Retraction Note: Hepatic stellate cell mediates transcription of TNFSF14 in hepatocellular carcinoma cells via H2S/CSE-JNK/JunB signaling pathway

Hepatic stellate cells (HSC) and hydrogen sulfide (H2S) both play important roles in the development of hepatocellar carcinoma (HCC). Whereas, in the microenvironment of HCC, whether HSC participate in regulating the biological process of HCC cells by releasing H2S remains elusive. In vitro, Flow cytometry (FCM), CCK-8, RNA-sequencing, Western blotting, RT-qPCR, immunofluorescence and ChIP assays were carried out in the HCC cells to investigate the effect of H2S on biological functions and JNK/JunB-TNFSF14 signaling pathway. Specimens from HCC patients were analyzed by RT-qPCR and Western blotting assays for evaluating the expression of TNFSF14 and CSE. Statistical analysis was used to analyze the correlation between TNFSF14 expression and clinical data of HCC patients. Based on the FCM and CCK-8 results, we found the LX-2 cells were able to induce HCC cells apoptosis through releasing H2S. RNA-sequencing, RT-qPCR, and Western blotting results showed that TNFSF14 gene was upregulated in both LX-2 and NaHS group. NaHS treated in HCC cells led to JNK/JunB signaling pathway activating and greater binding of p-JunB to its responsive elements on TNFSF14 promoter. Impairment of TNFSF14 induction alleviated LX-2 and NaHS induced apoptosis of HepG2 and PLC/PRF/5 cells. Furthermore, TNFSF14 expression in HCC tissues was lower than the adjacent tissue. HCC patients with low expression of TNFSF14 had higher malignant degree and poor prognosis. In summary, demonstration of the involvement of HSC-derived H2S in JNK/JunB mediated expression of TNFSF14 gene strongly indicates H2S palys an important role in the regulation of HCC apoptosis.


INTRODUCTION
Nowadays hepatocellular carcinoma (HCC) is the second most lethal cancer with persistently increasing mortality in the worldwide due to its highly complex and heterogeneous genetic aberrations [1,2]. More and more researches suggest that tumor microenvironment (TME) plays an important role in the processes of hepatocarcinogenesis, epithelial-mesenchymal transition (EMT), invasion, and metastasis [3]. TME comprises various cells, including hepatic stellate cells (HSC), liver stromal endothelial cells (LSEC), immune cells and cytokines like extracellular matrix (ECM) [4].
HSC are usually considered to foster the tumor microenvironment formation and contribute to a malignant tumor phenotype. Activated HSC contribute to hepatocarcinogenesis through the secretion of growth factors, and producing ECM and matrix metalloproteinase (MMPs) [5][6][7]. But, there are more and more studies implicated that HSC could play an anti-tumor role in HCC progression [8,9]. Hydrogen sulfide (H 2 S) is the most recently-discovered gas signal molecular, participated in regulating cell proliferation, apoptosis, oxidative stress, inflammation and angiogenesis [10]. H 2 S is mainly produced and cleared in the liver tissues by three enzymes: cystathionine-gamma-lyase (CSE), cystathionine-beta-synthase (CBS), and 3-mercapto-pyruvate sulfotransferase (MPST) [11]. Increasing evidence shows that H 2 S affect HCC through the regulation of cell proliferation [12], migration and invasion [13], angiogenesis and immune responses [14]. Intriguingly, our previous study found that exogenous H 2 S inhibit HCC progression through the PI3K/Akt/mTOR signaling pathway in vitro [15]. Our previous data found that activated HSC can produce H 2 S mainly regulated by CSE. Nowadays, the mechanisms involved in the capacity of HSC relased H 2 S to affect the HCC have not been worked out, all evidence up to this point suggests these mechanisms have great potential importance.
In this study, we investigate how activated HSC influence HCC fate. We found that activated HSC released H 2 S to induce HCC cell apoptosis. H 2 S induced cancer cell apoptosis through activating JNK/JunB signaling, which upregulated tumor necrosis factor superfamily member 14 (TNFSF14) expression.

MATERIALS AND METHODS Patients
A total of 45 patients with HCC who received hepatectomy were enrolled from Beijing You'an Hospital, affiliated to Capital Medical University from July 2010 to January 2013. HCC was diagnosed and graded according to the criteria of the American Association for the Study of Liver Diseases [36]. Liver samples were collected from cancerous and adjacent non-cancerous liver tissues and stored in liquid nitrogen for further experimental use. Clinic pathologic characteristics of these patients are shown in Table S1. The study protocol was approved by the Ethics Committee of Beijing You'an Hospital affiliated to Capital Medical University. The informed consent was obtained from all enrolled patients.

Cells
PLC/PRF/5 (ATCC® Cat. No. CRL-8024TM), HepG2 (ATCC® Cat. No. HB-8065TM) and LX-2 cells were cultured in DMEM medium supplemented with 10% FBS in a 37°C incubator with 5% CO 2 . The Transwell insert system was used for co-culture experiments to analyze the effect of LX-2 cells on HCC cells as previously described [37].

CCK-8
Cells proliferation was examined by a Cell Count KIT-8 assay (CCK8, AbMole, Houston, TX, America) according to the protocol. 2 × 10 3 cells with different treatments were pipetted into 96-well microplates for measurements. The absorbance values were measured under a wavelength of 450 nm on a Universal Microplate Reader (ELx800; Bio-Tek Instruments Inc, Winooski, VT, USA). The percentage of cell viability was calculated as: Cell proliferation % = [(A 450 sample-background)/ (A 450 control-backgroud)] × 100%.

Flow cytometry
Cell apoptosis was measured by FCM analysis using Annexin V-FITC/PI Apoptosis Detection Kit (40302ES20, Yeasen, Shanghai, China). Briefly, cells were collected and incubated with 5 µL Annexin V-FITC and 10 µL PI for 15 min at room temperature. Adjusted the volume to 500 µL with binding buffer. The apoptosis of HCC cells was analyzed with a FACScan-420 flow cytometer (BD Biosciences, Franklin Lakes, NJ, USA).

ELISA
The concentrations of H 2 S were examined by a Hydrogen sulfide Assay Kit (JEB-11780, Jin Yibai Biological Technology, Nanjing, China) according to the instruction.

RNA sequencing
RNA sequencing analysis was conducted by Novogene (Beijing, China). In brief, total RNA was isolated by TRIzol reagent (15596026, Thermo Fisher Scientific, Waltham, MA, USA). The RNA quantification and qualification were assessed using the RNA Nano 6000 Assay Kit of the Bioanalyzer 2100 system (Agilent Technologies, CA, USA). Sequencing of the libraries was carried out using the Illumina NovaSeq 6000 platform (Novogene, Beijing, China). All the raw data were uploaded on the Gene Expression Omnibus (GSE193660).

Quantitative real-time reverse transcription PCR (RT-qPCR)
Total RNA was isolated by TRIzol reagent and subsequently subjected to reverse transcription (4368814, High-Capacity cDNA Reverse Transcription Kits, Thermo Fisher Scientific, Waltham, MA, USA). RT-qPCR was performed using the TB Green (RR420A, Takara) on an ABI ViiA7 (GX-XVI R2, Applied Biosystems Life Tech, USA). The relative mRNA expression levels were calculated with 2 −ΔΔCt method and presented as mean foldchange of samples to control. The primers used in the study were listed in the Table 1.

Chromatin immunoprecipitation assay
The Chromatin immunoprecipitations (ChIP) assay was carried out according to a ChIP assay kit (#56383 CST). Cells were lysed with TNE buffer. Lysates were collected by centrifugation (15,000 g × 10 min) at 4°C. Aliquots were incubated with Protein G-Sepharose beads (GE Healthcare) for 1 h. After centrifugation (1000 g × 5 min), the supernatants were incubated with phospho-JunB (#8053 CST) and IgG(#2729 CST) antibodies overnight at 4°C. Next day, added Protein G-Sepharose beads for 8 h incubation. Then eluted the precipitate with the sample buffer. Finally, the immunoprecipitated chromatin was analyzed by PCR using primers targeting the human TNFSF14 promoter containing JunB binding sites. The primers used for ChIP-PCR are listed in Table 1.

Immunofluorescence staining
Cells were fixed in ice-cold 4% paraformaldehyde solution. Then incubated the cells with primary antibody (p-JunB) for overnight at 4°C in the wet box. Next day, the cells were washed and subsequently incubated with the secondary antibody conjugated with Alexa Fluor 488 for 1 h at room temperature. Cell nuclei were counterstained with DAPI. Cells were viewed and captured with a fluorescence microscope (Nikon Eclipse 80i; Nikon, Tokyo, Japan).

Transient transfection
The small interfering RNAs (siRNAs) technique and TNFSF14-expressing plasmid were used to evaluate the effect of TNFSF14 on H 2 S-induced HCC cell apoptosis. The siRNAs sequences were listed in the Table 1. Briefly, HCC cells  were seeded into the 6-well plates and cultured to the 80% confluence. SiRNAs and the plasmids were transfected to HCC cells according the instruction of Lipofectamine TM 3000 transfection reagent (L3000-015, Thermo Fisher Scientific, Waltham, MA, USA). After 24 h of transfection, the expression of TNFSF14 was evaluated by RT-qPCR and Western blotting.

Data analysis of public databases
The GSE14520 gene expression dataset was downloaded from the GEO database (http://www.ncbi.nlm.nih.gov/geo). According the annotation information in the platform, the probes were converted into the gene symbols. We divided the examined HCC patients into TNFSF14 high and low groups based on the median of the gene expression.

Statistical analysis
Each experiment was repeated at least three times independently. All data were presented as mean ± S.D. performed using the SPSS software (version 23.0). Differences between unpaired groups were compared by the Student's t-test or Mann-Whitney U test. For paired groups, the Wilcoxon

Activated hepatic stellate cells induce apoptosis of cancer cells
To examine how hepatic stellate cells (HSC) impact HCC cell behavior, we performed co-culture with activated HSC cell line LX-2 cells and HCC cell lines, HepG2 or PLC/PRF/5 cells. CCK-8 assays showed that cell viability of HCC cells was significantly reduced over time when the cells were co-cultured with LX-2 cells (Fig. 1A,  C). Flow cytometry analyses further demonstrated that co-culture with LX-2 cells induced apoptosis of HepG2 and PLC/PRF/5 cells.
The maximal apoptosis of these cancer cells was observed at 72 h following co-culture (Fig. 1B, D). These results suggest that activated HSC not only inhibit HCC cell proliferation, but also induce apoptosis of cancer cells.

LX-2 cells induce apoptosis of HCC cells through releasing H 2 S
How do activated HSC induce apoptosis of HCC cells? We hypothesized that H 2 S is a crucial mediator in the process given that activated HSC release H 2 S [16] and H 2 S is capable of inducing cancer cell apoptosis [15]. ELISA assay confirmed that LX-2 cells increased H 2 S release over time ( Fig. 2A). LX-2 cells released the same amount of H 2 S at 72 h as 10 -3 M NaHS did at 24 h (Fig. 2B). Subsequently, we observed the effects of NaHS, a donor of H 2 S, and DL-Propargylglycine (PPG), a CSE inhibitor, on LX-2 celldependent HCC cell apoptosis. As shown in Fig. S1, NaHS increased H 2 S release, whereas PPG dose-dependently inhibited mRNA and protein expression of CSE as well as H 2 S release in LX-2 cells. Flow cytometry analyses showed that NaHS resulted in HCC cell apoptosis (Fig. 2C, D). LX-2-dependent cancer cell apoptosis was significantly reduced by the administration of PPG (Fig. 2C, D). PPG alone did not affect HCC apoptosis (Fig. 2C, D). These results suggest that activated HSC result in HCC cell apoptosis through releasing H 2 S.

H 2 S upregulates pro-apoptotic factor TNFSF14 expression in HepG2 cells
To clarify how H 2 S results in HCC cell apoptosis, we performed RNA-seq in HepG2 cells, which were co-cultured with LX-2 cells or treated with NaHS. A heat-map shows different transcriptome cluster in cells with different administrations (Fig. 3A). LX-2 coculture and NaHS treatment respectively led to 1047 and 1183 gene upregulation as well 997 and 1144 gene downregulation in HepG2 cells. Among these altered genes, 659 up-regulated genes and 342 down-regulated genes were presented in HepG2 cells treated with either LX-2 co-culutre or NaHS (Fig. 3B). Based on the fold change, 30 genes with maximal fold changes were selected to further analyze (Fig. S2). Among these 30 genes, we identified a pro-apoptotic gene TNFSF14 (Fig. 3C). We examined TNFSF14 mRNA and protein in HepG2 and PLC/PRF/5 cells treated with different concentrations of NaHS. NaHS dose-dependently increased expression of TNFSF14 and reached a maximum at 10 -3 M treated 24 h in both HCC cells (Fig. 3D-G). Furthermore, we observed that TNFSF14 expression was up-regulated when HCC cells were co-cultured with LX-2 cells (Fig. 3H, I). The effect was inhibited by PPG administration (Fig. 3H, I). These results imply that TNFSF14 might be the key factor mediating H 2 S-induced HCC apoptosis.

TNFSF14 is indispensable for H 2 S-induced HCC cell apoptosis
To confirm the role of TNFSF14 in H 2 S-induced HCC cell apoptosis, we knocked down TNFSF14 expression in both HepG2 and PLC/ PRF/5. RT-qPCR and Western blotting analyses confirmed that NaHS and LX-2 co-culture-induced TNFSF14 mRNA and protein expression was inhibited by siRNA-TNFSF14 in both HepG2 (Fig. 4A, B) and PLC/PRF/5 cells (Fig. 4D, E). Flow cytometric analyses further showed that NaHS and LX-2 co-culture-induced apoptosis of cancer cells was significantly inhibited when TNFSF14 expression was knocked down (Fig. 4C, F). On the other hand, the effect of TNFSF14 was also confirmed through transfection of the tnfsf14 gene into HCC cells, which showed that transfection with pcDNA3.1-tnfsf14 results in apparent increase of apoptosis in HepG2 and PLC/PRF/5 cells with different treatments (Fig. S3).
These results suggest that TNFSF14 expression is required for H 2 S-induced HCC cell apoptosis.
H 2 S regulates TNFSF14 transcription through JNK/JunB signal To clarify how H 2 S regulates the expression of TNFSF14, we analyzed PROMO dataset and adopted JASPAR software to predict potential transcript factors binding to the TNFSF14 promoter. JunB got the highest scored among all the analyzed factors (Fig. 5A upper panel) and possessed multiple binding sites on the TNFSF14 promoter ( Fig.5A lower panel). ChIP assays confirmed the binding of p-JunB to the TNFSF14 promoter in both HepG2 and PLC/PRF/5 cells (Fig. 5B). The binding was increased by NaHS treatment (Fig.  5B). Immunofluorescence staining and Western blotting showed that administration of NaHS remarkably induced p-JunB expression and nuclear translocation in both HCC cell lines (Fig. 5C-D). Besides NaHS, LX-2 cell co-culture also increased p-JunB expression and nuclear translocation (Fig. 5E). JunB phosphorylation and nuclear translocation required upstream p-JNK. Administration of SP600125, a p-JNK inhibitor, completely inhibited NaHS-or LX-2 cell co-culture-induced p-JunB expression and nuclear translocation in both cancer cells (Fig. 5F). The effects of NaHS-or LX-2 cell co-culture-induced p-JunB expression and nuclear translocation was also inhibited by PPG (Fig. 5F). These results suggest that JNK/ JunB signaling mediates H 2 S to regulate TNFSF14 transcription.
TNFSF14 expression is associated with the clinical outcome of HCC Subsequently, we observed the association between TNFSF14 expression and clinical outcome of 45 HCC patients. TNFSF14 and CSE mRNA expression were significantly downregulated in tumor tissues compared with surrounding non-tumor tissues (Fig. 6A). Western blotting analyses performed in 6 patients confirmed the results (Fig. 6B). CSE mRNA expressions in tumor tissues were positively correlated with that of TNFSF14 (r = 0.438, P < 0.05, Fig.  6C). We further investigated the association between TFNSF14 expression and tumor characteristics. Patients with high TNFSF14 expression in HCC showed less portal invasion, lower Barcelona Clinic Liver Cancer (BCLC) HCC stages and longer survival and disease-free time than those with low TNFSF14 expression (Table 2 and Fig. 6D, E). These results were similar as the data in GSE 14520: TNFSF14 and CSE gene expression were decreased in tumor tissues compared with non-tumor tissues (Fig. S4). CSE gene expressions in HCC tissues were positively associated with TNFSF14 expression (r = 0.168, P <0.05, Fig. S4). In addition, patients with high TNFSF14 expression showed less alphafetoprotein (AFP) levels, lower ALT, lower BCLC HCC stages, and lower metastasis risk than those with low TNFSF14 expression (P < 0.05, Fig. S4). The patients with high TNFSF14 expression had longer survival and disease-free time (Fig. S4).

DISCUSSION
HSC, as the central modulators of the tumor microenvironment, have long been considered as a tumor promoter during hepatocarcinogenesis by the secretion of growth factors such as hepatocyte growth factor (HGF), platelet-derived growth factor (PDGF) and connective tissue growth factor (CTGF), altering components of the ECM, regulating angiogenesis and immune surveillance [17][18][19][20][21]. In stark contrast, HSC are a source of transforming growth factor beta (TGF-β), which exert tumorsuppressing activity by inducing cytostasis and apoptosis of hepatocytes in early tumor phases [8]. Recently, HSC were reported to be negative regulators of HCC progression through upregulating endosialin which is capable of inhibiting tumorpromoting cytokines, including IGF2, RBP4, DKK1, and CCL5 [9]. The findings of our study shed further insight into the increasing appreciation that HSC do not just act as pro-tumorigenic role, but may inhibit tumor progression. In our study, we found that HSC cells co-cultured with HepG2 and PLC/PRF/5 cells could promote cell apoptosis which was time-dependently manner. Further research found that HSC cells through releasing H 2 S induced HCC cells apoptosis, which was relieved by adding PPG (CSE inhibitor) to reduce the amount of H 2 S. As a gas signaling molecule, the role of H 2 S in HCC is still unclear. The previous reports shown that in the HepG2 and PLC/ PRF/5 cells, endogenous H 2 S/CSE pathway regulated cell growth mainly through EGFR pathway [12]. Another similar study found that in the HepG2 cells, H 2 S/CSE pathway was activated after irradiated, which led to the long-term cell invasion and tumor metastasis [22]. The function of exogenous H 2 S has also been studied in HCC cells. In the PLC/PRF/5 cells, 500 μM NaHS could induce cell proliferation, migration, and anti-apoptotic via activating NF-κB pathway [13]. Other researchers found that exogenous H 2 S effectively restricted the tumor development in H22 HCCbearing mice and HCC cell lines proliferation by blocking STAT3 and NF-κB pathways [14,23]. Similar with our pervious study, 10 −3 M NaHS treated HepG2 and HLE cells could inhibit cells proliferation, migration and pro-apoptotic via mTOR signaling pathway [15]. In this study, our results showed that both endogenous and exogenous H 2 S promoted HCC cells apoptosis, and the apoptosis was reduced when endogenous H 2 S was inhibited by adding PPG.
How do H 2 S regulate HCC cells apoptosis? Based on the RNAsequencing results, we found that both endogenous and exogenous H 2 S could up-regulate TNFSF14 in HepG2 cell. TNFSF14, mainly expressed on activated T cells, is activated in Natural Killer (NK) cells, and immature dendritic cells (DC) [24]. Previous studies reported that TNFSF14 act as an immunomodulatory factor to induce tumor regressions and apoptosis through enhancing T cells proliferation, differentiation, and cytokines secretion [25][26][27]. In human epithelial cancer cell lines, TNFSF14 promotes apoptosis and inhibits proliferation, which effect is independent of T cells [28][29][30][31]. However, there are few reports about the effect of TNFSF14 in HCC. In our study, we found that H 2 S induced HCC apoptosis partly due to the TNFSF14 gene. TNFSF14 mRNA and protein expression levels was upregulated in HepG2 and PLC/PRF/5 cells under H 2 S treatment. Knockdown of TNFSF14 reduced H 2 S-induced HCC apoptosis.
To figure out the precise molecular mechanism of H 2 S regulating the transcription of TNFSF14 gene, the transcription  (upper) and JunB binding sites on TNFSF14 promoter predicted with JASPAR website (lower). B ChIP assays for JunB binding onto TNFSF14 promoter were carried out in HepG2 (upper) and PLC/PRF/5 (lower) cells. The right panel showed the statistical analysis for the ChIP assay. C NaHS was used to treat HepG2 (left) and PLC/PRF/5 (right) cells for 0, 1, 2, 8, and 24 h. Expression of p-JunB in nuclear of HCC cells was detecting using immunofluorescence staining, and Nuclei and p-JunB were stained with DAPI (blue) and FITC (green), respectively. The image is representative of three independent experiments. D The effects of different time of NaHS treatment on the expression of p-JunB in the nuclear and cytoplasm were analyzed with Western blotting in the HepG2 cells and PLC/PRF/5 cells. E The expression levels of p-JunB protein were detected by Western blotting in the HCC cells under exogenous and endogenous H 2 S treatment. F Western blotting for the effects of PPG (CSE inhibitor) and SP600125 (p-JNK inhibitor) in the expressions of JNK/p-JNK, JunB/p-JunB, and TNFSF14 proteins in HepG2 and PLC/PRF/5 cells. All Bars, ± SD; n = 3; *P < 0.05 compared with control group in HepG2 cells, # P < 0.05 compared with control group in PLC/PRF/5 cells.
regulatory region of TNFSF14 gene was analyzed on the JASPAR website. We found transcription factor JunB binds to the TNFSF14 gene. The transcriptional activity of JunB is enhanced through phosphorylation-JNK [32]. In HCC cells, JNK/JunB exerts anti-tumor effects by promoting tumor cell apoptosis and inhibiting cell growth [33]. In this study, we found that H 2 S regulated TNFSF14 transcription mainly through activating JNK/JunB singal in both HepG2 and PLC/PRF/5 cells.
All these in vitro results demonstrated that H 2 S, released by LX-2 cells, participated in the JNK/JunB-TNFSF14 pathway to promote HCC cell apoptosis. Further validation was performed in the GEO dataset and our cohort, the results showed that patients with high CSE and TNFSF14 expression in HCC showed less portal invasion, lower BCLC stages and longer survival and disease-free time compared with lower TNFSF14 expression patients. Similar with our results was found in the primary colorectal cancer patients, TNFSF14 expression was decreased on tumor infiltrating lymphocytes in resected colorectal cancer liver metastases patients and significantly associated with overall survival and recurrence-free survival [34]. But in the nonalcoholic steatohepatitis (NASH)related HCC patients, NKT cells secreted TNFSF14 to promote NASH-to-HCC transition [35]. These discrepant researches suggested that TNFSF14 may play different roles in different liver diseases.
It is presently unclear how significant the impact of H 2 S mediated TNFSF14 expression through JNK/JunB pathway on the occurrence and high mortality of HCC. However, taking into consideration the effects of H 2 S in the HCC, it is reasonable to speculate that H 2 S has great potential significance and importance in HCC apoptosis. Taken our results together, HSC released H 2 S could upregulate the TNFSF14 expression through JNK/JunB axis to induce HCC cells apoptosis, and TNFSF14 was significantly associated with prognosis of HCC patients. Thus, enhance the circulating H 2 S to promote cells apoptosis may be a potential strategy for HCC. Nonetheless, further clinical study of this approach will be required.

DATA AVAILABILITY
The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.  N = 45). D, E Overall survival and disease-free survival curves showed that patients with high expression levels of TNFSF14 have better prognosis. All Bars, ± SD; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.