Berberine Derivative B68 Promotes Tumor Immune Clearance by Dual-Targeting BMI1 for Senescence Induction and CSN5 for PD-L1 Degradation

Abstract

Promoting tumor cell senescence arrests the cell cycle of tumor cells and activates the immune system to eliminate these senescent cells, thereby suppressing tumor growth. Nevertheless, PD-L1 positive senescent tumor cells resist immune clearance and possess the ability to secret various cytokines and inflammatory factors that stimulate the growth of tumor cells. Consequently, drugs capable of both triggering senescence in tumor cells and concurrently diminishing the expression of PD-L1 to counteract immune evasion are urgently needed. Here, a berberine derivative B68 is developed, which specifically induces tumor cell senescence by targeting BMI1. B68 also involves the degradation of PD-L1 by targeting CSN5, thereby disrupting the immunosuppressive PD-1/PD-L1 interaction and enabling rapid clearance of senescent tumor cells. This approach simultaneously inhibits tumor progression and activates T cell immunity, as evidenced by the robust antitumor response following B68-induced immunization of senescent cancer cells. Moreover, the synergistic effect of B68 with anti-CTLA4 therapy further enhances antitumor immunity, and its ability to induce senescence in cancer cells triggers a strong protective response by dendritic and CD8⁺ T cells. These findings provide a scientific basis for developing a new tumor treatment strategy based on senescence induction and lay the foundation for further preclinical research.

Keywords: BMI1; CSN5; PD‐L1; colorectal cancer; tumor cell senescence.

Figure 1

B68 induces senescence in colon cancer cells. A) RKO and HCT116 cells were cultured with different concentrations of B68, B1, and bleomycin (BLEO) for one week, and cellular senescence was assessed by SA‐β‐gal activity staining, and quantification of SA‐β‐gal activity staining is shown in B). Scale bars, 200 µm (n = 3, error bars represent SEM, mean ± SEM, one‐way ANOVA). C) Representative images of H3K9me3 staining of RKO and HCT116 cells after 7 days of treatment with B68. Scale bars, 100 µm. D) Quantitative results of H3K9me3 staining in RKO and HCT116 cells (n = 3, error bars represent SEM, mean ± SEM, one‐way ANOVA). E–H) The protein expression levels of p21 and p53 were detected by immunoblotting after the treatment of RKO and HCT116 cells with different concentrations of B68 for 7 days. The results of the quantification of the p21 and p53 protein levels are shown below (n = 3, error bars represent SEM, mean ± SEM, one‐way ANOVA). I) Quantitative RT‐PCR was conducted to assess the expression levels of mRNAs associated with the SASP following a 7‐day treatment of HCT116 cells with B68 at a concentration of 15 µm and bleomycin at 10 mg mL⁻¹ (n = 3, error bars represent SEM, mean ± SEM, one‐way ANOVA). n.s, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 2

B68 significantly inhibited the development of colorectal cancer induced by AOM/DSS. A) Azoxymethane (AOM‐12.5 mg kg⁻¹) was first injected into C57BL/6J (male) mice, followed by 2% or 2.5% glucose sodium sulfate (DSS) in the drinking water for 1 week and then regular drinking water for 2 weeks; this process was repeated three times as described in the “Animal experiment methods” section. B) Representative images of the colon and rectum after AOM/DSS treatment depicting the extent of the tumor load. C) The length of mouse colon in the control group, model group, AOM/DSS+B68 group and AOM/DSS+B1 group (n = 5, error bars represent SEM, mean ± SEM, Student's t‐test). D) Changes in the body weights of mice in the control, model, AOM/DSS+B68 and AOM/DSS+B1 groups during the experimental period (n = 5, error bars represent SEM, mean ± SEM, Student's t‐test). E) The number and diameter of tumors in the control group, model group, AOM/DSS+B68 group, and AOM/DSS+B1 group were measured at the end of the treatment (n = 5, error bars represent SEM, mean ± SEM, Student's t‐test). F) Representative micrographs of H&E staining of colon tumors in control, model, AOM/DSS+B68 and AOM/DSS+B1 mice after AOM/DSS treatment (scale bar, 100 µm or 500 µm). G) Representative images of IHC staining of cleaved caspase 3, Ki‐67, p16, p21, p53, and γ‐H2AX in the colon of model, AOM/DSS+B68 and AOM/DSS+B1 mice; scale bar, 100 µm. H) Quantitative statistical results of the immunohistochemical staining shown in Figure 2G (error bars represent SEM, mean ± SEM, Student's t‐test). n.s, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 3

B68 exerts antitumor effects by targeting BMI1‐induced senescence in colorectal cancer cells. A) KEGG analysis of differentially expressed mRNAs in B68‐treated RKO cells. B) The volcano plot illustrates the changes in mRNA levels after B68 treatment in RKO cells (red dots represent significant upregulation, and blue dots represent significant downregulation), with marked genes related to senescence and cell cycle (fold change ≥2, p value ≤0.05). C) Gene Ontology (GO) analysis of differentially expressed mRNAs in RKO cells after B68 treatment. D,E) KEGG enrichment of the potentially differentially expressed proteins associated with senescence and cell cycle in B68‐treated RKO cells. F) Heatmap showing the expression of the top 30 differentially expressed proteins in B68‐treated RKO cells, with senescence‐ and cell cycle‐related genes labeled (fold change ≥1.2, unique peptides ≥2). G) DARTS analysis of potential target proteins that bind to B68. H) Pull‐down analysis of potential target proteins that bind to B68. I) Venn analysis results of differentially expressed proteins that appear in the DARTS, pull‐down and proteomics results. Detection of SA‐β‐gal activity J), p21 and p53 protein expression K) following BMI1 gene knockdown in RKO and HCT116 cells. L) H3K9me3 expression in RKO and HCT116 cells after BMI1 gene knockdown was detected by immunofluorescence, and the results of the quantitative analysis are shown on the right (M). Scale bars, 100 µm (n = 3, error bars represent SEM, mean ± SEM, one‐way ANOVA). ****p < 0.0001.

Figure 4

B68 induces senescence in colorectal cancer cells by targeting BMI1. A) After BMI1 knockdown, B68 did not further increase SA‐β‐Gal activity in HCT116 cells. B) Quantitative results of SA‐β‐Gal staining are shown on the right (n = 3, error bars represent SEM, mean ± SEM, one‐way ANOVA). C) CETSA was used to determine the thermal stabilization of the BMI1 interaction with B68 at a series of temperatures from 62 to 83 °C, and the quantification of BMI1 is shown in (D) (n = 3, error bars represent SEM, mean ± SEM, Student's t‐test). E) Stability of different concentrations of B68 on BMI1 at 65 °C. F) Quantitative results of BMI1 intensity in (E) (n = 3, error bars represent SEM, mean ± SEM, Student's t‐test). G) Effect of B68 on the stability of the BMI1 protein at different protein hydrolase concentrations. H) Quantitative results of BMI1 intensity in (F) (n = 3, error bars represent SEM, mean ± SEM, Student's t‐test). I) Effect of different concentrations of B68 on BMI1 protein stability at a pronase‐to‐protein ratio of 1:300. The quantitative results for the BMI1 protein are shown on the right (J) (n = 3, error bars represent SEM, mean ± SEM, Student's t‐test). (K) Molecular docking of B68 with BMI1. L) Cellular MST assay of GFP‐tagged BMI1 upon overexpression of wild‐type BMI1 and disrupted mutants (Arg123 or Glu120). n.s, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 5

Immunization with senescent tumor cells promotes antitumor effects and immune surveillance. A) RNA‐seq analysis of the mRNA expression of genes related to antigen presentation mechanisms and immunity in MC38 cells in the control and B68 groups. B) Flow cytometry detection of H‐2K^b/D^b expression in B68‐treated MC38 and CT26 cells. C) After senescence was induced in MC38 cells with B68, the mRNA expression of the p16, p21, p53, B2 m, Tap2, and Nlrc5 genes was detected by qRT‐PCR (n = 3 independent experiments, error bars represent SEM, mean ± SEM, one‐way ANOVA). D) C57BL/6 mice were subcutaneously injected with either vector (PBS) or senescent MC38 cells. On day 7, the skin at the site of tumor inoculation was subjected to H&E staining and immunohistochemistry, with a scale = 100 µm. The quantitative results of immunohistochemistry are shown on the right E) (n = 3, error bars represent SEM, mean ± SEM, one‐way ANOVA). F) Flow cytometry analysis of CD80, CD86, and MHC‐II (I‐A/I‐E) expression in DCs cocultured with senescent MC38 cells and quantification of CD80, CD86, and MHC‐II (I‐A/I‐E) expression are shown on the right G) (n = 3, error bars represent SEM, mean ± SEM, one‐way ANOVA). H) C57BL/6J mice were subcutaneously injected with PBS or senescent MC38 cells on the left side, and on day 7, MC38 cells were injected on the right side, after which the growth of tumors on the right side of the mice was observed. I,J) Individual tumor growth curves of PBS‐treated mice or mice immunized with senMC38 cells. Tumor growth (number of animals that developed tumors out of the total) and tumor latency (mean ± SEM of the day on which the tumor appeared) are indicated for each group. One‐way ANOVA was performed, and the error bars represent SEM, mean ± SEM). n.s, not significant; **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 6

B68 targets CSN5 to decrease the expression of PD‐L1 on tumor cells. A–H) Immunoblotting for PD‐L1 expression in RKO and HCT116 cells treated with B68 in combination with the proteasome inhibitor MG132 (1 µm, 10 µm), chloroquine (20 µm, 40 µm), the lysosomal inhibitor bafilomycin A1 (200 nm, 800 nm) or the autophagy inhibitor 3‐methyladenine (1 mm, 5 mm). The quantitative results of PD‐L1 expression are shown in the figure below (n = 3, error bars represent SEM, mean ± SEM, one‐way ANOVA). I) Immunoprecipitation (IP) analysis of PD‐L1 ubiquitination in RKO cells treated with B68 and IB analysis with a ubiquitin antibody. The cells were treated with MG132 before ubiquitination analysis was performed. J) Molecular docking of B68 with CSN5. K) The binding of CSN5 to B68 was determined by MST after the overexpression of GFP‐tagged wild‐type CSN5 or disruptive mutants (Ser481 and Glu76). n.s, not significant; *p < 0.05, **p < 0.01, ****p < 0.0001.

Figure 7

B68 is capable of inducing senescence in colorectal cancer cells and exerting antitumor activity by diminishing PD‐L1 expression on tumor cells. A) MC38 cells were injected into C57BL/6 mice on day ‐5, and B68 (8 mg kg⁻¹) and B1 (8 mg kg⁻¹) were administered as indicated. B) Ex vivo observation of the tumors. C) Tumor weight, D) tumor volume, and E) weight of the treated mice. F) After the experiment, the tumors were removed, and frozen sections were prepared for SA‐β‐Gal staining, scale bar, 100 µm. G–J) Representative IHC staining results for PD‐L1, CD8, cleaved caspase‐3, Foxp3, Ki‐67, p16, p21, p53, and γ‐H2AX in tumor tissues and quantitative results of immunohistochemical analysis are shown on the right. Scale bar, 100 µm (error bars represent SEM, mean ± SEM, Student's t‐test). K) Western blot analysis of p21, p53 and PD‐L1 expression in the tumor tissues of C57BL/6J mice. L) The results of the quantitative analysis are shown in the right (error bars represent SEM, mean ± SEM, one‐way ANOVA). n.s, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 8

B68 in combination with CTLA4 antibody effectively inhibited tumor growth in vivo in immunoreactive mice. A) C57BL/6 mice harboring MC38 cell‐derived tumors were subjected to various treatments, including PBS, anti‐PD‐L1, anti‐CTLA4, B68 alone, and a combination of B68 with anti‐CTLA4. On day 5, MC38 cells were implanted into C57BL/6 mice, followed by the administration of B68, anti‐PD‐L1, or anti‐CTLA4 at the indicated time points. Subsequently, the tumor morphology B), tumor mass C), body weight D), and tumor volume E) were monitored and measured over a 14‐day period. F) Xenograft tumor growth curves of C57BL/6J mice in each treatment group (n = 5, error bars represent SEM, mean ± SEM, Student's t‐test). Flow cytometric analysis was utilized to quantify the presence of GzmB⁺ cells F), Gr‐1⁺CD11b⁺ myeloid cells G), and Foxp3⁺CD25⁺ regulatory T cells H) among the groups treated with PBS, anti‐PD‐L1, anti‐CTLA4, B68 alone, B68 combined with anti‐PD‐L1, or B68 combined with anti‐CTLA4, and the quantitative results of the flow cytometry analysis are shown on the right (n = 5, error bars represent SEM, mean ± SEM, Student's t‐test). M) Representative immunohistochemical images of p21 in the tumor tissue of mice in each group, and the quantitative results of immunohistochemistry are shown on the right. Scale bar, 100 µm (n = 3, error bars represent SEM, mean ± SEM, Student's t‐test). n.s, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 9

Clinical relevance of BMI1, PD‐L1, or CSN5 in colorectal cancer tissues. A) Representative fluorescence images of BMI1, CSN5, or PD‐L1 in paracancerous tissue (Normal) and tumor tissue (Tumor) of CRC patients. B) Quantification of BMI1, CSN5, or PD‐L1 expression in (A) (n = 3, error bars represent SEM, mean ± SEM, Student's t‐test). C) The survival of CRC patients stratified by the expression of CSN5 or PD‐L1 was compared by two‐sided log‐rank analysis. D,E) Scatter plot showing the correlation between PD‐L1 or CSN5 expression and infiltrating CD8⁺ T cells in COAD and READ patients detected by TIMER 2.0. F,G) Correlations between CSN5 and the steps of the cancer immune cycle in COAD and READ patients. **p < 0.01, ***p < 0.001, ****p < 0.0001.