Romidepsin (FK228) regulates the expression of the immune checkpoint ligand PD-L1 and suppresses cellular immune functions in colon cancer

The mouse colon cancer cell lines CT26 and MC38 were obtained from the ATCC (Manassas, VA, USA) and cultured in RPMI-1640 with 10% FBS. All cells were maintained under humidified condition (37 °C, 5% CO₂), and continual culture did not exceed 2 months. Romidepsin was purchased from Selleck Chemicals (Houston, TX, USA). An anti-PD-1 antibody was purchased from Biolegend (San Diego, CA, USA). Nontarget and BRD4-targeting siRNAs were purchased from RiboBio Co., Ltd. (Guangdong, China).

The following antibodies (Abs) were used for Western blotting: rabbit anti-pan-Akt mAb (#4691), anti-p-Akt (T308) mAb (#4056), anti-p-Akt (S473) mAb (#4060), anti-pan-ERK mAb (#4695), anti-p-ERK (T202/Y204) mAb (#4376), anti-BRD4 (#13,440), anti-H3(#4499), anti-H4(#13,919), anti-acetyl-H3(Lys9/Lys14) (#9677), anti-acetyl-H3 (Lys27) (#8173), anti-acetyl-H4 (Lys5) (#8647), anti-acetyl-H4 (Lys8) (#2594), anti-acetyl-H4 (Lys16) (#13,534) and anti-acetylated-lysine (#9441); all of these were purchased from Cell Signaling Technology (Danvers, MA, USA). Mouse anti-GAPDH mAb (G8795) was purchased from Sigma-Aldrich. Goat anti-PD-L1 (AF1019) was purchased from R&D Systems (Minneapolis, MN, USA).

Apoptosis and proliferation in CT26 and MC38 cells were analyzed using an Annexin V-FITC Apoptosis Detection Kit II (BD) and BrdU Cell Proliferation Assay Kit (BD), respectively. The results were measured using a FACS Canto II (BD). Representative results of independent assays are shown.

For the transient knockdown of BRD4 expression, CT26 and MC38 cells were seeded in 6-well plates and transfected with nontarget or BRD4-targeting siRNA pool using HiPerFect (Qiagen, German) according to the manufacturer's instructions. The working concentration of siRNAs was 100 nM.

RNA was extracted using TRIzol (Thermo Fisher Scientific), and cDNA was synthesized using a PrimeScript™ RT Reagent Kit (TaKaRa). The qPCR assays were performed using SYBR® Premix Ex Taq™ II (TaKaRa) and a QuantStudio™ 5 Real-Time PCR System (Thermo Fisher Scientific). GAPDH was used simultaneously as the internal control. One representative result of three independent assays is shown.

Cell lysis buffer (100 mM NaCl, 10 mM EDTA (pH 8.0), 50 mM Tris–Cl (pH 8.0) and 0.5% (v/v) Triton X-100) with EDTA-free complete protease and phosphatase inhibitors (Roche) was used for protein extraction. The lysates were separated on a 10% SDS-PAGE gel and transferred onto PVDF membranes. The targets were detected using an Amersham Imager 600 (GE). GAPDH was used as the loading control.

The CT26 and MC38 cells were treated with romidepsin for 24 h. Then, a Universal Magnetic Co-IP Kit (Active Motif) was used to harvest the acetylated proteins. Western blotting was used to detect the levels of histone 3 and histone 4 in the acetylated proteins.

This study was carried out in accordance with laboratory animal-guideline for Ethical review of animal Welfare and approved by Institutional Animal Care and Use Committee of Tianjin Medical University Cancer Institute & Hospital (approval No. 2014035). Six-week-old female BALB/C mice (Beijing Vital River Laboratory Animal Technology Co., Ltd) were acclimatized in the Laboratory Animal Resource Center at Tianjin Tumor Hospital. CT26 cells (1 × 10⁶) diluted in PBS were implanted subcutaneously (s.c.) in the flank regions of the mice. When the tumors became palpable, 32 mice were randomized into four experimental groups. Based on previous studies, the mice were given intraperitoneal (i.p.) injections of 300 μg per mouse anti-PD-1 (diluted in sterile PBS) on days 14, 16,18,20,22,24,26,28,30 post-tumor implantation and romidepsin (1 mg/kg dissolved in DMSO and diluted in sterile PBS) on days 15, 17,19,21,23,25,27,29. This schedule was repeated for a total of 30 days. The mice in the control group were treated with PBS. The mice were killed when the tumors reached a length ≥ 2 cm. Tumor volumes (mm³) were calculated with the following formula: tumor volume = 1/2 (length × width²), where the width and the length were the shortest and the longest diameters, respectively, as measured by calipers every 2 days. At the end of the study, the mice were killed. The tumors were weighed and prepared as lysates.

The following mouse Abs were used for flow cytometry analyses: anti-CD25 (101,908), anti-CD8 (100,706), anti-IFN-γ (113,604) and anti-PD-L1 (124,311); these antibodies were all purchased from Biolegend. Anti-CD4 (# 45–0042-82), anti-FOXP3 (#17–5773-82) and anti-IL-4 (# 25–7041-82) were purchased from Thermo Fisher. Gates were determined using isotype control Ab staining. Data were acquired with a FACS Canto II (BD) and expressed as the mean of five separate experiments.

The data were analyzed by two-tailed Student’s t test or ANOVA using GraphPad Prism software (La Jolla, CA). Comparisons between groups are presented as the mean ± SEM. Values of p < 0.05 were considered statistically significant.

The effects of romidepsin on the proliferation, cell cycle and apoptosis of the murine colon cancer cell lines were evaluated. CT26 and MC38 cells were treated with or without romidepsin for 24 h, then measured by Brdu cell cycle assay. We observed a dramatic decrease in the number of cells in the S + G2/M phase and a significant increase in the number of cells in the G0/G1 phase after romidepsin treatment (Fig. 1a). And the expression level of PCNA was also downregulated, which suggested the proliferation inhibition effects of romidepsin (Fig. 1b). Compared to the control cells, romidepsin-treated cells had a marginal increase in apoptotic events (Fig. 1c), which was validated by the increased caspase 3 cleavage (Fig. 1d). We further examined the responses of romidepsin in vivo using subcutaneous tumor-transplanted mice and CAC mice. Subcutaneous transplantation tumor mouse models were established by inoculating CT26 cells. After 16 days of treatment with romidepsin, tumor size and weight were obviously reduced compared with the control group (Fig. 2a). The expression of cleaved-caspase 3 in romidepsin-treated group was also upregulated (Fig. 2b).In the CAC model, colon cancer was induced by AOM and 3% DSS treatment for 70 days and then treated with romidepsin three times (Fig. 2c). The average number of tumors per mouse was used as criterion of efficacy, which was divided based on size (< 5 mm and ≥ 5 mm in the largest dimension). The average number of tumors ≥ 5 mm per mouse was lower in the romidepsin treatment group than in the control group, while there was no obvious difference in the number of tumors with sizes < 5 mm between the two groups (Fig. 2d). These results showed that romidepsin could inhibit colon cancer growth in vitro and in vivo by inhibiting cancer cell proliferation, inducing cell cycle arrest and promoting apoptosis.

Our study indicated that romidepsin arrested colon cancer growth directly. We further analyzed its effects on PD-L1 expression in cancer cells and immune responses. Following treatment with romidepsin, the qRT-PCR and Western blotting assay results showed that the cells experienced PD-L1 upregulation at both the mRNA and protein levels (Fig. 3a, b). These findings were further confirmed by evaluating the mean fluorescence intensity (MFI) of PD-L1 using flow cytometry (Fig. 3c). The ability of romidepsin to upregulate PD-L1 in vivo was also investigated using a subcutaneous transplantation tumor model and a CAC model. After treatment with romidepsin, PD-L1 expression in the tumors was assessed by qRT-PCR and Western blotting. Consistent with the in vitro data, romidepsin treatment could significantly increase the PD-L1 expression in tumors (Fig. 3d, e).

Because romidepsin enhanced PD-L1 expression in colon cancer cells, we examined the possible mechanisms by which romidepsin regulated PD-L1 expression. The histone acetylation state and expression levels of BRD4, a transcription factor for PD-L1, were analyzed before and after treatment with romidepsin. After treatment with romidepsin, the acetylation of histone 3 and histone 4 at different lysine sites was increased (Fig. 4a, b). In addition to the acetylation of histones 3 and 4, we detected the role of bromodomain-containing proteins (BRD4) in PD-L1 expression upregulation by romidepsin. BRD4 plays a critical role in regulating the transcription of PD-L1, whose binding sites are acetylated lysines 5 and 8 of histone H4 (H4K5ac/K8ac). Pools of three siRNAs (100 nM in total) were used to inhibit BRD4 expression in CT26 and MC38 cells (Fig. 4c). The effects of upregulating PD-L1 via romidepsin in the BRD4 knockdown group were much less than those in the control group (Fig. 4d), which indicated that BRD4 participated in regulating PD-L1 via romidepsin.

Because romidepsin increased the expression of PD-L1 in colon cancer cells, we next evaluated the influence of romidepsin on T cells, which was mediated by crosstalk between cancer cells and T cells through PD-L1/PD-1 signaling. Subcutaneous transplantation tumor models and CAC models were used to assess the effect. These mice were treated with romidepsin (R), anti-PD-1 antibody (P), combination treatment (R + P) or saline (control). The percentages of Foxp3 + Tregs in the blood, spleen, tumors and bone marrow were quantified by flow cytometry assays (Fig. 5a). Mice in the romidepsin group had a significantly higher percentage of Foxp3+ Tregs than the control group, and this effect was partially reversed when combined with the anti-PD-1 treatment (Fig. 5b). This phenomenon could be observed in subcutaneously transplanted tumor mice and in only the blood and tumors of CAC mice (Fig. 5b). Furthermore, to determine the role of romidepsin in Foxp3+ Treg regulation, we sorted CD4+ CD25+ T cells from the spleens of BABL/c mice by MACS. The expression of Foxp3 in CD4+ CD25+ T cells was significantly increased after treatment with romidepsin (Fig. 5c).

Intracellular T-helper 1/T-helper 2 (Th1/Th2) cytokines, including IFN-γ and IL-4, were detected by flow cytometric assays. In blood and tumors of subcutaneously transplanted tumor mice and CAC mice, the Th1/Th2 ratio (shown by IFN-γ/IL-4) was decreased after romidepsin treatment, which could be reversed when combined with anti-PD-1 treatment (Fig. 6a). IFN-γ secretion was examined to evaluate the activation of CD8+ T cells. Flow cytometry showed that romidepsin decreased the prevalence of CD8+ IFN-γ+ T cells. When combined with the anti-PD-1 antibody, the percentage of CD8+ IFN-γ+ cells were increased than romidepsin treatment alone (Fig. 6b).

The anti-tumor effects of romidepsin and anti-PD-1 antibody co-treatment were evaluated in subcutaneously transplanted tumor mice and in CAC mice. In subcutaneously transplanted tumor mice, the inhibitory effect of the co-treatment on the tumor growth was significantly increased compared with the treatment with romidepsin alone. And compared with anti-PD-1 antibody treatment, the tumor suppression effect of the combination therapy also showed an increasing trend (Fig. 7a). The incidences and multiplicity of colon neoplasms in CAC mice are shown in Fig. 7b. The average number of tumors ≥ 5 mm per mouse in the romidepsin combined with anti-PD-1 blockade treatment group tended to be less than that in the romidepsin treatment only group. In addition, compared with anti-PD-1 treatment, the group of combination also had an advantage in efficacy with no significant statistically difference.