RING-finger protein 6 enhances c-Myc-mediated Warburg effect by promoting MAD1 degradation to facilitate pancreatic cancer metastasis

Abstract

Aerobic glycolysis (the Warburg effect) promotes tumor metastasis; hence, drugs targeting its regulators are being developed. c-Myc, a critical transcription factor that regulates the Warburg effect, is involved in the tumorigenesis of many cancers, including pancreatic cancer (PC). However, the upstream regulating mechanisms of c-Myc in PC are unclear. Herein, we reported that E3 ubiquitin ligase RING-finger protein 6 (RNF6) was upregulated in PC tissues, and an elevated RNF6 level was closely associated with metastasis and poor prognosis in patients with PC. In functional experiments, RNF6 over-expression accelerated the metastatic ability of PC cells, whereas RNF6 knockdown impaired PC cell motility and invasiveness along with metastasis in an orthotopic mouse model. Furthermore, we found that RNF6 promoted PC cell metastasis by enhancing c-Myc-mediated aerobic glycolysis. Mechanistically, RNF6 increased the expression level of c-Myc by catalyzing the ubiquitination of Max-dimerization protein-1 (MAD1), a cellular antagonist of c-Myc. Lastly, RNF6 promoted the degradation of MAD1 via the ubiquitin-proteasome pathway, and this reduction in the MAD1 levels enabled c-Myc to promote the Warburg effect in PC. Our results demonstrate that RNF6 may be a novel biomarker in PC carcinogenesis, thereby indicating that targeting the RNF6/MAD1/c-Myc axis is a potential strategy for PC therapy.

Introduction

Pancreatic cancer (PC) is one of the most fatal malignancies worldwide, and its incidence is increasing annually, especially in industrialized countries [1]. Despite recent improvements in PC diagnosis and therapy, the 5-year survival rate of patients with PC has remained dismal (as low as 6%) [2]. A high rate of metastasis is the leading causes of therapy failure in patients with PC according to an autopsy series, which reported that distant metastasis occurs in 90% of cases [3,4]. Therefore, the identification of novel approaches or targets is urgently required for improving treatment to prevent distal metastasis of PC.

Aerobic glycolysis, also known as the Warburg effect, is a shift from oxidative phosphorylation to glycolysis and is considered to be the root of the occurrence and development of tumors [5]. An enhanced Warburg effect is a fundamental property of many human cancers, including PC [6]. In particular, the metastatic ability of PC has been shown to be closely correlated with enhanced aerobic glycolysis [7]. Protooncogene c-Myc is an important transcription factor that regulates the process of glycolysis, which directly mediates the expression of genes involved in the regulation of glucose metabolism [8,9]. Increasing evidence shows that c-Myc plays a key role in the tumorigenesis of many tumors [10], including those of PC [11]. Although the deregulation of c-Myc contributes to PC progression and aerobic glycolysis, the upstream regulating mechanisms of c-Myc remains unclear.

It has been reported that oncogene proteins are regulated by the ubiquitin-proteasome system (UPS). As a substrate interaction module, E3 ubiquitin ligases, have been attracting increasing attention [12]. RING-finger protein 6 (RNF6), a RING-domain E3 ubiquitin Ligase, mediates the ubiquitination of its target proteins and tags them for proteasomal degradation [13,14]. A large number of studies have shown that RNF6 functions as an oncogene in multiple cancers such as hepatocellular carcinoma [15], esophageal squamous cell carcinoma [16] and prostate cancer [17]. Additionally, high RNF6 level is an indicator of poor prognosis for colorectal cancer [18,19]. Importantly, we observed elevated RNF6 mRNA in pancreatic cancer tissues though the Cancer Genome Atlas (TCGA) dataset, which was analysed using GEPIA software (http://gepia.cancer-pku.cn/) (Figure S1). These data suggest that RNF6 may play a role in pancreatic cancer tumorigenesis and development. However, the molecular function of RNF6, its target protein substrates, and its clinical significance in PC are unclear.

In this study, we aimed to determine the role of RNF6 in the progression of PC. We further explored the mechanisms of c-Myc overexpression in PC and found that RNF6 positively regulates c-Myc expression by destabilising Max dimerisation protein 1 (MAD1), a key antagonist of c-Myc [20]. Our work identified an interplay among these proteins that promotes PC metastasis by enhancing the Warburg effect.

Materials and methods

Clinical specimens

Matched cancerous and normal pancreatic tissue specimens were obtained from 109 patients with PC admitted to the Second Affiliated Hospital of Nanchang University from 2015 to 2019. Written informed consent was obtained from all patients, and the research procedure was approved by the Ethics and Research Committees of the Second Affiliated Hospital of Nanchang University. All specimens from resection surgery were frozen and stored at -80°C for further analysis. The clinical characteristics of all patients are summarised in Table 1.

Table 1.

Correlation between RNF6 and clinicopathologic characteristics of 109 PAAD patients

Variables	clinicopathological characteristics	numbers	RNF6 high expression	RNF6 low expression	p value
Age					0.764
	<60	45	28	17
	≥60	64	38	26
Gender					0.896
	Female	49	30	19
	Male	60	36	24
Tumor site					0.485
	Pancreatic head	64	37	27
	Pancreatic body and tail	45	29	16
Tumor size					0.19
	<3 cm	40	21	19
	≥3 cm	69	45	24
Histologic grade					0.002**
	High differentiation	51	23	28
	Middle and low differentiation	58	43	15
TNM stage					<0.001***
	l-ll	49	20	29
	lll-lV	60	46	14
Lymph node metastasis					0.032*
	Negative	65	34	31
	Positive	44	32	12
Vessel invasion					<0.001***
	NO	60	27	33
	YES	49	39	10
CEA					0.596
	Normal	25	14	11
	Abnormal	84	52	32

^* P<0.05;

^** P<0.01;

^*** P<0.001.

Over-expression constructs, shRNA plasmids, and cell transfection

Vectors encoding RNF6, c-Myc, or MAD1 and plasmids encoding short hairpin RNAs (shRNAs) against either RNF6 or MAD1 were synthesised by GenePharma (Shanghai, China). PC cells were transfected with these overexpression constructs or shRNA plasmids using Lipofectamine 3000 (Invitrogen, USA) according to the manufacturer’s instructions.

In vivo metastasis assay

The nude mice (male BALB/c-nu/nu, 6-8 weeks old) were purchased from the Animal Center of Nanjing Medical University. 5×10⁵ PC cells were stably transduced with firefly luciferase gene and injected into the tail vein of BALB/c nude mice. After 6-8 weeks, for in vivo signal detection, the mice were anesthetized with isofluorane and then imaged in a Lumina Series III IVIS instrument (PerkinElmer, USA). Then, the mice were euthanized, and their lungs were removed and stained with haematoxylin and eosin stain (H&E) for pathological examination. The animal work was approved by the Ethics Committee for Animal Experiments of the Second Affiliated Hospital of Nanchang University.

Extracellular acidification rate (ECAR) and oxygen consumption rate (OCR)

Metabolic profiles were tested using the Glycolysis Stress Test and Mito Stress Test Kits using the Seahorse XFe96 Analyser according to the manufacturer’s instructions.

Mass spectrometry

PC cell lysates were collected and immunoprecipitated with RNF6 antibodies or IgG. Next, the bound proteins were resolved via sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), visualised using Coomassie blue staining, and then subjected to mass spectrometric analysis. Proteins were identified by comparing their sequences with those on the human RefSeq protein database.

Co-immunoprecipitation (Co-IP) and in vivo ubiquitination assay

Co-IP assays were performed as described previously [21]. For the in vivo ubiquitination assay, RNF6-overexpressing or RNF6-silenced PC cells were treated with MG132 for 4 h before being harvested. Then, the cell lysate was immunoprecipitated with anti-MAD1 antibodies, and the ubiquitination level of MAD1 was tested with an anti-Ub antibody. In addition, to detect MAD1 ubiquitination in HEK293 cells, the cells were transfected with Flag-MAD1, His-RNF6, and HA-Ub constructs and further incubated with MG132 (15 μM). To assess the MAD1 ubiquitination type, MAD1 mutants (mutations of all lysine residues) and two Ub mutants (K48R and K63R) were purchased from GenePharma.

Statistical analysis

All data were expressed as mean ± standard errors of means using GraphPad Prism 6 (GraphPad Software, USA). Significant differences were analysed using the Student’s t-test and two-tailed distributions. The Kaplan-Meier method was used to calculate the survival curve, and log-rank test was used to determine the significance. P<0.05 was considered significant. Other materials and methods are provided in the Supplementary Data.

Results

High expression of RNF6 is closely associated with metastasis and poor prognosis of patients with PC

To investigate the role of RNF6 in PC, we detected the RNF6 expression level in PC tissues and compared it with that in the adjacent normal tissues. As shown in Figure 1A and 1B, aberrant RNF6 expression occurred in PC tissues, whereas a weak positive signal was detected in corresponding normal tissues. In keeping with the IHC results, the western blot data showed that RNF6 protein expression was remarkably higher in PC tissues than in normal pancreatic tissues (Figure 1C and 1D). These results revealed that RNF6 is highly expressed in PC, suggesting that it may be involved in the progression of PC.

Figure 1.

High RNF6 level correlates with poor prognosis in PC patients. (A and B) Representative image (A) and quantification (B) of IHC staining of RNF6 in PC tissues and paired normal tissues. The image was captured at 400× magnification. Scale bar, 50 μm. **P<0.01. (C and D) Determination and quantification of RNF6 protein levels in PC tissues and paired normal tissues by western blotting assay. Tubulin was used as a loading control. **P<0.01. (E and F) Kaplan-Meier plots representing probabilities of progression-free and overall survival in 109 PC patients according to expression level of RNF6. Statistical analysis was conducted using Student’s t-test and Log Rank test.

To determine the pathologic significance of RNF6 expression in PC progression, we analysed the correlation between RNF6 expression and established PC prognostic factors (Table 1). No significant association was observed between RNF6 expression with age, sex, or tumour size in PC patients. However, RNF6 overexpression was significantly correlated with histologic grade (P=0.002), TNM stage (P<0.001), lymph node metastasis (P<0.001) and vessel invasion (P=0.032). As shown in Figure 1E and 1F, shorter overall survival (OS) and disease-free survival (DFS) in the high level of RNF6 expression group than those in the low level of RNF6 expression group were observed in our cohort of PC patients. Additionally, the results of multivariate Cox regression analysis indicate that RNF6 overexpression predicted shorter OS and RFS in PC patients (Table 2), supporting that high RNF6 level is an independent risk factor for PC progression.

Table 2.

Univariate and multivariate analyses of overall survival in PC patients

Parameters	Univariate analysis			Multivariate analysis
	HR	95% CI	P value	HR	95% CI	P value
Age (≥60 vs <60)	1.374	0.628-3.165	0.764	-	-	-
Sex (Female vs Male)	1.773	0.509-2.845	0.896	-	-	-
Tumor site (head vs body and tail)	1.668	0.737-2.665	0.485	-	-	-
Tumor size (<3 vs ≥3)	1.672	0.643-4.215	0.190	-	-	-
CEA (Normal vs Abnormal)	1.212	0.643-1.976	0.596	-	-	-
Histologic grade (High vs Middle/Low)	1.329	1.759-4.324	0.002*	1.553	1.248-2.375	0.031*
TNM stage (I-II vs III-IV)	2.908	1.835-5.522	<0.001*	1.834	1.459-5.332	0.008*
Lymph node metastasis (No vs Yes)	1.569	1.134-3.452	0.032*	1.764	1.365-3.897	0.013*
Vessel invasion (No vs Yes)	1.213	1.132-2.243	<0.001*	1.224	0.976-2.214	0.067
RNF6 expression (High vs Low)	3.513	2.216-6.432	0.003*	2.981	1.611-5.332	0.009*

HR, hazard ratio; CI, confidence interval;

^*P<0.05.

RNF6 promotes PC cell migration and invasion in vitro, and metastasis in vivo

To evaluate the impact of RNF6 expression on PC biology, we transfected two PC cell lines exhibiting high endogenous RNF6 expression (AsPC-1 and BxPC-3) with shRNF6 plasmids to knock down their RNF6 expression (Figure 2A-D). Migration and invasion assays indicated that RNF6 silencing significantly suppressed the mobility and invasiveness of PC cells (Figure 2E and 2F). Consistently, the data of RTCA assays also shown that RNF6 knockdown obviously inhibited the metastasis ability of AsPC-1 and BxPC-3 cells (Figure 2G and 2H). Conversely, we infected PANC-1 and CaPan-1 cells, which have intermediate levels of endogenous RNF6 expression, with vectors containing RNF6 (Figure S2A and S2B). We found that RNF6 upregulation promoted PANC-1 and CaPan-1 cell migration and invasion (Figure S2C and S2D). RTCA assays demonstrated that RNF6 overexpression enhanced the migratory activity of both PC cell lines, more so than that of control cells (Figure S2E and S2F).

Figure 2.

RNF6 promotes migration and invasion of PC cell in vitro and accelerates the metastasis of PC cell in vivo. A and B. The protein and mRNA levels of RNF6 were detected in four PC cells and the immortalised H6c7 line. Tubulin was used as a loading control. C and D. Western blot analyses were used to detect the expression level of RNF6 in AsPC-1 and BxPC-3 cells stably transfected with the RNF6-silenced vector. Tubulin was used as a loading control. E and F. Transwell migration and Transwell invasion assays of AsPC-1 and BxPC-3 cells transfected with RNF6 knockdown vector. The image was captured at 400× magnification. Scale bar, 50 μm. *P<0.05, **P<0.01. G and H. RTCA assays were performed to detect the metastasis ability of AsPC-1 and BxPC-3 cells transfected with RNF6-knockdown vector. I. BxPC-3/shRNF6 cells were injected into the tail vein of nude mice, and the incidence of liver metastasis were measured after 6-8 weeks. n=6, **P<0.01. J. Representative image (left; magnification: ×100, inset magnification: ×400) and quantification (right) of H&E staining of liver metastatic nodules. n=6/group. Scale bar, 50 μm. **P<0.01.

Next, we determined the effect of altered RNF6 expression on PC metastasis in vivo. As shown in Figure 2I, the in vivo tumor metastatic assay revealed that knockdown of RNF6 rescued the decreased incidence of liver metastasis. Additionally, H&E-stained serial liver sections revealed that RNF6 silencing markedly suppressed experimental liver metastasis (Figure 2J). However, overexpression of RNF6 increased liver metastasis (Figure S2G and S2H). These data clearly demonstrate that RNF6 promotes the invasion and metastasis of PC cells in vitro and in vivo, indicating that RNF6 might function as an oncogene for PC.

The Warburg effect is crucial for the oncogenic functions of RNF6

To determine the mechanism by which RNF6 influences the invasion and metastasis of PC, we applied RNA-Seq to find global changes in the transcriptome when RNF6 was knocked down in AsPC-1 cells. KEGG analysis revealed that the top downregulated gene set in RNF6-knockdown cells was associated with glycolysis (Figure 3A). Therefore, we assessed whether RNF6 could modulate the glycolytic phenotype in PC cells. As expected, RNF6 silencing reduced glucose uptake, lactate production and ATP generation in AsPC-1 and BxPC-3 cells (Figures 3B-D, S3A-C). Further, RNF6 knockdown also suppressed the extracellular acidification rate (ECAR), which indicated overall glycolytic flux (Figures 3E, S3D). We also examined OCR (cellular oxygen consumption rate), an indicator of mitochondrial respiration. RNF6-silenced PC cells displayed an increase in OCR (Figures 3F, S3E). In contrast, RNF6 overexpression significantly enhanced the glycolysis activity of PANC-1 and CaPan-1 cells (Figures 3G-K, S3F-J).

Figure 3.

RNF6 promotes migration and invasion of PC cells by modulating Warburg effect. A. KEGG analysis revealed that the top down-regulated gene set in RNF6-silenced cells was associated with glycolysis. B-D. Glucose consumption, lactate production, and ATP levels in AsPC-1/shRNF6 cells. Three independent experiments were performed. **P<0.01. E. ECAR data showing the glycolytic rate and capacity in RNF6-silenced PC cells. Glucose (10 mM), the oxidative phosphorylation inhibitor oligomycin (1.0 μM), and the glycolytic inhibitor 2-deoxyglucose (2-DG, 50 mM) were sequentially injected into each well at the indicated time points. All measurements were normalised to the cell number calculated using a crystal violet assay at the end of the experiment. **P<0.01. F. OCR results showing the basal respiration and maximum respiration in RNF6-silenced PC cells. Oligomycin (1.0 μM), the mitochondrial uncoupler carbonyl cyanide p-trifluoromethoxy phenylhydrazone (FCCP, 1.0 μM), and the mitochondrial complex I inhibitor rotenone plus the mitochondrial complex III inhibitor antimycin A (Rote/AA, 0.5 μM) were sequentially injected. All measurements were normalised to the cell number calculated using crystal violet assay at the end of the experiment. *P<0.05. G-I. Glucose consumption, lactate production, and ATP levels in PANC-1/His-RNF6 cells. *P<0.05, **P<0.01. J. ECAR data showing the glycolytic rate and capacity in RNF6-overexpressing PC cells. *P<0.05. K. OCR results showing the basal respiration and maximum respiration in RNF6-overexpressing PC cells. *P<0.05. L. Effects of 2-DG on the migration and invasion of PANC-1/His-RNF6 cells. The image was captured at 400× magnification. Scale bar, 50 μm. **P<0.01.

To determine whether the oncogenic role of RNF6 is dependent on the Warburg effect in PC, the glycolytic inhibitor 2-deoxy-D-glucose (2-DG) was used to abolish glycolysis in PANC-1 cells transfected with vectors containing RNF6. As show in Figure 3L, RNF6 overexpression increased the migration and invasion of PANC-1 cell, which was abolished by 2-DG. These findings suggest that RNF6 exerts its oncogenic role by enhancing the Warburg effect in PC.

RNF6 promotes the warburg effect by regulating c-Myc expression in PC

Next, we performed gene set enrichment analysis (GSEA) using the TCGA database to determine the possible associations between RNF6 and c-Myc. As shown in Figure 4A, the gene sets of Hallmark_C-MYC_Targets_V1 were obviously enriched in PC samples with high RNF6 level, indicating that c-Myc was closely associated with high RNF6 level in PC. Thus, we speculated that the Warburg effect and metastasis in the RNF6-mediated PC cells might have resulted from elevated levels of c-Myc. As expected, the qRT-PCR data indicated that RNF6 overexpression significantly upregulated c-Myc mRNA expression, whereas RNF6 knockdown suppressed the mRNA level of c-Myc in PC cells (Figure 4B). We also found that the protein levels of c-Myc increased in RNF6-overexpression cells, whereas they decreased in RNF6-silenced PC cells (Figure 4C). We then investigated the expression of c-Myc target glucose metabolism genes, including glucose transporter GLUT1, hexokinase 2 (HK2), phosphofructokinase (PFKM) and lactate dehydrogenase A (LDHA) in PC cells with RNF6 alteration. As shown in Figure 4D and 4E, the mRNA and protein levels of these genes decreased in RNF6-silenced PC cells, whereas the expression of these genes increased in RNF6-overexpression PC cells.

Figure 4.

RNF6 promotes glycolysis by regulating c-Myc transcription. A. GSEA comparing the gene sets of Myc targets in RNF6^high and RNF6^low PC patients. Data were obtained from the TCGA database. NES, normalised enrichment score. B. The mRNA levels of c-Myc in RNF6-overexpressing or RNF6-silenced PC cells were detected by qRT-PCR. *P<0.05, **P<0.01. C. Protein levels of c-Myc in RNF6-overexpressing or RNF6-silenced PC cells were detected by western blotting. Tubulin was used as a loading control. D. mRNA levels of c-Myc target glucose metabolism genes (LDHA, GLUT1, HK2 and PFKM) in RNF6-overexpressing or RNF6-silenced PC cells were detected by qRT-PCR. *P<0.05, **P<0.01. E. Protein levels of LDHA, GLUT1, HK2 and PFKM in RNF6-overexpressing or RNF6-silenced PC cells were detected by western blotting. Tubulin was used as a loading control. F. qRT-PCR analysis of c-Myc mRNA expression in PC tumours and paired normal tissues. **P<0.01. G and H. Determination and quantification of c-Myc protein levels in PC tissues and paired normal tissues by western blotting assay. Tubulin was used as a loading control. I. Scatter plots show a positive correlation between RNF6 and c-Myc at the protein level in PC. J. Representative IHC staining of RNF6 and c-Myc in PC tissues and paired normal tissues. The image was captured at 400× magnification. Scale bar, 50 μm.

To further determine the correlation between RNF6 and c-Myc, we detected the expression level of c-Myc in PC tissues. As shown in Figure 4F-H, the expression level of c-Myc was obviously higher in PC tissues than in normal pancreatic tissues. We also detected the expression level of c-Myc in 36 PC specimens with high RNF6 expression, using an IHC assay. And, the statistical analysis results indicated that the protein level of c-Myc was positively associated with the protein level of RNF6 in PC tissues (Figure 4I). Moreover, as shown in Figure 4J, c-Myc overexpression, based on expression in normal pancreatic tissues, occurred in 28 cases (77.8%). Overall, these findings suggest that c-Myc may mediate the regulatory function of RNF6 in PC cells.

The oncogenic effect of RNF6 in PC is dependent on enhancing c-Myc expression

To investigate whether c-Myc enhancement was responsible for the oncogenic functions of RNF6 in PC, we transfected c-Myc overexpression plasmid into RNF6-knockdwon PC cells and examined its effects on these cell biological functions. As shown in Figure 5A-D, c-Myc restoration abolished the decrease in glucose uptake, lactate production and ATP generation in RNF6-knockdown PC cells. Additionally, the restored expression of c-Myc also reversed ECAR and OCR loss in RNF6-silenced cells (Figure 5E and 5F). Furthermore, the reduction in LDHA, GLUT1, HK2, and PFKM expression in RNF6-silenced cells was also reversed by overexpression of c-Myc (Figure 5G). Moreover, the inhibitory effects of RNF6 silencing on migration and invasion were also reversed by c-Myc restoration (Figure 5H and 5I). These findings suggest that RNF6 exerts its oncogenic functions by upregulating c-Myc in PC.

Figure 5.

Oncogenic effect of RNF6 is dependent on c-Myc enhancement. A. After overexpression of c-Myc in RNF6 knockdown AsPC-1 cells, the protein levels of RNF6 and c-Myc were detected. Tubulin was used as a loading control. B-D. After overexpression of c-Myc in RNF6-silenced AsPC-1 cells, glucose uptake, the production of lactate and ATP were measured. E and F. After overexpression of c-Myc in RNF6-knockdown AsPC-1 cells, the ECAR and OCR were measured. G. After overexpression of c-Myc in RNF6-silenced AsPC-1 cells, protein levels of LDHA, GLUT1, HK2 and PFKM were detected. Tubulin was used as a loading control. H and I. After overexpression of c-Myc in RNF6-knockdown AsPC-1 cells, cell migration and invasive capacity were measured using the transwell assay. *P<0.05, **P<0.01.

RNF6 induces c-Myc expression by downregulating MAD1 expression

To determine how RNF6 modulates c-Myc expression, mass spectrometric analysis was used to identify RNF6-interacting proteins in PC cells. The above results indicate that RNF6 positively regulates the expression of c-Myc in PC cells. As an E3 ubiquitin ligase, RNF6 is responsible for protein degradation and recycling, and we next hypothesized that RNF6 may upregulate c-Myc expression level via degrading the expression of the negative regulator of c-Myc. 35 proteins were identified by mass spectrometry, we found that only the MAD1 (an important cellular antagonist of c-Myc) may interact with RNF6 (Figure 6A). Additionally, MAD1 has been linked to the progression of diverse tumors [22,23]. Further, the data of Co-IP assays confirmed the interaction between RNF6 and MAD1 in AsPC-1 and PANC-1 cells (Figure 6B). Moreover, the GST pull-down assay indicated that RNF6 could bind to MAD1 under cell-free conditions (Figure 6C). These results suggested that RNF6 directly interacted with MAD1 in PC cells.

Figure 6.

RNF6 upregulates c-Myc by downregulating MAD1 expression. A. A partial list of RNF6-associated proteins were indicated by immunoprecipitation-mass spectrometry. B. Representative western blot showing that RNF6 interacts with MAD1 in PC cells. PC cells were lysed, and an immunoprecipitation assay was performed with anti-MAD1, anti-RNF6, or anti-IgG antibodies, followed by western blot with the indicated antibodies. C. GST pull-down assay performed to detect the direct interaction between RNF6 and MAD1 in HEK293 cells. D and E. Protein levels of RNF6 and MAD1 in RNF6-overexpressing or RNF6-silenced PC cells were detected by western blotting. Tubulin was used as a loading control. F and G. After knockdown of MAD1 in RNF6-silenced PC cells, the protein levels of RNF6, c-Myc and MAD1 were detected. Tubulin was used as a loading control. H and I. After restoration of MAD1 in RNF6-overexpression PC cells, the protein levels of RNF6, c-Myc and MAD1 were detected. Tubulin was used as a loading control.

Next, we assessed whether RNF6 could regulate the expression level of MAD1 in PC cells. As shown in Figure 6D, RNF6 knockdown obviously upregulated the protein level of MAD1 in AsPC-1 and BxPC-3 cells, whereas RNF6 overexpression significantly reduced the protein level of MAD1 in PANC-1 and CaPan-1 cells (Figure 6E). However, the mRNA level of MAD1 was not affected by RNF6 alteration in PC cells (Figure S4), indicating that RNF6 regulates the expression of MAD1 via the post-translational modification.

Finally, our findings revealed that RNF6 silencing did not further suppress c-Myc expression in MAD1-silenced PC cells (Figure 6F and 6G). Conversely, ectopic expression of MAD1 obviously dampened c-Myc induction in RNF6-overexpression PC cells (Figure 6H and 6I). Collectively, these data indicate that MAD1 may be required for the regulatory role of RNF6 on c-Myc expression.

RNF6 mediates ubiquitination and degradation of MAD1

Prior studies have demonstrated that MAD1 can be polyubiquitinated and degraded by the UPS [24]. As an E3 ubiquitin ligase, RNF6 is responsible for protein degradation. Hence, we speculated that RNF6 mediated the reduction in MAD1 levels by modulating the ubiquitination level of MAD1. To test this notion, we first examined the degradation of MAD1 protein following treatment with cycloheximide (CHX). As shown in Figure 7A and 7B, RNF6 silencing significantly inhibited the degeneration of MAD1 protein in PC cells. Secondly, we found that the protein level of MAD1 was restored in RNF6-overexpression cells following treatment with the proteasome inhibitor MG132 (Figure 7C). Thirdly, RNF6 overexpression increased the ubiquitination level of MAD1, whereas RNF6 knockdown reduced the poly-ubiquitination level of MAD1 in PC cells (Figure 7D). Finally, all Lys mutation in MAD1 could abolish RNF6-mediated MAD1 poly-ubiquitination in PC cells (Figure 7E). Furthermore, K48R mutation on Ub almost completely eliminated RNF6-mediated MAD1 poly-ubiquitination, whereas K63R mutation on Ub had no effect (Figure 7F). Overall, these findings suggest that RNF6 serves as an E3 ubiquitin ligase responsible for MAD1 degradation via the ubiquitin-proteasome pathway in PC.

Figure 7.

RNF6 destabilises MAD1 by promoting MAD1 ubiquitination and degradation in PC cells. A and B. Representative western blot showing that RNF6 knockdown promotes degeneration of MAD1 protein in PC cells. PC cells were treated with cycloheximide (CHX, 100 mg/ml) for indicated time points and subjected to western blot analysis (left); the quantitative results of MAD1 protein level determination are shown (right). C. PC cells transduced with His-RNF6 were treated with 10 µM MG132. Cells were collected at 6 h and immunoblotted with the antibodies indicated. D. Knockdown or exogenous expression of RNF6 altered the ubiquitination of MAD1. The cells in each group were treated with proteasomal inhibitor MG132. Cell lysates were prepared and subjected to immunoprecipitation with anti-MAD1 antibodies. The level of ubiquitin-attached MAD1 was detected by western blotting with anti-Ub antibodies. E. Ubiquitination of wild-type MAD1 or the K-to-R mutant (mutations in all Lys site of MAD1 gene) in HEK293 cells. F. Determination of MAD1 ubiquitination type in HEK293 cells.

Discussion

RNF6, a main member of the RING finger E3 ligase family, has been shown to play critical roles in tumorigenesis in several types of cancers [15,18,25]. However, there is currently no information on the role and mechanism of RNF6 in PC. In this study, we found that RNF6 was highly expressed in PC tissues, and an increased in RNF6 expression was closely associated with malignant phenotype of patients with PC. In addition, we examined the role of RNF6 in PC tumorigenesis using cell culture approaches and animal-based tumour models. We found that RNF6 promotes PC cell migration and invasion in vitro, and metastasis in vivo, thereby implying that RNF6 may serve as an oncogene in PC. We believe, to the best of our knowledge, that this is the first study to explore the role of RNF6 in PC.

The Warburg effect, a hallmark of cancer cells, is a crucial process during cancer metastasis. Therefore, understanding the potential regulatory mechanisms of the Warburg effect in PC will be beneficial for the prevention and treatment of this malignant cancer. In the process of the Warburg effect, c-Myc plays a critical role through directly regulating the expression of glycolytic enzyme-related genes [9,26]. Our GSEA data suggested that the c-Myc target genes were closely associated with high RNF6 level in PC tissues. Hence, we attempted to investigate whether the pro-tumour role of RNF6 was dependent on the c-Myc mediated Warburg effect. Indeed, RNF6 silencing significantly suppressed the expression of c-Myc as well as the metabolic phenotype of aerobic glycolysis. In contrast, RNF6 overexpression dramatically enhanced the glycolytic activity of PC cells. Importantly, ectopic expression of c-Myc could abolish the reduced glycolysis and malignant phenotypes in RNF6-silenced PC cells, indicating that RNF6 exerted its oncogenic function via the c-Myc-mediated Warburg effect.

It is well known that the regulation mechanism of c-Myc is very complex. Ectopic expression of c-Myc in many cancers is caused by genetic amplifications, insertions, and translocations [27,28]. Various factors such as TIP30 [29], microRNA 145 [30], Fbxw7 [31] and SKP2 [32] can regulate c-Myc expression via transcriptional, post-transcriptional mechanisms and post-translational modifications. In the present study, we clarified the mechanisms by which RNF6 mediates the overexpression of c-Myc in PC. RNF6 may upregulate the expression of c-Myc by ubiquitylating and degrading MAD1, but not other regulators of c-Myc. Our data indicate that RNF6 alteration could regulate the expression level of c-Myc and its transcriptional repressor MAD1. Importantly, Mass spectrometry and Co-IP assay showed that RNF6 did not interact with c-Myc (Figure S5) but interacted with the MAD1 in PC cells. These results suggest that RNF6 indirectly upregulates c-Myc expression by disrupting MAD1 expression.

Studies have shown that MAD1 is a key negative regulator of c-Myc [33]. Moreover, the UPS-mediated degradation of MAD1 is an important mechanism for regulating the expression level of MAD1 in many cancers [34,35]. Consistent with these results, we found that RNF6 is involved in the MAD1 degradation process and serves as an E3 ubiquitin ligase for MAD1 in PC cells. This notion is supported by three lines of experimental evidence. First, RNF6 can decrease the half-life of MAD1. Second, treatment with MG132, a proteasome inhibitor, abolishes RNF6-induced MAD1 degradation in PC cells. Third, RNF6 silencing decreases the ubiquitination levels of MAD1, whereas RNF6 overexpression increase them.

In conclusion, we found that RNF6 interacted with MAD1 and increased its poly-ubiquitination level, thereby destabilising MAD1 and causing the activation of the c-Myc-mediated Warburg effect in PC (Figure 8). Additionally, a high RNF6 level was significantly associated with a poor outcome for patients with PC. Mechanistically, our results demonstrate that RNF6 contributes to PC metastasis by enhancing the MAD1/c-Myc axis-mediated Warburg effect, thus suggesting that targeting RNF6/MAD1/c-Myc axis is a potential strategy for PC therapy.

Figure 8.

Proposed model by which E3 ubiquitin ligase RNF6 promotes pancreatic cancer metastasis by enhancing c-Myc-mediated Warburg effect via destabilization of MAD1.

Disclosure of conflict of interest

None.