TP53INP1 Downregulation Activates a p73-Dependent DUSP10/ERK Signaling Pathway to Promote Metastasis of Hepatocellular Carcinoma

Liver cancer remains one of the most prevalent and deadliest cancer types worldwide. Hepatocellular carcinoma (HCC) accounts for over 75% of all liver cancer cases. Metastasis and postsurgical recurrence are common and represent major obstacles to the improvement of patient survival. HCC patients are often diagnosed at an advanced stage when curative therapy is no longer available and even after surgery, the prognosis of HCC remains unsatisfactory, with a 5-year postrecurrence rate at >70%. Metastasis is a complex multistep process involving alterations in the dissemination, invasion, survival, and growth of new cancer cell colonies, which are regulated by a complex network of intra- and intercellular signal transduction cascades (1). However, metastasis remains the most poorly understood component of cancer pathogenesis (2). Elucidation of the mechanisms underlying metastasis is fundamental for the development of new therapeutic treatments for advanced metastatic HCC.

Extracellular signal-regulated kinases (ERK) have been shown to play critical roles in malignant transformation and cancer metastasis (3). Oncogenic activation of ERKs can be induced by various mechanisms including transcriptional overexpression, mutations in upstream components of the MAP kinase pathway, such as RAS and BRAF, and downregulation of negative regulator dual-specificity MAP kinase phosphatases (DUSP; ref. 4). ERK plays a major role in invasion by inducing proteases that degrade the basement membrane, enhances cell migration, and increases cell survival. Activated ERK pathway has been shown to correlate with the expression of epithelial–mesenchymal transition (EMT) markers, a hallmark of metastasis. These findings collectively suggest that ERK plays a major role in tumor progression and metastasis. However, our knowledge of endogenous regulators of DUSP/ERK remains limited and how they work to promote HCC metastasis is also not known.

TP53INP1 is a stress-induced tumor suppressor gene with antiproliferative and proapoptotic activities (5, 6). It is an alternatively spliced gene encoding two protein isoforms (α and β), and when overexpressed, both isoforms exert a tumor suppressor function, mainly by inducing the transcription of target genes involved in cell-cycle arrest and p53-mediated apoptosis as part of the cell responses to genotoxic stress. Significant reduction or loss of TP53INP1 expression has been shown in a number of cancer types, including those of the stomach (7), breast (8), pancreas (9), esophagus (10), lung (11), melanocyte (12), colon (13), and blood (14). In relation to metastasis, TP53INP1 has only been implicated in a handful of studies including one report where they found transcriptional levels of TP53INP1 to be downregulated in metastatic lung of brain cancers (15). A more recent study led by our collaborator Dusetti and colleagues found TP53INP1 to reduce pancreatic cancer cell migration by regulating SPARC expression (16). TP53INP1 is a target gene of the transcription factor p53. Conversely, TP53INP1 has also been shown to play a role in cellular homeostasis through p53-dependent and p53-independent manners (5, 6). In addition to p53, TP53INP1, which is also a p73 target gene, can enhance transcriptional activity of p73 to induce cell-cycle arrest and cell death (17). Thus, TP53INP1 can exert its tumor suppressor function by inducing the transcription of both p53 and p73 target genes.

In our previous studies, we found that the initiation, growth and self-renewal of CD133⁺ liver tumors to be fine-tuned by a balance of miR-130b overexpression and TP53INP1 downregulation (18). This result suggests that TP53INP1 is a critical effector driving hepatocarcinogenesis. Nevertheless, to date, no studies have determined the function of TP53INP1 in HCC metastasis or the molecular mechanism by which TP53INP1 regulates invasion and metastasis in HCC. Here, we demonstrate that TP53INP1 is frequently downregulated in advanced-stage and metastatic human HCC tumors and that downregulation of TP53INP1 in HCC functionally promotes metastasis through ERK activation via a p73-dependent DUSP10 regulation. Findings from our study not only provide new insights into how HCC metastasis is regulated but also provide a new layer of mechanism by which DUSP10/ERK signaling is regulated by p73/TP53INP1 and also identify DUSP10 as a new transcriptional effector of p73.

Gene expression profiling studies involving multiple clinical samples were performed analyzing the expression of specific transcripts in two datasets available through Gene Expression Omnibus (GSE25097 and GSE40367; refs. 19, 20). In addition, human primary and matched metastatic HCC tissue samples were obtained from 37 patients undergoing hepatectomy at the Sun Yat-Sen University Cancer Centre in Guangzhou, China. Tissue samples were collected from patients who had not received any previous local or systemic treatment prior to operation. Use of human samples was approved by the committee for ethical review of research involving human subjects at the Sun Yat-Sen University Cancer Centre.

Human HCC cell lines Hep3B, SNU182, SK-Hep-1, and human hepatoblastoma cell line HepG2 were purchased from American Type Culture Collection. Human liver cell line LO2 and HCC cell lines PLC8024, QSG-7701, and QGY-7703 were obtained from the Institute of Virology, Chinese Academy of Medical Sciences, Beijing, China. Human HCC cell line HLE was obtained from Japanese Collection of Research Bioresources Cell Bank. Immortalized normal liver cell line, MIHA, was provided by Dr. J. R. Chowdhury, Albert Einstein College of Medicine, New York, New York (21). MHCC97L cells were obtained from Liver Cancer Institute, Fudan University, China (22). 293FT cells were purchased from Invitrogen. All cell lines used in this study were obtained between 2013 and 2016, regularly authenticated by morphologic observation and AuthentiFiler STR (Invitrogen) and tested for absence of mycoplasma contamination (MycoAlert, Lonza). Cells were used within 20 passages after thawing.

U0126 was purchased from Cell Signaling Technologies. Mitomycin C was purchased from Calbiochem.

Proteome Profiler Human Phospho-Kinase Array Kit was purchased from R&D Systems (ARY003B).

Total RNA was extracted using RNAisoPlus (Takara). For quantitative (q)RT-PCR of mRNA targets, cDNA was synthesized by PrimeScript RT Master Mix (Takara) and amplified with EvaGreen qPCR MasterMix-R (Applied Biological Materials) and primers listed in Supplementary Table S1. β-Actin was amplified as an internal control. Reactions were performed on an ABI Prism 7900 System (Applied Biosystems) with data analyzed using the ABI SDS v2.3 software (Applied Biosystems). Relative expression differences were calculated using the 2^−ΔΔCt method.

Protein lysates were quantified and resolved on a SDS-PAGE gel, transferred onto PVDF membrane (Millipore), and immunoblotted with a primary antibody, followed by incubation with a secondary antibody. Antibody signal was detected using an enhanced chemiluminescence system (GE Healthcare). The following antibodies were used: TP53INP1 (1:250, Genway Biotech, GWB-61D856), p-ERK1/2 (1:1,000, Cell Signaling Technology, 9101), total ERK (1:1,000, Cell Signaling Technology, 9102), DUSP10 (1:500, Cell Signaling Technology, 3483), p73 (1:1,000, Novus Biologicals, NB100-56674), BAX (1:1,000, Cell Signaling Technology, 2772), MDM2 (1:500, Santa Cruz Biotechnology, sc-965), and β-actin (1:5,000, Sigma-Aldrich, A5316).

Expression plasmids for shRNAs were made in a pLKO.1-puro vector (Sigma-Aldrich). The targeted sequences were: human TP53INP1 (464, 5′-CCGGCATAGATACTTGCACTGGTTTCTCGAGAAACCAGTGCAAGTATCTATGTTTTTTG-3′) and (3834, 5′-CCGGGCGCCATGTTTCTCAAAGTTTCTCGAGAAACTTTGAGAAACATGGCGCTTTTTTG-3′); human p73 (753, 5′- CCGGATCCGCGTGGAAGGCAATAATCTCGAGATTATTGCCTTCCACGCGGATTTTTTG-3′) and (1643, 5′- CCGGCCAAGGGTTACAGAGCATTTACTCGAGTAAATGCTCTGTAACCCTTGGTTTTTG-3′); human ERK1 (5′- CCGGCTATACCAAGTCCATCGACATCTCGAGATGTCGATGGACTTGGTATAGTTTTTG-3′) and ERK2 (5′- CCGGGACATTATTCGAGCACCAACCCTCGAGGGTTGGTGCTCGAATAATGTCTTTTTG-3′) and scrambled shRNA nontarget control (NTC; 5′-CCGGCAACAAGATGAAGAGCACAACTCGAGTTGGTGCTCTTCATCTTGTTGTTTTT-3′). Sequences were transfected into 293FT cells, packaged using MISSION lentiviral packaging mix (Sigma-Aldrich). The full-length complementary DNA of human DUSP10 was amplified in cDNA of human adult normal liver tissue RNA (BioChain) as a template using the following primers (forward 5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTGCCACCATGCCTCCGTCTCCTTTAGAC-3′; reverse 5′-GGGGACCACTTTGTACAAGAAAGCTGGGTCACACAACCGTCTCCACG-3′); and then cloned into the Gateway entry vector pDONR201. DUSP10 was then shuttled into the Gateway destination vector pEZ-Lv199 (GeneCopoeia). Sequences were transfected into 293FN cells, packaged using Lenti-Pac HIV Expression Packaging Mix (GeneCopoeia). Stable clones were selected with puromycin. The full-length complementary DNA of human TP53INP1 was amplified in cDNA of human adult normal liver tissue RNA (BioChain) as a template using the following primers (forward 5′- GGGGACAAGTTTGTACAAAAAAGCAGGCTGCCACCATGTTCCAGAGGCTGAATAAAATGT -3′; reverse 5′-GGGGACCACTTTGTACAAGAAAGCTGGGTTAGTAATTGTACTGACGCGGG -3′); and then cloned into the Gateway entry vector pDONR201. TP53INP1 was then shuttled into the Gateway destination vector pLenti CMV Blast DEST (706-1; Addgene plasmid #17451). Sequences were transfected into 293FN cells, packaged using LentiPac HIV expression packaging mix (GeneCopoeia). Stable clones were selected with blasticidin.

Migration and invasion assays were conducted in 24‐well Millicell hanging inserts (Millipore) and 24‐well BioCoat Matrigel Invasion Chambers (BD Biosciences), respectively. Cells resuspended in serum-free DMEM were added to the top chamber and the medium supplemented with 10% FBS was added to the bottom chamber as a chemoattractant. After 48 hours of incubation at 37°C, cells that migrated or invaded through the membrane (migration) or Matrigel (invasion) were fixed and stained with crystal violet (Sigma‐Aldrich). The number of cells was counted in 3 random fields under 20× objective lens and imaged using SPOT imaging software (Nikon).

Immunohistochemical staining of paraffin sections was carried out using a two-step protocol. After antigen retrieval, sections were incubated with the following antibodies against anti-human TP53INP1 (clone A25-E12; 6 μg/mL; ref. 9), anti-human p73 (1:500, Novus Biologicals, NB100-56674), anti-human DUSP10 (1:50, Cell Signaling Technology, 3483) and anti-human p-ERK1/2 (1:500, abcam; ab50011). Anti-mouse, -rabbit and -rat HRP–labeled polymer (DAKO) was used as secondary antibodies. Color detection was performed by liquid DAB+ substrate chromogen system (DAKO). Slides were counterstained with Mayer's hematoxylin. According to the intensity and total area of the staining, the expression of TP53INP1 was scored as either low (<30%), medium (30 to 60%), or high (>60%) expression.

Both fragments of the DUSP10 promoter regions S1 (−4,400 to −2,201 bp, carrying predicted site sequences ATTAAGTTTCAACATGTA and ATCATGTTACAACATCCA) and S2 (−2,200 to −1 bp, carrying predicted site sequences GGTATGTGCCTGCATGTA and GGCAAGGGGCGGCTTGCC) were amplified and cloned into the XhoI and HindIII sites of a pGL3 basic vector (Promega) for luciferase reporter assay. All PCR products cloned into the plasmid were verified by DNA sequencing to ensure that they were free of mutations and in the correct cloning direction. Primer sequences used listed in Supplementary Table S2.

Chromatin immunoprecipitation (ChIP) assay was performed using the MagnaChIP A Kit (Millipore). Briefly, cells were sonicated and lysed after cross-link treatment by 1% formaldehyde for 10 minutes. The crosslinked protein/DNA complex was immunoprecipitated by anti-p73 antibody or normal IgG bound to protein A magnetic beads. After overnight incubation at 4°C, the complex was eluted and DNA was purified. The immunoprecipitated DNA was quantified by qPCR using primer sequences designed to detect specific regulatory regions listed in Supplementary Table S3.

The study protocol was approved by and performed in accordance with the Committee of the Use of Live Animals in Teaching and Research at The University of Hong Kong. Metastasis was assessed by orthotopically injecting into the liver to observe for extrahepatic metastasis to the lung. Luciferase-‐labeled cells were injected into the left lobes of the livers of 6-week-old BALB/c nude mice (n = 6–10/group). Six to eight weeks after implantation, mice were administered with 100 mg/kg D-luciferin (Gold Biotechnology) via peritoneal injection 5 minutes before bioluminescent imaging (IVIS 100 Imaging System, Xenogen). Livers and lungs were harvested for ex vivo imaging and histologic analysis. Metastatic nodules in the lungs were counted.

Data were analyzed by SPSS 21.0 or GraphPad Prism 6.0 and shown as mean ± standard deviations, unless otherwise specified. Differences between groups were analyzed by an unpaired Student t test for continuous variables. Correlation between expressions was analyzed by the χ-square test. Statistical significance was defined as P ≤ 0.05.

As an initial attempt to explore whether TP53INP1 expression is associated with metastasis, we evaluated the expression of TP53INP1 transcripts in two public gene expression databases. We found that in advanced-stage HCC samples (stage IV of AJCC and TNM) that are more likely associated with recurrence and metastasis, the expression of TP53INP1 was significantly lower than that in early-stage samples (stages I, II, and III; GSE25097; ref. 19; Fig. 1A). In a second, independent data set (GSE40367; ref. 20) that compares metastatic free HCCs and HCCs with extrahepatic metastases, we also observed significantly lower expression of TP53INP1 in HCC samples with extrahepatic metastases (Fig. 1B). To confirm these observations experimentally, we carried out immunohistochemical analyses in 37 pairs of matched primary and metastatic HCC tissue samples. Consistently, TP53INP1 was found to be downregulated in metastatic HCC. Only 28 of the 37 metastatic HCC samples were strong or moderately positive for TP53INP1 and 9 were weak or negative. In contrast, moderate or strong immune positivity for TP53INP1 was present in all 37 out of 37 primary HCC cases, suggesting that a downregulation of TP53INP1 expression is involved in HCC metastasis (Fig. 1C). We then carried out Western blotting analyses in a panel of immortalized normal liver (MIHA and LO2), hepatoblastoma (HepG2) and HCC cell lines (SK-Hep1, HLE, SNU182, PLC8024, MHCC97L, Hep3B, QSG-7701, and QGY-7703). The expression of TP53INP1 was high in the immortalized normal liver and hepatoblastoma cells, while 7 of the 8 HCC cell lines examined displayed significantly lower or undetectable levels of TP53INP1 expression (Fig. 2A).

To assess the functional role of TP53INP1 in cancer cells, we knocked down expression of TP53INP1 in immortalized normal liver cells MIHA and HCC cells MHCC97L using two TP53INP1-specific shRNA lentiviruses (sh-TP53INP1 464 and 3834). As controls, we used lentiviruses expressing nonspecific shRNA (nontarget control, NTC). Efficient TP53INP1 knockdown was confirmed by Western blotting (Fig. 2B). We found that TP53INP1 shRNA-expressing cells had a significantly enhanced ability to migrate and invade compared with control cells (Fig. 2C and D). Similar results were also obtained when the same experiment was performed in the presence of mitomycin C, where cells were inhibited to proliferate, suggesting that TP53INP1-mediated migration and invasion are not a misinterpretation of the cells' altered ability to proliferate (Supplementary Fig. S1). To confirm these findings, we further examined the effects of TP53INP1 expression in an in vivo experimental metastasis model where cells were orthotopically injected into the liver for observation of metastasis to the lung. TP53INP1 suppression induced a potent increase in the ability of MHCC97L cells to not only form tumors in the liver, but also metastasize to the lung (sh-464: 7 of 10 tumors formed in the liver with 6 developing extrahepatic metastasis in the lung; sh-3834: 8 of 10 tumors formed in the liver with 4 developing extrahepatic metastasis in the lung). In contrast, MHCC97L control cells only resulted in tumor growth in the liver in 6 of 10 mice injected, with only 2 mice going on to develop lung metastasis (Fig. 2E; only 4 representative mice shown). Mice were sacrificed after 8 weeks and both livers and lungs were removed for histologic analyses. Hematoxylin and eosin (H&E) staining of the tumors confirmed the bioluminescence signals observed to indeed represent tumor cells and that there is altered ability of the cells to metastasize to the lung as evident by increased number of tumor nodules present there (Fig. 2E and F). Immunohistochemical analysis also found TP53INP1 expression to be preferentially expressed in the livers and lungs of the nontarget control xenografts (Fig. 2F). In addition, to rule out any potential off-target effects of our knockdown shRNAs, we performed experiments to rescue the effects of TP53INP1 shRNAs on migration and invasion by overexpressing TP53INP1 in the same cells. Overexpression of TP53INP1 in MHCC97L cells with TP53INP1 stably repressed rescued the ability of the cells to attenuate migration and invasion in both knockdown clones, further demonstrating the importance of TP53INP1 in regulating metastasis in HCC (Supplementary Fig. S2).

To elucidate the molecular mechanism of TP53INP1 in regulating HCC metastasis, a Proteome Profiler Human Phospho-Kinase Array Kit was utilized to compare the relative levels of 43 human protein kinase phosphorylation between HCC cells with or without TP53INP1 knocked down. Intensity of the spots on the array was quantified by ImageJ analyses and those spots that displayed >1.5-fold change between control and TP53INP1 knocked down cells were selected for further validation by Western blotting analyses. Altogether, 6 phospho-kinases were found altered, including pERK1/2 (T202/Y204 and T185/Y187), pGSK3β (S21/S9), pAMPK1α (T183), pAMPK2α (T172), p-p53 (S15), and p-WNK1 (T60; Fig. 3A), of which only p-ERK1/2 could be validated to be commonly increased in both MIHA and MHCC97L cells (Fig. 3B). To further validate the role of pERK1/2 signaling in TP53INP1-regulated metastasis, we analyzed the impact of introducing an ERK inhibitor (U0126) or stable shRNA against ERK1/2 into HCC cells with TP53INP1 stably repressed on these altered metastatic phenotype. Introduction of U0126 or sh-ERK1/2 in TP53INP1-suppressed HCC cells attenuated in vitro cell migration and invasion abilities (Figs. 3C and D, 4A; Supplementary Fig. S3), as well as lung metastasis in vivo (Fig. 4B–D), suggesting that ERK signaling is needed to drive metastasis in TP53INP1-deficient HCC. Immunohistochemical analysis also found p-ERK expression to be preferentially expressed in the livers and lungs of the nontarget control of sh-TP53INP1 xenografts (Fig. 4D). Note that 1 μmol/L and 10 μmol/L of ERK inhibitor U0126 was initially used to test which concentration was most appropriate for experimental use. At the end, 10 μmol/L concentration was chosen as it resulted in complete abolishment of ERK expression as evident by Western blotting analysis, with no sign of toxicity to the cells (data not shown).

Dual-specificity MAP kinase (MAPK) phosphatases (MKPs or DUSPs) are well-established negative regulators of MAPK/ERK signaling in mammalian cells and tissues. By virtue of their differential subcellular localization and ability to specifically recognize, dephosphorylate and inactivate different MAPK isoforms, they are key spatiotemporal regulators of pathway activity. The MKPs constitute a distinct subgroup of 11 catalytically active enzymes within the larger family of DUSPs, which all share a conserved cluster of basic amino acid residues involved in MAPK recognition (23–25). A screen of these DUSP members at the genomic level by qRT-PCR in HCC cells with or without TP53INP1 suppressed identified DUSP10/MKP-5 to be consistently downregulated in both MIHA and MHCC97L cells following TP53INP1 knockdown (Fig. 5A). This observation was further validated at the proteomic level by Western blotting where DUSP10 was found to be significantly downregulated (Fig. 5B), concomitant with p-ERK1/2 activation in TP53INP1 shRNA-expressing cells as compared with control cells (Fig. 3B). To further validate the role of DUSP10-mediated pERK1/2 signaling in TP53INP1 regulated metastasis, rescue experiments where DUSP10 was reintroduced into HCC cells with TP53INP1 stably repressed was carried out (Fig. 5C). Introduction of DUSP10 in TP53INP1-suppressed HCC cells resulted in a marked decrease in phosphorylated ERK (Fig. 5C) concomitant with attenuated abilities of HCC cells to migrate and invade in vitro (Fig. 5D and E), suggesting that DUSP10-mediated alteration of p-ERK in TP53INP1 low/absent HCC cells can indeed promote metastasis. Consistently, we also observed a significantly lower expression of DUSP10 in human HCC samples with extrahepatic metastases as compared with metastatic free HCC samples in the GSE40367 dataset (20). A positive correlation between TP53INP1 and DUSP10 expression was also found in the same sample cohort (R = 0.4152; P = 0.0001; Fig. 5F).

To determine the link between TP53INP1- and DUSP10-mediated ERK signaling in regulating HCC metastasis, the upstream region of DUSP10 (−1 to −4,400) was analyzed using JASPAR (http://jaspar.genereg.net). Four predicted binding sites of p73, which activity is known to be modified by TP53INP1 (17), was found in the upstream region of DUSP10 [two sites in S1 (A and B); and two sites in S2 (C and D)], with a high relative score of >0.75 (Fig. 6A). ChIP assays showed high physical binding affinity of endogenous p73 to DUSP10 in MHCC97L cells in two of the four predicted sites, namely site B (at −3716 to −3699) and site D (at −1337 to −1320; Fig. 6B, left). To delineate the involvement of TP53INP1 in the regulation of p73 activity and its subsequent binding to the promoter of DUSP10, we knocked down TP53INP1 in MHCC97L and repeated the ChIP assay again. Silencing of TP53INP1 attenuated binding of p73 to DUSP10 promoter in the same two binding sites (B and D; Fig. 6B, right), suggesting that TP53INP1 does indeed play a role in modulating the binding affinity of p73 to the DUSP10 promoter. Note that it has previously been reported that TP53INP1 can also alter the transactivation capacity of p73 on a number of genes, demonstrating a functional association between p73 and TP53INP1 (17). Notably, both MIHA liver and MHCC97L HCC cell lines that were used for functional experiments in our current study are either p53 absent (MIHA) or mutant (MHCC97L). Both cell types are, however, p73 wild-type. Luciferase reporter assays showed high transcriptional activity of endogenous p73 to DUSP10 in MHCC97L cells in both sites 1 and 2, as knockdown of p73 would decrease the activation of DUSP10 promoter by two folds (Fig. 6C). Stable knockdown of p73 in MHCC97L cells led to a marked decrease in DUSP10 and concomitant increase in pERK1/2 expression; while overexpression of DUSP10 in cells with p73 stably suppressed can cancel this effect (Fig. 6D). Further, we found stable TP53INP1 knockdown in MHCC97L cell to also result in a similar decrease in DUSP10 promoter activation (Fig. 6E). Immunohistochemical staining of xenograft tumors generated from HCC cells with and without TP53INP1 knockdown further validated these observations as TP53INP1 repressed tumors displayed elevated pERK1/2 concomitant with a decrease in DUSP10 (Supplementary Fig. S4). Note p73 expression levels remain unchanged in TP53INP1-repressed HCC cells, as evidenced by both Western blotting and IHC analyses (Fig. 6F and Supplementary Fig. S4). In addition, we also noted that in addition to DUSP10, knockdown of TP53INP1 would similarly lead to a marked downregulation of other known p73 targets, including MDM2 and BAX2 (Fig. 6F; ref. 17). Taken together, TP53INP1 can enhance p73 ability to drive DUSP10 transcription, thereby, altering downstream ERK signaling to drive HCC metastasis. In HCC tumors where p73 mutations are rarely observed, TP53INP1 downregulation promotes HCC metastasis through DUSP10 inactivation via p73-dependent DUSP10 promoter binding and regulation, resulting in activation of the ERK signaling pathway (Fig. 6G).

Metastasis is a major hallmark of cancer and yet remains the most poorly understood component of cancer pathogenesis (2). It is a complex multistep process involving alterations in the dissemination, invasion, survival, and growth of new cancer cell colonies, which are regulated by a complex network of intra- and intercellular signal transduction cascades (1). In this study, we demonstrate that TP53INP1 is frequently downregulated in advanced-stage and metastatic human HCC tumors and that downregulation of TP53INP1 in HCC promotes metastasis through DUSP10 inactivation via p73-dependent DUSP10 promoter binding and regulation, resulting in activation of the ERK signaling pathway. Findings from our study not only provide new insight into how HCC metastasis is regulated but also provide a new layer of mechanism by which DUSP10/ERK signaling is regulated by p73/TP53INP1. Note that because TP53INP1-mediated ERK1/2 activation can also lead to increased cell proliferation, we must take caution when we interpret our metastasis findings, such that we must ensure that the metastasis effect is not a by-product of the cells' altered proliferation potential. To address this, we repeated our migration and invasion assays again, in the presence of mitomycin C, a drug used to inhibit cell proliferation.

TP53INP1 is a stress-induced p53-target gene whose expression is modulated by transcription factors such as p53, p73, and E2F1 (6, 17, 26). It encodes two protein isoforms, TP53INP1α and TP53INP1β (5), which have similar functions and can induce cell-cycle arrest and apoptosis when overexpressed (6). In association with homeodomain-interacting protein kinase-2 (HIPK2), TP53INP1 phosphorylates p53 protein at serine 46, thereby enhancing p53 protein stability and its transcriptional activity, leading to transcriptional activation of p53-target genes, cell growth arrest and apoptosis upon DNA damage stress (27). The antiproliferative and proapoptotic activities of TP53INP1 indicate that TP53INP1 has an important role in cellular homeostasis and DNA damage response. TP53INP1 can be subcellularly localized in the nucleus or cytoplasm depending on the context. In addition to its role in the nucleus where it stimulates the transcriptional activity of p53 and p73 (17, 27), it also contributes to autophagy and regulation of energetic metabolism and reactive oxygen species (28–31).

Deficiency in TP53INP1 expression results in increased tumorigenesis. A number of studies have demonstrated a significant reduction of TP53INP1 expression during cancer formation of the stomach (7), breast (8), pancreas (9), esophagus (10), lung (11), melanocyte (12), colon (13), and T-cell leukemia (14); and that downregulation of TP53INP1 correlated with more aggressive clinicopathologic behaviors in several human cancer types (7–9, 11). TP53INP1-deficient mice exhibited exacerbated colitis-associated carcinogenesis (32), while TP53INP1 expression was found to be lost in rat preneoplastic liver lesions (33). In contrast to this, two recent studies published by the same group in 2012 have found TP53INP1 to be frequently overexpressed in prostate cancer and castration-resistant prostate cancer, and that its overexpression correlated with poor prognostic factors and is predictive of tumor relapse (34, 35), suggesting that TP53INP1 appears to play a dual role as both a tumor-suppressing and tumor-promoting gene and that its expression trend is cancer type specific. TP53INP1 downregulation in cancers is regulated at multiple levels by DNA methylation (10), the transcription factors c-myc (10) and n-myc (36), histone deacetylase 2 (36) as well as a plethora of miRNAs including miR-569 (37), miR-155 (9, 38–40), miR-182 (41), miR-93, miR-130b (14, 18), miR-30a, miR-205 (42–43), and miR-125b (11). Studies have not only demonstrated a functional tumor suppressive role of TP53INP1 but also a role in modulating cancer stem cell phenotypes (38), cisplatin and gemcitabine resistance (41, 44), as well as oxidative stress (45). In our previous studies, we found that the initiation, growth, and self-renewal of CD133⁺ liver tumors are regulated by a balance of miR-130b overexpression and TP53INP1 downregulation (18), yet to date, the role of TP53INP1 in HCC metastasis or the molecular mechanism by which TP53INP1 regulates migration and invasion in HCC has not been explored. Prior to findings presented in this study, only three reports have linked TP53INP1 to metastasis where they found TP53INP1 to reduce pancreatic cancer cell migration by regulating SPARC expression (16); that TP53INP1 is downregulated in distant lung metastasis of brain cancer (15); and TP53INP1 3′UTR to function as a competitive endogenous RNA (ceRNA) in repressing the metastasis of glioma cells by regulating miRNA activity (46). Specifically, using a mouse model of skin wound healing in TP53INP1 wild-type and deficient mice, our collaborators elegantly showed TP53INP1 to suppress cell migration in vivo. Similar observations were also noted in vitro in TP53INP1 wild-type and deficient mouse embryonic fibroblasts (MEF). Above studies collectively support a role of TP53INP1 in regulating metastasis.

As mentioned above, TP53INP1 encodes two protein isoforms (α and β; ref. 5). This current study did not examine these two isoforms separately, but just looked at the role of both isoforms collectively in HCC. The antibody used for Western blotting analysis binds to the N-terminus of TP53INP1, which detects both protein isoforms. However, it should be noted that a predominant 36-kDa band was observed in the Western blotting, which according to our previous experience and studies would represent the α isoform. We did observe a much weaker band at 55-kDa band, which in our experience would correspond to the β isoform of TP53INP1. However, this band was only detected upon extensive exposure. RT-PCR analysis on HCC cell lines, clinical samples and sh-TP53INP1 HCC cells using primers specific to just α isoform, β isoform or both α and β isoforms, revealed that expression levels were similarly expressed or unexpressed (data not shown). Whether the two isoforms are differentially expressed at the mRNA and protein level or would exert different functional roles in HCC would need to be further studied.

There is ample evidence to show that TP53INP1 can alter the transactivation capacity of a number of genes through both p53- and p73-dependent manners (17). P53 is mutated in approximately 30% of all liver cancers (47). But unlike p53, mutation of p73 is not a common event in HCC nor other human tumors. P73 was also not found to be differentially expressed in nontumor versus HCC (GSE25097) nor metastatic-free HCC versus HCC with extrahepatic metastasis (GSE40367; data not shown). Here, we have uncovered a novel mechanism by which TP53INP1 downregulation contributes to HCC metastasis, through a p73-dependent DUSP10/ERK signaling pathway. The immortalized normal liver and HCC cell lines that were used in this current study, namely MIHA and MHCC97L, respectively, were either p53 deleted or mutated, but were both p73 intact. Whether downregulation of TP53INP1 promotes HCC metastasis through a similar DUSP10/ERK mechanism in a p53-dependent manner needs to be further studied using appropriate cell lines that harbor wild-type p53. It is interesting to note that we were also able to predict five p53 putative binding sites on the DUSP10 promoter with a relative score higher than 0.75 (which is the same setting used for prediction of p73 binding sites on DUSP10), suggesting that TP53INP1 may also indeed control DUSP10/ERK pathway via a p53-dependent manner. If experimentally proven, TP53INP1 downregulation would be able to regulate DUSP10/ERK via both p53 and p73 means, uncovering a new mechanism for all p53 wild-type, mutated/deleted HCC tumors.

p63 also exhibits significant structural homology to p53 and p73, has been reported to bind to the same responsive element as p73 and plays a role in cancer metastasis (48). It would also be intriguing to study the possible involvement of p63 and TP53INP1-mediated suppression of metastasis. Toward this end, we went back to examine the p63 status in HCC tissues and found that p63 expression is largely absent in HCC (49, 50). With this, we cannot conclude that p63 has no role in TP53INP1-mediated DUSP10/ERK signaling, but at least in the context of HCC, where p63 expression is absent, chances are low.

Our study benefitted from the fast growing publicly available transcriptome datasets deposited in NCBI Gene Expression Omnibus. The two datasets used, namely GSE40367 (20) and GSE25097 (19), were chosen in particular as the clinical samples profiled are all representative of Asian ethnicity and are thus more relevant to the disease in our locality. In particular, the GSE40367 dataset was sampled from laser capture microdissected tissue of pure tumor cells of HCCs with extrahepatic metastases and metastasis-free HCCs. Samples of these are rare and of particular importance to studies like this that focuses specifically on HCC metastasis.

No potential conflicts of interest were disclosed.

Conception and design: K.-Y. Ng, T.K. Lee, S. Ma

Development of methodology: S. Chai

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): K.-Y. Ng, L.-H. Chan, S. Chai, M. Tong, N.P. Lee, Y. Yuan, D. Xie, N.J. Dusetti, A. Carrier, S. Ma

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): K.-Y. Ng, M. Tong, S. Ma

Writing, review, and/or revision of the manuscript: K.-Y. Ng, N.J. Dusetti, S. Ma

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S. Chai, X.-Y. Guan, D. Xie, S. Ma

Study supervision: S. Ma

We thank the Faculty Core Facility at the LKS Faculty of Medicine, The University of Hong Kong for providing and maintaining the equipment needed for animal imaging.

This work was supported in part by grants from Research Grants Council–General Research Fund (HKU_773412M), Collaborative Research Fund (C7027-14G), and the Croucher Foundation Innovation Award to S. Ma.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.