FUS-induced circRHOBTB3 facilitates cell proliferation via miR-600/NACC1 mediated autophagy response in pancreatic ductal adenocarcinoma

Tumor tissues and adjacent normal pancreas tissues were collected from PDAC patients who received pancreaticoduodenectomy at Pancreas Center, the First Affiliated Hospital of Nanjing Medical University, from Jul. 2014 to Dec. 2018. The 110 patients were followed regularly until 11th September 2020. All patients didn’t receive any chemotherapy or radiotherapy before surgery. All patients signed an informed consent that was supervised by the Hospital Ethics Committee before specimen collection. None of the 110 selected patients died within 1 month after surgery. The samples were excised from patients, immediately frozen in liquid nitrogen and stored until use. All the cancer and adjacent tissues were diagnosed by two pathologists independently. TNM stage classification complied with the TNM classification system of the International Union Against Cancer (8th edition). We used Kaplan Meier method to draw the overall survival curve according to the relative expression of circRHOBTB3 (or miR-600 or NACC1) and the cut-off value (Median of the expression) for circRHOBTB3 (or miR-600 or NACC1).

Human PDAC cell lines (BxPC-3, MiaPaca-2, CFPAC-1 and PANC-1) and human pancreatic ductal epithelial (HPNE) cells were purchased from the Cell Bank of Type Culture Collection of the Chinese Academy of Sciences in Shanghai, China. The cells were cultured in a humidified atmosphere at 37 °C with 5% CO2. All cells were cultured in Dulbecco’s modified Eagle’s medium (Life Technologies) with 10% fetal bovine serum (Wisent, Montreal, QC, Canada), 10 mM HEPES (Sigma, St Louis, MO), 2 mM L-glutamine (Sigma), 1 mM sodium pyruvate (Sigma), 100 U/ml of penicillin (Life Technologies) and 100 μg/ml streptomycin (Life Technologies).

Plasmids were transfected in to PDAC cells using Lipofectamine 3000 and P3000 (Invitrogen) according to the manufacturer’s instructions. SiRNAs, miRNA mimics or inhibitors were transfected into cells using Lipofectamine 3000 (Invitrogen, USA). To construct the circRHOBTB3 stably knocking down cell lines, we transfected the control vector or the circRHOBTB3 knocking down vector pLV3ltr-GFP-Puro-U6-siRNA (circRHOBTB3) into PANC-1 and MiaPaca-2 cells and then selected them with puromycin (Sigma, USA) for 2–3 weeks until circRHOBTB3 was stably knocked down in the cells.

The circRHOBTB3 overexpression vector was constructed by Obio Technology Corp., Ltd. (Shanghai, China) by inserting the sequence of human circRHOBTB3 cDNA into the pGL3-circ expression vector, with the empty plasmid used as a control. siRNA targeting circRHOBTB3 and the control siRNA were synthesized by RiboBio (Guangzhou, China).

The miR-600 mimics and inhibitor, NACC1-overexpressing vector and NACC1 siRNAs were synthesized by GenePharma (Shanghai, China). The pRL-SV40-circRHOBTB3 and NACC1 luciferase reporter was constructed by inserting circRHOBTB3 fragments and 3’UTR fragment of NACC1 downstream of the luciferase reporter gene in the reporter plasmid (GenePharma, Shanghai, China). The miR-600 complementary sequence “CTGTAAG” in circRHOBTB3 and the 3′UTRs of NACC1 were mutated to remove the complementarity. All of the constructs were verified by sequencing and the sequences are listed in Table S1.

Total RNA was isolated from clinical specimens or cell lines with TRIzol Reagent (Life Technologies, Carlsbad, CA, USA) according to the manufacturer’s instructions. After spectrophotometric quantification, RNA (500 ng) was reverse transcribed into cDNA following the protocol of the iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA, USA).

The circRNA and mRNA levels were normalized to that of 18S rRNA or GAPDH in tissue or cell lines, respectively. The miRNA was normalized to that of U6. Target gene expression was calculated using the 2^-ΔΔCT method. Each quantitative PCR assay was performed in triplicate and independently repeated three times. The sequences of the primers used in the present study are listed in Table S1.

For RNase R treatment, 2 μg of total RNA was incubated for 10 min at 37 °C with or without 3 U/μg RNase R (Epicentre Technologies, Madison, WI, USA). PDAC cells were treated with 5 μg/ml actinomycin D (Sigma-Aldrich, USA) and collected in a series of time intervals. The expression of circRHOBTB3 and the linear mRNA was detected by qRT-PCR.

Cytoplasmic and nuclear fractions were preparing using the reagents in a PARIS™ kit (AM1556, Thermo Fisher Scientific, Waltham, USA) according to the manufacturer’s protocol. Briefly, PDAC cells were lysed in Cell Fraction Buffer on ice for 10 min. After centrifugation at 500 g for 3 min at 4 °C, the supernatant was collected as the cytoplasmic fraction. Then, the pelleted nuclei were washed with Cell Fraction Buffer and used as the nuclear fraction.

Cy3-labelled circRHOBTB3 probes and fluorescein amidite (FAM)-labelled miR-600 probes were designed and synthesized by RiboBio. The sequences of the probes are listed in Table S1. A fluorescence in situ hybridization (FISH) kit (RiboBio) was used to detect the probe signals in PDAC cells and tissues according to the manufacturer’s instructions. Nuclei were stained with 4,6-diamidino-2-phenylindole (DAPI). All images were acquired with an LSM880 NLO (2 + 1 with BIG) confocal microscope system (Carl Zeiss).

We used a CCK-8 assay kit (Dojindo, Japan), clone formation assay and 5-Ethynyl-20- deoxyuridine (EdU) assay (Beyotime) to assess cell proliferation. The CCK-8 assay was performed by seeding the treated cells in 96-well plates at 1.0 × 10³ cells/well. Subsequently, each well was incubated with 100ul 10% CCK-8 solution for 2 h at 37 °C away from light and the absorbance of each well at 450 nm was measured with a microplate reader. For clone formation assay, cells were seeded in six-well plates (800 cells/well) and cultured in complete medium supplemented with 10% fetal bovine serum for 2 weeks. Then, the clusters were stained with 0.1% crystal violet (Beyotime) counted if their diameter was greater than 1 mm. The EdU assay was performed according to the manufacturer’s protocol. Briefly, PDAC cells were plated in 96-well plates and incubated with a 50 mM EdU solution for 2 h and then fixed in 4% paraformaldehyde. The cells were then permeabilized with 0.3% Triton for 10 min and then sequentially stained with Alexa Fluor 555 azide and Hoechst 33342. Subsequently, the EdU-treated cells were imaged and counted under an Olympus FSX100 microscope (Olympus, Tokyo, Japan).

A total of 1.0 × 10⁷ cells were harvested and lysed, and C-1 magnetic beads were incubated with the circRHOBTB3 probes or miR-600 probes (Life Technologies) at 25 °C for 2 h to generate probe-coated beads. The cell lysates were then incubated with the coated beads at 4 °C overnight. The RNA complexes bound to the beads were then eluted and extracted with a RNeasy Mini Kit (Qiagen) for qRT-PCR analysis. The biotinylated probes were designed and synthesized by RiboBio (Guangzhou, China) and the sequences of the probes are listed in Table S1.

The Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore, Billerica, MA, USA) was used to perform the RIP assay. PDAC cells were collected and lysed in RIP lysis buffer supplemented with protease and RNase inhibitors. The cell lysates were then incubated with IgG, anti-AGO2 or anti-FUS antibody-coated beads (Millipore) at 4 °C overnight. The immunoprecipitated RNAs were subsequently extracted with a RNeasy MinElute Cleanup Kit (Qiagen, Valencia, CA, USA) after treatment with proteinase K buffer. Finally, the RNA levels of the assayed genes were measured by qRT-PCR.

Proteins were extracted from treated PDAC cells with RIPA buffer containing proteinase inhibitor. The protein concentrations in the cell lysates were measured by the DC Protein Assay Kit (Bio-Rad). Then, the proteins were separated via electrophoresis using SDS-containing polyacrylamide gels and then transferred onto a polyvinylidene fluoride (PVDF) membranes (Millipore, Billerica, MA, USA). After blocking the membranes with 5% nonfat dry milk in 0.1% Tween (TBST) buffer at room temperature for 2 h, the membranes were incubated at 4 °C overnight with the appropriate primary antibody. Subsequently, the membranes were washed 3 times with TBST buffer, and the membranes were incubated with a corresponding HRP-labelled secondary antibody for 2 h at room temperature, after which they were washed 3 times with TBST buffer. Finally, the western blot signals were visualized using an enhanced chemiluminescence detection system with Chemiluminescence HRP Substrate (Millipore, WBKL0100). All the primary and secondary antibodies used in this study are listed in Table S2.

Four-week-old male nude mice (BALB/c) were purchased from the Animal Center of Nanjing Medical University (Nanjing, China). All animal experiments were conducted in compliance with animal protocols approved by Nanjing Medical University, and were carried out at the Animal Center of Nanjing Medical University. Six mice per group were subcutaneously injected in the inguinal region with 1 × 10⁶ circRHOBTB3 lentivirus treated cells. The tumor volume was measured every 5 days and calculated according to the following formula: volume = (width² × length)/2. After 25 days, all the mice were sacrificed, and then tumors were resected and collected. All experiments were performed following the relevant institutional and national guidelines and regulations.

Human tumor tissues and xenografts were first fixed in 4% paraformaldehyde and embedded in paraffin, and 5 μm thick sections were cut. Then, antigen retrieval was performed by incubating the samples with sodium citrate buffer (pH 6.0) for 20 min at 95 °C, after which the samples were blocked with 5% normal goat serum for 10 min at 20 °C. Subsequently, the sections were incubated with polyclonal antibodies against NAC1 or Ki-67 at 4 °C overnight and then incubated with secondary antibodies. The tissue sections were scanned, and the protein levels were calculated as positive cells/total cells by Halo v3.0.311.314. All the primary and secondary antibodies used in this section are listed in Table S2.

PANC-1 and MiaPaca-2 cells transfected with mCherry-GFP-LC3 lentivirus (GeneChem, China) were seeded into a 35-mm culture dish for confocal microscopy. The nucleus was stained with DAPI. Red and yellow puncta representing autolysosomes and autophagosomes, respectively, were detected by confocal microscopy (Carl Zeiss, Germany). At least 10 cells in each of three independent experiments were analyzed randomly.

We placed the cell pellet in a droplet of 2.5% glutaraldehyde in PBS buffer at pH 7.2 and fixed the cells overnight at 4 °C. The samples were then rinsed in PBS solution for 10 min three times and postfixed in 1% osmium tetroxide for 60 min at room temperature. Next, the samples were embedded in 10% gelatin, fixed in glutaraldehyde at 4 °C and cut into several blocks. Subsequently, the samples were dehydrated for 10 min in increasing concentrations of alcohol (30, 50, 70, 90, 95, and 100% × 3). Next, we exchanged alcohol with propylene oxide and infiltrated samples with increasing concentrations (25, 50, 75, and 100%) of Quetol-812 epoxy resin mixed with propylene oxide. Each step lasts at least 3 h. The samples were then embedded in pure Quetol-812 epoxy resin and polymerized at 35 °C for 12 h, 45 °C for 12 h, and 60 °C for 24 h. We cut samples into sections (100 nm) using an ultramicrotome and poststained them with uranyl acetate for 10 min and lead citrate for 5 min at room temperature. After that, we observed sections under a transmission electron microscope operated at 120 kV.

Data are presented as the means ± standard deviations (SDs). The general statistical analysis was performed using GraphPad Prism 8.0 (GraphPad Software, La Jolla, USA). Student’s t-test was used to analyze the differences between two groups, while two-way ANOVA was used for multiple groups. Correlations between circRHOBTB3, miR-600 and NACC1 expression and various clinicopathological or serological variables were analyzed by the Mann-Whitney U test. Survival distributions and overall survival (OS) rates were determined using the Kaplan-Meier method, and the significance of differences between survival rates was calculated by the log-rank test. The univariate and multivariate Cox proportional hazards model, which was used to estimate the adjusted hazard ratios and 95% confidence intervals, as well as identify independent prognostic factors, was performed by SPSS 20.0 (IBM, SPSS, Chicago, IL, USA).

Our previous studies had characterized circular RNA transcripts using RNA-seq analysis of ribosomal RNA-depleted total RNA from three pairs of pancreatic ductal carcinoma and adjacent normal tissues, and identified circNEIL3 as an oncogene for PDAC progression and metastasis [20]. In this research, we characterized a circRNA derived from RHOBTB3 gene, circRHOBTB3 (chr5:95091099–95,099,324, circBase ID: hsa_circ_0007444). Based on the circBase annotation, circRHOBTB3 is a 479 nt length circRNA derived from exon6 and exon7 of the parental gene through back-splicing. Sanger sequencing confirmed the junction site with the PCR product (Fig. 1a). The agarose gel electrophoresis of the PCR products showed that circRHOBTB3 was only amplified from cDNA by divergent primers, ruling out the possibility of genomic rearrangements and trans-splicing (Fig. 1b). The half-time of the circRHOBTB3 transcript was significantly longer than RHOBTB3 mRNA after treated with actinomycin D, which suppressed RNA transcription. Moreover, the RNase R assays showed that circRHOBTB3 was resistant to RNase R treatment, which is a 3′ to 5′ exoribonuclease, while the linear counterpart mRNA considerably degraded after the enzyme treatment, illustrating the circular form of circRHOBTB3. (Fig. 1c-d). Furthermore, we investigated the cellular localization of circRHOBTB3 in PDAC cell lines, nuclear and cytoplasm fractionation and FISH assays indicated that circRHOBTB3 is mainly located in cytoplasm (Fig. 1e-f). We also detected higher circRHOBTB3 expression in four PDAC cell lines and 110 PDAC samples relative to the adjacent normal tissue samples via qRT-PCR (Fig. 1g-h), and its higher expression was also associated with poor overall survival (OS) than lower groups based on median expression (Fig. 1i). Collectively, these results demonstrated that circRHOBTB3, located in the cytoplasm of PDAC cells, is a highly expressed and stable circRNA.

To investigate the biological functions of circRHOBTB3 in PDAC cells, three siRNAs that specifically targeted the junction sites of circRHOBTB3 were constructed and transfected into PANC-1 and MiaPaca-2 cells (Fig. 2a). Based on the sequence of si-circRHOBTB3–1, we constructed knockdown lentivirus package to acquire stably transfected cell lines (Fig. 2b). And pGL3-circRHOBTB3 plasmid was used to overexpress circRHOBTB3 (Fig. 2c). Consequently, circRHOBTB3 expression level was significantly downregulated or upregulated in both PANC-1 and MiaPaca-2 cell lines, while the parental gene RHOBTB3 expression was barely affected. Based on these cell lines with different circRHOBTB3 expression level, gain-of-function assays were performed to evaluate the effects of circRHOBTB3 on the malignant potentials of pancreatic cancer cells with CCK8, colony formation and EdU incorporation assays. As shown, knocking-down circRHOBTB3 in pancreatic cells significantly inhibited cellular proliferation and colony formation, while overexpression showed the opposite effects (Fig. 2d-f, Figure S1a-b).

To further investigate the effects of circRHOBTB3 on tumor growth in vivo, mouse models of xenograft tumor growth were performed. We found that the tumor weight and volume were considerably reduced upon circRHOBTB3 knockdown, indicating a vital role of circRHOBTB3 in tumor growth (Fig. 2g). Meanwhile, Ki67 staining of xenograft tumor tissue demonstrated that in PANC-1 and MiaPaca-2 cells, circRHOBTB3 knockdown significantly suppressed proliferative activity in vivo (Fig. 2h). Taking together, these findings demonstrated that circRHOBTB3 plays a promotive role in PDAC proliferation in vivo and vitro.

CircRNAs mainly exhibited their biological functions in tumor progression by sponging to miRNAs in abundant previous studies [21]. To determine the miRNAs interacting with circRHOBTB3, we overlapped the results obtained from three public database miRanda (http://​www.​microrna.​org), RNAhybrid (https://​bibiserv.​cebitec.​uni-bielefeld.​de/​rnahybrid/​) and Targetscan (http://​www.​targetscan.​org/​mamm_​31/​) (Fig. 3a). A total of 57 miRNAs with predicted capacity of binding circRHOBTB3 were suggested and 11 candidate miRNAs with top confidence were selected (Table S3). Then, we performed RNA pull-down assays to further confirm the binding counterpart miRNAs of circRHOBTB3. The efficiency of circRHOBTB3 probe was demonstrated in circRHOBTB3 overexpression and control vector transfected cells (Fig. 3b). Among the 11 selected miRNAs, miR-600 was specifically enriched by circRHOBTB3 biotin-labelled probe in PANC-1 and MiaPaca-2 cell lines (Fig. 3c). On the other hand, an RNA pull-down assay using biotin-labelled miR-600 probe was performed, showing circRHOBTB3 was significantly enriched with biotin-miR-600 probe but not with the control probe in pancreatic cancer cells (Fig. 3d). Moreover, the function of circRNA as miRNA sponge requires binding with AGO2 and miRNA to form a circRNA-AGO2-miRNA complex [22]. Anti-AGO2 RNA immunoprecipitation (RIP) experiments confirmed that both circRHOBTB3 and mi-600 could bind to AGO2, indicating circRHOBTB3 could act as a miR-600 sponge. (Fig. 3e). To confirm the sponge effect of circRHOBTB3, we conducted a dual-luciferase assay by co-transfection of miR-600 mimics and a circRHOBTB3-WT or mutated plasmid with a luciferase reporter in PANC-1 and MiaPaca-2 cells respectively (Fig. 3f). The results showed that the luciferase activity of wild-type reporter was considerably reduced after transfection of miR-600 mimics, whereas no changes was noticed in the mutant reporter, demonstrating that circRHOBTB3 acts as a miR-600 sponge specifically binding at the CTGTAAG sites (Fig. 3g). Next, we checked the miR-600 expression level in different pancreatic cell lines and found that miR-600 expression was relatively lower in PDAC cell lines than HPNE cells (Fig. 3h). We further detected miR-600 expression in 110 pairs of PDAC tissues and adjacent normal tissues and the results showed that PDAC tissues exhibit decreased expression of miR-600 than adjacent normal tissues (Fig. 3i). Furthermore, FISH assays demonstrated that circRHOBTB3 and miR-600 co-localized in cytoplasm in pancreatic cancer cell lines and PDAC tissue (Fig. 3j, k). Besides, miR-600 expression was not influenced by circRHOBTB3 overexpression or knockdown, However, we found a negative linear correlation between the expression level of circRHOBTB3 and miR-600 in 110 cases of PDAC tissues, suggesting that circRHOBTB3 function as miR-600 sponge without affecting the expression of miR-600 (Fig. 3l). More importantly, a Kaplan-Meier analysis revealed that lower miR-600 expression was associated with decreased OS in PDAC patients (Fig. 3m). Taking together, our results indicated that circRHOBTB3 functions as a miR-600 sponge in PDAC.

In order to further illustrating the underlying mechanism by which miR-600 regulating PDAC functions, we performed rescue experiments by co-transfecting PDAC cells with circRHOBTB3 KD lentivirus or overexpression plasmid along with miR-600 mimics and inhibitors. The efficiency of miR-600 mimics and inhibitor transfection on miR-600 expression level in PANC-1 was verified by qRT-qPCR (Fig. 4a). The results revealed that the miR-600 inhibitor significantly promoted the proliferation of PANC-1 and reversed the proliferation suppressive effects by circRHOBTB3 downregulation through CCK-8, EdU incorporation assays and colony formation (Fig. 4b-d). Similar results were observed in MiaPaca-2 cells transfected with miR-600 mimics and the circRHOBTB3 plasmid (Figure S2a-c). Collectively, these experiments demonstrated that miR-600 has an inhibitory effect on PDAC cells and may serve a crucial function downstream of circRHOBTB3.

As illustrated in considerable literature, miRNAs interact with the 3’UTR region of target genes mRNA to induce its degradation by forming a RISC complex [23]. To further investigate the target genes of miR-600 in PDAC cells, we performed another bioinformatics analysis using online databases miRWalk (http://​mirwalk.​umm.​uni-heidelberg.​de/​), TargetScan (http://​www.​targetscan.​org/​mamm_​31/​), miRDB (http://​mirdb.​org/​) and miRTarbase (http://​mirtarbase.​cuhk.​edu.​cn/​php/​index.​php) (Fig. 5a). Through overlapping the results obtained from these four subjects, we paid attention on three candidate genes most possibly targeted by miR-600 for further validation. qRT-PCR showed that only NACC1 expression was negatively regulated by miR-600 mimics or inhibitors consistently in pancreatic cancer cell lines (Fig. 5b). Next, to confirm that miR-600 directly bind to NACC1 3’UTR region, we constructed luciferase reporter plasmid comprising the 3’UTR of NACC1 mRNA. Luciferase reporter assays showed that the transfection with miR-600 mimics could significantly reduce the luciferase activity of the wild-type but not the mutant 3’UTR NACC1 construct in both cell lines (Fig. 5c). Notably, NACC1 shares the same binding region with circRHOBTB3 on miR-600, which indicating its role as a downstream target on circRHOBTB3-miR-600 axis. We performed qRT-PCR analysis and found that NACC1 expression is positively regulated by circRHOBTB3 (Fig. 5d). Furthermore, we detected NACC1 expression on both mRNA and protein levels upon circRHOBTB3 KD co-transfected with miR-600 inhibitor in PANC-1 cells. The results showed that NACC1 expression was downregulated by circRHOBTB3 knockdown but could be rescued by miR-600 inhibitor. The opposite regulation circumstance was validated in MiaPaca-2 cells with circRHOBTB3 overexpression vector and miR-600 mimics (Fig. 5e-f). Similar with circRHOBTB3, NACC1 is highly expressed in PDAC cell lines. qRT-PCR assay in human PDAC samples showed that NACC1 expression in cancerous tissue was relatively higher than that in normal tissue (Fig. 5g-h). Besides, the IHC staining of PDAC tissues and counterparts supported the same result on protein levels (Fig. 5i). Linear correlation analysis of the tissue expression level of NACC1, miR-600 and circRHOBTB3 suggested that NACC1 expression was negatively correlated with miR-600 but positively correlated with circRHOBTB3, which accorded with the circRHOBTB3/miR-600/NACC1 regulatory axis hypothesis (Fig. 5j). Moreover, patients with higher NACC1 expression levels had poorer OS, which indicated that NACC1 could also be a risk factor for overall survival in PDAC patients (Fig. 5k). Taking together, circRHOBTB3 restrained miR-600 and further induced NACC1 degradation, while both circRHOBTB3 and NACC1 may function as oncogene in PDAC.

Given that NACC1 is the downstream target of miR-600 and circRHOBTB3, we further evaluated whether circRHOBTB3 promoted PDAC progression through NACC1. We transfected PANC-1 and MiaPaca-2 cell lines with NACC1 siRNA and overexpression plasmid. The efficiency of two constructs was validated by qRT-PCR and Western blot (Fig. 6a-b). Functional experiments showed that NACC1 overexpression in PANC-1 cells could promote cellular proliferation and colony formation, while the proliferation rate is largely reduced by NACC1 knockdown in MiaPaca-2 cells. Subsequently, we found that the inhibitory effect on cellular proliferation of circRHOBTB3 knockdown could be considerably reversed by NACC1 overexpression in PANC-1 cells, while the proliferation promoting effects of circRHOBTB3 overexpression could be retarded by NACC1 knockdown in MiaPaca-2 cells (Fig. 6c-e, Figure S3a-c). Taken together, our results indicated that circRHOBTB3 promotes the progression of PDAC via the miR-600/NACC1 axis.

NAC1 has been reported to be associated with autophagy in ovarian cancer [16], we wondered whether it could mediate PDAC cells autophagy response for proliferation. Subsequently, the autophagy flux detection revealed that autophagy level was suppressed after circRHOBTB3 knockdown in PANC-1 cells. Moreover, the increased autophagy after NACC1 overexpression could be suppressed by circRHOBTB3 knockdown, which indicated that circRHOBTB3 could regulate autophagy level of PDAC cells via controlling NACC1 expression (Fig. 7a). On the other hand, we performed the opposite effect experiments in MiaPaca-2 cells and the results revealed the same that circRHOBTB3 could promote autophagy flux in PDAC cells via miR600/NACC1 axis (Figure S4a). Meanwhile, the transmission electron telescope showed that there were less autophagosomes upon circRHOBTB3 knockdown but could be reversed by NACC1 overexpression in PANC-1 cells, and NACC1 siRNA could restrain the autophagy promoting effect of circRHOBTB3 overexpression in MiaPaca-2 cells, which coordinated the results we had above (Fig. 7b, Figure S4b). Besides, PANC-1 cells with circRHOBTB3 knockdown showed higher p62 levels but lower LC3B II levels, while NACC1 overexpression could partially reverse the expression change of autophagy-related protein. On the contrary, the pro-autophagy role of circRHOBTB3 overexpression could be retarded by NACC1 knockdown with siRNA (Fig. 7c, Figure S4c). In the view of above, our experiments revealed that circRHOBTB3 could promote PDAC cells autophagy by upregulating NACC1.

Given that autophagy is an adaptive response upon inadequate energy supplying for PDAC progression. We wondered whether the proliferation promoting effects of circRHOBTB3 relies on regulation of autophagy. Consequently, we performed another rescue experiment using autophagy inhibitor 3-MA (3-methyladenine) with circRHOBTB3 overexpression. We found that 3-MA treatment reversed the effects of circRHOBTB3 on autophagy-associated protein levels in PANC-1 and MiaPaca-2 cells (Fig. 8a, Figure S5a. Next, we transfected mCherry-GFP-LC3B labeled PDAC cells with circRHOBTB3 plasmid and 3-MA, and observed that the red and yellow LC3 puncta were increased in PDAC cells after circRHOBTB3 overexpression but restrained by 3-MA applying (Fig. 8b, Figure S5b). Besides, the transmission electron microscope showed that 3-MA successfully inhibited the augmented autophagy levels in PDAC cells with circRHOBTB3 overexpression (Fig. 8c, Figure S5c). On the other hand, we found that 3-MA treatment reversed the promotive effects of circRHOBTB3 overexpression on PANC-1 and MiaPaca-2 cell proliferation (Fig. 8d-f, Figure S5d-f). Collectively, these data indicated that circRHOBTB3 promotes PDAC cell proliferation by accelerating autophagy response.

As reported in various literature, Akt/mTOR pathway is a common pathway regulating PDAC tumor progression, and acts as an inhibitory role in autophagy response [24, 25]. Consequently, we conducted western blot and the results revealed that circRHOBTB3 knockdown increased protein phosphorylation levels but not the expression of Akt and mTOR. These effects could be reversed by NACC1 overexpression in PANC-1 cells. On the opposite, MiaPaca-2 cells showed decreased activation of mTOR and Akt signaling upon circRHOBTB3 overexpression, which could be rescued by NACC1 knockdown (Fig. 7d, Figure S4d). To further examine whether circRHOBTB3 promotes autophagy in PDAC cells by inhibiting Akt/mTOR signaling axis, circRHOBTB3 knockdown cells were treated with Akt phosphorylation inhibitor MK-2206 or mTOR phosphorylation inhibitor Rapamycin. We observed that MK-2206 and Rapamycin treatment alleviated the inhibitory effects of circRHOBTB3 on autophagy flux in PANC-1 and MiaPaca-2 cell lines (Fig. 9a). In parallel to decreased phosphorylation levels of Akt and mTOR, p62 protein decreased while LC3B II increased by MK-2206 or Rapamycin treatment in both cell lines (Fig. 9b, c). On the other hand, the Akt signaling activator SC79 and mTOR signaling activator MHY1485 displayed the opposite effects in MiaPaca-2 cells with circRHOBTB3 overexpression (Fig. 9a). These data confirmed that circRHOBTB3 facilitates PDAC cells autophagy by inhibiting Akt/mTOR signaling axis.

Once illustrated that circRHOBTB3 regulates PDAC progression via miR600/NACC1 axis, we wondered what the initial factors and the upstream regulatory mechanism of circRHOBTB3 in PDAC are. Firstly, we identified 8 candidate circRNA biogenesis-associated RNA-binding proteins (EIF4A, QKI, DHX9, FUS, EIF4A3, PRPF, ADAR and SF3A), and constructed siRNAs for all of them. The efficiency of each siRNA was confirmed by qRT-PCR. The results showed that circRHOBTB3 expression was significantly reduced only when the RNA-binding protein FUS was downregulated, circRHOBTB3 expression exhibited reduced levels, indicating that FUS could be a mediating protein that promotes circRHOBTB3 biogenesis (Fig. 10a). As FUS is a DNA/RNA-binding protein that plays a role in various cellular processes such as transcription regulation, RNA splicing, RNA transport, DNA repair and damage response [26], it also binds to nascent pre-mRNAs and acts as a molecular mediator between RNA polymerase II and U1 small nuclear ribonucleoprotein thereby coupling transcription and splicing [27]. Therefore, it has been reported that FUS controlled back-splicing reactions leading to circRNA production [9]. Then we assumed that whether FUS could bind to pre-RHOBTB3 mRNA and then promote the back-splicing process of circRHOBTB3 production. We performed RNA immunoprecipitation (RIP) assays and found that FUS antibody could bind to pre-RHOBTB3 mRNA but not circRHOBTB3 or RHOBTB3 mRNA (Fig. 10b). On the other hand, pull-down assays results conformed to above that the pre-RHOBTB3 mRNA probe could precipitate FUS protein in PANC-1 and MiaPaca-2 cells (Fig. 10c). Taken together, we primarily validated the biogenesis promoting role of FUS in circRHOBTB3 biogenesis in PDAC cells. Given the discovery of the circRHOBTB3/miR-600/NACC1 axis in PDAC cells and that the characteristics of each subject have been illustrated, we next evaluated the clinical significance of circRHOBTB3, miR-600 and NACC1 in a cohort of 110 patients. Patients were divided into high and low groups based on the median expression. The relationship between circRHOBTB3, miR-600 and NACC1 expression and the clinical characteristics of PDAC patients are listed in Table 1. The results revealed that tumor size, vascular invasion, and clinical stage, especially T stage were significantly associated with at least one of these three genes. Further univariate and multivariate Cox regression analysis showed that circRHOBTB3, miR-600 and NACC1 expression levels were independent prognostic factors for PDAC patients, as were the tumor size and clinical stages (Table 2). Thereupon, we had a conclusion that FUS-mediated circRHOBTB3/miR-600/NACC1 axis is correlated with PDAC prognosis.