Twenty pairs of HCC and matched peritumor liver tissues were collected from patients who underwent surgical resection and were pathologically diagnosed with HCC at Zhongshan Hospital, Fudan University (Shanghai, China), and used to detect the mRNA and protein expression levels of SFKs by quantitative real-time PCR (qPCR) and western blot (WB). The clinical characteristics of the 20 resected patients with HCC were shown in Supplementary Table 1. This study was approved by the Research Ethics Committee of Zhongshan Hospital, Fudan University, and informed consent was obtained from each patient before use.

TIMER2.0 (http://timer.comp-genomics.org/) is a comprehensive tool that provides modules for systematic analysis of immune infiltrates and gene expression across multiple cancer types [24]. The Cancer Genome Atlas (TCGA) samples were used by this tool for data analysis. We compared the expression levels of SFKs between tumor and normal tissues in multiple cancers via the ‘Gene${}_{\text{-}}$DE Module’ of Cancer Exploration Component. Statistical significance evaluated by the Wilcoxon test was annotated by the number of stars (*p 0.05, **p 0.01, ***p 0.001). The ‘sCNA (Somatic Copy Number Alterations) Module’ and ‘Gene Module’ of Immune Associated Component were used to analyze the interrelated proportion of different sCNA states and the correlation between SFKs with the immune cell types, including B cells, macrophages, dendritic cells, neutrophils, CD4+ T cells and CD8+ T cells. Meanwhile, the‘purity adjusted’option was selected and the relationship between gene expression level and tumor purity appeared on the left-most panel. Partial Spearman’s correlation was used to perform association analysis and p 0.05 was considered statistically significant.

UALCAN (http://ualcan.path.uab.edu/index.html) is a free web resource for analyzing cancer OMICS data (TCGA, CPTAC and CBTTC) [25]. UALCAN allows users to perform gene expression analysis and identify biomarkers, as well as graphs representing expression profiles and patient survival information. The transcriptional expression of SFKs and their relationship with HCC stage and prognosis were analyzed through this website using TCGA dataset. The significance of the difference in the transcriptional levels was assessed by Student’s t-test, and p-value 0.05 was considered statistically significant.

cBioPortal (https://www.cbioportal.org/) is an open access tool for research, visualization, and analysis of multidimensional cancer genomic data [26]. The tool integrates data from 126 tumor genome studies, including large tumor studies such as TCGA, covering data from 28,000 samples, some of which also include clinical prognostic and phenotypic information. In this study, the genetic alterations in SFKs and their relationship with mRNA expression were obtained from cBioPortal based on the HCC dataset including 973 patients (TCGA, 372 samples; three studies, 601 samples).

The Kaplan-Meier plotter (https://kmplot.com/analysis/), an online tool that contains gene expression data and survival information of 21 cancer types, was used to assess the effect of SFKs mRNA levels on the overall survival (OS) of patients with HCC [27]. The resources used by this tool for HCC survival analysis were sourced from TCGA database. Information on the number-at-risk case, median mRNA expression level, hazard ratio (HR), 95% confidence interval (CI), and p-value were displayed on the Kaplan-Meier plotter web page. Statistical significance was considered when the p-value was 0.05.

GEPIA (http://gepia.cancer-pku.cn/index.html) is an interactive web application that provides fast and customizable functions based on TCGA and Genotype Tissue Expression (GTEx) data [28]. OS analysis of SFKs based on gene expression was achieved through GEPIA using these samples (TCGA, 369 tumor samples, 50 normal samples; GTEx, 110 samples). The log-rank test was used for hypothetical tests. The median expression level was used to divide the high- and low-expression cohorts. The Cox proportional HR and 95% CI information were included in the survival plot. p 0.05 was considered statistically significant.

The HCC cell line Huh7 was obtained from Shanghai Cell Bank, Type Culture Collection Committee, Chinese Academy of Sciences. Huh7 was cultured in DMEM medium (#D6429, Sigma-Aldrich, Taufkirchen, Germany) supplemented with 10% fetal bovine serum (Gibco, Carlsbad, CA) at 37 °C in a humidified incubator with 5% CO${}_2$.

Full-length cDNA of FYN or SRC were cloned into the empty lentiviral vector pCDH-CMV-MCS-EF1-Puro. Recombinant plasmids (psPAX2 and pMD2.G) were co-transfected into 293T cells using Lipofectamine 3000 (Invitrogen, Carlsbad, CA). The lentivirus was collected from the complete medium two days after transfection.

To establish stably overexpressing cell lines, an appropriate amount of virus was added to Huh7 cells, followed by puromycin selection 48 h later. To establish knockdown cells, small interfering RNA (siRNA) targeting FYN (siFYN: GGUGGAUACUACAUUACCA) and SRC (siSRC: GAAUCUGAUCAACAGUUUAUU) was used in huh7 cells. The overexpression and knockdown efficiencies of FYN and SRC were verified by WB.

Total RNA was extracted from 20 pairs of HCC and matched peritumor liver tissues using the RNA-easy Isolation Reagent (#R701-01, Vazyme, Nanjing, China). RNA was reverse-transcribed into cDNA using the HiScript® II Q RT SuperMix for qPCR (+gDNA wiper) kit (#R223-01, Vazyme, Nanjing, China). For qPCR, the ChamQ Universal SYBR qPCR Master Mix kit (#Q711-02, Vazyme, Nanjing, China) was used according to the manufacturer’s instructions. DNA amplification and detection were performed using a 7900 Sequence Detection System (Applied Biosystems, Foster City, CA). The primer sequences used are listed in SupplementaryTable 2. Relative gene expression levels were calculated using the 2${}^{-△△\text{Ct}}$ method and were normalized with GAPDH. Data were analyzed using GraphPad Prism (version 8.3.0, Dotmatics, Boston, MA, USA). p 0.05 was considered as statistically significant using Student’s t-test.

Total proteins were extracted in RIPA lysis buffer, separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and electrotransferred onto polyvinylidene fluoride (PVDF) membranes. The membrane was blocked with 5% nonfat milk for one hour at room temperature and then incubated with primary antibody ($\beta$-actin, #abs137975, Absin, Shanghai, China; FYN, #sc-365913, Santa Cruz, CA, USA; SRC, #2109, CST, Danvers, MA, USA) overnight at 4 ℃, followed by HRP-conjugated secondary antibody for one hour at room temperature. Immunoreactive bands on the membrane were detected using an enhanced chemiluminescence (ECL) reagent.

Huh7 cells were seeded in 96-well plates at 10${}^3$ cells/well and observed for seven days. Ten $\mu$L CCK8 solution (#8-500T, FTC Life Science, Shanghai, China) was added to each well and the plate was incubated for one hour at 37 °C. The absorbance of each well was measured at a wavelength of 450 nm using a Microplate Reader (AMR-100, Allsheng, Hangzhou, China).

For the colony formation assay, Huh7 cells were seeded in 6-well plate at 10${}^3$ cells per well. After incubation for two weeks, the cells were fixed with 4% paraformaldehyde (#P0099, Beyotime, Shanghai, China) and stained with 0.05% crystal violet (#C0121, Beyotime, Shanghai, China).

Transwell assays were performed in chambers with a pore size of 8 $\mu$m (#3422, Corning, NY, USA). For the cell migration assay, 1 $\times$ 10${}^4$ Huh7 cells in 200 $\mu$L serum-free DMEM were placed in the upper chamber, and 600 $\mu$L DMEM with 20% FBS was added to the lower chamber. After incubation for 48 h, the cells on the bottom surface of the chamber were fixed in 4% paraformaldehyde, and stained with 0.05% crystal violet (Beyotime, Shanghai, China). For the cell invasion experiments, each chamber was precoated with 50 $\mu$L BD Matrigel mixture (diluted at 1: 5 with DMEM) for one hour at 37 °C. 5 $\times$ 10${}^4$ Huh7 cells were added to the upper chamber and the subsequent steps were the same as those used in the cell migration assay. Five randomly selected fields were photographed using an IX71 inverted microscope (IX71, Olympus Corp, Tokyo, Japan), and the mean value was recorded.

BALB/c-nude mice (4–6 weeks, male) were purchased from the Shanghai SLAC Laboratory Animal Co., Ltd (Shanghai, China) and maintained under pathogen-free conditions. To establish the subcutaneous xenograft tumor model, the mice were randomly divided into the control group (n = 6) and the FYN overexpression group (n = 6). Cell suspensions (5 $\times$ 10${}^6$ cells) of Huh7 cells were subcutaneously injected into the flanks of each mouse and tumor size was measured with digital calipers every five days. About three weeks after injection, the mice were sacrificed and tumors were harvested. Tumor size was calculated using the formula: V = 0.5 $\times$ length $\times$ width${}^2$. All experiments were approved by the Animal Ethics Committee of School of Basic Medical Sciences, Fudan University.

Transcriptional expression levels of SFKs in HCC and matched normal tissues from the TIMER2.0 database were analyzed using the Wilcoxon test, and those from the UALCAN database and qPCR experiments were assessed using the Student’s t-test. The Student’s t-test was also used to analyze the relationship between the expression levels of SFKs and HCC stages. The effect of SFKs mRNA expression on the survival of patients with HCC was evaluated using the log-rank test. A partial Spearman’s correlation was used to perform an association analysis between SFKs expression levels and immune cells. GraphPad Prism (version 8.3.0, Dotmatics, Boston, MA, USA) was used for statistical analysis and figure creation. Cell proliferation, migration, invasion, tumor volume, and tumor weight were analyzed using the Student’s t-test. p 0.05 was considered statistically significant.

The working flow chart depicted in Fig. 1 and the clinical characteristics of HCC patients in each dataset shown in Supplementary Table 3. We first used TIMER2.0 tool to explore the expression of SFKs in multiple cancers, including HCC (Fig. 2A). The data revealed that the transcriptional levels of SFKs were frequently abnormal in many types of cancer. In HCC, LYN, SRC and SRM were significantly elevated in HCC tissues compared to normal tissues, while BLK, FGR, FYN and HCK were reduced. Other members, including BRK, FRK, LCK and YES, showed no significant differences in HCC. We next utilized the UALCAN website to verify seven SFK genes with altered mRNA expression levels (Fig. 2B). The results indicated that the expression levels of LYN, SRC and SRM were significantly higher in HCC tissues than in normal tissues, and the expression levels of FYN were markedly lower. However, the changes in the expression levels of BLK, FGR and HCK were not statistically significant. These results were partly consistent with those obtained using TIMER2.0.

Fig. 1.

Flow chart of the study. Blue, discovery cohort; Orange, validation cohort and experiments. The sample sources used by each analysis tool are shown under the tool.

Fig. 2.

The mRNA levels of Src family kinases (SFKs) in various types of cancer including Hepatocellular Carcinoma (HCC). (A) The mRNA expression data analysis from TIMER2.0 indicated that the mRNA levels of BLK, FGR, FYN and HCK were significantly down-regulated in HCC tissues, while LYN, SRC and SRM were up-regulated. *p 0.05, **p 0.01, ***p 0.001, Wilcoxon test. (B) 17Validation analysis from UALCAN on the relative mRNA level of BLK, FGR, FYN, HCK LYN, SRC and SRM in normal liver and HCC tissues. The expression levels of FYN were decreased in HCC tissues than that in normal liver tissues, while LYN, SRC and SRM were increased (Normal, n = 50; Primary tumor, n = 371). ns, not significant, *p 0.05, ***p 0.001, Student’s t-test. Red, overexpression genes; Blue, low expression genes.

To further verify the above results, we detected the mRNA expression levels of SFKs in 20 pairs of HCC and matched peritumor liver tissues using qPCR (Fig. 3). The results revealed that the expression levels of LYN, SRC and SRM were upregulated, and FYN was downregulated, whereas other SFK members did not show notable differences in HCC. The qPCR results were largely in agreement with the common parts of the TIMER2.0 and UALCAN.

Fig. 3.

The relative SFKs mRNA level in 20 paired HCC and adjacent nontumor tissues. Quantitative real-time PCR (qPCR) assay were employed to measure the mRNA levels in tissues. Relative expression of SFKs was normalized with GAPDH. Compared to adjacent normal tissues, the expression levels of FYN were reduced, but LYN, SRC and SRM were elevated in HCC. n = 20. ns, not significant, **p 0.01, ***p 0.001, Student’s t-test. Red, overexpression genes; Blue, low expression genes.

We then investigated the genetic alterations of SFKs in patients with HCC using the cBioPortal and TIMER2.0 databases. As shown in Fig. 4A, alterations in the frequencies and types of SFKs were determined in 973 samples from four HCC studies. The SFKs mutations, amplifications, deep deletions, and multiple alterations in 973 HCC samples occurred at 10.08%, with frequencies of 3.08% (30 cases), 3.60% (35 cases), 3.19% (31 cases) and 0.21% (2 cases), respectively. The percentages of gene changes in individual SFKs in HCC were displayed in Fig. 4B, and the mutation frequency ranged from 0.4% to 4% (BLK, 3%; BRK, 0.7%; FGR, 0.6%; FRK, 0.8%; FYN, 1.3%; HCK, 0.4%; LCK, 0.6%; LYN, 4%; SRC, 0.7%; SRM, 1.1%; YES, 0.8%).

Fig. 4.

Genetic alteration of SFKs genes in HCC. (A,B) The genetic alteration type and frequency of different expressed SFKs in different HCC datasets (cBioPortal). SFKs were altered in 98 samples of 973 patients with HCC, accounting for 10%, *p 0.05. (C) The relative proportion of different sCNA status of SFKs members in The Cancer Genome Atlas Liver Hepatocellular Carcinoma (TCGA-LIHC) (TIMER2.0). (D) The boxplot of correlation between the genetic alteration of SFKs and its corresponding mRNA expression (cBioPortal).

Next, the stacked bar plot from TIMER2.0 showed the relative proportion of different sCNA states of SFKs in HCC patients (Fig. 4C). Arm-level deletion and arm-level gain were the primary sCNA states, and arm-level deletions were observed mainly in BLK, FGR, FRK, FYN, LCK and YES, whereas arm-level gain was detected in BRK, HCK, LYN, SRC and SRM.

Furthermore, we explored the relationship between gene alterations and mRNA expression in the 973 HCC samples (Fig. 4D). The decrease in the mRNA expression level of FYN might be partly caused by the deep or shallow deletion of genes. The gain or amplification of genes could explain the increased mRNA expression levels of LYN, SRC and SRM. Combined with the above data, we speculated that genetic alterations in SFKs might lead to changes in mRNA expression.

We then assessed the correlation between SFKs mRNA expression and pathological stage in HCC using the UALCAN website. As shown in Fig. 5, SRC was highly expressed in HCC stage 1–3 compared to that in normal tissues, and BRK, FRK, FYN, LYN and SRM were differentially expressed in several tumor stages. In comparison to normal tissues, the expression levels of BLK, LCK and YES were diverse in certain tumor stages, but there was no difference in the expression levels of FGR and HCK in different tumor stages.

Fig. 5.

Correlations between SFKs expression and tumor stage in HCC patients. Histogram (UALCAN) showed that BLK, BRK, FRK, FYN, LCK, LYN, SRC, SRM and YES were correlated with HCC clinical cancer stage. The gene names are shown in red. *p 0.05, **p 0.01, ***p 0.001, Student’s t-test.

Furthermore, we analyzed the prognostic significance of SFKs expression in HCC using the Kaplan-Meier Plotter, GEPIA, and UALCAN websites. The OS curves were presented in Fig. 6A. Increased BLK (HR = 0.61, p = 0.01), BRK (HR = 0.68, p = 0.033), FYN (HR = 0.43, p = 2.7E-06), LCK (HR = 0.59, p = 0.0034), SRM (HR = 0.52, p = 0.00021), YES (HR = 0.68, p = 0.028) mRNA levels and decreased SRC (HR = 1.75, p = 0.0023) mRNA levels in HCC patients were strongly associated with better OS. However, the expression levels of FGR (HR = 1.18, p = 0.4), FRK (HR = 0.75, p = 0.099), HCK (HR = 1.23, p = 0.24) and LYN (HR = 0.77, p = 0.15) did not affect the OS of HCC patients.

Fig. 6.

Prognostic values of SFKs in HCC. (A) HCC patients with reduced mRNA expression levels of BLK, BRK, FYN, LCK, SRM and YES were significantly related to short overall survival (OS), while higher SRC mRNA expression was correlated with short OS. Cutoff: median (Kaplan-Meier plotter). (B) Low mRNA expression levels of FYN and LCK were significantly associated with short OS, conversely, high SRC mRNA expression was correlated with short OS in HCC patients. Cutoff: median (GEPIA). (C) Elevated FYN mRNA level in HCC was positively correlated with OS, while SRC was negatively correlated with OS (UALCAN). p 0.05 (genes labeled with red) was considered statistically significant.

GEPIA was used to verify the seven SFK genes that had an impact on prognosis (Fig. 6B). HCC patients with higher transcriptional levels of SRC were significantly associated with shorter OS (p = 0.016), whereas those with increased transcriptional levels of FYN (p = 0.0034) or LCK (p = 0.044) were markedly associated with better long-term OS. There was no significant correlation between the expression levels of the rest four genes and OS in patients with HCC. The results of FYN, LCK, SRC from the Kaplan-Meier Plotter and GEPIA were consistent.

UALCAN was used to further verify the influence of the mRNA expression levels of FYN, LCK and SRC on OS (Fig. 6C). Among these three genes, only FYN (p = 0.043) and SRC (p = 0.00049) mRNA expression levels were significantly associated with OS. The above results from the three websites indicated that the mRNA expression levels of FYN and SRC were remarkably correlated with prognosis in patients with HCC.

Interaction network and functional enrichment analysis revealed that SFKs might be closely associated with tumorigenesis, progression, and the immune microenvironment (Supplementary Fig. 1; Supplementary Table 4; Supplementary Table 5). We further explored the relationship between SFKs and immune infiltration levels in HCC using the TIMER2.0. As shown in Fig. 7, the expression levels of FGR, HCK, LCK, LYN and YES were positively in connection with the infiltration of B cells, macrophages, dendritic cells, neutrophils, CD4+ T cells and CD8+ T cells. Additionally, BLK expression was positively associated with the infiltration of B cells, dendritic cells, CD4+ T cells and CD8+ T cells. BRK expression was positively correlated with the infiltration of macrophages, neutrophils and CD4+ T cells. FRK expression was significantly positively correlated with the infiltration of macrophages, dendritic cells, neutrophils, CD4+ T cells and CD8+ T cells. FYN expression was positively connected with the infiltration of macrophages, dendritic cells, neutrophils and CD8+ T cells. SRC expression was positively related to the infiltration of B cells, macrophages, dendritic cells, neutrophils and CD4+ T cells. SRM expression was positively relevant to the infiltration of B cells, macrophages, neutrophils and CD4+ T cells. Taken together, the above results demonstrated that SFKs might play a dominant role in the immune infiltration of HCC.

Fig. 7.

Correlations between differentially expressed SFKs and immune cell infiltration (TIMER2.0). Spearman tests were used for analyzing the correlation coefficients (Rho). Positive correlation (p 0.05, Rho $>$ 0); Negative correlation (p 0.05, Rho 0); Not significant (p $>$ 0.05).

Multiple analysis tools combined with qPCR experiments suggested that among all SFK members, only FYN and SRC were dysregulated in expression and associated with OS, therefore, we mainly considered these two genes as candidate genes for predicting prognostic markers in HCC patients. The impact of FYN on HCC has rarely been reported, therefore, we first focused on identifying the effect of FYN on HCC in vitro and in vivo.

Western blot analysis revealed decreased protein expression levels in both HCC tissues (Supplementary Fig. 2A) and the human hepatoma cell line Huh7 (Supplementary Fig. 2B) by western blot analysis. We then generated FYN-overexpression or -knockdown Huh7 cell lines (Fig. 8A,B). Cell proliferation ability, detected by CCK-8 proliferation assay and colony formation assay, decreased in FYN-overexpression Huh7 cells and increased in FYN-knockdown Huh7 cells (Fig. 8C–F). Moreover, transwell migration and invasion experiments showed that overexpression of FYN inhibited cell migration and invasion, which was the opposite of FYN knockdown (Fig. 8G–J). Next, we examined the effects of FYN on HCC cell proliferation in vivo. In the subcutaneous xenograft tumor model, the tumor growth rate in mice injected with FYN-overexpression cells was significantly slower than that in mice injected with control Huh7 cells (Fig. 8K). FYN-overexpression Huh7 cells formed smaller tumors (365.5 $\pm$ 113.0 mm${}^3$ vs. 945.0 $\pm$ 156.7 mm${}^3$, Fig. 8L; 412.8 $\pm$ 64.79 mg vs. 918.3 $\pm$ 66.37 mg, Fig. 8M). Together, these data imply that overexpression of FYN, a potential tumor suppressor of HCC, suppresses tumor biological behavior in vitro and in vivo.

Fig. 8.

Overexpression of FYN suppresses tumor biological behavior in vitro and in vivo. (A,B) FYN overexpression (A) and knockdown (B) efficiency verification at the protein level. FYN protein levels were detected by western blot and relative levels were normalized to $\beta$-actin. Ctrl, control group; FYN, overexpression group; NC, negative control group; siFYN, knockdown group. (C–F) Overexpression of FYN inhibited the proliferation ability of Huh7 cells. CCK-8 (C,D) and colony formation assays (E,F) were used to detect proliferationin ability of Huh7 cells. Results were expressed as mean $\pm$ SD; n = 3; *p 0.05, **p 0.01, ***p 0.001; Student’s t-test. (G–J) Overexpression of FYN inhibited the migration and invasion of Huh7 cells. Transwell assays were performed to test the migration (G, H; left panel) and invasion (I, J; left panel) abilities of Huh7 cells. Scale bar, 100 $\mu$m. Quantification was achieved by ImageJ (version 1.53n, NIH, Bethesda, MD, USA) and shown in the right panels of G–J, respectively. Results were expressed as mean $\pm$ SD; n = 5; *p 0.05, ***p 0.001; Student’s t-test. (K) Overexpression of FYN suppressed tumor xenograft growth in vivo. Effects of FYN on tumor growth in subcutaneous xenograft tumor model. The tumors were excised on day 20. Three representative mice in each group were shown. n = 6. (L) The tumor volume progression of nude mice within the observation period (20 days). Results were expressed as mean $\pm$ SD; ***p 0.001; Student’s t-test. (M) The weight of tumors excised on day 20 were lighter in the FYN overexpression group than that in the control group. Values were mean $\pm$ SD; n = 6; ***p 0.001; Student’s t-test.

Several studies have shown that SRC has a tumor-promoting effect on HCC. Here, we studied its function through tumor biological behavior experiments in vitro. We first generated SRC-overexpression or -knockdown Huh7 cell lines (Fig. 9A,B). CCK-8 proliferation and colony formation assays revealed that knockdown of SRC suppressed cell proliferation, while overexpression of SRC promoted cell proliferation (Fig. 9C–F). Transwell migration and invasion assays demonstrated that knockdown of SRC inhibited cell migration and invasion, conversely, overexpression of SRC promoted cell migration and invasion (Fig. 9G–J). These results suggest that knockdown of SRC, a potential tumor promoter of HCC, suppresses the biological behavior of tumors in vitro.

Fig. 9.

Knockdown of SRC suppresses tumor biological behavior in vitro. (A,B) SRC overexpression (A) and knockdown (B) efficiency verification at the protein level. SRC protein levels were detected by western blot and relative levels were normalized to $\beta$-actin. (C–F) Knockdown of SRC inhibited the proliferation ability of Huh7 cells. CCK-8 (C,D) and colony formation assays (E,F) were performed to detect proliferationin ability of Huh7 cells. Results were expressed as mean $\pm$ SD; n = 3; *p 0.05, **p 0.01; Student’s t-test. (G–J) Knockdown of SRC inhibited the migration and invasion of Huh7 cells. Transwell assays were achieved to test the migration (G, H; left panel) and invasion (I, J; left panel) abilities of Huh7 cells. Scale bar, 100 $\mu$m. Quantification was achieved by ImageJ and shown in the right panels of G–J, respectively. Results were expressed as mean $\pm$ SD; n = 5; ***p 0.001; Student’s t-test.