitrogen). Lipofectamine RNAi MAX (Invitrogen) was used for siRNA-mediated knockdown experiments. Small interfering RNAs (siRNAs) against β-catenin, p68 and TCF4 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). CT26 cells were transfected using Lipofectamine LTX (Life Technologies), G418 (Calbiochem, Darmstadt, Germany) (500 μg/ml) was used for selection of stable cell line.

p68 was sub-cloned from pSG5-Myc vector into pGZ21dx, pEGFP-c3 and pIRES-hrGFP-1a. pBI-β-catenin was sub-cloned into pGZ21dx vector and human Wnt3a gene was cloned in pcDNA3.1-myc-his, respectively. RelA promoter region (1746 bp fragment) was PCR amplified using genomic DNA from HEK-293 cells and following standard protocol cloned into pGL3 basic vector. ALGGEN-PROMO (http://​alggen.​lsi.​upc.​es/​) was used for putative analysis of the transcription factors binding to this region. All the constructs were verified by restriction digestion and confirmed by sequencing. All primers were synthesized from IDT (Integrated DNA Technologies, Coralville, IA, USA). Sequences of the primers are given in Additional file 5: Table S1.

Deletion of TBE consensus sites (CTTTG [sites 2,4,6,7]/CAAAG [sites 1,3,5]) in the RelA promoter construct were performed using the QuickChange XL Site-Directed Mutagenesis kit (Agilent Technologies, Santa Clara, CA, USA), following manufacturer’s instructions. All deletion constructs were verified by DNA sequencing. Sequences of the primers are given in Additional file 5: Table S1.

Two different shRNAs targeting p68 and a control (scrambled) shRNA were subcloned in pMKO.1-Puro according to Addgene’s protocol. shRNA targeting β-catenin was purchased from Addgene (pLKO.1 puro shRNA β-catenin) (Addgene Cat no 18803). Scramble siRNA, p68, β-catenin and TCF4 siRNA (Santa Cruz Biotechnology, Inc., Dallas, TX, USA) were used at a final concentration of 30 nM.

Preparation of whole cell lysates (WCL), cytoplasmic and nuclear extracts and immunoblot analyses were performed as described previously [20, 21]. The following primary antibodies were used: p68 (Abcam); Cyclin D1, RelA/p65, Bcl-2, XIAP, Bcl-xL, Survivin, TCF4 (Cell Signaling Technology); α-Tubulin, Lamin B, Actin, GFP, β-catenin and GAPDH (Santacruz Biotechnology); HRP-tagged anti-rabbit and anti-mouse secondary antibodies (Cell Signaling Technology); HRP-tagged anti-goat secondary antibody (Sigma-Aldrich). GelQuant. Net software (http://​biochemlabsoluti​ons.​com/​GelQuantNET.​html) was used for densitometric scanning of immunoblots. Blots were evaluated by using Luminata Classico Western HRP Substrate (Millipore) following manufacturer’s protocol.

Total RNA was extracted using the Trizol reagent (Invitrogen, NY, USA), converted to cDNA (1 μg of RNA for each sample) using RevertAid H Minus First Strand cDNA Synthesis Kit (Thermo Scientific™) which was subsequently used for qRT-PCR analysis using FastStart Universal SYBR Green Master (Roche) on a ViiA 7 Real-Time PCR system (Applied Biosystems). In all the experiments, 18S rRNA served as the internal control. Quantification and Standard deviation (SD) calculations were based on three independent experiments in triplicates. Sequences of the primers are given in Additional file 5: Table S1.

Cells were transiently transfected with pGL3-RelA promoter luciferase reporter construct along with Renilla luciferase plasmid (pRL-TK) and respective gene constructs as per experimental interest and indicated in the relevant figure(s). Luciferase Assay was performed with the cell lysates 36 or 48 h post-transfection following the manufacturer’s protocol. The efficiency of transfection was normalized with Renilla luciferase expression. Quantification was based on three independent experiments. Luciferase activity was determined luminometrically in the Varioskan Flash Multimode Reader (Thermo Fisher Scientific, Waltham, MA, USA) by the dual luciferase assay system (Promega, Madison, WI, USA) following manufacturer’s instructions and as described previously [22].

The preparation of chromatin and the immunoprecipitation of such for ChIP assay were performed as described earlier [23, 24]. ChIP Quantitative RT-PCR (qRT-PCR) reactions were performed using SYBR Green master mix. Sequences of the primers are given in Additional file 5: Table S1.

For histology and IHC analysis, 5 μm thick sections from both human patient samples and murine allograft tumors were prepared from paraffin-embedded blocks. For histological analysis, the tissue sections were subjected to hematoxylin and eosin (H&E) staining following standard protocol. IHC was performed using SuperSensitive™ Polymer-HRP IHC Detection System (BioGenex), following manufacturer’s instructions and as described previously [19]. Images were captured using EVOS XL microscope (Life technologies) at 200X optical magnifications. A semiquantitative scoring method was employed for assessing the intensity of staining and generation of H-score which was used for further statistical analysis. Mayer’s Hematoxylin (#MHS1; Sigma, St. Louis, MO, USA) was used as a counter stain for nucleus. The following primary antibodies were used: p68, Snail (Abcam); RelA, TCF4, PCNA, Bcl-2, Bcl-xL, XIAP, Survivin and Vimentin (Cell Signaling Technology); β-catenin and Bcl-2 (Santacruz Biotechnology).

All animal care and experimentation were conducted according to the CPCSEA guidelines; with approval of the animal ethics committee at CSIR-Indian Institute of Chemical Biology. Murine adenocarcinoma cell line CT26 was used for generation of syngeneic tumor model.

At least three independent experiments were performed in triplicates to calculate means and standard deviations (SD). To determine the significance values in all the experiments, Student’s t-test (unpaired) was performed, using GraphPad QuickCals (http://​www.​graphpad.​com/​quickcalcs/​index.​cfm). H-scores were calculated manually. Mann–Whitney U-test was used to evaluate the differences in H-score values of all the concerned proteins between human normal colon and colon carcinoma samples, using R module Wilcoxon-Mann-Whitney Test calculator (https://​www.​wessa.​net/​rwasp_​Reddy-Moores%20​Wilcoxon%20​Mann-Witney%20​Test.​wasp). Distribution of H-scores are represented by Box plots, generated using web tool BoxPlotR (http://​boxplot.​tyerslab.​com/​). Spearman’s rank correlation coefficient (r_s) was used to determine the correlation between the expression levels of two different proteins (individual H-scores) in both human normal colon and colon carcinoma samples, using R module Spearman Rank Correlation calculator (http://​www.​wessa.​net/​rwasp_​spearman.​wasp/​). Throughout, the value of P < 0.05 was considered to be statistically significant (*), P < 0.01, P < 0.001 and P < 0.0001 were considered to be very significant (**), highly significant (***) and extremely significant (****), respectively.

p68 is reported to be overexpressed in colon cancer samples, implicating it as a favorable prognostic marker for colon carcinoma [1, 2]. Increased RelA expression in colon carcinoma samples has been substantiated by several studies [15]. To investigate the possible connection of p68 with RelA, we conducted immunohistochemical (IHC) analysis in normal and colon carcinoma tissue samples procured from Indian patients. The expression of p68 and RelA exhibited marked elevation in colon carcinoma samples compared to that of normal. Expression of PCNA served as a proliferation marker (Fig. 1a). The difference in staining intensities (measured by H-score) of p68 and RelA between normal and carcinoma samples (represented by Notched Box Plot) were statistically significant, as predicted by the Mann–Whitney U-test analysis (Fig. 1b). Also, the difference between the average H-scores of p68 and RelA in colon carcinoma samples was statistically significant compared to the normal samples (Fig. 1c). Staining intensities of p68 and RelA (measured by H-score) followed similar trends in both normal and carcinoma samples (Fig. 1d). H-scores of p68 and RelA was found to bear strong positive correlation (rs = 0.955), when statistically analyzed by Spearman’s rank correlation test (Fig. 1e). These results indicate conceivable interconnection between p68 and RelA.

To investigate whether p68 regulates RelA gene expression, p68 was overexpressed in multiple colorectal cancer (CRC) cell lines. In all the cases, p68 enhanced RelA protein expression. p68 co-activates β-catenin in regulating its target gene Cyclin D1 [4]. Hence, Cyclin D1 served as positive control for p68 overexpression, which also displayed similar trends of expression like RelA (Fig. 1f). To further delve into the possible effect of p68 on RelA, we overexpressed p68 in a dose-dependent manner in HCT 116 cells. The results demonstrated dose-dependent increment in the expression of RelA, and also that of Cyclin D1 following a similar course (Fig. 1g). Next, in a converse approach, we used multiple short hairpin RNA (shRNA) against p68 to deplete its endogenous level. Significant reduction in RelA expression was observed on knocking down p68. Cyclin D1 followed a corresponding trend (Fig. 1h). Altogether, these results indicate highly plausible interconnection between p68 and RelA in both clinical samples and CRC cell lines and reinforces our hypothesis of p68 being involved in regulating RelA gene expression.

To study whether the effect of p68 on RelA expression at the protein level is a consequence of its transcriptional regulation, we quantified its mRNA expression. In multiple CRC cell lines, p68 overexpression led to increase in mRNA expression of RelA (Fig. 2a). Conversely, upon p68 knockdown there was a significant reduction in RelA mRNA expression (Fig. 2b). Cyclin D1 (CCND1) served as a positive control for both [4]. These results indicate p68’s role in regulating RelA at the transcript level.

We next speculated that RelA promoter activity might be regulated by p68, as p68 is a co-activator of several oncogenic transcription factors that might bind on the RelA promoter. We studied the promoter region of RelA 2000 bp upstream to the transcriptional start site and found that this region contains seven putative β-catenin/TCF-LEF binding elements (TBEs). Wnt/β-catenin signaling is highly implicated in colon cancer progression and p68 is reported to interact with β-catenin and thereby, regulate the expression of its target genes [4, 5]. Hence, we hypothesized that p68 might regulate RelA gene expression via TBEs. To explore whether p68 positively modulated RelA gene expression by directly upregulating its promoter activity, we cloned 1746 bp fragment of the RelA promoter (231 bp upstream to the transcription start site). Our luciferase assay results demonstrated that p68 overexpression, significantly upregulated RelA promoter activity in HCT 116 and SW-480 cells (Fig. 2c). Conversely, the knockdown of p68 using multiple shRNAs and siRNAs showed prominent decrease of the RelA promoter activity in both HCT 116 and SW-480 cells (Fig. 2d and Additional file 1: Figure S1A). Hence, p68 elevates RelAexpression by transcriptional activation of its promoter.

Next, we conducted Chromatin Immunoprecipitation (ChIP) assay to gain deeper mechanistic insights into the modulation of RelA promoter activity mediated by p68 and β-catenin cooperation. Immunoprecipitations (IPs) were carried out using anti-p68, anti- β-catenin and anti-TCF4 followed by amplification of different TBEs sites on the RelA promoter region, as shown in the schematic diagram (Fig. 2e). The ChIP data in HCT 116 and SW-480 cells demonstrated that p68, β-catenin and TCF4 bind to the promoter region of RelA in vivo (Fig. 2f and Additional file 1: Figure S1B). Amplification of Cyclin D1 promoter region spanning TBEs sites served as a positive control. Upon p68 knockdown, the occupancy of β-catenin is sharply reduced (TBE site1–3) (Fig. 2g). This further highlights that p68 is indeed essential for the occupancy of β-catenin on the RelA promoter, for regulating its expression.

The RelA promoter possesses seven TBEs namely: TBE1 (AACAAAG; − 252 to − 259), TBE2 (CTTTGCT; − 329 to − 336), TBE3 (CGCAAAG; − 568 to − 575), TBE4 (CTTTGGG; − 973 to − 980), TBE5 (ACCAAAG; − 1268 to − 1275), TBE6 (CTTTGTG; − 1722 to − 1729) and TBE7 (CTTTGGG; − 1886 to − 1893). The positions are relative to the transcription start site. We generated RelA promoter constructs containing deletions (5 nucleotides deleted by SDM) of the seven TBEs, as shown in the schematic diagram (Additional file 1: Figure S1C). Results of the luciferase assays showed that deletion of the TBEs markedly reduced the RelA promoter activity in both HCT 116 and SW-480 cells. Deletion of TBE4 (TBEΔ4) and TBE7 (TBEΔ7) has the most prominent effect in HCT 116 cells. In SW-480 cells, increase of RelA promoter activity in the presence of p68 was almost abrogated in case of TBEΔ4, TBEΔ5, TBEΔ6 and TBEΔ7. Thus, p68 mainly exerts its function through β-catenin/TCF4 complex at TBE4 and TBE7 and all these sites are crucial for β-catenin-dependent regulation of the RelA promoter activity (Fig. 2h). Hence, we can conclude that p68 acts in association with β-catenin in mediating the transcriptional activation of the RelA promoter.

Canonical Wnt-β-catenin driven carcinogenesis is characterized by increased accumulation of β-catenin and TCF/LEF complexes in the nucleus leading to inappropriate activation of Wnt target genes; thus making β-catenin an enticing molecular target for CRC therapy [5]. From our previous results, we have established that p68 cooperates with β-catenin in regulation of RelA expression. We now wanted to study whether β-catenin has any direct role in regulating RelA. Hence, to investigate the possible connection of β-catenin with RelA, we conducted immunohistochemical (IHC) analysis in normal and colon carcinoma tissue samples procured from Indian patients. We also checked the expression status of TCF4, since TCF4 is the transcription factor to which β-catenin binds, on the promoter region of its target genes. The expression of β-catenin, TCF4 and RelA exhibited marked elevation in CRC samples compared to normal (Fig. 3a). The difference in staining intensities (measured by H-score) of β-catenin and TCF4 between normal and carcinoma samples (represented by Notched Box Plot) were extremely statistically significant, as predicted by the Mann–Whitney U-test analysis (Additional file 2: Figure S2A). The difference between the average H-scores of β-catenin and TCF4 in colon carcinoma samples was highly statistically significant as to that of the normal samples (Additional file 2: Figure S2B). Staining intensities of β-catenin, TCF4 and RelA (measured by H-score) followed similar trends in both normal and carcinoma samples (Fig. 3b). H-scores of β-catenin and RelA was found to bear strong positive correlation (rs = 0.947), when statistically analyzed by Spearman’s rank correlation test (Fig. 3c). These results indicate strong possible interconnection between β-catenin/TCF4 and RelA expression.

To investigate whether β-catenin regulates RelA gene expression in CRC cell lines, β-catenin was overexpressed in HT-29 and HCT-15. In both the cases, β-catenin enhanced RelA protein expression. Cyclin D1, a target gene of β-catenin [25] served as positive control for β-catenin overexpression, which also displayed similar trends of expression like RelA (Fig. 3d). Next, in a converse approach, we used short hairpin RNA (shRNA) against β-catenin to deplete its endogenous level. Significant reduction in RelA expression was observed upon knocking down β-catenin. Cyclin D1 followed a corresponding trend (Fig. 3e). To further delve into the possible effect of β-catenin on RelA, we decreased the endogenous level of β-catenin in a dose-dependent manner in HCT 116 cells using shRNA against it. The results demonstrated dose-dependent reduction in the expression of RelA, and also that of Cyclin D1 following a similar course (Fig. 3f). To study whether the effect of β-catenin on RelA at the protein level is a consequence of its transcriptional regulation, we quantified its mRNA expression. In multiple CRC cell lines, β-catenin overexpression led to increase in mRNA expression level of RelA (Fig. 3g and Additional file 3: Figure S3A). Conversely, upon β-catenin knockdown there was a significant reduction in RelA mRNA expression (Fig. 3h and Additional file 3: Figure S3B). Cyclin D1 and c-Myc (also a target gene of β-catenin) served as a positive control for both [26].

The requirement of p68 and β-catenin amalgamation in driving RelA gene expression is further supported by negation of β-catenin and p68 mediated increase in the mRNA levels of RelA in p68 or β-catenin depleted condition respectively (Additional file 3: Figure S3C and D). We next speculated that RelA promoter activity might be affected by β-catenin, as we have already established the occupancy of β-catenin on the promoter of RelA by ChIP assay. Our luciferase assay results demonstrated that upon knockdown of β-catenin endogenous level using either shRNA or siRNAs, a prominent decrease in RelA promoter activity was observed (Fig. 3i and Additional file 3: Figure S3E). Conversely, overexpression of β-catenin upregulated RelA promoter activity in both HCT 116 and SW-480 cells (Additional file 3: Figure S3F). Wnt3a, an important member of the Wnt family is a conserved, secreted signaling glycoprotein that influences cell behavior by stabilization and nuclear translocation of β-catenin. Substantial studies report Wnt3a expression to be significantly correlated with colorectal carcinoma’s clinical staging, metastasis and recurrence [27]. In this report, we observed Wnt3a overexpression led to significant upregulation of RelA promoter activity in both HCT 116 and SW-480 cells (Fig. 3j). Thus, based on the above results we could conclude that β-catenin elevates RelA expression by transcriptional activation of its promoter. Next, on Wnt3a and β-catenin overexpressed condition, luciferase assay was conducted using the seven mutant RelA promoters (one TBE site deleted in each mutant). Results of the luciferase assay were similar to that obtained on p68 overexpressed condition, i.e. marked reduction in RelA promoter activity, in both HCT 116 and SW-480 cells. Deletion of TBE4 (TBEΔ4) and TBE7 (TBEΔ7) showed the most prominent effect. Thus all the TBE sites are crucial for p68-β-catenin-dependent regulation of the RelA promoter activity especially sites TBE4 and TBE7 (Fig. 3k and Additional file 3: Figure S3G). Altogether, these results reinforce our hypothesis of p68 coactivating β-catenin/TCF4 complex in mediating the transcriptional activation of the RelA promoter.

Here, we pursued to investigate further, the outcome of p68-mediated RelA upregulation. In unstimulated cells, NF-κB complexes are ordinarily sequestered in the cytoplasm with inhibitory proteins IκBs. Upon stimulation, degradation of IκB, allows nuclear translocation of NF-κB complexes to promote the transcription of genes linked to a variety of physiological processes and multiple imperative aspects of oncogenesis [6]. We wanted to study whether p68 mediated upregulation of RelA has any effect on NF-κB signaling axis. We checked the nucleo-cytoplasmic distribution of RelA, in p68 overexpressed HCT 116 and SW-480 cells. The results clearly indicated an increase in the nuclear localization of RelA, implicating that this might have a role in upregulation and activation of NF-κB signaling pathway (Fig. 4a). From the above result, we postulated that p68 might have possible impact on NF-κB signaling axis as a result of RelA regulation. NF-κB as a transcription factor contributes to the progression of CRC by regulating the expression of diverse set of genes which act as a pivotal crossroad between cell proliferation and cell death; the two facets of the oncogenesis-governing equation. Bcl-2 [8], Bcl-xl [9], Survivin [11] and XIAP [10], regulated by NF-κB, act as prototype molecule at this crossroad and holds promise as target genes in CRC therapy [28, 29]. Overexpression of these genes is correlated with majority of the salient features of tumor development and progression [12, 13, 30]. To investigate further whether p68 affects the expression of NF-κB target genes, we checked the expression of genes responsible for enhanced survival (Survivin) and evasion of apoptosis (Bcl-2, Bcl-xL and XIAP). In multiple CRC cell lines upon p68 overexpression; all the selected NF-κB target genes displayed enhanced protein expression (Fig. 4b). In a converse approach, on depleting the endogenous level of p68 using shRNA against it, protein expression of all the selected NF-κB target genes markedly reduced in HCT 116 and SW-480 cell lines (Fig. 4c). Next, we depleted the endogenous level of p68 using siRNAs against it, in a dose dependent manner in HCT 116 cells. As anticipated the protein expression of RelA as well as, all the selected NF-κB target genes reduced in a dose dependent fashion (Fig. 4d). As we have already established the involvement of β-catenin in RelA gene regulation we wanted to investigate whether β-catenin also plays a role in activating the NF-κB signaling pathway, by regulating the expression of NF-κB target genes. Upon overexpression of β-catenin in HT-29 and HCT-15 cell lines, all the selected NF-κB target genes displayed enhanced protein expression (Fig. 4e). Depletion of the endogenous level of β-catenin using shRNA against it, led to reduction in the protein expression of the selected NF-κB target genes, in both HCT 116 and HT-29 cells (Fig. 4f). Since TCF4 is the transcription factor to which β-catenin binds, the ultimate downstream effector of Wnt/β-catenin signaling, we next checked the status of the selected NF-κB target genes upon depleting the endogenous level of TCF4 by using siRNAs against it in a dose dependent manner in HCT 116 cell line, as anticipated the expression of RelA along with the selected genes decreased in a dose dependent fashion (Fig. 4g). From the above results we could conclude that ‘p68- β-catenin/TCF4 complex’ plays an active role in regulating NF-κB signaling axis.

To gain deeper insights into p68 and β-catenin mediated regulation of NF-κB target genes; we pursued to explore further the interplay of p68 with β-catenin in regulating RelA-dependent gene expression. We observed that β-catenin mediated increase in the protein levels of RelA and the selected NF-κB target genes were abrogated in p68-depleted condition (Fig. 4h). This highlights that p68 and β-catenin collaboration is required for RelA expression, followed by NF-κB signal activation and expression of signature NF-κB dependent genes. In all the above experiments, Cyclin D1 served as a positive control for both p68 and β-catenin/TCF4 signaling. Thus the above results corroborate that; p68 along with β-catenin/TCF4 exerts a strong influence on NF-κB signaling pathway.

After having established the involvement of p68 and β-catenin synergism in regulating NF-κB signaling pathway, we were inclined to investigate the clinical relevance of this signaling cascade. To inspect the expression status of the NF-κB target genes in human colon carcinoma tissue samples, we performed IHC analysis of Bcl-2, Bcl-xL, Survivin and XIAP in the same set of normal and colon carcinoma tissue samples, used in Figs. 1a and 3a. The abundance of Bcl-2, Bcl-xL, Survivin and XIAP were enhanced in colon carcinoma samples compared to normal (Fig. 5a). The difference in staining intensities as measured by H-score, of these proteins between normal and carcinoma samples were also statistically significant, as predicted by the Mann–Whitney U-test analysis, signifying the activation of these genes in the genesis stage of colorectal carcinoma (Fig. 5b). H-scores of p68 and RelA followed similar trends with H-scores of Bcl-2, Bcl-xL, Survivin and XIAP in both normal and carcinoma samples (Fig. 5c). H-scores of p68 were found to bear strong positive correlation with H-scores of Bcl-2 (rs = 0.924), Bcl-xL (rs = 0.936), Survivin (rs = 0.932) and XIAP (rs = 0.944) (Fig. 5d); likewise H-scores of RelA were found to bear strong positive correlation with H-scores of Bcl-2 (rs = 0.926), Bcl-xL (rs = 0.924), Survivin (rs = 0.919) and XIAP (rs = 0.931) when statistically analyzed by Spearman’s rank correlation test (Fig. 5e). These results indicate highly conceivable interconnection between p68 and NF-κB target genes-Bcl-2, Bcl-xL, Survivin and XIAP.

To widen our study horizon in terms of clinical relevance of this signaling cascade we further investigated the relationship between the expression status of β-catenin with p68 and the NF-κB target genes. Interestingly, H-scores of β-catenin were also found to follow similar trends with H-scores of p68 and NF-κB target genes-Bcl-2, Bcl-xL, Survivin and XIAP in both normal and carcinoma samples (Additional file 4: Figure S4A). H-scores of β-catenin bearing strong positive correlation with H-scores of p68 (rs = 0.958) and downstream NF-κB target genes-Bcl-2 (rs = 0.923), Bcl-xL (rs = 0.931), Survivin (rs = 0.962) and XIAP (rs = 0.937) (Additional file 4: Figure S4B). These results augment our in vitro findings of enhancement of NF-κB signaling axis by p68 and β-catenin collaboration.

We extended our investigation in mice colorectal tumor to assess the physiological relevance of our in vitro findings. Primary tumors were generated using mice syngeneic adenocarcinoma cell line CT26, stably expressing either enhanced green fluorescent protein (EGFP)-p68 (CT26^p68) or the empty vector encoded EGFP (CT26^EV). The murine CT26 is a grade IV carcinoma cell line developed in 1975 by exposing BALB/c mice to N-nitroso-N-methylurethane (NMU) and is extensively used as a model system in investigation of colorectal carcinomas. The expression of RelA and NF-κB target genes were significantly elevated in CT26^p68 cells compared with CT26^EV (Fig. 6a). The primary tumor volume produced after 30 days from CT26^p68 cells was markedly higher (approximately 15 times) with enhanced vasculature than that produced from CT26^EV cells (Fig. 6c). Hematoxylin and eosin (H&E) staining was conducted to examine the gross morphological features of the tissue sections derived from these tumors and it revealed, the tumor mass formed by CT26^p68 to be large and more compact with higher cellular density in comparison to CT26^EV (Fig. 6b and c). IHC analysis showed heightened expression of p68, RelA and NF-κB target genes- Bcl-2, Bcl-xL and Survivin in the tissue sections derived from CT26^p68 tumors. PCNA served as proliferation marker, also showed increased expression (Fig. 6d). Highly significant difference was exhibited in the H-scores of each of the proteins in CT26^p68 tumor-derived tissues in comparison to CT26^EV (Fig. 6e). Taken together, these animal data suggest that p68 enhances primary tumor growth, leads to upregulation of RelA and subsequent activation of oncogenic NF-κB signaling axis and corroborates the correlation between p68 and expression of NF-κB target genes.

To investigate the impact of p68 in development of lung metastasis, we injected stable CT26^p68 and CT26^EV cells intravenously into the tail vein of BALB/c mice. After 30 days of inoculation, it was observed that the metastatic lung nodules developed in mice injected with stable CT26^p68 were much higher in number (approximately three times) and elevated in size (approximately four times) and possessed compact tissue morphology with high cellular density as observed by the histological analysis (H&E staining) of the lung tissue sections in comparison to that formed by CT26^EV cells (Fig. 7a and b). IHC analysis showed elevated expression of p68, RelA, NF-κB target genes- Bcl-2, Bcl-xL and Survivin in the tissue sections of CT26^p68-derived lung metastatic nodules. Proliferation marker PCNA also displayed heightened expression in the CT26^p68-derived metastatic nodules. EMT is characterized with increase expression of EMT markers like Vimentin and Snail; we observed augmented expression of these proteins in the lung tumor tissue derived from the mice intravenously injected with CT26^p68 in comparison to CT26^EV cells (Fig. 7c). Each of the examined proteins in CT26^p68 tumor-derived tissues of lung metastatic nodules exhibited marked difference in H-score from CT26^EV tumor-derived tissues (Fig. 7d). Thus, p68-mediated upregulation of RelA and consequent activation of NF-κB signaling was followed by the elevated expression of NF-κB target genes henceforth accelerating the process of tumor formation and metastasis.