Mutant Kras and mTOR crosstalk drives hepatocellular carcinoma development via PEG3/STAT3/BEX2 signaling

Background & Aims: Abnormal activation of mTOR through loss of tuberous sclerosis complex (Tsc) frequently occurs in hepatocellular carcinoma (HCC). Mutant Kras could induce aggressive HCCs. Here, we aim to identify the predictive or prognostic biomarkers for HCC patients with Kras mutant and mTOR hyperactivation, and to provide potential therapeutic approaches for this subtype of HCCs. Methods: We generated transgenic mice in which hepatocytic mTOR was hyperactivated through Tsc1 insufficiency with or without oncogenic KrasG12D. Bioinformatics and gain- or loss-of-function studies were used to illustrate the mechanisms underlying oncogenic pathway alterations. Transcriptional profiling was used to identify biomarker for the subtype of HCC. The therapeutic efficacy of targeting mTOR was tested in a liver orthotropic homogeneous murine model. Results: Oncogenic KrasG12D facilitated mTOR activation via the Mek/Erk/ROS axis, leading to HCC tumorigenesis and metastasis. Inhibition of Mek/Erk enhanced the anticancer effect of mTOR inhibitor via reduction of mTOR activity. Paternally expressed 3 (PEG3) was responsible for Kras/Erk- and mTOR-driven HCC. Elevated PEG3 protein interacted with STAT3 and promoted its transcriptional activity, resulting in the upregulation of proliferation- and metastasis-related proteins. Targeting mTOR significantly inhibited these actions in vitro and in vivo. Moreover, in clinical samples, PEG3 was identified as a new poor prognostic marker for HCC patients with Kras/Erk and mTOR hyperactivation. Conclusion: These findings reveal the underlying mechanism of hepatocytic Kras/Erk-driven mTOR activation and its downstream targets (PEG3 and STAT3) in HCC, identify PEG3 as a new prognostic biomarker for HCC with Kras/Erk and mTOR hyperactivation, and provide a potential therapeutic strategy for this subset of HCC patients.


Orthotopic tumorigenesis
First, primary liver cells were successfully isolated from Kras G12D ;Tsc1 fl/fl ;Alb-Cre mice. Then, 1 × 10 7 cells were suspended in 100 µl PBS/Matrigel (1:1) and transplanted to the hypodermis of 6-week-old male BALB/c-nude mice (GemPharmatech, Chengdu, China), which were first anesthetized by isoflurane. After two weeks, the nude mice were killed with an excessive anesthetic dose, and the tumor tissue was removed subcutaneously. The tumor was cut into 1 mm 3 pieces and retransplanted into the livers of 32 six-week-old male nude mice that were anesthetized and underwent laparotomy. All mice were were randomly divided into three groups. Thus, orthotopic tumorigenesis was finished [2]. One day after orthotopic tumorigenesis, rapamycin and sapanisertib treatment assays were conducted. Rapamycin and sapanisertib were reconstituted in ethanol at 20 mg/ml and 5 mg/ml, respectively, and then diluted in 5% Tween-80 and 5% PEG-400 to working concentrations of 4 mg/kg and 1 mg/kg, respectively. Mice were treated with rapamycin (4 mg/kg, i.p.) and sapanisertib (1 mg/kg, i.p.) every other day for 3 weeks [3][4][5]. Finally, nude mice were killed with an excessive anesthetic dose, and tumor volumes and tumor weights were evaluated. Experimenters conducting murine model construction and drug treatment and Experimenters conducting mouse death and data collection were two different groups of experimentalists and followed the blinding principle. All experiments were approved by the Institutional Animal Care and Use Committees of Army Medical University.

Lentivirus infection
KTC primary cells were successfully infected by sh-Cont or sh-Peg3 lentivirus according to the manufacturer's specification. 2 × 10 7 lentivirus and 100 µl HiTranG A were added on KTC primary cells. After 24 h infection, medium was replaced with complete medium. At 3 days after infection, the transfection efficiency was analyzed by fluorescence microscopy and Western blot.

HCC xenografts mouse model and lung metastasis model
For HCC xenografts model, infected KTC primary cells suspension (5 × 10 7 cells/ml, in 100 µl 1:1 PBS/Matrigel) were injected subcutaneously into the left flanks of 6-week-old male BALB/c-nude mice. Tumor size was recorded every 4 days.
Tumor volume was calculated with the formula of 0.5 × length × width 2 [6]. At 3 weeks after injection, the nude mice were euthanized. The tumors were imaged. The tumors were weighted. For lung metastasis model, the infected KTC primary cells (2 × 10 7 cells/ml, in 100 µl PBS) were injected into 6-week-old male BALB/c-nude mice via tail vein. After 6 weeks injection, the nude mice were euthanized and lung tissues were collected.

Small interfering RNA (SiRNA) transfection
The siRNA sequences for Peg3 and the negative control were synthesized by GenePharma Co. (Shanghai, China). The sequence for mouse si-Peg3 was as follows:

H&E staining
Liver tissues were fixd in 10% formalin for at least 24 h and embedded in paraffin. Then, 4-μm-thick tissue sections were cut and dewaxed, rehydrated and stained with hematoxylin for 30 seconds, followed by eosin staining for 20 seconds.
The slides were naturally dried and covered with neutral gum. Images were acquired with a light microscope (BX41, Olympus).

Immunohistochemistry (IHC)
Briefly, tissue slides were dried at 60 °C for 2 h, deparaffinized with xylene and rehydrated through a graded alcohol series. Endogenous peroxidase activity was blocked by incubation in 3% hydrogen peroxide for 30 minutes. Subsequently, antigen retrieval was carried out by immersing the slides in 10 mM sodium citrate buffer (pH 6.0) and maintaining them at a boiling temperature for 10 minutes. After cooling to room temperature, the slides were blocked in 1% blocking serum for 30 minutes and then incubated with the primary antibody at 4 °C in a humidified chamber.
Subsequently, the sections were incubated with biotinylated secondary antibody and then streptavidin-labeled horseradish peroxidase. Finally, the signal was developed with diaminobenzidine, and the slides were counterstained with hematoxylin. The staining intensity was determined using the Spot Denso function of AlphaEaseFC software. Positively stained cells were counted among a total of 500 cells on average from 3 independent tumors/inflammation tissues derived from 3 mice per group.

Sirius red staining
Tissue samples were fixed, embedded, dewaxed and rehydrated for H&E staining. Then, the tissue sections were stained with Sirius red solution for approximately 30 minutes. The slides were naturally dried in air at room temperature and covered with neutral gum. Images were acquired using the light microscope.

Western blot analysis (WB)
Proteins were extracted from cultured cells or tissues using PIPA buffer containing protease inhibitor cocktails. The relative band intensities from western blot experiments were normalized to the level of GAPDH. The relative band intensities were measured by a Quantity One software (V4.6.6) and added below the bands.

Co-immunoprecipitation (Co-IP) experiment
Proteins were extracted from HCC cells using 1 ml IP Lysis buffer. 200 μl was reserved for input detection. The remaining was centrifuged at 4 ˚C for 15 minutes at 12,000xg to remove cell debris. The cleared lysate was added with 20 μl of the protein A/G agarose bead (Santa Cruz, sc-2003) for 5 minutes at 4 ˚C with gentle agitation.
And then, the sample was centrifuged at 2500 rpm for 5 minutes to remove the protein A/G agarose bead. Next, another 60 μl of the protein A/G agarose bead and 5 μl anti-STAT3 antibody were added to the sample again and incubated overnight at 4˚C with gentle agitation. Tomorrow, the sample was centrifuged at 2500 rpm for 5 minutes to remove the supernatants, and the remaining beads were washed three times with the lysis buffer. The proteins were eluted from the beads with 50μl 1x loading buffer for 5 minutes at 100℃. The immunoprecipitated proteins were analyzed through WB as described before.

RNA isolation and reverse transcription-polymerase chain reaction (RT-PCR)
Total RNA was isolated by RNAiso Plus (Takara; 9180). RT-PCR was performed with the PrimeScript™ RT reagent kit (Takara; RR037A) according to the manufacturer's instructions. Quantitative real-time PCR (qPCR) was performed using TB Green™ Premix Ex Taq™ (Takara; RR420A). The results were obtained with CFX96TM Real-time System 3.1 software (Applied Bio-Rad). Gapdh was used as an internal control. Specific primer sequences are listed in Table S7.

Cell growth assay
For the cell viability assay, the cells were seeded in 96-well plates at a density of 8

8,000 cells (for primary cell) or 2,000 cells (for cell lines) per well overnight, treated
with the respective agents for the indicated duration and then exposed to 10 μl of CCK-8 solution per well for 1 h at 37 °C. Absorbance was measured at 570 nm with a microplate reader.

Cell migration assay
The cells were firstly seeded in 6-well plates at a density 2 × 10 5 cells per well overnight. Transfection with the si-Peg3 and nontargeting siRNA for 24 h. Cells were digested by trypsin and plated into Transwell plate without FBS overnight. Finally, cells were fixed by 4% paraformaldehyde for 10 minutes, and then stained by 0.1% crystal violet for 30 minutes. Images were acquired using the light microscope.

ROS analysis using DCFHDA staining
First, the dated medium was discarded, and the cells were washed with sterile PBS 3 times. Then, the cells were cultured with fresh medium containing 10 nM DCFHDA for 20 minutes at 37 °C in the dark. Next, the fresh medium was discarded again, followed by washing 3 times with sterile PBS. Images of ROS were observed with a fluorescence microscope with an equal exposure setting in each group. The fluorescence intensity was analyzed using Image-Pro Plus software in every group, and the production of ROS was indicated. Additionally, the cells were digested and collected. Flow cytometry was performed to evaluate superoxide production in cells.

ROS analysis using dihydroethidium (DHE) staining
Frozen sections of liver tissue were incubated with DHE at 37 °C for 40 minutes and then fixed with paraformaldehyde for 10 minutes. Next, DAPI staining solution was used to stain the nucleus for 10 minutes. Finally, images were captured using a fluorescence microscope at an excitation wavelength of 490 nm and an emission wavelength of 590 nm. The exposure time used for image acquisition of all sections was 30 ms. The intensity of DHE fluorescence was quantified by an analysis system.

High-throughput mRNA sequencing
Liver tumor nodules were obtained at 280 days from Tsc1 fl/fl ; Alb-Cre (TC); Kras G12D ;Alb-Cre (KC); and Tsc1 fl/fl ; Kras G12D ;Alb-Cre mice (KTC). Total RNA was carefully prepared, and RNA-seq analysis was performed (Genergy Bio, Shanghai, China). RNA was extracted and inspected, and a library was constructed. Then, bridge PCR amplification was performed on the cBot instrument to generate clusters, and 2 × 150 sequencing mode was used for sequencing on the Illumina sequencing platform.
Briefly, after a series of data analyses, the final sequencing results were obtained.
Limma package was used to analyze the differential gene expression between TC/KC and KTC mice [7]. The length data of whole genes were obtained from GENCODE (https://www.gencodegenes.org/). Thus, normalized read counts and Fragments Per Kilobase Million (FPKM), were calculated. p values were adjusted using the Benjamini-Hochberg procedure for multi-testing, and these false discovery rate (FDR) were computed [8]. Thus, these arguments were used to define the most significant change gene in KTC mice compared with TC or KC mice.

Bioinformatics analysis
The human HCCs were based on data generated by The Cancer Genome Atlas  [12] .
A co-immunoprecipitation experiment result of Peg3 knockout and wild type in mouse embryonic fibroblast cells using tandem mass spectrometry was obtained from an online literature's supplementary materials [13]. Protein expression difference analysis was also carried out using R limma package, followed volcano plot was drawn with R. Protein interaction networks were retrieved from STRING v.11.5 database (https://cn.string-db.org/) [14,15] and visualized using Cytoscape v.3.9.1 [16].