UBL4A inhibits autophagy-mediated proliferation and metastasis of pancreatic ductal adenocarcinoma via targeting LAMP1

PDAC and normal pancreatic tissues were obtained from 69 patients who underwent pancreatectomy in the Department of Pancreatic and Biliary Surgery (The First Affiliated Hospital of Harbin Medical University, Harbin, Heilongjiang, China) from January 2009 to January 2015. Informed consent was obtained from each patient prior to biopsy or surgery, and ethical approval for the use of human subjects was obtained from the Research Ethics Committee of the First Affiliated Hospital of Harbin Medical University. The patients’ clinical characteristics are shown in Table 1.

Flag-tagged lentiviral vectors encoding human UBL4A (Lv-UBL4A-Flag) and empty vectors were constructed in GV341 (GeneChem, Shanghai China). The lentiviruses with scrambled shRNA against UBL4A (Lv-shUBL4A) and the shRNA control (shCtrl) were constructed in GV112 (GeneChem, Shanghai China). After lentiviral infection, single-cell clones were selected by 2.5 μg/ml puromycin (Sigma-Aldrich, St. Louis, MO, USA) for 2 to 4 weeks. The SiRNA for LAMP1 (si-LAMP1) and the siRNA negative control (si-NC) were purchased from RIBOBIO (Guangzhou, China). To induce overexpression of LAMP1, human LAMP1 (NM_005561) cDNA was cloned into a plasmid (GeneCopoeia, Guangzhou, China). For transient transfection, the cells were seeded in six well plates, and 50 nm siRNA or 2 μg plasmids were transfected into the cells using Lipofectamine 2000 (Life Technologies Limited Paisley, Grand Island, NY, USA). The efficiency of all transfections was evaluated by qRT-PCR and/or western blotting. The target sequences of the lentiviruses and siRNAs are listed in Additional file 1: Table S1.

Total RNA was extracted and isolated from the cell lines and frozen tumor specimens using a AxyPrep Multisource Total RNA Miniprep Kit from Axygen (Corning, Suzhou, Jiangsu, China), and the first strand cDNA was synthesized using the Rever TraAce qPCR RT Kit Master Mix with gDNA Remover (FSQ-301, Toyobo Co. Ltd. Osaka, Osaka Prefecture, Japan) according to the manufacturer’s instructions. Quantitative real-time polymerase chain reaction (qRT-PCR) was performed as previously described [10]. Briefly, qRT-PCR (SYBR Green Assay, Roche Diagnostics GmbH, Indianapolis, IN, USA) was performed on a 7500 FAST Real-Time PCR System (Applied Biosystems). The relative expression levels of the mRNA were calculated and quantified using the 2^−ΔΔT method after normalization to the expression of the control. GAPDH served as the endogenous control. The primer sequences are described in Additional file 2: Table S2 and were purchased from Comate Bioscience (Institute of Biotechnology, Jilin, China).

Electron microscopy was performed as previously described [10]. Pancreatic cancer cell lines were fixed in 2.5% glutaraldehyde and postfixed in 1% osmium tetroxide buffer. Tissues were embedded in spur resin, and thin sections were cut. The sectioned grids were stained with a saturated solution of uranyl acetate and lead citrate. Sections were examined at 80 kV using a JEOL 1200EX transmission electron microscope.

The GFP-mRFP-LC3 lentivirus was purchased from GeneChem (Shanghai China). Pancreatic cancer cell lines (CFPAC-1 and PANC-1) cultured on covers lips were transduced with the control and the GFP-mRFP-LC3 lentiviral vectoros and were then selected with puromycin (Sigma-Aldrich, St. Louis, MO, USA) for one week. Stably transfected cells were infected with the LV-UBL4A-Flag, LV-shUBL4A, LAMP1 plasmid, si-LAMP1 and their respective controls. The cells were viewed under a fluorescence microscope. Theoretically, GFP is a stably folded protein and is relatively resistant to lysosomal proteases. However, the low pH level inside the lysosomes quenches the fluorescent signal of the GFP. Therefore, autophagosomes and autolysosomes were labeled yellow (mRFP and GFP) and red (mRFP only), respectively. The numbers of GFP and mRFP dots were determined by manual counting of fluorescent puncta in five high-power fields (20×, Olympus) and analyzed using Image-Pro Plus v6.0 software. We calculated the ratio of autolysosomes (red) to autophagosomes (yellow) per cell to evaluate the extent of autophagosomes maturation into autolysosomes.

The tumor cells transfected with Lv-UBL4A-Flag and Lv-shUBL4A as well as their controls were seeded on 24-well plates. The cells were fixed with 4% paraformaldehyde for 30 min and were permeabilized with 0.5% Triton X-100 for 20 min. After incubation for 2 h with anti-UBL4A (Proteintech, Wuhan, Hubei, China), anti-LAMP1 (Abcam, Shanghai, China) and anti-LC3B (Cell Signaling Technology, Danvers, MA, USA), the cells were washed with PBS three times. Then, the cells were incubated with secondary antibodies for 1 h (Bioss, Beijing, China), and 4′6-diamino-2-phenylindole (DAPI, Beyotime Biotechnology, Shanghai, China) was added to stain the cell nuclei. Finally, the cells were viewed with a fluorescence microscope (20×, Olympus). Information on the primary antibodies for immunofluorescence is provided in Additional file 3: Table S3.

Transfected cells were stained with Lyso-Tracker Red (Beyotime Biotechnology, Shanghai, China) according to the manufacturer’s instructions, and a fluorescence microscope (20×, Olympus) was used to observe fluorescence. The red signals (lysosome) were quantified and are shown in s bar graph with relevant statistics.

First, 2 × 10⁷ cells were lysed with 500 μL of ice-cold polysome lysis buffer (100 mM KCl; 4 mM MgCl₂; 10 mM HEPES, pH 7.0; 0.5% Nonidet P-40; 1 mM DTT; 100 units/ml RNase OUT; and 40 μL/mL complete protease inhibitor cocktail) for 10 min on ice. Then, the cell lysates were collected after centrifugation. UBL4A antibodies (Proteintech, Wuhan, Hubei, China) or control immunoglobulin (IgG) (Beyotime Biotechnology, Shanghai, China) were added to protein G agarose beads (EMD Millipore Corporation, Temecula, CA, USA) and allowed to bind while rotating at 4 °C overnight (for UBL4A and LAMP1) or 2 h (for IgG). The lysates were precleared with an IgG antibody and then incubated with precoated beads for 2 h at RT on a rotator. After the beads were washed, the complexes were boiled for 10 min at 100 °C and loaded on a gel for western blotting. Information on the primary antibodies for co-IP is provided in Additional file 3: Table S3.

Orthotopic tumor models were created as previously described [10]. The study protocol was approved by the Institutional Review Board of The First Affiliated Hospital of Harbin Medical University. Briefly, two luciferase-expressing cell lines (CFPAC-1-UBL4A and PANC-1-shUBL4A and their control cells) (5 × 106/0.2 mL) were injected into the right flanks of nude mice. Then, 1 mm3 pieces of tumor harvested from four mice were transplanted into four groups of mouse pancreatic tails. The animals were imaged weekly using the Night OWL II LB983 imaging system in vivo (Berthold Technologies GmbH & Co. KG, Germany). After 5 weeks of xenograft procedures, the mice were sacrificed, and the numbers of visible metastatic lesions in the gut, mesentery, liver, spleen and kidneys were recorded. The primary and metastatic pancreatic tumors were excised, weighed, and fixed in 4% paraformaldehyde.

Statistical analysis was performed with SPSS 19.0 software or GraphPad Prism 6.01 software. The data were shown as the mean ± standard deviation (SD). Pearson analysis and Kaplan–Meier survival analysis were used to evaluate the statistical significance, and the variance between the two groups was analyzed using Student t-tests. Differences were considered significant when *, P < 0.05; **, P < 0.01; ***, P < 0.001; and non-significant when P > 0.05.

Details on other experimental procedures are described in the Supplementary Methods (Additional file 9: Supplementary Methods).

Sixty-nine patients who underwent pancreatectomy for PDAC were included in this study. The expression of UBL4A in PDAC tissues was compared with that in normal tissues (18/69 patients) by qRT-PCR assays (Fig. 1a). Moreover, patients with high levels of UBL4A mRNA exhibited significantly longer overall survival than those whose tumors expressed a low UBL4A mRNA level (Fig. 1b). A clinical association analysis indicated that UBL4A expression was significantly associated with the TNM stage and lymph node metastasis in PDAC (Table 1). Moreover, compared with normal tissue samples, the level of UBL4A protein was lower in PDAC samples, as determined by immunohistochemistry (IHC) and western blot analysis (Fig. 1c-f). Regarding different cell lines, the UBL4A protein and mRNA levels were detected in HPDE cells, the human pancreatic duct epithelial cell line, and four PDAC cell lines (Fig. 1g-i). Together, all of these results demonstrate that the level of UBL4A is reduced in both PDAC tissues and cell lines and that elevated UBL4A is associated with longer survival.

To assess the role of UBL4A in PDAC cell proliferation and metastasis, four pancreatic cancer cell lines (SW1990, CFPAC-1, PANC-1 and BxPC-3) were introduced. Through lentivirus transfection, UBL4A was effectively overexpressed in the CFPAC-1 and SW1990 cells and silenced in the PANC-1 and BxPC-3 cells. Both the overexpression and knockdown efficiency of UBL4A were confirmed by qRT-PCR and western blot assays (Additional file 4: Figure S1a-f). Cell growth assays showed that UBL4A overexpression (LV-UBL4A) suppressed cell viability and blocked the DNA replication in the indicated PDAC cells. Moreover, knockdown of UBL4A (LV-shUBL4A) in the PANC-1 and BxPC-3 cells increased their proliferation and DNA replication capability (Fig. 2a, Additional file 4: Figure S1 g-j). Consistently, colony formation assays demonstrated that the UBL4A upregulation (LV-UBL4A) groups led to the generation of fewer and smaller colonies compared with the control groups and that suppression of UBL4A (LV-shUBL4A) increased the number of colonies in the PANC-1 and BxPC-3 cells relative to the control groups (Fig. 2b-c). To assess the effect of UBL4A on the potency of the migration and invasion of pancreatic cancer cells, transwell and wound healing assays were performed. Our results demonstrated that elevated expression of UBL4A led to a significant decrease in the migratory and invasive capabilities of the CFPAC-1 and SW1990 cells. In a corresponding finding, UBL4A knockdown led to the significant promotion of cell migration and invasion in the PANC-1 and BxPC-3 cells (Fig. 2d-j). It has been widely accepted that epithelial-to-mesenchymal transition (EMT) enables tumor cells to acquire the features of high motility and invasiveness [21]. Moreover, our previous study proved that EMT progression played a crucial role in pancreatic cancer metastasis [22]. Thus, the expression of epithelial marker (E-cadherin) and mesenchymal markers (N-cadherin and vimentin) was evaluated by western blotting, and the results were consistent with those described above (Fig. 2k-l). Together, these results indicate that UBL4A inhibits tumor proliferation and metastasis in pancreatic cancer cells.

Autophagy is important to support tumor growth in extreme environments [23]. Previous studies have demonstrated the tumorigenic roles of autophagy in pancreatic cancer [10, 24]. To determine whether autophagy is required for UBL4A-mediated tumor suppression, the correlation between autophagy and UBL4A was further tested, and the autophagy inhibitor chloroquine (CQ) was introduced to the experimental cells. LC3B (microtubule-associated protein 1 light chain 3β) plays a crucial role in functional autophagosome formation, and p62 has been suggested to act as a chaperone during the degradation of autophagosomes [25]. In addition, LC3BII, as detected by western blotting, and autophagosome formation, which is assessed by electron microscopy, have been regarded as standard markers for autophagy [14]. Our results showed that the levels of LC3BII and p62 were significantly elevated in the LV-UBL4A groups. In contrast, LV-shUBL4A decreased the LC3BII and p62 protein levels in PANC-1 and BxPC-3 cells (Fig. 3a, Additional file 5: Figure S2a-b). Consistent with the western blotting results, electron microscopy revealed a significant increase in autophagosomes in SW1990 and CFPAC-1 cells transfected with UBL4A-expressing lentiviral vectors compared with the control groups (Fig. 3b). These results indicate that UBL4A causes the accumulation of autophagosomes. In addition, UBL4A failed to induce changes in autophagy or in E-cadherin or N-cadherin expression in the presence of CQ (Fig. 3c-d, Additional file 5: Figure S2c-d). To explore whether the UBL4A inhibition of pancreatic cancer proliferation and metastasis was associated with autophagy suppression, EdU retention assays (Additional file 6: Figure S3a-c), transwell assays (Fig. 3e-g, Additional file 6: Figure S3d-f) and wound-healing assays (Fig. 3h-j, Additional file 6: FigureS3 g-i) were carried out. Consistently, CQ abolished the antitumor effects of UBL4A on tumor proliferation and metastasis in vitro. These findings consistently demonstrate that UBL4A suppresses tumor proliferation and metastasis through inhibition of autophagy, similar to the accumulation of autophagosomes.

To explore whether accelerated autophagosome synthesis or reduced autophagic vacuole maturation and degradation is responsible for UBL4A-induced autophagosome accumulation in cells, the autophagy-related genes ATG5 and ATG7, key molecules involved in autophagosome formation, were introduced [26, 27]. Western blot assays were subsequently used to analyze the expression levels of ATG5 and ATG7 among different groups. However, as shown in Additional file 7: Figure S4a, neither of these two ATGs exhibited noteworthy changes with the downregulation or upregulation of UBL4A. Thus, we speculate that UBL4A may play a role in autolysosome degradation instead of autophagosome formation. Consistently, our data showed that CQ abolished the effects of UBL4A on autophagy inhibition and tumor suppression and had no synergetic function with UBL4A (Fig. 3c-j, Additional file 6: Figure S3). These findings confirmed that UBL4A participated in regulating autolysosome degradation, which occurs during the late stage of autophagy. In addition, a tandem-labeled GFP-mRFP-LC3 reporter was used to measure autophagic flux. GFP is a stably folded protein that is resistant to lysosomal proteases. The low pH level inside lysosomes quenches the fluorescent signal of GFP. Autophagosomes and autolysosomes were labeled yellow (mRFP and GFP) and red (mRFP only), respectively [28]. As shown in Fig. 3k and m, the ratios of red dots (autolysosomes) to yellow dots (autophagosomes) per cell were significantly decreased in the UBL4A overexpressing groups compared with those in the control groups, indicating that autophagic flux was hampered after UBL4A upregulation. In contrast, suppression of UBL4A increased the ratio of red dots (autolysosomes) to yellow dots (autophagosomes) per cell in comparison with the control groups (Fig. 3l-m). Furthermore, CQ simultaneously blocked autophagic flux and eliminated the influence of UBL4A on autophagic flux (Fig. 3k-m). Together, these data indicate that UBL4A is a potent autophagic inhibitor that causes autophagosome accumulation due to impaired autophagic degradation and suppresses pancreatic cancer proliferation and metastasis by inhibiting autophagy.

Impaired autophagic degradation is closely related to the immaturity of autolysosomes in that it is caused by either defective fusion between autophagosomes and lysosomes or lysosomal dysfunction [29]. To determine the exact mechanism of UBL4A-induced impaired autophagic degradation, we used fluorescence microscopy to determine the collocation of LC3B and LAMP1. Both the upregulated and downregulated UBL4A groups exhibited collocation of LC3B (red puncta) and LAMP1 (green puncta), similar to their respective control groups (Fig. 4a-b). These results indicate that UBL4A may not affect autophagosome-lysosome fusion. Thus, we next attempted to explore whether UBL4A influenced lysosomal function. Lyso-Tracker Red DND-99 is a fluorescent probe with weak alkalinity that can selectively remain in acidic lysosomes to specifically label lysosomes. Our work showed that the intensity of Lyso-Tracker Red fluorescence was lower in UBL4A overexpressing cells but higher in UBL4A-knockdown cells compared with the controls. CQ remarkably reduced the fluorescence intensity as expected in the UBL4A upregulated and downregulated groups (Fig. 4c-f). These results indicate that UBL4A causes lysosomal alkalinization in PDAC cells. Moreover, lysosomal degradation depends on the amount and activity of hydrolases. Cathepsins are the most studied lysosomal hydrolases that participate in autophagic and lysosomal degradation [30], and the relationship between UBL4A and cathepsins is still unknown. We then investigated whether UBL4A affects the expression and enzymatic activity of cathepsin B (CTSB). Reduced expression of CTSB was observed at the protein level and primarily involved the mature forms of the protein in UBL4A overexpressing cells compared with the control groups. Consistently, UBL4A down-regulation in PANC-1 and BxPC-3 cells increased the activity of proteases compared to the control groups (Fig. 4g-h). Together, these data verify that UBL4A causes lysosomal dysfunction due to simultaneous lysosomal alkalinization and inhibition of the expression and maturation of CTSB. Moreover, the impaired autophagic degradation caused by UBL4A is attributed to lysosomal dysfunction rather than impaired fusion between autophagosomes and lysosomes.

To identify the putative downstream target of UBL4A, four online interactive protein databases, BIOGRID version 3.5.168, I2D (Interologous Interaction Database) version 2.9 [31], UALCAN [32] and STRING version 10.5, were used simultaneously. LAMP1, GET4 and FAF2 were found in all four databases (Fig. 5a). We focused on LAMP1 because of its involvement in lysosomal exocytosis, in movement of lysosomes along microtubules, and in the fusion of phagosomes with lysosomes [33]; it may also play a role in tumor cell metastasis [34]. Furthermore, LAMP1, as a lysosomal marker, had already been detected by immunofluorescence in our study (Fig. 4a, b), and its expression was positively correlated with UBL4A. To validate the potential correlation, the mRNA expression of UBL4A and LAMP1 in 69 PDAC samples was quantitatively analyzed by qRT-PCR. As shown in Additional file 7: Figure S4b, the LAMP1 mRNA levels were positively related to the expression of UBL4A in PDAC tissues (r² = 0.2879, P < 0.001). Western blotting showed that UBL4A inhibition significantly reduced the LAMP1 protein level and that increased UBL4A remarkably increased LAMP1 protein expression compared to their respective control groups (Fig. 5b, Additional file 7: Figure S4e). In addition, we found that UBL4A inhibition remarkably decreased the LAMP1 mRNA levels, and increased UBL4A dramatically upregulated LAMP1 mRNA expression compared to the control groups (Additional file 7: Figure S4c-d). For additional analyses, immunofluorescence (IF) staining was performed. UBL4A-overexpressing cells enhanced the expression of LAMP1, and suppression of UBL4A reduced the fluorescence intensity of LAMP1 (Fig. 5c-d). In addition, the collocation of UBL4A and LAMP1 was shown in IF staining assays, providing rational evidence for their interaction in the cytoplasm (Fig. 5c-d). To ascertain whether UBL4A directly interacts with LAMP1 in pancreatic cancer cells, UBL4A and LAMP1 were immunoprecipitated from cell lysates of the pancreatic cancer cells. As shown in Fig. 5e-f, the specific, direct interaction between UBL4A and LAMP1 was confirmed by a coimmunoprecipitation (co-IP) assay. These results indicate that UBL4A directly targets LAMP1 in PDAC.

To verify the role of LAMP1 in UBL4A-induced tumor suppression and inhibition of autophagy, a siRNA against LAMP1 and a LAMP1 overexpression plasmid were introduced. The silencing and overexpressing efficiencies were confirmed by western blotting and qRT-PCR. The expression of LAMP1 was notably decreased by si-LAMP1#2, and si-LAMP1#2 was chosen for further experiments (Additional file 8: Figure S5a-d). Knockdown of LAMP1 reversed the UBL4A-induced autophagy suppression, mesenchymal-to-epithelial transition (Fig. 6a-b), inhibiton of autophagic flux (Fig. 6e), inhibition of cancer cell migration and invasion (Fig. 7a-c), cell proliferation (Fig. 7g) and wound healing capacity (Fig. 7j). Consistently, LAMP1 overexpression reversed the stimulation of autophagy in UBL4A-knockdown cells (Fig. 6c-d). CTSB and N-cadherin levels (Fig. 6c-d), autophagic flux activation (Fig. 6f-g), cell migration and invasion (Fig. 7d-f), DNA replication activity (Fig. 7h) and wound healing ability (Fig. 7k-l) were also reversed by LAMP1 overexpression in UBL4A-knockdown cells. Together, these data indicate that LAMP1 restoration in pancreatic cancer cells reverses the UBL4A-induced antitumor effects and inhibition of autophagy. In summary, these results demonstrate that UBL4A inhibits autophagy-mediated proliferation and metastasis of pancreatic ductal adenocarcinoma by directly targeting LAMP1.

To assess the role of UBL4A on PDAC proliferation and metastasis in vivo, stably transfected cell lines (CFPAC-1-Vector, CFPAC-1-UBL4A, PANC-1-shCtrl and PANC-1-shUBL4A) were used to develop a pancreatic orthotopic tumor model. The animals were imaged weekly, and the sizes of the tumors were recorded at day 35. As shown in Fig. 8a, compared with the control groups, the average size of tumors in the UBL4A-overexpression group increased relatively slowly. High UBL4A expression induced a small tumor size and reduced tumor weight in the orthotopic tumor model (Fig. 8b-c). In contrast, larger and heavier pancreatic orthotopic tumors were observed in the UBL4A-knockdown groups relative to the controls (Fig. 8d-f). Furthermore, fewer visible metastatic nodes were found in the gut of mice administered cells of the UBL4A-overexpression groups compared with those administered the control groups. In contrast, knockdown of UBL4A promoted metastatic node formation in the gut compared with the controls (Fig. 8g-h). Additionally, immunohistochemical results showed that the UBL4A level was positively correlated with the expression of LAMP1, LC3B, p62, and E-cadherin and negatively correlated with CTSB, vimentin, and N-cadherin in pancreatic cancer patients (Fig. 8i). Moreover, increased expression of UBL4A caused increased levels of LAMP1, LC3B, p62, and E-cadherin and resulted in decreased levels of CTSB, vimentin, and N-cadherin in the orthotopic pancreatic cancer models (Fig. 8j). In contrast, the opposite effects were observed in the UBL4A-knockdown groups (Fig. 8k). Taken together, our data suggest that UBL4A inhibits tumor proliferation and metastasis in orthotopic pancreatic cancer models.