DNA primase subunit 1 deteriorated progression of hepatocellular carcinoma by activating AKT/mTOR signaling and UBE2C-mediated P53 ubiquitination

Background DNA primase subunit 1 (PRIM1) has been reported as a novel oncogene in several cancer types. However, its roles in hepatocellular carcinoma (HCC) remain unclear. This study aimed to investigate underlying mechanisms of PRIM1 and identify it as a potential molecular target for HCC. Methods Hub genes were screened between HCC tissues and normal liver tissues in 3 gene expression omnibus (GEO) datasets and the cancer genome atlas (TCGA). The expression features and prognostic value of one of the hub genes PRIM1 were analyzed by bioinformatic analyses and immunohistochemistry. Loss-of-function and gain-of-function studies were used to investigate the regulatory role of PRIM1 in HCC cells. Real-time (RT)-qPCR, western blotting, and ubiquitin immunoprecipitation assays were performed to explore the underlying mechanisms. The xenograft model was employed to detect the roles of PRIM1 in tumor growth in vivo. Finally, the 3D spheroid model was conducted to validate the role of PRIM1 in tumor growth and sorafenib resistance. Results The hub genes of HCC were screened in multiple bioinformatic datasets. PRIM1, as one of the hub genes, was significantly overexpressed in HCC tissues in mRNA and protein levels. In addition, high expression of PRIM1 indicated poor prognosis of HCC patients in TCGA, ICGC, and Nantong cohorts. Overexpression of PRIM1 promoted the proliferation, migration/invasion, and sorafenib resistance of HCC cells, with the decrease in apoptosis and cell cycle arrest. Mechanically, PRIM1 facilitated epithelial-mesenchymal transition (EMT) process and the activity of PI3K/AKT/mTOR signaling of HCC cells. Additionally, PRIM1 could cause the ubiquitination and degradation of P53 by upregulating Ubiquitin Conjugating Enzyme E2 C (UBE2C). Furthermore, knockdown of PRIM1 significantly inhibited the growth of xenograft tumors and HCC cells-derived spheroids with enhanced sorafenib resistance. Conclusion This study implies that PRIM1 may play a key role in the progression of HCC and may serve as a potential target for HCC treatment.


Introduction
Hepatocellular carcinoma (HCC) is one of the leading causes of cancer-related mortality, with ~ 1 million new cases worldwide. Risk factors of HCC include hepatitis B virus (HBV) or hepatitis C virus (HCV) infections, aflatoxin B1 exposure, alcohol abuse, and metabolic syndrome [1]. Surgery is optimal for the HCC patients at early stage. However, 50% of HCC patients were diagnosed at advanced stages, for whom ablation, transarterial chemoembolization are optional strategies [2]. Furthermore, systematic therapy such as multiple-target tyrosine kinase inhibitor sorafenib have shown efficacy in improving the outcomes of HCC patients at late stages. Regorafenib, cabozantinib, and ramucirumab were approved as second-line treatments for advanced HCC [3]. Despite of the remarkable progression in surveillance, diagnosis and management, the overall outcome of HCC patients remains poor due to increasing recurrence and chemoresistance [4]. Therefore, it is warranted to better investigate the biological and molecular processes in HCC and provide novel targets for HCC.
DNA primases synthesize primers for DNA polymerases-mediated DNA replication. DNA primases are also defined as DNA-dependent RNA polymerases. The polymerases complex includes the catalytic subunit POLA1, the regulatory subunit POLA2, as well as the primase complex subunits DNA Primase Subunit 1(PRIM1, 49 kDa) and DNA Primase Subunit 2 (PRIM2, 58 kDa) [5]. PRIM1 mapping at chromosome 12q13 is the smallest subunit within a polymerase α (Pol α) heterotetrameric complex, which plays a crucial role in DNA replication by facilitating catalytic function and synthesizing RNA primers as starting point for the replication of the leading and lagging strand, while PRIM2 had no similar enzymatic activities [6]. Baranovskiy AG et al. manifested the precise crystal structure and overall organization of PRIM1, suggesting the fundamental roles of dramatic conformational changes in its interaction with DNA and RNA, as well as the elongation of RNA synthesis [7]. Given its importance function in DNA replication, PRIM1 deficiency has been correlated with replicationassociated disorders. Parry DA et al. recently described a distinct form of microcephalic dwarfism associated with PRIM1 mutations [8]. Moreover, PRIM1 deficiency was also correlated with primary ovarian insufficiency [9] and type 2 diabetes mellitus [10]. In addition, missense mutation of PRIM1 may cause extensive apoptosis of retinal neurons through activation of the DNA damage checkpoint and tumor suppressor P53, with no alterations in cell proliferation [11]. For its potential roles in oncogenesis, amplifications of PRIM1 were firstly found in osteosarcoma [12]. Recently, emerging evidence has shown that PRIM1 acts as an oncogene that enhances the aggressive behaviors of tumor cells. Lee WH et al. manifested that depletion of PRIM1 inhibited the cell cycle progression and induced DNA damage of the breast cancer cells through inactivating G2/M cell cycle checkpoint [13]. Job et al. demonstrated that PRIM1 inactivation sensitized colorectal cancer cells to ATR and CHK1 inhibitors, suggesting PRIM1 or other primase subunits could be novel targets for individualized tumor therapeutic approaches [14]. Besides, PRIM1 was also identified as a predictor that indicated poor survival of tumor patients. Though recent studies described the abnormal expression features of PRIM1 [15], the mechanisms underlying its oncogenic roles have not been elucidated.
In the current study, we screened hub genes of HCC through integrated bioinformatic analyses. PRIM1 was further selected to comprehensively analyze its expression profiling and clinical values. In addition, we also investigated the roles of PRIM1 in phenotypic manifestations and the underlying molecular mechanisms. This study may provide new insights into the HCC progression and therapeutic targets for HCC treatment.

Clinical samples
Tissue samples of Nantong cohort were collected from 135 HCC patients who underwent surgery between January 2012 and October 2014 at Affiliated Hospital of Nantong University (Nantong, Jiangsu, China). The study was performed in compliance with the Helsinki Declaration and approved by the ethic committee of Affiliated Hospital of Nantong University (TDFY2018-025).

Data preprocessing
Microarray data of GSE25097, GSE121248, GSE1112790 was downloaded from the National Center for Biotech- nology Table 1.

Differential expression analysis
Linear models for microarray data (Limma; http://www. bioco nduct or.org/packa ges/relea se/bioc/html/limma .html) was the R package that analyzed gene expression microarray data, especially the use of linear models for the assessment of differential expression. DESeq2 (version 1.18.1) package in R was used for differential expression analysis of the sequencing data. Significantly differentially expressed genes were accepted as |log2FC|> 1 and adjusted P-value < 0.01.

Functional analysis
Pathway database. Reactome (http://www.react ome. org) is a manually curated open-data resource of human pathways and reactions, an archive of biological processes and a tool for discovering potential functions. Gene sets derived from the Reactome pathway database were downloaded from the MSigDB Collections. Enrichment test. Functional enrichment based on the Reactome and GO terms databases was assessed by hypergeometric test, which was used to identify a priori-defined gene sets that showed statistically significant differences between two given clusters. The test was performed by the cluster-Profiler package. We further corrected the test P-values by the Benjamini-Hochberg and less than 0.05 was considered as statistically significant. The GO enrichment was visualized using the GOplot package.

Survival analysis
The Kaplan-Meier analyses were performed in TCGA-LIHC and ICGC samples. Kaplan-Meier (KM) curve with a log rank test was used for presenting the results of survival analysis. The 'survival' R package (http://cran.rproje ct.org/web/packa ges/survi val/index .html) was subjected to survival analysis.

Gene set enrichment analysis (GSEA)
GSEA 3.0 software (http://www.broad .mit.edu/gsea) was performed to predict the potential functions and mechanisms mediated by PRIM1 in TCGA LIHC and ICGC databases. The reference was obtained from Molecular Signatures Database (MSigDB, http://softw are.broad insti tute.org/gsea/msigd b). HCC cohort was stratified into high and low PRIM1 groups based on the median

Transwell assays
The invasion and migration assay were performed by using the 8-μm Transwell chambers (Corning, Acton, MA, USA). For the invasion assay, 200 μL of MHCC97H or HepG2 cells of each group were plated in the upper chambers pre-coated with Matrigel (BD, CA, USA) mixed in DMEM, while the complete medium was plated in the lower chamber. After incubated for 24 h, the upper chambers were fixed in 4% paraformaldehyde and stained in the crystal violet solution. Then, the stained cells were observed under the microscope. For the migration assay, the same protocol was conducted without the administration of the Matrigel. Five fields were randomly selected to calculate the average number of invasion or migration cells.

Xenograft tumor assay
Four-week-old BALB/c nude mice were obtained from Animal Center of Nantong University (Nantong, China). HCC cells suspended at the density of 5 × 10 6 /100 μL serum-free DMEM were subcutaneously injected into the flanks of the mice. The xenograft tumors were monitored every 3 days for a total of 5 weeks until the sacrifice. The volume was measured as follows: volume = length × width 2 × 0.5. The procedures of this study were approved by the Animal Care and Use Committee of Nantong University.

Spheroid model construction
The 3D spheroid model was constructed as previously described [16].

Immunohistochemistry
Human liver and mouse specimens were fixed in formaldehyde and embedded in paraffin. Sections were dewaxed by xylene and dehydrated by serial diluted ethanol. Then the slides were incubated in 0.3% hydrogen peroxide and blocked with 1% BSA solution. Following that, the slides were incubated in anti-

Immunoprecipitation and protein stability assay
The MHCC97H and HepG2 cell lysates of each group were collected by radioimmunoprecipitation assay buffer (RIPA) for 30 min on ice. Then the lysates were incubated with the primary antibody against P53 (Santa Cruz Biotechnology, USA) overnight at 4 °C. Following that, the samples were incubated with anti-rabbit Ig-IP beads (Rockland Immunochemicals, USA) for 2.5 h at 4 °C. The beads were spun down and rinsed with PBS. Subsequently, the proteins were eluted from the beads by boiling in loading buffer (Bio-Rad) for 5 min and subjected to immunoblotting. For the protein half-life assay, MHCC97H or HepG2 cells in each group were administrated with CHX (100 μg/mL, Sigma, USA) and collected at the indicated time points. IB was used to determine the protein attenuation induced by CHX.

Flowcytometry
The effects of PRIM1 on apoptosis of HCC cells were analyzed by A FITC-Annexin V Apoptosis Detection Kit (BD Biosciences, USA) according to the manufacturer's instructions. Cells of each group were pre-suspended in binding buffer and incubated with Annexin V-FITC for 15 min. Following incubation with Propidium iodide (PI), the samples were detected by using a BD FACS Calibur flow cytometer (Becton-Dickinson, USA). For cell-cycle detection, the cells of each group were incubated with PI solution for 15 min and detected by using a BD FACS Calibur flow cytometer. The results were analyzed by Modfit software.

Reverse transcription quantitative polymerase chain reaction (RT-qPCR)
Total RNA of each group was extracted by TRIzol (Invitrogen, USA) according to the manufacturer's instructions. cDNA was synthesized by using TaKaRa PrimeScript RT regent kit (TaKaRa, Japan). RT-qPCR was conducted by using SYBR on an ABI 7500 qPCR apparatus (Applied Biosystems, Life Technologies, CA, USA).

Western blotting
Protein of each group extracted by RIPA was separated on a sodium dodecyl sulfate (SDS) gel and transferred onto a polyvinylidene difluoride (PVDF) membrane (Bio-Rad, Hercules, CA, USA). Subsequently, the membranes were blocked in 5% BSA for 2 h and incubated in primary antibodies at 4 °C overnight. Then the samples were further incubated in horseradish peroxidase (HPR) -conjugated secondary antibodies at room temperature for 2 h. The membranes were visualized by using enhanced chemiluminescence (ECL) kit (Millipore, MA, USA). GAPDH (Abcam, Cambridge, UK) was served as an internal control for western blotting.

Statistical analysis
Data are presented as the mean ± SEM. Statistical analysis was performed using SPSS 19.0 and GraphPad Prism 7.0 (CA, USA). The χ 2 test and Student's t-test were performed to evaluate the differences between two groups, while one-way or two-way ANOVA was used for multiple group comparisons. Kaplan-Meier curves with log-rank test was performed to evaluate the prognostic value of various variables. All experiments were repeated at least three times. P value < 0.05 was considered statistically significant.

Screening hub genes of HCC through integrated bioinformatic analyses
To identify the differentially expressed genes between HCC and normal samples, we performed a comprehensive analysis of the gene expression data from public domain GEO. Three datasets, i.e., GSE25097, GSE121248, and GSE1112790, with both tumor and normal samples were collected from the GEO database. By comparing tumor and normal samples, we identified, respectively 676, 332 and 1021 significantly up-regulated genes in each dataset, which resulted in 501 genes that were consistently overexpressed in three datasets. Hierarchical clustering is performed using the 501 genes across these datasets (Fig. 1a). Besides that, we also downloaded the expression data of HCC patients from TCGA. 7744 genes were shown to be significantly up-regulated in tumor. Finally, we identified 172 up-regulated genes across four datasets, as indicated in the Venn plot (Fig. 1b). Next, we examined the enriched functions associated with the 172 up-regulated genes in the Reactome pathway database, including activation of the pre-replicative complex, G1/S transition, and G2/M checkpoints (Fig. 1c). To gain more insights into their biological significance, we assessed the enrichment of gene ontology (GO) terms for biological processes (BPs), cellular components (CCs) and molecular functions (MFs). The significant enriched GO terms indicated the cell cycle process with a particular involvement of the G1/S phase transition as well as nuclear division (Fig. 1d). As expected, such annotations indicated the cell cycle processes that were also supported by the Reactome pathway analysis. Further investigation of the enriched Reactome pathways revealed the most frequently involved genes. As shown in Fig. 1e, PRIM1 was one of the top genes that were associated with most of the enriched pathways. Meanwhile, the Logrank test indicated that PRIM1 might have potentially prognostic significance (P = 0.0006) for HCC patients (Fig. 1f ). Then we focused on investigating the potential roles and mechanisms of PRIM1 in HCC.

Expression and prognostic value of PRIM1 identified in bioinformatic HCC cohorts
Firstly, we explored the expression features of PRIM1 in HCC by comprehensively analyzing multiple bioinformatic datasets. As shown in Fig. 2a-d, PRIM1 was overexpressed in HCC samples with advanced stages or grades, P53 mutation, and metastasis. Similarly, PRIM1 was significantly overexpressed in HCC samples of ICGC cohort (Fig. 2e). In addition, PRIM1 presented higher expression in advanced HCC cases than that of early HCC cases or pre-HCC cases (Fig. 2f-h). Moreover, overexpression of PRIM1 was also observed in HCC cases and HCC cells with sorafenib resistance in GSE63989, GSE109211, and GSE62813, respectively (Fig. 2i-k). Furthermore, we analyzed the prognostic value of PRIM1 in TCGA and ICGC datasets (Fig. 2l, m). High expression of PRIM1 indicated poor overall survival (OS) and recurrence free survival (RFS) of HCC patients in entire cohort or stratified sub-groups. Consistently, the univariate and multivariate analyses indicated that PRIM1 was an independent prognostic marker for OS and RFS of HCC patients (Additional file 1: Figure S1A-D).

Expression features and prognostic value of PRIM1 in Nantong HCC cohort
Next, we conducted immunohistochemistry to detect the protein expression of PRIM1 in 135 self-paired tissues in Nantong HCC cohort. As shown in Fig. 3a, higher staining intensity of PRIM1 was observed in HCC cases compared with para-cancerous tissues. In addition, the expression of PRIM1 was dynamically elevated with HCC staging. Then we analyzed the correlation of PRIM1 with the clinical characteristics of the 135 HCC patients. High expression of PRIM1 was correlated with tumor size, Edmondson-Steiner grade, ascites, and advanced TNM stage (Table 2). Then, we performed the Kaplan-Meier analyses to discover the prognostic value of PRIM1 (Fig. 3b). The result demonstrated that patients with high expression of PRIM1 had shorter OS and RFS compared with cases with low PRIM1 expression. Interestingly, the stratification analysis suggested that overexpression of PRIM1 also indicated shorter survival in patients with early stages and advanced stages. In accordance, the univariate and multivariate analyses showed that PRIM1 was an independent predictor for the OS and RFS of HCC patents in Nantong cohort (Tables 3, 4).

PRIM1 enhanced the aggressive behaviors and facilitated EMT phenotypes
Given the expression features and clinical significance of PRIM1, we further investigated the effects of PRIM1 on the biological behaviors of the HCC cells. As shown in Fig. 4a, differential expression of PRIM1 was observed in HCC cell lines at protein and mRNA levels. Then we chose MHCC97H and HepG2 to conduct Loss-offunction and gain-of-function studies, respectively. As elucidated in Fig. 4b, c, knockdown of PRIM1 significantly inhibited the colony formation, proliferation of MHCC97H cells, while ectopic expression of PRIM1 promoted colony formation and cell proliferation of HepG2 cells. Furthermore, PRIM1 could enhance the resistance of HCC cells against sorafenib (Fig. 4d). In addition, silencing or overexpressing PRIM1 increased or decreased the G2/M arrest and apoptosis of MHCC97H or HepG2 cells, respectively (Fig. 4e, f ). Remarkably, knockdown or overexpression of PRIM1 significantly enhanced or reduced the migration and invasion capacity of HCC cells (Fig. 4g). Consistently, PRIM1 could enhance the expression of Vimentin and N-cadherin and downregulate the expression of E-cadherin, suggesting its roles in promoting epithelial-mesenchymal transition (EMT) of HCC cells (Fig. 4h). Thus, PRIM1 might facilitate the aggressive behaviors of HCC cells.

PRIM1 activated the AKT/MTOR signaling of HCC cells
To further discover the underlying mechanisms, we conducted GSEA and simultaneously in the TCGA cohort and ICGC cohort (Fig. 5a, b). Similar with ORA (Additional file 2: Figure S2), PRIM1 was implicated in various tumor-related biological processes, including cell cycle, DNA replication, DNA repair, and autophagy. In addition, PRIM1 might regulate the activity of mTORC1 signaling and PI3K/AKT/mTOR signaling, which were previously defined as canonical metastasis and proliferation related pathways (Fig. 5c, d). Consistently, depletion of PRIM1 inhibited the expression of p-AKT, p-GSK-3β, and p-mTOR of MHCC97H cells (Fig. 5e). Ectopic expression of PRIM1 enhanced the expression of these markers, while AKT inhibitor LY294002 or mTOR inhibitor ink128 significantly abrogated the activation of the PI3K/AKT/mTOR signaling (Fig. 5f, g). Moreover, LY294002 or ink128 reversed the enhancement of EMT markers induced by exogenous PRIM1 (Fig. 5h). Thus, PRIM1 might promote the aggressive behaviors of HCC cells by activating PI3K/AKT/mTOR signaling.

PRIM1 promoted P53 ubiquitination and degradation
P53 is a vital tumor-suppressive signaling, for which ubiquitination is a common post-transcriptional modification. GSEA indicated that PRIM1 might be involved in regulating P53 signaling and ubiquitin mediated proteolysis (Fig. 6a). Thus, we further explored the regulatory  (Fig. 6b). In contrast, PRIM1 overexpression significantly inactivated P53 signaling of HepG2 cells, which was restored by protease inhibitor boriezomib (Fig. 6c). Moreover, silencing PRIM1 alleviated the degradation of P53 induced by CHX. Conversely, exogenous PRIM1 accelerated the degradation of P53 (Fig. 6d, e). Furthermore, we discovered the correlation of PRIM1 mediated P53 degradation with ubiquitination. As proposed, the weakened ubiquitination of P53 from the P53-lysates immunocomplex was detected in Kd-PRIM1 transfected MHCC97H cells compared with NC group following elevated P53 protein expression (Fig. 6f ). Conversely, the enhanced ubiquitination of p53 was observed in the p53-lysates immunocomplex in OE-PRIM1 transfected HepG2 cells compared with NC groups following decreased P53 expression (Fig. 6g). Taken together, PRIM1 promoted P53 ubiquitination and degradation of HCC cells.

PRIM1 facilitated P53 ubiquitination by upregulating UBE2C
Next, we further investigated the potential mechanisms of PRIM1-mediated P53 ubiquitination by screening the genes enriched in the item "ubiquitin mediated proteolysis". P53 was previously defined as a substrate of one of the top genes Ubiquitin Conjugating Enzyme E2 C (UBE2C) [17]. Firstly, we investigated the correlation of UBE2C with PRIM1 and expression features of UBE2C   Fig. 7a, PRIM1 mRNA level was positively correlated with UBE2C in both of TCGA and ICGC datasets. Similar with PRIM1, overexpression of UBE2C was observed in HCC tissues with P53 mutation or at advanced grades (Additional file 3: Figure S3A-D). Moreover, the expression level of UBE2C was higher in high dysplastic liver nodules and sorafenib resistant HepG2 cells (Additional file 3: Figure S3E and F). Additionally, high expression of UBE2C indicated poor survival of HCC patients in TCGA (Additional file 4: Figure S4A) and ICGC datasets (Additional file 4: Figure S4B). Then we conducted functional validations regarding the effects of UBE2C. Silencing UBE2C significantly repressed the proliferation and colony formation of MHCC97H cells; In contrast, overexpression of UBE2C enhanced the aggressive behaviors of HepG2 cells (Fig. 7c, d). Moreover, 3D spheroid model indicated that PRIM1 facilitated tumor growth and chemoresistance of sorafenib (Fig. 7e, f; Additional file 5: Figure S5). We hypothesized that UBE2C might serve as an important oncogene that participated in regulating PRIM1mediated P53 ubiquitination and related phenotypes of HCC. As shown in Fig. 7g, h, knockdown of PRIM1 inhibited UBE2C expression of MHCC97H cells in both of the mRNA and protein levels, while overexpressing PRIM1 enhanced the expression of UBE2C of HepG2 cells. Furthermore, overexpression or depletion of UBE2C abolished the alterations of P53 signaling induced by PRIM1 in HCC cells (Fig. 7i). Consistently, exogenous UBE2C restored proliferation, colony formation, and sorafenib resistance of MHCC97H cells inhibited by PRIM1 repression. In addition, knockdown of UBE2C abrogated the enhancement of these aggressive phenotypes following PRIM1 overexpression (Fig. 7j-l). Thus, PRIM1 might promote the aggressive behaviors via UBE2C-mediated P53 ubiquitination.

PRIM1 enhanced tumor growth in the xenograft model and 3D spheroid model
Ultimately, we validated the roles of PRIM1 in vivo and 3D models. As shown in Fig. 8a-c, the mean volume and weight of PRIM1-silenced MHCC97H-derived xenograft tumors were significantly less than these of NC group. In addition, the H&E staining showed that knockdown of PRIM1 alleviated the typically histological features of xenografts (Fig. 8d). Further immunohistochemical staining demonstrated that silencing PRIM1 significantly decreased the expression of UBE2C, Ki67, N-cadherin, P-AKT, and P-mTOR of the xenograft tumor tissues, with the increase of E-cadherin and P53 expression (Fig. 8e). Next, we constructed 3D spheroid model to further investigate the role of PRIM1. As elucidated in Fig. 8f, depletion of PRIM1 impeded the growth of the 3D spheroids derived from MHCC97H cells in day 1-5, while ectopic expression of UBE2C restored the growth of the spheroids. Besides, PRIM1 silenced MHCC97H-derived spheroids presented higher sorafenib sensitivity than the NC group. Further exogenous UBE2C enhanced the resistance of the spheroids against sorafenib (Fig. 8g). In consistent, silencing UBE2C abrogated the enhancement  c Following transfection of Kd-PRIM1 or OE-PRIM1 plasmid, the proliferation of HCC cells was detected by CCK-8 assay. d The sorafenib sensitivity was detected in PRIM1-silenced MHCC97H or PRIM1-overexpressed HepG2 cells compared with NC group. e, f Following transfection of Kd-PRIM1 or OE-PRIM1 plasmid, the cell cycle and apoptosis were detected by flowcytometry. g The migration and invasion of MHCC97H or HepG2 cells in each group were detected by transwell assay. h The expression of EMT markers were examined by western blotting with PRIM1 overexpression or downregulation. **P < 0.01; *P < 0.05 of growth and sorafenib resistance induced by PRIM1 overexpression in HepG2-derived spheroids (Additional file 6: Figure S6A and B). These evidences suggested that PRIM1 might contributed to tumor growth and sorafenib resistance.

Discussion
Various hub genes and signaling pathways maintain the intracellular homeostatic status. The abnormal expression of oncogenes is critical to hepatocarcinogenesis and HCC progression, which may be key targets with the potential to improve HCC treatment [18]. A better understanding of hub genes that involve in HCC carcinogenesis will contribute to identifying potentially therapeutic targets [19]. Firstly, we identified DEGs between HCC and normal liver tissues in GEO and TCGA datasets, which were implicated in multiple cell cycle and DNA replication related terms. Of them, PRIM1 was one of the top genes associated with most of the enriched pathways and predicted poor prognostic of HCC patients. In addition, the roles of PRIM1 in phenotypic manifestations of HCC and the underlying molecular mechanisms have not been well elucidated. Therefore, we focused on the investigation of PRIM1 from the aspects of expression profiling, clinical significance, effects on aggressive behaviors, and molecular mechanisms in HCC. Although emerging studies have indicated the potential roles of PRIM1 in the malignant behaviors of breast cancer [13], the expression features and functions of PRIM1 in HCC remain unclear. In this study, elevated mRNA and protein expression of PRIM1 were determined in HCC tissues from various bioinformatic datasets and tissue microarray. Overexpression of PRIM1 was correlated with aggressive clinical characteristics and poor survival of HCC patients. In the current study, the prognostic value of PRIM1 was validated in TCGA, ICGC and local Nantong cohorts, suggesting that it might be an attractive marker for HCC. Apart from its clinical implications, loss-of-function and gain-of-function studies suggested the contributing roles of PRIM1 in the malignant  phenotypes including proliferation, colony formation, invasion and sorafenib resistance, and EMT processes. To further explore the mechanisms, we conducted GSEA based on the TCGA and ICGC datasets. Apart from the enrichment in the biological processes such as cell cycle and DNA proliferation, the GSEA suggested that PRIM1 might be implicated in AKT/GSK-3β and mTOR signaling. Both in vitro and in vivo studies have demonstrated that PI3K/AKT/mTOR signaling plays key roles in regulating HCC initiation and progression [20]. It regulates proliferation, cancer stem cell properties, and invasion and migration of HCC via phosphorylating its downstream effectors [21]. As expected, PRIM1 could regulated the activity of PI3K/AKT/mTOR signaling. The inhibitors of AKT or mTOR reversed the modulation induced by PRIM1 overexpression. Moreover, the inhibitors also restored the activation of EMT markers induced by exogenous PRIM1. Therefore, it was supposed that PRIM1 positively regulated the activity of PI3K/AKT/mTOR signaling and its mediated malignant behaviors. A recent work using CRISPR-based-genomewide screening determined PRIM1 as a robust regulator of cAMP-PKA signaling. Further validation also confirmed that deletion of PRIM1 significantly blunted the activation of cAMP signaling via multi-stimuli [22]. As an important second messenger, cAMP is competent in multiple signaling transduction and pathway linking. For example, cAMP-PKA axis has been reported to facilitate PI3K-AKT and subsequent mTORC activation [23]. Hence, PRIM1 may regulate PI3K-AKT-mTOR partially through ensuring the activity of cAMP-PKA cascade. However, it is the speculation of the PRIM1-mediated mechanism, and the explicit regulatory network should be elucidated by further investigations.
P53 signaling is the most intensively studied tumorsuppressive hallmark that modulates multiple biological processes, such as proliferation, senescence, and apoptosis [24]. Gao et al. revealed that P53 mutation occurs in almost 60% HBV-related HCC patients, suggesting that mutation of the P53 gene abrogates the P53 tumorsuppressor function to deteriorate hepatocarcinogenesis [25]. Thus, it is an attractive strategy to reactivate P53 signaling in HCC patients with P53 mutation. According to the bioinformatic analyses in TCGA and ICGC datasets, PRIM1 might be implicated in the regulation of P53 signaling. Further molecular validation found that PRIM1 downregulated the expression of P53 and its downstream genes. Then we further discovered the molecular mechanisms of PRIM1 mediated P53 repression. Based on the GSEA, PRIM1 might participate in the ubiquitin mediated proteolysis. The protease inhibitor boriezomib reversed the inhibition of P53 induced by PRIM1 overexpression. Thus, we hypothesized that PRIM1 inactivated P53 via ubiquitination-mediated degradation. With the treatment of CHX, PRIM1 overexpression enhanced the attenuation of P53 protein expression of MHCC97H cells, while silencing PRIM1 alleviated the P53 degradation induced by CHX. PRIM1 silencing or overexpression decreased or enhanced p53 ubiquitination with the increase or decrease of p53 protein expression, respectively. It was consistent with our hypothesis that PRIM1 de-stabilized P53 in a ubiquitination-dependent manner.
Ubiquitination is a crucial post-translational modification for protein stability. By searching the enriched genes in GSEA analysis, UBE2C was selected for further validation due to a previous study that suggested P53 protein as its ubiquitination substrate [17]. Consistently, UBE2C, as a member of ubiquitin-conjugating enzyme (E2) family, has been reported to be abnormally expressed in HCC and involved in its initiation and progression [26,27]. Similar with the expression features of PRIM1, UBE2C was overexpressed in HCC tissues at advanced grades or sorafenib-resistant HCC cells. In both of the TCGA datasets and ICGC datasets, PRIM1 presented positive correlation with UBE2C in HCC samples. Overexpression or knockdown of UBE2C abrogated the alteration of P53 expression induced by PRIM1. In addition, silencing UBE2C inhibited the proliferation, colony formation, tumor spheroid growth, and sorafenib resistance induced by exogenous PRIM1. Therefore, we drew the hypothesis that PRIM1 promoted these aggressive phenotypes partially through UBE2C-mediated P53 degradation. Interestingly, Luo YD et al. recently noted that P53 haploinsufficiency and mTOR signaling activation identified a subset of aggressive HCC patients [28]. In the current study, we found that depletion of PRIM1 reactivated P53 signaling and decreased the activity of AKT/mTOR signaling. Based on such observation, targeting PRIM1 may provide a new strategy for treating aggressive subtype of HCC.

Conclusion
In conclusion, this study identified PRIM1 as a pivotal regulator that deteriorated the aggressive phenotypes of HCC cells by modulating AKT/mTOR signaling and UBE2C-mediated P53 inactivation, suggesting that PRIM1 might serve as a therapeutic target in the prevention and treatment of HCC.