3.4 PB inhibited liver fibrosis through downregulation of GLI1 expression
Recent studies have reported that a positive correlation exists between HSC activation and GLI1 overexpression (Chung et al., 2016; Guerrero-Juarez & Plikus, 2017; Zhang et al., 2017). Growing evidence indicate that GLI1 is a critical regulator of adult liver repair and hence, a potential diagnostic and/or therapeutic target in cirrhosis (Chen et al., 2020; Seki, 2016; Zhang et al., 2017). Based on the above, we investigated whether GLI1 was involved in the anti-hepatic fibrosis activity of PB. To determine whether PB regulated GLI1 activity, GLI-dependent luciferase activity was monitored, and the results showed that PB strongly repressed GLI-luciferase activities (Figure 4A). In addition, TGFβ1 stimulation significantly increased GLI1 expression, while PB treatment decreased GLI1 expression (Figure 4B, C). Similar results were also observed in PB treated mouse pHSCs (Figure 4D, E). However, RT-qPCR results showed that PB scarcely influenced the GLI2, GLI3 expression in LX-2 cells (Supplementary Fig. 2A). Furthermore, PB treatment downregulated GLI1 protein expression in BDL-induced fibrotic liver tissues (Figure 4F).
Next, we wonder whether PB mediated anti-fibrosis effect dependent on GLI1. First, we verified that knockdown ofGLI1suppressed the mRNA and protein expression of fibrogenic in LX-2 cells (Figure 5A, B). These results indicate that GLI1 indeed plays an important role in HSCs activation. Next, we measured Collagen I and αSMA expression in PB treated LX-2 cells with or without GLI1knockdown. As shown in Fig 5C, PB could not further reduce the Collagen I and αSMA expression in the context of GLI deficiency. In addition, the similar results were achieved when the cells were treated with GANT61, the most-widely used specific inhibitor of GLI1. GANT61 treatment decreased GLI-luciferase activity and GLI1 mRNA levels (Supplementary fig. 2B, C). However, the combination of PB and GANT61 showed no additive effect on the reduction of fibrogenic gene expression in LX-2 cells (Supplementary Fig. 2D). Furthermore, overexpression of GLI1 promoted Collagen I and αSMA expression, and reversed PB mediated inhibition of aforementioned proteins (Figure 5D). These results illustrated that PB inhibited HSC activation via downregulation of GLI1 expression.
3.5 PB blocked the nuclear localization of GLI1 in HSCs
Previous findings had shed light on GLI1 activity, which is tightly controlled through the regulation of nuclear import and the modulation of protein stability (Gulino, Di Marcotullio, Canettieri, De Smaele & Screpanti, 2012). GLI1 nucleus localization is critical for its growth-promoting function. Recently, Zhang et al. reported that GLI1 nuclear translocation leads to HSC contraction and cirrhotic portal hypertension (Zhang et al., 2020). We next tested if PB influenced the GLI1 nuclear localization. Expectedly, PB treatment blocked the translocation of GLI1 from the cytoplasm to the nucleus in a dose-dependent manner (Figure 6A). Nuclear and cytoplasmic fractionation analysis for GLI1 distribution reciprocated the immunofluorescence experiments in both LX-2 cells and mouse pHSCs (Figure 6B, C). Given that nuclear translocation of GLI1 resulting in increased GLI1 target gene expression (Kim, Kim, Cho, Kim, Kim & Cheong, 2010), we then measured GLI1 downstream target genes. The mRNA expression of well-known GLI1 downstream genes, including HHIP , CYCLIN D , CYCLIN E and C-MYC were dramatically down-regulated in both PB-treated LX-2 cells and BDL-PB treated mouse liver samples (Figure 6D, E). Similarly, GLI1 downstream genes were reduced in GANT61 treated LX-2 cells, whereas the combination of PB and GANT61 showed no additive effect on the reduction of downstream genes expression in LX-2 cells (Supplementary Fig. 2E). Taken together, these results indicated that PB repress the translocation of GLI1 from the cytoplasm to the nucleus, decreasing the expression of the downstream target genes.
3.6 PB inhibited LAP2 α-HDAC1 mediated deacetylation of GLI1
Published literatures suggest that GLI1 is temporarily inactive by acetylation (AcGLI1), therefore acetylation is an important modification to regulate cellular GLI1(Coni et al., 2017; Mirza et al., 2019). Thus, we speculated that PB may affect the acetylation of GLI to regulate its activity. Indeed, anti-acetylated lysine immunoprecipitates indicated that GLI1 are constitutively acetylated with PB treatment (Figure 7A). Acetylation directly inhibits the expression of transcription factors (Gurung, Feng & Hua, 2013), thus PB may inhibit liver fibrosis by upregulating the acetylation of GLI1, which inactivates it.
It has been reported that the acetylation of GLI1 is regulated by HDAC1(Canettieri et al., 2010; Falkenberg et al., 2016) and LAP2α(Mirza et al., 2019). Mechanistically, LAP2α recruits HDAC1 to GLI1, physically interacts with HDAC1, and scaffolds a complex with GLI1(Mirza et al., 2019). Next, we wonder whether PB showed any effect on LAP2α and HADC1 expression. However, no significant protein variations were observed in both LAP2α and HADC1 after PB treatment in LX-2 cells (Figure 7B). Given that LAP2α/HADC1 mediated GLI deacetylation play an important role in GLI activity, we next wonder whether PB affect LAP2α/GLI interaction or LAP2α/HADC1complex formation. As shown in Fig 7C, PB showed no effect on the interaction between LAP2α/GLI1, however, strongly inhibited the LAP2α/HDAC1 complex formation in HEK 293T cells (Figure 7 D). Congruently, endogenous LAP2α and HDAC1 interaction was significantly disrupted by the PB treatment in LX‐2 cells (Figure 7E), and this interaction was also reduced in CCl4-induced mouse fibrotic liver tissue treated with PB (Figure 7F).
Previous studies have demonstrated that nucleoplasmic complex LAP2α-HDAC1 could protect GLI1 from acetylation and promote GLI1 activation(Mirza et al., 2019). As described in the above co-IP results, confocal microscopy assay was used to determine the subcellular localization of HDAC1 (red) and LAP2α (green) to confirm this resultin situ . Our results showed that PB had no influence on the nuclear localization of LAP2α with or without TGFβ1 treatment. However, PB obviously decrease the nuclear localization of HDAC1 after TGFβ1 treatment (Figure 7G), indicating that PB administration may disturb the intracellular colocalization of LAP2α with HDAC1. These results indicated that, instead of interfering with LAP2α and GLI interaction, PB inhibited LAP2α-HDAC1 complex formation, which may responsible for PB mediated GLI inactivation.
Next, we employed the histone deacetylase inhibitor vorinostat to further validate the results. Vorinostat decreased the levels of fibrogenic markers which is in consistent with previous published results(Park et al., 2014), and no further reduction was observed when combination with PB (Figure 8A). Immunofluorescence assay further confirmed that PB and vorinostat downregulated GLI1 expression and nuclear localization (Figure 8B). In addition, PB and vorinostat increased the acetylation of GLI1 (Figure 8C). Due to the protective effect of LAP2α on GLI1 deacetylation, we next verify that whether LAP2α deficiency can attenuate HSC activation by treating LX-2 cells with LAP2α siRNA. The fibrogenic markers were suppressed in LAP2α siRNA treated LX-2 cells (Figure 8D, E). Similarly, combination of LAP2α silencing and PB treatment blocked nucleus translocation of GLI1 in LX-2 cells (Figure 8F). In addition, LAP2α-silencing and PB treatment increased GLI1 acetylation (Figure 8G). Taken together, our results indicate that PB interferes with LAP2α-HDAC1 complex formation, thereafter inhibits GLI1 deacetylation and downstream signaling. PB could be a promising therapeutic agent for liver fibrosis.