Inhibition of proteasome triggers M-phase arrest and multi-polar spindle formation

We first assessed whether inhibition of proteasome affected the fates of cells in M-phase. Staining of Ser-10-phosphorylated histone H3 (pH3-S10) was used to mark M-phase cells. Flow cytometry analysis showed that treatment with BTZ significantly increased the ratio of cells at M-phase (Fig. 1a; Supplementary Fig. 1a, b). And BTZ treatment also enhanced the proportion of G2-phase cells in a dose-dependent manner, and high dose of BTZ showed a reduced capability to increase the fraction of M-phase cells due to significant G2-phase arrest (Fig. 1a; Supplementary Fig. 1a, b). These findings were validated with the treatment of carfilzomib (CFZ) (Supplementary Fig. 1c), the second-generation proteasome inhibitor,25 which irreversibly binds to 20S-CR. We then validated the effects of proteasome inhibition by silencing the individual AAA+ ATPase or representative non-ATPase subunits of proteasomal 19S regulatory particle (Supplementary Fig. 1d, e). Knockdown of PSMC5, but not other PSMCs, mimicked the effects of BTZ and CFZ in increasing both G2- and M-phase populations in all three cell lines examined (Fig. 1b; Supplementary Fig. 1f, j). However, silencing non-ATPase subunits of 19S regulatory particle, like Rpn10, Rpn13 (ubiquitin receptors) or Rpn11 (de-ubiquitinating enzyme),18 failed to raise the population of G2-phase cells (Fig. 1c), implying that inhibiting the non-ATPase subunits of 19S-RP may not mimic the effects of BTZ.

Fig. 1

Inhibition of proteasome induces M-phase arrest, multi-polar spindle formation and mitotic catastrophe. a–c The effects of proteasome inhibition on cell cycle progression. SNU449 cells were treated with the indicated dose of BTZ (a) for 30 h or transfected with the indicated RNA duplexes for 60 h (b, c), then stained for Ser-10-phosphorylated histone H3 (pH3-S10) to indicate M-phase cells and stained with propidium iodide (PI) to indicate DNA content, followed by FACS for phase distribution of the cell cycle. d–i Inhibition of proteasome induced mitotic arrest, multi-polar spindle formation and ballooning bubbles from cell membranes. SNU449 subline that stably expressed histone H2B-EGFP and mCherry-α-tubulin were treated with vehicle or 30 nM BTZ (d, f, h), or transfected with NC or siPSMC5-1/2 (mixture of siPSMC5-1 and siPSMC5-2) for 24 h (e, g, i), followed by live-cell imaging for 46 h (d, f, h) or 70 h (e, g, i). For d (Vehicle, n = 25; BTZ, n = 14) and e (NC, n = 28; siPSMC5-1/2, n = 27), the time from nuclear envelope breakdown (NEBD) to the end of anaphase or cell death was designated as mitotic duration (right panel). White arrows indicate the large bubbles blowing from the plasma membrane. Scale bar, 5 μm. For f and g, cell death was determined by the emergence of pyroptosis characteristics or cell detachment, and the fractions of cells died at interphase or M-phase were quantified based on at least 118 cells in each group. For h (Vehicle, n = 25; BTZ, n = 14) and i (NC, n = 28; siPSMC5-1/2, n = 27), the fates of individual mitotic cell are shown. For d, e, h and i, the time point of NEBD was set as 0. j, k Inhibition of proteasome caused multi-polar spindle formation. SNU449 cells were treated with vehicle or 15 nM BTZ for 30 h (j), or transfected with the indicated RNA duplexes for 60 h (k), then stained for pericentrin (PCNT, red), α-tubulin (TUBA, green) and DAPI (blue) to indicate centrosome, spindle and chromosome, respectively. The proportion of mitotic cells possessing multi-polar spindles was calculated (right panel). Scale bar, 2.5 μm. Error bars: SEM from at least three independent experiments. One-way ANOVA (a–c and k) and Student’s t test (d, e and j) were used. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns not significant

The fates of mitotic cells after proteasome inhibition were then examined by the live-cell imaging under a time-lapse microscopy, using SNU449 subline stably expressing histone H2B-EGFP and mCherry-α-tubulin fusion proteins, which indicated chromosomes and microtubules, respectively. Upon treatment with BTZ or CFZ or silencing of PSMC5, a great majority of mitotic cells showed a significant extension of mitotic duration (mean time of vehicle vs. BTZ: 46 vs. 1857 min; vehicle vs. CFZ: 43 vs. 515 min; NC vs. siPSMC5: 40 vs. 530 min) (Fig. 1d; Supplementary Fig. 2a; Fig. 1e). Compared to control group, inhibition of proteasome resulted in a significantly increased cell death at M-phase (vehicle vs. BTZ: 0% vs. 8.8%; vehicle vs. CFZ: 0% vs. 12.1%; NC vs. siPSMC5-1/2: 0% vs. 14.4%) (Fig. 1f; Supplementary Fig. 2b; Fig. 1g), while a very low cell population died at interphase in both control group and proteasome-inhibiting group (vehicle vs. BTZ: 2.54% vs. 2.04%; vehicle vs. CFZ: 2.2% vs. 2.4%; NC vs. siPSMC5-1/2: 1.60% vs. 2.06%) (Fig. 1f; Supplementary Fig. 2b; Fig. 1g), indicating that most of the BTZ-, CFZ- or siPSMC5-treated cells which underwent mitotic arrest eventually died without exit from M-phase. We also observed that during mitotic arrest, multi-polar spindles appeared after spindle bipolarization (Fig. 1h; Supplementary Fig. 2c; Fig. 1i), and further immunofluorescent staining for microtubules and centrosomes confirmed that a large proportion of mitotic cells in BTZ- or siPSMC5-treated group displayed multi-polar spindles (Fig. 1j, k; Supplementary Fig. 3a–d). These results suggest that inhibition of proteasome may promote mitotic catastrophe by inducing aberrant spindle assembly and mitotic failure.

Inhibition of proteasome induces pyroptosis in M-phase via cGAS-caspase-3-GSDME cascade

Notably, live-cell imaging assays disclosed that the mitotic cells in BTZ/CFZ/siPSMC5 groups exhibited morphology features of pyroptosis, that is, cells swell and form balloon-like membrane structure (Fig. 1d; Supplementary Fig. 2a; Fig. 1e), which we termed mitotic pyroptosis. We therefore verified the effects of BTZ/siPSMC5 on the mitotic pyroptosis of different tumor cell lines, based on characteristic morphology, cleavage of gasdermin, and release of LDH. Compared to control group, BTZ treatment or PSMC5 silencing significantly increased the proportion of cells with pyroptosis morphology (Fig. 2a, b; Supplementary Fig. 4a, b) and induced the release of LDH (Fig. 2c, d; Supplementary Fig. 4c, d), with very few cells undergoing apoptosis (Supplementary Fig. 4e, f).

Fig. 2

Inhibition of proteasome induces mitotic pyroptosis via GSDME. a, b Proteasome inhibition induced morphology of pyroptosis. Five random fields in each well were captured and then subjected to analysis for the rate of cells with pyroptosis morphology. One of the five fields is shown as representative image for each group. Red arrows indicate the pyroptotic cells with large ballooning bubbles. The proportion of pyroptotic cells was calculated (right panel). Scale bar, 20 μm. c, d Proteasome inhibition stimulated LDH release. e, f GSDME silencing attenuated the proteasome inhibition-induced increase of pyroptotic cells. g, h GSDME knockdown abrogated proteasome inhibition-induced LDH release. For e–h, SNU449 cells were transfected with NC or siRNA targeting the indicated gasdermins (siGSDMs) for 24 h, then treated with 15 nM BTZ for another 48 h (e, g), or cells were co-transfected with siGSDMs and siPSMC5-1/2 for 72 h (f, h) before phase-contrast imaging (e, f) or LDH release assay (g, h). i, j Proteasome inhibition induced translocation of GSDME to the plasma membrane of multi-polar mitotic cell. White arrows indicate the clusterization of GSDME on cell membrane. Scale bar, 2.5 μm. k, l Proteasome inhibition induced the cleavage of caspase-3 and GSDME. SNU449 cells were treated with 15 nM BTZ for 48 h (a, c, i, k), or transfected with the indicated RNA duplexes for 72 h (b, d, j, l) before phase-contrast imaging (a, b), LDH detection (c, d), immunofluorescent staining for GSDME (Red), α-tubulin (TUBA, green) and chromosomes (DAPI, blue) (i, j), or immunoblotting (k, l). #, unspecific band. m Silencing caspase-3 but not caspase-1 blocked the BTZ-induced GSDME cleavage. n Silencing cGAS but not CHOP or IκBα attenuated the BTZ-induced cleavage of caspase-3 and GSDME. For m, n, SNU449 cells were transfected with NC or the indicated siRNA for 24 h, then treated with 15 nM BTZ for another 48 h before immunoblotting. o Ectopic expression of BCL-xL attenuated the BTZ-induced cleavage of caspase-3 and GSDME. SNU449-BCL-xL and its control line SNU449-Ctrl were treated with vehicle or 15 nM BTZ for 48 h before immunoblotting. Red arrows indicate the target band. p BTZ-induced cleavage of GSDME was enhanced by nocodazole but was inhibited by CDK1 inhibitor RO-3306. SNU449 cells were pretreated with vehicle, 50 ng/mL nocodazole or 10 μM RO-3306 for 6 h, followed by treatment with vehicle or 15 nM BTZ for another 48 h before immunoblotting. Error bars: SEM from at least three independent experiments. Student’s t test (a and c) and one-way ANOVA (b and d–h) were used. **P < 0.01; ***P < 0.001; ****P < 0.0001; ns not significant

We next explored which member of gasdermin family mediated the proteasome inhibition-induced mitotic pyroptosis. Examination on the levels of five gasdermin members revealed high level of GSDME in all three cell lines used in this study, and detectable levels of GSDMB and GSDMD in two of them (Supplementary Fig. 5a). Silencing of GSDME, but not GSDMB and GSDMD (Supplementary Fig. 5b), attenuated the roles of BTZ and siPSMC5 in increasing the fraction of cells with pyroptosis features (Fig. 2e, f) and in promoting LDH release (Fig. 2g, h; Supplementary Fig. 5c), suggesting GSDME may mediate BTZ/siPSMC5-induced pyroptosis. Consistently, accumulation of GSDME foci on plasma membrane was observed in the BTZ- or siPSMC5-induced multi-polar mitotic cells (Fig. 2i, j). And immunoblotting assays confirmed that both BTZ and siPSMC5 promoted the cleavage of GSDME (Fig. 2k, l; Supplementary Fig. 5d, e).

We further explored the mechanisms for BTZ-induced mitotic pyroptosis, especially how BTZ regulates GSDME cleavage. It has been reported that GSDME is cleaved by caspase-3,12 while GSDMD is cleaved by caspase-1.40 We found that BTZ- or siPSMC5-treatment increased the level of active caspase-3 (Fig. 2k, l; Supplementary Fig. 5d, e), and the depletion of caspase-3, but not caspase-1 (Supplementary Fig. 5f), blocked BTZ-induced GSDME cleavage (Fig. 2m), suggesting that caspase-3-GSDME cascade, but not caspase-1-GSDMD, mediates BTZ/siPSMC5-induced pyroptosis. It is shown that inhibition of proteasome can activate caspase-3 by upregulating IκBα or inducing ER-stress,25,41 and mitotic arrest can activate caspase-3 through cGAS signaling, that is, cGAS-activated STING induces IRF3 phosphorylation, leading to inhibition of BCL-xL, and in turn permeabilization of mitochondrial outer membrane and consequent caspase-3 activation.7 Therefore, we examined whether these pathways were involved in proteasome inhibitor-induced pyroptosis. The results showed that BTZ-induced caspase-3 activation and GSDME cleavage were attenuated by silencing cGAS (Supplementary Fig. 5g; Fig. 2n) or overexpression of BCL-xL (Fig. 2o). However, silencing of CHOP to inhibit ER stress signaling, or inhibition of IκBα to enhance NF-κB activity were unable to abrogate BTZ-induced caspase-3 and GSDME cleavage (Fig. 2n). Moreover, BTZ-induced caspase-3/GSDME activation was strengthened by nocodazole that triggered mitotic arrest, but it was inhibited by CDK1 inhibitor (RO-3306) that blocked mitotic entry (Fig. 2p). These results suggest that proteasome inhibitor-induced mitotic arrest may activate cGAS signaling, which inactivate BCL-xL, resulting in the activation of caspase-3-GSDME cascade and subsequent mitotic catastrophe in a form of mitotic pyroptosis.

Combined inhibition of proteasome and WEE family kinases displays synergistic effect in inducing mitotic pyroptosis and selectively killing cancer cells

We found that BTZ treatment arrested a large number of cells at G2-phase (Fig. 1a; Supplementary Fig. 1a, b). Therefore, we explored whether abrogation of G2/M checkpoint could facilitate the mitotic entry and subsequent mitotic catastrophe of BTZ-treated tumor cells. The levels of key regulators of G2/M checkpoint,42,43 including checkpoint kinase 1/2 (CHK1/2) and WEE family kinases (WEE1, PKMYT1) and their downstream effector (Y15-phosphorylated CDK1), were first analyzed. As shown, CHK1, WEE1, PKMYT1, but not CHK2, were upregulated upon BTZ treatment (Fig. 3a; Supplementary Fig. 6a). Consistently, WEE1/PKMYT1-induced phosphorylation at the Tyr15 site of CDK1 (pCDK1-Y15), which inactivated CDK1, was enhanced in BTZ-treated cells (Fig. 3a; Supplementary Fig. 6a). Next, cells were treated with BTZ first, then exposed to an inhibitor that repressed WEE1 alone (MK-1775) or inhibited both WEE1 and PKMYT1 (PD0166285), or exposed to an inhibitor that suppressed CHK1 alone (rabusertib) or inhibited both CHK1 and CHK2 (prexasertib) (Fig. 3b). The results showed that compared with BTZ monotreatment, sequential treatment with BTZ and MK-1775 or PD0166285 significantly reduced G2-phase population and increased M-phase cells, and PD0166285 showed a much stronger effect than MK-1775 (Fig. 3c; Supplementary Fig. 6b). Furthermore, overexpression of CDK1-T14A/Y15F, the dominant active mutant CDK1 that is resistant to WEE family kinase-induced inhibitory phosphorylation, increased the levels of pH3-S10 in BTZ-treated cells (Fig. 3d), which mimicked the effects of WEE family kinase inhibitors to promote mitotic entry. Notably, treatment with rabusertib or prexasertib could not alleviate the BTZ-induced G2-phase arrest (Fig. 3b, c; Supplementary Fig. 6b). Consistently, inhibition of ATR, the activator of CHK1/CHK2, by AZD6738 failed to relieve G2-phase arrest and promote mitotic entry upon BTZ treatment (Supplementary Fig. 6c, d).

Fig. 3

PD0166285 abrogates BTZ-induced G2-phase arrest and enhances BTZ-induced mitotic catastrophe. a BTZ increased the protein levels of CHK1, WEE1, PKMYT1 and Tyr15-phosphorylated CDK1. SNU449 cells were treated with BTZ at the indicated concentrations for 18 h before immunoblotting. b, c PD0166285 effectively alleviated BTZ-induced G2-phase arrest. Schematic diagrams of study design are shown in b. The triangles indicate the time points for the indicated treatment. SNU449 cells were pretreated with vehicle or 20 nM BTZ for 18 h, followed by treatment with vehicle or 0.5 μM of the indicated inhibitors for another 2 h before pH3-S10/PI staining and FACS (c). d Ectopic expression of the dominant active mutant CDK1-T14A/Y15F enhanced the BTZ-induced up-regulation of pH3-S10. SNU449-CDK1, SNU449-CDK1-T14A/Y15F and control line SNU449-Ctrl were treated with BTZ at the indicated concentrations for 30 h before immunoblotting. e Concurrent exposure to PD0166285 potentiated BTZ-induced accumulation of mitotic cells. SNU449 cells were treated with vehicle, 20 nM BTZ, 0.25 μM PD0166285, or BP-Combo (20 nM BTZ and 0.25 μM PD0166285) for 24 h before pH3-S10/PI staining and FACS. f–i PD0166285 amplified the effects of BTZ in inducing mitotic arrest and mitotic cell death. SNU449 subline that stably expressed histone H2B-EGFP and mCherry-α-tubulin were treated with 20 nM BTZ, 0.25 μM PD0166285, or BP-Combo, followed by live-cell imaging for a total of 2600 min. Representative images (f) and quantification of mitotic duration (g) are shown. White arrows indicate the large bubbles blowing from the plasma membrane (f). Scale bar, 5 μm. The cell death was determined by the emergence of pyroptosis characteristics or cell detachment, and the fractions of cell death at interphase or M-phase were quantified based on at least 146 cells in each group (h). In i, the fates of each cell within 2600 min are presented, each horizontal line represents one cell, and a fork in the line indicates cell division and cell fate of each daughter cell is also shown. The beginning time of BTZ treatment was set as 0. For a and d, red arrows indicate the target band. Error bars: SEM from at least three independent experiments. One-way ANOVA (c, e and g) was used. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns not significant

We then analyzed the expression level of representative proteasome subunits and WEE family kinases in patients of 17 cancer types from The Cancer Genome Atlas (TCGA).44 As shown, the mean expression level of a subset of proteasome subunit genes, including 20S core subunits (α1–7, β1–7) and 19S regulatory subunits (PSMC1–6), was significantly upregulated in various human malignancies (15/17, Supplementary Fig. 6e), compared with their normal counterparts. Interestingly, WEE1 was upregulated in only 3/17 cancer types, whereas PKMYT1 elevated in 16/17 cancer types (Supplementary Fig. 6e). These results suggest that the combined treatment of proteasome and WEE family inhibitors may be effective for a wide range of cancer types. And PD0166285, which can inhibit both PKMYT1 and WEE1, may be more powerful than MK-1775 in the combination treatment. Therefore, we chose the treatment of concurrent BTZ and PD0166285 exposure (designated as BP-Combo) for further study. As shown, the BP-Combo treatment had the same effect (Fig. 3e; Supplementary Fig. 6f) as sequential BTZ and PD0166285 treatment on the mitosis (Fig. 3c; Supplementary Fig. 6b). Consistently, combined PD0166285 treatment diminished BTZ-enhanced phosphorylation of CDK1-Y15 (Supplementary Fig. 6g). However, PD0166285 didn’t affect BTZ-induced accumulation of WEE1 (Supplementary Fig. 6h), indicating that PD0166285 may abrogate BTZ-induced G2-phase arrest via blocking the WEE1/PKMYT1-induced CDK1-T14/Y15 phosphorylation, without affecting proteasome activity.

The fates of cells treated with BP-Combo were further confirmed by live-cell imaging assay. Compared to control group, treatment with BTZ alone or BP-Combo, but not PD0166285 alone, induced significant M-phase arrest, multi-polar spindles and subsequent mitotic pyroptosis (Fig. 3f). The mean time of mitotic duration in the cells treated with vehicle, BTZ, PD0166285, or BP-Combo was 50, 377, 64, or 769 min, respectively (Fig. 3g). And longer mitotic duration was correlated with higher mitotic cell death (vehicle, BTZ, PD0166285, BP-Combo: 0%, 7.8%, 0%, 47.3%) (Fig. 3h, i). Notably, mitotic arrest and spindle multi-polarization appeared no earlier than three and four hours, respectively, after addition of BTZ (Fig. 3i, data not shown). The significant interphase extension upon BTZ treatment was diminished in BP-Combo group (Fig. 3i), which was reminiscent of the above findings that BP-Combo could overcome the BTZ-induced G2-phase arrest (Fig. 3c, e; Supplementary Fig. 6b, f). These data suggest that PD0166285 may augment the BTZ-induced mitotic catastrophe by releasing the cells from G2-phase arrest and promoting mitotic entry.

Next, we investigated whether PD0166285 enhanced the BTZ-induced mitotic pyroptosis. Compared with BTZ or PD0166285 monotreatment, BP-Combo displayed much stronger effects in increasing the fraction of cells with pyroptosis morphology (Fig. 4a; Supplementary Fig. 7a) or with Annexin V/PI double-staining (Fig. 4b; Supplementary Fig. 7b), and BP-Combo also had greater ability in enhancing LDH release (Fig. 4c; Supplementary Fig. 7c) and cytotoxic GSDME cleavage (Fig. 4d; Supplementary Fig. 7d). In contrast, very few apoptotic cells were observed in BTZ, PD0166285 and BP-Combo group (Supplementary Fig. 7e), indicating that pyroptosis but not apoptosis mainly contributed to the synergistic lethality effect. The synergistic effect was also verified using the combination of CFZ and PD0166285 (Supplementary Fig. 7f). Nevertheless, AZD6738, an ATR inhibitor, was unable to promote BTZ-induced pyroptosis (Supplementary Fig. 7g), in accord with its failure in promoting mitotic entry of BTZ-treated cells. These findings indicate that the combined treatment of BTZ with PD0166285 may have a synergistic effect in inducing mitotic catastrophe by promoting mitotic entry and pyroptosis.

Fig. 4

PD0166285 augments BTZ-induced pyroptosis. a PD0166285 enhanced the role of BTZ in increasing the proportion of cells with pyroptosis morphology. Red arrows indicate the pyroptosis cells with large bubbles. Five random fields in each well were captured and then subjected to analysis for the rate of cells with pyroptosis morphology. One of the five fields is shown as representative image for each group. Scale bar, 50 μm. b PD0166285 enhanced the role of BTZ in increasing the fraction of cells with Annexin V/PI double-staining. c PD0166285 promoted the effects of BTZ in promoting LDH release. d PD0166285 enhanced the role of BTZ in promoting GSDME cleavage. Red arrows indicate the target band. For a–d, SNU449 (left panel) and HeLa (right panel) cells were exposed to vehicle, 20 nM BTZ, 0.25 μM PD0166285, or BP-Combo for 30 h (SNU449) or 24 h (HeLa) before phase-contrast imaging (a), Annexin V/PI staining (b), LDH release assay (c) and immunoblotting (d). Error bars: SEM from at least three independent experiments. One-way ANOVA (a–c) was used. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns not significant

We then used Bliss combination index (Bliss CI, value below 1 indicates synergy) to evaluate the combinatorial effect of BTZ and PD0166285 in killing cancer cells at different concentrations. Significant synergistic lethality between BTZ and PD0166285 was observed in different solid tumor cells originated from various tissue types, including the liver (SNU449, Huh1, HepG2, SK-Hep-1 and Hepa1-6), cervix (HeLa) and bone (U2OS), and also in transformed human bronchial epithelial cell (HBERST), as evidenced by Bliss CI value below 1 (0.21–0.74, Supplementary Table 1). Importantly, much higher IC50 and no cooperative lethality between BTZ and PD0166285 were detected in immortalized cell lines (HBE, 293T, L02, LX2) and normal SF (Supplementary Table 1). The Loewe plots also revealed synergistic effects of BTZ and PD0166285 in killing cancer/transformed cells rather than immortalized/normal cells (Fig. 5 and Supplementary Fig. 8). Furthermore, the IC50 of BTZ and PD0166285 in the combined treatment was much lower than that in BTZ or PD0166285 monotreatment across tested cancer/transformed cell lines (Supplementary Table 1).

Fig. 5

BTZ and PD0166285 shows synergistic effect in killing cancer cells but not immortalized cells. a–g Various cancer cell lines were sensitive to BP-Combo treatment. h–l Immortalized cell lines and normal cells were resistant to BP-Combo treatment. Cancer cell lines from hepatoma (SNU449, Huh1, HepG2, Hepa1-6), cervical cancer (HeLa) and osteosarcoma (U2OS), transformed human bronchial epithelial cell line (HBERST), immortalized cell lines (HBE, 293T, L02, LX2) and normal cell (SF) were treated for 48 h with a combination of BTZ and PD0166285 at the indicated concentration. Cell survival was measured by Alamar Blue assay. Cooperativity screens (upper panels) and Loewe plots (down panels) for the synergistic effect of BTZ and PD0166285 are shown based on at least three independent experiments. In upper panels, color bars indicate the percentage of surviving cells in BP-Combo-treated group, which was normalized to untreated group. In down panels, color bars indicate synergy score in the Lowe plots; a score greater than 0 indicates synergism, and less than 0 indicates antagonism

Collectively, BP-Combo treatment may be effective for a wide range of solid tumor types and have a potential to specifically eliminate cancer cells, at least partly via mitotic pyroptosis.

Combined inhibition of proteasome and WEE family kinases significantly represses tumor growth and metastasis in vivo

Next, we explored the effect of BP-Combo on tumor development. The in vitro studies revealed that compared with BTZ or PD0166285 monotreatment, BP-Combo showed a much stronger inhibition on colony formation of tumor cells (Fig. 6a, b; Supplementary Fig. 9a, b). The in vivo studies were conducted by administering BTZ or/and PD0166285 at the early or late stage of subcutaneous xenograft development in immunodeficient mice. We found that early treatment with BP-Combo, from day one after cancer cell inoculation, dramatically inhibited the formation of HeLa xenografts, whereas BTZ or PD0166285 monotreatment only showed modest effect (Fig. 6c; Supplementary Fig. 9c). And late treatment with BP-Combo, beginning when tumor volumes reached appropriately 50 mm3, also exerted a more prominent inhibitory role on xenograft growth of both HeLa (Fig. 6d, e; Supplementary Fig. 9d) and Hepa1-6 (Fig. 6f, g; Supplementary Fig. 9e) cancer cells, compared with BTZ or PD0166285 monotreatment. We found that combined treatment with PD0166285 did not increase the concentration of BTZ in xenograft tissue (Supplementary Fig. 9f). Notably, compared with BTZ or PD0166285 group, xenografts from BP-Combo group showed a much higher level of mitotic marker pH3-S10 and cleaved GSDME (Fig. 6h), suggesting the enhanced mitotic arrest and pyroptosis. To verify the in vivo function of GSDME in BP-Combo treatment, Hepa1-6 cells which stably expressing shNC or shGsdme (Supplementary Fig. 9g) were injected subcutaneously into immunocompetent mice. As shown, the role of BP-Combo treatment in repressing xenograft growth was attenuated after GSDME in Hepa1-6 was knocked down (Fig. 6i, j and Supplementary Fig. 9h). These findings imply that the anti-tumor effect of BP-Combo depends on GSDME-mediated pyroptosis in cancer cells.

Fig. 6

BTZ and PD0166285 has a synergistic effect in suppressing growth of subcutaneous tumor xenografts. a, b BP-Combo showed much stronger effect than BTZ or PD0166285 monotreatment in repressing colony formation of tumor cells. SNU449 and HeLa cells were treated with vehicle, BTZ or PD0166285 alone, or with BP-Combo for 10 days before staining with 0.1% crystal violet. The representative images (a) and colony quantification (b) are shown. Scale bar, 2 mm. Error bars: SEM from at least three independent experiments. c–g BP-Combo showed much stronger effect than BTZ or PD0166285 monotreatment in inhibiting subcutaneous tumor xenograft development. The BALB/c nude mice were subcutaneously injected with HeLa cells, and then intraperitoneally injected with the indicated inhibitors one day after tumor cell implantation (early treatment, c), or when tumor volumes reached ~50 mm3 (late treatment, d, e). Early treatment: n = 5 (vehicle), 6 (BTZ), 6 (PD0166285) and 6 (BP-Combo). Late treatment: n = 3 (vehicle), 3 (BTZ), 3 (PD0166285) and 4 (BP-Combo). For f, g, C57BL/6J mice were subcutaneously injected with Hepa1-6 cells, and then intraperitoneally injected with the indicated inhibitors when tumor volumes reached ~50 mm3. n = 4 for each group. h BP-Combo had much stronger effect than BTZ or PD0166285 monotreatment in increasing pH3-S10 level and inducing GSDME cleavage in tumor xenografts. Mouse xenograft tissues from late treatment groups in Figures d and e were analyzed. i, j Silencing of GSDME diminished the anti-tumor effect of BP-Combo. C57BL/6J mice were subcutaneously injected with Hepa1-6-shNC or Hepa1-6-shGsdme cells, and then intraperitoneally administered with vehicle or BP-Combo treatment when tumor volumes reached ~50 mm3. n = 3 for each group. One-way ANOVA (b, c, e, g) and two-way ANOVA (d, f, i, j) were used. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns not significant

We also assessed the therapeutic efficacy of dual proteasome and WEE kinase inhibition on the development of autochthonous liver cancers in immunocompetent mice, which were hydrodynamically injected with plasmids expressing myr-AKT and NRasG12V or c-Myc and sgTP53. Drug treatment was performed at 26 days after hydrodynamic injection of plasmids. Compared with vehicle-treated group, BP-Combo significantly reduced the incidences of both myr-AKT/NRasG12V- (vehicle vs. BP-combo: 9/9 vs. 1/7) and c-Myc/sgTP53-induced liver tumors (vehicle vs. BP-combo: 6/7 vs. 4/10), and also decreased the number and size of tumor nodules (Fig. 7a–c), suggesting an inhibitory role of BP-Combo on liver cancer formation and growth. For c-Myc/sgTP53 injection model, in comparison with vehicle control group, BP-Combo decreased the incidence of pulmonary metastasis (vehicle vs. BP-combo: 5/7 vs. 3/10) as well as the number and size of metastatic foci (Fig. 7d, e), and greatly prolonged mouse survival (Fig. 7f). Both picric acid staining and histology analysis revealed no pulmonary metastasis (Supplementary Fig. 10a, b) and no mice died at the end of experiment in AKT/NRasG12V injection model.

Fig. 7

BP-Combo inhibits the growth and metastasis of mouse autochthonous liver tumors. a–c BP-Combo suppressed the growth of mouse autochthonous liver tumors. The day when C57BL/6J mice were hydrodynamically injected with the indicated plasmids was set as day 0 (upper panels). The tumor incidences and photographs of the livers (left panels), the number of macroscopic tumor nodules in the livers (middle panel) and the maximal diameter of macroscopic tumor nodules (right panels) are shown in a and b. The representative images of hematoxylin-eosin (H&E) staining and the numbers of microscopic tumor foci in the livers of c-myc/sgTP53-injected mice are shown in c. Scale bars, 1 cm (a, b) and 100 μm (c). d, e BP-Combo inhibited pulmonary metastasis of c-Myc/sgTP53-induced liver tumors. Photographs of the lungs (left panel) and the number of macroscopic metastatic nodules (right panel) are shown in d. H&E staining (left panel), the number (middle panel) and maximal diameter (right panel) of microscopic pulmonary metastatic foci are shown in e. Metastasis rates are indicated under the images (e). Scale bars, 1 mm (d) and 100 μm (e). Met. metastases. f BP-Combo improved the survival of mice with c-Myc/sgTP53-induced liver tumors. Student’s t test (a–e) and the log-rank test (f) was used. *P < 0.05; **P < 0.01

Collectively, BP-Combo may represent an effective regimen in repressing in vivo tumor development.

Combined inhibition of proteasome and WEE family kinases has no obvious side effects in the treated mice

We further validated the role of BP-Combo in liver orthotopic xenograft model and evaluated its potential toxicities in immunocompetent mice. Consistent with the above findings, BP-Combo significantly inhibited the growth (Fig. 8a, b) and metastasis (Fig. 8c, d) of Hepa1-6 xenografts.

Fig. 8

BP-Combo represses the growth and metastasis of liver orthotopic xenografts and has no obvious toxicity on normal tissues. a, b BP-Combo suppressed the growth of liver orthotopic xenografts. Hepa1-6 cells were inoculated under the capsule of the left hepatic lobe of C57BL/6 mice. Vehicle or BP-Combo treatment was intraperitoneally administered at the indicated time. The tumor incidences and photographs of dissected livers (a) and the tumor volume (b) are shown. c, d BP-Combo suppressed metastasis of liver xenografts. The representative images of H&E staining and metastasis rates (left panel) and the number of the intrahepatic (c) or pulmonary (d) metastatic foci (right panel) are shown. Scale bar, 50 μm (c) and 25 μm (d). Met. metastases. e, f The proportion of hematopoietic stem and progenitor cells were not affected in BP-Combo-treated mice. Bone marrow cells were isolated from the vehicle or BP-Combo-treated mice and analyzed by flow cytometry to detect CD117+Sca1+ population (e) and CD117+Lin– population (f). g BP-Combo did not change the lengths of ileum villi. H&E staining of the ileum villi are shown (left panel) and the villi lengths were calculated (right panel). Scale bar, 25 μm. BP-Combo did not affect the size and tissue structure of kidney (h) and heart (i). Scale bar, 5