Introduction

Hepatocellular carcinoma (HCC) is the cause of high number of cancer-related deaths worldwide. As per the latest annual GLOBOCON statistics, the mortality rate of hepatocellular carcinoma among all cancer types is 8.2% (781,631 deaths in 2018); this figure very close to new HCC incidence of 841,080 cases in 2018 (Bray et al. 2018). The lack of markers for its early diagnosis results in usually only at advanced stage detection of HCC that makes the patient’s survival limited and higher mortality rate seems inevitable (Hemming et al. 2016). During the last 10 years, the emergence of HCC has been in proportion to the prevalence and perpetuation of exposure to hepatitis B and C (Hemming et al. 2016). However, now, obesity and NAFLD either viral- or alcoholic-induced hepatitis are slowly becoming the primary causes of HCC (Younes and Bugianesi, 2018). The mass awareness about hepatitis A and hepatitis B vaccines and availability of high effective of direct antiviral therapies for HCV may have substantial impact in reducing the appearance of HCC in future (Hemming et al. 2016). It is well reported that most of the chronic diseases related to liver cancer are caused by changes at the phenotypic and genotypic levels in tumor cells. Further exploration of pathogenesis leading to HCC suggested that certain signaling pathways and molecular alterations have regulatory impact in HCC development by promoting cell growth and survival (Avila et al. 2006). Alterations in cell signaling pathways play a major role in tumorigenesis (Sever and Brugge, 2015), but mechanisms affecting cell signaling leading to hepatocarcinogenesis are still not fully understood (Samuele et al. 2013). The main lineaments of HCC are inflammation, ROS (reactive oxygen species), fibrosis, and cirrhosis that induce genetic and epigenetic mutations leading to activation of oncogenes and inhibition of tumor suppressor genes (Samuele et al.2013). P53/RB, Wnt/β-catenin, PI3K/PTEN/Akt/mTOR, JAK/STAT, MAPKs, RAS, ERK1/2, NF-κB, c-Myc, IGF1 and AP-1, retinoblastoma protein (pRb) chromatin regulation, oxidative and endoplasmic reticulum stress signaling, TGF-β, and HGF/c-Met are the main altered signaling pathways involved in HCC development (Aravalli et al. 2013; De Minicis. et al. 2013: Chiba et al. et al. 2015; Bakiri et al. 2017).

To identify the key players in pathogenesis of HCC, our laboratory has been trying to understand the signal transduction processes that lead to HCC initiation and progression, and to identify the tissue specific proteins that may have potential to serve as biomarker for early diagnosis and therapeutic targets of HCC. The present study was carried out using liver tissues from HCC-bearing male Wistar rats (Ranjpour et al. 2018). We earlier reported the development of a novel rodent model for study of HCC (Malik et al. 2013). Total tissue proteins were resolved on 2D gels, the differentially expressed proteins were analyzed by PD Quest, and protein spots of interest were characterized. The mRNA expression for the characterized proteins was quantified to validate the protein data obtained by 2D-Electrophoresis. Further, a global gene network has been developed to reveal the interactions among genes coding for the experimentally identified proteins and their interactors associated with signaling pathways during HCC progression.

Methods

The animal model for study of HCC was previously developed in our laboratory (Malik et al. 2013). The liver tissues were taken from HCC-bearing animals (Ranjpour et al. 2018). The study included 36 animals (Wistar rats) divided into two groups: control group and treated group, each group having 18 animals. The animals in treated group were further subdivided into three groups: 1 month, 2 month, and 4 month (6 animals in each group). The carcinogens DEN and 2-AFF (acetyl amino fluorine) were used to induce HCC. A single dose (200 mg/kg body weight) of DEN was given intravenously to each animal of treated group to initiate HCC, followed by oral doses of 2-AAF (150 mg/kg body weight) on the alternate days of first week of each month to promote HCC development. The control groups were treated with normal saline at the same schedule as the treated groups. The animals were sacrificed at the end of each month; liver tissues were excised and preserved in RNA later solution for further study (Ranjpour et al. 2019).

Proteomic profiling of proteins

We commenced our investigation based on omics approaches, wherein the liver tissues were preserved in RNA later solution (RNA stabilization reagent from QIAGEN) and kept at − 80 °C until used. (100–150 mg) Tissues were homogenized using Polytron PT3100 homogenizer in the presence of urea lysis buffer (8 M urea, 65 mM CHAPS, 65 mM DTT, 2 M thiourea, 33 Mm Tris, and 6 mM PMSF). The homogenate was centrifuged (10,000 rpm for 10 min) (Chaudhary et al. 2013). The proteins extracted from the supernatant fraction were estimated by Bradford method (Bradford 1976). Aliquots containing 120 μg proteins were focused on 11-cm, pH 4–7 IPG strips (GE Healthcare, Immobilize TM Dry Strip) at 500 V for 1 h, at 1000 V for 1 h, and finally at 6000 V for a total of 35,000 V-hour. The total proteins were then fractionated using 12% gradient SDS polyacrylamide gels electrophoresis for 2D analysis of proteins. The resolved proteins were silver stained, analyzed by PD-Quest software (Bio-Rad), and the proteins of interest were characterized by MALDI-TOF-MS analysis as described earlier (Ranjpour et al. 2018).

Characterization of proteins by MALDI-TOF-MS/MS analysis

The protein spots identified by PD Quest analysis were excised and transferred to sterilized eppendorf tubes, destained by soaking the gel pieces in 15 mM potassium ferricyanide and 50 mM sodium thiosulphate, three times 10 min each. ACN (Acetonitrile) and a Speed Vac were used to thoroughly dehydrate and dry the gel pieces. These were further incubated for 1 h in DTT for rehydration followed by incubation in iodoacetamide for 45 min. The gel pieces were incubated for 10 min with ammonium bicarbonate, completely dehydrated with ACN, and dried. The proteins in gel pieces were digested with 150–200 ng trypsin in 50 mM ammonium bicarbonate for 16 h at 37 °C. The peptides were extracted with the peptide extraction solution and re-suspended in 20 mM of Tris-acetate buffer pH 7.5. The HCCA matrix was mixed with peptides that produced in 1:1 ratio and 2 μl of the mixture solution was applied on MALDI plate. Finally, samples were dried and analyzed by MALDI TOF/TOF Ultraflex III instrument version 3.2. The peptide mass fingerprint was fed into the Mascot server for a peptide mass fingerprint search in the NCBInr Rattus norvegicus database for protein characterization (Ranjpour et al. 2018).

Validation of expression of experimental protein findings using real-time PCR

Total RNA was isolated from 50 mg of liver tissue from control and treated group using Mini Sure Spin total RNA isolation kit (Nucleopore, Genetix Biotech Asia), RNA concentration was measured using NanoDrop 2000c Spectrophotometer 56 (Thermo Fisher Scientific), and the purity of RNA was assessed by A260/A280 ratio. A sample containing 1500 ng RNA was reverse transcribed using cDNA synthesis kit (Verso cDNA Synthesis Kit, Thermo Scientific). The integrity of cDNA was confirmed by PCR amplification using GAPDH specific primers. Further, the primers were designed against the specific genes coding for the MALDI-TOF characterized proteins (Table 1). The PCR reaction mixture contained 0.5 μl cDNA, 1× Taq buffer, 1 U Taq polymerase, 1.5 mM MgCl2, 0.4 mM dNTPs, and 0.4 μM each primers; the amplification conditions were (i) 96 °C for 5 min (initial denaturation); (ii) 35 cycles of 94 °C for 30 s, 58 °C for 30 s, 72 °C for 30 s; and (iii) final extension 72 °C for 10 min. Amplified products were analyzed on 1.5% agarose gels by electrophoresis in 1× TAE buffer system and the DNA bands were visualized using UV gel documentation system (Alpha Imager HP System). Furthermore, the quantification of transcripts was carried out using Roche LightCycler 480 Real-Time PCR System using SYBR green chemistry. The data was analyzed using LightCycler 480 Software (Version 1.5). The 25 μl reaction mixture contained 12.5 μl of 2× Maxima SYBR Green qPCR Master Mix (Thermo Fisher Scientific), 0.4 μM of each primer (Table 1), 200 ng cDNA and nuclease free water. The thermal cycling conditions were (i) 95 °C for 10 min (initial denaturation), (ii) 45 cycles of 95 °C for 30 s, 56 °C for 30 s, and 72 °C for 30 s, followed by (iii) melt curve analysis at temperature range of 55–95 °C.

Table 1 The primers sequence of corresponding genes of MALDI-TOF identified proteins along with housekeeping gene primers
Full size table

Mapped networks for SYNE1, MMP10, MTG1, and their interacting partners

Our major goal was to understand the functional involvement of the three selected proteins and their interacting partners with etiopathophysiology of HCC. We took the advantage of software GENEMANIA database to map our network (Warde-Farley et al. 2010). All the genes coding for selected proteins and their associated partners were mapped.

The statistical analysis

In the real-time analysis, all the cDNA samples were run in triplicates. One-way ANOVA and post hoc Dunnet test were carried out by using GraphPad Prism (version 8); at P ˂ 0.05, the difference between treated and control group was considered to be significant.

Results

Identification of the differentially expressed proteins

We earlier reported histological confirmation of HCC initiation at 1 month and tumorigenesis at 4 months after carcinogen treatment (Ranjpour et al. 2018). Total liver proteins of control and tumor-bearing rats were resolved on 2D gels (Fig. 1, A and B). Protein profiles were compared and analyzed using PD-Quest software. Differentially expressed proteins were identified, based on the difference in intensity of the spots. Three protein spots, namely, A, C, and E whose expression was elevated in tumor tissue, were selected for further studies. These spots were assigned unique SSP numbers by PD-Quest analysis (Fig. 1 C). The protein spots of interest were characterized to be nesprin-1 isoform X6 (Nesprin-1), stromelysin-2 precursor (Stromelysin-2), and mitochondrial ribosome–associated GTPase 1-isoform X2 respectively (Fig. 2). These proteins and their genes are listed in Table 2. The peptide sequence analysis of these proteins has been represented in Table 3.

Fig. 1
figure 1

Total tissue proteins were resolved on 2D gels of control (A) and treated (B) animals. Some of the differentially expressed protein spots (target proteins in this study) are marked as A, C, and E. (C) Master gel of control and 4th month in which the selected protein spots are given unique SSP numbers, namely, A: 6006; C: 1001; and E: 7215. The pop up graphs on master gel compared the protein expression based on intensity

Full size image
Fig. 2
figure 2

Mascot Score Histogram of MALDI-TOF-MS-identified protein spots. The search against Mascot Score Histogram identified the proteins and the score for each protein is shown in this figure. Whereas spot E: Nesprin-1 isoform X6, spot C: Stromelysin-2 precursor and spot A: Mitochondrial ribosome-associated GTPase 1 isoform X2

Full size image
Table 2 Name of characterized proteins by MALDI-TOF and their related genes
Full size table
Table 3 The peptide sequence analysis of SYNE1, MMP10, and MTG1 (Mr, average molecular mass of the peptide in kilodalton; Expt, experimentally determined molecular mass; Calc, theoretically calculated mass of peptide based on atomic mass; and Ppm, parts per million)
Full size table

Transcript quantification of differentially expressed proteins

The qPCR analysis revealed elevated expression of the transcripts for all the target proteins in tumor tissues. A 2.9-fold increase in expression of transcript for MMP10 was seen at 1-month post-carcinogen treatment. The mRNA expression continued to increase with disease progression and showed 3.7-fold elevation at 2-month and 5.4-fold increase at 4-month post-carcinogen treatment. Similarly, quantification of mRNA for MTG1 revealed 2.3-fold increase at 1-month, 4.7-fold increase at 2-month, and 5.1-fold increase at 4-month post-carcinogen treatment. On the other hand, the analysis of mRNA expression for SYNE1 revealed 1.33-fold increase at 1 month after carcinogen treatment. However, its transcript expression then plateaued and remained elevated almost at the same level during HCC progression and tumorigenesis. The analysis revealed 1.28-fold increase at 2-month and 1.74-fold increase at 4-month post-carcinogen treatment (Fig. 3).

Fig. 3
figure 3

Expression analysis of SYNE1, MMP10, and GTPase1 in control and carcinogen-treated groups by real-time PCR

Full size image

Network analysis of genes coding for the experimentally identified proteins

Further study on these proteins demonstrated their association with a number of interacting partners indicating the contribution of these proteins in pathophysiology of HCC. We analyzed the functional association of the dysregulated genes coding for experimentally identified proteins and their interacting partners to deduce their possible role in HCC development. The resulting global gene network using functional GENEMANIA database (Fig. S1) (supplementary data) showed following observations:

  1. 1-

    Co-expression of MMP10 and MTG1; SYNE1 and KDR; MTG1 with MTOR, CUL7, CUL9, MMP10, and KLC1.

  2. 2-

    SYNE1 gene was found to interact with FOS, JUN, EGFR, VEGFA, MMP3, HGF, KLC1, ATF2, MLF2, FKLPB7, TELO2, MAPK8, MAPK9, and MAPKAP1.

  3. 3-

    The network analysis revealed direct interaction between MMP10 and MMP3, which further connect MPP10 to JUN, FOS, SYNE1, and EGFR; furthermore, JUN connects MMP10 to MAPK8, MAPK9, and TP53. The direct interaction among EGFR and MTOR has also been found. Moreover, the network also revealed the direct interaction of MTG1 with CUL7 and CUL9. These further connect MTG1 to FOS, JUN, EGFR, VEGFA through TP53 (Fig. 4) (Table 4).

Fig. 4
figure 4

Network diagram showing interactions among the gene coding for identified proteins and their neighbors involve in cancer pathways

Full size image
Table 4 The significant reported HCC-associated genes showing interactions with SYNE1, MMP10, and MTG1
Full size table

Discussion

The report describes the elucidation of molecular events during HCC progression. Using proteomic and genomic approaches, we have investigated the possible role of the three proteins and their genes, namely, nesprin-1 isoform X6 (SYNE1), stromelysin-2 precursor (MMP10), and mitochondrial ribosome-associated GTPase 1-isoform X2 (MTG1). We report that these proteins interact with several signaling pathways involved in HCC development.

The level of nesprin-1 isoform X6 was found to be elevated. Nesprin-1 is an outer nuclear membrane protein that links the nucleus to actin cytoskeleton (Zhang et al 2009). It has been reported that nesprin-1 plays a critical role in cell proliferation and apoptosis of mesenchymal stem cells (MSCs) (Yang et al. 2013). SYNE1 is large gene (0.5 Mb) that encodes different isoforms of nesprin-1 by alternative splicing. This plays a critical role in cell structural integrity, cell signaling, and nuclear organization (Doherty et al. 2010). In contrary to our study, SYNE1 gene was found to be downregulated in ovarian cancer (Doherty et al. 2010) and oral cancer (Shah et al. 2018). On the other hand, our observations are in sync with several studies wherein SYNE1 gene has been reported to be upregulated in colorectal cancer (Melotte et al. 2015) and lung cancer (Tessema et al. 2008). Nuclear envelope encoding gene–associated mutations have been implicated as one of the important reasons of age-associated diseases and cancer (Kim et al. 2015). Mutations in SYNE1 have been reported to be associated with several cancers including head and neck squamous cell cancer (Eun et al. 2017), colorectal cancer (Jesinghaus et al. 2015), NAFLD-related HCC (Ki Kim S et al. 2016), glioblastoma (Masica and Karchin 2011), and gastric cancer (Cong et al. 2018).

Diane Zhang has reported the association of SYNE1, CLU7, and CUL9 to liver cancer (Zhang 2016). The represented data propose the interaction between SYNE1, CUL7, and CUL9 through p53 and VEGFR. It has been documented that function of CUL9 in tumor suppression is largely mediated through p53 whereas deletion of CUL9 had no significant effect on cell-cycle progression, but affected impaired DNA damage-induced apoptosis (Pei et al. 2011). On the other hand, SYNE1 has interaction with JUN and FOS. A study carried out on DEN-induced liver cancer-bearing mice showed that dysregulated expression of FOS is a key element occurring during HCC development that induces inflammation, hepatocyte proliferation, DNA damage response activation, and premalignant transformation (Latifa et al. 2017). Our study was performed on the liver with DEN-induced HCC whereas, DEN is a carcinogen which induces single-strand breaks in DNA in hepatocytes and also causes changes in several enzymes involved in DNA repair (Malik et al. 2013). All together, we conclude that high expression of SYNE1 may induce liver cancer formation through interactions with liver cancer-associated genes such as JUN, FOS, VEGFR, P53, CUL7, and CUL9 suggesting that SYNE1 could have a role in cancer development pathways.

Upregulation of stromelysin-2 (MMP10) expression has been reported in our study and its elevation is associated with HCC angiogenesis (García-Irigoyen et al. 2015). MMP10 is a cancer-related gene belonging to stromelysins (MMP-3, 7, 10, 11, 26, and 27) that displays hydrolyzing ability of a broad range of complex extracellular matrix (ECM) proteins like collagen type III, IV, V, elastin, proteoglycans, and glycoproteins (Zhang and Chen, 2017). Degradation and renewal of liver tissue are associated with the ECM-related pathways (Duarte et al. 2015). Many extracellular factors such as cytokines, growth factors, and cell contact to ECM regulate the expression of MMP genes (Westermarck and Kähäri, 1999). The invasion of MMP10 in head and neck squamous cell carcinoma is presumably associated with partial inhibition of p38 MAPK (Deraz et al. 2011). The p38α (MAPK14) has been speculated to directly affect tumor invasion and angiogenesis as it can induce expression of the matrix metalloproteinases such as MMP1, MMP3, and MMP13 which regulate matrix renewal and degradation by metastatic cancer cells whose subsequent activity is needed for HCC proliferation (Wagner and Nebreda, 2009). In the present study, the deduced network reveals the interaction of MMP3 with p53 through HGF, as well as cross-talk between MMP10 with JUN and FOS through MMP3, which connect MMP10 with MAPK8 and MAPK9 (Fig. 4). Stromelysin-1 (MMP-3) induces HGF-mediated invasion of HCC (Inserm 2002). There are several signal cascades that control the expression of MMP10 including MAPK and JAK/STAT pathways (Cui et al. 2012). MMP10 gene contains DNA-binding sites for AP-1 and STAT3 at the 5′ regulatory region (Cui et al. 2012). AP-1 is a dimer composed of FOS and JUN family members and its activity is regulated by interactions with some kinases and transcriptional co-activators (Karin et al. 1997). Co-expression of FOS and JUN is suspected to be associated with tumor progression in HCC (Yuen et al. 2001). Our observations correlate with another finding wherein JNK/JUN pathway was reported to target genes of metalloproteinases such as MMP10. The oncogenic functions of JNKs are mostly based on their ability to phosphorylate JUN and therefore, to activate AP-1; however, their suppression function is also related to their pro-apoptotic activity (Wagner and Nebreda, 2009). On the other hand, our network revealed interactions of MMP10 and STAT3 through MTG1 and MTOR, whereas STAT3 plays a pivotal role in HCC survival, growth, angiogenesis, and metastasis (Jung et al. 2017). Altogether, based on our data obtained from the network, we hypothesize that signal transduction among JNK/JUN and STAT3 may induce the enhanced expression of MMP10 through MMP3 leading to HCC progression by induction of angiogenesis.

We also report an increase in the mitochondrial ribosome-associated GTPase 1 expression in our study. The generated network revealed direct interactions among MTG1 with MMP10 and with MTOR, CUL7, and CUL9. MTG1 is potentially associated with several cancers; however, further studies are needed to find out the exact role of this protein in cancer induction or progression (Liu and Pan 2016). MTG1 plays a role in regulation of mitochondrial ribosome assembly and translational activity (Kotani et al. 2013).

According to GENE CARD database (www.genecards.org), Gene Ontology shows the annotations related to MTG1, which includes GTP binding and GTPase activity. Certain small GTP-binding proteins control the enzymatic activity of mitogen-activated protein kinases (MAPKs) such as ERKs, JNKs, and p38 kinase subfamilies (Teramoto et al. 1996) in signaling pathways associated with cancer progression (Latifa et al. 2017). The present study shows the interactions of MTG1 with MAPK8 and MAPK9 through MTOR. We may conclude that MTG1 may have a role in HCC formation through its relation with MAPKs and JNKs.

Altogether, we conclude that elevation in expression of SYNE1, MMP10, and MTG1 may play critical role in HCC development through their interrelationship with CUL 7 and 9, TP53, VEGFR, JNK/JUN, STAT3, FOS, MAPKs, MTOR, and HGF.

Conclusion

Three proteins, namely SYNE1, MMP10, and MTG1 have been found to be overexpressed in HCC and the interactions among these are shown to be through JUN, STAT3, FOS, TP53, VEGFR, CUL9, CUL7, HGF, and MAPKs Table 4. These proteins play pivotal regulatory roles in signaling pathways associaed with HCC progression and thus could be targeted as potential biomarkers for early detection of HCC. However, further research is needed to validate the data obtained from this study using human samples.