CPI‐203 improves the efficacy of anti‐PD‐1 therapy by inhibiting the induced PD‐L1 overexpression in liver cancer

Abstract Hepatocellular carcinoma (HCC) is one of the commonest lethal malignancies worldwide, and often diagnosed at an advanced stage, without any curative therapy. Immune checkpoint blockers targeting the programmed death receptor 1 (PD‐1) have shown impressive antitumor activity in patients with advanced‐stage HCC, while the response rate is only 30%. Inducible PD‐L1 overexpression may result in a lack of response to cancer immunotherapy, which is attributed to a mechanism of adaptive immune resistance. Our study investigated that the overexpression of PD‐L1 promoted the invasion and migration of liver cancer cells in vitro, and the induced overexpression of PD‐L1 in the tumor microenvironment could weaken the effects of anti‐PD‐1 immunotherapy in a BALB/c mouse model of liver cancer. CPI‐203, a small‐molecule bromodomain‐containing protein 4 (BRD4) inhibitor, which can potently inhibit PD‐L1 expression in vitro and in vivo, combined with PD‐1 antibody improved the response to immunotherapy in a liver cancer model. Cell transfection and chromatin immunoprecipitation assay manifested that BRD4 plays a key role in PD‐L1 expression; CPI‐203 can inhibit PD‐L1 expression by inhibiting the BRD4 occupation of the PD‐L1 promoter region. This study indicates a potential clinical immunotherapy method to reduce the incidence of clinical resistance to immunotherapy in patients with HCC.

response rates are generally low (20%-40%). [7][8][9] Combination therapy with the immune checkpoint blockers is being researched. [10][11][12][13][14][15] However, it is too early to conclude whether a meaningful clinical benefit is obtained in combination therapy, because almost 59% of patients have to discontinue this immunotherapy due to serious immune-related adverse events (irAEs). 16 The exhaustion of antigen-specific CD8 + T cells during persistent infections and cancer has been well documented, and PD-1 is a marker of CD8 + T-cell exhaustion. 17 IFNγ is a primary factor released by activated CD8 + T cells, and plays an important role in the host defense mechanism. 18 Current findings indicate that IFNγ has dual roles: one as a marker of antitumor immunity, the other as an inducer of immune escape through various mechanisms, such as PD-L1 expression, 19 which serves as a potent immune escape mechanism in cancer cells. [20][21][22] Bromodomain-containing protein 4 (BRD4) is an epigenetic reader of histones for acetylated lysine, 23 which is overexpressed in many solid tumors, and predicts malignant progression as well as poor patient prognosis. 24 Several BRD4 inhibitors have been developed, mainly for acute leukemia, lymphoma, and multiple myeloma.
But this has not been reported in patients with liver cancer. 25,26 Recent studies have revealed that BRD4 plays an important role in the regulation of PD-L1 expression in cancer, 27,28 so, in this study, we investigated the antitumor effects of a BRD4 inhibitor in combination with PD-1 antibody in HCC.
Human antibodies specific for BRD4, GAPDH, and β-actin were purchased from Cell Signaling Technology. Human anti-PD-L1 antibody was purchased from Abcam. Mouse antibody specific for PD-L1 as well as recombinant human IFNγ were purchased from Proteintech.

| Cell cytotoxic assay
HepG2 cells were treated with the specified compounds at final concentration gradients for the indicated durations. MTT (working concentration 0.5 mg/mL) was added, and the cells were incubated for 4 h at 37°C. The medium was removed and DMSO was added, and the plates were placed in a shaking bed for mixing at a low speed for 15 min. Then, absorbance was detected at 570 nm using a microplate reader.
When visible clones appeared, the medium was discarded, and the wells were washed twice with PBS. The HepG2 clones were fixed with 96% methanol for 20 min, stained with 1% crystal violet for 30 min, scanned, and photographed using an enzyme-linked immunodot analyzer.

| Transwell assay
The cells were washed twice with PBS and resuspended in serumfree medium for adding into the upper chamber. The lower chambers were filled with medium containing 10% FBS. After incubation at 37°C for 24 h, the cells located at the underside of the chamber were washed with PBS, fixed with methanol, and stained with 0.5% crystal violet. Cells were imaged by inverted microscope and calculated.

| RNA extraction and quantitative real-time PCR
Total mRNA was extracted using an RNA extraction kit. Reverse transcription was carried out using the PrimeScript RT Reagent Kit.
qRT-PCR was conducted on a 96 Real-Time System and analyzed on the LightCycler 480 System. The mRNA levels of genes were normalized to the GAPDH gene and calculated using the 2 -∆ ∆CT method.
The primer sequences for PD-L1, BRD4, and GAPDH are shown in Table S1.

| Western blotting
Cells were lysed using RIPA buffer, centrifuged at 12 000 rpm at 4°C for 15 min, and the supernatant was collected. Protein samples were separated by SDS-PAGE and transferred to a PVDF membranes.
After blocking, the membranes were probed with a primary and a secondary antibody. Chemiluminescence signals were detected using an ECL reagent kit, and a multifunctional imaging system (ProteinSimple) was used for image acquisition.

| Cell transfection
Small interfering RNAs (siRNAs) targeting BRD4 were designed by GenePharma, as shown in Table S2, and were transfected using the Lipo3000 transfection reagent according to the manufacturer's protocol; mRNA was collected for qRT-PCR analysis. HepG2 cells were infected with pcDNA3.1-HA-BRD4 plasmid or pcDH-PD-L1 plasmid, with sequence as shown in Table S3, using the Polyjet TM transfection reagent according to the manufacturer's instructions, and mRNA or protein was collected for qRT-PCR or western blot (WB) analysis.

| Chromatin immunoprecipitation (ChIP) assay
Cells were cross-linked and lysed; the chromatin was broken into 200-400-bp fragments using an ultrasonic cell disruptor.
Immunoprecipitation was performed with the BRD4 antibody at 4°C for at least 12 h, and then incubated with protein A agarose/salmon sperm DNA beads at 4°C for 1 h. The DNA was reverse-cross-linked and extracted using a DNA extraction kit, and qRT-PCR was performed to estimate DNA sequence levels using the PD-L1 promoter primer. The primer sequences used for the PD-L1 promoter are shown in Table S1.
The Renilla plasmid pRL-TK was used as an internal control. HepG2

| Mouse tumor implantation
Male BALB/c mice (6-8 wk old) were obtained from the Animal the mice were randomly assigned to control or experimental groups.

| ELISA assay
Serum samples were added to the 96-well plates, and incubated with antibody for 2 h. The HRP conjugate working solution was added, and samples incubated for another 45 min.
3,3′,5,5′-Tetramethylbenzidine (TMB) was added under dark conditions, and the terminating solution was added according to the experimental conditions. Dual-wavelength absorption values at 450 and 570 nm were determined using an enzyme plate analyzer and the concentration was calculated.

| Flow cytometry assay
Tumors were digested with PBS containing 2 mg/mL collagenase

| Immunohistochemistry and H&E staining
The sections were deparaffinized in toluene and rehydrated in a gradient series of ethanol. Endogenous peroxides were quenched with 0.3% H 2 O 2 /methanol. After blocking with goat serum, the slides were incubated with anti-PD-L1 antibody and appropriate secondary antibodies. Slides were visualized using a DAB kit, followed by counterstaining with hematoxylin. Brown membranous staining was considered to be positive for PD-L1 expression.
For H&E staining, the tissue sections were subjected to xylol deparaffinization, rehydration, and stained with hematoxylin solution. Here, 1% HCl solution was used for differentiation. The sections were then stained with eosin solution, dried overnight, and the results were observed using an inverted fluorescence microscope.

| Data presentation and statistical analysis
GraphPad Prism 5.0 was used for making charts. Statistical analysis was performed using SPSS 22.0 software. Parametric data are presented as mean ±standard error of the mean (SEM). We used the Student t test for two-group comparison or one-way ANOVA for multiple comparisons. P < .05 was considered significant: P-values were defined as *P < .05, **P < .01, and ***P < .001.

| PD-L1 overexpression in the tumor microenvironment leading to immune escape
BALB/c mice were used to examine the anti-PD-1 immunotherapeutic effects in liver cancer. PD-1 antibody (αPD-1) was administrated 3 times ( Figure S1A,B). The results showed no statistically significant difference in tumor volume and tumor weight between the αPD-1 group and the control group ( Figure S1C,D). ELISA was used to detected levels of IFNγ in the serum, and the result showed that it was significantly higher in the αPD-1 group than in the control group

| CPI-203 inhibited PD-L1 expression in the tumor microenvironment
Analysis of the Cancer Genome Atlas (TCGA) database using the online database UALCAN (http://ualcan.path.uab.edu) showed that BRD4 was highly expressed in liver cancer tissues ( Figure S2A). In addition, BRD4 expression in an hepatic cell line (L02) was lower than in hepatocarcinoma cell lines ( Figure S2B). JQ1 is the most recognized and most widely used BRD4 inhibitor, however its clinical application is limited by poor water solubility, short half-life, and relatively high toxicity. CPI-203 is a new optimized compound that is based on the JQ1 structure ( Figure S2C), and water solubility been improved, with extended half-life, and is more targeted. MTT assay was used to explore the proliferation toxicity of JQ1 and CPI-203 in the L02 line, and the results showed that CPI-203 is less toxic than JQ1  Figure S3B-D), was weaker than the initial cell lines ( Figure S3E-G). These results indicated that CPI-203 manifested an antitumor effect by inhibiting the expression of PD-L1, and it is a suitable candidate for combination immunotherapy.

| Significant antitumor effects of CPI-203 in combination with anti-PD-1 immunotherapy in liver cancer
The antitumor effect of CPI-203 combined with PD-1 antibody on liver cancer was evaluated in vivo. To fully demonstrate the therapeutic effect of the PD-1 antibody and CPI-203, the administration of the PD-1 antibody and CPI-203 was extended to 14 d, among them, PD-1 antibody was administrated 5 times ( Figure 3A).  Figure 3D). The necrosis of tumoral tissue was more obvious in the combination group than in the other groups ( Figure 3F). Also, the results showed that the expression of PD-L1 significantly increased after the PD-1 antibody was administered, however when proliferation was detected using MTT. B, Cells were seeded into six-well plates, and cultured in different concentrations of CPI-203 for about 2 wk. The clones were scanned using an enzyme-linked immunodot analyzer. C, CPI-203 was administrated in vivo, the tumor tissue single cells were stained by antibody, the percentage expression of PD-L1 was detected by FCM (*P < .05) of the treatment. We speculated that the initial failure to achieve an antitumor effect in the PD-1 antibody group might be attributable to the overexpression of PD-L1 in the tumor microenvironment.
With prolonged administration of PD-1 antibody, the antitumor effect was achieved only when sufficient PD-1 antibodies were aggregated in the tumor microenvironment to block the binding of PD-1and PD-L1. The problems that should be also noted are that with prolonged administration of PD-1 antibody, IFNγ was increased synchronically, also overexpression of PD-L1 was induced by IFNγ. Immunotherapy efficacy and resistance will coexist, and which is dominant may be affected by the immune status of different individuals. However, CPI-203 can block the induced expression of PD-L1 and exerted synergistic antitumor effects initially. In this way, not only could immunotherapy resistance be avoided, but also dose increase in PD-1 antibody could be avoided, and the incidence of immune-related adverse events (irAEs) would be decreased.
We also detected the proportion of CD4 + and CD8 + T lymphocytes in the tumoral tissues by flow cytometry (FCM). Results indicated that there was no change between the groups in the percentage of CD3 + CD4 + T and CD3 + CD8 + T lymphocytes in the tumor and spleen tissues ( Figure S4C,D). These suggested that the antitumor effect of the combination therapy was not dependent on the increase in cytotoxic T lymphocytes, but relied on the activation of cytotoxic T lymphocytes that originally existed in the tumor microenvironment.
IFNγ in the serum of all treatment groups was monitored using ELISA. The results showed that combination group did not significantly decrease or increase the serum IFNγ ( Figure S4B). This result indicated that the inhibitory effect of CPI-203 on PD-L1 expression was not dependent on the inhibition of IFNγ, it was a direct regulation of PD-L1 expression.

| Adverse events monitored for the CPI-203 plus anti-PD-1 therapy
Adverse events must be given equal attention when considering the improvement of therapeutic efficacy, especially the irAEs.
First, the body weight of the mice was monitored during treatment and showed no significant changes in either group ( Figure S4A).
The lung tissues H&E staining showed no obvious immune-related pneumonia in any of the therapy groups ( Figure 4A), Colon tissues H&E staining showed that some follicular lymphocyte aggregated in the basal area of intestinal mucosa in the PD-1 antibody group; while this was not observed in the combination treatment group ( Figure 4B). Simultaneously, we examined the biochemical indices of the liver and kidney using a biochemical analyzer and conducted routine blood examination to determine the bone marrow function.
All of these indicators showed no statistical differences among the groups ( Figure 4C,D). Therefore, combination therapy with CPI-203 and anti-PD-1 for liver cancer is clinically feasible. Next, the dual-luciferase reporter gene assay results showed that IFNγ increased the PD-L1 promoter activity, whereas CPI-203 significantly inhibited PD-L1 promoter activity in HepG2 cells, regardless of IFNγ stimulation ( Figure 5I). This can infer that CPI-203 has an inhibitory effect for PD-L1 at the PD-L1 promoter region.

| CPI-203 inhibited the BRD4 occupancy of the PD-L1 promoter region
siRNAs that targeted BRD4 were used to investigate the role of BRD4 in regulating PD-L1 expression in HepG2 (low expression of the PD-L1 gene) and HepG2.2.15 (relatively high expression of the PD-L1 gene) cell lines. When BRD4 was knocked down in HepG2 and F I G U R E 3 Antitumor effects of CPI-203 combined with PD-1 antibody in BALB/c mice. A, Schematic diagram of drug administration in BALB/c mice, there were 5 groups (n = 8): Isotype (IgG2a), CPI-203, αPD-1, and CPI-203 plus αPD-1 group. PD-1 antibody was administrated every 3 d for a total of 5 times; CPI-203 was administrated once daily, 14 times in total; IgG2a was the same as anti-PD-1. B, Photograph of the tumors in each group. C, Tumor volume was calculated using the formula and were analyzed. (*P < .05, **P < .01, ***P < .001). D, Tumor volume was calculated during the administration. E, Tumor single cells were stained with mouse anti-PD-L1-PE conjugated antibody, the percentage expression of PD-L1 was analyzed by FCM. F, H&E staining of the tumor tissues in different treatment groups HepG2.2.15 cell lines, the expression of PD-L1 at mRNA level was suppressed both with or without IFNγ stimulation ( Figure 6A-C). In addition, when BRD4 was overexpressed in HepG2 and HepG2.2.15 cell lines, the expression of PD-L1 at the mRNA level was increased ( Figure 6D,E). Therefore, we inferred that BRD4 plays an important role in the regulation of PD-L1 expression.
Next, ChIP-qPCR was used to detect the mechanism of CPI-203 on PD-L1 expression. The results showed that CPI-203 significantly

| D ISCUSS I ON
The liver is an important immune regulatory organ of the body, which adapts to the environment of antigen exposure through the development of innate immune tolerance. Continuous exposure to extrinsic antigen and active hepatitis B/C usually results in the depletion of T-cell function. Therefore, immunotherapy can be used for the treatment of liver cancer. However, only a few patients with HCC can benefit from clinical immunotherapy. To solve this problem, we selected PD-1 antibody to treat liver cancer in BALB/c mice in vivo. At first, no obvious antitumor effect of immunotherapy was observed when PD-1 antibody was administrated 3 times. Possibly, immunotherapy encountered primary or acquired resistance; a previous publication has reported that the establishment of resistance relevant to immunotherapeutic failure may predate immunotherapy. 31 Next, the study showed that IFNγ was increased significantly in the PD-1 antibody group, suggesting that PD-1 antibody dose pro-  10,20,40,80, or 160 ng/mL) for 24 h, then PD-L1 expression was detected using WB and qRT-PCR. C, D, HepG2 cells were treated with IFNγ (100 ng/mL) or IFNγ combined with BRD4 inhibitors JQ1 and CPI-203 (5 μmol/L) for 24 h and 48 h, then the PD-L1 expression was detected using WB and qRT-PCR (*P < .05). E, HepG2 cells were treated by IFNγ (100 ng/mL) with or without CPI-203 (5 μmol/L) for 48 h, then proteins were collected for WB to detect phosphorylation of STAT1. F, Expression of PD-L1 protein in HepG2 and MDA-MB-231 cells was detected by WB. G, HepG2 was treated with IFNγ (100 ng/mL) or IFNγ combined with CPI-203 (1 μmol/L, 2 μmol/L, 5 μmol/L) for 48 h. H, MDA-MB-231 cells were treated with CPI-203 (1 μmol/L, 2 μmol/L, 5 μmol/L) for 48 h. I, PD-L1 promoter activity was detected using double luciferase reporter gene assay in HepG2 cells, treated with IFNγ (100 ng/mL) or CPI-203 (5 μmol/L) for 48 h deserves further study in the future. 46 Our studies generally concluded that PD-L1 expression is closely related to the treatment of PD-1 immune checkpoint inhibitors, inhibiting the induced PD-L1 expression by combination therapy can improve the immunotherapy effect.

ACK N OWLED G M ENTS
This work was financially supported by the National Natural Science

CO N FLI C T O F I NTE R E S T
We have no conflicts of interest to declare.