Potential conflict of interest: Nothing to report.
Supported by the Hong Kong Research Grants Council Collaborative Research Fund (HKU 1/06C and HKU 7/CRG/09).
Hepatocellular carcinoma (HCC) is a prevalent cancer with an extremely high mortality rate attributed to HCC metastasis, which is the major cause of tumor recurrence and organ failure. Presence of tumor thrombi in the portal veins (venous metastases) is a clinicopathological feature of metastatic HCCs. In this study, we analyzed the microRNA (miRNA) expression profiles of nontumorous livers, primary HCCs, and venous metastases in the same livers from 20 HCC patients by way of TaqMan low-density array (TLDA) and identified the precise alterations of miRNA expression from nontumorous livers to primary HCCs and venous metastases globally. By unsupervised clustering analysis, nontumorous livers were distinctly segregated from primary HCCs and venous metastases, whereas no discernible difference in the expression pattern could be found between primary HCCs and venous metastases. However, a marked global reduction of miRNA expression levels was detected in venous metastases, as compared with primary HCCs. These data suggest that miRNA deregulation is an early event in liver carcinogenesis and the later global miRNA down-regulation aggravates the preexisting miRNA deregulation to further promote HCC metastasis. Conclusion: Our study has enriched the current understanding of the deregulation of miRNAs in HCC progression and highlighted the sequential and distinctive alterations of miRNA expression in primary HCC and venous metastasis formation. (HEPATOLOGY 2012;)
Hepatocellular carcinoma (HCC), the most common primary liver cancer, is an extremely lethal disease that causes about 700,000 deaths worldwide annually. 1 The high mortality rate of HCC is mainly related to intra- and extrahepatic metastasis leading to liver and extrahepatic organ failures. Although tumor resection or liver transplantation is the most efficient treatment option for HCC patients, the majority of patients are unsuitable for the operation because either metastasis has already occurred at the time of diagnosis or patients have poor liver function. Furthermore, HCC is highly recurrent and intra-hepatic metastasis is the major cause of tumor relapse after resection. 2 Therefore, metastasis remains a major obstacle to HCC treatment. Understanding of the mechanisms underlying metastasis is important to the development of better diagnostic platforms and therapeutic options for HCC patients. Metastasis is a complicated disease which involves multiple steps: invasion out of the primary tumor site, intravasation, survival in the circulatory system, extravasation, and colonization at secondary site. Cancer cells that are able to accomplish these steps have higher metastatic capability due to accumulation of genetic and epigenetic alterations including microRNA (miRNA) expression changes 3
miRNAs are a class of small non-protein coding RNAs. miRNAs are transcribed into initial transcripts called pri-miRNAs that are subsequently cleaved by Drosha, a ribonuclease, to form pre-miRNAs, hairpins of approximately 60-70 nucleotides, that are exported to the cytoplasm by exportin 5. Then, pre-miRNA is digested by Dicer, RNaseIII, into mature miRNAs. Mature miRNAs are short single-stranded RNAs, approximately 18-25 nucleotides long, that can be incorporated into an miRNA-induced silencing complex (miRISC), forming perfect or imperfect matches with the 3′ untranslated region of their target mRNAs and causing mRNA degradation or translational repression. 4 It is estimated that miRNAs regulate the expression of one-third of human genes, thereby directing a wide repertoire of biological mechanisms in cells. 5 Accumulating evidence has demonstrated that miRNAs contribute to aberrant gene expression in cancer initiation and progression. Over the last decade, miRNA profilings of human cancers and their corresponding normal tissues have identified a substantial number of oncomirs, miRNAs that contribute to cancer development by targeting tumor suppressor genes or oncogenes. 6 On the other hand, knowledge about metastamirs, miRNAs that regulate cancer metastasis, is relatively insufficient. 7, 8 Array-based miRNA profiling of a breast cancer cell line and its highly metastatic derivative cell line employed to identify metastamirs showed loss of miR-335 in the metastatic derivative, and its re-expression significantly inhibited breast cancer metastasis by targeting the progenitor cell transcription factor SOX4 and extracellular matrix component tenascin C. 9 In HCC, by comparing the miRNA profiles of 241 cases of human HCCs and their corresponding nontumorous liver tissues, a 20-miRNA aggressive tumor signature was identified based on patients' clinicopathological information. This miRNA signature significantly predicted metastasis-free HCC and HCC with venous metastases as well as tumor recurrence. 10
These studies have provided important insight for miRNA involvement in metastasis based on mouse model and clinicopathologic correlation. Development of HCC is a multistep process advancing from chronic hepatitis, cirrhosis, primary HCC, to metastatic HCC. In this study, we sought to understand the alterations of miRNA expression leading to HCC metastasis by comparing the miRNA profiles of human nontumorous livers, primary HCCs, and venous metastases of the same livers from 20 patients. miRNA profiles of these samples brought two important insights to the understanding of hepatocarcinogenesis. First, we found that miRNA deregulation is an early event in which discernible difference was observed between miRNA profiles of primary HCCs and nontumorous liver through clustering analysis. However, no major change was observed between the miRNA profiles of primary HCCs and venous metastases. Second, unlike other cancers, no global miRNA down-regulation was detected in primary HCCs. Instead, marked global miRNA down-regulation was detected in venous metastases, and this global miRNA down-regulation could exacerbate the preexisting miRNA deregulation in primary HCCs and facilitate metastasis formation.
Twenty advanced HCC cases with intrahepatic metastasis as diagnosed with the presence of venous thrombi were selected from a large cohort of approximately 400 HCC patients who underwent curative surgical resection at Queen Mary Hospital, Hong Kong, between 1999 and 2009. Formalin-fixed and paraffin-embedded sections were retrieved from these cases, and the presence of venous thrombi was reviewed by an experienced liver pathologist (IOLN). In our center, HCCs with gross tumor thrombosis in the portal vein are often inoperable. We therefore selected only those cases in which the venous thrombi were large enough for microdissection. Twenty HCC cases with medium-sized (3-10 mm) HCC thrombi were selected for this study. All patients were of Chinese origin, with a mean age of 51.5 years; 19 were male and one was female. Eighteen patients had chronic hepatitis B virus (HBV) infection, as shown by the positive serum HBV surface antigen status, and the remaining two patients had chronic hepatitis C viral (HCV) infection, as shown by the positive serum anti-HCV status. There were no patients with both HBV and HCV infection. Liver cirrhosis was present in 13 patients. The demographic data and clinicopathological features of the patients are described in Supporting Table 1. Use of human samples was approved by the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster.
miRNA Extraction From Formalin-Fixed Paraffin-Embedded Samples.
Paraffin sections were cut at a thickness of 4 μm, dewaxed, rehydrated, and lightly stained with hematoxylin. Nontumorous livers and their corresponding primary HCC tissues and venous thrombi were examined and microdissected with a 25-gauge needle under a dissecting microscope as described. 11, 12 Four to eight consecutive tissue sections were cut in each case to obtain enough microdissection material for further evaluation. miRNA extraction was performed with an miRNeasy FFPE kit (Qiagen, Valencia, CA). The RNA samples were quantified using a Quant-iT Ribo-Green RNA Assay kit (Invitrogen, Carlsbad, CA) and kept at −80°C for later use.
Quantitative Reverse-Transcription Polymerase Chain Reaction for Mature miRNAs.
Two hundred nanograms of miRNA was used as starting material for reverse-transcription (RT). The RT reaction was performed using an miRNA RT kit with human miRNA Megaplex RT primers (Applied Biosystems, Foster City, CA). Before real-time polymerase chain reaction (PCR), miRNA was subjected to preamplification in Megaplex-Pre-Amp reaction (Applied Biosystems) using one-third of miRNA RT products as starting material. Ten percent of preamplified miRNA was then used for real-time PCR in TaqMan Low-Density Array (TLDA) Human MicroRNA version 2. The above experiments were performed as described in the manufacturer's protocol (Applied Biosystems).
Quantitative RT-PCR data were analyzed with RQ Manager 1.2 with the standard procedures (Applied Biosystems). Problematic wells such as those that were not amplified, had higher relative noise, or exhibited off-scale fluorescence signal were automatically omitted for further analysis. Ct values were determined at 0.1ΔRn threshold level after automatic baseline calibration. Expression levels of individual miRNA were determined by −ΔCt approach relative to the average Ct of four normalization controls (U6, RNU24, RNU44 and RNU48). Expression changes of paired samples were determined by ΔΔCt approach. Unsupervised clustering analysis was done with GeneCluster and Treeview software. 13 The differential expression of miRNAs between paired primary HCCs and their corresponding nontumorous livers, as well as paired venous metastases and their corresponding primary HCCs were analyzed by paired t test. A test was considered statistically significant when the P value was less than 0.05 or adjusted by Bonferroni correction in multiple tests. Pathway analysis of miRNAs was performed with DIANA-mirPath software available from http://diana.cslab.ece.ntua.gr/pathways/. 14 TargetScan 5 was selected for the miRNA target prediction algorithm. Enrichment score was presented by −ln (P value).
Quantification of miRNA Expression.
To investigate the miRNA expression change in metastasis formation of human HCC, we analyzed the miRNA expression profiles of paired nontumorous livers, primary HCCs, and venous metastases from 20 HCC patients. miRNA was extracted from microdissected FFPE samples (Fig. 1) and analyzed with quantitative RT-PCR-based TLDA. It is well recognized that appropriate reference genes for normalization is critical for genome-wide quantitative gene expression analyses. 15, 16 Housekeeping small RNAs, such as 18S and small nucleolar RNAs (snoRNAs) that were stably expressed across different tissue types are commonly used as endogenous controls in miRNA profiling studies. However, it is still uncertain whether these reference genes are differentially expressed between normal and tumor samples. With this concern, we normalized the miRNA expression data with a panel of four commonly used endogenous controls (U6, RNU24, RNU44, and RNU48) in order to avoid systematic bias introduced by particular reference snoRNA. Among the 664 miRNA species included in the array, 374 miRNAs had expression detected in more than 50% of the specimens, including 212 miRNAs that were consistently detected in all specimens, and were therefore selected for further investigations. On the other hand, 290 miRNAs with a detection rate less than 50%, including 151 miRNAs that were not detected in any samples tested, were excluded from the study (Supporting Fig. 1A,B).
miRNA Expression Profiles in Nontumorous Livers, Primary HCCs, and Venous Metastases.
The miRNA expression profiles in the paired nontumorous livers, primary HCCs, and venous metastases were analyzed by unsupervised clustering approach. As shown in Fig. 2, the nontumorous liver samples exhibited a distinct miRNA expression profile and was clearly separated from their corresponding primary HCCs and venous metastases in the clustering analysis. However, the clustering analysis was not able to segregate venous metastases from the primary HCC samples. Primary HCCs and venous metastases from the same patients often clustered together, indicating that the miRNA expression profiles of primary HCCs and corresponding venous metastases were similar.
miRNA Deregulation in Primary HCC.
To further investigate the miRNA deregulation in hepatocarcinogenesis and metastasis, the expression changes of individual miRNA (i.e., ΔΔCt) were plotted against their statistical significance (P value, paired t test) across different sample groups in volcano diagrams. To maintain the statistical stringency in multiple comparisons, tests were considered significant when P < 1.34 × 10−4 (based on Bonferroni correction). When comparing 20 pairs of primary HCCs with their corresponding nontumorous liver samples, significant deregulation was observed in 30 miRNAs, 23 of which were significantly down-regulated in primary HCC samples, and 7 of which were up-regulated (Fig. 3A and Table 1). miR-139-5p and miR-18a were the most significantly down- and up-regulated miRNAs, respectively. These 30 miRNAs, representing the most deregulated miRNAs in primary HCC, could be used as a specific miRNA signature for primary HCC.
Table 1. MicroRNA Significantly Deregulated in Primary HCCs and Venous Metastases
To determine whether miRNA deregulation contributes to HCC metastatic growth, we compared the miRNA expression levels between paired primary HCCs and venous metastases by volcano plot as described above. However, we found that no miRNA reached the Bonferroni adjusted significance level, indicating that there was no significant deregulation of individual miRNAs between primary HCCs and venous metastasis.
However, a global trend of miRNA down-regulation was evidently observed in venous metastases (Fig. 3B). Using a one-sample t test to interrogate the global miRNA expression changes between the nontumorous livers, primary HCC, and venous metastases, we found that venous metastases exhibited a significant global miRNA down-regulation of approximately 0.5ΔΔCt (equivalent to 30% of miRNA expression) when compared with either the nontumorous livers and primary HCCs (P < 0.0001 for each). In contrast, despite the finding that individual miRNAs were frequently deregulated in primary HCCs, there was no directional trend of global miRNA up- or down-regulation between nontumorous livers and primary HCCs (P = 0.2115) (Fig. 4).
miRNA Deregulation During HCC Metastasis Growth.
Similar to the primary HCC samples, a subset of significantly deregulated miRNA was identified when comparing HCC venous metastases to their corresponding nontumorous livers. In total, 70 miRNAs were deregulated in venous metastases. Predominantly, 65 miRNAs were down-regulated in venous metastases, but only five miRNAs were found to be up-regulated in this comparison (Fig. 3C and Table 1). Interestingly, the deregulated miRNA subset identified from venous metastases covered most of the deregulated miRNAs identified in primary HCCs (25/30, 83%) (Fig. 5A), and this observation was consistent with the unsupervised clustering analysis as illustrated in Fig. 1. These findings indicate that the pattern of miRNA deregulation was likely to be already established during primary HCC development and substantial qualitative change of miRNA expression might not be required for HCC metastasis.
On the other hand, for the subset of venous metastases-specific down-regulated miRNAs, we observed a consistent stepwise down-regulation from nontumorous livers, to primary HCCs, to venous metastases (Fig. 5B). Interestingly, the seven miRNAs that were found to be up-regulated from nontumorous livers to primary HCCs also had reduced expression in venous metastases, further strengthening the species-independent global miRNA down-regulation from primary HCC to venous metastases (Supporting Fig. 2). In silico analysis predicted that these miRNAs preferentially participated in regulating the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways that are implicated in cancer development and metastasis, and, in particular, those related to focal adhesions, adheren junctions, and actin cytoskeletal regulation (Table 2). This finding indicates that global miRNA down-regulation in primary HCCs may facilitate HCC metastasis. Taken together, these findings suggest a sequential miRNA deregulation during HCC development and metastasis. Qualitative change on miRNA expression pattern contributes to HCC development, while the subsequent global miRNA down-regulation promotes liver cancer metastasis by exacerbating the preexisting miRNA deregulation in primary HCCs.
Table 2. Pathway Enrichment Analysis
KEGG Pathway Name
KEGG Pathway Number
Number of Genes Regulated by the miRNA Subset
−ln (P value)
Abbreviations: KEGG, Kyoto Encyclopedia of Genes and Genomes; MAPK, mitogen-activated protein kinase; miRNA, microRNA.
Renal cell carcinoma
MAPK signaling pathway
Insulin signaling pathway
Wnt signaling pathway
ErbB signaling pathway
Regulation of actin cytoskeleton
Hepatocarcinogenesis is a multistep process driven by an accumulation of molecular alterations from background liver disease, such as chronic hepatitis and cirrhosis, with or without going through a premalignant intermediate stage known as dysplastic nodule, to early and advanced HCCs. Metastasis is considered a late stage of HCC progression and is the major cause of the high mortality in HCC patients. In the past decade, studies have deciphered the molecular alterations along this multistep hepatocarcinogenesis. 17 Most of these studies have focused on genome abnormalities or transcriptome changes. Results obtained from these studies have established the relationship between gradual accumulation of molecular alterations and stepwise HCC progression. 18, 21 However, previous findings have also indicated that systematic changes in chromosomal deletion or global gene expression are unlikely to be involved in the metastatic formation of primary HCC. 19, 22 We have previously shown that there was no significant difference in allelic losses between primary HCCs and their corresponding intrahepatic metastases, although this absence of major allelic losses in this transformation to a metastatic phenotype may not exclude small-scale chromosomal losses or gene deletions. 19
Recently, the involvement of miRNA deregulation in human carcinogenesis has been increasingly recognized. Previous high-throughput array analyses have clearly demonstrated that the miRNA expression profile is substantially altered in cancer samples. 23-27 miRNA deregulation is an early event in human hepatocarcinogenesis and has been profoundly found in premalignant dysplastic nodules. 11, 28 Emerging evidence has further linked miRNA deregulation to cancer metastasis. Recently, several metastatic suppressive miRNAs have been proposed and characterized in cellular or animal models. 12, 29-33 However, the global miRNA expression in relation to metastasis formation of HCC remains elusive. Because HCC patients with detectable metastasis are often inoperable and thus clinical samples of HCC metastases are extremely rare, direct evidence from comparing clinical primary HCCs and HCC metastases is still missing.
HCC metastasis is characterized by intrahepatic spreading through the portal vein system. Tumor thrombi in the portal or hepatic veins (venous metastases) represent metastatic HCC cells that have acquired molecular changes that enable them to detach from the primary tumor mass, invade the blood vessel, and survive in the circulatory system. 34, 35 In this study, we performed a global expression analysis to investigate the miRNA expression changes in HCC metastasis formation by examining a series of clinical specimens consisting of 20 sets of paired nontumorous livers, primary HCCs, and venous metastases. By way of unsupervised clustering analysis, we found that the global miRNA expression profiles between tumor and nontumorous liver samples are substantially different. This observation was consistent with previous studies and suggests a critical role of miRNA deregulation in liver carcinogenesis. 23-27 However, the miRNA expression profiles of primary HCC and venous metastases were similar. Thus, unlike HCC formation from normal hepatocytes, a substantial miRNA profile change may not be required for later metastatic growth. Consistent with our present observation, a recent study has identified a panel of miRNA metastatic signature by comparing the miRNA expression profiles of primary HCCs obtained from patients with or without metastasis. This miRNA metastatic signature was able to stratify HCC patients according to their metastatic potential and predict tumor recurrence. 10 Therefore, it is likely that miRNA deregulation at an early stage of HCC development predisposes to later metastatic growth of primary HCC.
One interesting finding revealed by this study was the global miRNA down-regulation in the venous metastases. It has previously been reported in other cancers that global miRNA down-regulation was a common feature of human cancers. 36 However, this issue remains controversial in human HCCs, and inconsistent findings have been reported in different studies. The major reason for this inconsistency might be attributed to the various profiling platform and different reference controls involved. In this study, we found no global miRNA down-regulation in primary HCC samples, but it was evident in the venous metastases. It is unlikely that this finding was due to systematic bias introduced by the reference controls, because a strictly consistent trend was observed when the global miRNA expression was normalized against four reference controls (U6, RNU44, RNU48, and RNU24) as a panel or individually (Supporting Fig. 3). The mechanisms behind this global miRNA down-regulation in HCC venous metastases remain an interesting topic to be explored. Recently, mutations on TARBP2 (Trans-activation-responsive RNA-binding protein) and exportin 5 have been reported to impair miRNA maturation in human cancers with microsatellite instability. 37, 38 We speculate that the global miRNA down-regulation in venous metastases could be related to the malfunctioning of miRNA biogenesis machinery, and further investigations are much awaited.
In agreement with the global miRNA down-regulation observed in our venous metastasis samples, we noticed a progressive miRNA deregulation accumulating in the process of HCC formation. As compared with their corresponding nontumorous livers with a stringent statistical criterion, 30 and 70 significantly deregulated miRNAs were identified from the primary HCCs and venous metastases. The subset of deregulated miRNAs identified in venous metastases not only covered most of that identified in primary HCC but also encompassed 45 additional miRNAs that were not found in primary HCCs. Typically, the expression of these miRNAs was progressively decreased from nontumorous livers to primary HCCs and further down-regulated in venous metastases. Many deregulated miRNAs identified here have been shown to be involved in cancer metastasis. For example, we previously reported that miR-139-5p, one of the most down-regulated miRNAs in both primary HCC and venous metastases, was associated with various pathological metastatic features and poor prognosis of HCC patients. Overexpression of miR-139 significantly suppressed HCC cell invasion in vitro and lung metastasis in vivo. 12 In addition, we have reported that miR-125b and miR-145 functionally suppress cell motility in different HCC cell lines. 11, 29 Importantly, pathway enrichment analysis revealed that these miRNAs might have profound functions in regulating various cancer-related pathways, particularly those related to cell motility and adhesion. Thus, further down-regulation of these miRNAs might facilitate HCC metastasis. Further delineating the prognostic significance of these miRNAs individually or as a signature panel in a larger cohort may shed light on clinical HCC stratification and prediction of postoperative tumor recurrence in HCC patients.
Based on the above findings, we propose a sequential miRNA deregulation model involved in HCC development and metastasis. Because miRNA deregulation is an early event in liver carcinogenesis, accumulation of aberrant miRNA expressions drives HCC formation. The later global miRNA down-regulation exacerbates the preexisting miRNA deregulation and promotes metastasis formation by deregulating critical cell motility–associated pathways, which may consequently result in clonal selection that promotes cancer cells to detach from the primary HCC mass, survive in the blood stream, and form venous metastasis in the veins.