Development of an Anti-canine PD-L1 Antibody and Caninized PD-L1 Mouse Model as Translational Research Tools for the Study of Immunotherapy in Humans

Abstract

Immune checkpoint blockade therapy, one of the most promising cancer immunotherapies, has shown remarkable clinical impact in multiple cancer types. Despite the recent success of immune checkpoint blockade therapy, however, the response rates in patients with cancer are limited (∼20%-40%). To improve the success of immune checkpoint blockade therapy, relevant preclinical animal models are essential for the development and testing of multiple combination approaches and strategies. Companion dogs naturally develop several types of cancer that in many respects resemble clinical cancer in human patients. Therefore, the canine studies of immuno-oncology drugs can generate knowledge that informs and prioritizes new immuno-oncology therapy in humans. The challenge has been, however, that immunotherapeutic antibodies targeting canine immune checkpoint molecules such as canine PD-L1 (cPD-L1) have not been commercially available. Here, we developed a new cPD-L1 antibody as an immuno-oncology drug and characterized its functional and biological properties in multiple assays. We also evaluated the therapeutic efficacy of cPD-L1 antibodies in our unique caninized PD-L1 mice. Together, these in vitro and in vivo data, which include an initial safety profile in laboratory dogs, support development of this cPD-L1 antibody as an immune checkpoint inhibitor for studies in dogs with naturally occurring cancer for translational research. Our new therapeutic antibody and caninized PD-L1 mouse model will be essential translational research tools in raising the success rate of immunotherapy in both dogs and humans.

Significance: Our cPD-L1 antibody and unique caninized mouse model will be critical research tools to improve the efficacy of immune checkpoint blockade therapy in both dogs and humans. Furthermore, these tools will open new perspectives for immunotherapy applications in cancer as well as other autoimmune diseases that could benefit a diverse and broader patient population.

FIGURE 1

Working flow chart for production and validation of the cPD-L1 antibody. A, Immunization of antigen, cPD-L1 protein. B, Establishment of hybridomas (over 2,000 clones). C, Evaluation of the cPD-L1 specific antibody by the live cell–based antibody binding assay. D, Therapeutic antibody selection by cPD-L1/cPD-1 blockade assay. E,In vivo validation of therapeutic efficacy of cPD-L1 antibodies in the caninized PD-L1 mice. F, Production of the cPD-L1 chimeric antibody. G, Validation of the cPD-L1 chimeric antibody. H, Initial safety profile and PK analysis. PK, pharmacokinetic.

FIGURE 2

The high-throughput screening of therapeutic antibodies. A, Schematic diagram of the cPD-L1 antibody binding assay. BT549 cells expressing cPD-L1 were seeded on 96-well or 384-well plates. cPD-L1 antibodies (from hybridomas) and Alexa Fluor 488–conjugated anti-mouse IgG Fc-specific secondary antibody were added, and then green fluorescence signal was measured to quantify the amount of bound PD-L1 antibody by IncuCyte S3. B, A representative result of the cPD-L1 antibody binding assay. Kinetic graphs from each well of a 96-well plate showing quantitative binding of cPD-L1 antibodies on BT549 cells expressing cPD-L1 at 6-hour time intervals. The positive clones are highlighted in red (A2, I6, and I11). C, Representative images (at 18 hours) of cPD-L1 antibody binding. Green fluorescent merged images of cPD-L1–expressing cells are shown. D, Schematic diagram of the cPD-L1/cPD-1 blockade assay. BT549 cells expressing cPD-L1 were seeded on 96-well or 384-well plates. cPD-1-human IgG Fc (hFc) protein, Alexa Fluor 488–conjugated anti-human IgG Fc-specific secondary antibody and/or cPD-L1 antibody were added, and then green fluorescence signal was measured to quantify the amount of bound PD-1 protein by IncuCyte S3. E, A representative result of the cPD-L1/cPD-1 blockade assay. Kinetic graphs from each well of a 96-well plate showing quantitative binding of cPD-1 protein on BT549 cells expressing cPD-L1 at 3-hour intervals after the addition of cPD-L1 antibodies. The positive clones that blocked the interaction of cPD-L1/cPD-1 proteins are highlighted in red (A4 and B8). F, Representative images (at 18 hours) of the cPD-L1/cPD-1 blockade. Green fluorescent merged images of cPD-L1–expressing cells are shown. Note the lack of fluorescence due to the antibody binding to PD-L1 and blocking the interaction with cPD-1.

FIGURE 3

cPD-L1 antibodies enhance antitumor immunity in the caninized PD-L1 syngeneic mouse model. A, Knock-in strategy of the caninized PD-L1 mice (c57BL/c background). B, Validation of cPD-L1 protein expression in the MB49^cPD-L1 cells. Flow cytometric analysis of membrane located mPD-L1 and cPD-L1 protein in MB49 cells expressing cPD-L1 (MB49^cPD-L1) or MB49 parental cells. C, Immunofluorescence staining and protein expression pattern of mPD-L1 and cPD-L1 in MB49 or MB49^cPD-L1 tumor masses from the caninized PD-L1 mice. DAPI, nuclear counterstaining. Scale bar, 100 μm. D, Interaction of cPD-1 or mPD-1 protein with cPD-L1 or mPD-L1 protein with or without cPD-L1 antibody, 12C. His-tagged canine or mPD-L1 protein was immobilized on the Ni-NTA 96-well plate, and HRP-conjugated anti-human IgG Fc-specific secondary with mPD-1-hFc or cPD-1-hFc protein was added. OD₄₅₀ was measured to quantify the amount of bound PD-1 protein. E, Binding of cPD-L1 antibodies, 12C and 3C, with human PD-L1 (hPD-L1), mPD-L1, and cPD-L1 proteins. His-tagged human PD-L1, mPD-L1, or cPD-L1 protein was immobilized on the Ni-NTA 96-well plate, and anti-cPD-L1 antibodies, 12C or 3C, with HRP-conjugated anti-canine IgG-specific secondary was added. OD₄₅₀ was measured to quantify the amount of bound PD-L1 antibodies. Ab, antibody. F, Tumor growth of MB49^cPD-L1 in the caninized PD-L1 mice treated with cPD-L1 antibody, 12C or 3C. The IgG isotype of 12C and 3C antibodies is mouse IgG1 which is equivalent to human IgG4. Tumors were measured at the indicated timepoints (n = 8 per group). At the endpoint, the tumors were dissected. G–I, Immunofluorescence staining, and protein expression pattern of CD8 and granzyme B in MB49 tumor masses from IgG-, 12C-, or 3C-treated mice. DAPI, nuclear counterstaining. Scale bar, 100 μm. Representative images of immunostaining of CD8 and granzyme B in the MB49 tumor mass (G). CD8 (H) and granzyme B (I) were quantified using Gen5 software (BioTek). n = 10. Treatment with the PD-L1 antibody did not affect kidney function (serum creatinine; J) or liver enzyme activity (ALT; K), measured in blood collected at the end of the experiment. ALT, alanine aminotransferase.

FIGURE 4

The quality attributes of the purified anti-cPD-L1, 1210E4 (12C) chimeric antibody. A, Canine PD-L1, 12C10E4 chimeric antibody expression construct, pTRIOZ-cIgG2-cPD-L1 12C10E4. B, SDS-PAGE analysis of 12C chimeric antibody purity under nonreducing and reducing (2-mercaptoethanol) conditions. HC, heavy chain; LC, light chain; SM, protein size marker. C,IEF analysis of 12C chimeric antibody. Standard, pI standard. D, Peptide mapping analysis. The peptide mapping of cPD-L1 12C chimeric antibody. The 12C chimeric antibody was enzymatically digested with trypsin on S-trap micro columns from Protifi after reduction and alkylation. Peptides were then separated and analyzed by RP-LC/MS-MS. The resultant mass spectrometric data were analyzed using the PEAK PTM workflow in the PEAKS X PRO Studio 10.6 software package from Bioinformatics solutions Inc. to map the detected MS1 and MS2 ions to the amino acid sequence of antibody. The sequence coverage of heavy and light chains was 100% (453 of 453 amino acids) and 98.2 (223 of 227 amino acids), respectively. E,SEC analysis of 12C chimeric antibody. Standard, SEC standard. F, MALDI-MS profiling of permethylated N-glycans released from PNGase F-treated 12C chimeric antibody. The masses of indicated glycan species represent the [M + Na+] values.

FIGURE 5

Evaluation of the caninized cPD-L1 chimeric antibody. A, 12C antibody binding on the BT549^cPD-L1 cells. B, Flow cytometric analysis of the 12C chimeric antibody on the BT549^cPD-L1 cells. cIgG serves as a negative control. C, 12C chimeric antibody binding to cPD-L1 and cPD-L2. D, Binding affinity (K_D) analysis of 12C chimeric antibody by Octet. E, EC₅₀ of 12C chimeric antibody, 12C10E4. EC₅₀ = 0.419 μg/mL. The bound cPD-1 protein was quantified by measuring green fluorescence at the IncuCyte S3. F and G, Canine IO Panel (NanoString) analyses were used to query changes in gene expression upon activation of cPBMCs from three healthy pet dogs. The RNA from the resting and activated PBMCs was used for NanoString work. The Canine IO panel was used to query the changes in approximately 700 genes. Groupwise analyses were conducted using “Rosalind.” There were 65 genes that were differentially expressed when comparing control PBMCs with activated PBMCs (P < 0.05, FC > 1.5) including 30 upregulated and 35 downregulated genes. In the heatmap, each column consists of data from one sample. IFNγ (H) and TNFα (I) concentrations were analyzed in the activated canine PBMCs. J and K, Flow cytometric analysis of cPD-L1 protein expression on the K9TCC or the nuclear-restricted RFP-expressing K9TCC (K9TCC^nRFP) cells using the 12C chimeric antibody. The endogenous PD-L1 expression was stimulated by 50 ng/mL canine IFNγ for 12 hours. cIgG served as a negative control. L, The quantitative RT-PCR analysis of cPD-L1 (CD274) mRNA expression in the K9TCC or K9TCC^nRFP cells. M, The 12C chimeric antibody enhances the tumor cell killing. Canine bladder cancer, K9TCC cells were cocultured with cPBMCs that were activated with CD3 antibody (100 ng/mL) and IL2 (10 ng/mL) at a ratio of 1 tumor cell: 15 cPBMCs. The live tumor cell count at 72 hours is shown in the bar graph. N, IFNγ concentrations were analyzed in the medium from the coculture of the K9TCC cells and activated cPBMCs with/without the 12C chimeric antibody treatment.

FIGURE 6

Pharmacokinetic analysis of 12C chimeric antibody in laboratory dogs. A, Schematic diagram of ELISA for pharamacokinetic analysis. The concentration of the 12C chimeric antibody was measured in the 2 mg/kg (B) or 5 mg/kg (C) 12C antibody-treated dogs’ serum.