A BCL-xL/BCL-2 PROTAC effectively clears senescent cells in the liver and reduces MASH-driven hepatocellular carcinoma in mice

Abstract

Accumulation of senescent cells (SnCs) plays a causative role in many age-related diseases and has also been implicated in the pathogenesis and progression of metabolic dysfunction-associated steatotic liver disease (MASLD). Senolytics that can selectively kill SnCs have the potential to be developed as therapeutics for these diseases. Here we report the finding that 753b, a dual BCL-xL/BCL-2 proteolysis-targeting chimera (PROTAC), acts as a potent and liver-tropic senolytic. We found that treatment with 753b selectively reduced SnCs in the liver in aged mice and STAM mice in part due to its sequestration in the liver. Moreover, 753b treatment could effectively reduce the progression of MASLD and the development of hepatocellular carcinoma (HCC) in STAM mice even after the mice developed substantial metabolic dysfunction-associated steatohepatitis (MASH) and hepatic fibrosis. These findings suggest that BCL-xL/BCL-2 PROTACs have the potential to be developed as therapeutics for MASLD to reduce MASH-driven HCC.

Extended Data Fig. 1.. Evaluation of 753b-induced degradation of the BCL-2 family proteins in WI-38 cells and 753b senolytic activity against renal epithelial cells (RECs), human umbilical vein endothelial cells (HUVECs), and preadipocytes (PACs) in vitro.

A. Representative western blotting images of the levels of BCL-xL, BCL-2, BCL-w, MCL-1, and von Hippel-Lindau (VHL) in NC WI-38 cells after they were treated with increasing concentrations of 753b in a cell culture for 24 h. B. Densitometric analyses of BCL-xL, BCL-2, BCL-w, and MCL-1 expression in NC WI-38 cells from A are presented. DC₅₀, drug concentration causing 50% degradation of protein of interest; D_max, the maximum level of degradation of protein of interest. C. The levels of VHL, BCL-xL, BCL-2, BCL-w, and MCL-1 in NC and IR-SnC WI-38 cells and human platelets (PLTs) from three donors (P1-3) were detected by western blotting. Similar results from NC and IR-SnC WI-38 cells were observed in a separate assay. D-E. Cell viability analyses show that 753b is more potent than ABT263 against IR-SnC and REP-SnC REC (D) and HUVEC (E) but less toxic to their non-senescent counterparts. The viability of NC, IR-SnC and REP-SnC REC and HUVEC was determined 72 h after treatment with increasing concentrations of ABT263 and 753b. EC₅₀, half-maximal effective concentration. The data presented are mean ± SD (n = 6 technical replicates) of a representative assay. EC₅₀, half-maximal effective concentration. F & G. Cell viability analyses show that 753b is not senolytic, but dasatinib and quercetin (D+Q) are, against IR-SnC PAC. The viability of IR-SnC PAC was determined 72 h after treatment with increasing concentrations of ABT263 and 753b (F), or with vehicle (VEH), low D+Q (1 μM D plus 20 μM Q) and high D+Q (10 μM D plus 200 μM Q) (G). The data presented are mean ± SD (n = 3 technical replicates) of a representative assay. β-actin was used as a loading control in A and C.

Extended Data Fig. 2.. 753b has no effect on the levels of Cdkn2a expression in the lung, kidney and fat tissues but reduces hepatic expression of SASP factors in naturally aged mice.

A. The levels of Cdkn2a mRNA in the lung, kidney and inguinal fat from untreated young mice and naturally aged mice treated with VEH and 753b. B. The levels of Cxcl12, Ccl5, Ccl2, Cxcl10, Mmp3, Mmp13, Il6, Il1a, Tnfa, and Tnfsf11 in the liver tissue from untreated young mice and naturally aged mice treated with VEH and 753b. The data are presented as means ± SEM (n = 7, 6, and 7 mice per group for young mice, VEH- and 753b-treated aged mice, respectively) and were analyzed by one-way ANOVAs with Šídák's multiple comparisons test or Tukey's multiple comparisons test.

Extended Data Fig. 3.. 753b reduces splenic expression of SASP factors in naturally aged mice.

The levels of Il1b, Serpine1, Mmp3, Mmp13, Cxcl12, Ccl5, Ccl2, Cxcl10, Il6, Il1a, Tnfa, and Tnfsf11 in the spleens from untreated young mice and naturally aged mice treated with VEH and 753b. The data are presented as means ± SEM (n = 8, 6, and 7 mice per group for young mice, VEH- and 753b-treated aged mice, respectively) and were analyzed by one-way ANOVAs with Šídák's multiple comparisons test or Tukey's multiple comparisons test.

Extended Data Fig. 4.. Characterization of SnCs in the liver from STAM mice and additional evaluations of 753b treatment on STAM mice.

A. A cartoon indicates distribution of zone 1, 2, and 3 hepatocytes in liver lobules along with blood flow across the periportal to pericentral axis. Periportal hepatocytes are in zone 1 that consists of portal veins, hepatic arteries, and bile ducts. B-C. SA-β-gal staining was combined with immunohistochemistry to characterize types of SnCs in the livers from STAM mice 8 weeks after STZ and 4 weeks after HFD. Antibodies against the pericentral hepatocyte marker Cyp2E1 (B), periportal hepatocyte marker GP6Cα (C), hepatocyte marker HNF4α (D), and biliary epithelial cell marker CK19 (E) were used for the stainings. Representative images of the stainings are presented on the left (scale bar = 100 μm) and higher magnification images of the marked area on the left images are presented on the right for C-E. Data presented in A-E are from one representative experiment and three independent experiments were performed with similar results. F. The levels of Cdkna1 mRNA in the tumor free liver tissues from VEH-treated and 753b-treated STAM mice on P150. The data are presented as means ± SEM (n = 5 mice/group) and were analyzed by a two-tailed, unpaired Student’s t-test. G. The levels of selected SASP mRNA in the tumor free liver tissues from STAM mice on P150. The data are presented as means ± SEM (n = 2 and 3 mice for VEH and 753b group, respectively) and were analyzed by a two-tailed, unpaired t-tests. H. Photo of reprentative VEH-treated and 753b-treated STAM mice on P150. I. Whole body weight of STAM mice on P150. Data are presented as means ± SEM (n = 5 mice per group) and were analyzed by a two-tailed, unpaired Student’s t-test. J. Blood levels of glucose in VEH- and 753b-treated STAM mice after IP injection of insulin one week before the termination of the experiment on P150. Data are presented as means ± SEM (n = 5 mice per group) and analyzed by two-way ANOVA.

Extended Data Fig. 5.

Diagram illustrating the time-dependent progression of NAFLD and development of HCC in STAM mice and different 753b treatment schedules and their effects on HCC development and progression.

Extended Data Fig. 6.. The effects of early and delayed treatments with 753b on the selective markers of hepatic inflammation, necroptosis, and macrophase activation in the livers from STAM mice.

A. The levels of Ccl2, Ccl5, Mlkl, Ripk3, and Itgax/Cd11c mRNA in the tumor free liver tissues from STAM mice on P150 after receiving earlier VEH or 753b treatment as shown in Fig. 6A. The data are presented in A as means ± SEM (n = 5 mice/group) and were analyzed by a two-tailed, unpaired Student’s t-tests. B. Western blotting image of αSMA and Type 1 procollagen in the tumor free liver tissues (left panel), and that of Gpc3 expression in the whole liver tissues (right panel), from STAM mice on P150 after receiving delayed VEH or 753b treatment as shown in Fig. 7A. C. The levels of Ccl2, Ccl5, Il6, Serpine1, Mmp3, Mmp13, Mlkl, Ripk3, and Itgax/Cd11c mRNA in the tumor free liver tissues from STAM mice on P150 after receiving delayed VEH or 753b treatment as shown in Fig. 7A. The data are presented in C as means ± SEM (n = 5 mice/group) and were analyzed by a two-tailed, unpaired Student’s t-tests.

Extended Data Fig. 7.. 753b is not cytotoxic to HCC cells in vitro.

A. Cell viability of human HCC cells, HepG2 and Huh7, 72 h after treatment with increasing concentrations of ABT263 and 753b in cell culture. EC₅₀, half-maximal effective concentration. The data presented are mean ± SD (n = 6 technical replicates) of a reprentative assay. Similar results were observed in two additional assays. B. Representative western blotting images of BCL-xL, BCL-2, BCL-w and MCL-1 in HepG2 cells after they were treated with increasing concentrations of ABT263 and 753b for 16 h. β-actin was used as a loading control.

Fig. 1.. 753b is a potent and broad-spectrum senolytic agent in vitro.

A-B. 753b degraded BCL-xL in NC and IR-SnC WI-38 cells in a dose-dependent manner. C. Densitometric analyses of BCL-xL expression in NC and IR-SnC WI-38 cells. Data are presented as mean from two independent assays. DC₅₀, drug concentration causing 50% degradation of protein of interest; D_max, the maximum level of degradation of protein of interest. D-E. 753b degraded BCL-xL in NC and BCL-xL and BCL-2 in IR-SnC WI-38 cells in a time-dependent manner. Data presented is from one experiment. Similar results were observed in a repeat. F. Treatment with MG132 (a 26S proteasome inhibitor), VHL ligand (VHL-L) or ABT263 2h prior to 753b treatment abrogated the degradation of BCL-xL in NC (left) and BCL-xL and BCL-2 in IR-SnC (right) WI-38 cells. Data presented is from one experiment. Similar results were observed in a repeat. G. 753b is more potent to kill IR-SnC and REP-SnC WI-38 cells but less toxic to NC WI-38 cells than ABT263. EC₅₀, half-maximal effective concentration. The data presented in the left and middle panels are mean ± SD (left panle: n = 6, 6, 3 technical replicates for NC, IR-SnC, REP-SnC, respectively; middle panle: n = 3, 6, 4 technical replicates for NC, IR-SnC, REP-SnC, respectively) of a representative assay out of three independent assays and that on the right panel are mean ± SEM of multiple independent assays (n = 28 and 30 biological replicates for ABT263 and 753b, respectively). A two-tailed, unpaired Student’s t-test was used. H. ABT263 and 753b induces apoptosis in IR-SnC WI-38 cells but not in NC WI-38 cells evidenced by immunoblot analysis of caspase 3 activation/cleavage. Data presented is from one experiment.

Fig. 2.. 753b effectively reduces senescence burden in naturally aged mice in a tissue-specific manner.

A. The scheme of the experimental design. B. Blood platelet (PLT) counts in 753b-treated and vehicle (VEH)-treated aged mice 24 h after the first (left) and last (right) treatment and in control young mice. C. Whole body weight and the liver weight as a percentage of the whole body weight in young and VEH- and 753b-treated aged mice were measured at the end point of study. D. The levels of Cdkn2a mRNA in the liver and spleen from young and VEH- and 753b-treated aged mice. E. The levels of Il1b and Serpine1 mRNA in the liver from young and VEH- and 753b-treated aged mice. F. Representative SA-β-gal staining (blue color) images of liver tissue sections from young and VEH- and 753b-treated aged mice (scale bar = 300 μm) (top panels) and a higher magnification of the selected areas from the top panels (lower panels). G. Percentage of SA-β-gal staining areas in the liver tissue sections from young and VEH- and 753b-treated aged mice. H. The mean concentrations of 753b in the liver, spleen, lung, kidney, and perigonadal fat tissues of mice (n = 2) were monitored over 168 h after a single IP administration of 5 mg/kg 753b. The data presented in Fig. 2B are means ± SEM and were analyzed by Krustal-Wallis one-way ANOVA with Dunn’s post-hoc test; the data presented in Fig. 2C, D, E, and G are means ± SEM and were analyzed by one-way ANOVA with Šídák's multiple comparisons test or Tukey's multiple comparisons test. For Fig. 2B-C, n = 5, 6, and 7 mice/group for young, VEH, and 753b groups, respectively, except n = 3 for young on the right panel of Fig. 2B; for Fig. 2D, n = 8, 6, and 7 mice/group for young, VEH, and 753b groups, respectively; for Fig. 2E, n = 7, 6, and 7 mice/group for young, VEH, and 753b groups, respectively; and for Fig. 2G, n = 5 mice/group for young, VEH, and 753b groups, respectively.

Fig. 3.. 753b is a more potent senolytic than ABT263 against senescent murine hepatocytes.

A. Representative images of phase contrast, EdU stainig, nuclear staining with Hoechst 33342 (Hoechst), and the overlays of these two stainings of non-senescent (NC) and irradiation-induced senescent (IR-SnC) AML-12 cells (scale bar in the overlay = 300 μm) and images of a higher magnification of the selected areas from the overlay images are shown on the left and percentages of EdU⁺/ Hoechst⁺ cells are presented on the right. B. Representative images of SA-β-gal staining of NC and IR-SnC AML-12 cells (scale bar = 300 μm) and images of a higher magnification of the selected areas from the SA-β-gal staining images are shown on the left and percentages of SA-β-gal⁺ cells are presented on the right. C. The levels of Cdkn2a, Cdkn1a, Ccl2, Tnfa, Il1b, Serpine1, Cxcl12, Mmp3, Tgfb1 and Il6 mRNA in NC and IR-SnC AML-12 cells. D. 753b is more potent against IR-SnC AML-12 cells than ABT263. EC₅₀, half-maximal effective concentration. The data presented are mean ± SD (n = 6 technical replicates) of a representative assay. Similar results were observed in two more independent assays. E. Representative immunoblot images show that treatment with 753b (0.3 μM) but not ABT263 (0.3 μM) can degrade BCL-xL in NC AML-12 cells and BCL-xL and BCL-w in IR-SnC AML-12 cells. The data presented in all the bar graphs are means ± SEM (n = 3 independent assays) and were analyzed by a two-tailed, unpaired Student’s t-test.

Fig. 4.. 753b reduces hepatic cellular senescence, steatosis, fibrosis, and injury in STAM mice.

A. Experimental design. STZ, streptozocin; HFD, high fat diet; ITT, insulin tolerance test; MRI, Magnetic resonance imaging; ALT, alanine aminotransferase; AST, aspartate aminotransferase; VEH, vehicle-treated STAM mice; and 753b, 753b-treated STAM mice. B. Representative images of SA-β-gal staining (dark blue color) (left panel) and percentage of SA-β-gal staining areas in the liver tissues from mice on P90. Scale bar = 150 μm. C. The levels of Cdkn2a mRNA in the tumor-free liver tissues from mice on P90 and P150. D. Western blot images (left) and densitometric quantification (right) of the BCL-xL levels in the livers from mice on P150. E. Representative oil red O staining images (left panels) and quantification (right panel) of the percentage of areas with the staining in the images of the livers from STAM mice on P90. Scale bar = 50 μm. F. The levels of triglyceride in the livers on P90. G. The relative expression of selective fibrosis-related genes in the tumor-free liver tissues from STAM mice on P150. H. Representative sirius red staining (left panels) and the quantification of the staining in the tumor-free liver tissues from STAM mice on P150. Scale bars = 200 μm. I. Western blotting images (left panel) of αSMA and type I pro-collagen in the tumor-free liver tissues from STAM mice on P150. The levels of αSMA (middle panel) and type I pro-collagen (right panel) relative to the levels of α-tubulin were quantified by densitometry. J. The levels of hydroxyproline in the tumor-free liver tissues from STAM mice on P150. K. The serum levels of ALT and AST from STAM mice on P150. Data presented in all the bar graphes are means ± SEM (n = 5 mice per group) except in the left panel of Fig. 4C (n = 3 mice for VEH group and n = mice 4 for 753b group) and D (n =3 mice per group) and were analyzed by a two-tailed, unpaired Student’s t-test.

Fig. 5.. 753b administration reduces tumorigenesis in STAM mice.

A. Representative MRI T1 pre-contrast and T2-weighted images of VEH- and 753b-treated STAM mice. Tumor lesions (indicated by yellow arrows) were determined based on hypointense signals in T1 and hyperintense signals in T2-weighted images. B. Images of morphological overviews (upper panel) of the whole livers and representative microscopic images (lower panel) of the HE staining of the liver tissues from five VEH- and 753b-treated STAM mice. Tumors in the lower panel are circled by the white dotted lines. Scale bar = 150 μm. C. The number (upper panel) and volume (lower panel) of HCC lesions were quantified according to the MRI T1 and T2-weighted images. Data for each mouse were analyzed based on 5-6 slices of images that covered the whole liver lobes. D. The ratios between liver and body weight of VEH- and 753b-treated STAM mice. E. Gpc3 expression in the liver tissues from VEH- and 753b-treated STAM mice was measured by western blotting (left panel) and quantified by densitometry (right panel). Data presented in all the bar graphes are means ± SEM (n = 5 mice per group) and were analyzed by a two-tailed, unpaired Student’s t-test. MRI assays were done three days prior to the termination of the study on P150, and all other assays were done on P150.

Fig. 6.. Early treatment (P45-90) with 753b has no effect on hepatic cellular senescence, steatosis, fibrosis, and injury, and HCC development in STAM mice.

A. Diagram illustrating the experimental design for the early 753b treatment. B. Left panel: Representative images of SA-β-gal staining (green color) in the liver tissues from VEH- and 753b-treated STAM mice. Scale bar = 150 μm. Right panel: Percentage of SA-β-gal staining areas in the liver tissue sections from VEH- and 753b-treated STAM mice. C. The levels of Cdkn2a and Cdkn1a mRNA in the tumor-free liver tissues from VEH- and 753b-treated STAM mice. D. Representative oil red O staining images (left panels) and quantification (right panel) of the percentage of areas with oil red O staining in the images of the livers from VEH- and 753b-treated STAM mice. Scale bar = 150 μm. E. Representative sirius red staining (left panels) and the quantification of the staining in the tumor-free liver tissues from VEH- and 753b-treated STAM mice. Scale bars = 200 μm. F. The relative expression of selective fibrosis-related genes in the tumor-free liver tissues from VEH- and 753b-treated STAM mice. G. Left: Images of morphological overviews of the whole livers from VEH- and 753b-treated STAM mice. Tumors in the images are circled by the white dotted lines. Right: The number of HCC lesions as seen on the left images. H. Gpc3 expression in the liver tissues from VEH- and 753b-treated STAM mice was measured by western blotting and quantified by densitometry. I. The serum levels of ALT and AST from VEH- and 753b-treated STAM mice. Data presented in all the bar graphes are means ± SEM (n = 5 mice per group) except in H (n = 4 mice per group) and were analyzed by a two-tailed, unpaired Student’s t-test. All the assays were done on P150.

Fig. 7.. Delayed treatment (P90-150) with 753b reduces hepatic cellular senescence, steatosis, fibrosis, and injury, and inhibits HCC development in STAM mice.

A. Diagram illustrating the experimental design for the delayed 753b treatment. B. Left panel: Representative images of SA-β-gal staining (green color) in the liver tissues from VEH- and 753b-treated STAM mice. Scale bar = 150 μm. Right panel: Percentage of SA-β-gal staining areas in the liver tissue sections from VEH- and 753b-treated STAM mice. C. The levels of Cdkn2a and Cdkn1a mRNA in the tumor-free liver tissues from VEH- and 753b-treated STAM mice. D. Representative oil red O staining images (left panels) and quantification (right panel) of the percentage of areas with oil red O staining in the images of the livers from VEH- and 753b-treated STAM mice. Scale bar = 150 μm. E. Representative sirius red staining (left panels) and the quantification of the staining in the tumor-free liver tissues from VEH- and 753b-treated STAM mice. Scale bars = 200 μm. F. The relative expression of selective fibrosis-related genes in the tumor-free liver tissues from VEH- and 753b-treated STAM mice. G. The levels of αSMA (middle panel) and type I pro-collagen (right panel) relative to the levels of β-actin were measured by Western blotting and quantified by densitometry. H. Left: Images of morphological overviews of the whole livers from VEH- and 753b-treated STAM mice. Tumors in the images are circled by the white dotted lines. Right: The number of HCC lesions as seen on the left images. I. Gpc3 expression in the liver tissues from VEH- and 753b-treated STAM mice was measured by western blotting and quantified by densitometry. J. The serum levels of ALT and AST from VEH- and 753b-treated STAM mice. Data presented in all the bar graphes are means ± SEM (n = 5 mice per group except n = 4 for G and I) and were analyzed by a two-tailed, unpaired Student’s t-test. All the assays were done on P150.

Fig. 8.. 753b dose not inhibt HCC progression.

A. Diagram illustrating the experimental design for the further delayed or later 753b treatment in STAM mice. B. Left: Images of morphological overviews of the whole livers from VEH- and 753b-treated STAM mice. Tumors in the images are circled by the white dotted lines. Right: The number of HCC lesions as seen on the left images. C. Gpc3 expression in the liver tissues from VEH- and 753b-treated STAM mice was measured by western blotting and quantified by densitometry. The data presented in the bar graphs of B and C are means ± SEM (n = 4 mice per group) and were not statistically significant analyzed by a two-tailed, unpaired Student’s t-test. D. Diagram illustrating the experimental design for the HTVi-induced HCC with or without 753b treatment. HTVi, hydrodynamic tail vein injection (HTVi); AFP, alpha-fetoprotein. E. Plasma levels of AFP in VEH- and 753b-treated mice after HTVi. Data presented are means ± SEM (n = 7 mice per group). F. Images of morphological overviews of the whole livers from VEH- and 753b-treated mice.