GLS-010 is protected by patent #WO 2017/025051 A1 (29). It is manufactured at a concentration of 30 mg/mL and >99% purity. It is stable for 3 years at a storage temperature of 2-8°C.

The dynamic affinity of the anti-PD-1 antibody to human PD-1 was assessed by the SPR technique. The antibodies, including GLS-010 and nivolumab, were diluted to 2 μg/mL with 1× PBS/0.01% Tween 20 and injected (at 30 μL/min) to the channel of a sensor chip pre-coated with 10 μg/mL of protein A. The chip was rotated by 90° and rinsed by running buffer until the baseline was stable. The analyte WBP305.hPro1.ECD.His (containing the His-labeled extracellular fusion domain of human PD-1 protein) was diluted by the same buffer to different concentrations (0.625, 1.25, 2.5, 5, and 10 nmol/L). The buffer and the five concentrations of analytes were eluted through the chip channel successively, at a flow rate of 100 μL/min. The association phase for the samples was 240 s, followed by the dissociation phase of 600 s. The data were fitted by the Langmuir analysis 1:1 model to obtain the association rate constant (k_a), dissociation rate constant (k_d), and affinity constant (KD).

FACS was used to measure the affinity of the anti-PD-1 antibody to the cellular surface PD-1 protein. The WBP305.CHO-S.hPro1.C6 cells (an engineered cell line expressing the human PD-1 protein) or activated CD4⁺ T cells were incubated with PD-1 antibodies (2-fold serially diluted from 66.7 nmol/L to 0.01 nmol/L) at 4°C for 1 h. After the cells were washed, a PE-labeled goat-anti-human antibody was used to measure the binding of the anti-PD-1 antibodies to the cells. Nivolumab was used as a positive control, and non-transfected cells were used as a negative control. The mean fluorescence intensity (MFI) of the cells was measured by flow cytometer and quantitatively analyzed by the FlowJo software (BD Biosciences, Franklin Lake, NJ, USA).

The affinity of the antibody to human- (human PD-1.ECD.His), cynomolgus monkeys- (cynomolgus PD-1.ECD.His), and mouse- (mouse PD-1.ECD.His) derived PD-1 was assessed by ELISA. The ELISA plates were pre-coated with human PD-1.ECD.His, cynomolgus PD-1.ECD.His, and mouse PD-1.ECD.His. The plates were blocked with 2% BSA-containing PBS, and 100 μL of testing antibody (GLS-010, nivolumab, isotype control, and negative control) at 6.67 nmol/L was added. After incubation and rinsing, HRP-labeled goat anti-human IgG antibody was added, incubated at room temperature, rinsed, and developed with TMB substrate for 10 min. The reaction was stopped with 2 M HCl, and the absorbance was read at 450 nm using a microplate spectrophotometer within 15 min.

Plates were pre-coated with 1 μg/mL of human PD-1.ECD.mFc, CD28.ECD.mFc, CTLA-4.ECD.His, or ICOS.ECD.mFc at 4°C overnight. After 1 h of blocking, testing antibodies were added to the plates at 66.7 nmol/L. The plate was incubated at ambient temperature for 2 h. The binding of the antibodies to the immobilized protein was detected by the HRP-labeled goat anti-human IgG antibody. The color was developed by dispensing 100 μL of TMB substrate and then stopped using 100 μL of 2 N HCl. The absorbance was read at 450 nmol/L and 540 nmol/L using a microplate spectrophotometer.

FACS was used to evaluate the competitive blocking capability of the testing antibody to PD-L1 against PD-1 on the cell surface. WBP305.CHO-S.hPro1.C6 was incubated with anti-PD-1 antibodies (2-fold serially diluted from 66.7 to 0.01 nmol/L) at 4°C for 1 h. A constant concentration of PD-L1.ECD.mFc (containing mouse mFc-labeled extracellular fusion domain of human PD-L1) was added. The MFI of the cells was measured by flow cytometry and quantitatively analyzed by the FlowJo software (BD Biosciences, Franklin Lake, NJ, USA).

ELISA was used to evaluate the competitive blocking capability of the testing antibody to PD-L2 against PD-1. The ELISA plates were pre-coated with human PD-1-mFc (containing mouse Fc-labeled extracellular fusion domain of human PD-1 protein), incubated at 4°C overnight. After 1 h of blocking by PBS containing 2% BSA the next day, constant concentration of human PD-L2.ECD.His (5 μg/mL, containing His-labeled extracellular fusion domain of human PD-L2) and testing antibodies (3-fold serial dilution from 66.7 nmol/L to 0.001 nmol/L) were pre-mixed and added to the plates and incubated for 1 h at room temperature. HRP-labeled goat anti-His antibody was added after rinsing, incubated at room temperature, rinsed, and developed with TMB substrate for 10 min. The reaction was stopped using 2 M HCl, and the absorbance was read at 450 nm using a microplate spectrophotometer within 15 min.

The experiments were performed according to the “Methods for treating cancer using anti-PD-1 antibodies” (patent #US-9084776-B2). For allogeneic MLR, DCs and CD4⁺ T cells from different donors were mixed at a ratio of 1:10 DC:T cells in the presence or absence of the study antibodies. For autologous MLR, PBMCs were treated with CMV pp65 peptide and a low dose of rhIL-2 (20 U/mL) for 5 days before CD4⁺ T cell isolation. In the meanwhile, DCs were generated by inducing monocytes from the same donor’s PBMC. After 5 days’ culture, the DCs were harvested and pulsed with CMV pp65 peptide for one hour, and then co-cultured with CMV pre-treated CD4⁺ T cells at a ratio of 1:10 DC:T in the presence or absence of the study antibodies. MLR was set up in 96-well round-bottom plates using a complete RPMI-1640 medium. CD4⁺ T cells, various concentrations of antibodies (166.75 nM, 66.7 nM, 6.67 nM, 0.667 nM, 0.0667 nM and 0.00667 nM), and DCs were incubated at 37°C, 5% CO₂ for five days. The supernatant was harvested for IFN-γ production detection, and the cells were harvested to measure proliferation by ³H-TDR.

Antibody-dependent cell-mediated cytotoxicity (ADCC) refers to the killing of target cells by natural killer (NK) cells mediated by the binding of the Fab segment of specific antibody to target cells and the binding of the Fc segment to CD16 on the surface of NK cells. The activated CD4⁺ T cells(2×10⁴ cells/well) and effector cells (PBMC,1×10⁶ cells/well) were mixed at a ratio of 1:50 and incubated with various concentrations of testing anti-PD-1 antibodies at 37°C with 5% CO₂ for 5 h. Then, the lysis of the target cells was assessed by the LDH cytotoxicity kit. The BT474 cells (American Type Culture Collection (ATCC), Manassas, VA, USA) treated with trastuzumab were used as a positive control. The percentage of target cell lysis was calculated using the formula: % of lysis = (OD_sample – OD_PBMC – OD_{target background})/(OD_{target lysis} – OD_{target background}) * 100%.

Complement-dependent cytotoxicity (CDC) refers to the cell lysis effects induced by the membrane attack complex formed by the binding of specific antibodies to the corresponding cell surface antigens, which triggers the canonical complement activation pathway. The activated CD4⁺ T cells (2×10⁴ cells/well) were used as the target cells, mixed with various concentrations of anti-PD-1 antibodies. Then normal human serum complements (A112, Quidel Corp., San Diego, CA, USA) diluted to certain concentrations were added. The cells were cultured at 37°C with 5% CO₂ for 3 h. The cell lysis was assessed by the CellTiter-Glo cell viability kit. The target Ramos cells (American Type Culture Collection (ATCC), Manassas, VA, USA) were treated with rituximab (synthesized according to the C2B8 sequences described by the Genentech company in the patent US5736137) were used as the positive control. The percentage of target lysis was calculated as (1 – RFU_sample/RFU_{total cells}) * 100%.

The animal experiments were approved by the animal use committee of Crown Bioscience. During the experiments, the raising and using of the animals strictly abided by the AAALAC guidelines. The murine colon carcinoma MC38 cells were subcutaneously inoculated to the right flank of the hPD-1 knock-in mice (1×10⁶/mouse). When the tumor size reached amount 150 mm³ (which was considered day 0), the mice were randomly divided into four groups: IgG isotype control group, 10 mg/kg GLS-010 group, 20 mg/kg GLS-010 group, and positive control group (treated with 20 mg/kg pembrolizumab), with eight mice in each group. After treatment, the efficacy assessment was performed by evaluating the tumor volume and tumor growth inhibitory rate (T/C% and TGI%). Safety assessment was performed by evaluating the bodyweight change and death of the mice.

Forty-eight cynomolgus monkeys were selected and randomly divided into four groups according to the doses of GLS-010: control group (0 mg/kg), low dose group (5 mg/kg), moderate dose group (25 mg/kg), and high dose group (100 mg/kg). GLS-010 was administered once per week for 26 continuous weeks, followed by a recovery phase of 4 weeks.

The blood samples were collected at 1 ± 0.5 h after the last dosing on D50 for all the animals. In addition, the blood samples of the recovery phase were collected in the last week in the recovery phase (on D206 and D207). The blood samples were stained, washed, fixed, and measured about the occupancy of the receptors on the surface of the activated T cells by flow cytometry.

Nine male and nine female cynomolgus monkeys were selected and randomly divided into the low, moderate, and high dose groups and received an intravenous injection of 2, 6, and 18 mg/kg GLS-010, respectively. The samples for PK assessment were collected before (0) and at 0.25, 0.5, 1, 2, 4, 8, 24, 48, 72, 144, 312, 480, 648, and 816 h after dosing, respectively. The samples for ADA assessment were collected at -168, 0 (before dosing), 312, 648, and 984 h. WinNonlin Version 6.2.1 software (Certara, Princeton, NJ, USA) was used to process the serum levels of GLS-010 with the non-compartment model. The PK parameters were calculated by the linear and logarithmic trapezoidal methods.

Statistical differences between groups were evaluated using one-way analysis of variances (ANOVA) followed by the Dunnett’s test. All the data were analyzed using SPSS 19.0 (IBM, Armonk, NY, USA).

The association and dissociation diagrams of GLS-010 to the human PD-1 protein were analyzed; it was found that GLS-010 had a high affinity to the PD-1 protein, and the KD was 1.75×10^-10 M, while the KD of nivolumab was 1.16×10^-9 M (Table 1 and Figures 1A, B). GLS-010 had a relatively high concentration-dependent binding ability to both human PD-1-engineered CHO-S cells and activated human CD4⁺ T cells, for which the EC₅₀ was 0.77 nmol/L and 0.62 nmol/L, respectively (Figures 1C, D). Thus, GLS-010 could bind to human PD-1.ECD.His and cynomolgus PD-1.ECD.His, but not to mouse PD-1.ECD.His (Figure 1E). GLS-010 and nivolumab could specifically bind to the human PD-1 protein but not to the other proteins of the family (such as CD28, CTLA-4, and ICOS) (Figure 1F). Therefore, the results showed that GLS-010 had a high affinity and binding specificity to PD-1 without affecting the other members of the same family, indicating the potential of GLS-010 as a targeted therapy against PD-1.

Both GLS-010 and the positive control nivolumab effectively blocked the binding of PD-L1 to cell-surface PD-1 in CHO-S cells. The IC₅₀ of GLS-010 was 0.58 nmol/L (Figure 2A). In addition, both GLS-010 and the positive control nivolumab could effectively block the binding of PD-L2 to PD-1. The IC₅₀ of GLS-010 was 0.67 nmol/L (Figure 2B).

The findings of the allogeneic and autologous MLR showed that GLS-010 could enhance the secretion of IFN-γ by CD4⁺ T cells and promote CD4⁺ T cells proliferation. In addition, the activation of CD4⁺ T cells by the anti-PD-1 antibody was concentration-dependent (Figures 3A–D).

The effector cells (PBMC) showed no cytotoxic effects on activated CD4⁺ T cells in the presence of GLS-010 or positive control antibodies, and thus GLS-010 showed no evident ADCC effects, suggesting that GLS-010 could not mediate lysis of PD-1-positive lymphocytes by NK cells (Figure 4A). In addition, no evident lysis of activated CD4⁺ T cells was found in the presence of GLS-010 or positive control antibodies, and normal human serum complements, suggesting that GLS-010 had no evident CDC effects (Figure 4B).

In this study, the GLS-010 injection (10 and 20 mg/kg) showed statistically significant anti-tumor effects in the human PD-1 knock-in mouse model of MC38 tumors (Figure 5A). In addition, the effects of lower and equal concentrations of GLS-010 injection were comparable to the positive control (pembrolizumab) group, while the TGI% of these two concentrations of GLS-010 injection were both higher, numerically than pembrolizumab (Table 2 and Figures 5A, B).

After an intravenous injection of GLS-010, the RO rate on the D50 were saturated in the cynomolgus monkeys in the low-, moderate-, and high-dose groups, compared with the control group (CD3⁺T: 95.54% to 93.25%; CD8⁺T: 106.03% to 92.96%; and CD4⁺T: 119.72% to 77.98%). The RO rate on D206 (the last week of the recovery phase) in the low-, moderate-, and high-dose groups were lower than the D50 (CD3⁺T: 64.50% to 48.53%; CD8⁺T: 58.87% to 40.12%; CD4⁺T: 66.26% to 49.07%) but were higher than in the control group. In the recovery phase, GLS-010 was detected in the low-, moderate-, and high-dose groups on D190, D197, D204, and D211 (Table 3 and Figure 6A), suggesting that GLS-010 could exert long-term effects.

After a single intravenous injection of 2, 6, and 18 mg/kg GLS-010, the curves of the individual and average serum GLS-010 concentrations were not different between male and female animals (Table 4 and Figure 6B). With the increase of dose from 2 to 18 mg/kg, the systemic exposure level of GLS-010 (AUC_0-last) and C₀ increased proportionally. In contrast, the proportion of the increase in the exposure to GLS-010 (AUC_0-last and C₀) was higher than the proportion of the increase in the dose (Table 3).

In the 2 mg/kg group, the positive rate of ADA was 33.3%, of which the ADA-positive rate in female animals was 40%, and the antibody titer was higher in female animals than in males. In the 6 mg/kg group, the positive rate of ADA was 20%, and the antibody titer was slightly higher in male animals. In the 18 mg/kg group, the positive rate of ADA was 10%.

GLS-010 was administered by single intravenous infusion to cynomolgus monkeys at doses of 0, 100, 300 and 1000 mg/kg, respectively. A slight decrease in body weight, leukocytes, and neutrophils were observed only in 1000 mg/kg group for 14 days (Supplemental Data 1). Therefore, it is considered that the maximum tolerated dose of GLS-010 injection administered by single intravenous infusion is not reached.

No abnormal manifestation was observed in the animals in the 26 continuous weeks dosing study, including clinical manifestations, local irritation, body weight, food intake, body temperature, ophthalmic examination, blood routine, coagulation, serum biochemical parameters, urine examination, blood pressure, electrocardiogram, respiratory monitoring, immune analysis (serum IL-2, 4, 5, 6, 10, and 13; TNF-α; IgA and IgM; complement C3c and C4. Supplemental Data 2).