Abstract

   Microvascular invasion (MVI) has been widely valued in the field of
   liver surgery because MVI positivity indicates poor prognosis in
   hepatocellular carcinoma (HCC) patients. However, the potential
   molecular mechanism underlying the poor prognosis of MVI‐positive HCC
   patients is unclear. Therefore, this study focused on identifying the
   key genes leading to poor prognosis in patients with a high degree of
   malignancy of HCC by examining the molecular signaling pathways in
   MVI‐positive HCC patients. Through RNA sequencing, TOX high mobility
   group box family member 3 (TOX3) was demonstrated to be significantly
   highly expressed in MVI‐positive HCC tissues, which was associated with
   poor prognosis. The results of in vivo and in vitro showed that TOX3
   can promote the oncogenesis and development of HCC by targeting key
   molecules of the MAPK and EMT signaling pathways. The IP‐MS results
   indicated that proteasome degradation of TOX3 in HCC cells is
   potentially mediated by a tripartite motif containing 56 (TRIM56, an E3
   ligase) in HCC cells. Inhibiting TRIM56 enhances TOX3 protein levels.
   Overall, our study identified TOX3 as a key gene in the MAPK and EMT
   signaling pathways in HCC, and its overexpression confers significant
   proliferation and invasiveness to tumor cells.

   Keywords: growth and metastasis of liver cancer, Hepatocellular
   carcinoma, insulin‐like growth factor binding protein 3, microvascular
   invasion, TOX high mobility group box family member 3, tripartite motif
   containing 56
     __________________________________________________________________

   The depletion of the tripartite motif containing 56 in hepatocellular
   carcinoma (HCC) may potentiate the accumulation of TOX high mobility
   group box family member 3 (TOX3), and further regulate the classical
   signaling pathways of MAPK and EMT by increasing the expression level
   of insulin‐like growth factor binding protein 3, eventually promoting
   the growth and metastasis of HCCs.

   graphic file with name MCO2-5-e510-g006.jpg

1. INTRODUCTION

   Hepatocellular carcinoma (HCC) is the predominant primary hepatic
   malignant tumor with high postoperative recurrence rates and poor
   prognosis.[44] ^1 Recently, microvascular invasion (MVI) has been
   widely valued in the field of liver surgery, because it occupies an
   important position among many adverse factors.[45] ^2 , [46]^3 , [47]^4
   , [48]^5 Previous studies have indicated that MVI serves as a precursor
   to both intrahepatic and extrahepatic metastasis of HCC, representing a
   significant manifestation of highly malignant HCC cells and signifying
   their heightened invasiveness. To date, the molecular mechanism
   underlying the poor prognosis of MVI‐positive HCC patients has not been
   determined. The recognition of key cancer‐promoting molecules could
   contribute to understanding the oncogenesis and development mechanism
   of HCC and identifying potential therapeutic targets.

   The protein encoded by TOX high mobility group box family member 3
   (TOX3) contains a high mobility group box.[49] ^6 TOX3 may participate
   in the bending and releasing of DNA, as well as altering chromatin
   structure to promote transcription by binding to the eukaryotic cell
   promoters.[50] ^7 , [51]^8 , [52]^9 , [53]^10 It can be regulated by
   binding phosphorylation proteins and can also play a role as a
   transcriptional costimulatory factor.[54] ^8 However, the understanding
   of the role of TOX3 in tumors and its related mechanisms is in its
   infancy. According to the previously published literature, TOX3 can
   change the methylation status of its promoter and reduce the expression
   level of BRCA1, thus potentiating the growth and metastasis of breast
   cancer.[55] ^11 Additionally, TOX3 can mediate the expression of the
   apoptosis‐related molecule RhoB and enhance the proliferation of
   colorectal tumors.[56] ^12 Moreover, TOX3 can also strengthen the
   metastatic ability of lung cancer cells through regulating EMT and
   Hippo‐related signaling pathways.[57] ^13

   The tripartite motif containing 56 (TRIM56, an E3 ligase) is related to
   the degree of malignancy of liver, lung, hematological, and ovarian
   cancers. This motif inhibits the occurrence and metastasis of malignant
   tumors by targeting the degradation of certain carcinogenic
   proteins.[58] ^14 , [59]^15 , [60]^16 , [61]^17 Yang et al.
   demonstrated that the TRIM56 expression level is down‐regulated in HCC
   tissues versus adjacent normal liver tissues, and the degree of
   downregulation of TRIM56 expression is closely related to tumor stage
   and prognosis. Mechanistically, TRIM56 can inhibit the oncogenesis and
   development of HCC by regulating the targeted degradation of key
   molecules in the Wnt signaling pathway.[62] ^14 Yan et al. have shown
   that the binding of NEAT1 to the Wnt component DVL2 results in the
   interaction of the protein with TRIM56, which enhances the degradation
   of the DVL2 through the TRIM56‐mediated ubiquitin‐proteasome, thereby
   inhibiting the Wnt signaling pathway.[63] ^15 Lu et al. indicated that
   TRIM56 can inhibit the invasion and development of lung cancer.[64] ^16
   Zhao et al. demonstrated that the downregulation of TRIM56 expression
   dramatically strengthens the invasive ability of ovarian malignant
   tumor cells, while overexpression of TRIM56 can reduce the invasive
   ability. Mechanically, TRIM56 can inhibit the invasive ability of
   malignant ovarian tumor cells by targeting Vimentin for
   degradation.[65] ^17

   In this study, high‐throughput RNA sequencing of HCC tissues revealed
   that TOX3 in MVI‐positive HCC tissues was highly expressed, thereby it
   was regarded as the focus molecule. Abnormal expression of TOX3 could
   strengthen oncogenesis and metastasis in HCC both in vitro and in vivo
   through the MAPK and EMT pathways, whereas proteasome degradation of
   TOX3 is potentially mediated by the E3 ligase TRIM56. This study
   demonstrated that TOX3 could represent a promising therapeutic target
   for addressing the oncogenesis and development of HCC. Thus, this study
   contributes to the understanding of the mechanisms underlying the
   oncogenesis and development of HCC and aids in the identification of
   potential therapeutic targets.

2. RESULTS

2.1. TOX3 in MVI‐positive HCC tissues was highly expressed, and negative
prognosis in HCC patients was linked to TOX3 upregulation

   To explore the specific molecular mechanism underlying the poor
   prognosis caused by HCC with MVI, we collected 6 patients’ HCC tissues
   and paired adjacent normal liver tissues (three HCCs with MVI and three
   HCCs without MVI; Table [66]S1). High‐throughput RNA sequencing was
   performed on a total of 12 tissue samples to identify and validate the
   target genes (Figure [67]1A). Based on the sequencing results above,
   the differentially expressed genes (DEGs) were further screened using
   specific statistical analysis conditions: the difference multiple fold
   change > 2 or < 0.5 and corrected p‐value (adj p) < 0.05. Figure [68]1B
   shows the screening process. The expression abundance of TOX3 was
   consistent across adjacent normal tissue, HCC without MVI, and HCC with
   MVI tissues and showed a gradual increase among the three groups (Table
   [69]S1). Additionally, the expression abundance of TOX3 was high in
   tissue samples, and it was found that there were substantial
   differences in TOX3 among the groups through the verification of the
   validation cohorts (Table [70]S1 and Figure [71]1C,D). Thus, TOX3 was
   ultimately selected as the focus of this study.

FIGURE 1.

   FIGURE 1
   [72]Open in a new tab

   TOX high mobility group box family member 3 (TOX3) is upregulated in
   hepatocellular carcinoma (HCC) tissues with microvascular invasion
   (MVI) and associated with poor prognosis in HCC patients. (A) Schematic
   model presenting the process to identify and verify the target genes.
   (B) Flow diagram for selecting candidate genes through the intersection
   of tumor versus normal tissues and MVI versus NMVI HCC sequencing
   results. (C) The expression of TOX3 in HCC tissues and adjacent normal
   tissues from 72 patients with HCC. (D) The above samples were divided
   into the HCC with MVI group (n = 32) and HCC without MVI group
   (n = 40). The difference in TOX3 expression between the two groups was
   analyzed. (E) The OS of HCC patients with high TOX3 expression versus
   low expression of HCC by TCGA database. (F) The OS of HCC patients with
   high TOX3 expression versus low TOX3 expression in 72 patients above.
   *p < 0.05, **p < 0.01, ***p < 0.001.

   The present study examined TOX3 expression in tumor and adjacent normal
   liver tissues from 72 HCC patients who underwent selected hepatectomy
   at our center. The validation results demonstrated that the TOX3
   expression level was markedly upregulated in cancer tissues versus
   adjacent normal liver tissues (Figure [73]1C). According to the
   pathological results of the above patients, we further divided them
   into an MVI‐positive HCC group (n = 32) and an MVI‐negative HCC group
   (n = 40). Subgroup analysis further demonstrated that TOX3 in
   MVI‐positive HCC tissues was further highly expressed compared with
   that in the MVI‐negative HCC tissues (Table [74]S2 and Figure [75]1D).
   Based on the statistical analysis of the TCGA database, patients with
   high TOX3 expression had a worse prognosis (Figure [76]1E).
   Furthermore, real‐time quantitative polymerase chain reaction (RT‐qPCR)
   detection of HCC tissue was used to detect HCC tissue and assess the
   relationship between the TOX3 expression level and patient survival
   rate. Patient groups were divided into two groups based on the median
   TOX3 expression. Analysis of Kaplan‒Meier survival curves revealed a
   significantly poorer long‐term prognosis among HCC patients with high
   TOX3 versus low TOX3 expression (Figure [77]1F). Additionally, we
   further performed enrichment analysis (GO_BP gene terms) for the DEGs
   between HCC tissues and paired adjacent normal liver tissues (Figure
   [78]S1A), and between the HCC patients in the MVI‐positive group and
   MVI‐negative group (Figure [79]S1B), respectively. The results of
   enrichment analysis above mainly focused on cell proliferation,
   adhesion, and migration (Figure [80]S1A,B).

2.2. TOX3 strengthens oncogenesis and development of HCC in vitro and in vivo

   TOX3 knockout in an HCC cell line (HCCLM3) was performed via the use of
   CRISPR/Cas9 (sgRNA1 and sgRNA2) to identify the role of TOX3 in the
   oncogenesis and development of HCC. The proliferation of HCCLM3 cells
   was suppressed after TOX3 gene knockout compared with that in the
   negative control group (the NC group) (Figure [81]2A,B,E,F). Moreover,
   a cell competition test between TOX3 knockout cells and control cells
   was performed, which showed that TOX3 knockout resulted in a decrease
   in the survival advantage of HCC cells (Figure [82]2I,J). Moreover,
   TOX3 knockout resulted in a substantial decrease in the migratory
   capacity of HCCLM3 cells. (Figure [83]2C,D,G,H). The overexpression of
   TOX3 in HCCLM3 cells induced by lentivirus transduction significantly
   enhanced the proliferation and migratory capacity of HCC cells
   (Figure [84]2K–N). To evaluate the antitumor effect of inhibiting Tox3
   expression compared with that of classical targeted drugs, we used
   lenvatinib, which is the classical therapeutic target drug for HCC. The
   results indicated that, compared with that of lenvatinib, the
   inhibitory effect of TOX3 knockout on HCCLM3 cells was not inferior to
   the effect of lenvatinib on the IC[50] (50 µM) against HCCLM3 (Figure
   [85]S1C). To evaluate the degree to which TOX3 knockdown inhibited
   normal liver cell proliferation after TOX3 knockdown, we used shRNA to
   knock down TOX3 in HCCLM3 cells and normal liver cells (THLE‐2 cells).
   The results demonstrated that after knocking down TOX3, the
   proliferation capacity of HCCLM3 cells substantially decreased, but the
   proliferation ability of THLE‐2 cells was relatively unaffected (Figure
   [86]S1D).

FIGURE 2.

   FIGURE 2
   [87]Open in a new tab

   TOX high mobility group box family member 3 (TOX3) promotes the
   proliferation and migration of HCCLM3 cells in vitro. (A) and (E) After
   the TOX3 gene of HCCLM3 cells was knocked out by sgRNA1 and sgRNA2,
   respectively, the proliferation activity of the HCCLM3 cells was
   detected with a CCK‐8 kit. (B, F) The TOX3 gene of HCCLM3 cells was
   knocked out by sgRNA1 and sgRNA2, respectively, and the ability of
   colony formation was detected. (K, L) After overexpression of the TOX3
   gene in HCCLM3, the proliferation capacity of cells was detected with a
   CCK‐8 kit, and the ability of colony formation was detected,
   respectively. (I, J) Coculturing WT cells (NC, no fluorescence) and
   HCCLM3 cells (mCherry, red fluorescence) of KO‐TOX3‐sgRNA1 and
   KO‐TOX3‐sgRNA2 respectively. The percentage of fluorescence positive
   rate was detected by flow cytometry. (C, G) After the TOX3 gene of
   HCCLM3 cells was knocked out by sgRNA1, Wound healing assay, and vitro
   cell migration assay were performed to detect the migratory capacity of
   the cells. (D, H) After the TOX3 gene of HCCLM3 was knocked out by
   sgRNA2, a wound healing assay, and vitro cell migration assay were
   performed to detect the migratory capacity of the cells. (M, N) After
   overexpression of the TOX3 gene in HCCLM3, wound healing assay and
   vitro cell migration assay were performed to detect the migratory
   capacity of the cells. *p < 0.05, **p < 0.01, ***p < 0.001.

   To explore whether the knockout or overexpression of the TOX3 gene
   affects the proliferative capacity of HCC cells in vivo, HCCLM3 cells
   (including the TOX3 knockout group, TOX3 overexpression group, and NC
   group) were used to establish the subcutaneous tumor model in nude
   mice. The proliferative capacity of subcutaneous tumors in nude mice
   substantially decreased after knockout of the TOX3 gene (KO‐TOX3)
   (Figure [88]3A), and the size and weight of the subcutaneous tumors
   substantially decreased (Figure [89]3B,C). Additionally, the
   proliferation capacity of the nude mice in the subcutaneous tumor model
   was substantially enhanced after TOX3 was overexpressed (OE‐TOX3)
   (Figure [90]3A), and the size and weight of the subcutaneous tumors
   increased significantly after TOX3 was overexpressed (Figure [91]3B,C).
   The lung metastasis model of HCC in nude mice was established by using
   HCCLM3 cells, and the results of small animal live imaging system
   chemiluminescence showed that the fluorescence signal in the lung area
   was substantially decreased after the TOX3 knockout (Figure [92]3D),
   and the lung metastases after the TOX3 knockout were also decreased
   (Figure [93]3E). The liver orthotopic implantation model of nude mice
   was established by using HCCLM3 cells. The results of small animal live
   imaging system chemiluminescence showed that the fluorescence signal
   was substantially decreased in the abdominal area of the nude mouse
   model after the TOX3 knockout (Figure [94]3F), and H&E staining of
   liver in situ tumor sections showed that the number of hepatic in situ
   tumors was also decreased after TOX3 knockout (Figure [95]3G).

FIGURE 3.

   FIGURE 3
   [96]Open in a new tab

   TOX high mobility group box family member 3 (TOX3) promotes the
   proliferation and migration ability of HCCLM3 cells in vivo. (A) The
   proliferative capacity of subcutaneous tumors in nude mice in the
   TOX3‐overexpression group, TOX3‐knockout group, and control group. (B)
   After the establishment of subcutaneous tumors, the tumor size of each
   group was measured weekly, and the tumor volume was calculated. (C) The
   subcutaneous tumor weights of each group. (D) For the establishment of
   the lung metastasis model, a small animal live imaging system
   chemiluminescence was performed to compare the fluorescence signals in
   the lung area of nude mice in the KO‐TOX3 group and NC group. (E) Lung
   metastasis was stained with H&E staining. (F) For the establishment of
   the liver orthotopic implantation model, a small animal live imaging
   system chemiluminescence was performed to compare the fluorescence
   signals in the abdominal area of nude mice in the KO‐TOX3 group and NC
   group. (G) Liver in situ tumors were stained with H&E staining.
   *p < 0.05, **p < 0.01, ***p < 0.001.

2.3. TOX3 regulates HCC oncogenesis and development by targeting insulin‐like
growth factor binding protein 3

   To clarify the mechanism by which the TOX3 gene regulates the
   oncogenesis and metastasis capacity in HCC, HCCLM3 cell lines in the
   TOX3 knockout group, overexpression, and NC groups were subjected to
   high‐throughput RNA sequencing. The differential expression of genes in
   the three groups was consistent in each group, and there were
   significant differences among the three groups (Figure [97]4A). GSEA
   revealed that the HALLMARK_KRAS_SIGNALING_UP and
   HALLMARK_KRAS_SIGNALING_DN pathways were downregulated in the TOX3
   knockout and overexpression groups, respectively (Figure [98]4B,C).
   Pathway enrichment analysis revealed that the regulation of TOX3 could
   significantly affect signaling pathways related to tumor growth and
   metastasis, such as the MAPK pathway and the EMT (Figure [99]4D,E). To
   further verify whether overexpression of TOX3 can directly upregulate
   the MAPK signaling pathway, we performed an enrichment analysis of
   GO_BP gene terms on genes that were statistically significant
   upregulation in the TOX3‐OE group. The MAPK pathway was significantly
   enriched, indicating that overexpression of TOX3 can upregulate the
   MAPK signaling pathway (Figure [100]S1E). To identify the downstream
   target genes of TOX3, we screened for genes that exhibited significant
   upregulation in the KO‐TOX3 group while displaying significant
   down‐regulation in the OE‐TOX3 group; conversely, genes that were
   significantly downregulated in the KO‐TOX3 group but significantly
   upregulated in the OE‐TOX3 group were identified. This screening
   process yielded a total of 16 genes. Additionally, to pinpoint robust
   target genes, we employed the robust rank aggregation (RRA) algorithm
   to integrate the two sets of DEGs. Based on the RRA score, our analysis
   suggests that insulin‐like growth factor binding protein 3 (IGFBP3) may
   be the downstream target gene of TOX3 (Figure [101]4F). The sequencing
   results demonstrated that the level of IGFBP3 was 0 after TOX3
   knockout, and the level of IGFBP3 increased after TOX3 was
   overexpressed (Figure [102]S1F). This result was confirmed by RT‒qPCR
   (Figure [103]S1G). Additionally, a Violin plot revealed that with the
   increase in IGFBP3 expression, the tumor stage was later
   (Figure [104]4G). TCGA analysis demonstrated that IGFBP3 upregulation
   dramatically impacted HCC patient survival outcomes including the
   overall survival (OS) rate and tumor‐free survival rate (DFS)
   (Figure [105]4H,I).

FIGURE 4.

   FIGURE 4
   [106]Open in a new tab

   TOX high mobility group box family member 3 (TOX3) regulates signaling
   pathways involved in hepatocellular carcinoma (HCC) growth and
   metastasis. (A) PCA analysis of HCCLM3 cell lines in the TOX3 knockout
   group, overexpression group, and NC group were subjected. (B) and (C)
   GSEA showed that HALLMARK_KRAS_SIGNALING_UP and
   HALLMARK_KRAS_SIGNALING_DN pathways were down‐regulated in the TOX3
   knockout and overexpression groups, respectively. (D) and (E) The
   results of the signaling pathway (KEGG and HALLMARK gene terms)
   enrichment analysis of differential expression genes of TOX3‐KO and
   TOX‐OE both showed that the regulation of TOX3 could significantly
   affect signaling pathways related to tumor growth and metastasis, such
   as MAPK and EMT. (F) The results of RRA analysis for the downstream
   target of TOX3. (G) Violin plot showing the relation of IGFBP3
   expression with the tumor stage of HCC. (H) and (I) The relationship
   between the IGFBP3 expression with OS and DFS in HCC patients was
   analyzed by the TCGA database. *p < 0.05, **p < 0.01, ***p < 0.001.

   To further clarify whether the regulatory effect of TOX3 on the
   malignant biological behavior of HCC cells depends on IGFBP3, we
   knocked down IGFBP3 in TOX3‐overexpressing HCCLM3 cells and
   overexpressed IGFBP3 in TOX3‐knockout HCCLM3 cells. According to our
   study, the oncogenesis and migratory capacity of HCCLM3 cells in the
   IGFBP3 knockdown group (OE‐TOX3‐sh‐IGFBP3) were substantially
   suppressed compared with the OE‐TOX3‐sh‐NC group
   (Figure [107]5A,C,[108]E,G). In addition, compared with those in the NC
   group (KO‐TOX3‐OE‐NC), the oncogenesis and migratory capacity of HCCLM3
   cells with suppressed proliferation and development after TOX3 knockout
   was restored by the overexpression of IGFBP3 in the experimental group
   (KO‐TOX3‐OE‐IGFBP3) (Figure [109]5B,D,[110]F,H).

FIGURE 5.

   FIGURE 5
   [111]Open in a new tab

   TOX high mobility group box family member 3 (TOX3) regulates the
   hepatocellular carcinoma (HCC) oncogenesis and development by targeting
   IGFBP3. (A) and (B) Knocking down IGFBP3 in TOX3‐OE HCCLM3 cells and
   overexpressing IGFBP3 in TOX3‐KO HCCLM3 cells, respectively, the
   proliferation activity of cells was detected with a CCK‐8 kit. (C) and
   (D) Knocking down IGFBP3 in TOX3‐OE HCCLM3 cells and overexpressing
   IGFBP3 in TOX3‐KO HCCLM3 cells, respectively, the ability of colony
   formation was detected. (E) and (F) Knocking down IGFBP3 in TOX3‐OE
   HCCLM3 cells and overexpressing IGFBP3 in TOX3‐KO HCCLM3 cells,
   respectively, wound healing assay was detected. (G) and (H) Knocking
   down IGFBP3 in TOX3‐OE HCCLM3 cells and overexpressing IGFBP3 in
   TOX3‐KO HCCLM3 cells, respectively, vitro cell migration assay was
   performed to detect the migratory capacity of the cells. *p < 0.05,
   **p < 0.01, ***p < 0.001.

2.4. TOX3 regulates the key molecules of the MAPK and EMT signaling pathways
by targeting IGFBP3

   The results of high‐throughput RNA sequencing showed that the
   regulatory effect of TOX3 on the malignant biological behavior of HCC
   cells involved mainly focused on signaling pathways related to tumor
   growth and metastasis, such as the KRAS, MAPK, and EMT pathways
   (Figure [112]4D,E and Figure [113]S1E). In addition, a series of
   high‐quality related studies have shown that IGFBP3 can regulate
   signaling pathways related to proliferation and metastasis, including
   the MAPK, PI3K‐Akt, and EMT pathways.[114] ^18 , [115]^19 , [116]^20 ,
   [117]^21 Thus, western blot (WB) was employed to clarify whether
   IGFBP3, which is regulated by TOX3, can affect the expression of key
   molecules in the MAPK and EMT pathways. The results indicated that the
   knockdown of IGFBP3 (OE‐TOX3‐sh‐IGFBP3 HCCLM3) led to a significant
   decrease in the key proteins p‐ERK1/2, vimentin, Snail, Cluding, and
   Slug in the MAPK and EMT pathways (Figure [118]6A). Additionally,
   IGFBP3 upregulation substantially increased the expression of p‐ERK1/2,
   Snail, and Cluding (Figure [119]6B). Additionally, immunohistochemical
   (IHC) staining indicated that TOX3 knockout substantially downregulated
   IGFBP3, Ki‐67, vimentin, and N‐cadherin expression but upregulated
   E‐cadherin expression (Figure [120]6C). These experimental results are
   consistent with previous high‐quality studies, indicating that IGFBP3
   can target key molecules involved in regulating MAPK and EMT signaling
   pathways.

FIGURE 6.

   FIGURE 6
   [121]Open in a new tab

   TOX high mobility group box family member 3 (TOX3) regulates key
   molecules in the MAPK and EMT signaling pathways by targeting IGFBP3,
   and the E3 ligase TRIM56 potentially mediates the proteasomal
   degradation of TOX3. (A) Effect of IGFBP3 knockdown on the expression
   of p‐ERK1/2, Vimentin, Snail, Cluding, and Slug in the MAPK and EMT
   pathways. (B) Effect of IGFBP3 overexpression on the expression of
   p‐ERK1/2, Snail, and Cluding in the MAPK and EMT pathways. (C) IHC
   staining about the expression of IGFBP3, Ki‐67, Vimentin, N‐cadherin,
   and E‐cadherin in the KO‐TOX3 and NC groups. (D) MG132 was added to the
   TOX3‐overexpressing HCCLM3 cells, and the TOX3‐FLAG expression was
   detected at 6, 12, and 24 h. (E) IP‐MS revealed the proteins
   interacting with TOX3 and explored the ability of the E3 ubiquitin
   ligase to target TOX3 degradation. (F) The expression of TRIM56 between
   the normal liver tissue (GETx) and hepatocellular carcinoma tissues
   (TCGA‐LIHC). (G) The heatmap of GSVA scores of the ubiquitin‐proteasome
   system and autophagy between TRIM56‐high and TRIM56‐low TCGA‐LIHC tumor
   samples, samples divided according to the median of the mRNA expression
   of TRIM56. Downregulation of TRIM56 significantly decreased the
   activity of the ubiquitin‐proteasome system but had no effect on
   autophagy. (H) TRIM56 was decreased and the expression of TOX3 was
   detected. *p < 0.05, **p < 0.01, ***p < 0.001.

2.5. Proteasome degradation of TOX3 is potentially mediated by the E3 ligase
TRIM56

   During this experiment, we found that there was a dramatic difference
   in the level of TOX3 transcription and protein expression. To explore
   the posttranslational ubiquitination of TOX3, we tried to intervene in
   the targeted degradation process of the intracellular proteasome system
   by using the proteasome inhibitor MG132. Consequently, TOX3 was
   upregulated after MG132 intervention and substantially increased with
   the prolongation of MG132 treatment in TOX3‐overexpressing HCCLM3 cells
   (Figure [122]6D). These results suggest that the TOX3 protein undergoes
   intracellular degradation via the ubiquitin‐proteasome system. To
   further investigate the specific ubiquitin ligase responsible for
   targeting the degradation of TOX3, we sequenced the protein spectrum of
   immunoprecipitated proteins by using the protein complex combined with
   Flag to pull down TOX3 in TOX3 overexpressing HCCLM3 cells. The
   sequencing results of IP‐MS showed that the TRIM56 potentially
   participated in the targeted proteasomal degradation of TOX3
   (Figure [123]6E). Furthermore, we further analyzed the expression level
   of TRIM56 in the normal liver tissue (GETx) and HCC tissue (TCGA‐LIHC)
   samples, and the results indicated that the expression level of TRIM56
   in HCC tissues was markedly down‐regulated than that in the normal
   liver tissues (Figure [124]6F). Heatmap of GSVA scores of the autophagy
   and ubiquitin‐proteasome systems between TRIM56‐high and TRIM56‐low
   TCGA‐LIHC tumor samples. Samples were divided according to the median
   of the expression level of TRIM56. TRIM56 downregulation significantly
   decreased the activity of the ubiquitin‐proteasome system but had no
   effect on autophagy (Figure [125]6G and Figure [126]S2A). Finally,
   small interfering RNA (siRNA) was used to silence TRIM56 in
   OE‐TOX3‐FLAG HCCLM3 cells, which demonstrated that the downregulation
   of TRIM56 can increase the expression level of TOX3 (Figure [127]6H).
   This part of the experimental results preliminarily explored the
   upstream regulatory mechanism of TOX3, and the results showed that
   proteasomal degradation of TOX3 is potentially mediated by the E3
   ligase TRIM56.

3. DISCUSSION

   The degree of malignancy of HCC is high, resulting in a poor prognosis
   for patients, and the underlying mechanisms of hepatocarcinogenesis and
   development remain incompletely understood. At present, much‐published
   literature has demonstrated that MVI positivity indicates poor
   prognosis in HCC patients and it occupies an important position among
   many adverse factors.[128] ^2 , [129]^3 , [130]^4 , [131]^22 Previous
   studies have shown that MVI serves as a precursor to both intrahepatic
   and extrahepatic metastasis of HCC, representing a significant
   manifestation of highly malignant HCC cells and signifying their
   heightened invasiveness.[132] ^23 However, the molecular mechanism
   contributing to the poor prognosis in MVI‐positive HCC patients remains
   unclear. Consequently, MVI was used as the entry point for identifying
   highly malignant markers of HCC to determine the key genes that could
   lead to poor prognosis in patients with a high malignancy degree of
   HCC. HCC tissues from patients with MVI, HCC tissues without MVI, and
   adjacent normal tissues were sequenced, and the key target gene TOX3
   was screened out after statistical analysis. The results of expanded
   sample size verification and a series of in vitro and in vivo
   functional phenotyping experiments demonstrated that TOX3 significantly
   promoted the growth and metastasis of HCC.

   Previous studies have shown that the proteins encoded by TOX3 may
   participate in the bending and release of DNA and changes in chromatin
   structure. Furthermore, other published studies have indicated that
   TOX3 can act as a transcriptional costimulatory factor, which can
   promote the transcription of downstream genes.[133] ^7 , [134]^8 ,
   [135]^9 , [136]^10 , [137]^24 The sequencing results of the TOX3
   overexpression group, knockout group, and NC group showed that TOX3
   could regulate many biological processes (BPs) and molecular‐related
   functions (MFs) related to cell proliferation and metastasis. These
   include but are not limited to cell proliferation, angiogenesis, cell
   adhesion, cell connection, extracellular matrix construction, or growth
   factor activity. Considering the phenotypes observed in the
   aforementioned cells and animal models, it can be inferred that TOX3
   enhances the malignant biological behavior of HCC cells by affecting
   cell proliferation, exfoliation, and metastasis. Moreover, our
   enrichment analysis demonstrated that the regulation of TOX3 could
   significantly affect classical signaling pathways related to KRAS,
   MAPK, EMT, etc. In addition, IGFBP3 was identified as a downstream
   target. According to previous studies, TOX3 can promote the oncogenesis
   and development of breast cancer, colorectal cancer, and lung cancer
   through the MAPK, EMT, and Hippo‐related signaling pathways.[138] ^12 ,
   [139]^13 Thus, the differential gene enrichment analysis of our study
   is relatively reliable. Interestingly, based on previous studies,
   IGFBP3 targets the regulation of classical signaling pathways and key
   molecules such as MAPK, EGFR, AKT, and EMT to regulate tumor
   proliferation and metastasis.[140] ^18 , [141]^19 , [142]^20 , [143]^21
   IGFBP3 is the main member of the IGFBP family, especially IGFBP3
   secreted into the circulatory system occupies an important
   position.[144] ^18 Much literature has reported that IGFBP3 can
   regulate cell proliferation and migratory capacity by binding to cell
   surface receptors and stimulating its activity.[145] ^25 , [146]^26 ,
   [147]^27 Song et al. showed that galectin‐3 can regulate the expression
   of IGFBP3 and vimentin through β‐catenin and substantially strengthen
   the angiogenesis and the development of HCC, thereby IGFBP3 markedly
   upregulates the proliferation and migration of HCC cells.[148] ^20
   Janet et al. have also indicated that IGFBP3 enhances the oncogenesis
   and development of breast cancer; mechanically, it can upregulate the
   EGFR phosphorylation and activate the p44/42 and p38 MAPK classical
   signaling pathways.[149] ^19 Deobrat et al. demonstrated that the M6A
   reader YTHDF2 can upregulate IGFBP3 expression through MYC, which
   subsequently modulates the AKT and EGFR through IGFBP3 to enhance the
   growth of glioma. Notably, the overexpression of IGFBP3 can restore the
   tumorigenesis of glioma caused by YTHDF2 deletion in vivo.[150] ^28 Liu
   et al. showed that IGFBP3 can directly regulate the key molecules of
   the AKT, EMT, and MAPK signaling pathways, thereby influencing the
   oncogenesis and development of renal carcinoma.[151] ^29 Therefore,
   according to previous studies, pathway enrichment analysis, WB assays,
   and IHC staining in this study, it is plausible to suggest that TOX3
   up‐regulates IGFBP3 expression level by binding to the CRE site in the
   promoter of the downstream gene IGFBP3. Thus, it promotes the malignant
   biological behavior of HCC by regulating the classical pathway of tumor
   growth and metastasis. Moreover, according to the results of this
   study, the inhibitory effect of downregulating Tox3 expression on the
   proliferative capacity of HCC tumor cells was dramatically greater than
   that on human normal hepatocytes. Therefore, HCC‐targeted therapeutic
   drugs could be further developed from TOX3, especially for highly
   malignant HCC patients with MVI positivity, which may improve the
   survival rate of these high‐risk patients with recurrence.

   For the regulation of the upstream molecular mechanism of TOX3, this
   study demonstrated that TOX3 potentially undergoes posttranslational
   ubiquitin modification and undergoes intracellular degradation through
   the ubiquitin‐proteasome system. However, E3 ubiquitin ligases are
   critical for the intracellular degradation of substrate proteins in the
   proteasome. Additionally, many published literature have indicated that
   the E3 ligase is related to tumorigenesis and tumor development.[152]
   ^30 , [153]^31 , [154]^32 , [155]^33 The IP‐MS results showed that
   TRIM56 potentially participated in the targeted proteasomal degradation
   of TOX3. Subsequent employment use of siRNA to knock down TRIM56
   resulted in an increase in the expression of TOX3, leading to the
   conclusion that proteasomal degradation of TOX3 is potentially mediated
   by the E3 ligase TRIM56. Notably, some published literature has
   demonstrated that TRIM56 can target and degrade some oncoproteins,
   thereby inhibiting the occurrence and development of malignant
   tumors.[156] ^14 , [157]^15 , [158]^16 , [159]^17 Yang et al. showed
   that the expression level of TRIM56 in HCC tissues is downregulated
   compared with adjacent normal liver tissues, and the downregulation
   degree of TRIM56 expression is closely related to tumor stage and
   prognosis. Mechanistically, TRIM56 can inhibit the oncogenesis and
   metastasis of HCC by targeting the degradation of proteins related to
   the Wnt signaling pathway.[160] ^14 Therefore, the depletion of TRIM56
   in HCC may potentiate the accumulation of TOX3 and further regulate the
   classical signaling pathways of MAPK and EMT by upregulating the IGFBP3
   expression level, eventually promoting the oncogenesis and development
   of HCC (Figure [161]7).

FIGURE 7.

   FIGURE 7
   [162]Open in a new tab

   Schematic model presenting the mechanism by which TOX high mobility
   group box family member 3 (TOX3) promotes the oncogenesis and
   development of hepatocellular carcinoma (HCC) cells.

4. CONCLUSIONS

   In summary, TOX3 in HCC tissues was upregulated, and compared with
   MVI‐negative HCC tissues, TOX3 expression level in MVI‐positive HCC
   tissues was markedly further increased. TOX3 overexpression was
   associated with a more malignant degree of HCC and negative prognosis
   in HCC patients. TOX3 up‐regulates IGFBP3 expression level and fosters
   the oncogenesis and development of HCC via the MAPK and EMT pathways.
   Additionally, the E3 ligase TRIM56 potentially mediates the proteasomal
   degradation of TOX3. The above studies could lead to the identification
   of potential therapeutic targets for HCC treatment.

5. MATERIALS AND METHODS

   All clinical tissue specimens and related clinical information involved
   in this study were in accordance with the principles of the Declaration
   of Helsinki. The animal experiment scheme was approved by the Animal
   Ethics Committee of our hospital (20220302033).

5.1. Clinical samples

   The specimens of the sequencing group (including 12 MVI‐positive and
   ‐negative HCC and paired liver normal tissues) were collected from the
   operating room of our center (Table [163]S1), and the samples of the
   validation group were collected from the biological sample bank of our
   hospital (Table [164]S2), which included HCC patients received
   selective curative hepatectomy from 2018 to 2019 (2021‐DI‐009). The
   inclusion criteria were pathological diagnosis of HCC and no adjuvant
   treatment before the operation. The exclusion criteria were patients
   combined with another type of cancer, and the patients’ data were
   incomplete.

5.2. Cell culture

   The HCC cells HCCLM3 and the Human liver immortalized cells THLE‐2
   involved are all human cell lines commonly used. HCCLM3 cells were
   purchased from the cell bank of the Shanghai Institute of Life
   Sciences, and 293T cells were purchased from the cell bank of Shanghai
   Type Culture Collection (Chinese Academy of Sciences). THLE‐2 cells
   were purchased from Wuhan Procell Life Science & Technology Co. 293T
   and HCCLM3 cells were cultured in 90% DMEM high glucose medium
   supplemented with 10% fetal bovine serum (FBS) and 1%
   penicillin/streptomycin. THLE‐2 cells were cultured in a Complete
   medium of THLE‐2 including BEGM and 10% FBS.

5.3. Animals

   The animals used in the animal experiment were BALB/Nude male mice,
   which were purchased from Huafukang Company and Gempharmatech Company.

5.4. Single‐cell clone selection

   We used the CRISPR‐cas9 system to obtain complete knockout clones of
   TOX3 by transfecting lentivirus for CRISPR targeting of the TOX3 gene
   into HCC cells. Two target sequences including TAACTATATGAATATGGCTG
   (sgRNA1) and CAGTACATCCGGCATGCCTA (sgRNA2), were used for TOX3. Stably
   transfected cells were selected using puromycin and single‐cell clones
   were then isolated using a flow cytometer.

5.5. CCK‐8

   Approximately 1.6×103 cells were suspended and added to each well of a
   96‐well plate. A CCK‐8 proliferation detection kit from the MCE Company
   was used to detect cell proliferation. The absorbance OD value of the
   cells was detected by an EonTM Microplate Reader with a wavelength of
   450 nm.

5.6. Colony formation

   We seeded 2000 cells per well in a 6‐well plate and allowed them to
   incubate for 2 weeks. After fixation with 4% paraformaldehyde, we
   stained the colonies with 0.1% crystal violet.

5.7. Wound healing assay

   We seeded cells into a 6‐well plate and created scratches with a 10 µL
   pipette tip. The rate of cell migration was recorded under a
   microscope, and the location of mapping at each time point was as
   consistent as possible.

5.8. In vitro cell migration assay

   We seeded 5×104 cells in the upper Transwell chamber with an 8 µm pore
   size. The inner chamber contained serum‐free DMEM, while 600 µL of DMEM
   containing 10% FBS was added to each lower chamber. After 48 h, we
   fixed and stained the cells on the lower surface of the chamber with 4%
   paraformaldehyde and 0.1% crystal violet respectively, then recorded
   the images.

5.9. Cell competition test

   TOX3‐KO cells and negative control (NC) cells were cocultured.[165] ^34
   TOX3‐KO cells expressed the red fluorescent protein mCherry, while the
   NC cells had no fluorescence. 1×10[166] ^6 TOX3‐KO and NC cells were
   obtained, respectively; a total of 2×106 cells were fully mixed in the
   same aseptic centrifuge tube. 1/3 of the cells above was used to detect
   the ratio of red fluorescent cells (TOX3‐KO) by flow cytometry, and the
   remaining cells were transferred into triplicate wells of a 6‐well
   plate for coculture. Additionally, the cocultured mixed cells were
   detected by flow cytometry at an interval of one day, the percentage of
   fluorescence positive rate was analyzed, and the above operation was
   repeated.

5.10. Protein immunoprecipitation combined with mass spectrometry

   Immunoprecipitation (IP) antibody Flag was added to the target protein
   and then flipped overnight at 4°C for at least 16 h. The complex was
   added to the pretreated magnetic beads (Protein A/G, MCE) to incubate
   with flipping and binding. The mixture was placed on a magnetic frame,
   and IP lysis buffer was used to wash the magnetic beads, then the
   proteins were denatured and the magnetic beads were separated under the
   setting conditions of 100°C and 10 min. The tube was placed on the
   magnetic frame, and the pull‐down complex protein was collected. The IP
   proteins were subjected to electrophoresis and sent for mass
   spectrometry sequencing analysis.

5.11. Animal experiment

   We used five‐week‐old male BALB/c nude mice established tumor models.
   In detail, the subcutaneous tumor model of liver cancer, the liver
   orthotopic implantation model, and the lung metastasis model were
   established with different numbers of HCC cells. 2 × 106 HCC cells
   within 100 µL suspension were established the subcutaneous tumor model
   by injecting subcutaneously HCC cells into the right chest and abdomen
   of nude mice, and a semicircular colliculus could be formed locally.
   The volume of the subcutaneous tumor and the final tumor weight were
   recorded. 1×106 HCC cells within 100 µL suspension were established the
   liver orthotopic implantation model by injecting into the liver of nude
   mice, and the red fluorescence from mCherry in the HCC cell line was
   strong; thus, the fluorescence technique of small animal live imaging
   systems could be selected for detection and recording. 1.5 × 106 HCC
   cells within 125 µL suspensions were established in the lung metastasis
   model by injecting intravenously into the tail of nude mice, and the
   tumor metastases in vivo would be recorded by small animal live imaging
   system chemiluminescence mode.

5.12. RNA sequencing data analysis

   First, we used TrimGalore (v0.6.7) to perform quality trimming on the
   raw sequencing data. Subsequently, we aligned the processed clean reads
   to the human genome (GRCh38) using Kallsito (v0.46.1) to obtain the
   gene expression matrix. Following this, we employ DESeq2 (v1.38.3)
   software for differential expression analysis, with the criteria for
   selecting differentially expressed genes as follows: (1) absolute value
   of log2 (fold change) > 1; (2) adjusted p‐value < 0.05. Both GSEA
   analysis and pathway enrichment analysis of differentially expressed
   genes are conducted using clusterProfiler (v4.6.0), the KEGG, HALLMARK,
   and GO‐BP gene terms used for analysis were downloaded from MSigDB
   ([167]https://www.gsea‐msigdb.org/gsea/msigdb). Enrichment of
   up‐regulated genes in the TOX3‐OE group was performed in the KOBAS‐i
   database ([168]http://bioinfo.org/kobas/). RobustRankAggreg (v1.2.1)
   was used for integrating differential expression genes. Gene terms
   related to the ubiquitin‐proteasome system and autophagy were
   downloaded from the MSigDB database, and the molecular pathway activity
   of tumor samples in the TCGA‐LIHC dataset was quantified via GSVA
   (v1.46.0).

   The mRNA expression matrix of the TCGA‐LIHC dataset and its
   corresponding clinical information are obtained from UCSC Xena
   ([169]https://xena.ucsc.edu/). The mRNA expression matrix of normal
   liver tissue was obtained from the GTEx dataset and can be also
   downloaded from UCSC Xena.

5.13. Statistical analysis

   The data of this study was analyzed by GraphPad Prism 8.0, ImageJ, and
   ipwin 32. The schematic diagram of this paper is completed by using the
   BioRender online drawing website. Student's T‐test was used for the
   analysis of the continuous variables, and the Pearson chi‐square test
   or Fisher exact probability method was used for the analysis of the
   counting data. The survival rate of patients with liver cancer was
   statistically analyzed by the Kaplan‒Meier method, and the difference
   in survival rate was compared by the log‐rank test. p‐Values < 0.05
   were considered significant.

AUTHOR CONTRIBUTIONS

   Designed this work (L.F., M.X.L, W.Y.G., and S.H.B.). Analyzed this
   data (P.Y.F. and Y.J.). Wrote this paper (P.Y.F. and Y.J.). Revised
   this paper (P.Y.F., Y.J., T.L.Y., L.B., L.F., W.Y.G., M.X.L., and
   S.H.B.). Took part in some major experiments (P.Y.F., Y.J., C.K.F,
   Z.H.L., Z.W., and T.L.Y). All authors have read and approved the final
   manuscript.

CONFLICT OF INTEREST STATEMENT

   The authors declare no conflict of interest.

ETHICS STATEMENT

   The clinical tissue specimens’ acquisition was approved by the
   biological sample bank of West China Hospital of Sichuan University
   (2021‐DI‐009). The animal experiment scheme was approved by the Animal
   Ethics Committee of West China Hospital of Sichuan University
   (20220302033).

Supporting information

   Supporting Information
   [170]MCO2-5-e510-s001.docx^ (1.5MB, docx)

ACKNOWLEDGEMENTS