PVT1 was significantly upregulated in GBC tissues and associated with poor prognosis of GBC patients
Firstly, we performed qRT-PCR to identify the expression levels of PVT1 in 55 pairs of GBC tissues and adjacent normal tissues. As shown in Fig. 1A, PVT1 was significantly upregulated in GBC tissues compared to the adjacent normal tissues. Moreover, 55 GBC patients were divided into two groups according to the median ratio of the relative PVT1 expression in tumor tissues: the high group (n = 30, fold change ≥ mean ratio) and the low group (n = 25, fold change < mean ratio). Pearson chi-square tests were used to analyze the relationship between patients’ clinicopathologic features and PVT1 expression level. It was demonstrated that high expression of PVT1 was related with lymph node metastasis, histological grade, and TNM stage but not gender, age and tumor size of GBC patients (Table 1).
Table 1
Correlation between PVT1 expression and clinicopathologic features of GBC patients.
Characteristics | Case number | PVT1 expression | p-Value |
High (n = 30) Low (n = 25) |
Gender | | | | 0.79 |
Male | 30 | 17 | 13 | |
Female | 25 | 13 | 12 | |
Age | | | | 0.282 |
≤60 | 27 | 17 | 10 | |
>60 | 28 | 13 | 15 | |
Tumor size | | | | 0.285 |
≤5 cm | 26 | 12 | 14 | |
> 5 cm | 29 | 18 | 11 | |
Lymph node metastasis | | | | 0.032* |
Yes | 31 | 21 | 10 | |
No | 24 | 9 | 15 | |
Histological grade | | | | 0.016* |
Well and morderately | 23 | 8 | 15 | |
Poorly and others | 32 | 22 | 10 | |
TNM stage | | | | 0.011* |
I-II | 20 | 6 | 14 | |
III–IV | 35 | 24 | 11 | |
*p < 0.05 |
Furthermore, we conducted Kaplan-Meier survival analysis and found that patients with high PVT1 levels had a shorter survival compared to those with low levels (Fig. 1B). Univariate survival analysis showed that lymph node metastasis, histological grade, TNM stage and high PVT1 expression were prognostic factors. Multivariate Cox regression analysis demonstrated that only lymph node metastasis, TNM stage and high PVT1 expression were independent prognostic factors for GBC patients (Table 2).
Table 2
Univariate and multivariate analysis of prognostic factors for overall survival in GBC patients.
Characteristics | Univariate analysis | | multivariate analysis |
HR | P-value | | HR(95%CI) | P-value |
Gender | 0.550 | 0.458 | | | |
Age | 1.445 | 0.229 | | | |
Tumor size | 3.592 | 0.058 | | | |
Lymph node metastasis | 9.061 | 0.003** | | 11.395(2.624–19.484) | 0.001** |
Histological grade | 5.482 | 0.019* | | 1.295(0.503–3.329) | 0.592 |
TNM stage | 12.393 | < 0.001*** | | 14.512(6.592–19.141) | < 0.001*** |
PVT1 expression | 13.485 | < 0.001*** | | 0.322(0.152–0.680) | 0.003** |
*p < 0.05,**p < 0.01, *** p < 0.001 |
PVT1 was upregulated in GBC cell lines and promoted GBC cells proliferation in vitro
Then the expression of PVT1 in five GBC cell lines (SGC-996, EHGB-1, NOZ, OCUG-1, GBC-SD) and a human non-tumorigenic biliary epithelial cell line (H69) was evaluated by qRT-PCR. It was found that PVT1 was significantly upregulated in GBC cell lines than H69 (Fig. 2A). To further study the function of PVT1, we transfected two PVT1 siRNAs into GBC-SD cell line, which had relatively high expression of PVT1. We also designed pcDNA-PVT1 for ectopic expression in SGC-996 cell line, which had relatively low expression of PVT1. The efficiency of pcDNA-PVT1 and siRNAs was examined by qRT-PCR as shown in Fig. 2B and 2C. So we chose si-PVT1-1 and pcDNA-PVT1 for the following experiments.
Next, CCK-8 and colony formation assays were conducted to demonstrate the effect of PVT1 on GBC cell proliferation. We found that knockdown of PVT1 significantly attenuated the proliferation and the cloning ability of GBC-SD cells, whereas, overexpression of PVT1 promoted the proliferation and the cloning ability of SGC-996 cells (Fig. 2D and 2E).
Then flow cytometry analysis was conducted to investigate whether the influence of PVT1 on GBC cells proliferation was due to its regulation on cell cycle. And we found that PVT1 downregulation led to a significant G1/G0 phase arrest in GBC-SD cells and vice versa in SGC-996 cells when PVT1 was overexpressed (Fig. 2F). Furthermore, we examined the expression of CDK4 and CyclinD1 which were related with G1/G0 phase by western blotting, and it showed that CDK4 and CyclinD1 were decreased in GBC-SD cells transfected with si-PVT1-1, and vice versa in SGC-996 cells transfected with pcDNA-PVT1 (Fig. 2G). These results suggested that PVT1 acted as an oncogene that promoted the proliferation of GBC cells.
Knockdown of PVT1 inhibited tumor growth in vivo
To further confirm the function of PVT1 on tumorigenesis in vivo, sh-PVT1 was designed, and the significant knockdown efficiency of sh-PVT1 was shown in Fig. 3A. Then the GBC-SD cells transfected with either sh-NC or sh-PVT1 were injected subcutaneously into nude mice. As shown in Fig. 3B and 3C, knockdown of PVT1 obviously inhibited the tumor development. Meanwhile, the tumor weight from sh-PVT1 group was significantly lighter than that from the sh-NC group (Fig. 3D). Moreover, we conducted qRT-PCR and confirmed the downregulation of PVT1 in tissues from sh-PVT1 group (Fig. 3E). Furthermore, immunohistochemical staining demonstrated decreased Ki67 positivity in tumors from sh-PVT1 group (Fig. 3F). Taken together, these results suggested that knockdown of PVT1 could inhibit GBC cell proliferation in vivo.
PVT1 inhibited miR-18b-5p expression via promoting DNA promoter methylation
LncRNAs always exercise their function through regulating or interacting with microRNAs[14, 15]. To determine whether PVT1 promotes GBC cells proliferation by regulating microRNA, we compared the expression profiles of microRNAs in GBC-SD cells transfected with si-PVT1-1 or si-NC by miRNA microarray assay. As shown in Fig. 4A, the variation of microRNAs expression between GBC-SD/si-PVT1-1 and GBC-SD/si-NC was exhibited by the scatter and volcano plots. In total, 22 upregulated microRNAs and 46 downregulated microRNAs with p value < 0.05 were identified (supplementary Table 2). The differently expressed microRNAs were also shown by hierarchical clustering analysis in Fig. 4B.
Among the aberrantly upregulated microRNAs, we aimed at screening out the microRNAs that were related to the proliferation of GBC cells. We selected 5 candidates with highest fold change and knocked down their expression with microRNA inhibitor respectively. In these 5 microRNAs, we found that only miR-18b-5p inhibited the proliferation ability of GBC-SD cells by CCK-8 assays (data not shown).
To further verify the negative regulation of miR-18b-5p by PVT1, qRT-PCR was conducted. It was found that knockdown of PVT1 markedly enhanced the expression of miR-18b-5p in GBC-SD cells, and vice versa in SGC-996 cells when PVT1 was overexpressed (Fig. 5A). LncRNAs always regulate miRNAs by serving as “miRNA sponges”, but we did not find the potential binding sites of PVT1 and miR-18b-5p by analyzing their sequences, which indicated that PVT1 may inhibited miR-18b-5p expression by other ways. DNA methylation was an important way to regulate gene expression[16], we supposed that the inhibition of miR-18b-5p by PVT1 may be due to the DNA promoter methylation. Then the CpG island location of miR-18b-5p promoter regions was predicted by http://www.urogene.org/ as shown in Fig. 5B, suggesting a potential involvement of DNA methylation in the regulation of miR-18b-5p expression. Furthermore, bisulfate sequencing PCR (BSP) was conducted to validate the effect of PVT1 on DNA methylation level of miR-18b-5p promoter region, it was demonstrated that knockdown of PVT1 in GBC-SD cells significantly attenuated DNA methylation level at the CpG island of miR-18b-5p promoter region, and vice versa in SGC-996 cells when PVT1 was upregulated (Fig. 5C). Moreover, the DNA methylation inhibitor 5-aza-dC was used to further examine the role of DNA methylation in the regulation of miR-18b-5p expression, and it showed that 5-aza-dC treatment significantly increased the expression of miR-18b-5p in GBC-SD and SGC-996 cells (Fig. 5D). These results suggested that PVT1 inhibited miR-18b-5p expression via DNA promoter methyaltion.
PVT1 promoted miR-18b-5p DNA promoter methylation by recruiting DNMT1 via EZH2
To explore the mechanism by which PVT1 promoted miR-18b-5p DNA promoter methylation, we firstly conducted qRT-PCR to determine the subcellular localization of PVT1 in GBC-SD and SGC-996 cells. It demonstrated that PVT1 expressed higher in the nucleus than the cytosol in both cell lines (Fig. 6A), suggesting that PVT1 may function through binding with RNA-binding proteins directly and regulate gene expression at the transcriptional level.
In the process of DNA promoter methylation, DNMTs which included DNMT1, DNMT3A, and DNMT3B played crutial roles[17]. To identify the DNMT which was responsible for miR-18b-5p DNA promoter methylation, we designed siRNAs of DNMT1, DNMT3A, and DNMT3B. QRT-PCR found that the expression of miR-18b-5p in GBC-SD and SGC-996 cells increased only when the cells were transfected with siRNAs of DNMT1 (Supplementary Fig. 1A). Therefore, we concluded that it was DNMT1 that executed the DNA promoter methylation of miR-18b-5p. Furthermore, to validate whether PVT1 was required for DNMT1 binding to the promoter region of miR-18b-5p, ChIP was conducted, and it showed that PVT1 increased the binding of DNMT1 to the miR-18b-5p promoter region (Fig. 6B).
Then we performed RIP to validate the association of PVT1 with DNMT1, interestingly, the results showed that PVT1 could not directly bind to DNMT1 in both GBC-SD and SGC-996 cells (Supplementary Fig. 1B). It had been reported that many lncRNAs located in the nucleus could bind with EZH2, a catalytic subunit of PRC2, and then regulate downstream target genes[18]. We supposed that PVT1 may function through binding with EZH2, so RIP was conducted, and it was found that PVT1 could directly bind to EZH2 in GBC-SD and SGC-996 cells (Fig. 6C). In addition, We performed Co-IP assays to evaluate whether EZH2 was responsible for DNMT1 binding to the miR-18b-5p promoter region, as shown in Fig. 6D, EZH2 could physically interact with DNMT1 in GBC cells, and the interaction could be hampered when PVT1 was downregulated. Collectively, these findings validated that PVT1 promoted miR-18b-5p DNA promoter methylation by recruiting DNMT1 via EZH2.
HIF1A was downregulated by miR-18b-5p and promoted GBC cells proliferation in vitro
MiR-18b-5p had been reported to act as a tumor suppressor and downregulate downstream target mRNAs by binding with their 3’UTRs in many cancers[19, 20]. So to explore the mechanism of miR-18b-5p’s tumor-suppressing function, three miRNA databases were used in the present study. There were 30 candidate genes in the overlapped fraction of three databases as shown in Fig. 7A and Supplementary table 3. Among the 30 candidate genes, only 9 genes (IGF1, TNRC6B, TAOK1, MAP3K1, CTGF, RNF146, SMAD2, HIF1A, ESR1) were involved in cell proliferation and related signal pathways. Furthermore, qRT-PCR determined that only the expression of HIF1A was significantly downregulated by miR-18b-5p mimic in GBC cells (data not shown).
In addition, we analyzed the expression of HIF1A and miR-18b-5p in 20 GBC tissues, a significant inverse correlation was found as shown in Fig. 7B. To further confirm whether HIF1A was the direct target of miR-18b-5p in human GBC cells, we performed luciferase reporter assays and the wildtype or mutant 3’UTRs of HIF1A mRNA was cloned and inserted into a luciferase reporter vector (Fig. 7C). The results showed that the luciferase activity of the wildtype but not mutant 3’UTRs was significantly decreased by miR-18b-5p in both GBC-SD and SGC-996 cells (Fig. 7D). These findings suggested that HIF1A was the downstream target gene of miR-18b-5p.
Next, we designed HIF1A siRNA and transfected it to GBC-SD cell, the efficiency of si-HIF1A was examined by qRT-PCR as shown in Fig. 7E. It was found that knockdown of HIF1A significantly attenuated the proliferation and the cloning ability of GBC-SD cells by CCK-8 and colony formation assays (Fig. 7F and 7G). Flow cytometry analyses demonstrated that downregulation of HIF1A significantly increased G1/G0 phase arrest (Fig. 7H). Moreover, the expression of CDK4 and CyclinD1 was markedly decreased in GBC-SD cells transfected with si-HIF1A (Fig. 7I). These results showed that HIF1A promoted GBC cells proliferation in vitro.
PVT1 regulated GBC cell proliferation via HIF1A
Furthermore, rescue assays were conducted to confirm whether PVT1 regulated GBC cells proliferation via HIF1A. First, we designed pcDNA-HIF1A for ectopic expression in GBC-SD cell line, and qRT-PCR verified the efficiency of pcDNA-HIF1A (Fig. 8A). Then, GBC-SD cells were cotransfected with si-PVT1-1 and pcDNA-HIF1A. CCK-8 and colony formation assays showed that pcDNA-HIF1A partially rescued the attenuated proliferation and cloning ability of GBC-SD cells caused by si-PVT1-1 (Fig. 8B and 8C). Flow cytometry analyses demonstrated that G1/G0 phase arrest induced by si-PVT1-1 could be partially reversed by pcDNA-HIF1A (Fig. 8D). These data suggested that PVT1 executed its function in GBC cells via HIF1A.