Anti-VEGF-PD1 bispecific antibody with novel structure and use thereof

Abstract

The disclosure relates to an anti-VEGF-PD1 bispecific antibody with a novel structure and a use thereof, which belongs to the technical field of molecular immunology. The CDR-H1 in the heavy chain variable region of the antibody is an amino acid sequence expressed by SEQ ID NO: 1, the CDR-H2 is an amino acid sequence expressed by SEQ ID NO: 2, the CDR-H3 is an amino acid sequence expressed by SEQ ID NO: 3, and the CDR-L in the light chain variable region of the antibody is an amino acid sequence expressed by SEQ ID NO:4.

Claims

1. An anti-VEGF-PD1 bispecific antibody Ps3Vm comprising the dsFv-PD1-linker-VEGF-H chain structure of SEQ ID NO: 9 and the anti-VEGF light chain of SEQ ID NO: 10.

2. A pharmaceutical composition comprising the bispecific antibody of claim 1 and a pharmaceutically acceptable carrier.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 is a graph showing a SDS-PAGE electrophoresis result of PD-1 and VEGF antigen (where A is VEGF antigen; B is PD-1 antigen).

(2) FIG. 2 is a graph showing an electrophoretic detection result of anti-PD-1 humanized antibody PDAB.

(3) FIG. 3 is a graph showing an electrophoretic detection result of anti-VEGF humanized antibody Avastin.

(4) FIG. 4 is a graph showing an electrophoretic detection result of the bispecific antibody A3P4.

(5) FIG. 5 is a graph showing a SEC detection result of the bispecific antibody A3P4.

(6) FIG. 6 is a schematic view of a protein structure of the bispecific antibody Vs3P4.

(7) FIG. 7 is a graph showing an electrophoretic detection result of the bispecific antibody Vs3P4.

(8) FIG. 8 is a graph showing a SEC detection result of the bispecific antibody Vs3P4.

(9) FIG. 9 is a schematic view of a protein structure of the bispecific antibody Ps3Vm.

(10) FIG. 10 is a graph showing an electrophoretic detection result of the bispecific antibody Ps3Vm.

(11) FIG. 11 is a graph showing a SEC detection result of the bispecific antibody Ps3Vm.

(12) FIG. 12 is a graph showing a comparison of the relative binding activity of PDAB, A3P4, Vs3P4, and Ps3Vm with respect to PD1-His by using ELISA.

(13) FIG. 13 is a graph showing a comparison of the relative binding activity of Avastin, A3P4, Vs3P4, and Ps3Vm with respect to rHuVEGF by using ELISA.

(14) FIG. 14 is a graph showing identification of the specificity of PDAB, A3P4, Vs3P4, Ps3Vm, Avastin and PD1 in binding epitopes by using competitive ELISA.

(15) FIG. 15 is a graph showing identification of the specificity of PDAB, A3P4, Vs3P4, Ps3Vm, Avastin and VEGF in binding epitopes by using competitive ELISA.

(16) FIG. 16 is a graph showing the change of amount of IL-2 secreted by T cells induced by Nivolumab, PDAB, Vs3P4, Ps3Vm, and IgG1 in vitro in relation to the change of concentration of antibody.

(17) FIG. 17 is a graph showing the change of amount of IFN-γ secreted by T cells induced by Nivolumab, PDAB, Vs3P4, Ps3Vm, and IgG1 in vitro in relation to the change of concentration of antibody.

(18) FIG. 18 is a graph showing the weight change of a mouse model initially constructed.

(19) FIG. 19 is a graph showing the change of tumor volume of a mouse model initially constructed.

DESCRIPTION OF THE EMBODIMENTS

(20) In order to make the disclosure more comprehensible, the disclosure will be further described below in conjunction with the embodiments and accompanying drawings. The following embodiments are only to illustrate the disclosure but not to limit it. The materials, reagents, instruments and methods used in the following examples are all conventional materials, reagents, instruments and methods in the art unless otherwise specified, and can be obtained through commercial channels.

Example 1 Preparation of PD1 and VEGF Antigens and Antibodies

(21) 1. Construction of Expression Vector for PD-1 Antigen

(22) In the cDNA of human PD-1 synthesized by Kingsray Corporation in Nanjing, the GeneID is 5133 and the cDNAID is NM_005018.2. After synthesizing the PD-1 gene in the extracellular region, an Fc purification tag was added to obtain PD-1-mFc, and Xba I was introduced at both ends. Two restriction enzyme splice sites of Bam HI were connected to the pTT5 expression plasmid, which was verified by sequencing. The sequenced plasmid was transfected into Trans10 (purchased from Beijing Quanshijin Biotechnology Co., Ltd.), and the single clone was picked and inoculated into 1 liter of LB liquid medium. When the OD.sub.600 was 1, the cells were collected by centrifugation, and a plasmid maxiprep kit (purchased from Qiagen) was used to extract the plasmid.

(23) 2. Construction of Expression Vector for VEGF Antigen

(24) The amino acid corresponding to the gene VEGF (NCBI Gene ID: 7422) was integrated with the Fc protein fragment mFc (Ig gamma-2A chain C region) of IgG of the mouse to obtain VEGF-mFc. In order to improve the expression efficiency of the target gene in the 293F cell expression system, the sequence was optimized, and Xba I was introduced at both ends. Two restriction enzyme splice sites of Bam HI were connected to the pTT5 expression plasmid, which was verified by sequencing. The sequenced plasmid was transfected into Trans10 (purchased from Beijing Quanshijin Biotechnology Co., Ltd.), and the single clone was picked and inoculated into 1 liter of LB liquid medium. When the OD.sub.600 was 1, the cells were collected by centrifugation, and a plasmid maxiprep kit (purchased from Qiagen) was used to extract the plasmid.

(25) 3. Expression and Purification of PD-1 and VEGF Antigens

(26) Transfect 293F cells (purchased from Invitrogen) with the correct expression vector identified by sequencing, which was conducted at a temperature of 37 degrees with 5% of CO.sub.2, and culture at 130 rpm/min for 7 days. Then, the supernatant was collected by centrifugation. The supernatant was centrifuged at 4000 rpm for 10 min, and then filtered with a 0.45 μm filter membrane; the filtrate was added with 400 mM of NaCl; and the pH was adjusted to 8.0. After the sample was filtered again through a 0.2 μm filter membrane, load the sample to a 5 mL HiTrap Protein A column equilibrated with PBS (137 mM of NaCl, 2.7 mM of KCl, 10 mM of Na.sub.2HPO.sub.4, 2 mM of KH.sub.2PO.sub.4, pH7.4). After the sample was loaded, use PBS for washing; the flow rate was 5 mL/min, and the UV monitoring result was at the standard level. Buffer B (1M Glycine, pH 3.5) was eluted at a flow rate of 1 mL/min. The flow-out peak was collected and neutralized with Tris to pH 7.5, and subjected to SDS-PAGE detection. The SDS-PAGE electrophoresis result is as shown in FIG. 1. The elution peak was concentrated and changed into PBS with an ultrafiltration concentration tube, thereby obtaining an antigen.

(27) 4. Construction of Anti-PD1 Humanized Antibody

(28) (1) Antigen-Immunized Mice and Hybridoma Screening

(29) In this experiment, three 8-week-old female BALB/c mice were selected, and the mice were immunized with a mixture of PD-1 extracellular domain antigen and Freund's complete adjuvant by intraperitoneal injection; the process was performed once a week in a total of 3 times. One week after the last immunization, the serum titers of the mice were measured. After the conditional titers were greater than 8K, the immunization was boosted once. The result showed that all 3 mice met the titers (the dilution value corresponding to the OD.sub.450 value greater than 2 times the negative control and greater than 0.25 is the titer of the antibody, and the requirement is met as long as the titer is greater than or equal to 8K). After 3 days, the mice are sacrificed, the spleens of the mice were taken, and the spleen cell population was obtained after grinding. The ELISA test results of mouse serum titer are shown in Table 1.

(30) TABLE-US-00001 TABLE 1 ELISA detection of 20871 mouse immune serum Serial number of mice/ Dilution comparison 1K 2K 4K 8K 16K 32K 64K 128K M1 1.23  1.04 0.508 0.427 0.281 0.189 0.103 0.067 M2 1.124 1.01 0.861 0.546 0.294 0.171 0.127 0.094 M3 1.254  1.149 0.918 0.545 0.325 0.18  0.116 0.088 Positive 2.549 control Negative 0.048 control

(31) The B cells of anti-human PD-1 antibody were screened by flow cytometry (FACS), placed in RPMI1640 medium, added with myeloma cells (SP2/0) and mixed, and cell integration was performed using 50% PEG solution. The integrated cells were appropriately diluted, divided and cultured in multiple 96-well culture plates, and HAT selective medium was added to kill unintegrated B cells and myeloma cells to obtain hybridoma cells. After cultured for 2 weeks, the 96-well plate cell culture supernatant was collected, combined with PD-1 antigen-coated 96-well microplate for 1 hour, added with anti-mouse/HRP secondary antibody and incubated for 1 hour, and finally added with TMB color reagent for 10 minutes. The light absorption value at 450 nm was measured with a microplate reader, and the hybridoma cells with binding activity to PD-1 were selected (primary screening: 12 pieces of 96-well plates to obtain 42 wells with OD value ≥0.5). Subsequently, flow cytometry (FACS) screening was performed to select hybridoma cells with PD-1/PD-L1 blocking activity. Then sub-cloning by limiting dilution method was carried out, and the cells with limited dilution were cultured in 96-well plates. When the clones grew to ⅙ of the full wells, the monoclones and polyclones were labeled, and the monoclones were detected by ELISA. After the detection, the monoclone with the highest OD value was then diluted into 96-well plates and subcloned again as described above. This process was repeated several times until the positive well ratio was 100%. The plant was successfully constructed, and an anti-PD-1 mouse monoclonal antibody cell strain was finally obtained. The result of subcloning by limiting dilution method is shown in Table 2, and the result of affinity identification is shown in Table 3.

(32) TABLE-US-00002 TABLE 2 Positive clone well plate position Serial Positive 96-well 384-well OD number clone plate plate value 1 2G8-1N8 2G8 1N8 1.022

(33) TABLE-US-00003 TABLE 3 Affinity identification antigen 0.1 μg/mL antigen 0.01 μg/mL antigen 0.001 μg/mL Plate Serial Dilution Dilution Dilution Dilution Dilution Dilution Dilution Dilution Dilution position number 1:3 1:9 1:27 1:3 1:9 1:27 1:3 1:9 1:27 1N8 1 0.98 0.837 0.793 0.124 0.181 0.108 0.11 0.149 0.11

(34) (2) Anti-PD-1 Murine Antibody Variable Region Gene Retrieval

(35) The anti-PD-1 hybridoma clones were selected, the total RNA was extracted using the Trizol method, and reverse transcription PCR was performed using antibody-specific (Isotype) specific primers or universal primers to respectively argument genes in the antibody light chain variable region (VL) and heavy chain variable region (VH), then connected to cloning vectors for DNA sequencing analysis. Finally, the complete DNA sequences of VL and VH were obtained and translated into corresponding amino acid sequences. The amino acid sequences of the heavy chain and light chain of the anti-PD-1 murine antibody are SEQ ID NO: 13-14 respectively; wherein, the CDR-H1, CDR-H2 and CDR-H3 amino acid sequences in the heavy chain variable region are SEQ ID NO: 15-17 respectively, the CDR-L1, CDR-L2 and CDR-L3 amino acid sequences in the light chain variable region are SEQ ID NO: 18-20 respectively.

(36) (3) Humanized Transformation of Variable Region Gene of Anti-PD-1 Murine Monoclonal Antibody

(37) (a) Humanization of Heavy Chain

(38) First, Ig Blast (http://www.ncbi.nlm.nih.gov/igblast) was used to analyze human germline genes with high homology to the VH gene of the mouse PD-1 antibody. The result showed that the heavy chain IGHV3-23 had 83% homology at the amino acid level, so it was selected as a candidate gene template for the heavy chain variable region. The CDR-H1, CDR-H2 and CDR-H3 of the mouse PD-1 antibody were numbered according to the Kabat numbering rule, and the corresponding CDR region amino acid sequence was introduced into the framework region of IGHV3-23. The amino acid No. 49 (S->T) and No. 78 (T->N) in the framework region were back-mutated to the original sequence of mouse PD-1 antibody. Then, the heavy chain CDR H1 No. 33 (G->D) and H2 No. 56 (S->R) were subjected to additional mutations, thereby completing the humanization of the heavy chain variable region. The heavy chain amino acid sequence of the anti-PD-1 humanized antibody is SEQ ID NO: 21; wherein, the CDR-H1, CDR-H2, and CDR-H3 amino acid sequences of the heavy chain variable region are SEQ ID NO: 22-24, respectively.

(39) (b) Humanization of Light Chain

(40) First, Ig Blast (http://www.ncbi.nlm.nih.gov/igblast) was used to analyze human germline genes with high homology to the VL gene of the mouse PD-1 antibody. The result showed that the light chain IGKV1-16 had 86% homology at the amino acid level, so it was selected as a candidate gene template for the light chain variable region. The CDR-L1, CDR-L2 and CDR-L3 of the mouse PD-1 antibody were numbered according to the Kabat numbering rule, and the corresponding CDR region amino acid sequence was introduced into the framework region of IGKV1-16. The amino acid No. 83 (F->M) in the framework region was back-mutated to the original sequence of mouse PD-1 antibody. Then, the light chain CDR L1 No. 31 (S->T) and No. 34 (S->A), L2 No. 56 (D->L) were additionally mutated to complete humanization of the light chain variable region. The light chain amino acid sequence of the anti-PD-1 humanized antibody is SEQ ID NO: 25; wherein, the CDR-L1, CDR-L2 and CDR-L3 amino acid sequences of the light chain variable region are SEQ ID NO: 26-28, respectively.

(41) (4) Affinity Maturation of Anti-PD-1 Humanized Antibodies

(42) An antibody mutant library was designed for the five CDR regions (L1, L3, H1, H2, and H3) of the anti-PD-1 humanized antibody, and the mutation sites covered all non-conserved sites of the CDR regions. A single chain antibody (scFv) gene was obtained by SOE-PCR reaction, after DNA gel recovery and digestion, it was connected with the digested pCANTAB-5E phage display vector to electrotransform TG1 competent bacteria to obtain 5 CDR-containing mutations single chain antibody library. By infecting M13KO7 helper phage to produce recombinant phage, a total of three rounds of elutriation were performed to retain and enrich antibody-binding mutants with strong binding ability. In each round of elutriation, the recombinant phage and the biotin-labeled recombinant human PD-1 antigen were combined for 2 hours, then streptavidin magnetic beads were added for 30 minutes, and 2% of TPBS, 1% of TPBS and PBS were used in sequence for washing for 5 times, 5 minutes per washing. After the elutriation, TG1 cells were immediately used for infection for the next round of preparation of recombinant phage. After three rounds of elutriation, the enriched TG1 monoclone were selected to prepare the recombinant phage supernatant, which was combined with a 96-well microtiter plate coated with 1 μg/mL PD-1 antigen for 1 hour, added with M13/HRP secondary antibody and incubated for 1 hour, and finally added with OPD to carry out a color reaction for 10 minutes. The light absorption value at 490 nm was measured with a microplate reader. After analyzing the data, calculate the relative affinity of antibody-containing mutants, and select 3, 6, and 5 clones with significantly improved affinity from the L3, H1, and H3 mutant libraries, respectively, and finally select one clone PDAB with the highest affinity from the H3 mutant library for the next study. The electrophoresis result is shown in FIG. 2.

(43) 5. Construction of Anti-VEGF Humanized Antibody

(44) The anti-VEGF humanized antibody used in this experiment was bevacizumab (Avastin, bevacizumab) launched by Roche (Genentech) in 2004. The antibody sequence (CN101210051A) was obtained from a public protein sequence website such as a patent website. The cDNA of the light chain and the heavy chain of VEGF antibody was artificially synthesized, and the synthesized cDNA was cloned into the pTT5 plasmid, and the plasmid construction was verified by sequencing. The sequenced plasmid was transfected into Trans10 (purchased from Beijing Quanshijin Biotechnology Co., Ltd.), and the single clone was picked and inoculated into 1 liter of LB liquid medium. When the OD.sub.600 was 1, the cells were collected by centrifugation, and a plasmid maxiprep kit (purchased from Qiagen) was used to extract the plasmid. The VEGF heavy chain expression vector and light chain expression vector (1:1) identified by sequencing were co-transfected into 293F cells, which was performed at a temperature of 37 degrees with 5% of CO.sub.2, and cultured at 130 rpm/min for 7 days. The supernatant was collected by centrifugation. The supernatant was centrifuged at 4000 rpm for 10 min, and filtered with a 0.45 μm filter membrane, and the filtrate was collected; the filtrate was added with 400 mM of NaCl; the pH was adjusted to 8.0. After the sample was filtered again through a 0.2 μm filter membrane, the sample was loaded to a 5 mL HiTrap MabSelect column (purchased from GE) that had been equilibrated with PBS (137 mM of NaCl, 2.7 mM of KCl, 10 mM of Na.sub.2HPO.sub.4, 2 m of MKH.sub.2PO.sub.4, pH7.4). After the sample was completely loaded, rinse with PBS at a flow rate of 5 mL/min, and the UV monitoring result is at a standard level. Buffer B (1M Glycine, pH3.5) was eluted at a flow rate of 1 mL/min. The flow-out peak was collected and neutralized with Tris to pH7.5, and subjected to SDS-PAGE detection. The SDS-PAGE non-reducing electrophoresis detection result is shown in FIG. 3. The elution peak was concentrated with an ultrafiltration concentration tube, and the solution was changed into PBS with a desalting column to obtain antibody VEGF protein.

Example 2 Preparation of Candidate Bispecific Antibodies

(45) 1. Preparation of scFv-VEGF-Linker-PD1-H Chain Structure Bispecific Antibody (A3P4):

(46) On the basis of existing anti-VEGF humanized antibodies, the heavy chain and light chain variable region genes are extracted and connected with peptides to form a single chain antibody scFv-VEGF. The scFv-VEGF was cloned into the N-terminus of the anti-PD1 antibody heavy chain to construct a bispecific antibody with scFv-VEGF-linker-PD1-H chain structure. The heavy chain expression vector and the light chain expression vector of anti-PD1 antibody were co-transformed into 293F cells, and the supernatant was collected and purified. SDS-PAGE was used to identify molecular weight and purity (see FIG. 4). By using SEC, it was detected that there are more antibody dimers in this sequence (see FIG. 5).

(47) 2. Preparation of dsFv-VEGF-Linker-PD1-H Chain Structure Bispecific Antibody (Vs3P4):

(48) On the basis of the original experiment, the new structure is redesigned. The heavy chain and light chain variable region genes of anti-VEGF humanized antibody were extracted and VH44cys and VL100cys mutations were performed (intra-chain disulfide bonds were increased to improve aggregation), and peptide chains were connected to form single-chain antibody dsFv-VEGF. The dsFv-VEGF was cloned into the N-terminus of the anti-PD1 antibody heavy chain to construct a bispecific antibody with a dsFv-VEGF-linker-PD1-H chain structure (see FIG. 6). The heavy chain expression vector and anti-PD1 antibody light chain expression vector were co-transformed into 293F cells, and the supernatant was collected and purified. SDS-PAGE was used to identify molecular weight and purity (see FIG. 7), and SEC detection was performed (see FIG. 8).

(49) 3. Preparation of dsFv-PD1-Linker-VEGF-H Chain Structure Bispecific Antibody (Ps3Vm):

(50) This is the third structural optimization design. On the basis of the existing anti-PD1 humanized antibody, the heavy chain and light chain variable region genes were extracted and VH44cys and VL100cys mutations were performed (intra-chain disulfide bonds were increased to improve aggregation), and peptide chains were connected to form single-chain antibody dsFv-PD1. The dsFv-PD1 was cloned into the N-terminus of the anti-VEGF antibody heavy chain to construct a bispecific antibody with a dsFv-PD1-linker-VEGF-H chain structure (see FIG. 9). The heavy chain expression vector and anti-VEGF antibody light chain expression vector were co-transformed into 293F cells, and the supernatant was collected and purified. SDS-PAGE was used to identify molecular weight and purity (see FIG. 10), and SEC detection was performed (see FIG. 11). The amino acid and nucleotide sequences of the heavy chain of the Ps3Vm antibody are SEQ ID NO: 9 and SEQ ID NO: 11, respectively, and the amino acid and nucleotide sequences of the CDR-H1, CDR-H2 and CDR-H3 in the heavy chain variable region are SEQ ID NO: 1-3 and SEQ ID NO: 5-7; the light chain amino acid and nucleotide sequences of the Ps3Vm antibody are SEQ ID NO: 10 and SEQ ID NO: 12, respectively. The amino acid and nucleotide sequences of CDR-L in the light chain variable region are SEQ ID NO: 4 and SEQ ID NO: 8, respectively.

Example 3 Measurement of Affinity of Bispecific Antibodies

(51) 1. Affinity of Bispecific Antibody to PD-1

(52) The enzyme-labeled plate was coated with PD-1-mFc, blocked with 1% of BSA, and the antibodies PDAB, A3P4, Vs3P4, and Ps3Vm of different concentrations were added to the enzyme-labeled plate respectively. After incubation at 37° C., the enzyme-labeled secondary antibody was added for incubation at 37° C. for 30 minutes. The light absorption value at 450 nm was measured with a microplate reader. The binding result of antibodies PDAB, A3P4, Vs3P4, Ps3Vm and antigen PD-1 showed that antibodies PDAB, A3P4, Vs3P4, and Ps3Vm can effectively bind to PD-1 protein, and the binding efficiency is dose-dependent. The results are shown in FIG. 12 and Table 4.

(53) TABLE-US-00004 TABLE 4 Binding efficiency of antibodies PDAB, A3P4, Vs3P4, Ps3Vm and PD-1 protein Concentration (ng/mL) PDAB A3P4 Vs3P4 Ps3Vm 1000 1.515 1.527 1.392 1.366 1.313 1.332 1.554 1.518 500 1.47  1.474 1.269 1.287 1.243 1.218 1.473 1.455 250 1.321 1.374 1.22  1.198 1.191 1.167 1.315 1.32  125 1.251 1.209 1.146 1.151 1.052 0.943 1.227 1.25  62.5 1.079 1.088 0.827 0.874 0.948 0.884 1.114 1.125 31.25 0.684 0.674 0.443 0.409 0.597 0.574 0.893 0.89  15.625 0.561 0.447 0.253 0.307 0.38  0.36  0.693 0.711 7.8125 0.235 0.245 0.167 0.157 0.174 0.186 0.487 0.547 3.90625 0.151 0.136 0.102 0.107 0.088 0.103 0.28  0.269 1.953125 0.063 0.06  0.043 0.043 0.042 0.05  0.143 0.175 0.9765625 0.038 0.032 0.034 0.03  0.023 0.028 0.081 0.112 0.48828125 0.023 0.024 0.014 0.017 0.014 0.02  0.043 0.06  EC50 33.9 47.92 36.58 20.01

(54) 2. Affinity of Bispecific Antibodies to VEGF

(55) The enzyme-labeled plate was coated with VEGF-mFc, blocked with 1% of BSA, and the antibodies Avastin, A3P4, Vs3P4, and Ps3Vm of different concentrations were added to the enzyme-labeled plate respectively. After incubation at 37° C., the enzyme-labeled secondary antibody was added for incubation at 37° C. for 30 minutes. The light absorption value at 450 nm was measured with a microplate reader. The binding result of antibodies Avastin, A3P4, Vs3P4, Ps3Vm and antigen VEGF showed that antibodies Avastin, A3P4, Vs3P4, and Ps3Vm can effectively bind to VEGF protein, and the binding efficiency is dose-dependent. The results are shown in FIG. 13 and Table 5.

(56) TABLE-US-00005 TABLE 5 Binding efficiency of antibodies Avastin, A3P4, Vs3P4, Ps3Vm and VEGF protein Concentration (ng/mL) Avastin A3P4 Vs3P4 Ps3Vm 2000 3.212 3.341 1.997 1.927 2.04  1.992 2.977 2.987 1000 3.178 3.157 1.825 1.762 1.898 1.766 2.581 2.68  500 2.888 2.817 1.51  1.394 1.581 1.501 2.248 2.324 250 2.448 2.384 1.156 1.066 1.095 1.097 1.776 1.763 125 1.773 1.687 0.707 0.635 0.605 0.636 1.309 1.281 62.5 1.008 1.062 0.417 0.363 0.377 0.351 0.911 0.755 31.25 0.619 0.617 0.212 0.196 0.208 0.199 0.47  0.433 15.625 0.349 0.335 0.122 0.108 0.103 0.109 0.264 0.273 7.8125 0.179 0.177 0.069 0.061 0.055 0.055 0.163 0.156 3.90625 0.098 0.092 0.036 0.037 0.033 0.034 0.081 0.085 1.953125 0.054 0.056 0.022 0.022 0.019 0.022 0.05  0.05  0.9765625 0.035 0.036 0.013 0.016 0.016 0.016 0.039 0.049 EC50 127.1 258.4 258.7 186.4

Example 4 Measurement of Specificity of Bispecific Antibodies

(57) 1. The Specificity of Bispecific Antibodies to PD-1

(58) The enzyme-labeled plate was coated with PD-1-mFc, blocked with 1% of BSA, and the antibodies PDAB, A3P4, Vs3P4, Ps3Vm, and Avastin of different concentrations were mixed with PD-1-mFc, respectively. After incubation at 37° C., the enzyme-labeled secondary antibody was added for incubation at 37° C. for 30 minutes. The light absorption value at 450 nm was measured with a microplate reader. The binding result of antibodies PDAB, A3P4, Vs3P4, Ps3Vm, and Avastin and antigen PD-1 showed that antibodies PDAB, A3P4, Vs3P4, Ps3Vm, and Avastin can effectively compete with PDL-1 to bind to PD-1 protein, and the binding efficiency is dose-dependent. The result is shown in FIG. 14.

(59) 2. Specificity of Bispecific Antibodies to VEGF

(60) The enzyme-labeled plate was coated with VEGF-mFc, blocked with 1% of BSA, and the antibodies Avastin, A3P4, Vs3P4, Ps3Vm, and PDAB of different concentrations were mixed with VEGF-A-hFc, respectively. After incubation at 37° C., the enzyme-labeled secondary antibody was added for incubation at 37° C. for 30 minutes. The light absorption value at 450 nm was measured with a microplate reader. The binding result of antibodies Avastin, A3P4, Vs3P4, Ps3Vm, and PDAB and antigen VEGF showed that antibodies Avastin, A3P4, Vs3P4, Ps3Vm, and PDAB can effectively compete with VEGF-A to bind to VEGF protein, and the binding efficiency is dose-dependent. The result is shown in FIG. 15.

Example 5 Candidate Bispecific Antibodies Induce T Cells to Secrete IL-2 In Vitro

(61) The Ficoll centrifugation method (purchased from GE) and CD4+ T cell enrichment column (purchased from R&D Systems) were used to prepare fresh PBMC and purify human T cells. Plate the cells into a 96-well flat bottom plate, after overnight cultivation, add six different concentrations of antibodies NIVO, PDAB, Vs3P4 and Ps3Vm in an amount of 0.0096, 0.048, 0.24, 1.2, 6, and 30 μg/mL respectively. The same type control antibody IgG1 of six different concentrations were added as a negative control. After 3 days of culture, the supernatant was collected, and the secretion level of the supernatant IL-2 was measured by using a Luminex apparatus (purchased from LifeTechnology) and a cytokine IL-2 detection kit (purchased from BD Biosciences). The result is shown in FIG. 16. The result showed that the bispecific antibodies Vs3P4 and Ps3Vm can effectively stimulate the function of T cells to secrete the cytokine IL-2, and the stimulation is related to antibody concentration, whereas the isotype control antibody cannot promote proliferation of T cells and secretion of IL-2.

Example 6 Candidate Bispecific Antibodies Induce T Cells to Secrete IFN-γ In Vitro

(62) The Ficoll centrifugation method (purchased from GE) and CD4+ T cell enrichment column (purchased from R&D Systems) were used to prepare fresh PBMC and purify human T cells. The monocytes were purified by using Miltenyi CD14 monocyte purification kit, and DC cells were generated after monocytes were cultured with GM-CSF and IL-4 (both purchased from PeproTech) for 7 days. Plate the cells into a 96-well flat bottom plate, after overnight cultivation, each culture with a total volume of 200 μL contains 10e5 purified T cells and 10e4 dendritic cells. Add six different concentrations of antibodies NIVO, PDAB, Vs3P4 and Ps3Vm in an amount of 0.0096, 0.048, 0.24, 1.2, 6, and 30 μg/mL respectively. The same type control antibody IgG1 of six different concentrations were added as a negative control. The cells were cultured for 5 days at 37° C. After 5 days, 100 μL of culture medium was taken from each culture for measurement of cytokine IFN-γ. The level of IFN-γ was measured by using OptEIA ELISA kit (purchased from BD Biosciences). The result is shown in FIG. 17. The result showed that the bispecific antibodies Vs3P4 and Ps3Vm can effectively stimulate the function of T cells to secrete the cytokine IFN-γ, and the stimulation is related to concentration, whereas the isotype control antibody cannot promote proliferation of T cells and secretion of IFN-γ.

Example 7 Candidate Bispecific Antibody Inhibits Tumor Growth in Mice

(63) 1. Preliminarily Constructed Mouse Model, Select PBMC Cells Suitable for the Experiment

(64) The PBMC cells, human colon cancer Colo-205 cells, B-NDG mice used in this experiment are commonly available in the industry.

(65) Human colon cancer Colo-205 cells purchased from the Chinese Academy of Sciences were cultured above 6.0*10.sup.7, and B-NDG mice (2.0*10.sup.6 cells each, 30 mice in total) subcutaneously inoculated with the cells were purchased from Biocytogen. The mice were fed normally, and when the tumor grew to a size of 100 mm.sup.3, the human PBMC cells purchased from different sources were intraperitoneally injected into each of the B-NDG severely immunodeficient mice purchased from Biocytogen at 1*10.sup.7. The growing condition of the tumor was observed until the tumor was formed successfully (select 10 groups of PBMC cells and inject each group of the cells into 3 mice for parallel experiments).

(66) The experimental results are shown in Table 6 below and FIG. 18 and FIG. 19 (all figures are averages):

(67) TABLE-US-00006 TABLE 6 Change of weight and tumor size in preliminarily constructed mouse model DAY 0 3 7 10 15 18 21 G1 Body 20.30 20.20 20.10 19.70 19.10 17.77 18.00 Weight (g) Tumor 102.73 209.05 599.98 644.22 1223.34 1581.46 1918.73 Size (mm.sup.3) G2 Body 19.03 19.07 19.13 18.17 18.43 17.70 17.97 Weight (g) Tumor 116.63 214.43 579.18 805.71 1242.41 1830.42 2160.05 Size (mm.sup.3) G3 Body 19.10 18.56 18.07 17.57 18.10 16.23 16.90 Weight (g) Tumor 121.03 237.41 534.89 692.74 999.83 1647.12 1974.83 Size (mm.sup.3) G4 Body 19.73 19.65 19.47 19.03 19.27 17.03 17.70 Weight (g) Tumor 103.10 138.20 404.44 575.30 897.40 1471.85 1634.99 Size (mm.sup.3) G5 Body 20.13 20.01 19.93 18.90 18.60 18.45 19.60 Weight (g) Tumor 100.92 205.06 432.67 709.04 1056.45 1425.82 1872.85 Size (mm.sup.3) G6 Body 18.97 18.74 18.43 18.83 17.77 17.07 17.23 Weight (g) Tumor 93.56 137.69 487.36 591.19 1199.08 1307.07 1439.73 Size (mm.sup.3) G7 Body 19.90 19.05 18.40 18.03 17.23 16.20 15.87 Weight (g) Tumor 109.67 153.39 590.91 794.74 1358.81 1660.09 1721.64 Size (mm.sup.3) G8 Body 20.33 19.13 18.60 18.37 19.10 18.00 15.10 Weight (g) Tumor 115.43 211.55 555.03 793.15 1273.73 1550.99 1872.78 Size (mm.sup.3) G9 Body 19.97 19.20 18.43 18.03 18.10 15.93 16.93 Weight (g) Tumor 120.50 180.85 511.27 619.23 978.95 1290.83 1339.49 Size (mm.sup.3) G10 Body 19.67 19.53 19.13 18.67 18.03 17.25 18.10 Weight (g) Tumor 108.68 197.87 611.62 797.08 1662.81 1891.89 2096.03 Size (mm.sup.3)

(68) 2. Use the Selected PBMC Cells to Build Animal Models

(69) PBMC cells (G1, G2, G8, G10) with successful matching were selected and injected into B-NDG-b2m MHC knockout severely deficient mice (1*10.sup.7 cells per mouse) purchased from Biocytogen. Meanwhile, the mice were subcutaneously inoculated with human colon cancer Colo-205 cells to observe whether tumors are formed successfully, which is a pre-experiment. Only 8 mice were injected and inoculated (as two sets of parallel experiments). Tumorigenicity was observed and the PBMC cells that were successfully formed into tumors were selected for the next stage of experiment.

(70) The change in tumor size is shown in Table 7 below (all numbers are averages in the table): Tumor Volume (mm.sup.3)

(71) TABLE-US-00007 TABLE 7 Tumor volume changes of constructed animal model with selected PBMC cells DAY 0 3 7 10 15 18 21 G1 100.23 230.56 548.73 668.39 1268.45 1684.54 2025.68 G2 108.35 203.21 502.72 851.94 1054.26 1563.81 1954.29 G8 113.51 211.55 465.84 712.43 946.25 1458.48 1743.65 G10 100.62 198.36 600.26 900.25 1356.31 1965.32 2200.25

(72) 3. Animal Model Construction for Experiments

(73) The above PBMC cells (G10) were selected and injected into B-NDG-b2m MHC knockout severely deficient mice (1*10.sup.7 cells per mouse in 4 groups, 6 mice per group) purchased from Biocytogen. Meanwhile, the mice were subcutaneously inoculated with human colon cancer Colo-205 cells to observe whether tumors are formed successfully. The mice were randomly divided into 4 groups according to growth of tumor. Negative control group (intraperitoneally injected with saline), Vs3P4 group (subjected to tail vein injection of Vs3P4 antibody in an amount of 3 mg/kg), Ps3Vm group (subjected to tail vein injection of Ps3Vm antibody in an amount of 3 mg/kg), positive control group bevacizumab (subjected to tail vein injection in an amount of 3 mg/kg). The mice were administered with dose every 3 days for a total of 21 days. The tumor volume changes are shown in the table below: Tumor Volume (mm.sup.3)

(74) TABLE-US-00008 TABLE 8 Tumor volume changes in animal models for experiments Days 0 3 7 10 15 18 21 Saline 400.23 609.56 812.26 1009.32 1478.26 1993.54 2365.82 Vs3P4 403.26 454.35 353.02 256.51 202.56 194.32 203.45 Ps3Vm 400.24 469.35 260.24 134.87 108.46 156.36 147.51 Avastin 408.58 498.69 338.56 305.45 289.5 306.43 324.62 antibody

(75) The disclosure tested anti-VEGF-PD1 bispecific antibodies with three different structures, and respectively tested their antibody effects from molecular, cellular, biological aspects. The results show that: bispecific antibody Ps3Vm (with VEGF as the skeleton with insertion of dsFv-PD1 monomer) has the best test effect, can effectively bind to PD-1 and VEGF protein, and can effectively compete with PDL-1 to bind to PD-1 protein and compete with VEGF-A to bind to VEGF protein, while can effectively stimulate the function of T cells and secretion of cytokines IL-2 and IFN-γ. In contrast, the isotype control antibody cannot promote proliferation of the T cells and secretion of IL-2 and IFN-γ. In addition, the bispecific antibody Ps3Vm can also significantly inhibit growth of tumor in mice.

(76) Although the disclosure has been disclosed in the above embodiments, it is not intended to limit the disclosure, and those skilled in the art can make some modifications and refinements without departing from the spirit and scope of the disclosure. Therefore, the scope of the disclosure is subject to the definition of the scope of the appended claims.