The capacity of targeted anticancer agents to exert immunomodulatory effects provides a strong rationale to develop novel agents suitable for combinatorial regimens with immunotherapy to improve clinical outcomes. In this study, we developed a dual-targeting PI3K and HDAC inhibitor BEBT-908 that potently inhibits tumor cell growth and potentiates anti-PD1 therapy in mice by inducing immunogenic ferroptosis in cancer cells. Treatment with BEBT-908 promoted ferroptotic cell death of cancer cells by hyperacetylating p53 and facilitating the expression of ferroptotic signaling. Furthermore, BEBT-908 promoted a proinflammatory tumor microenvironment that activated host antitumor immune responses and potentiated immune checkpoint blockade therapy. Mechanistically, BEBT-908–induced ferroptosis led to upregulation of MHC class I and activation of endogenous IFNγ signaling in cancer cells via the STAT1 signaling pathway. The dual PI3K/HDAC inhibitor BEBT-908 is a promising targeted therapeutic agent against multiple cancer types that promotes immunogenic ferroptosis and enhances the efficacy of immunotherapy.

Significance:

The dual PI3K/HDAC inhibitor BEBT-908 elicits potent antitumor responses, effectively inducing immunogenic ferroptosis of tumor cells and potentiating cancer immunotherapy.

Immunotherapy has shown great promise in cancer therapy. However, only a small subset of patients with advanced cancers respond to single-agent immune checkpoint blockade (ICB). Many efforts are therefore underway to develop combination approaches that improve the effectiveness and patient response rates of immunotherapy (1). Targeted anticancer agents are developed on the basis of the strategy of selective disruption of molecular alterations in tumor cells essential for oncogenesis. Intriguingly, many targeted anticancer agents that are either at earlier stages of clinical evaluation or already approved for use in patients with cancer have shown the ability to enhance cancer immunotherapy. Examples include BCR-ABL kinase inhibitor (2), KRasG12C inhibitor (3), CDK4/6 inhibitors (4), and VEGFR inhibitors (5, 6). The capacity of targeted anticancer agents exerting immunomodulatory effects provides a strong rationale to develop novel agents suitable for combinatorial regimens with immunotherapy.

Activated PI3K signaling cascade (7) and dysregulated histone deacetylase (HDAC; ref. 8) are two critical signaling pathways impacting human malignancies and thus very important targets for drug development. However, clinical experience has indicated that targeting a single oncoprotein is often not sufficient to achieve desired therapeutic efficacy due to the development of resistance either by functional redundancy of alternative pathways or genetic mutations that enable drug resistance in surviving cancer cells (9, 10). Synergistic strategies of combining HDAC inhibitor (HDACi) with PI3K–AKT–mTOR pathway inhibitors have presented promising antineoplastic effects in both preclinical and clinical studies (11–13). For example, HDAC inhibition can destabilize AKT (14), increase oxidative stress (15), or upregulate of tumor suppressor gene FOXO1 in MYC-driven medulloblastoma (16) to modulate the sensitivity of tumors, and in some instances overcome resistance to PI3K–AKT–mTOR pathway inhibitors. Furthermore, pharmacologic or genetic inhibition of these functional elements not only limits mitogenic signaling in cancer cells but also mediate a panel of therapeutically cancer cell–dependent immunostimulatory effects, including improved antigen presentation on MHC molecules, immunogenic cell death, secretion of proinflammatory cytokines. On the basis of these evidence, we posit that simultaneous inhibition of both HDAC and PI3K pathways may be more potent at inhibiting tumor cell growth and better able to synergize with ICB therapy.

In this study, we developed a biofunctional PI3K and HDACi (BEBT-908). BEBT-908 can kill tumor cells potently as a monotherapeutic agent with favor pharmacokinetic characteristics. Furthermore, it induces ferroptotic cancer cell death, which leads to increased tumor cell surface MHC I expression and a proinflammatory tumor microenvironment with robust intratumoral infiltration of immune effector cells that enhanced ICB therapy significantly.

Cell lines

MC38 mouse colon adenocarcinoma cell line was obtained from Kerafast. All other human cancer cell lines were obtained from Cell Culture Facility of Chinese Academy of Sciences or Guangzhou Cellcook Biotech Co. Ltd. Cells were cultured in RMPI 1640 (Gibco) supplemented with 10% FBS, 100 units/mL penicillin, and 100 μg/mL streptomycin antibiotics at 37°C in a humidified 5% CO2 incubator. All human cell lines were authenticated by the short tandem repeat Multi-amplification Kit (PowerPlex 16HS System). All cell lines were periodically tested for Mycoplasma contamination by use of a universal Mycoplasma detection kit (ATCC).

Chemicals

BEBT-908 was designed and synthesized by Guangzhou BeBetter Medicine Technology Co. LTD. In brief, BEBT-908 was synthesized from starting materials compound 6, SM3, and SM4 (Supplementary Fig. S1). Compound 6 was reacted with starting material SM3 in the presence of acetonitrile and N,N-diisopropylethylamine to form compound 7. Compound 7 was then reacted with the boronic acid starting material SM4 under Suzuki coupling conditions to form compound 8, which was then reacted with hydroxylamine to form BEBT-908 free form. Final compound BEBT-908 was formed as a trihydrochloride salt by treatment of BEBT-908 free form with hydrogen chloride in methanol. The chemical structure of BEBT-908 was confirmed by nuclear magnetic resonance and mass spectrometry (MS). The purity of BEBT-908 is more than 98%, which was analyzed by high performance liquid chromatography methods. Starting materials SM3 is commercially available, compound SM4 was prepared through three-step synthesis starting from commercially available 2, 5-Dibromopyridine (SM2) and compound 6 was prepared through five-step synthesis starting from commercially available 3-aminothiophene-2-carboxylate (SM1). The trihydrochloride salt form of BEBT-908 was used in the study.

Reference HDACis LBH589 and vorinostat (SAHA), PI3K inhibitor GDC-0941, ferroptosis inducer (erastin, RSL3), ferroptosis inhibitor (ferrostatin-1) were obtained from Selleck. For combination treatments, SAHA and GDC-0941were added to the cells at a 1:1 molar ratio with the IC50 values were calculated as the concentrations of single agents.

Molecular cloning

CRSIPR constructs were cloned following previously published method (17). Single-guided RNA (sgRNA) sequences targeting human GPX4: sgRNA1-CGTGTGCATCGTCACCAACG, sgRNA2-ACTCAGCGTATCGGGCGTGC. HCT116 cells were infected with GPX4 sgRNA lentiviral vector, and selected with 1 μg/mL puromycin. The knockout efficiency was determined by Western blot using GPX4 antibody (1:2,000; 14432-1; Proteintech Group Inc.).

In vitro PI3K and HDAC enzyme activity assay

Activity of PI3Kα was measured using an ADP-Glo luminescent kinase assay. Enzyme activity was assayed using a complex of N-terminal GST-tagged recombinant full-length human p110α (GenBank Accession No. U79143) and untagged recombinant full-length human p85α (XM_043865) that were coexpressed in a Baculovirus infected Sf9 cell expression system. Similarly, purified PI3Kα E545K, PI3Kα H1047R, PI3Kβ(NM_006219), PI3Kδ(NM_005026), and PI3Kγ (AF327656) were made from Sf9 cell expression system to assess the potency and specificities of BEBT-908.

Total HDAC activities of class I and II HDACs were assessed using Biomol Color de Lys system. HDAC subtype specificity assays were performed at BPS Bioscience following the standard operating procedure.

Off-target kinase profiling study

High concentration (10 μmol/L) BEBT-908 was used to evaluate the inhibitory activity against a panel of kinases. Profiling was done following standard operating procedures of Eurofins Cerep (France). Each experiment was carried out in duplicate. Those with over 80% response were further evaluated.

Cell viability assay

Cancer cells were plated at 2,000 to 3,000 per well in 96-well flatted bottom white cell culture plates (Corning) with various concentrations of BEBT-908, SAHA, GDC-0941 or combinations thereof. Growth viability was determined by CellTiter-Glo luminescent viability assay (Promega) on 72 hours after drug treatment. Each treatment condition was represented by three individual experiments. Relative viability was calculated by normalizing raw luminescence counts to DMSO control (Ctrl.) treated cells. Half maximal inhibitory concentration (IC50) values were calculated using Graphpad/Prism 8 software and sigmoidal dose–response curve fitting.

Reactive oxygen species detection and malondialdehyde assay

Cells were treated with DMSO (Ctrl.), 8 nmol/L BEBT-908, and 572 nmol/L of SAHA+GDC-0941 for 24 hours. Reactive oxygen species (ROS) production was detected using the probes DCFDA (Abcam) by flow cytometer Cytoflex (Beckman).

Malondialdehyde assay (MDA) content in tumor cells or tissues was measured by Lipid Peroxidation MDA Assay Kit (Beyotime) to monitor lipid peroxidation.

qRT-PCR

The primers used for qRT-PCR in this study were listed on Supplementary Table S1.

Immunofluorescence staining and IHC assay

Tumor tissues were fixed, embedded into paraffin and sectioned following standard procedures. The sections were incubated with the anti-mouse CD45 antibody at 4°C overnight, sequentially stained with FITC-conjugated rat anti-mouse secondary antibody for 1 hour at room temperature, then mounted with mounting medium (Vector Laboratories). Images were captured by use of Leica SP-5 confocal microscope. For IHC assay, tumor sections were incubated with primary and horseradish peroxidase (HRP)-conjugated secondary antibody (primary antibodies used for IHC were listed in Supplementary Table S2), then visualized with HRP/DAB system. Images were captured with light microscope (Olympus, BX43). The percentage of positive staining area in each image was calculated by use of ImageJ (NIH) with a standard procedure (18). In brief, immunostaining biomarker was detected in digital slides, and immunostaining detection was treated as a color image analysis problem and built statistical color models using a large number of labeled positive and negative immunostaining pixels. By implementing the statistical models in different color spaces, the opponent chromaticity signals were shown and effectively characterized the color distributions of the immunostaining biomarkers.

Mouse pharmacokinetic study

Pharmacokinetic studies were determined in female SCID mice with Daudi xenografts (n = 3). BEBT-908 compound was intravenously administrated at dose of 25 or 100 mg/kg. Plasma was collected into EDTA containing tubes by standard centrifugation techniques. Timepoints for determination of pharmacokinetic parameters were 5 minutes, 30 minutes, 1 hour, 3 hours, 6 hours, and 24 hours. Plasma concentrations were determined via LC/MS-MS using a six-point standard curve between 1 and 1,000 ng/mL prepared in mouse plasma. Samples above the curve were diluted with blank mouse plasma. Pharmacokinetic analysis was performed with WinNonlin (version 5.3). In the meantime, the tumors of indicated timepoints were harvested and homogenated to determine the concentration of compound in tumor tissues.

Tumor-infiltrating lymphocytes analysis and MHC I analysis

About 2 × 105 MC38 tumor cells were inoculated subcutaneously into C57BL/6 mice. A total of 100 mg/kg BEBT-908 was intravenously injected every other day for four times initiating from day 4 after tumor cell inoculation. Tumors were harvested on day 12 after inoculation, weighted, mechanically minced and incubated with 50 μg/mL DNase I (Sigma) and 2 mg/mL collagenase P (Sigma) for 20 minutes at 37°C. The dissociated cells were filtered, blocked with an anti-CD16/32 antibody and stained with indicated surface antibodies for 20 minutes on ice. Dead cells were marked by use of Live/Dead Fixable Violet dye (Thermo Fisher Scientific). Intracellular antibodies were added after fixation and permeabilization following the manufacturer's instructions (BioLegend). For MHC I detection in vivo, MC38 with stable GFP expression cells were inoculated into C57BL/6 mice and treated with the same procedure. Tumor cells were identified with GFP and stained with anti-H-2Kb/H-2Db antibody. The samples were analyzed and quantified using flow cytometer Cytoflex (Beckman). The antibodies used in the study were listed in Supplementary Table S2.

RNA sequencing and bioinformatics analysis

Total RNA from treated Daudi cells was extracted and quantified using a Nano Drop and Agilent 2100 bioanalyzer. Library preparation and sequencing analysis (BGISEQ-500, pair-end 150 bp) were performed by BGI RNA sequencing (RNA-seq; ShenZhen). The expression levels of genes were calculated by RSEM (v1.2.12). The heatmap was drawn by heatmap (v1.0.8) according to the gene expression in different samples. Essentially, differential expression analysis was performed using the DESeq2(v1.4.5) with Q value ≤ 0.05. To obtain biological insights of the phenotypic changes, Kyoto Encyclopedia of Genes and Genomes (KEGG; https://www.kegg.jp/) enrichment analysis of annotated different expression gene was performed by Phyper (https://en.wikipedia.org/wiki/Hypergeometric_distribution) based on Hypergeometric test. The significant levels of terms and pathways were corrected by Q value with a rigorous threshold (Q ≤ 0.05) by Bonferroni.

Tumor models and treatments

Xenografts for Daudi and H2122 in severe combined immunodeficient (SCID) mice were approved by Curis Inc. Institutional Animal Use and Care Committee and carried out by use of 6–8 weeks old female SCID mice from Charles River Laboratories. Other animal experiments conducted in this study were approved by Sun Yat-sen University Institutional Animal Use and Care Committee. C57BL/6, and athymic nude mice were from Sun Yat-sen University animal facility. Mice were housed in an environmentally controlled room (temperature 23°C ± 2°C, relative humidity 30%–70%), fed irradiated laboratory rodent diet, and provided with sterilized water. Tumor cells were inoculated into 6–8 weeks old mice subcutaneously. When the volume of xenografts reached around 100–200 mm3, tumor-bearing mice were randomly separated into different groups including control, BEBT-908, or anti-PD1 antibody group. The in vivo formation of BEBT-908 contains 30% sulfobutyl ether-beta-cyclodextrin sodium (30% cyclodextrin) and 1 mol/L NaOH. The formation with or without BEBT-908 was intravenously administrated every other day for totally four times. Anti-PD1 antibody (Clone RMP1-14, BioXcell) were intraperitoneally administrated at a dose of 100 μg/mice on day 5, and 8 after MC38 cell inoculation. No blinding was used in our tumor growth delay experiments. Tumor size were measured by use of a caliper and calculated using the formula Volume = (length)(width)2/2. Endpoint was defined as the time when a progressively growing tumor reached 15 mm in the longest dimension. Tumor samples were collected at the indicated times and processed for bioanalysis.

Statistical analysis

Statistical analyses were performed using GraphPad Prism 8 software and statistical significance was defined as P value less than 0.05. All results are shown as mean ± SEM. Comparisons between two groups were conducted by use of unpaired two-sided Student t test. Two-way ANOVA was used for multiple comparisons in tumor growth experiments. Log-rank test was used for mouse survival analysis.

Data availability

RNA-seq data are accessioned in Bioproject database (accession PRJNA692104).

A novel dual PI3K/HDACi BEBT-908 potently inhibited tumor growth in vitro

To obtain a highly potent PI3K and HDAC bifunctional inhibitor, we synthesized and optimized a series of dual-targeting PI3K and HDACis by incorporating HDAC inhibitory functionality into PI3K inhibitor pharmacophore, and discovered that BEBT-908 (Fig. 1A) exhibits a predominately isoform-selective activity against PI3Kα kinase (Fig. 1B) with mean IC50 value of 9.3 nmol/L, which compared favorably with that of pictilisib (GDC-0941), a pan-PI3K inhibitor. BEBT-908 also showed significant inhibition activities against HDAC1, HDAC2, HDAC3, HDAC10, and HDAC11 with average IC50 value from 0.9 to 2.7 nmol/L, which compared favorably with pan-HDACis vorinostat (SAHA) and panobinostat (LBH589; Fig. 1C). We further investigated the undesirable off-target kinase activity of BEBT-908 in vitro by use of a commercial kinase panel assay (Eurofins Cerep) following standard procedure. Kinase profiling study demonstrated that extremely high concentration (10 μmol/L) of BEBT-908 treatment showed minimal potential off-target inhibitory effect on TRKA, MST4, and mTOR kinase with inhibitory rate of 65%, 75%, and 99%, respectively (Supplementary Fig. S2A). The off-target kinase effect on mTOR of BEBT-908 was further determined to have an IC50 value of 227.5 nmol/L (Supplementary Fig. S2B), which is over 24-fold higher than that of on-target PI3Kα. These data suggested BEBT-908 was a selective bifunctional PI3K and HDACis. To determine the effect of BEBT-908 on cell proliferation, we treated tumor cells with PI3K inhibitor (GDC-0941) and HDACi (SAHA) alone or in combination, or with BEBT-908 for 72 hours in multiple cancer cell lines with indicated PI3KCA, P53-mutant status (Supplementary Fig. S2C). Our data showed that BEBT-908 was at least over 50-fold of more potent when compared with individual or combination treatment of PI3K inhibitor (GDC-0941) and HDACi (SAHA) in human Burkitt lymphoma Daudi cells (Fig. 1D), non–small cell lung cancer H2122 cells (Fig. 1E), and colorectal cancer HCT116 cells (Fig. 1F). BEBT-908 potently reduced cell viability in a wide array of cancer types (hematologic cancer, Fig. 1G; lung cancer, Fig. 1H; and colorectal cancer, Fig. 1I) at single-digit nanomolar concentration in a time-dependent manner (Fig. 1J and K). Further evidence for the specificity of BEBT-908 was obtained by Western blot analysis. AKT and eukaryotic initiation factor 4E binding protein-1 (4EBP1) are the downstream effectors of PI3K signaling to regulate multiple cellular processes through its phosphorylation. BEBT-908 significantly inhibited phosphorylation of AKT and 4EBP1 (Fig. 1L). Furthermore, BEBT-908 treatment induced hyperacetylation of proteins including histone H3, tubulin, and P53 in a dose-dependent manner (Fig. 1M; Supplementary Fig. S2D–S2F) in either P53 wild-type or P53-mutant cancer cells. These data demonstrated that dual PI3K/HDACi BEBT-908 is a highly potent anticancer agent.

Figure 1.

The structure and bifunctional PI3K and HDACs inhibition activities of BEBT-908. A, The structure of bifunctional PI3K and HDAC inhibitor, BEBT-908. B and C, Heatmap of half maximal inhibitory concentration (IC50) values for inhibitors targeting PI3K kinases (B) and HDACs (C). n/a, an IC50 could not be detected. D–F, Cell viability of Daudi (D), H2122 (E), and HCT116 (F) cells after treatment with serial doses of SAHA, GDC-0941, and BEBT-908 for 72 hours. Data are mean ± SEM; n = 3 biological replicates. G–I, The effect of BEBT-908 on cell viability in hematologic cancer (G), lung cancer (H), and colorectal cancer (I) cell lines after 72 hours treatment. Data are mean ± SEM; n = 3 biological replicates. J and K, Cell viability after BEBT-908 treatment at a dose of 1.6 nmol/L in Daudi (J) and 8 nmol/L in HCT116 (K) cells at indicated times. Data are mean ± SEM; n = 3 biological replicates. Western blotting analysis of phosphorylated AKT (pAKT), 4EBP1 (p4EBP1; L) and acetylated Histone H3 (Ac-H3), tubulin (Ac-Tub), P53 (Ac-P53; M) after treatment with BEBT-908 and reference compounds for 6 hours in H460 cells. Tubulin was used as an internal loading control. The experiments were independently repeated one additional time with similar results.

Figure 1.

The structure and bifunctional PI3K and HDACs inhibition activities of BEBT-908. A, The structure of bifunctional PI3K and HDAC inhibitor, BEBT-908. B and C, Heatmap of half maximal inhibitory concentration (IC50) values for inhibitors targeting PI3K kinases (B) and HDACs (C). n/a, an IC50 could not be detected. D–F, Cell viability of Daudi (D), H2122 (E), and HCT116 (F) cells after treatment with serial doses of SAHA, GDC-0941, and BEBT-908 for 72 hours. Data are mean ± SEM; n = 3 biological replicates. G–I, The effect of BEBT-908 on cell viability in hematologic cancer (G), lung cancer (H), and colorectal cancer (I) cell lines after 72 hours treatment. Data are mean ± SEM; n = 3 biological replicates. J and K, Cell viability after BEBT-908 treatment at a dose of 1.6 nmol/L in Daudi (J) and 8 nmol/L in HCT116 (K) cells at indicated times. Data are mean ± SEM; n = 3 biological replicates. Western blotting analysis of phosphorylated AKT (pAKT), 4EBP1 (p4EBP1; L) and acetylated Histone H3 (Ac-H3), tubulin (Ac-Tub), P53 (Ac-P53; M) after treatment with BEBT-908 and reference compounds for 6 hours in H460 cells. Tubulin was used as an internal loading control. The experiments were independently repeated one additional time with similar results.

Close modal

BEBT-908 treatment alters the global transcriptome patterns and results in ferroptosis

To evaluate the effect of BEBT-908 on global cellular transcriptome of treated cells, RNA-seq analysis was performed in tumor cells treated with DMSO (Ctrl.), BEBT-908 or combined SAHA and GDC-0941 treatment for 5 days. Compared with DMSO, combination (SAHA+GDC-0941) and BEBT-908 treatment induced clear gene expression changes (Fig. 2A and B). Intriguingly, there were a lot of differences in gene expression between combination (SAHA + GDC-0941) and BEBT-908–treated cells despite being possessing equivalent IC50 values to either of their targets (Fig. 2B; Supplementary Fig. S3A). We then performed a KEGG pathway analysis on the upregulated genes (Supplementary Fig. S3B) and the downregulated genes (Supplementary Fig. S3C) in comparison of BEBT-908–treated and combination-treated cells. Compared with the combination (SAHA+GDC-0941) treatment, bifunctional PI3K/HDACi BEBT-908 treatment induced higher level expression of transcripts associated with cell-cycle G2–M checkpoint arrest (Supplementary Fig. S3B and S3D) and relative lower level apoptosis-related gene expression (Supplementary Fig. S3C). The effects of BEBT-908 on G2–M cell-cycle arrest and apoptosis were confirmed by flow cytometry analysis of BEBT-908 or SAHA+GDC-0941–treated cells (Supplementary Fig. S3E–S3H).

Figure 2.

BEBT-908 inhibition of PI3K and HDACs altered global gene expression and resulted in ferroptotic cell death. A, Heatmap and hierarchical cluster analysis of the whole transcriptome in DMSO (Ctrl.), 89 nmol/L SAHA+GDC-0941, or 0.8 nmol/L BEBT-908–treated Daudi cells on day 5. B, Venn diagrams showing the overlap in differentially expressed genes in reponse to drug treatment of each comparison. Gene numbers in each section are shown in parentheses. C, KEGG analysis showing enrichment signaling pathways in BEBT-908–treated cells when compared with DMSO-treated cells. D, Western blotting analysis showing downregulation of ferroptosis-related factors (SLC7A11, GPX4, Nrf2) in HCT116 cells treated with DMSO (Ctrl.), 236 nmol/L SAHA+GDC-0941, 4 nmol/L BEBT-908, and 10 μmol/L erastin for 5 days. GAPDH was used as the internal loading control. The experiments were independently repeated one additional time with similar results. E–H, RT-PCR analyses of the expression of TRFC (E), KEAP1 (F), NOXO1 (G), and GCLM (H) in HCT116 cells treated with DMSO (Ctrl.), 236 nmol/L SAHA+GDC-0941, 4 nmol/L BEBT-908, and 10 μmol/L erastin for 5 days. n = 4 biological replicates. Data represent mean ± SEM. P values were determined by unpaired two-sided t test. **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. I and J, ROS (I) and MDA (J) generated in Ctrl., SAHA+GDC-0941, BEBT-908 at dose of indicated IC50 and 10 μmol/L erastin-treated HCT116 cells on day5. n = 3 biological replicates. Data represent mean ± SEM. P values were determined by unpaired two-sided t test. **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. K and L, GPX4 knockdown (K) increased BEBT-908–induced cell death in HCT116 cells (L). Data represent mean ± SEM. n = 3 biological replicates. P values were determined by unpaired two-sided t test. *, P < 0.05. M, Ferroptosis inhibition by 10 μmol/L ferrostatin-1 (Ferr1) suppressed BEBT-908–induced cell death in HCT116 cells. Data represent mean ± SEM. n = 3 biological replicates. P values were determined by unpaired two-sided t test. ***, P < 0.001.

Figure 2.

BEBT-908 inhibition of PI3K and HDACs altered global gene expression and resulted in ferroptotic cell death. A, Heatmap and hierarchical cluster analysis of the whole transcriptome in DMSO (Ctrl.), 89 nmol/L SAHA+GDC-0941, or 0.8 nmol/L BEBT-908–treated Daudi cells on day 5. B, Venn diagrams showing the overlap in differentially expressed genes in reponse to drug treatment of each comparison. Gene numbers in each section are shown in parentheses. C, KEGG analysis showing enrichment signaling pathways in BEBT-908–treated cells when compared with DMSO-treated cells. D, Western blotting analysis showing downregulation of ferroptosis-related factors (SLC7A11, GPX4, Nrf2) in HCT116 cells treated with DMSO (Ctrl.), 236 nmol/L SAHA+GDC-0941, 4 nmol/L BEBT-908, and 10 μmol/L erastin for 5 days. GAPDH was used as the internal loading control. The experiments were independently repeated one additional time with similar results. E–H, RT-PCR analyses of the expression of TRFC (E), KEAP1 (F), NOXO1 (G), and GCLM (H) in HCT116 cells treated with DMSO (Ctrl.), 236 nmol/L SAHA+GDC-0941, 4 nmol/L BEBT-908, and 10 μmol/L erastin for 5 days. n = 4 biological replicates. Data represent mean ± SEM. P values were determined by unpaired two-sided t test. **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. I and J, ROS (I) and MDA (J) generated in Ctrl., SAHA+GDC-0941, BEBT-908 at dose of indicated IC50 and 10 μmol/L erastin-treated HCT116 cells on day5. n = 3 biological replicates. Data represent mean ± SEM. P values were determined by unpaired two-sided t test. **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. K and L, GPX4 knockdown (K) increased BEBT-908–induced cell death in HCT116 cells (L). Data represent mean ± SEM. n = 3 biological replicates. P values were determined by unpaired two-sided t test. *, P < 0.05. M, Ferroptosis inhibition by 10 μmol/L ferrostatin-1 (Ferr1) suppressed BEBT-908–induced cell death in HCT116 cells. Data represent mean ± SEM. n = 3 biological replicates. P values were determined by unpaired two-sided t test. ***, P < 0.001.

Close modal

To understand the potency of BEBT-908 in suppressing tumor cell growth, we further analyzed the signaling pathways in BEBT-908–treated cells in comparison with those of vehicle control. As shown in Fig. 2C, BEBT-908 treatment induced robust enrichment of transcripts such as those of P53 signaling pathway, ferroptosis, and glutathione metabolism. Ferroptosis is a form of cell death that differs from apoptosis and results in iron-dependent accumulation of lipid peroxide (19–21), and the tumor suppressor P53 is closely associated with the sensitivity to ferroptosis (22–24). Our results suggested that BEBT-908–induced hyperacetylation of P53 at lysine 370 residue might sensitize cancer cells to ferroptosis. Previous reports demonstrated that intact or acetylated P53 could directly bind the promoter region of cystine/glutamate reverse transporter system xc subunit SLC7A11 (xCT) to repress its transcription (22, 24), which is essential for ferroptosis induction. We then detected the expression of two key regulators of ferroptosis, SLC7A11 and glutathione peroxidase 4 (GPX4) in DMSO, SAHA+GDC-0941, BEBT-908, and erastin (a type 1 ferroptosis inducer as positive control) treated cells. Our results indicated that BEBT-908 treatment significantly repressed the expression of SLC7A11 and GPX4 (Fig. 2D) in HCT116 cancer cells when compared with DMSO control or SAHA+GDC-0941 combination treatment. Inactivation of SLC7A11 and GPX4 was shown previously to be sufficient to induce ferroptosis (20, 21, 25). Furthermore, BEBT-908 treatment inhibited the expression of nuclear factor erythroid 2–related factor 2 (Nrf2; Fig. 2D), a negative regulator of ferroptosis. Nrf2 positively regulates many anti-ferroptosis genes, including those involved in preventing lipid peroxidation and the accumulation of free iron in the cells (26, 27). In addition, we confirmed the upregulation of iron metabolism-related genes (TRFC, KEAP1, and NOXO1) and the downregulation of redox glutathione metabolism-related gene (GCLM) in BEBT-908–treated HCT116 cells (Fig. 2EH). BEBT-908–treated HCT116 cells were also found to accumulate lipid ROS (Fig. 2I) and a product of lipid peroxidation, MDA (Fig. 2J), which promotes ferroptosis. BEBT-908–mediated activation of ferroptotic signaling (Supplementary Fig. S3I–S3M) and accumulation of ROS (Supplementary Fig. S3N), MDA (Supplementary Fig. S3O) were confirmed in human Daudi cancer cells by Western blot analysis, RT-PCR and flow cytometry, respectively. Genetic knockdown of GPX4 expression (Fig. 2K) increased BEBT-908–induced cell death (Fig. 2L). We also demonstrated that pharmacological inhibition of ferroptosis by ferrostatin-1–alleviated BEBT-908 induced cell death (Fig. 2M), indicating ferroptosis to be a major mode of cell death induced by BEBT-908.

BEBT-908 treatment delays tumor growth in vivo

Our data showed that PI3K inhibitor, HDACi alone or in combination treatment only has modest antitumor ability at MTD in Daudi xenograft SCID mice model (Supplementary Fig. S4A). To evaluate the antitumor efficacy of BEBT-908 in vivo, we first investigated the pharmacokinetic property of BEBT-908 in Daudi xenograft model. After intravenously administration BEBT-908, the plasma, and tumor tissues were harvested in various timepoints. LC/MS-MS assay revealed that the maximum concentration observed (Cmax) in plasma was 8,5517.0 ng/mL, the area under the concentration–time curve (AUC) in plasma was 18,039.9 ng/mL*hour, and the half-life (t1/2) in plasma was 5.24 hours (Fig. 3A; Supplementary Fig. S4B). The Cmax, AUC, and t1/2 in tumor tissues was 14,526.1 ng/g, 17157.9 ng/g*hour, and 7.9 hours, respectively (Fig. 3B; Supplementary Fig. S4B). Twenty-four hours after administration, the concentration of BEBT-908 in tumors remained 100 ng/g (around 200 nmol/L), which was over 200-fold of its IC50in vitro. The pharmacokinetic data suggested that BEBT-908 was suitable for intravenously administration every other day in next experiments of antitumor efficacy evaluation.

Figure 3.

The pharmacokinetic and antitumor activity of BEBT-908 in human diffuse large B-cell lymphoma xenograft model. A, Plasma concentration of BEBT-908 after intravenous administration of BEBT-908 in Daudi mice xenografts. n = 3 mice per group. Data represent mean ± SEM. B, The concentration of BEBT-908 in tumor tissues after intravenous administration of BEBT-908 in Daudi mice xenografts. n = 3 mice per group. Data represent mean ± SEM. C and D, SCID mice with established Daudi tumors were treated with indicated doses of BEBT-908 by intravenous injections. BEBT-908 treatment significantly inhibited tumor growth (C) without body weight loss of host mice (D). n = 8 mice per group. Data present mean ± SEM. P value was determined by two-way ANOVA.

Figure 3.

The pharmacokinetic and antitumor activity of BEBT-908 in human diffuse large B-cell lymphoma xenograft model. A, Plasma concentration of BEBT-908 after intravenous administration of BEBT-908 in Daudi mice xenografts. n = 3 mice per group. Data represent mean ± SEM. B, The concentration of BEBT-908 in tumor tissues after intravenous administration of BEBT-908 in Daudi mice xenografts. n = 3 mice per group. Data represent mean ± SEM. C and D, SCID mice with established Daudi tumors were treated with indicated doses of BEBT-908 by intravenous injections. BEBT-908 treatment significantly inhibited tumor growth (C) without body weight loss of host mice (D). n = 8 mice per group. Data present mean ± SEM. P value was determined by two-way ANOVA.

Close modal

Therefore, we carried out antitumor efficacy evaluation of BEBT-908 in human diffuse large B-cell lymphoma xenograft model. BEBT-908 significantly inhibited the growth of human Burkitt lymphoma (Daudi) at all doses, and caused almost complete regression of tumors (6 of 8 tumor free mice) at 100 mg/kg dose group (Fig. 3C). The percent treatment/control (T/C) values (28) were calculated to be −6.3% for the 25 mg/kg group (P < 0.001, ANOVA) and 0% for the 50 mg/kg group (P < 0.001, ANOVA). The regression is calculated to be −78.4% for the 100 mg/kg group (P < 0.001, ANOVA). No body weight loss or other side effect was observed for the BEBT-908 single-agent treatment group (Fig. 3D). Our data suggested that dual PI3K/HDACi BEBT-908 might be an effective anticancer agent for patients with advanced hematologic malignancies as a monotherapy.

Furthermore, we investigated the antitumor efficacy of BEBT-908 in solid tumor model. BEBT-908 also significantly inhibited the growth of non–small cell lung cancer (H2122; Fig. 4A), colorectal cancer (HCT116, MC38; Fig. 4B and C) in immune-deficient mice (HCT116) or in syngeneic immune-competent C57BL/6 mice (MC38) without significant body weight loss (Supplementary Fig. S4C–S4E). Of note, two of eight MC38 tumors in syngeneic C57BL/6 mice completely regressed after BEBT-908 treatment. IHC staining analysis of Ki67, Acetyl-p53, Nrf2, and GPX4 in tumor sections (Fig. 4DG) indicated that BEBT-908–treated tumors had significantly lower levels of cell proliferation (Ki67), higher levels of ferroptotic cell death, and MDA accumulation in tumor tissues (Supplementary Fig. S4F and S4G). Our data suggested that dual PI3K/HDACi BEBT-908 has great potency against solid malignancies such as colorectal cancer through ferroptotic cell death.

Figure 4.

BEBT-908 treatment inhibits solid tumor growth through ferroptosis. A–C, Mice with established H2122 tumors (SCID mice; A), HCT116 tumors (nude mice; B), and MC38 (C57BL/6; C) were treated with BEBT-908 by intravenous injections. n = 8 mice per group. Data present mean ± SEM. P values were determined by two-way ANOVA. D–G, IHC staining of a cell proliferation marker (Ki67; D) and ferroptosis-related proteins (Acetyl-p53, E; Nrf2, F; and Gpx4, G) in Ctrl.- and BEBT-908–treated HCT116 tumor sections at the end of animal experiments. Scale bars, 200 μm. Quantitative analysis of IHC staining of cell proliferation marker (Ki67) and ferroptosis-related proteins (Acetyl-p53, Nrf2, and GPX4) in the right of respective panel by use of ImageJ with the described procedure in Materials and Methods. Data were collected from four random fields of for each of the five tumor samples and are presented as mean ± SEM. P values were determined by unpaired two-sided t test. ****, P < 0.0001.

Figure 4.

BEBT-908 treatment inhibits solid tumor growth through ferroptosis. A–C, Mice with established H2122 tumors (SCID mice; A), HCT116 tumors (nude mice; B), and MC38 (C57BL/6; C) were treated with BEBT-908 by intravenous injections. n = 8 mice per group. Data present mean ± SEM. P values were determined by two-way ANOVA. D–G, IHC staining of a cell proliferation marker (Ki67; D) and ferroptosis-related proteins (Acetyl-p53, E; Nrf2, F; and Gpx4, G) in Ctrl.- and BEBT-908–treated HCT116 tumor sections at the end of animal experiments. Scale bars, 200 μm. Quantitative analysis of IHC staining of cell proliferation marker (Ki67) and ferroptosis-related proteins (Acetyl-p53, Nrf2, and GPX4) in the right of respective panel by use of ImageJ with the described procedure in Materials and Methods. Data were collected from four random fields of for each of the five tumor samples and are presented as mean ± SEM. P values were determined by unpaired two-sided t test. ****, P < 0.0001.

Close modal

Activation IFNγ signaling and upregulation MHC I expression in ferroptotic cancer cells exposed to BEBT-908

Because pharmacologic inhibition of HDACi or PI3Ki was shown to induce activation of the IFN pathway and increased chemokines (29–31), we next examined whether BEBT-908 induces immune-related changes in cancer cells. Gene enrichment analysis of RNA-seq data showed that BEBT-908 treatment upregulated IFNγ response pathway (Supplementary Fig. S5A). Heatmap analysis of individual gene indicated that IFNγ response-related genes were significantly increased in BEBT-908 treatment Daudi cells (Fig. 5A). We further confirmed the upregulation of IFNγ-related genes (IFI27, IFIT1, IFNG, and CXCL10) in SAHA+GDC-0941, and BEBT-908–treated HCT116 cells (Fig. 5B) as well as Daudi cells (Supplementary Fig. S5B). The endogenous IFNγ release further determined by flow cytometry (Fig. 5C and D). Previous study demonstrated that exogenous IFNγ stimulation modulates MHC I expression (32–34). Therefore, we investigated the MHC I expression of BEBT-908–treated tumor cells with autonomic IFNγ activation. As shown in Fig. 5E, the MHC-related genes were upregulated in BEBT-908–treated cells. qRT-PCR analysis confirmed that MHC I gene (HLA-A, TAP1 for human, and H-2K1 for mouse) expression was significantly increased in BEBT-908–treated human Daudi cancer cells (Supplementary Fig. S5C and S5D) and mouse MC38 cancer cells, respectively (Fig. 5F), when compared with DMSO, or SAHA+GDC-0941 treatment. Flow cytometry analysis further confirmed MHC I (H-2Kb/H-2Db) upregulation (Supplementary Fig. S5E and S5F, Fig. 5G) and PD-L1 upregulation (Supplementary Fig. S5G and S5H) on BEBT-908–treated cells in vitro and in xenografts. Upregulation of MHC I in tumor cells suggested that BEBT-908–treated cancer cells may become more immunogenic and susceptible to T cell–mediated killing.

Figure 5.

BEBT-908 treatment induces immunogenic ferroptosis of cancer cells. A, Heatmap analysis of RNA-seq data showing the expression of IFNγ pathway–related genes in DMSO (Ctrl.)- and BEBT-908–treated Daudi cells on day 5. B, qRT-PCR data showing the expression of IFNγ signaling pathway–related genes in Ctrl., 236 nmol/L SAHA+GDC-0941, and 4 nmol/L BEBT-908–treated HCT116 cells on day 5. C and D, Flow cytometry analysis (C) and quantitation (D) of IFNγ cytokine expression in Ctrl., 1656 nmol/L SAHA+GDC-0941, and 8 nmol/L BEBT-908–treated MC38 cells on day 1. E, Heatmap analysis of RNA-seq data showing the expression of MHC-related genes in DMSO (Ctrl.)- and BEBT-908–treated Daudi cells on day 5. F, qRT-PCR data showing the expression of MHC I–related gene (H-2K1) in Ctrl.- and 8 nmol/L BEBT-908–treated MC38 cells on day 5. G, FACS analysis of H2-Kb/H2-Db expression on the surface of Ctrl.- and BEBT-908–treated MC38 tumor tissues on day 12 after inoculation. Data in B, D, F, and G are presented as mean ± SEM. P values were determined by unpaired two-sided t test; n = 4 biological replicates for B and F; n = 6 for D; n = 6 tumors for G. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. H, IHC staining of phosphorylated STAT1 (pSTAT1) in Ctrl.- and BEBT-908–treated MC38 tumors on 12 days after inoculation. Scale bars, 200 μm. I, Quantitative data of IHC staining of pSTAT1 in Ctrl.- and BEBT-908–treated MC38 tumors 12 days after inoculation. Data were collected from five random fields for each of the five tumor samples and are presented as mean ± SEM. P values were determined by unpaired two-sided t test. ****, P < 0.0001. J, Western blotting analysis pSTAT1 expression in MC38 cells treated with Ctrl., 8 nmol/L BEBT-908, 10 μmol/L erastin, and 1 μmol/L RSL3 for 1 hour. GAPDH was used as an internal loading control. K and L, qPCR analysis MHC I (H-2K1) expression (K) and cell surface MHC I (H2-Kb/H2-Db) expression (L) of MC38 cells treated with Ctrl., 8 nmol/L BEBT-908, 10 μmol/L erastin, and 1 μmol/L RSL3 for 24 hours. Data presented as mean ± SEM. P values were determined by unpaired two-sided t test. n = 3 biological replicates for K, and n = 4 for L. *, P < 0.05; ***, P < 0.001. M, FACS analysis of H2-Kb/H2-Db expression on the surface of Ctrl., BEBT-908, and ferroptosis inhibitor (Ferr-1)–treated MC38 cells at dose of indicated IC50 on day 1 after treatment.

Figure 5.

BEBT-908 treatment induces immunogenic ferroptosis of cancer cells. A, Heatmap analysis of RNA-seq data showing the expression of IFNγ pathway–related genes in DMSO (Ctrl.)- and BEBT-908–treated Daudi cells on day 5. B, qRT-PCR data showing the expression of IFNγ signaling pathway–related genes in Ctrl., 236 nmol/L SAHA+GDC-0941, and 4 nmol/L BEBT-908–treated HCT116 cells on day 5. C and D, Flow cytometry analysis (C) and quantitation (D) of IFNγ cytokine expression in Ctrl., 1656 nmol/L SAHA+GDC-0941, and 8 nmol/L BEBT-908–treated MC38 cells on day 1. E, Heatmap analysis of RNA-seq data showing the expression of MHC-related genes in DMSO (Ctrl.)- and BEBT-908–treated Daudi cells on day 5. F, qRT-PCR data showing the expression of MHC I–related gene (H-2K1) in Ctrl.- and 8 nmol/L BEBT-908–treated MC38 cells on day 5. G, FACS analysis of H2-Kb/H2-Db expression on the surface of Ctrl.- and BEBT-908–treated MC38 tumor tissues on day 12 after inoculation. Data in B, D, F, and G are presented as mean ± SEM. P values were determined by unpaired two-sided t test; n = 4 biological replicates for B and F; n = 6 for D; n = 6 tumors for G. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. H, IHC staining of phosphorylated STAT1 (pSTAT1) in Ctrl.- and BEBT-908–treated MC38 tumors on 12 days after inoculation. Scale bars, 200 μm. I, Quantitative data of IHC staining of pSTAT1 in Ctrl.- and BEBT-908–treated MC38 tumors 12 days after inoculation. Data were collected from five random fields for each of the five tumor samples and are presented as mean ± SEM. P values were determined by unpaired two-sided t test. ****, P < 0.0001. J, Western blotting analysis pSTAT1 expression in MC38 cells treated with Ctrl., 8 nmol/L BEBT-908, 10 μmol/L erastin, and 1 μmol/L RSL3 for 1 hour. GAPDH was used as an internal loading control. K and L, qPCR analysis MHC I (H-2K1) expression (K) and cell surface MHC I (H2-Kb/H2-Db) expression (L) of MC38 cells treated with Ctrl., 8 nmol/L BEBT-908, 10 μmol/L erastin, and 1 μmol/L RSL3 for 24 hours. Data presented as mean ± SEM. P values were determined by unpaired two-sided t test. n = 3 biological replicates for K, and n = 4 for L. *, P < 0.05; ***, P < 0.001. M, FACS analysis of H2-Kb/H2-Db expression on the surface of Ctrl., BEBT-908, and ferroptosis inhibitor (Ferr-1)–treated MC38 cells at dose of indicated IC50 on day 1 after treatment.

Close modal

We next focused on the molecular mechanism involved in upregulation of MHC I expression in BEBT-908–treated cells. It was reported that IFNγ-mediated phosphorylation of STAT1 triggers higher-order chromatin remodeling of MHC locus for activation of MHC genes (35). Therefore, we hypothesized that BEBT-908–mediated SLC7A11 downregulation activated STAT1 phosphorylation to increase MHC I expression on ferroptotic cancer cells. We first examined phosphorylation of STAT1 (pSTAT1) in tumor sections by IHC staining. Our results indicated that BEBT-908 treatment significantly increased pSTAT1 expression on BEBT-908–treated MC38 tumors when compared with those of control (Fig. 5H and I). Western blot analysis also confirmed that BEBT-908 treatment increased phosphorylated STAT1 in MC38 cells (Fig. 5J, second lane from left). Interestingly, erastin, a type 1 ferroptosis inducer that selective inhibiting xCT transporter, also stimulated pSTAT1 expression (Fig. 5J, third lane from left). However, RSL3, a type II ferrtoptosis inducer that directly inhibits GPX4 had no effect on STAT1 phosphorylation (Fig. 5J, right lane). These results are consistent with the upregulation of MHC I expression on BEBT-908–treated, and erastin-treated cells on transcription level (Fig. 5K) and protein level (Fig. 5L). Our findings suggested that different ferroptotic effectors have disparate mechanisms in regulating tumor immunity. On the other way, when we introduced ferroptosis inhibitor in BEBT-908–treated cells, BEBT-908–mediated MHC I upregulation was significantly attenuated by the ferroptosis inhibitor in both human Daudi, HCT116 cells (Supplementary Fig. S5I and S5J) and mouse MC38 cells (Fig. 5M). A previous study focused on immunotherapy-activated CD8+ T lymphocytes to demonstrate that CD8+ T cell–derived IFNγ induced tumor cell ferroptotic cell death through STAT1/SLC7A11 signaling (36). Our results therefore demonstrated that the dual PI3K/HDACi BEBT-908 not only induced ferroptosis in tumor cells, but also the increased ferroptotic signaling stimulated MHC I expression on tumor cells.

BEBT-908 promotes a proinflammatory tumor microenvironment and synergizes with immunotherapy

Next, we attempted to quantify tumor-infiltrating lymphocytes in control and BEBT-908–treated tumors by immunofluorescence staining, flow cytometry, and qRT-PCR. Immunofluorescence staining indicated that BEBT-908 treatment caused an overall increase tumor infiltration of CD45+ leukocytes (Fig. 6A and B). Of particular interest was the observation of CD45+ cells staying mostly in the periphery in control tumors (Fig. 6A, top) but moving into tumor cell-rich areas in the BEBT-908–treated tumors (Fig. 6A, bottom). Flow cytometry analyses confirmed significant increase in intratumoral CD45+ leukocytes (Fig. 6C), CD8+ cytotoxic T cells (CTL; Fig. 6D) in BEBT-908–treated tumors, as well as intratumoral expression of CD8a (qRT-PCR; Supplementary Fig. S6A). In contrast, no significant increases in CD4+ Th cells (Supplementary Fig. S6B), CD4+Foxp3+ regulatory T (Treg; Supplementary Fig. S6C), or natural killer cells (Supplementary Fig. S6D) were observed. The ratio of CTLs versus Tregs in the BEBT-908–treated tumors were significantly increased (Fig. 6E). Consistently, the number of granzyme B+ (GZMB+) and IFNγ+ CTLs also increased significantly in BEBT-908–treated tumors (Fig. 6F and G), as well as intratumoral mRNA expression of GZMB and IFNG (qRT-PCR; Fig. 6H and I). These data demonstrated that bifunctional PI3K/HDACi BEBT-908 treatment not only made tumor cells more immunogenic for lymphocyte recognition, but also led lymphocytes infiltrating in tumors for CTL-mediated killing.

Figure 6.

BEBT-908 treatment induces a proinflammatory tumor microenvironment and potentiates ICB therapy. A, Immunofluorescene staining of CD45 in Ctrl. and BEBT-908–treated MC38 tumors on day 12 after inoculation. DAPI was used for nucleus staining. Scale bars, 50 μm. B, Quantitative data of immunofluorescence staining of CD45+ in Ctrl.- and BEBT-908–treated MC38 tumors 12 days after inoculation. Data were collected from four tumor samples and are presented as mean ± SEM. P values were determined by unpaired two-sided t test. ***, P < 0.001. C and D, Quantitative estimate of the numbers of CD45+ (C) and CD8+ (D) lymphocytes per mg of tumor tissue in Ctrl.- and BEBT-908–treated MC38 tumors as determined by flow cytometry. E, Ratio of CD8+ T/CD4+Foxp3+ Treg cells in Ctrl.- and BEBT-908–treated MC38 tumors. F and G, Average numbers of tumor-infiltrating GZMB+CD8+ T(F) and IFNγ+ CD8+ T (G) cells per mg of tumor tissue in Ctrl.- and BEBT-908–treated tumors. H and I, The qRT-PCR data showing the expression of GZMB (H) and IFNG (I) in Ctrl.- and BEBT-908–treated MC38 tumors on day 12 after inoculation. Data of CI are presented as mean ± SEM. P values were determined by unpaired two-sided t test; n = 8, 7 for Ctrl. and BEBT-908 group of C–G, respectively. n = 5 for H and I. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. J–L, Experimental protocol (J), tumor growth curve (K), and overall survival (L) of combination treatment of BEBT-908 (100 mg/kg, every other day for four times) and anti-PD1 antibody (clone RMP1–14, BioXcell, 100 μg/mice, two times) on MC38 tumors in syngeneic mice. n = 8 mice per group. Data in K are presented as mean ± SEM. P values were determined by two-way ANOVA (K) and log-rank test (L). *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. M and N, Tumor-free survivors of combined BEBT-908 and anti-PD1 antibody treatment were rechallenged with a lethal number of parental MC38 cells. M and N, Individual tumor growth (M) and overall survival (N) of host mice after rechallenge. Age-matched naïve C57BL/6 mice (WT) were used as control. n = 5 mice per group. P values were determined by unpaired two-sided t test on day 10 (M) and log-rank test (N). **, P < 0.01; ****, P < 0.0001.

Figure 6.

BEBT-908 treatment induces a proinflammatory tumor microenvironment and potentiates ICB therapy. A, Immunofluorescene staining of CD45 in Ctrl. and BEBT-908–treated MC38 tumors on day 12 after inoculation. DAPI was used for nucleus staining. Scale bars, 50 μm. B, Quantitative data of immunofluorescence staining of CD45+ in Ctrl.- and BEBT-908–treated MC38 tumors 12 days after inoculation. Data were collected from four tumor samples and are presented as mean ± SEM. P values were determined by unpaired two-sided t test. ***, P < 0.001. C and D, Quantitative estimate of the numbers of CD45+ (C) and CD8+ (D) lymphocytes per mg of tumor tissue in Ctrl.- and BEBT-908–treated MC38 tumors as determined by flow cytometry. E, Ratio of CD8+ T/CD4+Foxp3+ Treg cells in Ctrl.- and BEBT-908–treated MC38 tumors. F and G, Average numbers of tumor-infiltrating GZMB+CD8+ T(F) and IFNγ+ CD8+ T (G) cells per mg of tumor tissue in Ctrl.- and BEBT-908–treated tumors. H and I, The qRT-PCR data showing the expression of GZMB (H) and IFNG (I) in Ctrl.- and BEBT-908–treated MC38 tumors on day 12 after inoculation. Data of CI are presented as mean ± SEM. P values were determined by unpaired two-sided t test; n = 8, 7 for Ctrl. and BEBT-908 group of C–G, respectively. n = 5 for H and I. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. J–L, Experimental protocol (J), tumor growth curve (K), and overall survival (L) of combination treatment of BEBT-908 (100 mg/kg, every other day for four times) and anti-PD1 antibody (clone RMP1–14, BioXcell, 100 μg/mice, two times) on MC38 tumors in syngeneic mice. n = 8 mice per group. Data in K are presented as mean ± SEM. P values were determined by two-way ANOVA (K) and log-rank test (L). *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. M and N, Tumor-free survivors of combined BEBT-908 and anti-PD1 antibody treatment were rechallenged with a lethal number of parental MC38 cells. M and N, Individual tumor growth (M) and overall survival (N) of host mice after rechallenge. Age-matched naïve C57BL/6 mice (WT) were used as control. n = 5 mice per group. P values were determined by unpaired two-sided t test on day 10 (M) and log-rank test (N). **, P < 0.01; ****, P < 0.0001.

Close modal

BEBT-908 treatment induced immunogenicity of tumor cells, and robust tumor infiltration of CTLs, suggesting a potentially synergistic benefit of combining BEBT-908 target therapy and immunotherapy for solid tumors. To evaluate whether BEBT-908 could synergize with ICB therapy, we carried out tumor growth experiments with a murine anti-PD1 immune checkpoint inhibitor in syngeneic mice inoculated with MC38 cells (Fig. 6J). While administration of BEBT-908 alone delayed tumor growth of MC38 tumors, its efficacy was enhanced significantly when combined with an anti-PD1 antibody, with long-term survival of some host mice (Fig. 6K and L). In fact, BEBT-908 combination with anti-PD1 antibody treatment resulted in durable cures in 5 of 8 mice (Supplementary Fig. S6E–S6H). Furthermore, we examined whether those mice remained tumor-free after initial tumor cell inoculation and treatment could resist a rechallenge with parental tumor cells (Fig. 6J). Lethal doses of wild-type tumor cells were injected into those tumor-free mice after the initial challenge. Our data showed that most of (four of five) cured survivors rejected the wild-type tumor cell rechallenge (Fig. 6M and N), indicating an antitumor immune memory. These data demonstrated that dual PI3K/HDACi BEBT-908 could enhance anti-PD1 antibody therapy, and sometimes induce complete tumor regression in a murine colorectal tumor model.

Abnormally activated PI3K and dysregulated HDAC are two well-recognized targets for cancer therapy. However, targeting either one of the pathways often has only modest efficacy. Here we report a dual PI3K/HDACi BEBT-908, that targeting PI3K and HDAC signal pathways simultaneously, potently inhibits tumor growth by promoting immunogenic ferroptotic cell death, and this immunogenic ferroptosis of cancer cells potentiates ICB therapy (Fig. 7).

Figure 7.

A schematic diagram illustrating BEBT-908–mediated ferroptosis and stimulation of proinflammatory pathways.

Figure 7.

A schematic diagram illustrating BEBT-908–mediated ferroptosis and stimulation of proinflammatory pathways.

Close modal

Combinatorial inhibition of critical genetic and epigenetic driver signaling pathways in cancer cells is in theory a promising strategy for cancer treatment. However, such combinations of HDACis and PI3K–AKT–mTOR pathway inhibitors are usually limited in clinical due to adverse toxicity caused by incompatible pharmacokinetic properties of individual agent (7, 37, 38). In the current study, the novel bifunctional PI3K/HDACi BEBT-908 displays synergistic anti-proliferation activities against multiple cancer types including hematologic malignancies and solid cancers with favorable pharmacokinetic properties in mouse, when compared with PI3Ki and HDACi reference compounds and their combinations.

Previously, preclinical development of targeted anticancer agents has largely hinged on human cell lines maintained in vitro or xenografts into severely immunodeficient hosts. In fact, abundant correlative data support that most targeted anticancer agents currently approved for use in patients with cancer generally mediated immunostimulatory effects (39, 40). Unexpected immune functions of the targeted drug in clinic suggest that the actual mechanism of action of the drug may be more complicated than initially anticipated. Therefore, the candidate agents in development should be evaluated in immunocompetent hosts at the preclinical stage to examine potential immunomodulatory mechanisms.

While pharmacologic inhibition of PI3K signaling or HDACs induce cell-cycle arrest, apoptotic cell death, and other pathways activation including JAK/STAT3, MAPK/ERK, NFκB in previous studies (41–44), and current study (Supplementary Fig. S3B and S3C), the dual PI3K/HDACi BEBT-908–mediated ferroptosis signaling activation is a critical mechanism for tumor killing. Moreover, the ferroptotic cells has been shown a panel of cancer cell–dependent immunostimulatory effects, including antigen presentation on class I MHC molecules, and secretion of proinflammatory cytokines. The evaluated MHC I expression on ferroptotic cancer cells suggests that BEBT-908 treatment makes cancer cells more susceptible to T cell–mediated killing. Furthermore, immunogenic ferropototic cell death and proinflammatory tumor microenvironment is conducive to immunotherapy. Indeed, BEBT-908 synergized with anti-PD-1 antibody in colorectal cancer model and resulted in durable cures over half of tested tumor-bearing mice. These data suggested a novel clinical strategy of combining anti-PD1 antibody with BEBT-908 for solid cancer treatment.

Taken together, our data provide strong evidence for the bifunctional PI3K/HDACi BEBT-908 to stimulate immunogenic ferroptosis of cancer cells. BEBT-908 might be an effective anticancer agent for patients with advanced hematologic malignancies and solid tumors either as a monotherapy or as a powerful immunostimulator to enhance ICB therapy.

X. Cai reports a patent for ZL2014104799422 issued. C. Qian reports personal fees and other support from Guangzhou BeBetter Medicine Technology Co., Ltd.; grants from National Science and Technology Major Project of China (2016ZX09101002), grants from Guangdong Zhujiang Talents Programme (2014ZT05Y232), and Guangzhou Municipal Science and Technology Project (201909020004) during the conduct of the study; in addition, C. Qian has a patent for ZL2010800109771 issued and licensed to Curis Inc. and a patent for ZL2014104799422 issued. X. Liu reports grants from Guangdong Basic and Applied Basic Research Foundation grant, Shenzhen Science and Technology Program grant, and Fundamental Research Funds for the Central Universities outside the submitted work. No disclosures were reported by the other authors.

F. Fan: Formal analysis, methodology, writing–original draft. P. Liu: Formal analysis, visualization, methodology, writing–original draft. R. Bao: Methodology. J. Chen: Investigation. M. Zhou: Investigation. Z. Mo: Investigation. Y. Ma: Investigation. H. Liu: Validation, visualization, methodology. Y. Zhou: Formal analysis. X. Cai: Conceptualization. C. Qian: Conceptualization, resources, supervision, funding acquisition. X. Liu: Conceptualization, resources, formal analysis, supervision, funding acquisition, writing–original draft, project administration, writing–review and editing.

This study is supported partially by National Science and Technology Major Project of China (2016ZX09101002 to C. Qian), Guangdong Pearl River Talents Plan (2014ZT05Y232 to C. Qian), Guangzhou Municipal Science and Technology Project (201909020004 to C. Qian), Guangdong Basic and Applied Basic Research Foundation grant (2020B1515020054 to X. Liu), Shenzhen Science and Technology Program grant (JCY20190807154813511 to X. Liu), Fundamental Research Funds for the Central Universities (19ykpy144 to X. Liu).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Yu
JX
,
Hodge
JP
,
Oliva
C
,
Neftelinov
ST
,
Hubbard-Lucey
VM
,
Tang
J
. 
Trends in clinical development for PD-1/PD-L1 inhibitors
.
Nat Rev Drug Discov
2020
;
19
:
163
4
.
2.
Tanaka
A
,
Nishikawa
H
,
Noguchi
S
,
Sugiyama
D
,
Morikawa
H
,
Takeuchi
Y
, et al
Tyrosine kinase inhibitor imatinib augments tumor immunity by depleting effector regulatory T cells
.
J Exp Med
2020
;
217
:
e20191009
.
3.
Canon
J
,
Rex
K
,
Saiki
AY
,
Mohr
C
,
Cooke
K
,
Bagal
D
, et al
The clinical KRAS(G12C) inhibitor AMG 510 drives anti-tumour immunity
.
Nature
2019
;
575
:
217
23
.
4.
Goel
S
,
DeCristo
MJ
,
Watt
AC
,
BrinJones
H
,
Sceneay
J
,
Li
BB
, et al
CDK4/6 inhibition triggers anti-tumour immunity
.
Nature
2017
;
548
:
471
5
.
5.
Finke
JH
,
Rini
B
,
Ireland
J
,
Rayman
P
,
Richmond
A
,
Golshayan
A
, et al
Sunitinib reverses type-1 immune suppression and decreases T-regulatory cells in renal cell carcinoma patients
.
Clin Cancer Res
2008
;
14
:
6674
82
.
6.
Motzer
RJ
,
Penkov
K
,
Haanen
J
,
Rini
B
,
Albiges
L
,
Campbell
MT
, et al
Avelumab plus axitinib versus sunitinib for advanced renal-cell carcinoma
.
N Engl J Med
2019
;
380
:
1103
15
.
7.
Fruman
DA
,
Chiu
H
,
Hopkins
BD
,
Bagrodia
S
,
Cantley
LC
,
Abraham
RT
. 
The PI3K pathway in human disease
.
Cell
2017
;
170
:
605
35
.
8.
Li
Y
,
Seto
E
. 
HDACs and HDAC inhibitors in cancer development and therapy
.
Cold Spring Harb Perspect Med
2016
;
6
:
a026831
.
9.
Anderson
GR
,
Wardell
SE
,
Cakir
M
,
Crawford
L
,
Leeds
JC
,
Nussbaum
DP
, et al
PIK3CA mutations enable targeting of a breast tumor dependency through mTOR-mediated MCL-1 translation
.
Sci Transl Med
2016
;
8
:
369ra175
.
10.
Martz
CA
,
Ottina
KA
,
Singleton
KR
,
Jasper
JS
,
Wardell
SE
,
Peraza-Penton
A
, et al
Systematic identification of signaling pathways with potential to confer anticancer drug resistance
.
Sci Signal
2014
;
7
:
ra121
.
11.
Morel
D
,
Jeffery
D
,
Aspeslagh
S
,
Almouzni
G
. 
Postel-Vinay S. Combining epigenetic drugs with other therapies for solid tumours - past lessons and future promise
.
Nat Rev Clin Oncol
2020
;
17
:
91
107
.
12.
Knight
ZA
,
Lin
H
,
Shokat
KM
. 
Targeting the cancer kinome through polypharmacology
.
Nat Rev Cancer
2010
;
10
:
130
7
.
13.
Zhang
K
,
Lai
F
,
Lin
S
,
Ji
M
,
Zhang
J
,
Zhang
Y
, et al
Design, Synthesis, and biological evaluation of 4-Methyl quinazoline derivatives as anticancer agents simultaneously targeting Phosphoinositide 3-kinases and histone deacetylases
.
J Med Chem
2019
;
62
:
6992
7014
.
14.
Chen
CS
,
Weng
SC
,
Tseng
PH
,
Lin
HP
,
Chen
CS
. 
Histone acetylation-independent effect of histone deacetylase inhibitors on Akt through the reshuffling of protein phosphatase 1 complexes
.
J Biol Chem
2005
;
280
:
38879
87
.
15.
Malone
CF
,
Emerson
C
,
Ingraham
R
,
Barbosa
W
,
Guerra
S
,
Yoon
H
, et al
mTOR and HDAC inhibitors converge on the TXNIP/Thioredoxin pathway to cause catastrophic oxidative stress and regression of RAS-driven tumors
.
Cancer Discov
2017
;
7
:
1450
63
.
16.
Pei
Y
,
Liu
KW
,
Wang
J
,
Garancher
A
,
Tao
R
,
Esparza
LA
, et al
HDAC and PI3K antagonists cooperate to inhibit growth of MYC-driven medulloblastoma
.
Cancer Cell
2016
;
29
:
311
23
.
17.
Liu
X
,
Bao
X
,
Hu
M
,
Chang
H
,
Jiao
M
,
Cheng
J
, et al
Inhibition of PCSK9 potentiates immune checkpoint therapy for cancer
.
Nature
2020
;
588
:
693
8
.
18.
Shu
J
,
Qiu
G
,
Ilyas
M
,
Kaye
P
. 
Biomarker detection in whole slide imaging based on statistical color models
.
MIDAS J
2010
.
19.
Dixon
SJ
,
Lemberg
KM
,
Lamprecht
MR
,
Skouta
R
,
Zaitsev
EM
,
Gleason
CE
, et al
Ferroptosis: an iron-dependent form of nonapoptotic cell death
.
Cell
2012
;
149
:
1060
72
.
20.
Stockwell
BR
,
Angeli
JPF
,
Bayir
H
,
Bush
AI
,
Conrad
M
,
Dixon
SJ
, et al
Ferroptosis: a regulated cell death nexus linking metabolism, redox biology, and disease
.
Cell
2017
;
171
:
273
85
.
21.
Dixon
SJ
,
Stockwell
BR
. 
The hallmarks of ferroptosis
.
Annu Rev Cancer Biol
2019
;
3
:
35
54
.
22.
Wang
SJ
,
Li
D
,
Ou
Y
,
Jiang
L
,
Chen
Y
,
Zhao
Y
, et al
Acetylation is crucial for p53-mediated ferroptosis and tumor suppression
.
Cell Rep
2016
;
17
:
366
73
.
23.
Jiang
L
,
Kon
N
,
Li
T
,
Wang
SJ
,
Su
T
,
Hibshoosh
H
, et al
Ferroptosis as a p53-mediated activity during tumour suppression
.
Nature
2015
;
520
:
57
62
.
24.
Li
T
,
Kon
N
,
Jiang
L
,
Tan
M
,
Ludwig
T
,
Zhao
Y
, et al
Tumor suppression in the absence of p53-mediated cell-cycle arrest, apoptosis, and senescence
.
Cell
2012
;
149
:
1269
83
.
25.
Yang
WS
,
SriRamaratnam
R
,
Welsch
ME
,
Shimada
K
,
Skouta
R
,
Viswanathan
VS
, et al
Regulation of ferroptotic cancer cell death by GPX4
.
Cell
2014
;
156
:
317
31
.
26.
Chen
D
,
Tavana
O
,
Chu
B
,
Erber
L
,
Chen
Y
,
Baer
R
, et al
NRF2 is a major target of ARF in p53-independent tumor suppression
.
Mol Cell
2017
;
68
:
224
32
.
27.
Hayes
JD
,
Dinkova-Kostova
AT
. 
The Nrf2 regulatory network provides an interface between redox and intermediary metabolism
.
Trends Biochem Sci
2014
;
39
:
199
218
.
28.
Wu
W
,
Bi
C
,
Credille
KM
,
Manro
JR
,
Peek
VL
,
Donoho
GP
, et al
Inhibition of tumor growth and metastasis in non-small cell lung cancer by LY2801653, an inhibitor of several oncokinases, including MET
.
Clin Cancer Res
2013
;
19
:
5699
710
.
29.
Topper
MJ
,
Vaz
M
,
Chiappinelli
KB
,
DeStefano Shields
CE
,
Niknafs
N
,
Yen
RC
, et al
Epigenetic therapy ties MYC Depletion to reversing immune evasion and treating lung cancer
.
Cell
2017
;
171
:
1284
300
.
30.
Hopewell
EL
,
Zhao
W
,
Fulp
WJ
,
Bronk
CC
,
Lopez
AS
,
Massengill
M
, et al
Lung tumor NF-kappaB signaling promotes T cell-mediated immune surveillance
.
J Clin Invest
2013
;
123
:
2509
22
.
31.
Guha
M
,
Mackman
N
. 
The phosphatidylinositol 3-kinase-Akt pathway limits lipopolysaccharide activation of signaling pathways and expression of inflammatory mediators in human monocytic cells
.
J Biol Chem
2002
;
277
:
32124
32
.
32.
Magner
WJ
,
Kazim
AL
,
Stewart
C
,
Romano
MA
,
Catalano
G
,
Grande
C
, et al
Activation of MHC class I, II, and CD40 gene expression by histone deacetylase inhibitors
.
J Immunol
2000
;
165
:
7017
24
.
33.
Kortenhorst
MS
,
Wissing
MD
,
Rodriguez
R
,
Kachhap
SK
,
Jans
JJ
,
Van der Groep
P
, et al
Analysis of the genomic response of human prostate cancer cells to histone deacetylase inhibitors
.
Epigenetics
2013
;
8
:
907
20
.
34.
Chandrasekaran
S
,
Sasaki
M
,
Scharer
CD
,
Kissick
HT
,
Patterson
DG
,
Magliocca
KR
, et al
Phosphoinositide 3-kinase signaling can modulate MHC class I and II expression
.
Mol Cancer Res
2019
;
17
:
2395
409
.
35.
Christova
R
,
Jones
T
,
Wu
PJ
,
Bolzer
A
,
Costa-Pereira
AP
,
Watling
D
, et al
P-STAT1 mediates higher-order chromatin remodelling of the human MHC in response to IFNgamma
.
J Cell Sci
2007
;
120
:
3262
70
.
36.
Wang
W
,
Green
M
,
Choi
JE
,
Gijon
M
,
Kennedy
PD
,
Johnson
JK
, et al
CD8(+) T cells regulate tumour ferroptosis during cancer immunotherapy
.
Nature
2019
;
569
:
270
4
.
37.
Qian
C
,
Lai
CJ
,
Bao
R
,
Wang
DG
,
Wang
J
,
Xu
GX
, et al
Cancer network disruption by a single molecule inhibitor targeting both histone deacetylase activity and phosphatidylinositol 3-kinase signaling
.
Clin Cancer Res
2012
;
18
:
4104
13
.
38.
Dreyling
M
,
Santoro
A
,
Mollica
L
,
Leppa
S
,
Follows
GA
,
Lenz
G
, et al
Phosphatidylinositol 3-kinase inhibition by copanlisib in relapsed or refractory indolent lymphoma
.
J Clin Oncol
2017
;
35
:
3898
905
.
39.
Mahoney
KM
,
Rennert
PD
,
Freeman
GJ
. 
Combination cancer immunotherapy and new immunomodulatory targets
.
Nat Rev Drug Discov
2015
;
14
:
561
84
.
40.
Karachaliou
N
,
Gonzalez-Cao
M
,
Sosa
A
,
Berenguer
J
,
Bracht
JWP
,
Ito
M
, et al
The combination of checkpoint immunotherapy and targeted therapy in cancer
.
Ann Transl Med
2017
;
5
:
388
.
41.
Will
M
,
Qin
AC
,
Toy
W
,
Yao
Z
,
Rodrik-Outmezguine
V
,
Schneider
C
, et al
Rapid induction of apoptosis by PI3K inhibitors is dependent upon their transient inhibition of RAS-ERK signaling
.
Cancer Discov
2014
;
4
:
334
47
.
42.
Shankar
S
,
Srivastava
RK
. 
Histone deacetylase inhibitors: mechanisms and clinical significance in cancer: HDAC inhibitor-induced apoptosis
.
Adv Exp Med Biol
2008
;
615
:
261
98
.
43.
Curigliano
G
,
Shah
RR
. 
Safety and tolerability of Phosphatidylinositol-3-Kinase (PI3K) inhibitors in oncology
.
Drug Saf
2019
;
42
:
247
62
.
44.
Kim
YK
,
Lee
EK
,
Kang
JK
,
Kim
JA
,
You
JS
,
Park
JH
, et al
Activation of NF-kappaB by HDAC inhibitor apicidin through Sp1-dependent de novo protein synthesis: its implication for resistance to apoptosis
.
Cell Death Differ
2006
;
13
:
2033
41
.