Antitumor Effect of Korean Red Ginseng through Blockade of PD-1/PD-L1 Interaction in a Humanized PD-L1 Knock-In MC38 Cancer Mouse Model

Blocking immune checkpoints, programmed death-1 (PD-1) and its ligand PD-L1, has proven a promising anticancer strategy for enhancing cytotoxic T cell activity. Although we previously demonstrated that ginsenoside Rg3, Rh2, and compound K block the interaction of PD-1 and PD-L1, the antitumor effect through blockade of this interaction by Korean Red Ginseng alone is unknown. Therefore, we determined the effects of Korean Red Ginseng extract (RGE) on the PD-1/PD-L1 interaction and its antitumor effects using a humanized PD-1/PD-L1-expressing colorectal cancer (CRC) mouse model. RGE significantly blocked the interaction between human PD-1 and PD-L1 in a competitive ELISA. The CD8+ T cell-mediated tumor cell killing effect of RGE was evaluated using murine hPD-L1-expressing MC38 cells and tumor-infiltrating hPD-1-expressing CD8+ T cells isolated from hPD-L1 MC38 tumor-bearing hPD-1 mice. RGE also reduced the survival of hPD-L1 MC38 cells in a cell co-culture system using tumor-infiltrating CD8+ T cells as effector cells combined with hPD-L1 MC38 target cells. RGE or Keytruda (positive control) treatment markedly suppressed the growth of hPD-L1 MC38 allograft tumors, increased CD8+ T cell infiltration into tumors, and enhanced the production of Granzyme B. RGE exhibits anticancer effects through the PD-1/PD-L1 blockade, which warrants its further development as an immunotherapy.


Introduction
During cancer progression, tumors evolve and develop a variety of mechanisms to evade immune surveillance and suppress the antitumor immune response. The basic mechanism of tumor immune evasion involves immune checkpoint pathways. Immune checkpoints have been considered antitumor therapeutic targets as they regulate immune evasion in cancer cells [1].
Immune checkpoint inhibitors (ICIs) have appeared as innovative and effective immunotherapies for cancer. ICIs assist the patient's immune system in mounting an antitumor response. Among the various ICIs, programmed cell death (PD-1)/programmed cell death ligand 1 (PD-L1) blockers exhibit remarkable therapeutic effects against various cancers [2,3]. Currently, the Food and Drug Administration (FDA) has approved the PD-1/PD-L1 blockers atezolizumab, avelumab, cemiplimab, durvalumab, nivolumab, and pembrolizumab for cancer treatment [4]. However, side effects related to resistance and immunity have been reported, and treatment is expensive. Therefore, the development of an effective and safe PD-1/PD-L1 blocker is required, and many studies are currently underway [5].
Korean Red Ginseng (KRG) is processed by steaming or drying ginseng, and contains a high concentration (60-70%) of carbohydrates including starch. In addition, it contains ginsenosides, which are various physiologically active substances that do not exist in other plants, nitrogen-containing compounds, such as proteins, peptides, and alkaloids, and fat-soluble polyacetylenes, polysaccharides, and flavonoids [6]. In KRG, there are changes in the type and concentration of ginsenoside, a compound unique to ginseng, in the process of processing red ginseng, and there are physicochemical changes in polysaccharide, the most abundant compound in Korean red ginseng. In particular, it has a low moisture content of less than 15.5% and is widely used as a health functional food because it can be stored for a long time by gelatinizing starch.
KRG is a traditional herbal medicine that has been used in East Asia to treat various ailments, and is commonly used as an herbal medicine to treat allergies and inflammatory and cancer diseases [7]. Numerous studies have reported that KRG exhibits various biological activities, such as immunity enhancement [8][9][10], antiinflammation [11,12], antifatigue [13][14][15][16], antiobesity [17,18], antioxidant [19,20], menopausal relief [19,21], and anticancer effects [10,[22][23][24]. Moreover, KRG inhibited endocrine disrupting chemicalinduced endometriosis in a mouse model [25] and induced both extrinsic and intrinsic apoptotic pathways in breast cancer cells and non-malignant breast cells [26]. It has been reported that the non-saponin fraction of KRG regulates immune homeostasis in old mice [27], improves sarcopenia by regulating chronic inflammation and oxidative stress, and suppresses obesity and dyslipidemia in castrated mice fed a high-fat diet [28]. In addition, fatigue recovery effects through the control of 5-HT and corticosterone have been reported in a sleep-deprived mouse model [15]. Furthermore, KRG reduces the production of allergic and inflammatory cytokines and inhibits trans-epidermal water loss, and synergistic effects have been reported to inhibit atopic dermatitis when used in combination with probiotics [29]. However, the immuno-oncologic effects of KRG via immune checkpoint remains largely undefined.
In a previous study, we found that Rg3, Rh2, and compound K exhibited the potential to disrupt the PD-1/PD-L1 interaction in competitive ELISA and protein-ligand docking simulations [30]. Although Rg3, Rh2, and compound K are major components of RGE, RGE is primarily used as a health supplement rather than as a single compound, such as Rg3, Rh2, or compound K. The effects of RGE on the PD-1/PD-L1 interaction and immune system have not been thoroughly examined. In the present study, we determined the effect of RGE alone on the PD-1/PD-L1 interaction and its antitumor effects against a humanized PD-1/PD-L1-expressing CRC mouse model.

RGE Enhances Tumor-Infiltrating CD8 + Cell-Mediated CRC Cell Killing
The inhibition of the hPD-1/hPD-L1 protein interaction by RGE prompted us to termine how RGE affects hPD-1-expressing CD8 + T cell-mediated, hPD-L1-express CRC cell killing in a co-culture model. Therefore, we established CD8 + T cell-media CRC cell killing assays using murine hPD-L1-expressing MC38 cells and tumor-infilt ing hPD-1-expressing CD8 + T cell isolated from hPD-L1 MC38 tumor-bearing hPD-1 m To confirm whether RGE affects the viability of hPD-L1 MC38 cells, the cells were trea with diverse concentrations of RGE for 72 h. RGE exhibited a cytotoxicity on the hPD MC38 cells at a concentration of 1000 μg/mL (Figure 2A), whereas the IC50 of RGE on hPD-L1 MC38 cells was 656.6 μg/mL. RGE had no cytotoxic effect on tumor-infiltra hPD-1-expressing CD8 + T cells up to a concentration of 300 μg/mL ( Figure 2B). In addit RGE treatment of CD8 + T cells increased the release of IL-2, suggesting that CD8 + T c are activated by RGE ( Figure 2C).
Considering the non-cytotoxic concentrations (300 μg/mL) of RGE on hPD-L1 M and hPD-1 CD8 + T cells, a co-culture cell system was established using tumor-infiltra CD8 + T cell as effector cells in conjunction with hPD-L1 MC38 cells as target cells to el date the CD8 + T cell-mediated antitumor effect of RGE. As shown in Figure 2D, the ad tion of tumor-infiltrated CD8 + T cells at 5-fold excess of the hPD-L1 MC38 cell num RGE treatment effectively inhibited the amount of surviving hPD-L1 MC38 cells. The totoxicity of hPD-L1 MC38 cells co-cultured with tumor-infiltrated CD8 + T cells was creasingly high in a concentration-dependent manner ( Figure 2E). Moreover, RGE tr ment apparently activated cytotoxic CD8 + T cells, which are key mediators of cancer killing, resulting in the elevated release of GrB ( Figure 2F). The results of the CD8 + T c mediated MC38 cell killing assay indicates that RGE efficiently enhanced the activatio tumor-infiltrating hPD-1-expressing CD8 + T cell immune function through inhibition the hPD-1/hPD-L1 interaction.
Upregulation of PD-L1 in cancer cells is a way to evade immune surveillance. reported that the PD-L1 antibody binds to PD-L1 on the outer membrane then interna the PD-L1 [31]. In addition, the PD-L1 is degraded in proteasome or lysosomes. RGE hibited PD-L1 expression in MDA-MB-231 cells, the breast cancer cell line, which hig express PD-L1 ( Figure S1).

RGE Enhances Tumor-Infiltrating CD8 + Cell-Mediated CRC Cell Killing
The inhibition of the hPD-1/hPD-L1 protein interaction by RGE prompted us to determine how RGE affects hPD-1-expressing CD8 + T cell-mediated, hPD-L1-expressing CRC cell killing in a co-culture model. Therefore, we established CD8 + T cell-mediated CRC cell killing assays using murine hPD-L1-expressing MC38 cells and tumor-infiltrating hPD-1-expressing CD8 + T cell isolated from hPD-L1 MC38 tumor-bearing hPD-1 mice. To confirm whether RGE affects the viability of hPD-L1 MC38 cells, the cells were treated with diverse concentrations of RGE for 72 h. RGE exhibited a cytotoxicity on the hPD-L1 MC38 cells at a concentration of 1000 µg/mL (Figure 2A), whereas the IC 50 of RGE on the hPD-L1 MC38 cells was 656.6 µg/mL. RGE had no cytotoxic effect on tumor-infiltrating hPD-1-expressing CD8 + T cells up to a concentration of 300 µg/mL ( Figure 2B). In addition, RGE treatment of CD8 + T cells increased the release of IL-2, suggesting that CD8 + T cells are activated by RGE ( Figure 2C).
Considering the non-cytotoxic concentrations (300 µg/mL) of RGE on hPD-L1 MC38 and hPD-1 CD8 + T cells, a co-culture cell system was established using tumor-infiltrating CD8 + T cell as effector cells in conjunction with hPD-L1 MC38 cells as target cells to elucidate the CD8 + T cell-mediated antitumor effect of RGE. As shown in Figure 2D, the addition of tumor-infiltrated CD8 + T cells at 5-fold excess of the hPD-L1 MC38 cell number, RGE treatment effectively inhibited the amount of surviving hPD-L1 MC38 cells. The cytotoxicity of hPD-L1 MC38 cells co-cultured with tumor-infiltrated CD8 + T cells was increasingly high in a concentration-dependent manner ( Figure 2E). Moreover, RGE treatment apparently activated cytotoxic CD8 + T cells, which are key mediators of cancer cell killing, resulting in the elevated release of GrB ( Figure 2F). The results of the CD8 + T cell-mediated MC38 cell killing assay indicates that RGE efficiently enhanced the activation of tumor-infiltrating hPD-1-expressing CD8 + T cell immune function through inhibition of the hPD-1/hPD-L1 interaction.
Upregulation of PD-L1 in cancer cells is a way to evade immune surveillance. It is reported that the PD-L1 antibody binds to PD-L1 on the outer membrane then internalize the PD-L1 [31]. In addition, the PD-L1 is degraded in proteasome or lysosomes. RGE inhibited PD-L1 expression in MDA-MB-231 cells, the breast cancer cell line, which highly express PD-L1 ( Figure S1).

Antitumor Effect of RGE in a Humanized PD-1/PD-L1 Knock-In Tumor Model
We demonstrated RGE suppression of hPD-L1 MC38 tumor growth in vivo by activated CD8 + T cells through inhibition of the hPD-1/hPD-L1 interaction. Next, we implanted hPD-L1-expressing MC38 cells in humanized PD-1 knock-in mice to evaluate the antitumor activity of RGE. Once the tumor volume reached 100 mm 3 (day 14), the mice were administered oral RGE at a dose of 100 or 300 mg/kg every day, whereas Keytruda (hPD-1 antibody) was administered intraperitoneally at dose of 2.5 mg/kg twice a week. RGE 300 mg/kg and Keytruda treatment delayed tumor progression compared with the vehicle group ( Figure 3A). RGE or Keytruda treatment markedly suppressed the growth of hPD-L1 MC38 allograft tumors, exhibiting a stronger inhibition compared with the vehicle group as monitored by decreased tumor weight and volume ( Figure 3B,C). Tumor suppression rate in the RGE 300 mg/kg or Keytruda group was 60% and 70% on day 16, respectively ( Figure 3D). The body weight of the mice treated with RGE or Keytruda

Antitumor Effect of RGE in a Humanized PD-1/PD-L1 Knock-In Tumor Model
We demonstrated RGE suppression of hPD-L1 MC38 tumor growth in vivo by activated CD8 + T cells through inhibition of the hPD-1/hPD-L1 interaction. Next, we implanted hPD-L1-expressing MC38 cells in humanized PD-1 knock-in mice to evaluate the antitumor activity of RGE. Once the tumor volume reached 100 mm 3 (day 14), the mice were administered oral RGE at a dose of 100 or 300 mg/kg every day, whereas Keytruda (hPD-1 antibody) was administered intraperitoneally at dose of 2.5 mg/kg twice a week. RGE 300 mg/kg and Keytruda treatment delayed tumor progression compared with the vehicle group ( Figure 3A). RGE or Keytruda treatment markedly suppressed the growth of hPD-L1 MC38 allograft tumors, exhibiting a stronger inhibition compared with the vehicle group as monitored by decreased tumor weight and volume ( Figure 3B,C). Tumor suppression rate in the RGE 300 mg/kg or Keytruda group was 60% and 70% on day 16, respectively ( Figure 3D). The body weight of the mice treated with RGE or Keytruda showed no significant change compared with the vehicle group throughout the experiments ( Figure 3E). These results indicate that RGE effectively suppresses tumor growth of hPD-L1 MC38 allograft tumors in hPD-1 mice without affecting body weight.
showed no significant change compared with the vehicle group throughout the experiments ( Figure 3E). These results indicate that RGE effectively suppresses tumor growth of hPD-L1 MC38 allograft tumors in hPD-1 mice without affecting body weight.

RGE Activates Tumor-Infiltrating CD8 + Cells in hPD-1/hPD-L1 MC38 Tumor Tissues
Considering the suppression of hPD-L1 MC38 tumor growth by RGE in in vitro, in vivo, and co-culture cell models, and the activation of tumor-infiltrating CD8 + T cells by RGE treatment, we examined CD8 + T cell infiltration into hPD-L1 MC38 tumor tissues following RGE or Keytruda treatment. As expected, the infiltration of CD8 + T cells and CD4 + T cells into tumors was markedly increased in the group treated with RGE 300

RGE Activates Tumor-Infiltrating CD8 + Cells in hPD-1/hPD-L1 MC38 Tumor Tissues
Considering the suppression of hPD-L1 MC38 tumor growth by RGE in in vitro, in vivo, and co-culture cell models, and the activation of tumor-infiltrating CD8 + T cells by RGE treatment, we examined CD8 + T cell infiltration into hPD-L1 MC38 tumor tissues following RGE or Keytruda treatment. As expected, the infiltration of CD8 + T cells and CD4 + T cells into tumors was markedly increased in the group treated with RGE 300 mg/kg or Keytruda ( Figure 4A,B,D). In addition, intratumoral infiltrating cytotoxic CD8 + T cells produced more GrB in the group treated with RGE 300 mg/kg or Keytruda ( Figure 4A−C). Taken together, our findings indicate that RGE enhances the abundance of CD8 + T cell infiltration on hPD-L1 MC38 tumors and showed greater antitumor activity in the hPD-1/hPD-L1 knock-in MC38 CRC tumor allograft mouse model. mg/kg or Keytruda ( Figure 4A,B,D). In addition, intratumoral infiltrating cytotoxic C T cells produced more GrB in the group treated with RGE 300 mg/kg or Keytruda (Fi 4A−C). Taken together, our findings indicate that RGE enhances the abundance of CD cell infiltration on hPD-L1 MC38 tumors and showed greater antitumor activity in hPD-1/hPD-L1 knock-in MC38 CRC tumor allograft mouse model.

Identification of Ginsenoside Rg3 of RGE by HPLC-DAD Analysis
The content of ginsenoside was greatly increased by RGE; thus, ginsenoside Rg3 selected as a standard for analysis. As shown in Figure 5

Identification of Ginsenoside Rg3 of RGE by HPLC-DAD Analysis
The content of ginsenoside was greatly increased by RGE; thus, ginsenoside Rg3 was selected as a standard for analysis. As shown in Figure 5, ginsenoside Rg3 was wellseparated and demonstrated good resolution, while minimizing interference from other analytes. Based on a comparison with the chromatogram, the retention time (tR; 30.25 min of standard, 30.24 min of RGM) and UV spectrum (203 nm) of the standard and ginsenoside Rg3 in RGE were detected.

Validation of the Analytical HPLC Method
The calibration curve for ginsenoside Rg3 was established by plotting the peak area versus the concentration. The linear correlation coefficient (R 2 ) for the calibration curve was greater than 0.9999 (Table 1). Based on this curve, the amount of ginsenoside Rg3 was determined to be 6.0423 ± 0.0228 mg/mL.

Discussion
The interaction of PD-1/PD-L1 via extracellular domains is a well-known immune checkpoint associated with the escape of cancer cells from immune-mediated destruction by inhibiting the anticancer effect of T cells [32]. Blocking the PD-1/PD-L1 interaction is an important anticancer target and the immunotherapeutic effect is higher in cancer cells overexpressing PD-L1 [33]. CRC with a microsatellite instability-high (MSI-H)/deficient DNA mismatch repair (dMMR) phenotype contains a high tumor mutational burden and tumor-infiltrating lymphocyte (TIL) activation. This indicates the clinical significance of TILs in MSI-H/dMMR CRC patients, suggesting that they may respond to ICIs over a long time period [34]. Treatment of MSI-H/dMMR CRC targeting PD-1 or PD-L1 inhibition by FDA-approved ICIs, including atezolizumab, avelumab, pembrolizumab, and nivolumab, has been reported [35]. Pembrolizumab (PD-1 inhibitor) was FDA-approved as a first-line treatment for patients with unresectable or refractory MSI-H/dMMR CRC [36]. Although these monoclonal antibodies show a remarkable therapeutic effect in clinical trials, their use is limited because of immune-related adverse reactions, difficulty in manufacturing, poor delivery, restricted permeability into tumor tissues because of their large size, high production cost, and severe side effects resulting from a long biological half-life [37,38]. In contrast, small molecule inhibitors have many advantages, including high tumor penetration, favorable oral bioavailability, fewer side effects, self-administration, lower manufacturing cost, and shorter biological half-life [39]. Clinical studies indicate that small molecule inhibitors are biologically safe and can block tumor growth as effectively as PD-1/PD-L1-based ICI monotherapy [40].

Validation of the Analytical HPLC Method
The calibration curve for ginsenoside Rg3 was established by plotting the peak area versus the concentration. The linear correlation coefficient (R 2 ) for the calibration curve was greater than 0.9999 (Table 1). Based on this curve, the amount of ginsenoside Rg3 was determined to be 6.0423 ± 0.0228 mg/mL.

Discussion
The interaction of PD-1/PD-L1 via extracellular domains is a well-known immune checkpoint associated with the escape of cancer cells from immune-mediated destruction by inhibiting the anticancer effect of T cells [32]. Blocking the PD-1/PD-L1 interaction is an important anticancer target and the immunotherapeutic effect is higher in cancer cells overexpressing PD-L1 [33]. CRC with a microsatellite instability-high (MSI-H)/deficient DNA mismatch repair (dMMR) phenotype contains a high tumor mutational burden and tumor-infiltrating lymphocyte (TIL) activation. This indicates the clinical significance of TILs in MSI-H/dMMR CRC patients, suggesting that they may respond to ICIs over a long time period [34]. Treatment of MSI-H/dMMR CRC targeting PD-1 or PD-L1 inhibition by FDA-approved ICIs, including atezolizumab, avelumab, pembrolizumab, and nivolumab, has been reported [35]. Pembrolizumab (PD-1 inhibitor) was FDA-approved as a first-line treatment for patients with unresectable or refractory MSI-H/dMMR CRC [36]. Although these monoclonal antibodies show a remarkable therapeutic effect in clinical trials, their use is limited because of immune-related adverse reactions, difficulty in manufacturing, poor delivery, restricted permeability into tumor tissues because of their large size, high production cost, and severe side effects resulting from a long biological half-life [37,38]. In contrast, small molecule inhibitors have many advantages, including high tumor penetration, favorable oral bioavailability, fewer side effects, self-administration, lower manufacturing cost, and shorter biological half-life [39]. Clinical studies indicate that small molecule inhibitors are biologically safe and can block tumor growth as effectively as PD-1/PD-L1-based ICI monotherapy [40].
In this study, we focused on the use of herbal medicines as small molecule inhibitors for tumor immunotherapy by evaluating the medicinal herb, Korean Red Ginseng, which is known to contain active components. The results indicate that RGE affects CRC through different mechanisms. Kim et al. demonstrated that RGE decreased CRC tumor cell invasion and migration as well as hypoxia-induced epithelial to mesenchymal transition by inhibiting the activation of the NF-κB and ERK1/2 signaling pathways [41]. Jeong et al. reported that RGE increased apoptosis in human CRC cells through Noxa activation by generating reactive oxygen species and endoplasmic reticulum stress [42]. Although various CRC studies have shown that RGE is useful for treating CRC, the anticancer effect of RGE on CRC through T cell activation has not been defined.
In the present study, we demonstrated that RGE treatment results in anti-CRC activity through CD8 + T cell-mediated antitumor immunity. First, we showed the capacity of RGE to disrupt the PD-1/PD-L1 interaction by a PD-1/PD-L1 protein binding ELISA. We found increased release of IL-2 in tumor-infiltrated CD8 + T cells by RGE treatment. Moreover, in a co-culture cell model, the antitumor immune effect of RGE occurred by enhancing hPD-1 tumor-infiltrated CD8 + T cell activity and killing hPD-L1 MC38 CRC cells with non-toxic doses of hPD-1 + CD8 + T cells. We also confirmed that tumor-infiltrating CD8 + T cells activated by RGE release GrB to kill MC38 cells. We further examined the antitumor effect of RGE by establishing humanized PD-1/PD-L1 knock-in mice tumor-bearing MC38 allografts. The results demonstrated that RGE and Keytruda (anti-PD-1 antibodies) markedly suppressed MC38 tumor growth. RGE and Keytruda not only significantly enhanced CD8 + T cell infiltration into tumor tissues, but released GrB produced by CD8 + T cells into the tumor microenvironment. Taken together, RGE enhances the potential antitumor immunogenic T cell response by modulating the PD-1/PD-L1 axis in CRC.
Ginsenosides, a group of dammarane triterpenoids, include ginsenoside Rg3, ginsenoside Rh2, ginsenoside Rh4, and ginsenoside Rk1. They are derived from RGE and these active compounds modulate PD-L1 expression, and thus, exhibit immunomodulatory potential against immune checkpoints [39]. Ginsenoside Rg3 attenuates PD-L1 expression in cisplatin-resistant human lung cancer cells [43]. Ginsenoside Rh4 inhibits aerobic glycolysis and inhibits PD-L1 expression through the AKT/mTOR pathway in esophageal cancer [44]. Ginsenoside Rk1 downregulate PD-L1 expression in human lung cancer cells by inhibiting the NF-κB pathway in an A549 xenograft mouse model [45]. In addition, Ginsenoside Rh2 suppressed melanoma tumor growth and increased the survival in a B16F10 tumor model by enhancing CD4 + and CD8 + T cell infiltration into the tumor tissue [46]. In the present study, we confirmed that the content of ginsenoside Rg3, the predominant compound, was 6 mg/mL, confirming that it was in high abundance in RGE. In our previous study, among 12 ginsenosides, we found that ginsenoside Rg3, Rh2, and compound K were potent immune checkpoint inhibitors of human PD-1/PD-L1 binding axis by in vitro competitive ELISAs and in silico protein-ligand docking simulations [30]. Thus, ginsenosides in RGE may activate anticancer T cell immunity by targeting immune checkpoints and modulating the tumor microenvironment to improve the immune response, leading to CRC cell death. RGE has been prescribed in Korea for the treatment of cancer patients. We demonstrated the immunological effects of RGE, rather than individual ginsenosides, through PD-1/PD-L1 blockade, thus providing further evidence for its use as a cancer treatment.

Preparation of RGE
Korean Red Ginseng was provided by the JADAM Co. (Yangju, Korea). The dried rootlet of Korean Red Ginseng (100 kg) was mixed with 800 L of 70% ethanol for 12 h. After immersion, the RG was extracted with 70% ethanol at 60 • C for 12 h using an electric extractor (Daesung, Seoul, Korea) and the extract was filtered through a 100-mesh sieve. After filtration, the RG was condensed at 60 • using a rotary evaporator (Daesung, Seoul, Korea). The free-dried extract powder (30 kg; yield 30%; abbreviated as RGE) was dissolved in water and analyzed for Rg3 content in RGE or used for experiments.

Tumor-Infiltrating CD8 + T Cell Isolation and Stimulation
Tumor-infiltrated CD8 + T cells were isolated from hPD-L1 MC38 tumor tissues of humanized PD-1 mice. To obtain single cell suspensions, tumor tissues were minced with scissors and digested in Roswell Park Memorial Institute (RPMI) 1640 medium containing 0.5 mg/mL type IV collagenase (Sigma-Aldrich, St. Louis, MO, USA), 10% (v/v) FBS, and 1% antibiotics at 37 • C for 1 h in an incubator containing 5% CO 2 . The digestion was quenched with excess media and the samples were filtered through a 100-µm, 70-µm, and 40-µm mesh cell strainer (SPL Life Sciences, Pocheon, Korea). The tumor-infiltrated CD8 + T cells were purified by immunomagnetic negative selection (#19853, STEMCELL Technologies, Inc., Vancouver, BC, Canada). The CD8 + T cells were cultured in RPMI 1640 medium containing 10% (v/v) FBS and 1% antibiotics at 37 • C in a humidified atmosphere containing 5% CO 2 . The T cell receptors were then stimulated using Mouse CD3/CD28 T Cell Activator (Life Technologies, Carlsbad, CA, USA) at 37 • C for 72 h.

Cell Viability Assay
Cell viability was assessed using the methylthiazolyldiphenyl-tetrazolium bromide (MTT, Sigma-Aldrich) assay. Briefly, the cells were plated in 96-well flat bottom tissue culture plates at a density of 5 × 10 3 cells/well and cultured with the indicated concentrations of RGE at 37 • C for 72 h. MTT reagent (0.5 mg/mL) was added and the plates were incubated in the dark at 37 • C for 4 h. The formazan was dissolved by adding DMSO and the optical density was quantified at 540 nm using a SpectraMax i3 microplate reader (Molecular Devices, San Jose, CA, USA).

IL-2 Measurement Assay
IL-2 released by activated tumor-infiltrating CD8 + T cells was assessed by a sandwich ELISA (#555148, BD Biosciences, San Diego, CA, USA). Briefly, an anti-mouse IL-2 antibody was coated onto 96-well plates (#3590, Corning, New York, NY, USA) with 0.1 M sodium carbonate (pH 9.5). Biotin antibody and streptavidin-HRP conjugate were added to each well and incubated at RT for 1 h. The relative absorbance was determined using a SpectraMax i3 microplate reader at 450 nm.

Co-Culture Experiments with Tumor-Infiltrated CD8 + T Cells and MC38 Cells
For co-culture experiments, hPD-L1 MC38 cells (5 × 10 4 cells/1.9 cm 2 ) as target cells were co-cultured with activated tumor-infiltrated CD8 + T cells (2.5 × 10 5 cells/1.9 cm 2 ) as effector cells at a target: effector cell ratio of 1:5 and treated with RGE (100-300 µg/mL) at 37 • C for 72 h. After co-culture for 72 h, the co-cultured hPD-L1 MC38 cell viability was measured using the cell counting kit-8 assay (Dojindo Molecular Technologies, Inc., Rockville, MD, USA). A lactate dehydrogenase (LDH) assay kit (#ab65393, Abcam, Cambridge, UK) was used to measure the cytotoxic activity on target cells via effector cells according to the manufacturer's instructions. The absorbance of the formazan products was determined at 450 nm using a SpectraMax i3 microplate reader.

Granzyme B Measurement Assay
Measurement of mouse GrB released by cytotoxic CD8 + T cells in the co-culture supernatants was quantified by a sandwich ELISA (#88-8022, Thermo Fisher Scientific, Waltham, MA, USA). Briefly, an anti-mouse GrB antibody was coated onto 96-well plates (#9018, Corning) with PBS. The biotin antibody and Av-HRP were added to each well and incubated at RT for 1 h. The relative absorbance was determined using a SpectraMax i3 microplate reader at 450 nm.

In Vivo RGE Treatment
The in vivo antitumor effect of RGE was evaluated using the hPD-L1 MC38 CRC model. Male tumor-bearing humanized PD-1 knock-in mice were randomly distributed into four groups (n = 5 per group) and RGE treatments were initiated when the tumor volume reached 100 mm 3 (day 14 post inoculation). Mice were injected with PBS (vehicle) or RGE (100, 300 mg/kg) daily by oral administration or Keytruda (anti-PD-1 human antibody, 2.5 mg/kg) twice per week by intraperitoneal injection for 21 days. Mice were euthanized after day 22. Tumor volume was calculated according to the formula V t = (length × width 2 )/2. Tumor suppression rate was calculated with the formula (V c − V t )/V c × 100%, where V c (vehicle) and V t (other treatment groups) are the tumor volumes.

Statistical Analysis
Data are presented as the mean ± standard deviation (SD) and were performed using GraphPad Prism software (GraphPad Software, Inc., La Jolla, CA, USA). Statistical comparisons were done using a one-way analysis of variance, followed by Tukey's post hoc test, which was used for comparisons between multiple groups, as indicated. p-values < 0.05 were considered statistically significant. All experiments, except animal studies, were conducted in at least three independent cases.

Conclusions
We discovered that RGE inhibits PD-1/PD-L1 interaction by in vitro protein binding ELISA and suppresses cell growth of CRC by increasing tumor-infiltrated CD8 + T cells immune function in co-culture cell model. In addition, we established that RGE effectively reduced hPD-L1 MC38 tumor growth via inducing the CD8 + T cell infiltration and GrB release in the tumor microenvironment in allograft tumor humanized PD-1/PD-L1 knockin mouse models. Based on our results, RGE, as a potential antitumor drug with blockade of the immune checkpoint, PD-1/PD-L1 axis, warrants further preclinical investigation in patients with CRC as a potential cancer immunotherapy.