Shikonin induced pyroptosis in tumour cells via GSDME
To investigate the anticancer effects of shikonin (SK), BGC and SGC cells were stimulated with SK. SK suppressed BGC and SGC cell proliferation in a time- and dose-dependent manner (Fig. 1A and Sup Fig. 1A and B). BGC and SGC cells stimulated with SK showed PI-positive staining and lactate dehydrogenase (LDH) release (Fig. 1A and Sup Fig. 1C). These cells showed typical morphological features of pyroptosis (Fig. 1B). Both GSDMD and GSDME are involved in pyroptotic pathways, and RIPK3 is involved in necroptotic pathways. We first investigated whether these molecules are involved in SK-induced cell death. GSDME knockdown significantly inhibited LDH release, whereas GSDMD or RIPK3 knockdown failed to inhibit LDH release (Fig. 1C). We then examined GSDME expression in several cancer cell lines. GSDME was highly expressed in SGC and BGC cells but not in AGS cells (Fig. 1D). SK treatment induced the cleavage of GSDME in SGC and BGC cells (Fig. 1E). In GSDME knockdown cells, SK stimulation led to an apoptotic morphology and less LDH release (Fig. 1F and G). GSDME cleavage was also abolished after GSDME knockdown (Fig. 1H). The overexpression of WT GSDME in AGS cells also induced GSDME cleavage and LDH release after SK stimulation (Sup Fig. 1D and E). Collectively, these results suggest that SK can induce pyroptosis in some tumour cells, and this induction is mediated by GSDME.
BAX/caspase-3 activation was involved in SK-induced pyroptosis
We next investigated the molecular mechanisms of the GSDME-mediated pyroptosis induced by SK. We pretreated cells with a pan-caspase inhibitor, Z-VAD-FMK, and then stimulated them with SK. Z-VAD-FMK treatment significantly prevented cell swelling (Fig. 2A), LDH release (Fig. 2B and C) and GSDME cleavage (Fig. 2B and C). These results suggest that caspase may be involved in GSDME-mediated pyroptosis. Previous studies have found that caspase-3 can cleave GSDME in GSDME-mediated pyroptosis [7]. To further clarify the essential role of caspase-3 in SK-induced pyroptosis, we knocked down caspase-3 in SGC and BGC cells. As shown in Fig. 2D and E, caspase-3 knockdown significantly prevented cell death and reduced GSDME cleavage. BAX activation led to pore formation on the mitochondrial outer membrane, which would induce the activation of the caspase cascade [18]. We then investigated whether BAX activation induces caspase-3 cleavage and is involved in SK-induced pyroptosis. After BAX knockdown (Sup Fig. 2A), both caspase-3 and GSDME cleavage were reduced compared with the control (Fig. 2F). LDH release was also diminished in BAX-knockdown cells (Fig. 2F). Taken together, these data indicated that SK treatment activated the BAX/caspase-3/GSDME cascade to induce pyroptosis.
SK induced protective autophagy in tumour cells
It has been reported that some natural compounds can modulate crosstalk between apoptosis and autophagy [19]. However, it is unknown whether there is crosstalk between pyroptosis and autophagy after SK stimulation. We first examined the expression level of LC3-II, an important marker of autophagosomes, by western blotting. As shown in Fig. 3A, after SK stimulation, the LC3-II expression level was markedly increased in a time-dependent manner. We further examined autophagic flux by transducing mRFP-GFP-LC3 lentivirus into cells. SK stimulation increased the number of red puncta (autophagolysosomes) (Fig. 3B), indicating that SK stimulation promotes autophagic flux. Next, we investigated whether SK-induced autophagy affected GSDME-mediated pyroptosis. Beclin-1 and ATG5, key molecular regulators of autophagy, were knocked down in cells by shRNA (Sup Fig. 3A and 3B). SK stimulation markedly increased the change in pyroptotic morphology (Fig. 3C), cell death (Fig. 3D, G and Sup Fig. 3D), caspase-3 and GSDME cleavage compared with the control group (Fig. 3E, F and Sup Fig. 3C). Collectively, these findings demonstrated that autophagy exerts cytoprotective effects in SK-induced pyroptosis.
SK induced pyroptosis and autophagy through ROS-initiated signalling
SK stimulation can reportedly induce cell death via the generation of ROS in some tumour cells [20]. However, whether SK-elevated ROS generation is associated with GSDME-mediated pyroptosis is still unknown. We first measured the ROS levels in tumour cells after SK treatment. SK treatment significantly increased ROS levels in cells (Fig. 4A). Importantly, treatment with NAC, a ROS scavenger, dramatically reduced ROS production in SK-treated cells (Fig. 4A). Moreover, NAC treatment also markedly attenuated cell death (Fig. 4B), the change in pyroptotic morphology (Fig. 4C), caspase-3 and GSDME cleavage compared with the control (Fig. 4D). We further investigated whether ROS generation in SK-treated cells was affected by caspase-3 and GSDME. As shown in Fig. 4E and F, SK-induced ROS generation was not affected in caspase-3- or GSDME-knockdown cells compared with the control groups. This result suggested that ROS are upstream signals of caspase-3 and GSDME. Taken together, these data indicate that SK induces pyroptosis through ROS-initiated signalling to activate caspase-3 and GSDME.
The regulation of autophagy by ROS has been observed in cancer cells, and autophagy, in turn, can reduce oxidative damage [4]. However, whether SK-induced protective autophagy is also regulated by ROS is still unknown. We first treated cells with NAC and examined LC3-II expression and autophagic flux after SK treatment. The NAC-treated cells showed lower LC3-II expression than the control cells (Fig. 4G). The number of red puncta was also reduced in the NAC-treated cells (Fig. 4G). Collectively, these results suggest that SK-induced ROS generation also regulates autophagy and ultimately affects GSDME-mediated pyroptosis.
MAPK14/p38α-dependent modulation of SK-induced autophagy and pyroptosis in SK-induced pyroptosis
To identify the potential factors that regulate GSDME-mediated pyroptosis in SK-treated cells, we performed a high-sensitivity mass spectrometry assay that compared SK-treated cells and control cells. The differentially expressed proteins between SK-treated cells and control cells that are involved in cell death and autophagy are shown in Fig. 5A. One of the proteins, MAPK14, showed significantly decreased expression in the SK-treated group. MAPK14 was previously found to be a stress-activated protein kinase that can sense ROS and affect autophagy [21, 22]. We then investigated whether MAPK14 was involved in SK-induced pyroptosis and autophagy. In MAPK14-knockdown cells (Sup Fig. 4A), the number of red puncta and LC3-II expression were increased after SK treatment (Fig. 5B, C and E), suggesting that MAPK14 modulates autophagy. Next, we examined whether MAPK14 affects GSDME cleavage and pyroptosis. In MAPK14-knockdown cells, cell viability was increased compared with control cells after SK treatment (Fig. 5D). Caspase-3 and GSDME cleavage were also reduced in MAPK14-knockdown cells (Fig. 5E). Taken together, these data indicated that MAPK14 modulates SK-induced autophagy and eventually affects pyroptosis.
SK treatment decreased tumorigenicity and induced GSDME-mediated pyroptosis in vivo
Next, we sought to investigate the effect of SK administration on tumour growth in vivo by using a mouse model. We first examined whether SK treatment induces GSDME-mediated pyroptosis in a mouse cell line. EMT6 cells, which highly express GSDME, were used to analyse cell death and GSDME cleavage [7]. SK treatment significantly induced caspase-3, GSDME cleavage, and cell death in EMT6 cells (Fig. 6A and B). However, in GSDME-knockdown cells (Sup Fig. 4B), caspase-3, GSDME cleavage, and cell death were reduced compared with the control cells. Next, we ascertained the effect of SK on the growth of xenografts in immune-competent BALB/c mice. As shown in Figure 6C, SK administration significantly inhibited the growth of EMT6 solid tumours from Day 17. Moreover, the tumour volume and weight of SK-treated mice were significantly lower than those of the control mice (Fig. 6D and E). Immunohistochemical analysis of the xenografts showed that SK administration significantly enhanced the expression of caspase-3 and GSDME (Fig. 6F and G). Collectively, these results suggested that SK inhibits the growth of tumours in vivo by inducing pyroptosis.