Background
Liver cancer is the second leading cause of cancer death for men in less developed countries. In more developed countries, it is the sixth leading cause of cancer death among men [
1] and hepatocellular carcinoma (HCC) is the most prevalent and malignant type of liver cancer. Radiofrequency ablation (RFA) is commonly used to treat nonresectable and small liver tumors (≤3 cm) and may provide tumor clearance and increase quality of life. However, clinical research results indicate that local recurrence after RFA of liver tumors varies between 2 to 60% [
2‐
4], which is greater than after liver resection. The literature is almost unanimous that local hepatic tumor recurrence after RFA is sooner than after liver resection, why this occurs is not clear.
RFA temperature distribution effectively divides the treatment area into an application, central, transition, and reference zones. For the application and central zones, 60 °C is applied and causes immediate necrosis [
5]. Transition zone temperature is 42–60 °C, which is insufficient to kill tumor cells, so some cells survive and can form subsequent tumors. The reference area has no effect on tumor cells.
After RFA, HCC cell biological behavior is modified, and sub-lethal RFA may actually increased malignant transformation of HCC [
6] and facilitate rapid progression of residual hepatic VX2 carcinoma [
7]. Also, rapid progression of residual tumors via hypoxia inducible factor-1α (HIF-1α)/vascular endothelial growth factor A (VEGFA) pathways can occur and epithelial-mesenchymal transition (EMT) markers have been reported to be expressed at these recurrence sites after RFA treatment [
8]. Also, insufficient RFA promotes EMT of HCC cells [
9] and malignancy [
10]. After insufficient RFA, tumor-associated endothelial cells have enhanced angiogenesis and invasiveness of residual HCC [
11], and this may promote tumor angiogenesis via HIF-1α/VEGFA pathway [
12]. Incomplete RFA stimulates proliferation of residual renal carcinoma cells [
13] and enhances invasiveness and metastasis of residual HCC cancers [
14].
Long noncoding RNA (lncRNA) is generally defined as a transcript larger than 200 nucleotides that lacks protein-coding functions. Recently, lncRNAs have been shown to have critical regulatory roles in cancer biology, including genetic imprinting, immune response, tumorigenesis, cellular development, and metabolism [
15,
16]. HCC lncRNA expression measured with microarray confirmed that hundreds of lncRNAs were abnormally expressed in HCC tissues [
17]. Furthermore, invasion- and metastasis-related lncRNA of HCC have also been identified [
18]. However, the role of lncRNA in biological behavioral changes in residual HCC cells in the transition zone of RFA treatment is still unclear.
In the present work, cultured SMMC-7721 cells were subjected to hyperthermia (50 °C) for 10 min to mimic the transition zone produced by RFA, and this is an accepted model at this time [
10,
11]. Then, we measured changes in lncRNA expression in HCC cells and estimated the biological significance of these lncRNAs.
Methods
Cell culture
Human SMMC-7721, HepG2, and MHCC97-H HCC cells were obtained from the Cell Bank of the Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences (Shanghai, China). Cells were maintained in high-glucose Dulbecco’s modified Eagle’s medium (DMEM) (GIBCO, Invitrogen, Australia) supplemented with 10% fetal bovine serum (FBS) (GIBCO, Invitrogen) and 1% penicillin-streptomycin (GIBCO, Invitrogen) at 37 °C in a 5% CO2 atmosphere.
Sub-lethal heat treatment
SMMC-7721, HepG2, and MHCC97-H cells were sub-lethally heated (50 °C) for 10 min. Heat treatments were carried out by sealing the culture bottle with Parafilm, and submerging the plates in a water bath (HH · W21 · 600S, Shanghai, China) set at 50 °C and returned to the incubation chamber (Series II Water Jacket, Thermo-Scientific, Waltham, MA) for 24 h at 37 °C. Three independent experiments were performed.
Total RNA was extracted from sub-lethally heat-treated HCC cells using RNAiso Plus Reagent (TaKaRa, Dalian, China) according to the manufacturer’s instructions. RNA quantification and quality were assured with a NanoDrop ND-2000 spectrophotometer (Thermo-Scientific). RNA integrity and gDNA contamination were confirmed with agarose gel electrophoresis (Additional file
1: Figure S1).
Microarray analysis
Sample preparation and microarray hybridization and assay were performed by CapitalBio Corporation (Beijing, China). Briefly, RNA was purified with mirVana miRNA Isolation Kit (Ambion, Austin, TX) according to the manufacture’s protocol. cDNA labeled with a fluorescent dye (Cy5 and Cy3-dCTP) was produced using Eberwine’s linear RNA amplification method and subsequent enzymatic reaction as described [
19] with improvements using a CapitalBio cRNA Amplification and Labeling Kit (CapitalBio) for greater yields of labeled cDNA. cRNA amplification and labeling is depicted in Additional file
2: Figure S2. lncRNA + mRNA array data were analyzed for summarization, normalization and quality control with GeneSpring software V13.0 (Agilent). To select differentially expressed genes, we used threshold values of ≥2 and ≤ −2-fold changes and a Benjamini-Hochberg corrected
p value were <0.05 (multiple testing, Benjamini-Hochberg method). Data were Log2-transformed and median centered using genes and the Adjust Data function of CLUSTER 3.0 software then analyzed with hierarchical clustering with average linkages. Finally, we performed tree visualization using Java Treeview (Stanford University School of Medicine, Stanford, CA).
Construction of the lncRNA-mRNA gene co-expression network
The lncRNA-mRNA co-expression network was constructed based on the correlation between differentially expressed lncRNAs and mRNAs. For each gene pair, a Pearson correlation coefficient was calculated and significant correlation pairs were selected to construct the network. lncRNAs and mRNAs with coefficients
>0.99 were selected for network design using the open-source bioinformatics software Cytoscape. In a network analysis, a degree centrality is defined as the links one node has with other nodes. A degree is the simplest and most important measures of a gene centrality within a network and this establishes relative importance [
20]. A yellow ellipse represents selected seven up-regulated lncRNAs and a green ellipse represents mRNAs. Red lines represent positive correlations and blue lines represent negative correlations.
Gene ontology and pathway analysis
KOBAS 2.0 (KEGG Orthology Based Annotation System) was used to perform GO and pathway analysis. Its purpose is to identify significantly enriched pathways and diseases for a set of genes or proteins, using pathway and disease information from multiple databases. In the present study, p < 0.05 was a threshold to eliminate pathways related to sub-lethally heat treated HCC cells.
Liver specific lncRNA analysis
Quantitative reverse transcription PCR (qRT-PCR) validation
Total RNA was reverse-transcribed using a PrimeScript RT reagent Kit with gDNA Eraser (Perfect Real Time) (RR047A, TaKaRa, Dalian, China) with a GeneAmp PCR system 2700 (Applied Biosystems, Singapore). Real-time PCR amplification was used to measure lncRNA and mRNA with SYBR Premix Ex Taq II (RR820A, TaKaRa, Dalian, China) with a CFX96 Real-Time System (C1000 Thermal Cycler, Bio-Rad, Singapore) according to the manufacturer’s protocol. Amplification conditions were 95 °C for 30 s, followed by 40 cycles of 95 °C for 5 s and 60 °C for 30 s or 95 °C for 5 s, 55 °C for 30 s and 72 °C for 30 s, with a final extension at 65 °C for 5 s. Primers used are shown in Additional file
3: Tables S1 and Additional file
4: S2. GAPDH was used as a control. lncRNA and mRNA were calculated using the formula 2
−ΔΔCt, as previously reported [
21].
Statistical analysis
All assays were repeated a minimum of three times. Data were analyzed using SPSS 17.0 software. All quantitative data were expressed as means ± standard deviations (SD). Gene expression in sub-lethally heat-treated HCC cells was compared to non-treated HCC cells. A Student’s t-test was used to compare data between the two groups (p < 0.05 was considered a statistically significant difference).
Discussion
RFA is commonly used for treating HCC and this usually produces a transition zone in which the temperature does not achieve therapeutic efficacy, so it is sub-lethal to these cells. These residual tumor cells can return and rapidly produce subsequent tumors, representing a significant drawback for RFA. Recently, research suggests that lncRNAs are important in complicated diseases such as cancer [
22]. Several lncRNA have been identified as being involved in the development and progression of HCC, such as lncRNA-HEIH [
23], HOTAIR [
24,
25], lncRNA MVIH [
26], lncRNA-Dreh [
27], HOTTIP/HOXA13 [
28,
29], URHC [
30], PCNA-AS1 [
31], UFC1 [
32], HULC [
33], CCAT1 [
34], PVT1 [
35], ANRIL [
36,
37], ZEB1-AS1 [
38], PANDAR [
39], DANCR [
40]. Some of these lncRNA were also identified in our microarray data. For example, PVT1 and DANCR were upregulated in sub-lethally heat-treated HCC cells. Research indicates that BC017743, ENST00000395084, NR_026591, NR_015378, and NR_024284 were up-regulated, whereas NR_027151, AK056988, and uc003yqb.1 were down-regulated in HCC samples and adjacent non-tumor tissues according to microarray analysis [
17]. The relationship between lncRNAs and HCC cell invasion and metastasis was identified using the same method [
18]. Long noncoding RNA expression in TGF-β2-induced epithelial-mesenchymal transition has been identified [
41] and investigators have reviewed the important role for lncRNA in hepatitis B virus-induced liver cancer [
42], but few studies exist to describe changes in lncRNA in sub-lethally heat-treated HCC cells that recover after treatment.
Here, we report that 782 lncRNA and 1281 mRNA were significantly differentially expressed compared to untreated SMMC-7721 cells. Bioinformatic analyses confirmed that differentially expressed lncRNAs and mRNAs using GO, as well as pathway and co-expression network analysis had value for target and transcript factor prediction. Subsequent analysis of hepatic-specific lncRNAs and validation of microarray results by qRT-PCR were also performed. Data from GO and pathway analysis indicated that differentially expressed mRNAs and lncRNAs are related to metabolic pathways, indicating that metabolic pathways in sub-lethally heat-treated HCC cells undergo energy changes that allow adaptation and survival. We uploaded these data to the GEO web site (accession # GSE99351).
We selected seven lncRNA and seven mRNA for qRT-PCR analysis and these data are consistent with microarray data. But the mRNA expression is not the same in the three HCC cells, probably because different cell lines have different biological behaviors and expression profiles. Then, we constructed lncRNA-mRNA co-expression networks for seven lncRNAs. ENST00000570843.1 expression was the most changed, so in the future, we will study its role in proliferation and recurrence of residual cancer cells after RFA treatment.
We previously reported that RFA combined with sorafenib can prolong patient survival [
43]. However, chemo-resistance is common and influences patient prognosis. Here, we report [
44] data for lncRNA expression in chemo-resistant HCC cells in which we identified 120 differentially expressed lncRNAs, of which 61 were up-regulated and 59 were down-regulated. The underlying pathways of these differences in expression were related to cell death, proliferation, and cellular response to stimuli, and these included the p53, ErbB, and MAPK pathways. There is an interest in using lncRNAs as biomarkers of cancer as some investigators report that lncRNA expression data obtained with microarray from plasma of HCC and chronic hepatitis B patients had markers that may help diagnose HCC. Specifically, they identified lncRNA-urothelial carcinoma associated-1 (lncRNA-UCA1) and WD repeat containing antisense to TP53 (WRAP53) as novel biomarkers for noninvasive diagnosis of HCC [
45]. Serum expression of uc001ncr and AX800134 may also be a biomarker for diagnosing HCC, especially for patients with AFP
<400 ng/ml or with early-stage disease (BCLC 0 + A) [
46]. Thus, lncRNA may be novel therapeutic targets for cancer [
47].
Our choice of 50 °C heat applied for 10 min was selected based on observations that this best simulated the RFA transition zone when treating HCC. Also, tumor cells did not die after this heat application but their proliferation and invasiveness was enhanced after culture. This may cause more rapid tumor recurrence compared to surgical resection. The timing of 10 min best approximates clinical RFA applications. Finally, this heat applied at this time did not kill all cancer cells, but they recovered after 3 weeks (data not shown). After application of 47 °C for 10 min, cancer cells did not have altered morphology and did not change significantly after a lengthy culture. Therefore, our temperature and duration of application may best replicate heat stress experienced by residual cancer cells in the RFA transition zone.
Study limitations include sub-lethal RFA and lncRNA expression measurements after a short interval. lncRNA expression changes may not occur this rapidly in clinical specimens. In addition, we could not obtain HCC tissue specimens during or after RFA treatment because HCC biopsies are controversial in our institution (rarely produced). Therefore, HCC tissue specimens were not used for microarray analysis. So, we will quantify ENST00000570843.1 in blood samples (see above). In summary, we described global expression profiles of lncRNAs and mRNAs in sub-lethally heat-treated HCC cells using lncRNA and mRNA microarrays. With bioinformatics predictions, we identified target genes that are correlated with differentiation of seven candidate lncRNAs. Further studies of functional analysis of these lncRNA are needed to provide more conclusive evidence about their regulatory mechanisms. Collectively, our results suggest that lncRNAs are important regulators of transition zone tumor recurrence and may guide further investigation into mechanisms of lncRNAs that regulate residual cancer cell return after RFA treatment.