GBA overexpression confers poor prognosis on HCC patients and is associated with impaired autophagic degradation in human HCC tissues
The microarray data in Fig. 1a show that the level of GBA mRNA was significantly higher in human HCC tissues than in noncancerous liver tissues (cancerous vs. noncancerous, 7.541.10 vs. 5.731.02, P=0.03), and this result was validated by real-time quantitative PCR using our clinical cohorts (cancerous vs. noncancerous, 0.070.07 vs. 0.010.01, P<0.001; Fig. 1b). To evaluate the clinical significance of GBA in human HCC, the median value (0.506) of GBA expression in all 99 human HCC tissues was used to divide the patients into the high and low GBA expression groups (n=50 and 49, respectively). Statistical analysis showed that more HCC patients in the high GBA expression group than in the low GBA expression group showed high preoperative serum AFP levels (0.01 a Signal values indicating GBA mRNA expression detected by analysis of a microarray containing three pairs of noncancerous and cancerous liver tissues collected from three HCC patients (cancerous vs. noncancerous, P=0.03). b Expression levels of GBA mRNA detected by quantitative real-time PCR using 99 HCC and 18 noncancerous liver tissues (cancerous vs. noncancerous, P<0.001). Each dot represents information from one patient. The data are presented as the meanSEM values; p values were calculated by two-tailed unpaired t test. c, d KaplanMeier curves of overall (P=0.027) and disease-free (P=0.038) survival of HCC patients stratified according to the GBA mRNA expression level based on our clinical cohort; e, f KaplanMeier curves of overall (P=0.015) and disease-free (P=0.040) survival of HCC patients based on the TCGA dataset, respectively. The data are presented as proportions; p values were calculated by the log-rank test. g Immunohistochemical analysis of the GBA and SQSTM1/p62 protein expression levels in clinical HCC and noncancerous liver tissues. Original magnifications, 100 and 400; scale bars, 100m and 25m. h Immunofluorescence analysis of the subcellular localization and expression level of LC3B protein in clinical HCC and noncancerous liver tissues. Original magnification, 200; scale bar, 50m. i Semiquantitative results of the immunoreactive scores of GBA and SQSTM1/p62 proteins and the intensity of LC3B puncta in clinical HCC and noncancerous liver tissues. *p<0.05, **p<0.01 and ***p<0.001 compared with the noncancerous liver tissues. To investigate the function of GBA in HCC, two human HCC cell lines, HepG2 and MHCC-97H, which were demonstrated to have high GBA expression (Supplementary Fig. 1a), were transfected with three si-GBA plasmids (si-GBA-313, si-GBA-1255 and si-GBA-1620) to reduce GBA expression. Then, the western blot data demonstrated that the expression levels of GBA in all si-GBA transfection groups were markedly lower than those in the si-NC transfection group (all P<0.05, Fig. 2a and Supplementary Fig. 2a). Measurement of the cell proliferation ability by a CCK-8 assay indicated that the OD values of both HepG2 and MHCC-97H cells transfected with the si-GBA plasmids were significantly lower than those transfected with the si-NC plasmid at different time points (HepG2 cells: after 48h, MHCC-97H cells: after 24h, all P<0.05; Fig. 2b and Supplementary Fig. 2b). In addition, the knockdown of GBA effectively induced apoptosis in HCC cells by increasing the percentages of early-stage and late-stage apoptotic cells (both HepG2 and MHCC-97H cells, P<0.01; Fig. 2c, Supplementary Fig. 2d and Supplementary Fig. 2g). Moreover, evaluation of cell viability by a colony formation assay showed that the more colonies formed in the si-NC transfection groups than in the si-GBA transfection groups (both HepG2 and MHCC-97H cells, P<0.01; Fig. 2c, Supplementary Fig. 2e and Supplementary Fig. 2h). We also observed G1 arrest in both HepG2 and MHCC-97H cells when GBA expression was inhibited (all P<0.05, Fig. 2c, Supplementary Fig. 2c and Supplementary Fig. 2f). In contrast, the inhibitory effects of GBA knockdown on HCC cell proliferation, cell cycle progression and cell colony formation and its inductive effect on apoptosis were significantly reversed by treatment with the GBA activator LTI-291 (all P<0.05, Fig. 4a). a Protein expression levels of GBA in HepG2 cells transfected with three si-GBA plasmids (si-GBA-313, si-GBA-1250 and si-GBA-1620), as determined by western blot analysis. Representative column charts showing relative protein expression. b Viability of HepG2 cells transfected with si-GBA-313 and si-GBA-1620, as determined by a CCK-8 assay. c Flow cytometric, apoptosis and cloning efficiency analyses of the cell cycle distribution of HepG2 cells transfected with si-GBA-313 and si-GBA-1620. The data are expressed as the meanSD values. *p<0.05, **p<0.01, ***p<0.001; si-GBA groups compared with the si-NC group. d Protein expression levels of GBA, SQSTM1/p62 and LC3B in HepG2 cells of different groups (BAF: 10nM), as determined by western blot analysis. GAPDH was used as the internal control. e TEM images of different groups of cells (LTI-291: 10nM, BAF: 10nM). The red arrows indicate mitochondria, the green arrows indicate rough endoplasmic reticulum, and the yellow arrows indicate autophagosomes. Original magnifications, 8000 and 20,000; scale bars, 2m and 500nm. The data are expressed as the meanSD values. *p<0.05, **p<0.01, and ***p<0.001 compared with the si-NC group; #p<0.05, ##p<0.01, and ###p<0.001 compared with the si-NC-BAF group; @p<0.05, @@p<0.01, and @@@p<0.001 compared with the si-GBA group. It is worth noting that the protein expression levels of LC3B and SQSTM1/p62 were significantly increased in HCC cells with si-GBA transfection, and these levels were further increased by combination treatment with si-GBA transfection and the late-stage autophagy inhibitor BAF compared to those in cells with si-GBA transfection or BAF treatment alone (LC3B protein, all P<0.05 when using two si-GBA plasmids; SQSTM1/p62 protein, all P<0.05 when using the si-GBA-313 plasmid, Fig. 2d). In addition, transmission electron microscopy was used to observe the ultramicrostructure of HCC cells in the different treatment groups. Remarkably, numerous double- or multimembrane autophagosomes were found in GBA knockdown cells, and the number of these structures was increased by combination treatment with si-GBA transfection and BAF (all P<0.01, Fig. 2e). In contrast, the number of autophagosomes in HCC cells treated with the GBA activator LTI-291 was lower than that in HepG2 cells with or without si-GBA transfection (Fig. 2e). Notably, combination treatment with LTI-291 and BAF significantly reversed the increase in autophagosome accumulation induced by BAF alone (P<0.01, Fig. 2e). Consistent with these findings, reduced protein expression levels of LC3B and SQSTM1/p62 were found in HCC tissues from tumor-bearing nude mice with GBA activation by LTI-291 treatment (Fig. 7), which was in line with the observations made by transmission electron microscopy (Fig. 6). These findings suggest that GBA knockdown may result in impaired autophagic degradation in HCC cells, which was partially reversed by activation of GBA. The ART probe (Fig. 3a), with a minimalist alkyne-containing linker incorporated into ART29, was employed to identify cellular targets of ART by large-scale chemoproteomic experiments. HepG2 cells were treated with the ART probe. Proteins in the cell lysates were conjugated to biotin-N3, and the resulting probe-labeled proteins were affinity-purified and identified by LCMS/MS. Control experiments with the probes in the presence of the corresponding parent competitor (ART) were carried out concurrently, and the results were used to distinguish between real targets and background labeling. The identified protein hits were analyzed on corresponding volcano plots showing the -log2 of the competition ratio (probe/probe with excess competitor) plotted against the statistical significance (log10 p value). Proteins with a p value less than 0.05 and a competition ratio greater than 1.5 were considered to be significant hits. As expected, GBA was successfully identified as one of the direct targets of ART (Fig. 3b) and further validated by a pulldown assay followed by WB (Fig. 3c), which proved the direct binding of ART to GBA. Consistent with the above findings, the SPR data showed that ART bound strongly to GBA with a KD value of 40M (Fig. 3d, e). a Chemical structure of the ART probe. b Mass spectrometry-based profiling of the ART probe (20M) in the presence of excess ART (10). c Pulldown/western blot results for target validation of the ART probe (20M). d BIAcore surface plasmon resonance (SPR) kinetic analyses of ART binding to GBA. Sensorgram and saturation curve of the titration of ART on GBA immobilized on a CM5 chip. e Binding curves of ART to GBA were fit to a steady-state affinity model to obtain the KD value. To explore the effects of ART on HCC cell malignancy, the cytotoxicity of ART was evaluated in HepG2 and MHCC-97H cells by a CCK-8 assay. As shown in Supplementary Fig. 3a, e, ART induced cytotoxicity in these two HCC cell lines in a dose-dependent manner and was also confirmed to reduce the number of cell colonies in the colony formation assay (all P<0.05, Supplementary Fig. 3d for HepG2 cells; Supplementary Fig. 3h for MHCC-97H cells). In addition, G1 arrest was observed in both HepG2 and MHCC-97H cells treated with various concentrations of ART (all P<0.05, Supplementary Fig. 3b for HepG2 cells; Supplementary Fig. 3f for MHCC-97H cells). We also found that ART increased the apoptosis rate of these two HCC cell lines in a dose-dependent manner (all P<0.05, Supplementary Fig. 3c for HepG2 cells; Supplementary Fig. 3g for MHCC-97H cells). All these effects were similar to the influence of GBA knockdown in HCC cells, as shown in Fig. 2 and Supplementary Fig. 2. The above data indicate that HepG2 cells were more sensitive than MHCC-97H cells to ART. Further study was conducted mainly in HepG2 cells. To further determine whether GBA might be responsible for the inhibitory effects of ART in HCC cell malignancy, the GBA activator LTI-291 (10nM) was used in combination with ART (20M, 0.5 IC50). The inhibitory effects of ART on HCC cell proliferation, cell cycle progression and cell colony formation, as well as its inductive effect on apoptosis, were reversed by combined treatment with LTI-291 (10nM) (all P<0.05, Fig. 4a). a Proliferation of HepG2 cells treated with the GBA activator LTI-291 (10nM) alone or in combination with ART (20M, 0.5 IC50), as evaluated using a CCK-8 assay. Apoptosis of HepG2 cells treated with the GBA activator LTI-291 (10nM) alone or in combination with ART (20M, 0.5 IC50), as determined by a TUNEL assay. Flow cytometric analysis of the effects of LTI-291 treatment alone or in combination with ART on the cell cycle distribution of HepG2 cells. Colony formation ability of HepG2 cells treated with the GBA activator LTI-291 alone or in combination with ART, as evaluated by a colony formation assay. The data are expressed as the meanSD values. *p<0.05, **p<0.01, and ***p<0.001 compared with the blank control group; #p<0.05, ##p<0.01, and ###p<0.001 compared with the LTI-291 monotherapy group; @p<0.05, @@p<0.01, and @@@p<0.001 compared with the ART monotherapy group. (b) Transmission electron microscopy images of HepG2 cells treated with ART at different concentrations (ART: 20M, 0.5 IC50; LTI-291: 10nM; BAF: 10nM). The red arrows indicate mitochondria, the green arrows indicate endoplasmic reticulum, and the yellow arrows indicate autophagosomes. Original magnifications, 8000 and 20,000; scale bars, 2m and 500nm. (c) Protein expression levels of GBA, SQSTM1/p62 and LC3B in HCC cells treated with ART at different concentrations (40M, IC50), as determined by western blot analysis. GAPDH was used as the internal control. The data are expressed as the meanSD values. *p<0.05, **p<0.01, and ***p<0.001 compared with the blank control HepG2 cell group. ART promoted the conversion of LC3B-I to LC3B-II and elevated the protein expression of SQSTM1/p62 but reduced the protein expression of GBA in HCC cells (all P<0.05, Fig. 4c). Interestingly, ART treatment significantly increased the number of autophagosomes in HCC cells, and this increase was enhanced by combined treatment with si-GBA transfection but reversed by the combined treatment with LT1-291 (all P<0.05, Fig. 4b), as evaluated by transmission electron microscopy. To determine whether the accumulation of autophagosomes induced by ART treatment in HCC cells was due to impaired degradation, we measured autophagic flux with BAF, which is an inhibitor of late-stage autophagy, and found that the number of autophagosomes in HCC cells treated with the combination of BAF and ART or the combination of BAF, ART and si-GBA transfection was greater than that in cells treated with ART alone (both P<0.05), Fig. 4b. To verify the hypothesis that ART can attenuate inflammation-induced hepatic carcinogenesis, a stable rat model with DEN-induced progression of hepatic fibrosis to cirrhosis to cancer was established and used to assess the therapeutic effects of ART. No rats treated with ART (28.8mg/kg) died, and DEN-induced injury in rats resulted in progressive liver fibrosis and cirrhosis followed by HCC over different time periods. In detail, repeated administration of DEN (0.01%) for 8 weeks caused fibrosis, and advanced fibrosis and cirrhosis occurred after 12 weeks. All rats in the model group had marked neoplastic lesions by 18 weeks of DEN administration. Multiple white nodules were observed on the liver surface in mice in the DEN model group, which were effectively attenuated by the administration of ART (Supplementary Fig. 4a). Histologically, we observed normal hepatic lobular architecture, normal hepatocytes and a sinusoidal architecture without fibroplasia and inflammatory cell infiltration but only scattered foci of fatty degeneration in normal control group and normal-ART group rats (marked with green arrows in Supplementary Fig. 4b). After administration of DEN for 8 weeks, slight microscopic fibrosis occurred and continued to develop into cirrhosis, as well as hemorrhagic necrosis with inflammatory cell foci around fibrotic tissue, at the 12th week (marked with red arrows in Supplementary Fig. 4b). Subsequently, the incidence of tumors was markedly increased. All the rats in the DEN model group developed HCC by week 18. The white nodules observed on the surface of the livers were histologically confirmed to be HCC; the cells in these nodules had clear, eosinophilic or hyperbasophilic cytoplasm, and all had enlarged and hyperchromatic nuclei (marked with blue arrows in Supplementary Fig. 4b). Interestingly, administration of ART effectively reversed the pathological changes in the liver by reducing the inflammatory level and malignancy degree at the indicated times (Supplementary Fig. 4b). These findings were in line with the above gross anatomical observations. In addition, the body weights and liver indices of rats in the DEN model and DEN+ART treatment groups were significantly lower and higher, respectively, than those in the normal control and normal-ART groups at the indicated times (all P<0.05, Supplementary Fig. 5a, b). There were no significant differences in body weights or liver indices between the normal control and normal-ART groups or between the DEN model and DEN+ART groups (Supplementary Fig. 5a, b). Moreover, the serum levels of several liver function markers, such as ALT, AST and ALP, in the DEN model group were significantly higher than those in the normal control and normal-ART groups and remained high until week 18 (all P<0.05, Supplementary Fig. 6ac). After DEN administration was begun, the serum level of the fibrosis marker HA was high, with a statistically significant increase from baseline, from week 12 to week 18 (all P<0.05, Supplementary Fig. 6d). A gradual increase in the tumor-related serum marker AFP was observed beginning in week 12, and its serum levels were significantly higher than those in the normal control and normal-ART groups (all P<0.01, Supplementary Fig. 6e). Importantly, administration of ART effectively reduced the increasing serum levels of ALT, AST, ALP, HA and AFP induced by DEN administration (all P<0.05, Supplementary Fig. 6ae). To determine whether the inhibitory effects of ART on inflammation-induced hepatic carcinogenesis were associated with GBA-mediated autophagic degradation, the expression levels and subcellular localization of GBA, SQSTM1/p62 and LC3B proteins in cancerous and noncancerous liver tissues of the different groups were assessed by immunohistochemical and immunofluorescence assays. Figure 5a, b shows the stronger positive cytoplasmic staining of GBA and SQSTM1/p62 proteins induced by DEN. Following administration of ART, the protein expression level of GBA was significantly decreased but that of SQSTM1/p62 was markedly enhanced (all P<0.05, Fig. 5a, b). Consistent with these findings, our immunofluorescence assay showed a marked number of GFP-LC3B puncta in the cytoplasm of tumor cells in liver tissues after DEN administration, and the number of these puncta was also increased by ART (all P<0.05, Fig. 5c). a, b Immunostaining and immunoreactive scores of GBA and SQSTM1/p62 proteins in liver tissues of rats with DEN-induced HCC that were treated with or without ART (20M, 0.5 IC50). Original magnifications, 100 and 400; scale bars, 100m and 25m. c Immunofluorescence analysis of LC3B protein expression in liver tissues of rats with DEN-induced HCC that were treated with or without ART (20M, 0.5 IC50). Original magnification, 200; scale bar, 50m. The data are expressed as the meanSD values. *p<0.05, **p<0.01, and ***p<0.001 compared with the normal control group; #p<0.05, #p<0.01, and ###p<0.001 compared with the DEN model group. To determine the anti-HCC effect of ART in vivo, an orthotopic tumor model, which may naturally represent a native tumor because of the growth of tumor cells in the liver, was established in nude mice, and the mice were then treated with ART (low dose: 5mg/kg, moderate dose: 10mg/kg, high dose: 20mg/kg). As shown in Fig. 6a, administration of ART effectively decreased the tumor size compared with that in the model group. Next, we examined the regulation of autophagy by ART in tumor samples. Consistent with the in vitro findings, administration of ART significantly enhanced autophagosome accumulation in tumor tissues according to transmission electron microscopy observation (P<0.05, Fig. 6b). Interestingly, the number of autophagosomes in mice treated with the combination of ART and LTI-291 was decreased and increased compared to that in mice treated with ART or LTI-291 alone, respectively (Fig. 6b). We further revealed the colocalization of GBA with LC3 and SQSTM1/p62 proteins in tumor tissues from mice in the different groups by immunofluorescence assays. Notably, in tumor samples, the protein levels of LC3B and SQSTM1/p62 were increased but the level of GBA was reduced by ART (Fig. 7a, b), and both of these changes were reversed by LTI-291 (Fig. 7a, b), implying an association between the anticancer effects of ART and GBA-mediated autophagic degradation. a Generation of neoplastic lesions in the different groups. b Transmission electron microscopy images of the different groups. The red arrows indicate mitochondria, the green arrows indicate rough endoplasmic reticulum, and the yellow arrows indicate autophagosomes. Original magnifications, 8000 and 20,000; scale bars, 2m and 500nm. a, b Double immunofluorescence staining of GBA/p62 and GBA/LC3B proteins in liver tissues of orthotopic tumor-bearing mice in the different groups (ART: 20M, 0.5 IC50; LTI-291: 10nM; BAF: 10nM). Original magnification, 200; scale bar, 100m. The data are expressed as the meanSD values. *p<0.05, **p<0.01, and ***p<0.001 compared with the normal control group; #p<0.05, ##p<0.01, and ###p<0.001 compared with the HepG2 group; +p<0.05, ++p<0.01, and +++p<0.001 compared with the HepG2+LTI-291 group. Read the original post:
Preclinical investigation of artesunate as a therapeutic agent for hepatocellular carcinoma via impairment of glucosylceramidase-mediated autophagic...