Skip to main content
Download PDF

HERC5 is a prognostic biomarker for post-liver transplant recurrent human hepatocellular carcinoma

Journal of Translational Medicine volume 13, Article number: 379 (2015) Cite this article

Abstract

Background and aims

Orthotopic liver transplantation (OLT) can be an effective treatment option for certain patients with early stage hepatocellular carcinoma (HCC) meeting Milan, UCSF, or Hangzhou criteria. However, HCC recurrence rates post-OLT range from 20 to 40 %, with limited follow-up options. Elucidating genetic drivers common to primary and post-OLT recurrent tumors may further our understanding and help identify predictive biomarkers of recurrence—both to ultimately help manage clinical decisions for patients undergoing OLT.

Methods

Whole exome and RNA sequencing in matched primary and recurrent tumors, normal adjacent tissues, and blood from four Chinese HCC patients was conducted. SiRNA knockdown and both qRT-PCR and Western assays were performed on PLCPRF5, SNU449 and HEPG2 cell lines; immunohistochemistry and RNA Sequencing were conducted on the primary tumors of Chinese HCC patients who experienced tumor recurrence post-OLT (n = 9) or did not experience tumor recurrence (n = 12).

Results

In three independent HCC studies of patients undergoing transplantation (n = 21) or surgical resection (n = 242, n = 44) of primary tumors (total n = 307), HERC5 mRNA under-expression correlated with shorter: time to tumor recurrence (p = 0.007 and 0.02) and overall survival (p = 0.0063 and 0.023), even after adjustment for relevant clinical variables. HERC5 loss drives CCL20 mRNA and protein over-expression and associates with regulatory T cell infiltration as measured by FOXP3 expression. Further, matched primary and recurrent tumors from the 4 HCC patients indicated clonal selection advantage of Wnt signaling activation and CDKN2A inactivation.

Conclusions

HERC5 plays a crucial role in HCC immune evasion and has clinical relevance as a reproducible prognostic marker for risk of tumor recurrence and survival in patients.

Background

Hepatocellular carcinoma (HCC) is the most common primary liver malignancy and third leading cause of cancer deaths worldwide, with Hepatitis B virus a major etiological factor [1, 2]. Beyond sorafenib (Bayer HealthCare Pharmaceuticals, Inc.; Onyx Pharmaceuticals, Inc, Germany) which is only effective in a small patient population, there is no approved treatment for HCC. Patients have limited options, and orthotopic liver transplantation (OLT) is viable for certain early stage HCC cases, though it is only efficacious in a subset meeting Milan, Toronto, or UCSF clinical criteria [3–5]. The HCC recurrence rates after OLT range from 20 to 40 %, and treatment options after recurrence are limited [3–5]. To date, certain clinicopathologic variables such as tumor size and absence of macroscopic vascular invasion are used to predict risk of recurrence, though success of these factors vary from study-to-study [6]. In an effort to improve the prediction of HCC recurrence, molecular profiling has been applied in many studies.

Multiple transcriptomic and proteomic studies have been conducted to help understand the link between molecular mediators and factors of etiology, tumorigenesis, disease course, and/or other variables related to survival and recurrence in HCC. In fact, from 2003 to 2010, over 14 studies have identified gene signatures from a minimum of 12 genes or proteins to 186, totally more than 934 genes for purposes of predicting survival and/or recurrence in HCC patients [7–21]. More recently, Kim et al., developed a 233 gene signature to discern early from late tumor recurrence in primarily HBV-positive HCC [22], while Zheng et al., combined a 122 gene hepatic stellate cell signature with clinical variables for a prognostic index to predict overall survival in HCV-positive cirrhosis or HCC patients [23]. In contrast to gene or protein signatures, single analyte prognostic gene expression markers such as TNF-related apoptosis-inducing ligand (TRAIL) mRNA was shown to associate with tumor growth and survival, though the latter result did not show statistical significance, and melanoma-associated antigen-D2 (MAGE-D2) mRNA was identified by Hashimoto et al., as a prognostic factor for disease-specific survival following curative hepatectomy [24, 25] The commonality of genes across these studies is low, primarily due to factors of: heterogeneity within HCC populations, degraded RNA isolated from formalin-fixed tissues, differences in clinical stages and etiologies, small sample sizes, lack of independent validation, and basic analytical strategy used to identify predictive genes. Beyond gene or protein expression patterns shown in these studies, the underlying genetic role in HCC recurrence and how it influences pathway modulation has not been explored—something that can greatly enhance our ability to accurately predict tumor recurrence in HCC.

Recent sequencing studies have advanced our knowledge of genetic oncodrivers in HCC, identifying the most recurrent functional impacting mutations in genes and frequently modulated pathways such as Wnt signaling, G1/S cell cycle signaling, apoptosis, and JAK/STAT signaling [26–29]. Additional work has helped elucidate both sites and functional effects of the viral-host genome integration for HBV within HCC patients [30, 31]. These studies have provided a foundation for the genetic landscape of primary tumors in HCC patients though the genetic basis leading to tumor recurrence remains poorly understood, particularly somatic variation shared between the primary and recurring tumors and mechanisms supporting certain predictors of survival or disease recurrence, beyond statistical correlates.

In this study, we used an integrated omics strategy to identify a hemizygous DNA deletion and concordant mRNA under-expression of HERC5, an IFN-induced HECT-type E3 protein ligase gene associated with shorter: time to tumor recurrence and overall survival in HCC patients. The downstream immune-pathological impact from loss of HERC5 was also determined. Additionally, this study indicates a clonal selection advantage in the genetic changes in Wnt signaling in the recurrent tumors, relative to primary tumors of HCC patients. Assessing risk of such outcomes in HCC is a significant unmet need and a predictive biomarker to help manage clinical decisions has high relevance for patients potentially undergoing OLT.

Patients and methods

Patients and samples

To adhere to REporting recommendations for tumour MARKer prognostic studies (REMARK) reporting of clinical specimens, 21 patients within the Hangzhou criteria [4] who underwent OLT at Renji Hospital from 2008 to 2012 were retrospectively included in this study. No stratification or matching was used for patient inclusion in this study. Informed written consent was obtained from each patient and the study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki as reflected in a priori approval by the Ethics Committee of Renji Hospital. No donor organs were obtained from executed prisoners or other institutionalized persons. Within 24 months after OLT, 9 patients had recurrent liver tumors or remote metastasis following OLT, while the remaining 12 patients were tumor free. All clinical variables considered are provided in Table 1 and further detailed in Additional file 1: Table S1. The primary tumor (PT) and normal adjacent tissue (PNAT) were collected from all patients (n = 21). The recurrent tumor (RT), normal adjacent tissue from donor (RNAT), and recipient blood (PB) were collected from 4 of 9 recurrent patients.

Table 1 The summary of clinical information for 21 Chinese HCC patients
Full size table

DNA sequence, read mapping and variant calling

DNA exome sequencing (WES) was generated by Beijing Genomics Institute (BGI) using the Illumina standard library preparation and sequencing protocols [30]. Paired-end 90mer sequence FASTQs for both data types were provided to MedImmune. WES data was available from four patients, all of whom experienced tumor recurrence post-OLT with the following specimens: PT, RT, PNAT, RNAT, and PB. QC and both patient-level and summarized variants results are provided in Additional file 1: Tables S2–S5. Detailed explanations of somatic and germline variant and indel calling is provided in Additional file 2, in addition to the following methods: Patient identity, Clonal relationship value derivation, Donor tissue presence in recurrent tumors, Somatic copy number variation (CNV) analysis, Specificity of HERC5 prognostic correlation among genes in chr4q, and Integrated pathway analysis (Additional file 1: Tables S6–S8).

RNASeq read mapping and differential expression analysis

RNASeq data was generated by BGI using the Illumina standard library preparation and sequencing protocols [30]. Paired end 90mer sequence FASTQs were provided to MedImmune. Sequence data was QCd using FastQC (v0.10.1), with average read count per mate 50 million. Reads were mapped to reference (UCSC hg19; Feb 2009 release; Genome Reference Consortium GRCh37) using TopHat2 (v2.0.9; 32) using the human reference gtf annotation file (GRCh37.68). Transcript counts were calculated/normalized using htseq-count and DESeq (v1.12.1; 33). DESeq’s negative binomial distribution was used to calculate p-values and fold changes between PT and PNAT as well as RT and RNAT using p < 0.01 and |fold| >2 as a threshold for the four patients. These results were used in the pathway analyses and combined with the same genes harboring copy number (CN) amplified or deleted regions (see Additional file 2 for CN calling methods). Unadjusted p-values were utilized to simply identify the most differential transcripts within a single patient (PT vs. PNAT) using the fold change magnitude as a primary gene ranking. Since p value calculations were conducted within each patient, there was no replication and statistical power was not adequate to warrant multiple testing adjustment. Tumor cell prevalence was evaluated using ABSOLUTE [34] and verified against pathology assessments for each tumor. RNASeq data was available from 21 patients (9 experienced tumor recurrence post-OLT and 12 did not) with the following specimens: PT and PNAT.

Time-to-event analyses

Time-to-event analyses were used to correlate the expression of the four genes (NAA11, HERC5, DDX60, and HERC6) identified with tumor recurrence among the 21 Chinese patients’ primary tumors from this study (n = 9 experienced tumor recurrence; n = 12 did not experience tumor recurrence). The expression of each gene in the normal adjacent tissue was subtracted from the tumor expression for each patient (PT-PNAT) individually, then each gene was cut at the median into high or low expression groups. In alignment with REMARK criteria, Kaplan–Meier (KM) analysis, univariate Cox proportional hazards (PH) regression, and multivariate Cox PH regression analyses were conducted adjusting for HBV status post-OLT, age (binary), gender, and tumor grade. These four variables were the most relevant for potential confounding factors with a molecular prognostic. Cirrhosis status was positive for all but one patient, so this variable was not used in the analysis of these 21 patients. HERC5 was the only significant correlate (p < 0.05) with tumor recurrence among the four genes using a KM model (Fig. 3a; Table 2).

Table 2 Univariate Kaplan–Meier (KM) and Cox proportional hazards (PH) models and multivariate Cox PH models for the 21 Chinese HCC patients in this study, HCC patients from Roessler et al. study, and HCC patients from Boyault et al. study
Full size table

Then, correlation between under-expression of HERC5 and both progression-free survival (PFS) and overall survival was conducted in a publically available HCC microarray dataset [35, E-TABM-36]. This HERC5 under-expression was also tested again with both HCC recurrence and overall survival in an additional publically available microarray dataset [36, GSE14520]. As was conducted for the 21 HCC patients in this study, the three different analysis models (KM model, univariate Cox PH regression, and multivariate Cox PH regression) were used to test association of HERC5 under-expression with outcomes in these additional studies (Table 2). For study [36], liver tissue from healthy donors was available, so HERC5 (219863_at) was divided into high or low patient groups using mean-2 standard deviations (SD) of the normal liver distribution (n = 239) as the cut point (n = 62 and 180 HCC patients in low or high groups, respectively [18 patients were missing clinical data]). The difference between these groups was assessed using the grouping coefficient p-value, hazard ratio, and likelihood ratio test, in multivariate Cox PH regression with the available variables of age, cirrhosis (binary), gender (binary), HBV/HCV status (active viral replication chronic carrier = 2; chronic carrier = 1; no = 0), and TNM staging (I, II, or III). The same variables were also assessed individually in univariate Cox PH regression. Both overall survival and time to tumor recurrence were assessed in two separate analyses (Fig. 3b; Table 2). For study [35], no normal healthy tissue (matched nor independent subject as in study [36]) was available, so HERC5 was cut into high or low groups using the median of the HCC primary tumor expression values (n = 20 and 24 HCC patients in low or high groups, respectively [four patients were missing clinical data]). Then overall survival and PFS between high and low patient groups was computed adjusting for the available variables of gender (binary), age, and HBV (titer negative = 0; titer positive = 1) status (Fig. 3c; Table 2). Note that all available clinical variables in both microarray validation studies were analyzed with both univariate and multivariate Cox PH regressions with HERC5 to assess contribution of these variables to prognostic outcomes. All model summaries are reported in Table 2.

Biological significance of HERC5 loss in HCC

HERC5 siRNA transfection experiments, Microarray study of HERC5 siRNA knockdown in HCC cell lines, qRT-PCR (TaqMan) Validation, and Cell culture, ELISA assays and FOXP3 IHC assay on PT and PNAT samples are described in Additional file 2.

Results

Identification of tumor origin and estimation of donor cell contamination in the recurrent tumors of HCC patients post-OLT

Cases of donor-transmitted malignancies in cadaveric organ transplants are very rare [37, 38] in that recurrent HCC cases post-OLT are likely tumors derived from the recipient, thus the recurrent tumor (RT) genetic composition should match the primary tumor (PT). Previous studies have used microsatellite markers or CNVs to answer this question of tumor origin post-OLT or resection in HCC, as it has important implications for clinical and therapeutic strategies [37–39]. We implemented a derivation of the clonal relationship value [CR, 39] to determine the tumor origin in the RTs for the four patients—all four were of recipient origin (Additional file 2, Additional file 3: Figure S5).

Single nucleotide variants (SNVs) were used to assess donor tissue contamination within each RT. Since all four RTs were determined to originate from the recipient, any clonality difference between the PTs and RTs, such as clonal frequencies is due to purity of the RT biopsy [40]. Donor cell contamination in the capture of the RT can dilute the magnitude of somatic differences. Elimination of recipient versus donor-identified differences [normal adjacent tissue from donor (RNAT) vs. recipient blood (PB)] controls for some of this contamination, though the distribution of cancer clones will still vary between the PTs and RTs. The malignancy in patient HCC11 occurred in the lung, thus, both primary and metastatic tumors were of recipient origin, serving as a negative control against the other three patients. Using two independent approaches—somatic SNVs or germline SNPs, for each patient (Additional file 2), we estimated the proportion of donor cell contamination in the RTs as: HCC1 = 72–86 %; HCC4 = 3–9 %; HCC5 = 48–64 %; and HCC11 = 0 % (recipient = donor; Additional file 3: Figure S6A–B).

Somatic single nucleotide variant and insertion/deletion identification in primary and recurrent tumors of 4 HCC patients

We first asked whether there were potential major drivers shared between the PTs and RTs in HCC. Following quality control of the WES and RNASeq data (Additional file 3: Figures S2, S3), we identified 1,145 somatic variants using a stringent selection approach and controlling for donor-recipient differences (i.e. elimination of RNAT vs. PB variants; Additional file 2) in both PTs and RTs including 616 somatic nonsilent (nonsynonymous, stop-gain, stop-loss, or frameshift substitution) SNVs or insertion/deletions (indels) affecting 567 genes (Additional file 1: Tables S3–S5). On average 123 and 121 nonsilent somatic SNVs and 5 and 4.3 indels were identified, respectively in the PTs and RTs—an average of 96 SNVs and 3.5 indels shared between these tumors (Additional file 1: Table S3). The somatic SNV distribution in PT and RT specimens had highest occurrences of C > T/G > A and lowest occurrences of T > G/A > C (Additional file 3: Figure S4), consistent with a previous HCC report of coding exons [27], and PT and RT pairs for each patient had similar distributions of transition/transversion substitutions, though it is interesting that HCC11 showed highest prevalence of T > A/A > T. The nonsilent-to-silent SNV rate was average of 2.8 in the PTs and 2.3 in the RTs—a lower ratio in the RTs due to donor-tissue contamination ( Additional file 1: Table S3, Additional file 2). Nonsilent variants shared between PTs and RTs had a higher proportion of clones compared to each unique set (Additional file 3: Figure S1A-D) and variant allele frequencies (Vfs) for these shared SNVs were significantly higher (mean = 39.4) than those unique to PTs (mean = 23.4; p < 0.001 all four patients; Additional file 1: Table S3; Additional file 3: Figure S1E), suggesting shared somatic mutations to likely be driver mutations. The SNVs in the PTs called using WES were confirmed at certain loci with adequate depth and quality using WGS and RNASeq data (Additional file 1: Table S4).

Somatic copy number variant (CNV) detection in the primary, recurrent, and shared tumors of HCC patients

Somatic copy number (CN) amplifications or deletions were selected for uniqueness to the PT, RT, or shared between the two, all within at least 3 of 4 patients (Fig. 1; “Patients and methods”). Among the hemizygous (hemi) or homozygous (homo) amplifications identified, those that have been observed in a previous study of primary tumors in HCC using comparative genomic hybridization [39] are indicated by an asterisk (*). Those amplifications that were common to both the PT and RT in this study included: 1q* (hemi in all patients), 6p* (homo in 2 patients; hemi in 1 patient), 8q* (homo in 1 patient; hemi in 2 patients), 17q* (hemi in 3 patients), and 20p* (hemi in 3 patients). Common hemi or homo deletions identified in PT and RT included: 4q* (hemi in 3 patients) and 17p* (hemi in 3 patients), and amplified regions unique to RTs included 17q (hemi in all patients) and 20q (hemi in all patients). Thus, the vast majority of somatic CN amplifications or deletions were shared between the PT and RT, with two short regions having unique CN amplifications in the RT.

Fig. 1
figure 1

Somatic copy number amplifications (red), deletions (green), or neutral (yellow) identified in the primary (inner track) and recurrent (outer track) tumors of a HCC1, b HCC4, c HCC5, and d HCC11; e both shared and unique regions of CN amplifications or deletions in 3/4 patients across the genome. Color code is as follows: unique to primary tumors (red amplification; blue deletion); unique to recurrent tumors (brown amplification); and shared by the primary and recurrent tumors (green amplification; purple deletion)

Full size image

An integrated genetics and genomics analysis identifies Wnt-signaling pathway activation and tumor suppressor gene CDKN2A inactivation shared in both PTs and RTs of HCC patients

Using somatic variants, CNVs, and gene expression from PTs and RTs (Additional file 3: Figure S7; “Patients and methods”), Wnt/β-catenin signaling was identified as the most altered pathway in both PTs and RTs across all four patients (Additional file 1: Table S6). Although few affected genes were shared between patients, all identified genetic changes had the potential to prevent ubiquitin-mediated degradation of beta-catenin in the cytoplasm, drive activation of beta-catenin in the nucleus (Fig. 2), and subsequent loss of cell-cycle control by inactivation of CDKN2A.

Fig. 2
figure 2

Wnt/β-catenin signaling and directly related pathways affected by genetic-driven and/or gene expression activation or suppression in primary and/or recurrent tumors of 4 HCC patients. Determination of this pathway as most activated using genetic and genomic data is described in “Patients and methods” while other top ranked pathways are presented in Additional file 1: Table S6. Each of the 4 patients’ primary or recurrent tumors are represented as a quadrant on each pathway node

Full size image

Specifically, deleterious somatic genetic or genomic alterations shared between PTs and RTs included: HCC1: nonsilent mutations in TP53 and AKT2, with activation of upstream gene signatures in CTNB1, E2F1and TCF4 and suppression of CDKN2A upstream gene signature; HCC4: CDH11 and TGFBR3 stop codon mutations, DNA amplification in POU5F1 and UBD, and DNA deletion in GNAO1with activation of E2F1 and TXB2 upstream gene signatures and suppression of upstream gene signatures CDKN2A and RB1; HCC5: DNA deletion in CDH1 and a nonsilent mutation in TP53 with activation of E2F1 and MYC upstream gene signatures and suppression of upstream gene signatures CDKN2A and TP53; HCC11: activating mutation S45Y in CTNNB1, a stop codon mutation in LRP1, and a nonsilent mutation in the ANK2 domain of CDKN2A, with DNA amplification in SOX9. All four patients converge on shared inactivation of a key cell-cycle control tumor suppressor gene CDKN2A in PTs and RTs Using the same strategy to identify the most enriched pathways unique to RTs in all patients, cell cycle signaling was identified (Additional file 1: Table S7).

HERC5 within chromosome 4q somatic CN deletion shows mRNA under-expression and predicts risk of survival and tumor recurrence in the primary tumors of HCC patients

Next we focused on the largest region of somatic CN deletion shared in both the PTs and RTs on chromosome 4q (~107 Mbp). Genes were evaluated within this region to identify a single molecular biomarker to predict HCC tumor recurrence and survival. WES and RNASeq from the four patients as well as RNASeq from PTs and PNATs from a larger population of HCC Chinese patients who experienced tumor recurrence (n = 9 including the 4 described above) or did not (n = 12) were used to identified a predictor of tumor recurrence and survival. Using RNASeq data from the 21 PTs and PNATs, DESeq-normalized log2 gene counts in the PTs were scaled by the matched PNATs within each patient. Then these fold changes for each patient were used in a contrast between patients that experienced tumor recurrence (n = 9) and patients who did not (n = 12). A total of 273 genes were identified with |fold| >2 and p < 0.01 (Additional file 1: Table S9). To provide a large enough set of genes trending with under-expression in the recurrent patients, no multiple testing was implemented on these raw p-values. Further, the small sample sizes did not allow multiple testing corrections. Then genes within regions of DNA deletions in matched PTs and RTs in chromosome 4q (542 genes) and the 110 genes under-expressed (of the 273 genes both under- and over-expressed) in the PTs of HCC patients who experienced tumor recurrence (n = 9), relative to those who did not (n = 12; Additional file 1: Table S9) were intersected. From this analysis, the following genes were identified: NAA11, HERC5, DDX60, and HERC6 which were evaluated individually for association with tumor recurrence using the 21 Chinese patient primary tumors.

For each of the four genes, the 21 Chinese patients were categorized into high/low gene groups using the median fold change expression (Fig. 3a). Kaplan–Meier log-rank tests indicated a significantly shorter time to recurrence in the gene HERC5 low group (p = 0.007), while the other three genes did not significantly correlate with time to recurrence (p > 0.05), and after adjustment for HBV status pre-OLT, tumor grade, age, and gender, the trend remained for HERC5 (p = 0.07; HR = 7.29 CI95 = [0.85, 62.62]; Fig. 3a; “Patients and methods”).

Fig. 3
figure 3

a left HERC5 distribution of fold change values (log2 scale) for 21 Chinese patient primary tumors, blue line median of patients; right Kaplan–Meier (KM) curves comparing HERC5 low (n = 11) to high expression (n = 10) predicting recurrence; b left HERC5 distribution between normal liver (red; n = 239) and HCC tumors (green; n = 247; 19), blue line mean(normals)-2SD; middle Kaplan–Meier (KM) curves comparing HERC5 low (n = 62) to high expression (n = 180) predicting recurrence; right Same as middle predicting overall survival; c left HERC5 distribution for HCC tumors, blue line median of 65 patients (20); middle KM curves comparing HERC5 low (n = 20) to high expression (n = 24) predicting PFS; right Same as middle predicting overall survival. p log-rank test, p* Cox PH regression model, HR hazard ratio

Full size image

Two additional independent HCC datasets including primary tumors from patients who underwent resection were then evaluated [35, 36] where HERC5 expression was categorized into high/low patient groups using healthy control liver tissue [36] or the median of the HCC patient distribution (no normal liver tissue was available in this study) [35]. After adjustment for cirrhosis status, gender, HBV/HCV status and TNM staging, HERC5 was a significant predictor of overall survival (p = 0.004; HR = 2.02 CI95 = [1.26, 3.25]) and HCC recurrence (p = 0.004; HR = 1.80 CI95 = [1.20, 2.69]) in one dataset [36] with median time to recurrence in the HERC5 low group = 23 months versus HERC5 high group = 49 months (Fig. 3b; Table 2). Then in another independent HCC patient cohort [35], after adjustment for age, gender, and HBV status, HERC5 under-expression was associated with shorter overall survival (p = 0.02; HR = 3.31 CI95 = [1.22, 8.96]) and PFS (p = 0.01; HR = 3.80 CI95 = [1.38, 10.43]) (Fig. 3c; Table 2). To verify that the lack of correlation between the other three candidate genes alternative to HERC5 (i.e. NAA11, DDX60, and HERC6) and outcomes was not due to the small sample size used in the first patient dataset (n = 21), similar analyses were conducted on each gene independently using both larger HCC patient cohorts [35, 36] and no significant associations were observed for any of these three genes with any of the outcomes. An additional evaluation of specificity of the HERC5 correlation with prognosis was conducted among all genes that were located within the deleted region of chromosome 4q using study [36] (Additional file 2). A total of 262 genes were both within the chromosome 4q region and present on the microarray from this study, of which 118 were under-expressed in the HCC tumor biopsies compared to tissue from the normal controls. A multivariate Cox PH model was calculated independently on each of the 118 genes, where each gene was cut into high or low groups based on the same criteria that was described previously (mean-2 SD of the normal liver distribution). HAND2 had the most significant association with HCC recurrence, although it had a highly unbalanced distribution of high/low patients (3.6 % of patients in the low group). HERC5 was the second most significant correlate with HCC recurrence (Additional file 3: Figure S8), demonstrating biological specificity of this gene independent of other genes with CN deletions in chromosome 4q.

In a separate study of HCV-positive cirrhosis patients, molecular subgroupings of patients were identified and shown to correlate with good versus poor prognosis [41]. HERC5 was significantly over-expressed (p < 0.0001) in the good prognosis (n = 109) compared to poor prognosis group (n = 107) (Additional file 3: Figure S9).

HERC5 loss induces CCL20 mRNA and protein and associates with FOXP3 positive Treg infiltration in HCC

HERC5 loss was next evaluated for biological significance in HCC. By studying the whole transcriptome expression profile of the HCC cell line PLCPRF5 with HERC5 siRNA knockdown, CCL20 was the most over-expressed transcript (fold = 5.8), and genes coding classic regulators of CCL20 such as TNF, NFKB1, or TRIM32 showed no change (Additional file 1: Table S10, Additional file 2). TaqMan qRT-PCR and ELISA assays confirmed the overexpression of CCL20 at the transcript and protein level with two additional HCC cell lines (SNU449 and HEPG2; Fig. 4). The difference of under-expression in HERC5 in the recurrent patient PTs (mean = −0.3 fold) was significantly lower than that of PTs from patients who did not recur (mean = 1.1 fold; p = 0.001), confirming the results from the previous two array studies [35, 36] in this modest-sized study of Chinese patients. In addition, HERC5 and CCL20 mRNAs were significantly negatively correlated (p = 0.0003) in the PTs of HCC patients who experienced tumor recurrence and not in the PTs of HCC patients who did not experience tumor recurrence (p = 0.49; Fig. 5a). Immunostaining of FOXP3 in the primary tumors of the 21 HCC patients indicated significantly higher expression in the patients who experienced recurrence as compared to those who did not (Fig. 5c; p = 0.05).

Fig. 4
figure 4

Up-regulated CCL20 expression after HERC5 siRNA knockdown in HCC cell lines. a qRT-PCR analysis of HERC5 mRNA expression in three HCC cell lines 96 h after transfection of negative control siRNA and HERC5 siRNA. Shown is relative expression as copy numbers per 106 endogenous control genes (average of expression levels of ACTB, GAPDH, and UBC); b qRT-PCR analysis of CCL20 mRNA expression in three HCC cell lines 96 h after transfection of negative control siRNA and HERC5 siRNA; c Secreted CCL20 in HCC cell culture medium was measured by ELISA 96 h after siRNA transfection. Standard deviations are represented for each bar

Full size image
Fig. 5
figure 5

Negative correlation between HERC5 and CCL20 mRNA and the association of Tregs infiltration in PTs of recurrent patients. a Log2 fold changes HERC5 and CCL20 mRNA between and PNATs in the patients who experienced tumor recurrence (n = 9) or the patients who did not (n = 12). A Student’s paired t-test between HERC5 and CCL20 log2 fold change values for the recurrent patients have p = 0.0003 and p = 0.49 for the non-recurrent patients. Negative correlation between HERC5 and CCL20 mRNA is present in the recurrent patients and not observed in the patients who did not recurrent. b Examples of FOXP3 IHC in PTs and PNATs of a recurrent patient and a non-recurrent patient at ×20 magnification (530 × 460 µm in size per field). c FOXP3 IHC scores were calculated using the average numbers of FOXP3 positive stained lymphocytes in10 randomly selected fields at ×20 magnification (530 × 460 µm in size per field) in hepatocellular carcinoma for each sample. FOXP3 IHC score in PTs of recurrent patients (n = 9) were significantly higher than that in non-recurrent patients (n = 9) by Welch’s modified t-test (p = 0.05)

Full size image

Discussion

Our results describe the genetic and genomic heterogeneity between patients with HCC and demonstrate clonal persistence in tumor recurrence post-OLT. A unique study design inclusive of PTs and RTs (plus normal adjacent tissue and recipient blood), with an integration of somatic SNVs, CNVs, and transcript profiling and well-defined phenotypic spectrum allowed us to elucidate key cancer driver genes and delineate those critical gatekeepers of cancer initiation and progression. This study is the first to show a clonal advantage in RTs post-OLT compared to PTs in HCC, where Wnt/β-catenin signaling activation and tumor suppressor gene CDKN2A inactivation occurs in both tumors.

DNA CN loss at chromosome 4q was identified as one of the only two hemizygous deleted regions shared between the primary and recurrent tumors in 3 of 4 Chinese HCC patients. This region accounts for ~107 Mbp in length and similar deletions have been associated with either poor prognosis or advanced disease stages in pancreatic, colorectal, non-small cell lung cancer (NSCLC), and HCC tumors [42–46]. Within NSCLC specifically, FISH assays identified the primary region of 4q21.2-22.1 to be associated with poor prognosis [44, 46, 47], and a recent study from this same group showed that hypermethylation of HERC5 promoter (located at 4q22.1), and thus under-expression of the gene correlated with: positive disseminated tumor cells in the bone marrow, brain metastasis, and poor survival in both stage I adenocarcinoma and metastatic lung cancer patients [46]. Our results presented here in primary tumors of HCC patients are in agreement with these reports, underscoring the prognostic significance of HERC5 under-expression, as we have demonstrated with microarray or RNASeq technologies in primary tumors of three independent cohorts of HCC patients.

Cancer cells can avoid an immune response by disabling components of the immune system—a process well known as immune evasion. These cells can paralyze activated immune effector cells such as infiltrating cytotoxic T lymphocytes (CTLs) and natural killer (NK) cells by secreting TGF-β or other immunosuppressive factors [48, 49] such as the cell-cycle control tumor suppressor gene CDKN2A. However, the genetic basis behind recruitment of inflammatory cells that are actively immunosuppressive, such as regulatory T cells (Tregs) and myeloid-derived suppressor cells (MDSCs) is largely unknown. CCL20 has been shown to be the only chemokine significantly up-regulated in HCC tissue [50], where it can selectively recruit Tregs to the tumor, contributing to an immunosuppressive tumor microenvironment and leading to poor prognosis in HCC [51, 52], Both the genetic basis and mechanism for the source of highly secreted CCL20 in HCC still remains to be elucidated. Our results suggest that loss of HERC5 is the genetic precursor for CCL20 over-expression and associates with increased Treg infiltration in HCC, one of the underlying mechanisms of immune evasion in HCC. Further, pathway analysis of the top induced genes (|fold| >2, 129 up-regulated/78 down-regulated) following knockdown of HERC5 indicated activation of the IL17A pathway, as driven by over-expression of chemokine genes CXCL13, CXCL15, and CXCL16, with suppression of the Th2 cytokine IL-5, supported by the down-regulation of genes TMF1, PDIA6, ELL2 and APT1B1. Elevated serum IL17 levels in primary HCC patients have been shown to correlate with risk of tumor recurrence following curative hepatectomy [53], while suppressed IL5 mRNA expression was correlated with poor survival in cervical cancer [54]. The modulation of such pathways when HERC5 is silenced support an immunosuppressed environment for patients, thus influencing poor prognosis. HERC5 is an interferon-induced HECT-type E3 protein ligase that mediates type I interferon-induced ISGylation of protein targets; reduction in endogenous HERC5 blocks the IFN-induced ISG15 conjugation [55]. Thus host anti-viral responses are activated by the presence of HERC5 and reduced by its absence, suggesting an association with regulation of innate immune responses, a potential critical function leading to tumor recurrence in HCC. This hypothesis requires rigorous evaluation in future studies.

Unlike multiple transcriptomic studies that have developed gene signatures to predict tumor recurrence or survival in HCC [7–25], HERC5 was not solely identified as a statistical correlate with outcome, rather, this gene is located within a large somatic copy number deletion on chromosome 4q in both PTs and RTs and was suppressed in HCC patients who experienced tumor recurrence, compared to those who did not. Unfortunately in this study, HERC5 DNA loss was only evaluated in the 4 HCC patients with matched primary and recurrent tumor specimens, since DNA sequencing was not conducted on the remaining 17 HCC patients. Although, a recent study of 185 HCC patients showed that 28 % of patients’ primary tumors had a DNA deletion in HERC5, confirming a similar prevalence to patients with mRNA under-expression in this study (26 %; [56]; Additional file 3: Figure S9). Three independent studies with available comprehensive clinical and either RNASeq or microarray data were used to validate this clinical association totally 307 HCC patients. Taken together, these data confirm the robustness of our findings. HERC5 is also not present in any gene signature among the 14 catalogued in the Liverome database [21].

Studies have shown varied levels of consistency for classifying patients with high risk for HCC recurrence using clinicopathologic variables such as tumor size, vascular invasion, tumor state, tumor grade, and alpha-fetoprotein levels and data presented here suggest a single transcript as an important variable to couple to these relevant clinical factors for identifying HCC candidates for OLT. A single transcript also provides a pragmatic diagnostic assay for clinical use.

In a disease with no established adjuvant treatments, substantial shortage of donor organs, and high financial burden, identifying eligible patients for OLT with low recurrence risk at the molecular level accompanied by current clinical criteria, has potential to significantly improve patient clinical outcomes. This study is a first in HCC using comprehensive genetics and genomics patient profiling supported by large independent patient cohorts to provide evidence for such a biomarker and characterize the shared genetic drivers between primary and recurrent HCC. We believe that the translational study design and analytical strategy presented in this work will inspire other genetics studies beyond HCC recurrence and into other primary malignancies where metastases to other organs systems are observed.

Conclusion

This study provides evidence for a clonal selection advantage in the recurrent tumor, as compared to primary tumor in HCC patients and Wnt/β-catenin signaling was identified as the most regulated oncodriver pathway in matched primary and recurrent tumors. HERC5 was also identified as a prognostic biomarker for both survival and tumor recurrence in HCC patients in three independent HCC patient cohorts. Genetic- and genomic-driven under-expression of this gene is associated with CCL20 induction, suggested to increase Treg infiltration and ultimately poor prognosis in HCC patient post-OLT.

Abbreviations

HCC:

hepatocellular carcinoma

PT :

primary tumor

PNAT :

primary normal adjacent tissue

PB :

recipient blood

RT :

recurrent tumor

RNAT :

recurrent normal adjacent tissue

HBV:

hepatitis B virus

HCV:

hepatitis C virus

OLT:

orthotopic liver transplantation

WES:

whole exome sequencing

RNASeq:

RNA sequencing

Vf:

variant frequency

SNV:

single nucleotide variant

indel:

insertion or deletion

CNV:

copy number variation

HR:

hazard ratio

CI:

confidence interval

UCSF:

University of California San Francisco

CR:

clonal relationship value

WGS:

whole genome sequencing

Qphred:

Phred sequence value score

VCF:

variant call format

LRR:

Log2 R ratios

BAF:

B-allele frequency

CTLs:

cytotoxic T lymphocytes

NK:

natural killer

Tregs:

regulatory T cells

MDSCs:

myeloid-derived suppressor cells

HERC5:

HECT And RLD Domain Containing E3 Ubiquitin Protein Ligase 5

CCL20:

chemokine ligand 20

References

  1. El-Serag HB, Rudolph KL. Hepatocellular carcinoma: epidemiology and molecular carcinogenesis. Gastroenterology. 2007;132:2557–76.

    Article  CAS  PubMed  Google Scholar 

  2. Parkin DM. The global health burden of infection-associated cancers in the year 2002. Int J Cancer. 2006;118:3030–44.

    Article  CAS  PubMed  Google Scholar 

  3. Chang MH, You SL, Chen CJ, Liu CJ, Lee CM, Lin SM, et al. Decreased incidence of hepatocellular carcinoma in hepatitis B vaccines: a 20-year follow-up study. J Natl Cancer Inst. 2009;101:1348–55.

    Article  CAS  PubMed  Google Scholar 

  4. Zheng SS, Xu X, Wu J, Chen J, Wang WL, Zhang M, et al. Liver transplantation for hepatocellular carcinoma: Hangzhou experiences. Transplantation. 2008;85:1726–32.

    Article  CAS  PubMed  Google Scholar 

  5. Welker MW, Bechstein WO, Zeuzem S, Trojan J. Recurrent hepatocellular carcinoma after liver transplantation—an emerging clinical challenge. Transpl Int. 2013;26:109–18.

    Article  PubMed  Google Scholar 

  6. Zimmerman MA, Ghobrial RM, Tong MJ, Hiatt JR, Cameron AM, Hong J, Busuttil RW. Recurrence of hepatocellular carcinoma following liver transplantation: a review of preoperative and postoperative prognostic indicators. Arch Surg. 2008;143:182–8 (discussion 188).

    Article  PubMed  Google Scholar 

  7. Hoshida Y, Villanueva A, Kobayashi M, Peix J, Chiang DY, Camargo A, et al. Gene expression in fixed tissues and outcome in hepatocellular carcinoma. N Engl J Med. 2008;359:1995–2004.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  8. Iizuka N, Oka M, Yamada-Okabe H, Nishida M, Maeda Y, Mori N, et al. Oligonucleotide microarray for prediction of early intrahepatic recurrence of hepatocellular carcinoma after curative resection. Lancet. 2003;361:923–9.

    Article  CAS  PubMed  Google Scholar 

  9. Iizuka N, Tamesa T, Sakamoto K, Miyamoto T, Hamamoto Y, Oka M. Different molecular pathways determining extrahepatic and intrahepatic recurrences of hepatocellular carcinoma. Oncol Rep. 2006;16:1137–42.

    CAS  PubMed  Google Scholar 

  10. Kurokawa Y, Matoba R, Takemasa I, Nagano H, Dono K, Nakamori S, et al. Molecular-based prediction of early recurrence in hepatocellular carcinoma. J Hepatol. 2004;41:284–91.

    Article  CAS  PubMed  Google Scholar 

  11. Lee JS, Chu IS, Heo J, Calvisi DF, Sun Z, Roskams T, Durnez A, et al. Classification and prediction of survival in hepatocellular carcinoma by gene expression profiling. Hepatology. 2004;40:667–76.

    Article  CAS  PubMed  Google Scholar 

  12. Lee JS, Chu IS, Mikaelyan A, Calvisi DF, Heo J, Reddy JK, et al. Application of comparative functional genomics to identify best-fit mouse models to study human cancer. Nat Genet. 2004;36:1306–11.

    Article  CAS  PubMed  Google Scholar 

  13. Matoba K, Iizuka N, Gondo T, Ishihara T, Yamada-Okabe H, Tamesa T, et al. Tumor HLA-DR expression linked to early intrahepatic recurrence of hepatocellular carcinoma. Int J Cancer. 2005;115:231–40.

    Article  CAS  PubMed  Google Scholar 

  14. Okamoto M, Utsunomiya T, Wakiyama S, Hashimoto M, Fukuzawa K, Ezaki T, et al. Specific gene-expression profiles of noncancerous liver tissue predict the risk for multicentric occurrence of hepatocellular carcinoma in hepatitis C virus-positive patients. Ann Surg Oncol. 2006;13:947–54.

    Article  PubMed  Google Scholar 

  15. Tanaka S, Arii S, Yasen M, Mogushi K, Su NT, Zhao C, et al. Aurora kinase B is a predictive factor for the aggressive recurrence of hepatocellular carcinoma after curative hepatectomy. Br J Surg. 2008;95:611–9.

    Article  CAS  PubMed  Google Scholar 

  16. Tsuchiya M, Parker JS, Kono H, Matsuda M, Fujii H, Rusyn I. Gene expression in nontumoral liver tissue and recurrence-free survival in hepatitis C virus-positive hepatocellular carcinoma. Mol Cancer. 2010;9:74.

    Article  PubMed Central  PubMed  Google Scholar 

  17. Wang SM, Ooi LL, Hui KM. Identification and validation of a novel gene signature associated with the recurrence of human hepatocellular carcinoma. Clin Cancer Res. 2007;13:6275–83.

    Article  CAS  PubMed  Google Scholar 

  18. Woo HG, Park ES, Cheon JH, Kim JH, Lee JS, Park BJ, et al. Gene expression-based recurrence prediction of hepatitis B virus-related human hepatocellular carcinoma. Clin Cancer Res. 2008;14:2056–64.

    Article  CAS  PubMed  Google Scholar 

  19. Yokoo H, Kondo T, Okano T, Nakanishi K, Sakamoto M, Kosuge T, et al. Protein expression associated with early intrahepatic recurrence of hepatocellular carcinoma after curative surgery. Cancer Sci. 2007;98:665–73.

    Article  CAS  PubMed  Google Scholar 

  20. Yoshioka S, Takemasa I, Nagano H, Kittaka N, Noda T, Wada H, et al. Molecular prediction of early recurrence after resection of hepatocellular carcinoma. Eur J Cancer. 2009;45:881–9.

    Article  CAS  PubMed  Google Scholar 

  21. Lee L, Wang K, Li G, Xie Z, Wang Y, Xu J, et al. Liverome: a curated database of liver cancer-related gene signatures with self-contained context information. BMC Genomics 2011;12 Suppl 3:S3-2164-12-S3-S3.

  22. Kim JH, Sohn BH, Lee H-S, Kim S-B, Yoo JE, Park Y-Y, et al. Genomic Predictors for recurrence patterns of Hepatocellular Carcinoma: model derivation and validation. PLoS Med. 2014;11:e1001770.

    Article  PubMed Central  PubMed  Google Scholar 

  23. Zhang DY, Goossens N, Guo J, Tsai MC, Chou HI, Altunkaynak C, et al. A hepatic stellate cell gene expression signature associated with outcomes in hepatitis C cirrhosis and hepatocellular carcinoma after curative resection. Gut. 2015. doi:10.1136/gutjnl-2015-309655.

    Google Scholar 

  24. Piras-Straub K, Khairzada K, Trippler M, Baba HA, Kaiser GM, Paul A, et al. TRAIL expression levels in human hepatocellular carcinoma have implications for tumor growth, recurrence and survival. Int J Cancer. 2015;136:E154–60. doi:10.1002/ijc.29139 (Epub 2014 Aug 23).

    Article  CAS  PubMed  Google Scholar 

  25. Hashimoto R, Kanda M, Takami H, Shimizu D, Oya H, Hibino S, et al. Aberrant expression of melanoma-associated antigen-D2 serves as a prognostic indicator of hepatocellular carcinoma outcome following curative hepatectomy. Oncol Lett. 2015;9:1201–6.

    PubMed Central  PubMed  Google Scholar 

  26. Fujimoto A, Totoki Y, Abe T, Boroevich KA, Hosoda F, Nguyen HH, et al. Whole-genome sequencing of liver cancers identifies etiological influences on mutation patterns and recurrent mutations in chromatin regulators. Nat Genet. 2012;44:760–4.

    Article  CAS  PubMed  Google Scholar 

  27. Totoki Y, Tatsuno K, Yamamoto S, Arai Y, Hosoda F, Ishikawa S, et al. High-resolution characterization of a hepatocellular carcinoma genome. Nat Genet. 2011;43:464–9.

    Article  CAS  PubMed  Google Scholar 

  28. Kan Z, Zheng H, Liu X, Li S, Barber TD, Gong Z, et al. Whole-genome sequencing identifies recurrent mutations in hepatocellular carcinoma. Genome Res. 2013;23:1422–33.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  29. Li M, Zhao H, Zhang X, Wood LD, Anders RA, Choti MA, et al. Inactivating mutations of the chromatin remodeling gene ARID2 in hepatocellular carcinoma. Nat Genet. 2011;43:828–9.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  30. Guo G, Sun X, Chen C, Wu S, Huang P, Li Z, et al. Whole-genome and whole-exome sequencing of bladder cancer identifies frequent alterations in genes involved in sister chromatid cohesion and segregation. Nat Genet. 2013;45:1459–63.

    Article  CAS  PubMed  Google Scholar 

  31. Jiang Z, Jhunjhunwala S, Liu J, Haverty PM, Kennemer MI, Guan Y, et al. The effects of hepatitis B virus integration into the genomes of hepatocellular carcinoma patients. Genome Res. 2012;22:593–601.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  32. Kim D, Pertea G, Trapnell C, Pimentel H, Kelley R, Salzberg SL. TopHat2: accurate alignment of transcriptomes in the presence of insertions, deletions and gene fusions. Genome Biol. 2013;14(4):R36.

    Article  PubMed Central  PubMed  Google Scholar 

  33. Anders S, Huber W. Differential expression analysis for sequence count data. Genome Biol. 2010;11(10):R106.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  34. Carter SL, Cibulskis K, Helman E, McKenna A, Shen H, Zack T, et al. Absolute quantification of somatic DNA alterations in human cancer. Nat Biotechnol. 2012;30:413–21.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  35. Boyault S, Rickman DS, de Reynies A, Balabaud C, Rebouissou S, Jeannot E, et al. Transcriptome classification of HCC is related to gene alterations and to new therapeutic targets. Hepatology. 2007;45:42–52.

    Article  CAS  PubMed  Google Scholar 

  36. Roessler S, Jia HL, Budhu A, Forgues M, Ye QH, Lee JS, et al. A unique metastasis gene signature enables prediction of tumor relapse in early-stage hepatocellular carcinoma patients. Cancer Res. 2010;70:10202–12.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  37. Kim B, Woreta T, Chen PH, Limketkai B, Singer A, Dagher N, et al. Donor-transmitted malignancy in a liver transplant recipient: a case report and review of literature. Dig Dis Sci. 2013;58:1185–90.

    Article  PubMed  Google Scholar 

  38. Cankovic M, Linden MD, Zarbo RJ. Use of microsatellite analysis in detection of tumor lineage as a cause of death in a liver transplant patient. Arch Pathol Lab Med. 2006;130:529–32.

    PubMed  Google Scholar 

  39. Chen YJ, Yeh SH, Chen JT, et al. Chromosomal changes and clonality relationship between primary and recurrent hepatocellular carcinoma. Gastroenterology. 2000;119(2):431–40.

    Article  CAS  PubMed  Google Scholar 

  40. Myron KH, McBride MA, Cherikh WS, Spain PC, Marks WH, Roza AM. Transplant tumor registry: donor related malignancies. Transplantation. 2002;74:358–62.

    Article  Google Scholar 

  41. Hoshida Y, Villanueva A, Sangiovanni A, Sole M, Hur C, Andersson KL, et al. Prognostic gene expression signature for patients with hepatitis C-related early-stage cirrhosis. Gastroenterology. 2013;144:1024–30.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  42. Zhang H, Ma H, Wang Q, Chen M, Weng D, Wang H, et al. Analysis of loss of heterozygosity on chromosome 4q in hepatocellular carcinoma using high-throughput SNP array. Oncol Rep. 2010;23:445–55.

    Article  CAS  PubMed  Google Scholar 

  43. Danner BC, Gerdes JS, Jung K, Sander B, Enders C, Liersch T, et al. Comparison of chromosomal aberrations in primary colorectal carcinomas to their pulmonary metastases. Cancer Genet. 2011;204:122–8.

    Article  CAS  PubMed  Google Scholar 

  44. Wrage M, Ruosaari S, Eijk PP, Kaifi JT, Hollmén J, Yekebas EF, et al. Genomic profiles associated with early micrometastasis in lung cancer: relevance of 4q deletion. Clin Cancer Res. 2009;15:1566–74.

    Article  CAS  PubMed  Google Scholar 

  45. Luebke AM, Baudis M, Matthaei H, Vashist YK, Verde PE, Hosch SB, et al. Losses at chromosome 4q are associated with poor survival in operable ductal pancreatic adenocarcinoma. Pancreatology. 2012;12:16–22.

    Article  CAS  PubMed  Google Scholar 

  46. Wrage M, Hagmann W, Kemming D, Uzunoglu FG, Riethdorf S, Effenberger K, et al. Identification of HERC5 and its potential role in NSCLC progression. Int J Cancer. 2015;136:2264–72.

    Article  CAS  PubMed  Google Scholar 

  47. Uzunoglu FG, Dethlefsen E, Hanssen A, Wrage M, Deutsch L, Harms-Effenberger K, et al. Loss of 4q21.23-22.1 is a prognostic marker for disease free and overall survival in non-small cell lung cancer. PLoS One. 2014;9:e113315.

    Article  PubMed Central  PubMed  Google Scholar 

  48. Yang L, Pang Y, Moses HL. TGF-beta and immune cells: an important regulatory axis in the tumor microenvironment and progression. Trends Immunol. 2010;31:220–7.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  49. Shields JD, Kourtis IC, Tomei AA, Roberts JM, Swartz MA. Induction of lymphoidlike stroma and immune escape by tumors that express the chemokine CCL21. Science. 2010;328:749–52.

    Article  CAS  PubMed  Google Scholar 

  50. Rubie C, Frick VO, Wagner M, Rau B, Weber C, Kruse B, et al. Enhanced expression and clinical significance of CC-chemokine MIP-3 alpha in hepatocellular carcinoma. Scand J Immunol. 2006;63:468–77.

    Article  CAS  PubMed  Google Scholar 

  51. Huang F, Geng XP. Chemokines and hepatocellular carcinoma. World J Gastroenterol. 2010;16:1832–6.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  52. Ding X, Wang K, Wang H, Zhang G, Liu Y, Yang Q, et al. High expression of CCL20 is associated with poor prognosis in patients with hepatocellular carcinoma after curative resection. J Gastrointest Surg. 2012;16:828–36.

    Article  PubMed  Google Scholar 

  53. Wu J, Du J, Liu L, Li Q, Rong W, Wang L, et al. Elevated pretherapy serum IL17 in primary hepatocellular carcinoma patients correlate to increased risk of early recurrence after curative hepatectomy. PLoS One. 2012;7:e50035.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  54. Punt S, Houwing JJ, Schulkens IA, Thijssen VL, Osse EM, de Kroon CD, et al. A high IL6/IL17 ratio combined with low IL5 expression is correlated with poor survival in squamous cervical cancer. J Immunotherap Cancer. 2014;2:P232. doi:10.1186/2051-1426-2-S3-P232.

    Article  Google Scholar 

  55. Wong JJ, Pung YF, Sze NS, Chin KC. HERC5 is an IFN-induced HECT-type E3 protein ligase that mediates type I IFN-induced ISGylation of protein targets. Proc Natl Acad Sci USA. 2006;103:10735–40.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  56. Guichard C, Amaddeo G, Imbeaud S, Ladeiro Y, Pelletier L, Maad IB, et al. Integrated analysis of somatic mutations and focal copy-number changes identifies key genes and pathways in hepatocellular carcinoma. Nat Genet. 2012;44:694–8.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

Download references

Authors’ contributions

YY, FX, QX, JG, ZD, YG: study concept and design, acquisition of data, study supervision; BWH, JH, CM, WZ, XY, PB, YS, DS, MK: analysis and interpretation of data, statistical analysis; BWH, JH: drafting manuscript. All authors read and approved the final manuscript.

Competing interests

BWH, JH, CM, WZ, XY, PB, ZX, YS, ZL, DS, MK, ZD, HH, YG, YY: full time employees of MedImmune/Astra Zeneca and own stock in Astra Zeneca. FX, YX, JG, QX: no conflicts to declare.

Author information

Author notes

    Authors and Affiliations

    1. Department of Liver Surgery and Liver Transplantation, Renji Hospital, Shanghai Jiaotong University School of Medicine, Shanghai, 201203, China

      Feng Xue, Yun Xia & Qiang Xia

    2. Translational Bioinformatics, MedImmune, Gaithersburg, MD, 20878, USA

      Brandon W. Higgs, Jiaqi Huang, Chris Morehouse, Wei Zhu, Xin Yao, Philip Brohawn, Zhan Xiao, Yinong Sebastian, Zheng Liu, Dong Shen, Mike Kuziora & Yihong Yao

    3. Asia and Emerging Markets iMed, AstraZeneca R&D, 199 Liangjing Road, Shanghai, 201203, China

      Zhengwei Dong, Hulin Han & Yi Gu

    4. Shanghai Cancer Institute, Renji Hospital, Shanghai Jiaotong University School of Medicine, Shanghai, 201203, China

      Jianren Gu

    Authors
    1. Feng Xue
      View author publications

      You can also search for this author in PubMed Google Scholar

    2. Brandon W. Higgs
      View author publications

      You can also search for this author in PubMed Google Scholar

    3. Jiaqi Huang
      View author publications

      You can also search for this author in PubMed Google Scholar

    4. Chris Morehouse
      View author publications

      You can also search for this author in PubMed Google Scholar

    5. Wei Zhu
      View author publications

      You can also search for this author in PubMed Google Scholar

    6. Xin Yao
      View author publications

      You can also search for this author in PubMed Google Scholar

    7. Philip Brohawn
      View author publications

      You can also search for this author in PubMed Google Scholar

    8. Zhan Xiao
      View author publications

      You can also search for this author in PubMed Google Scholar

    9. Yinong Sebastian
      View author publications

      You can also search for this author in PubMed Google Scholar

    10. Zheng Liu
      View author publications

      You can also search for this author in PubMed Google Scholar

    11. Yun Xia
      View author publications

      You can also search for this author in PubMed Google Scholar

    12. Dong Shen
      View author publications

      You can also search for this author in PubMed Google Scholar

    13. Mike Kuziora
      View author publications

      You can also search for this author in PubMed Google Scholar

    14. Zhengwei Dong
      View author publications

      You can also search for this author in PubMed Google Scholar

    15. Hulin Han
      View author publications

      You can also search for this author in PubMed Google Scholar

    16. Yi Gu
      View author publications

      You can also search for this author in PubMed Google Scholar

    17. Jianren Gu
      View author publications

      You can also search for this author in PubMed Google Scholar

    18. Qiang Xia
      View author publications

      You can also search for this author in PubMed Google Scholar

    19. Yihong Yao
      View author publications

      You can also search for this author in PubMed Google Scholar

    Corresponding authors

    Correspondence to Brandon W. Higgs or Qiang Xia.

    Additional information

    Feng Xue, Brandon W. Higgs and Jiaqi Huang contributed equally to this work

    Additional files

    Rights and permissions

    Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

    Reprints and permissions

    About this article

    Check for updates. Verify currency and authenticity via CrossMark

    Cite this article

    Xue, F., Higgs, B.W., Huang, J. et al. HERC5 is a prognostic biomarker for post-liver transplant recurrent human hepatocellular carcinoma. J Transl Med 13, 379 (2015). https://doi.org/10.1186/s12967-015-0743-2

    Download citation

    • Received:

    • Accepted:

    • Published:

    • DOI: https://doi.org/10.1186/s12967-015-0743-2

    Keywords

    Journal of Translational Medicine

    ISSN: 1479-5876

    Contact us