Abstract
Background/Aim: Anti-programmed death-1 (PD-1)/PD-ligand 1 (PD-L1) antibody is a successful treatment for patients with solid cancers; however, there are several disadvantages that need to be resolved. Oral small molecule anti-PD-1/PD-L1 inhibitors have been developed and have good bioavailability. Materials and Methods: Potent anti-PD-1/PD-L1 inhibitor candidates from the Shizuoka small compound library were screened and investigated for their antitumor activities in vitro and in vivo using a humanized mouse model. A search for small compounds that inhibit PD-1/PD-L1 binding among 67,395 compounds through three rounds of screening procedures identified six compounds. Results: The two compounds (SCL-1 and SCL-2), which have as a key chemical structure of triazolopyridazin backbone with a piperazine residue on the aromatic ring and 1,3-diphenyl pyrazoline with hydrazinylphthalazine were selected based on in vitro assays and absorption, distribution, metabolism, and excretion (ADME) scoring and subjected to in vivo experiments using a humanized NOG mouse model. SCL-1 and SCL-2 exhibited moderate inhibitory activities against PD-1/PD-L1 binding compared to an anti-PD-1 antibody, with SCL-1 exerting markedly weaker cytotoxic effects on target cells than the other compounds. In in vivo experiments, SCL-1 exerted significant antitumor effects on PD-L1+ SCC-3 tumors, which were dependent on CD8+ T cell infiltration and PD-L1 expression in tumors. A pharmacokinetic study revealed that it has good bioavailability and distribution as an oral reagent. Conclusion: SCL-1 is a novel small compound that inhibits PD-1/PD-L1 binding and exerts potent antitumor effects. Thus, it has potential as an oral reagent for cancer immunotherapy.
- Small chemical compound library
- in silico screening
- ELISA for inhibition of PD-1/PD-L1 binding
- MHC-dKO NOG mouse
- ADME evaluation
Since the successful development of antibodies that block immune checkpoints, a breakthrough in cancer immunotherapy, the application of immune checkpoint blockade (ICB) therapy against programmed death-1 (PD-1)/PD-ligand 1 (PD-L1) molecules has extended to a number of solid cancers, including early-stage melanoma, advanced-stage non-small cell lung cancers, and breast cancers (1-4). However, there are several disadvantages that need to be resolved, such as limited clinical responses in 20-30% of patients with solid cancers, poor bioavailability and distribution because of resistance mechanisms, including autoantibodies against immune checkpoint antibody production, and autoimmune-based adverse effects (5).
The protein crystal structure of the PD-1/PD-L1 complex was recently resolved and revealed that this interaction was attributed to the large hydrophobic surfaces of the extracellular domains (6, 7). Zak et al. identified three hot spots on the PD-L1 surface (6), two of which function as drug-binding pockets. Based on the structure of the interface between PD-1 and PD-L1, researchers at Bristol-Myers Squibb (BMS) discovered a disclosed scaffold structure of BMS-202 binding to PD-L1, which consists of a substituted heterocycle with methyl biphenyl substructure (8). The biological activities of these derivatives, which are considered the first series of true small-molecule inhibitors of PD-1/PD-L1 binding, were verified using a homogeneous time-resolved fluorescence (HTRF) binding assay (9). Based on the structures of BMS-202 and BMS-8, the best lead compounds, BMS-1001 and BMS-1116 were developed to exhibit potent restorative activity for anti-CD3-mediated T cell activation (10, 11).
In recent years, several small-molecule drugs that target PD-1/PD-L1 binding have been identified, such as peptide-based molecules (heterocyclic peptidomimetic compounds) from Aurigene Ltd. (Karnataka, India) (12), D-peptide antagonists (13), macrocyclic peptides by BMS, and non-peptide small-molecule inhibitor derivatives from the BMS series (e.g., A22, PDI-1, and P-18) (14-16). Other small chemical molecules that inhibit the PD-1/PD-L1 interaction have been identified and used in phase I/II clinical trials (17-22). CA-170 not only binds to PD-L1, but also antagonizes VISTA or TIM3 binding (21). However, promising findings and structural information have not yet been reported; thus, the significance of these compounds has not yet reached the level of immune checkpoint antibodies. In contrast, INCB086550 was reported to exhibit antitumor effect in several microsatellite instability-high (MSI-H) solid tumor patients and the clinical trial is still on-going (22).
In the present study, we screened potent anti-PD-1/PD-L1 inhibitor candidates from the Shizuoka small compound library and investigated their antitumor activities in vitro and in vivo using a humanized mouse model. We identified two small-molecule compounds that target PD-1/PD-L1 protein interaction.
Materials and Methods
Cell lines and antibodies. The characteristics of the SCC-3 cell line (JCRB 0115; JCRB Cell Bank, Osaka, Japan) have been previously described (23). SCC-3-PD-L1 knockout (KO) subclones were prepared using the genome editing method and clone #11 was used for in vivo experiments. The human lymphoma cell line SCC-3 was subjected to CRISPR/Cas9-mediated KO of PD-L1 (sgRNA: ID#CRISPR1021720_SG, Thermo Fisher Scientific, Waltham, MA, USA). Briefly, the sgRNA and Cas9 complex were transduced into SCC-3 cells using the Neon™ Transfection System via electroporation (Invitrogen), and cells were seeded at a single-cell level on 96-well microplates using FACSAria (BD Biosciences, Franklin Lakes, NJ, USA). Proliferated clones were stained with anti-PD-L1 Ab (clone: 29E.2A3, BioLegend Inc., San Diego, CA, USA) and screened by flow cytometry. PD-L1-deleted clones were validated by Sanger sequencing using a DNA sequencer (3500 xL Genetic Analyzer, Applied Biosystems, Thermo Fisher Scientific).
The antibodies, used in the flow cytometric analysis of in vivo experiments, Anti-CD3-PerCP (HIT3a), anti-CD4-PE or anti-CD4-PE-Cy5 (RPA-T4), anti-CD8-PE-Cy5 (HIT8a), anti-CD45-FITC (2D1), anti-CD45RA-FITC (HI100), anti-CD45RO-APC (UCHL1), anti-CD56-PE (B159), anti-CD127-PE-Cy7 (A019D5), and anti-CXCR3-APC (1C6) were purchased from BD Pharmingen (San Jose, CA, USA). The anti-FoxP3-PE (hFOXY) antibody for human cell labeling was purchased from eBioscience, Inc. (San Diego, CA, USA). The anti-TIM3-PE (F38-2E2) antibody was purchased from Miltenyi Biotech (Bergisch, Gladbach, Germany). An anti-mouse CD45 antibody for labeling the mouse cells was purchased from BD Pharmingen. Anti-PD-1-APC (EH12.2H7), anti-CD39-PE-Cy7 (A1), and anti-CD62L-PerCP (DREG-56) antibodies were purchased from BioLegend. Splenocytes and peripheral blood cells were isolated using an ACK lysis buffer. Tumor-infiltrating lymphocytes (TILs) were separated from control or antibody-treated tumors using anti-human CD45-microbeads (Miltenyi Biotec) with the autoMACS system (Miltenyi Biotec). The staining method was previously described (23). Human cells were identified by gating the mouse CD45-PI- and human CD45+ fractions. Anti-CD4 and anti-CD8 antibodies for the in vivo CD4/CD8 antibody-blocking experiment were purchased from Bio X Cell (Lebanon, NH, USA).
In silico screening of the small compound library. We performed structure-based virtual screening to identify small molecules that inhibit the protein-protein interactions of PD-1/PD-L1. The crystal structure of the PD-1/PD-L1 complex (PDB ID: 4zqk) was used as the target molecule (6). The search regions of PD-1 and PD-L1 to be docked were set at the interface of the complex. The 3D structures of 67,395 in-house small molecules were generated using Balloon and Open Babel software (24). Molecular docking and scoring were performed using AutoDock Vina (25). In addition, the binding free energies of the three docking poses of each ligand were estimated separately using the MM/GBSA method. Based on computational screening targeting PD-1 and PD-L1, 1,674 compounds with lower scoring energies or binding free energies were selected for in vitro studies.
Establishment of enzyme-linked immunosorbent assay (ELISA) to screen inhibitory activities against PD-1/PD-L1 binding. An ELISA system that specifically screens inhibitory activity against PD-1/PD-L1 binding was prepared. Biotinylated human PD-1 extracellular domain (PD-1ex) and human PD-L1 extracellular domain recombinant proteins were produced using the Expi293F expression system (Thermo Fisher Scientific). A PD-1-tetramer probe was prepared by the step reaction of horseradish peroxidase-conjugated streptavidin (Thermo Fisher Scientific) to the biotinylated PD-1ex protein. After the incubation of the PD-1-tetramer probe with small compounds for 30 min, the complex was added to PD-L1-coated 96-well immobilizer amino plates (Nunc, Thermo Fisher Scientific), and then incubated for 1 h. TMB substrate (BD Biosciences) was added, and a specific color was developed. The inhibition rate was calculated using the following formula: percentage of inhibition=[optical density (OD) value with inhibitors – OD value of the background)]/[OD value with no inhibitors – OD value of the background] ×100. An inhibition rate >30% was considered positive. We evaluated the ELISA system using recombinant soluble PD-L1 protein and the anti-human PD-1 antibody nivolumab purchased from Selleck Chemicals (Houston, TX, USA) as positive controls.
Cell-based PD-1/PD-L1 blockade bioassay. A cell-based PD-1/PD-L1 blockade bioassay (Promega, Madison, WI, USA) was performed in triplicate by co-culturing PD-L1/T cell receptor (TCR) activator-expressing CHO cells and PD-1 expressing Jurkat cells for 6 h in the presence of varying concentrations of each inhibitor. The values obtained represent the TCR-mediated luminescence values of the activation of the NFAT pathway due to the inhibition of PD-1/PD-L1 binding at a dose of 50 μM of each inhibitor. The anti-PD-1 antibody (nivolumab) was used as a positive control.
Cell proliferation assay. A cell proliferation assay was performed using the WST-1 assay (Dojin Kagaku Corp., Kumamoto, Japan) as previously described (26). Briefly, 1×104 SCC-3 or anti-human CD3 antibody-stimulated Jurkat cells were seeded on a 96-well microculture plate (Corning Inc., Corning, NY, USA), and compounds at concentrations ranging between 0.25 and 100 μM were added. After four days, the WST-1 substrate was added to the culture, and the OD was measured at 450 and 620 nm using a multimode plate reader (Nivo, PerkinElmer Inc., Waltham, MA, USA).
In vivo experiments using humanized major histocompatibility complex (MHC)-double KO (dKO) NOG mice. MHC-dKO NOG mice were kindly provided by Dr. Mamoru Ito at the Central Institute for Experimental Animals (Kawasaki, Japan). All animals were cared for and treated according to the guidelines for the welfare and use of animals in cancer research, and the experimental procedures were approved by the Animal Care and Use Committee of Shizuoka Cancer Center Research Institute. The in vivo experiments using PBMCs derived from cancer patients and healthy volunteers were approved by the Institutional Review Board of Shizuoka Cancer Center, Shizuoka, Japan. All patients provided written informed consent.
The method for humanized MHC-dKO NOG mouse production was previously reported (23). Briefly, eight-week-old MHC-dKO NOG mice were irradiated with X-rays on day 0, and 1×107 human PBMCs from cancer patients were intravenously administered to each mouse via the tail vein. On day 14, small compounds, such as SCL-1 and SCL-2 (50 and 100 mg/kg), were orally administered 10 times over 14 days (days 1-5 and 8-12). The antitumor activities of the small compounds were evaluated by measuring the tumor volumes. Tumor volumes were calculated based on the National Cancer Institute formula: tumor volume (mm3)=length (mm) x [width (mm)]2 ×1/2. During organ harvest, mouse peripheral blood was collected from the retro-orbital venous plexus using heparinized pipettes. Three days after the last administration of the compounds, spleens and tumors were harvested from the control and small compound-treated groups. Blood and spleen cells from one set of three mice were used in in vitro assays, including flow cytometry and PCR analysis of cytokine and chemokine expression. Tumors from another set of three mice were primarily used for the tumor-infiltrating lymphocyte (TIL) analysis and immunohistochemistry (IHC).
In the CD4 and CD8 antibody blocking experiment, an anti-CD4 or anti-CD8 antibody (Bio X Cell, 2 mg/kg) was intraperitoneally injected into the humanized MHC-dKO NOG mice with SCC-3 tumors on day 4 and day 7, and the antitumor effects of the small compounds were evaluated. To investigate the influence of PD-L1 expression on the antitumor effects of the small compounds, humanized MHC-dKO NOG mice transplanted with PD-L1-KO SCC-3 tumors (clone #11) were used.
Pharmacokinetic study. Three BALB/cA nude mice transplanted with SCC-3 tumors and three BALB/cA mice were used in the in vivo pharmacokinetic study. After oral administration of SCL-1 (100 mg/kg) to tumor-bearing and regular BALB/cA mice, blood and tumor samples were collected at 5 min, 30 min, 1 h, 2 h, and 4 h. Plasma and tumor homogenates were treated with 50% methanol and 50% acetonitrile, and centrifuged at 3,000×g for 15 min. The supernatants were collected and subjected to LC-MS/MS (API4000, AB Sciex, Framingham, MA, USA).
Real-time PCR of cytokine/chemokine genes in small compound-treated SCC-3 tumors. Total RNA was isolated from peripheral blood cells, spleen cells and SCC-3 tumors using the NucleoSpin RNA kit (MACHEREY NAGEL, Düren, Germany). Real-time PCR analysis of cytokine/chemokine genes was performed using QuantStudio 12K (Thermo Fisher Scientific) as previously described (23). Expression of the following genes was performed: PD-L1, IL-2, IL-6, IL-10, tumor necrosis factor (TNF)-α, interferon (IFN)-γ, IL-12A, transforming growth factor (TGF)-β1, VEGF-A, colony-stimulating factor (CSF)-1, CCL3, CCL4, CCL5, CCL22, CXCL9, CXCL10, CXCL12, CXCR3, CCR5, CCR7, CD62L, PD-1, TIM3, CTLA4, TCF7, TOX, CD127, arginase1 (ARG1), CD11B, CD11C, CD39, CD73, CD83, GZMB, PRF1, and GAPDH.
Immunohistochemistry (IHC). Anti-CD4 (4B12) and anti-CD8 (C8/144B) antibodies (Thermo Fisher Scientific), anti-FoxP3 antibody (236A/E7, Abcam, Cambridge, MA, USA), and anti-CD204 antibody (SRA-C6, TransGenic Inc., Kobe, Japan) were used for IHC. Positive cell numbers in ten fields of view of sections of each tumor at a high magnification (×200) stained with various antibodies were calculated using the image-analyzing software, WinRoof (Mitani Corporation, Tokyo, Japan).
Statistical analysis. The significance of differences was analyzed using Student’s t-test and Mann–Whitney U-test. p-Values<0.05 were considered significant. Absorption, distribution, metabolism, and excretion (ADME) score analysis was performed using StarDrop software (ver. 7.1, Optibrium Ltd., Cambridge, UK).
Results
Selection of small chemical compounds using a virtual PD-1/PD-L1 docking algorithm. The binding free energies of the three docking simulations for each ligand were calculated. Based on computational screening, 1,674 compounds showing lower scoring energies or binding free energies were selected for in vitro studies (Figure 1).
Schema of screening procedures for the Shizuoka small molecule library. After 3-rounds of screening of 67,395 compounds, six compounds were ultimately identified. Based on ADME scores, two compounds were selected and used in in vivo experiments.
ELISA for screening the inhibitory activities against PD-1/PD-L1 binding. ELISA to screen the inhibitory activities of small compounds against PD-1/PD-L1 binding was established as the 2nd screening method using the PD-1 tetramer probe (Figure 2A and B). The PD-L1 soluble protein and anti-PD-1 antibody as positive controls exhibited significant dose-dependent inhibitory activities (Figure 2C). Additionally, anti-PD-1 antibody showed a dose-dependent recovery of TCR-mediated NFAT pathway activation due to the inhibition of PD-1/PD-L1 binding in cell-based PD-1/PD-L1 blockade bioassay (Figure 2D). The inhibitory activity of 1,674 small compounds selected was assayed thrice using ELISA; in comparison with the control without inhibitors, 26 compounds exhibited an inhibition rate >30% in two out of the three assays (Figure 3). Six compounds exhibited inhibition rates >30% in all three assays (Figure 3, Table I).
Establishment of ELISA for evaluating the inhibitory activities against PD-1/PD-L1 protein binding. (A) The structure of the PD-1-tetramer probe. (B) The reaction mechanism of ELISA for detecting the inhibitory activities of small compounds against PD-1/PD-L1 binding. (C) The results of ELISA for the soluble PD-L1 protein and anti-PD-1 antibody as the positive control. (D) Dose-response data of anti-PD-1 antibody in the cell-based TCR signal recovery assay. The inhibitory activity of each compound was measured in triplicate in 96-well microculture plates. After the first round of screening, 1674 compounds identified were evaluated thrice by ELISA.
Inhibitory activities of 26 small compounds against PD-1/PD-L1 binding selected from three-times ELISA. The 1,674 small compounds selected were subjected to ELISA three times and compared to the control value without inhibitors. An inhibition rate >30% was considered positive. Twenty-six compounds exhibited an inhibition rate >30% in two out of three assays. Particularly, six compounds (SCL-1~SCL-6) exhibited inhibition rates >30% in all three assays. Each column shows the mean±SD value from three-times ELISA. C: the control value without inhibitors. *p<0.05, **p<0.01.
Various assay data from six selected compounds.
Other immunological assays. The six compounds selected were subjected to other immunological assays, such as cell-based PD-1/PD-L1 blockade and cell proliferation assays. In the cell-based T cell receptor-signal recovery assay, SCL-1 showed a moderate positive value, while other compounds showed negative values at a dose of 50 μM. In the cell proliferation assay, SCL-1, 5, and 6 did not exhibit direct cytotoxicity. The other compounds exhibited moderate cytotoxicity. In contrast, the compounds from the BMS-series exhibited strong cytotoxicity (Table I).
ADME score analysis. ADME score analysis of the six compounds identified after the 3rd screening was performed using StarDrop software (27). Based on the results from individual categories, the six compounds were ranked in a higher score order as follows (Table II): SCL-1 > SCL-5 > SCL-3 > SCL-6 > SCL-4 > SCL-2. Based on their availability in terms of chemical synthesis, SCL-1 and SCL-2 were synthesized and used in vivo.
Absorption. distribution. metabolism, and excretion (ADME) score analysis of selected 6 componds with PD-1/PD-L1 binding inhibiting activity using StarDrop software.
Structure of representative chemical small compounds with common structural formula. The chemical structures of SCL-1 and SCL-2 compounds are shown in Figure 4. They have the chemical structures of (2-methylpiperidin-1-yl)-[1-[3-(trifluoromethyl)-[1,2,4]triazolo[4,3-b]pyridazin-6-yl] piperidin-4-yl]methanone and N-[(Z)-[4-(5-phenyl-3,4-dihydro pyrazol-2-yl)phenyl]methylideneamino]phthalazin-1-amine. The key structure SCL-1 for inhibition in the ELISA assay was a triazolopyridazin backbone with piperazine residue on the aromatic ring (Figure 4A and B), because SCL-1 and related analogs showed positive activity in ELISA assay. In particular, SCL-1 with 2-methyl piperazine linked to an amido bond possessed strong antitumor activity in vivo.
Chemical structures of selected small molecule compounds (SCL-1 and SCL-2). (A) Formula structure of specific series of p SCL-1 compound derivatives. (B) The structures of SCL-1 and SCL-2 small molecule compounds. (C) Computationally predicted binding mode of SCL-1 with PD-1. The PD-1 amino acids associated with the docking between SCL-1 and PD-1/PD-L1 interface are indicated on the panel. Our virtual screening was based on the X-ray crystal structure of human PD-1/PD-L1 complex (PDB ID: 4ZQK12).
On the other hand, SCL-2 also showed inhibitory activity on the ELISA assay and formed hydrazone core linkage between 1,3 diphenyl pyrazoline and 1-hydrazinylphthalazine. The important core structure is the 1,3-diphenylpyrazoline backbone, based on the assay results of SCL-2 analogs. The substituent of phthalazine residue was not so effective in increasing the inhibitory activity.
In addition, the computationally predicted binding mode of SCL-1 with PD-1 is shown in Figure 4C. Molecular docking results suggested that SCL-1 might bind near the PD-1/PD-L1 interface and form hydrogen bonds with Ser62, Asn66, and Lys78 of PD-1.
In vivo experiments using humanized MHC-dKO NOG mice. The following in vivo experiments were performed. The antitumor effects of SCL-1 (50 mg/kg) and SCL-2 (100 mg/kg) were investigated in humanized SCC-3 tumor-bearing MHC-dKO NOG mice (Figure 5A). The effects of anti-CD4/CD8 antibody blocking in vivo were examined in humanized mice (Figure 5B). The antitumor effects of SCL-1 on PD-L1 KO SCC-3 tumor (clone #11)-bearing humanized mice were assessed (Figure 5C). SCL-1 exerted stronger antitumor effects on SCC-3 tumors than SCL-2 (Figure 6A). SCL-1 promoted the engraftment of transplanted human PBMCs in the spleens and increased the number of infiltrated CD45+ human PBMCs and CD4+ and CD8+ T cells inside the tumors by 5-, 9-, and 4-fold per gram of tumor, respectively (Table III and Table IV).
In vivo experimental protocols using humanized MHC-double knockout NOG mice. (A) The administration of SCL-1 or SCL-2 to humanized mice. (B) The administration of SCL-1 to humanized mice treated with anti-CD4/CD8 antibody blockade. (C) The administration of SCL-1 to humanized mice bearing human PD-L1 gene-knockout (KO) SCC-3 tumors. MHC-double KO NOG mice (N=6 from each group) were irradiated with X-rays and human PBMCs were intravenously administered on day 0. After the transplantation of tumors on day 1, small compounds were orally administered 10 times over 14 days. Three days after the last administration of compounds, spleens, tumors, and blood were harvested from three mice in each group.
Effects of SCL-1 on SCC-3 tumor-bearing humanized mice. (A) Antitumor effects of SCL-1 or SCL-2 compound on SCC-3 tumors. The ratio of tumor/control volumes and body weight of mice are shown. Circle in green: SCL-1, circle in red: SCL-2. *p<0.05. Each point shows the mean value of 6 mice. (B) Effects of T cell depletion on the antitumor effects of SCL-1. Anti-CD4 or anti-CD8 mAb (Bio X Cell) was intraperitoneally injected at 100 μg/body/day into mice from days 7 to 9. Tumor volumes and the ratio of tumor/control volumes are shown. Circle in blue: control, circle in red: SCL-1, circle in green: anti-CD4 antibody-treated, circle in purple: anti-CD8 antibody-treated. *p<0.05. Each point shows the mean value of 6 mice. (C) Effects of PD-L1 gene expression on the antitumor effects of SCL-1. PD-L1 gene KO SCC-3 tumors instead of regular SCC-3 tumors were transplanted into humanized mice on day 1. In the tumor volume panel, circle in blue: parental SCC-3 tumor-control, circle in red: parental SCC-3 tumor-SCL-1 treated, circle in green: PD-L1 KO SCC-3 tumor-control, circle in purple: PD-L1 KO SCC-3 tumor-SCL-1 treated. In the T/C panel, circle in red: PD-L1 KO SCC-3 tumor-SCL-1 treated, circle in blue: parental SCC-3 tumor-SCL-1 treated. *p<0.05. Each point shows the mean value of 6 mice.
Flow cytometry analysis of tumor from control or compounds-treated mice bearing SCC-3 tumors.
Flow cytometry analysis of spleen from control or compounds-treated mice bearing SCC-3 tumors.
Anti-CD8 blocking markedly suppressed the antitumor effects of SCL-1 (Figure 6B) by significantly reducing the number of tumor-infiltrating CD8+ T cells (data not shown).
In the PD-L1-deleted SCC-3 tumor (clone #11)-bearing mice, SCL-1 did not induce antitumor responses (Figure 6C), which revealed that its antitumor effects were dependent on PD-L1 expression in the tumor. Parental SCC-3 cells showed a high level of PD-L1 expression by flow cytometry, but PD-L1-deleted SCC-3 cells (clone #11) showed no PD-L1 expression (data not shown).
IHC findings. More foci of lymphoid cell infiltration with necrotic changes were observed in the SCL-1-treated group than in the control group (Figure 7A). CD8+ T cells were more frequently identified in the SCL-1-treated group; however, FoxP3+ lymphocytes and CD204+ immune cell numbers were not significantly different (Figure 7B).
Effects of SCL-1 on immune cell infiltration in SCC-3 tumors. (A) Images of control and SCL-1-treated tumors stained with hematoxylin and eosin (H-E), and the anti-CD8 and anti-FoxP3 antibodies. Magnification: ×200 for H-E, ×200 for CD8 and FoxP3 staining. (B) Effects of SCL-1 on the number of infiltrating immune cells in SCC-3 tumors. Ten fields of view in sections of each tumor at a high magnification (×200) were selected and evaluated using image-analysis software (WinRoof). The positive cell count per field was compared between the control and SCL-1-treated groups. **p<0.01, significant difference.
Real-time PCR of cytokine/chemokine genes in small compound-treated SCC-3 tumors. In SCL-1-treated SCC-3 tumors, the expression of CXCL9 and CXCR3, which are effector T cell-attracting chemokines, was slightly upregulated. In contrast, the expression levels of CCL22 and TIM3, regulatory T cell-enhancing genes, were slightly downregulated (Figure 8). No significant shift in T cell-associated cytokine gene profiling was observed, which was partly attributed to the timing of harvesting already being past the peak of the immune (cytokine) response induced by SCL-1.
Cytokine/chemokine gene expression profiling in humanized mice treated with SCL-1. Expression of various genes in SCC-3 tumors (A) and PBMCs (B) from SCL-1-treted mice. The expression levels in the control group were rated as 1.0. Column in blue: the control group, column in orange: the SCL-1-treated group. Each column represents the mean value of triplicate experiments.
Pharmacokinetic study. In regular BALB/cA mice, the levels of SLC-1 in blood reached a maximum (4.63 μg/ml) at 30 min after its administration and T1/2 was 1.75 h. On the other hand, in SCC-3 tumor-bearing BALB/cA nude mice, maximum plasma concentration (Cmax) values in blood and tumors were 8.23 μg/ml (equivalent to 20 μM) and 4.63 μg/ml (equivalent to 10 μM), respectively. The T1/2 was shorter in tumor-bearing mice than in regular BALB/cA mice (Figure 9).
Pharmacokinetic profile in tumor-bearing nude mice and Balb/cA mice treated with SCL-1. Following the oral administration of SCL-1 (100 mg/kg) to tumor-bearing and regular BALB/cA mice, blood and tumor samples were collected after 5 min, 30 min, 1 h, 2 h, 4 h, 8 h, and 24 h. (A) Standard-curve of SCL-1 obtained from the measurement of the peak area by LC-MS. (B) Plasma levels of SCL-1 in Balb/c mice following SCL-1 oral administration (N=3). (C) Summary of plasma and intratumoral levels of SCL-1 in SCC-3 tumor-bearing Balb/cA nude mice (N=3) or regular Balb/c mice treated with SCL-1 oral administration (N=3).
Discussion
Since the successful application of ICB therapy in the clinical oncology field in 2012, extensive efforts have been dedicated to the development of small compound-based novel inhibitors of PD-1/PD-L1 binding, such as the biphenyl skeleton compound by BMS (8) and peptide-based small compounds by Origene (12, 13). However, except for drugs in ongoing clinical trials, promising small compounds as drug candidates with PD-1/PD-L1 antagonistic activities have yet to be identified. Small deleted compounds have some advantages over existing anti-immune checkpoint antibodies: 1) better bioavailability with oral administration, 2) superior in vivo distribution and metabolism in tumors. Therefore, continuous efforts are being made to identify small compounds with potent antitumor activities through immunological mechanisms.
Based on the recent history of drug development, most studies have focused on the synthesis of original BMS compounds, such as BMS-202 and 1166-associated derivatives, and an in silico docking model based on the PD-1/PD-L1 molecule docking structure (9, 10). Therefore, traditional and regular in silico screening methods for small compound libraries have been reported in a few studies. Patil et al. performed in silico screening of small compounds using the Autodock Vina algorithm and NCI database molecules based on the docking of NCI compounds to the PD-L1-binding site on PD-1 (28). Their findings demonstrated that the top four compounds significantly inhibited PD-1/PD-L1 binding, with the most active compound achieving an inhibition rate of >50% at a dose of 25 μM. However, data from in vitro and in vivo assays have not been reported.
In the present study, we identified six small compounds and used two molecules for in vivo experiments. The selected compounds were classified according to Lipinski’s rule of five for the bioavailability index. These compounds bound to PD-1 and/or PD-L1 in in vitro assay and were visualized by computer simulation; specifically, SCL-1 presented chemical stability in blood according to LCMS analysis and no cytotoxicity in a proliferation assay. On the other hand, SCL-2 did not show sufficient blood levels in a pharmacokinetic study (data not shown), which may suggest that the degradation of SCL-2 by the hydrolysis of hydrazone and/or the binding to plasma proteins resulted in poor pharmacokinetic results. In general, the hydrazone structure is usually classified as PAINS (29), an unfavorable structure for drug discovery and is easily decomposed by nucleophiles to form covalent bonds or alkaline condition. This result also supported the ADME prediction of plasma protein binding (Table II) using StarDrop (Optibrium, Cambridge, UK) (30).
Based on the ADME score and pharmacokinetic analyses (Figure 9 and Table II), SCL-1 has potential as an oral small-compound drug (good metabolism, distribution, and stability in serum) because Cmax levels in the plasma and tumors of tumor-bearing nude mice were 20 and 10 μM, respectively, which were estimated to be sufficient for biological activity. Furthermore, we confirmed that SCL-1 exerted weak activity in the ELISA for PD-1/PD-L1 binding inhibition and cell-based PD-1/PD-L1 blockade bioassay (TCR signal recovery) but showed no direct cytotoxicity in vitro and strong antitumor effects in in vivo experiments. Considering the differences in activity profiling from other reported small compounds, such as BMS-series-associated derivatives, the following needs to be confirmed: 1) no obvious direct cytotoxicity against target cell lines, 2) strong inhibitory activity against PD-1/PD-L1 binding in ELISA and cell-based assays, which are not critical factors, and 3) in vivo antitumor activity and safety based on immunological mechanisms. We investigated whether the antitumor effects in vivo involved immunological mechanisms using PD-L1 KO SCC-3 tumors and anti-CD4/CD8 antibody blockade. The results revealed that PD-L1 KO and anti-CD4/CD8 blockade significantly suppressed the antitumor effects of SCL-1; therefore, immunological mechanisms were clearly involved in tumor responses in the in vivo model.
Two issues need to be considered regarding the biological activity of SCL-1: the off-target effects of SCL-1 on specific kinase activity only moderately inhibited c-Met kinase (data not shown), and SCL-1 exerted inhibitory effects on the expression of other checkpoint molecules (Tim3 and CD39) on T cells in addition to promoting effector T cell infiltration into tumors (Table III).
In the case of TIM3 gene expression in SCL-1-treated SCC-3 tumors, the TIL marker analysis by FACS and real-time PCR both showed lower expression levels than in control tumors. In our previous study (31), we found that TIM3 functions as an immunosuppressive immune checkpoint receptor associated with a poor prognosis in cancer patients. TIM3 has recently been identified as an important factor for enhancing regulatory T cell function (32). In the present study, the downregulation of TIM3 is an important observation in terms of its dual functions on regulatory T cells as an enhancing factor and on effector T cells as an exhaustive marker; therefore, the inhibitory activity of SCL-1 on TIM3 warrants further intensive study.
CD39 converts ATP to AMP, and CD73 converts AMP to adenosines, which can induce together immunosuppressive effects in regulatory T cells (33, 34). These findings suggest that SCL-1 might have modulatory effects on suppressive immune checkpoint molecules expressed on regulatory T cells and trigger antitumor activity through an immune-metabolic mechanism. Importantly, Sade-Feldman et al. reported that the combination of a TIM-3 blocking antibody with an anti-CD39 small-molecule inhibitor significantly reduced tumor size in a B16F10 mouse melanoma model (35).
In the present study, we identified a potent inhibitor of PD-1/PD-L1 binding through three rounds of screening of a small compound library. Based on an investigation of structure-activity relationships using crystal structure analysis, more lead compounds will be discovered in the near future.
Acknowledgements
The Authors thank Dr. Koji Maruyama for excellent assistance in maintaining the NOG-dKO mice in the animal facility. This work was supported by a grant to Tadashi Ashizawa from JSPS KAKENHI (grant no. 18K07315), Japan.
Footnotes
Authors’ Contributions
YA participated in the design of the study and drafting of the manuscript and was responsible for completing the study. AI, HM, CM and AK performed immunological in vitro experiments. TA performed in vivo experiments. Takayuki Ando contributed to the synthesis and supply of small compounds. YI performed in silico-based screening of small compounds and developed a docking program of PD-1/PD-L1. RK performed the statistical analysis. TS contributed to the preparation of pathological specimens and performed pathological examination. KY reviewed the manuscript. All Authors read and approved the final draft.
Conflicts of Interest
The Authors declare that they have no conflicts of interest.
- Received August 10, 2022.
- Revision received September 16, 2022.
- Accepted September 22, 2022.
- Copyright © 2022 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
References
In this issue
Jump to section
Related Articles
Cited By...
- In Vivo Antitumor Activity of the PD-1/PD-L1 Inhibitor SCL-1 in Various Mouse Tumor Models
- MHC class I and II-deficient humanized mice are suitable tools to test the long-term antitumor efficacy of immune checkpoint inhibitors and T-cell engagers
- SOX10 Inhibits T Cell Recognition by Inducing Expression of the Immune Checkpoint Molecule PD-L1 in A375 Melanoma Cells