Treatment and prevention of cancer using HER3 antigen-binding molecules

11208498 · 2021-12-28

Assignee

Inventors

Cpc classification

International classification

Abstract

Methods for treating or preventing a cancer in a subject are disclosed, wherein the cancer comprises cells having a mutation resulting in increased expression of a ligand for HER3, wherein the method comprises administering a therapeutically or prophylactically effective amount of an antigen-binding molecule which is capable of binding to HER3 to the subject.

Claims

1. A method of treating a cancer in a subject, wherein the cancer comprises cells having a mutation resulting in increased expression of a ligand for HER3, wherein the method comprises administering a therapeutically or prophylactically effective amount of an antigen-binding molecule which is capable of binding to HER3 to the subject, and wherein the antigen-binding molecule comprises: (i) a heavy chain variable (VH) region incorporating the following CDRs: HC-CDR1 having the amino acid sequence of SEQ ID NO:43 HC-CDR2 having the amino acid sequence of SEQ ID NO:46 HC-CDR3 having the amino acid sequence of SEQ ID NO:51; and (ii) a light chain variable (VL) region incorporating the following CDRs: LC-CDR1 having the amino acid sequence of SEQ ID NO:91 LC-CDR2 having the amino acid sequence of SEQ ID NO:94 LC-CDR3 having the amino acid sequence of SEQ ID NO:99.

2. The method according to claim 1, wherein the ligand for HER3 comprises an amino acid sequence having at least 60% sequence identity to the EGF-like domain of an NRG.

3. The method according to claim 1, wherein the mutation results in increased expression of the EGF-like domain of an NRG at the cell surface.

4. The method according to claim 1, wherein the cancer comprises cells having an NRG gene fusion.

5. The method according to claim 4, wherein the NRG gene fusion is selected from CLU-NRG1, CD74-NRG1, DOC4-NRG1, SLC3A2-NRG1, RBPMS-NRG1, WRN-NRG1, SDC4-NRG1, RAB2IL1-NRG1, VAMP2-NRG1, KIF13B-NRG1, THAP7-NRG1, SMAD4-NRG1, MDK-NRG1, TNC-NRG1, DIP2B-NRG1, MRPL13-NRG1, PARP8-NRG1, ROCK1-NRG1, DPYSL2-NRG1, ATP1B1-NRG1, CDH6-NRG1, APP-NRG1, AKAP13-NRG1, THBS1-NRG1, FOXA1-NRG1, PDE7A-NRG1, RAB3IL1-NRG1, CDK1-NRG1, BMPRIB-NRG1, TNFRSF10B-NRG1, MCPH1-NRG1 and SLC12A2-NRG2.

6. The method according to claim 4, wherein the NRG gene fusion is selected from CLU-NRG1, CD74-NRG1, SLC3A2-NRG1 or VAMP2-NRG1.

7. The method according to claim 1, wherein the cancer comprises cells expressing HER3.

8. The method according to claim 1, wherein the cancer derives from the lung, breast, head, neck, kidney, ovary, pancreas, prostate, uterus, gallbladder, colon, rectum, bladder, soft tissue or nasopharynx.

9. The method according to claim 1, wherein the cancer is selected from lung cancer, non-small cell lung cancer, lung adenocarcinoma, invasive mucinous lung adenocarcinoma, lung squamous cell carcinoma, breast cancer, breast carcinoma, breast invasive carcinoma, head and neck cancer, head and neck squamous cell carcinoma, renal cancer, renal clear cell carcinoma, ovarian cancer, ovarian serous cystadenocarcinoma, pancreatic cancer, pancreatic adenocarcinoma, pancreatic ductal adenocarcinoma, prostate cancer, prostate adenocarcinoma, endometrial cancer, uterine carcinosarcoma, gallbladder cancer, cholangiocarcinoma, colorectal cancer, bladder cancer, urothelial bladder cancer, sarcoma, soft tissue sarcoma, neuroendocrine tumor and neuroendocrine tumor of the nasopharynx.

10. The method according to claim 1, wherein the cancer is selected from lung cancer, non-small cell lung cancer, lung adenocarcinoma, invasive mucinous lung adenocarcinoma and lung squamous cell carcinoma.

11. The method according to claim 1, wherein the antigen-binding molecule comprises: (i) a VH region incorporating the following CDRs: HC-CDR1 having the amino acid sequence of SEQ ID NO:41 HC-CDR2 having the amino acid sequence of SEQ ID NO:45 HC-CDR3 having the amino acid sequence of SEQ ID NO:48; and (ii) a VL region incorporating the following CDRs: LC-CDR1 having the amino acid sequence of SEQ ID NO:88 LC-CDR2 having the amino acid sequence of SEQ ID NO:92 LC-CDR3 having the amino acid sequence of SEQ ID NO:95.

12. The method according to claim 1, wherein the antigen-binding molecule comprises: a VH region comprising an amino acid sequence having at least 70% sequence identity to the amino acid sequence of SEQ ID NO:36; and a VL region comprising an amino acid sequence having at least 70% sequence identity to the amino acid sequence of SEQ ID NO:83.

13. The method according to claim 1, wherein the antigen-binding molecule comprises: a VH region incorporating the following framework regions (FRs): HC-FR1 having the amino acid sequence of SEQ ID NO:53 HC-FR2 having the amino acid sequence of SEQ ID NO:59 HC-FR3 having the amino acid sequence of SEQ ID NO:66 HC-FR4 having the amino acid sequence of SEQ ID NO:71.

14. The method according to claim 1, wherein the antigen-binding molecule comprises: a VL region incorporating the following framework regions (FRs): LC-FR1 having the amino acid sequence of SEQ ID NO:104 LC-FR2 having the amino acid sequence of SEQ ID NO:110 LC-FR3 having the amino acid sequence of SEQ ID NO:120 LC-FR4 having the amino acid sequence of SEQ ID NO:125.

15. The method according to claim 1, wherein the antigen-binding molecule comprises a heavy chain comprising the amino acid sequence of SEQ ID NO:171.

16. The method according to claim 1, wherein the antigen-binding molecule comprises a light chain comprising the amino acid sequence of SEQ ID NO:177.

Description

BRIEF DESCRIPTION OF THE FIGURES

(1) Embodiments and experiments illustrating the principles of the invention will now be discussed with reference to the accompanying figures.

(2) FIGS. 1A and 1B. Histograms showing staining of cells by anti-HER3 antibodies as determined by flow cytometry. Histograms show staining of HEK293 cells (which do not express HER3), or HEK293 HER3 overexpressing cells (HEK293 HER3 O/E) by (1A, 1B) anti-HER3 antibody clone 10D1 and (1B) anti-HER3 antibody clone LJM716.

(3) FIGS. 2A and 2B. Histograms showing staining of cells by anti-HER3 antibodies as determined by flow cytometry. Histograms show staining of HEK293 cells (which do not express HER3), or HEK293 HER3 overexpressing cells (HEK293 HER3 O/E) by (2A, 2B) anti-HER3 antibody clone 4-35-B2 and (2B) anti-HER3 antibody clone LJM716.

(4) FIGS. 3A and 3B. Histograms showing staining of cells by anti-HER3 antibodies as determined by flow cytometry. Histograms show staining of HEK293 cells (which do not express HER3), or HEK293 HER3 overexpressing cells (HEK293 HER3 O/E) by (3A, 3B) anti-HER3 antibody clone 4-35-B4 and (3B) anti-HER3 antibody clone LJM716.

(5) FIG. 4. Histograms showing staining of cells by anti-HER3 antibodies as determined by flow cytometry. Histograms show staining of HEK293 cells (which do not express HER3), or HEK293 HER3 overexpressing cells (HEK293 HER3 O/E) by anti-HER3 antibody clone 10A6.

(6) FIGS. 5A and 5B. Graphs showing the results of ELISA analysis of binding of anti-HER3 antibody clone 10D1 to (5A) human, mouse, rat and cynomolgus macaque HER3, and (5B) human EGFR and human HER2. EC.sub.50 values are shown.

(7) FIGS. 6A and 6B. Graphs showing the results of ELISA analysis of binding of anti-HER3 antibody clone 4-35-B2 to (6A) human, mouse, rat and cynomolgus macaque HER3, and (6B) human EGFR and human HER2.

(8) FIGS. 7A and 7B. Graphs showing the results of ELISA analysis of binding of anti-HER3 antibody clone 4-35-B4 to (7A) human HER3, human EGFR and human HER2, and (7B) human, mouse, rat and cynomolgus macaque HER3.

(9) FIG. 8. Representative sensorgram showing the results of analysis of affinity of binding of anti-HER3 antibody clone 10D1 to human HER3. Kon, Koff and K.sub.D are shown.

(10) FIG. 9. Representative sensorgram showing the results of analysis of affinity of binding of anti-HER3 antibody clone 4-35-B2 to human HER3.

(11) FIG. 10. Representative sensorgram showing the results of analysis of affinity of binding of anti-HER3 antibody clone 4-35-B4 to human HER3.

(12) FIG. 11. Graph showing the results of analysis of stability of anti-HER3 antibody clone 10D1 by Differential Scanning Fluorimetry analysis.

(13) FIG. 12. Graph showing the results of the analysis of recombinantly-expressed anti-HER3 antibody clone 10D1 by size exclusion chromatography.

(14) FIG. 13. Images showing the results of the analysis of anti-HER3 antibody clone 10D1 expression by SDS-PAGE and western blot. Lanes: M1=TaKaRa protein marker Cat. No. 3452; M2=GenScript protein marker Cat. No. M00521; 1=reducing conditions; 2=non-reducing conditions; P=positive control: human IgG1, Kappa (Sigma Cat. No. 15154). For western blot, the primary antibodies used were goat anti-human IgG-HRP (GenScript Cat No. A00166) and goat anti-human kappa-HRP (SouterhnBiotech Cat No. 2060-05).

(15) FIGS. 14A and 14B. Representative sensorgram and table showing the results of analysis of competition between different anti-HER3 antibody clones for binding to HER3.

(16) FIG. 15. Graph showing the results of analysis of the inhibition of interaction between HER3 and HER2 by anti-HER3 antibody clone 10D1 as determined by ELISA.

(17) FIGS. 16A and 16B. Table and histograms showing gene and protein expression of EGFR protein family members and their ligands by different cancer cell lines.

(18) FIG. 17. Images showing the results of analysis of the effect of anti-HER3 antibody clone 10D1 treatment on the HER3-mediated signalling in N87 and FaDu cells by phospho-western blot. UN=untreated; T=treated with anti-HER3 antibody clone 10D1.

(19) FIG. 18. Images and graph showing the results of analysis of the effect of anti-HER3 antibody clone 10D1 treatment on the HER3-mediated signalling in FaDu cells using the Phosphoprotein Antibody Array assay kit. Untreated=untreated FaDu cells; Treated=FaDu cells treated with anti-HER3 antibody clone 10D1.

(20) FIGS. 19A and 19B. Graphs showing the percent confluence of cells relative to an untreated control condition (100%), for the indicated cells lines as determined by CCK8 assay, following incubation in the presence of anti-HER3 antibody clone 10D1. (19A) Shows the results obtained for N87 cells, and (19B) shows the results obtained for FaDu cells.

(21) FIG. 20. Graph showing the results of analysis of tumour volume over time in a N87 cell-line derived mouse gastric carcinoma model. Anti-HER3 antibody clone 10D1 was administered IP, biweekly at 500 μg per dose for a total of 10 doses. A control treatment group received an equal volume of PBS (vehicle).

(22) FIG. 21. Graph showing the results of analysis of tumour volume over time in a N87 cell-line derived mouse gastric carcinoma model. Anti-HER3 antibody clone 4-35-B2 was administered IP, weekly at 11 mg/kg per dose for a total of 4 doses. A control treatment group received an equal amount of isotype control antibody (isotype).

(23) FIG. 22. Graph showing the results of analysis of tumour volume over time in a SNU16 cell-line derived mouse gastric carcinoma model. Anti-HER3 antibody clone 10D1 was administered IP, biweekly at 500 μg per dose for a total of 9 doses. A control treatment group received an equal volume of PBS (vehicle).

(24) FIG. 23. Graph showing the results of analysis of tumour volume over time in a FaDu cell-line derived mouse model of head and neck squamous cell carcinoma. Anti-HER3 antibody clone 10D1 was administered IP, weekly at 500 μg per dose for a total of 4 doses. Control treatment groups received an equal volume of PBS (vehicle), or the same dose of an isotype control antibody (isotype).

(25) FIG. 24. Graph showing the results of analysis of tumour volume over time in a FaDu cell-line derived mouse model of head and neck squamous cell carcinoma. Anti-HER3 antibody clone 10D1 was administered IP, biweekly at 500 μg per dose for a total of 8 doses. A control treatment group received an equal volume of PBS (vehicle).

(26) FIG. 25. Graph showing the results of analysis of tumour volume over time in an OvCAR8 cell-line derived mouse model of ovarian carcinoma. Anti-HER3 antibody clone 10D1 was administered IP, biweekly at 500 μg per dose for a total of 9 doses. A control treatment group received an equal volume of PBS (vehicle).

(27) FIG. 26. Graph showing the results of analysis of tumour volume over time in a HCC-95 cell-line derived mouse model of squamous lung cell carcinoma. Anti-HER3 antibody clone 10D1 was administered IP, biweekly at 500 μg per dose for a total of 4 doses. A control treatment group received an equal volume of PBS (vehicle).

(28) FIG. 27. Graph showing the results of analysis of tumour volume over time in an A549 cell-line derived mouse model of lung adenocarcinoma. Anti-HER3 antibody clone 10D1 was administered IP, biweekly at 500 μg per dose for a total of 10 doses. A control treatment group received an equal volume of PBS (vehicle).

(29) FIG. 28. Graph showing the results of analysis of tumour volume over time in an A549 cell-line derived mouse model of lung adenocarcinoma. Anti-HER3 antibody clone 4-35-B2 was administered IP, biweekly at 500 μg per dose for a total of 4 doses. A control treatment group received an equal volume of PBS (vehicle).

(30) FIG. 29. Graph showing the results of analysis of tumour volume over time in an ACHN cell-line derived mouse model of renal cell carcinoma. Anti-HER3 antibody clone 10D1 was administered IP, biweekly at 500 μg per dose for a total of 7 doses. A control treatment group received an equal volume of PBS (vehicle).

(31) FIG. 30. Histogram showing staining of cells by anti-HER3 antibody clone 10D1_c89 as determined by flow cytometry. Histograms show staining of HEK293 cells (which do not express HER3), or HEK293 HER3 overexpressing cells (HEK293 HER3 O/E).

(32) FIG. 31. Histogram showing staining of cells by anti-HER3 antibody clone 10D1_c90 as determined by flow cytometry. Histograms show staining of HEK293 cells (which do not express HER3), or HEK293 HER3 overexpressing cells (HEK293 HER3 O/E).

(33) FIG. 32. Histogram showing staining of cells by anti-HER3 antibody clone 10D1_c91 as determined by flow cytometry. Histograms show staining of HEK293 cells (which do not express HER3), or HEK293 HER3 overexpressing cells (HEK293 HER3 O/E).

(34) FIGS. 33A and 33B. Graphs showing the results of ELISA analysis of binding of anti-HER3 antibody 10D1 variant clones to human HER3. (33A) shows binding of anti-HER3 antibody clones 10D1 (referred to as 10D1P), 10D1_c75, 10D1_c76, 10D1_c77, 10D1_c78, 10D1_11B (referred to as v11b78L), 10D1_c85, 10D1_c85o1, 10D1_c85o2, 10D1_c87, 10D1_c89, 10D1_c90, 10D1_c91, 10D1_c93, LJM716 and hlgG (negative control). (33B) shows the same data as 33A, but for clones 10D1_c89, 10D1_c90, 10D1_c91 and LJM716 only.

(35) FIGS. 34A and 34B. Graphs showing the results of the analysis of recombinantly-expressed anti-HER3 antibody 10D1 variant clones by size exclusion chromatography. (34A) shows results for anti-HER3 antibody clones 10D1_c93, 10D1_c75, 10D1_c76, 10D1_c77, 10D1_c78, 10D1_11B (referred to as C78 v11b), 10D1_c85, 10D1_c85o1, 10D1_c85o2, 10D1_c89, 10D1_c90, 10D1_c91 and 10D1_c93. (34B) shows the same data as 33A, but for clones 10D1_c89, 10D1_c90, 10D1_c91 and only.

(36) FIGS. 35A to 35C. Graphs showing the results analysis of stability of anti-HER3 antibody 10D1 variant clones by Differential Scanning Fluorimetry analysis. (35A) shows results for anti-HER3 antibody clones LJM716 (also referred to as Elgemtumab), 10D1 (referred to as 10D1 (parental)), 10D1_c75, 10D1_c76, 10D1_c77 and 10D1_c78. (35B) shows results for 10D1_c85o2, 10D1_c87, 10D1_c89, 10D1_11B (referred to as c78_V11B), 10D1_c85 and 10D1_c85o1. (35C) shows results for 10D1_c90, 10D1_c91 and 10D1_c93.

(37) FIGS. 36A to 36M. Representative sensorgrams showing the results of analysis of affinity of anti-HER3 antibody 10D1 variant clones to human HER3. Kon, Koff and K.sub.D are shown. (36A) shows results for clone 10D1_c89, (36B) shows results for clone 10D1_c90, (36C) shows results for clone 10D1_c91, (36D) shows results for clone 10D1_c11B, (36E) shows results for clone 10D1_c85o2, (36F) shows results for clone 10D1_c87, (36G) shows results for clone 10D1_c93, (36H) shows results for clone 10D1_c76, (36I) shows results for clone 10D1_c77, (36J) shows results for clone 10D1_c78, (36K) shows results for clone 10D1_c75, (36L) shows results for clone 10D1_c85, and (36M) shows results for clone 10D1_c85o1.

(38) FIG. 37. Table summarizing properties of anti-HER3 antibody 10D1 variant clones relevant to safety and developability.

(39) FIGS. 38A and 38B. Bio-Layer Interferometry (38A) and thermostability (38B) analysis of Fc-modified anti-HER3 antibody clone 10D1 comprising GASDALIE and LCKC substitutions in CH2 region. (38A) BLI shows a representative sensorgram showing the results of analysis of affinity of binding to FcγRIIIa by Fc-modified anti-HER3 antibody clone 10D1. Kon, Koff and K.sub.D are shown.

(40) FIGS. 39A and 39B. Representative sensorgrams showing results of analysis of affinity of binding to FcγRIIIa by (39A) non-Fc-modified anti-HER3 antibody clone 10D1 and (39B) Fc-modified anti-HER3 antibody clone 10D1 comprising GASD substitutions in CH2 region. Kon, Koff and K.sub.D are shown.

(41) FIG. 40. Graph showing the results of analysis of stability of anti-HER3 antibody clone 10D1 GASD variant by Differential Scanning Fluorimetry analysis.

(42) FIGS. 41A and 41B. Tables showing the binding affinity for mouse and human Fc receptors of anti-HER3 antibody clones 10D1F.FcA and 10D1F.FcB (GASDALIE-LCKC variant) compared to silent variant N297Q, isoform variants, and commercially available antibodies. ND=K.sub.D Not Determined due to low binding affinity.

(43) FIGS. 42A and 42B. Histograms showing staining of cells by anti-HER3 antibodies as determined by flow cytometry. Histograms show staining of HEK293 cells (which do not express HER3), or HEK293 HER3 overexpressing cells (HEK293 HER3 O/E) by (42A) anti-HER3 antibody clone 10D1F.FcA and (42B) anti-HER3 antibody clones 10D1 and LJM-716.

(44) FIG. 43. Graph showing the results of ELISA analysis of binding of anti-HER3 antibody clone 10D1F.FcA to human EGFR (HER1) and human HER2. EC.sub.50 values are shown.

(45) FIG. 44. Histogram showing staining of cells by anti-HER3 and anti-HER4 antibodies as determined by flow cytometry. Histogram shows staining of HEK293 HER4 overexpressing cells by anti-HER3 antibody clone 10D1F.FcA, anti-HER3 antibodies LJM-716 and MM-121, and commercial anti-HER4 antibody.

(46) FIG. 45. Graph showing the results of ELISA analysis of binding of anti-HER3 antibody clone 10D1F.FcA to human, mouse, rat and cynomolgus macaque HER3. E050 values are shown.

(47) FIGS. 46A and 46B. Representative sensorgrams showing the results of analysis of affinity of binding of anti-HER3 antibody clones (46A) 10D1F.FcA and (46B) 10D1F.FcB to human HER3. Kon, Koff and K.sub.D are shown.

(48) FIGS. 47A and 47B. Graph showing the results of analysis of stability of anti-HER3 antibody clones (47A) 10D1F.FcA and (47B) 10D1F.FcB by Differential Scanning Fluorimetry analysis.

(49) FIGS. 48A and 48B. Graph showing the results of purity analysis of anti-HER3 antibody clones (48A) 10D1F.FcA and (48B) 10D1F.FcB by size exclusion chromatography.

(50) FIGS. 49A and 49B. Representative sensorgram and table showing the results of analysis of competition for binding to HER3 between anti-HER3 antibody clone 10D1F.FcA and anti-HER3 antibodies M-05-74 and M-08-11.

(51) FIG. 50. Graph and tables showing the results of pharmacokinetic analysis of anti-HER3 antibody clone 10D1 in mice.

(52) FIGS. 51A to 51F. Graphs showing the effect of anti-HER3 antibody clone 10D1 treatment on blood cell counts (51A), electrolyte indices (51B) and indices of hepatoxicity, nephrotoxicity and pancreatic toxicity (51C-51F) in mice. Left bars represent vehicle control, right bars represent 10D1 treatment. Dotted lines indicate the end points of the Charles River reference range. Indices of hepatoxicity, nephrotoxicity and pancreatic toxicity include alanine aminotransferase (ALT), aspartate transaminase (AST), blood urea nitrogen (BUN), creatinine (CREA), alkaline phosphatase (ALP), glucose (GLU), calcium (CAL), total bilirubin (BIL), total protein (TPR) and albumin (ALB).

(53) FIG. 52. Graph and table showing the results of analysis of the inhibition of interaction between HER2 and HER3 by anti-HER3 antibody clone 10D1F.FcA and antibodies MM-121, LJM-716 and Roche M05 as determined by ELISA.

(54) FIG. 53. Graph and table showing the results of analysis of the inhibition of interaction between EGFR and HER3 by anti-HER3 antibody clone 10D1F.FcA and antibodies MM-121 and LJM716, as determined by ELISA.

(55) FIG. 54. Graph and table showing the results of analysis of the ability of anti-HER3 antibody clones 10D1F.FcA (10D1F.A), 10D1F.FcB (10D1F.B), 10D1F-hIgG1 (N297Q) and anti-HER3 antibodies LJM-716 and Seribantumab (MM-121), to induce antibody-dependent cell-mediated cytotoxicity (ADCC). EC.sub.50 values are shown.

(56) FIGS. 55A to 55C. Images showing the results of analysis of the effect of anti-HER3 antibody treatment on HER3-mediated signalling in (55A) N87, (55B) FaDu and (55C) OvCar8 cells by phospho-western blot.

(57) FIGS. 56A and 56B. Graph and tables showing the results of pharmacokinetic analysis of anti-HER3 antibody clones (56A) 10D1F.FcA and (56B) 10D1F.FcB in mice. Parameters: maximum concentration (C.sub.max), T.sub.max, AUC (0-336 hr), AUC (0-infinity), Half-life (t.sub.1/2), Clearance (CL), Volume of distribution at steady state (V.sub.ss).

(58) FIGS. 57A to 57D. Graph and tables showing the results of pharmacokinetic analysis of anti-HER3 antibody clones 10D1F.FcA and 10D1F.FcB at (57A) 10 mg/kg, (57B) 25 mg/kg, (57C) 100 mg/kg and (57D) 250 mg/kg in rats. Parameters: maximum concentration (C.sub.max), T.sub.max, AUC (0-336 hr), AUC (0-infinity), Half-life (t.sub.1/2), Clearance (CL), Volume of distribution at steady state (V.sub.ss).

(59) FIGS. 58A to 58F. Graphs showing the effect of treatment of anti-HER3 antibody clone 10D1F.FcA or 10D1F.FcB at 200 ug (˜10 mg/kg), 500 ug (˜25 mg/kg), 2 mg (˜100 mg/kg), or 5 mg (˜250 mg/kg) on (58A, 58B) red blood cell indices, (58C) white blood cell indices, (58D) hepatotoxicity, (58E) kidney and pancreatic indices, and (58F) electrolyte indices.

(60) FIGS. 59A to 59D. Graphs showing the effect of treatment anti-HER3 antibody clone 10D1F.FcA on percentage tumour inhibition in in vitro mouse cancer models using N87 cells (gastric cancer), HCC95 cells (lung cancer), FaDu cells (head and neck cancer), SNU-16 cells (gastric cancer), A549 cells (lung cancer), OvCar8 cells (ovarian cancer), ACHN cells (kidney cancer) and HT29 cells (colorectal cancer) in comparison to (59A & 59B) anti-HER3 antibodies seribantumab and LJM-716 and (59C & 59D) EGFR family therapies cetuximab, trastuzumab and pertuzumab.

(61) FIG. 60. Graph showing the results of analysis of tumour volume over time in an A549 cell-line derived mouse model of lung adenocarcinoma after biweekly treatment with the indicated concentrations of antibodies for six weeks (n=6, vehicle control n=8). Antibody administration is indicated by triangles along x-axis.

(62) FIG. 61. Graph showing the results of analysis of tumour volume over time in a FaDu cell-line derived mouse model of head and neck squamous cell carcinoma after weekly treatment with the indicated concentrations of antibodies for six weeks (n=6). Antibody administration is indicated by triangles along x-axis.

(63) FIG. 62. Graph showing the results of analysis of tumour volume over time in a OvCAR8 cell-line derived mouse model of ovarian carcinoma after weekly treatment with the indicated concentrations of antibodies for six weeks (n=6). Antibody administration is indicated by triangles along x-axis.

(64) FIGS. 63A to 63D. Box plots showing the results of analysis of pathway activation by gene set enrichment analysis, for cancer cell lines treated with 10D1F.FcA, LJM-716 or seribantumab in in vitro phosphorylation assays. 63A shows the results obtained from N87 cells, 63B shows the results obtained from A549 cells, 63C shows the results obtained from OvCar8 cells and 63D shows the results obtained from FaDu cells.

(65) FIG. 64. Images showing the results of analysis of the effect of anti-HER3 antibody treatment on HER3-mediated signalling in A549 cells by phospho-western blot, at the indicated time points.

(66) FIG. 65. Graph showing the results of analysis of inhibiting of HER2:HER3 interaction by 10D1F.FcA or pertuzumab, as determined by PathHunter Pertuzumab Bioassay. IC50 (M) values are shown.

(67) FIGS. 66A to 66C. Histograms showing the results of analysis of expression of EGFR, HER2 and HER3 by (66A) BCPAP (66B) BHT101 and (66C) SW1736 cells. 1=unstained cells, 2=isotype control, 3=cetuximab, 4=trastuzumab, and 5=10D1F.FcA.

(68) FIGS. 67A to 67C. Graphs showing the results of analysis of the ability of different anti-ErbB antibodies to inhibit proliferation of BRAF.sup.V600E mutant thyroid cancer cell lines in vitro. 67A shows the results obtained for BHT101 cells, 687 shows the results obtained for BCPAP cells, and 67C shows the results obtained for SW1736 cells.

(69) FIGS. 68A to 68C. Graphs showing the results of analysis of the ability of 10D1F.FcA alone, or in combination with vemurafenib, to inhibit proliferation of BRAF.sup.V600E mutant thyroid cancer cell lines in vitro. 68A shows the results obtained for BHT101 cells, 68B shows the results obtained for SW1736 cells, and 68C shows the results obtained for BCPAP cells.

(70) FIGS. 69A to 69C. Tables showing representative hematological profiles of BALB/c mice following administration of 10 mg/kg, 25 mg/kg, 100 mg/kg or 250 mg/kg of 10D1F.FcA or an equal volume of PBS. 69A shows results of analysis of the red blood cell compartment, 69B shows results of analysis of the white blood cell compartment, and 69C shows results of analysis of correlates of liver, kidney and pancreas function, and levels of electrolytes. RBC=red blood cell, MVC=mean corpuscular volume, MCH=mean corpuscular haemoglobin, MCHC=mean corpuscular haemoglobin concentration, WBC=white blood cell, ALB=albumin, ALT=alanine aminotransferase, ALP=alkaline phosphatase, CREA=creatinine, BUN=blood urea nitrogen, GLU=glucagon, AMY=amylase, NA=sodium, K=potassium, P=phosphorus and CA=calcium.

(71) FIGS. 70A to 70C. Tables showing representative hematological profiles of SD rats at the indicated time points, following administration of 250 mg/kg 10D1F.FcA or an equal volume of PBS. 70A shows results of analysis of the red blood cell compartment, 70B shows results of analysis of the white blood cell compartment, and 70C shows results of analysis of correlates of liver, kidney and pancreas function, and levels of electrolytes. RBC=red blood cell, MVC=mean corpuscular volume, MCH=mean corpuscular haemoglobin, MCHC=mean corpuscular haemoglobin concentration, WBC=white blood cell, ALB=albumin, ALP=alkaline phosphatase, CREA=creatinine, BUN=blood urea nitrogen, GLU=glucagon, AMY=amylase, NA=sodium, K=potassium, P=phosphorus and CA=calcium.

(72) FIG. 71. Images showing the results of analysis of the effect of 10D1F.FcA treatment on HER3-mediated signalling in vivo in cells of FaDu or OvCar8 cell-derived tumors, as determined by phospho-western blot.

(73) FIG. 72. Box blot showing the results of analysis of internalisation of different anti-ErbB antibodies by the indicated cell lines.

(74) FIGS. 73A and 73B. Histograms and tables showing the results of analysis of internalisation of 10D1F.FcA or trastuzumab by the indicated cell lines at different time points, as determined by flow cytometry. 73B shows median fluorescence intensity and percentages of PE-positive cells determined from the histograms shown in 73A.

(75) FIG. 74. Graph showing the results of analysis of tumour volume over time in a N87 cell-line derived mouse model of gastric cancer after biweekly treatment with the indicated concentrations of the indicated anti-ErbB antibodies, for six weeks (n=6). Antibody administration is indicated by triangles along x-axis.

(76) FIGS. 75A and 75B. Images showing immunohistochemical staining of malignant and normal human tissues using 10D1F.FcA. 75A and 75B shows staining of different tissues.

(77) FIG. 76. Images showing immunohistochemical staining of A549 tumor xenograft cryosections by 10D1F or a rabbit polyclonal anti-HER3 antibody, at the indicated magnifications. Secondary-only control stainings are shown.

(78) FIGS. 77A and 77B. Bar charts showing the results of analysis of the ability of the indicated anti-ErbB antibodies to inhibit in vitro proliferation of the indicated cancer cell lines at the serum concentrations the antibodies reach at C.sub.max following IP administration to mice at 25 mg/kg. 77A and 77B show results obtained using different cell lines.

(79) FIGS. 78A to 78F. Representative sensorgrams showing results of analysis of affinity of binding to human HER3 by anti-HER3 antibodies 10D1F.A (78A, 78B), MM-121 (78C, 78D) and LJM-716 (78E, 78F), in the absence of human NRG1 (78A, 78C, 78E), and in the presence of human NRG1 (78B, 78D, 78F). Kon, Koff and K.sub.D) are shown.

(80) FIG. 79. Graph showing the results of analysis of tumour volume over time in a patient-derived xenograft model of ovarian cancer comprising CLU-NRG1 fusion, after biweekly treatment with 25 mg/kg of 10D1F or isotype-matched control antibody (n=10 per treatment group). Antibody administration is indicated by triangles along x-axis.

EXAMPLES

(81) In the following Examples, the inventors describe the generation of novel anti-HER3 antibody clones targeted to specific regions of interest in the HER3 molecule, and the biophysical and functional characterisation and therapeutic evaluation of these antigen-binding molecules.

Example 1: HER3 Target Design and Anti-HER3 Antibody Hybridoma Production

(82) The inventors selected two regions in the extracellular region of human HER3 (SEQ ID NO:9) for raising HER3-binding monoclonal antibodies.

1.1 Hybridoma Production

(83) Approximately 6 week old female BALB/c mice were obtained from InVivos (Singapore). Animals were housed under specific pathogen-free conditions and were treated in compliance with the Institutional Animal Care and Use Committee (IACUC) guidelines.

(84) For hybridoma production, mice were immunized with proprietary mixtures of antigenic peptide, recombinant target protein or cells expressing the target protein.

(85) Prior to harvesting the spleen for fusion, mice were either boosted with antigen mixture for three consecutive days or only for a single day. 24 h after the final boost total splenocytes were isolated and fused with the myeloma cell line P3X63.Ag8.653 (ATCC, USA), with PEG using ClonaCell-HY Hybridoma Cloning Kit, in accordance with the manufacturer's instructions (Stemcell Technologies, Canada).

(86) Fused cells were cultured in ClonaCell-HY Medium C (Stemcell Technologies, Canada) overnight at 37° C. in a 5% CO.sub.2 incubator. The next day, fused cells were centrifuged and resuspended in 10 ml of ClonaCell-HY Medium C and then gently mixed with 90 ml of semisolid methylcellulose-based ClonaCell-HY Medium D (StemCell Technologies, Canada) containing HAT components, which combines the hybridoma selection and cloning into one step.

(87) The fused cells were then plated into 96 well plates and allowed to grow at 37° C. in a 5% CO.sub.2 incubator. After 7-10 days, single hybridoma clones were isolated and antibody producing hybridomas were selected by screening the supernatants by Enzyme-linked immunosorbent assay (ELISA) and Fluorescence-activated cell sorting (FACs).

1.2 Antibody Variable Region Amplification and Sequencing

(88) Total RNA was extracted from hybridoma cells using TRIzol reagent (Life Technologies, Inc., USA) using manufacturer's protocol. Double-stranded cDNA was synthesized using SMARTer RACE 573′ Kit (Clontech™, USA) in accordance with the manufacturer's instructions. Briefly, 1 μg total RNA was used to generate full-length cDNA using 5′-RACE CDS primer (provided in the kit), and the 5′ adaptor (SMARTer II A primer) was then incorporated into each cDNA according to manufacturer's instructions. cDNA synthesis reactions contained: 5× First-Strand Buffer, DTT (20 mM), dNTP Mix (10 mM), RNase Inhibitor (40 U/μl) and SMARTScribe Reverse Transcriptase (100 U/μl).

(89) The race-ready cDNAs were amplified using SeqAmp DNA Polymerase (Clontech™, USA). Amplification reactions contained SeqAmp DNA Polymerase, 2× Seq AMP buffer, 5′ universal primer provided in the 5′ SMARTer Race kit, that is complement to the adaptor sequence, and 3′ primers that anneal to respective heavy chain or light chain constant region primer. The 5′ constant region were designed based on previously reported primer mix either by Krebber et al. J. Immunol. Methods 1997; 201: 35-55, Wang et al. Journal of Immunological Methods 2000, 233; 167-177 or Tiller et al. Journal of Immunological Methods 2009; 350:183-193. The following thermal protocol was used: pre-denature cycle at 94° C. for 1 min; 35 cycles of 94° C., 30 s, 55° C., 30 s and 72° C., 45 s; final extension at 72° C. for 3 min.

(90) The resulting VH and VL PCR products, approximately 550 bp, were cloned into pJET1.2/blunt vector using CloneJET PCR Cloning Kit (Thermo Scientific, USA) and used to transform highly competent E. coli DH5α. From the resulting transformants, plasmid DNA was prepared using Miniprep Kit (Qiagene, Germany) and sequenced. DNA sequencing was carried out by AITbiotech. These sequencing data were analyzed using the international IMGT (ImMunoGeneTics) information system (LeFranc et al, Nucleic Acids Res. (2015) 43 (Database issue): D413-22) to characterize the individual CDRs and framework sequences. The signal peptide at 5′ end of the VH and VL was identified by SignalP (v 4.1; Nielsen, in Kihara, D (ed): Protein Function Prediction (Methods in Molecular Biology vol. 1611) 59-73, Springer 2017).

(91) Four monoclonal anti-HER3 antibody clones were selected for further development: 10D1, 10A6, 4-35-B2, and 4-35-B4.

(92) Humanised versions of 10D1 were designed in silico by grafting of complementarity determining regions (CDRs) into VH and VL comprising human antibody framework regions, and were further optimized for antigen binding by yeast display method.

(93) For yeast display, humanized sequences were converted into single-chain fragment variable (scFv) format by polymerase chain reaction (PCR) and used as templates to generate mutant libraries by random mutagenesis. Mutant PCR libraries were then electroporated into yeast together with linearized pCTcon2 vector to generate yeast libraries. The libraries were stained with human HER3 antigen and sorted for top binders. After 4-5 rounds of sorting, individual yeast clones were sequenced to identify unique antibody sequences.

(94) TABLE-US-00003 Antibody clone VH/VL sequence 10A6  VH = SEQ ID NO: 157  VL = SEQ ID NO: 164 10D1 VH = SEQ ID NO: 24 VL = SEQ ID NO: 74 10D1_c75 VH = SEQ ID NO: 25 VL = SEQ ID NO: 75 10D1_c76 VH = SEQ ID NO: 26 VL = SEQ ID NO: 76 10D1_c77 VH = SEQ ID NO: 27 VL = SEQ ID NO: 77 10D1_c78v1 VH = SEQ ID NO: 28 VL = SEQ ID NO: 78 10D1_c78v2 VH = SEQ ID NO: 29 VL = SEQ ID NO: 78 10D1_11B VH = SEQ ID NO: 30 VL = SEQ ID NO: 78 10D1_c85v1 VH = SEQ ID NO: 31 VL = SEQ ID NO: 79 10D1_c85v2 VH = SEQ ID NO: 32 VL = SEQ ID NO: 79 10D1_c85o1 VH = SEQ ID NO: 33 VL = SEQ ID NO: 80 10D1_c85o2 VH = SEQ ID NO: 34 VL = SEQ ID NO: 81 10D1_c87 VH = SEQ ID NO: 35 VL = SEQ ID NO: 82 10D1_c89 VH = SEQ ID NO: 36 VL = SEQ ID NO: 83 10D1_c90 VH = SEQ ID NO: 37 VL = SEQ ID NO: 84 10D1_c91 VH = SEQ ID NO: 38 VL = SEQ ID NO: 85 10D1_c92 VH = SEQ ID NO: 39 VL = SEQ ID NO: 86 10D1_c93 VH = SEQ ID NO: 40 VL = SEQ ID NO: 87 4-35-B2  VH = SEQ ID NO: 127  VL = SEQ ID NO: 135 4-35-84  VH = SEQ ID NO: 143  VL = SEQ ID NO: 150

Example 2: Antibody Production and Purification

2.1 Cloning VH and VL into Expression Vectors

(95) DNA sequences encoding the heavy and light chain variable regions of the anti-HER3 antibody clones were subcloned into the pmAbDZ_IgG1_CH and pmAbDZ_IgG1_CL (InvivoGen, USA) eukaryotic expression vectors for construction of human-mouse chimeric antibodies.

(96) Alternatively, DNA sequence encoding the heavy and light chain variable regions of the anti-HER3 antibody clones were subcloned into the pFUSE-CHIg-hG1 and pFUSE2ss-CLIg-hk (InvivoGen, USA) eukaryotic expression vectors for construction of human-mouse chimeric antibodies. Human IgG1 constant region encoded by pFUSE-CHIg-hG1 comprises the substitutions D356E, L358M (positions numbered according to EU numbering) in the CH3 region relative to Human IgG1 constant region (IGHG1; UniProt: P01857-1, v1; SEQ ID NO:176). pFUSE2ss-CLIg-hk encodes human IgG1 light chain kappa constant region (IGCK; UniProt: P01834-1, v2).

(97) Variable regions along with the signal peptides were amplified from the cloning vector using SeqAmp enzyme (Clontech™, USA) following the manufacturer's protocol. Forward and reverse primers having 15-20 bp overlap with the appropriate regions within VH or VL plus 6 bp at 5′ end as restriction sites were used. The DNA insert and the vector were digested with restriction enzyme recommended by the manufacturer to ensure no frameshift was introduced and ligated into its respective plasmid using T4 ligase enzyme (Thermo Scientific, USA). The molar ratio of 3:1 of DNA insert to vector was used for ligation.

2.2 Expression of Antibodies in Mammalian Cells

(98) Antibodies were expressed using either 1) Expi293 Transient Expression System Kit (Life Technologies, USA), or 2) HEK293-6E Transient Expression System (CNRC-NRC, Canada) following the manufacturer's instructions.

(99) 1) Expi293 Transient Expression System

(100) Cell Line Maintenance:

(101) HEK293F cells (Expi293F) were obtained from Life Technologies, Inc (USA). Cells were cultured in serum-free, protein-free, chemically defined medium (Expi293 Expression Medium, Thermo Fisher, USA), supplemented with 50 IU/ml penicillin and 50 μg/ml streptomycine (Gibco, USA) at 37° C., in 8% CO.sub.2 and 80% humidified incubators with shaking platform.

(102) Transfection:

(103) Expi293F cells were transfected with expression plasmids using ExpiFectamine 293 Reagent kit (Gibco, USA) according to its manufacturer's protocol. Briefly, cells at maintenance were subjected to a media exchange to remove antibiotics by spinning down the culture, cell pellets were re-suspended in fresh media without antibiotics at 1 day before transfection. On the day of transfection, 2.5×10.sup.6/ml of viable cells were seeded in shaker flasks for each transfection. DNA-ExpiFectamine complexes were formed in serum-reduced medium, Opti-MEM (Gibco, USA), for 25 min at room temperature before being added to the cells. Enhancers were added to the transfected cells at 16-18 h post transfection. An equal amount of media was topped up to the transfectants at day 4 post-transfection to prevent cell aggregation. Transfectants were harvested at day 7 by centrifugation at 4000×g for 15 min, and filtered through 0.22 μm sterile filter units.

(104) 2) HEK293-6E Transient Expression System

(105) Cell Line Maintenance:

(106) HEK293-6E cells were obtained from National Research Council Canada. Cells were cultured in serum-free, protein-free, chemically defined Freestyle F17 Medium (Invitrogen, USA), supplemented with 0.1% Kolliphor-P188 and 4 mM L-Glutamine (Gibco, USA) and 25 μg/ml G-418 at 37° C., in 5% CO.sub.2 and 80% humidified incubators with shaking platform.

(107) Transfection:

(108) HEK293-6E cells were transfected with expression plasmids using PEIpro™ (Polyplus, USA) according to its manufacturer's protocol. Briefly, cells at maintenance were subjected to a media exchange to remove antibiotics by centrifugation, cell pellets were re-suspended with fresh media without antibiotics at 1 day before transfection. On the day of transfection, 1.5-2×10.sup.6 cells/ml of viable cells were seeded in shaker flasks for each transfection. DNA and PEIpro™ were mixed to a ratio of 1:1 and the complexes were allowed to form in F17 medium for 5 min at RT before adding to the cells. 0.5% (w/v) of Tryptone N1 was fed to transfectants at 24-48 h post transfection. Transfectants were harvested at day 6-7 by centrifugation at 4000×g for 15 min and the supernatant was filtered through 0.22 μm sterile filter units.

(109) Cells were transfected with vectors encoding the following combinations of polypeptides:

(110) TABLE-US-00004 Antigen- biding molecule Polypeptides Antibody  [1] 10D1 VH-CH1-CH2-CH3 (SEQ ID NO: 216) + anti-HER3 clone 10D1 IgG1 10D1 VL-C.sub.K (SEQ ID NO: 217)  [2] 10A6 VH-CH1-CH2-CH3 (SEQ ID NO: 222) + anti-HER3 clone 10A6 IgG1 10A6 VL-C.sub.K (SEQ ID NO: 223)  [3] 4-35-B2 VH-CH1-CH2-CH3 (SEQ ID NO: 218) + anti-HER3 clone 4-35-82 IgG1 4-35-B2 VL-C.sub.K (SEQ ID NO: 219)  [4] 4-35-84 VH-CH1-CH2-CH3 (SEQ ID NO: 220) + anti-HER3 clone 4-35-84 IgG1 4-35-84 VL-C.sub.K (SEQ ID NO: 221)  [5] 10D1_c75 VH-CH1-CH2-CH3 (SEQ ID NO: 187) + anti-HER3 clone 10D1_c75 IgG1 10D1_c75 VL-C.sub.K (SEQ ID NO: 188)  [6] 10D1_c76 VH-CH1-CH2-CH3 (SEQ ID NO: 189) + anti-HER3 clone 10D1_c76 IgG1 10D1_c76 VL-C.sub.K (SEQ ID NO: 190)  [7] 10D1_c77 VH-CH1-CH2-CH3 (SEQ ID NO: 191) + anti-HER3 clone 10D1_c77 IgG1 10D1_c77 VL-C.sub.K (SEQ ID NO: 192)  [8] 10D1_c78v1 VH-CH1-CH2-CH3 (SEQ ID NO: 19 ) + anti-HER3 clone 10D1_c78v1 IgG1 10D1_c78v1 VL-C.sub.K (SEQ ID NO: 195)  [9] 10D1_c78v1 VH-CH1-CH2-CH3 (SEQ ID NO: 194) + anti-HER3 clone 10D1_c78v2 IgG1 10D1_c78v2 VL-C.sub.K (SEQ ID NO: 195) [10] 10D1_11B VH-CH1-CH2-CH3 (SEQ ID NO: 196) + anti-HER3 clone 10D1_11B IgG1 10D1_11B VL-C.sub.K (SEQ ID NO: 195) [11] 10D1_c85v1 VH-CH1-CH2-CH3 (SEQ ID NO: 197) + anti-HER3 clone 10D1_c85v1 IgG1 10D1_c85v1 VL-C.sub.K (SEQ ID NO: 199) [12] 10D1_c85v2 VH-CH1-CH2-CH3 (SEQ ID NO: 198) + anti-HER3 clone 10D1_c85v2 IgGl 10D1_c85v2 VL-C.sub.K (SEQ ID NO: 199) [13] 10D1_c85o1 VH-CH1-CH2-CH3 (SEQ ID NO: 200) + anti-HER3 clone 10D1_c85o1 IgG1 10D1_c8501 VL-C.sub.K (SEQ ID NO: 201) [14] 10D1_c85o2 VH-CH1-CH2-CH3 (SEQ ID NO: 202) + anti-HER3 clone 10D1_c85o2 IgG1 10D1_c85o2 VL-C.sub.K (SEQ ID NO: 203) [15] 10D1_c87 VH-CH1-CH2-CH3 (SEQ ID NO: 204) + anti-HERS clone 10D1_c87 IgG1 10D1_c87 VL-C.sub.K (SEQ ID NO: 205) [16] 10D1_c89 VH-CH1-CH2-CH3 (SEQ ID NO: 206) + anti-HER3 clone 10D1_c89 IgG1 10D1_c89 VL-C.sub.K (SEQ ID NO: 207) [17] 10D1_c90 VH-CH1-CH2-CH3 (SEQ ID NO: 208) + anti-HER3 clone 10D1_c90 IgG1 10D1_c90 VL-C.sub.K (SEQ ID NO: 209) [18] 10D1_c91 VH-CH1-CH2-CH3 (SEQ ID NO: 210) + anti-HER3 clone 10D1_c91 IgG1 10D1_c91 VL-C.sub.K (SEQ ID NO: 211) [19] 10D1_c92 VH-CH1-CH2-CH3 (SEQ ID NO: 212) + anti-HER3 clone 10D1_c92 IgG1 10D1_c92 VL-C.sub.K (SEQ ID NO: 213) [20] 10D1_c93 VH-CH1-CH2-CH3 (SEQ ID NO: 214) + anti-HER3 clone 10D1_c93 IgG1 10D1_c93 VL-C.sub.K (SEQ ID NO: 215)

2.3 Antibody Purification

(111) Affinity Purification, Buffer Exchange and Storage:

(112) Antibodies secreted by the transfected cells into the culture supernatant were purified using liquid chromatography system AKTA Start (GE Healthcare, UK). Specifically, supernatants were loaded onto HiTrap Protein G column (GE Healthcare, UK) at a binding rate of 5 ml/min, followed by washing the column with 10 column volumes of washing buffer (20 mM sodium phosphate, pH 7.0). Bound mAbs were eluted with elution buffer (0.1 M glycine, pH 2.7) and the eluents were fractionated to collection tubes which contain appropriate amount of neutralization buffer (1 M Tris, pH 9). Neutralised elution buffer containing purified mAb were exchanged into PBS using 30K MWCO protein concentrators (Thermo Fisher, USA) or 3.5K MWCO dialysis cassettes (Thermo Fisher, USA). Monoclonal antibodies were sterilized by passing through 0.22 μm filter, aliquoted and snap-frozen in −80° C. for storage.

2.4 Antibody-Purity Analysis

(113) Size Exclusion Chromatography (SEC):

(114) Antibody purity was analysed by size exclusion chromatography (SEC) using Superdex 200 10/30 GL columns (GE Healthcare, UK) in PBS running buffer, on a AKTA Explorer liquid chromatography system (GE Healthcare, UK). 150 μg of antibody in 500 μl PBS pH 7.2 was injected to the column at a flow rate of 0.75 ml/min at room temperature. Proteins were eluted according to their molecular weights.

(115) The result for anti-HER3 antibody clone 10D1 ([1] of Example 2.2) is shown in FIG. 12.

(116) The results obtained for the different 10D1 variant clones are shown in FIG. 34.

(117) Sodium-Dodecyl Sulfate Polyacrylamide Gel Electrophoresis (SDS-PAGE):

(118) Antibody purity was also analysed by SDS-PAGE under reducing and non-reducing conditions according to standard methods. Briefly, 4%-20% TGX protein gels (Bio-Rad, USA) were used to resolve proteins using a Mini-Protean Electrophoresis System (Bio-Rad, USA). For non-reducing condition, protein samples were denatured by mixing with 2× Laemmli sample buffer (Bio-Rad, USA) and boiled at 95° C. for 5-10 min before loading to the gel. For reducing conditions, 2× sample buffer containing 5% of β-mercaptoethanol (βmE), or 40 mM DTT (dithiothreitol) was used. Electrophoresis was carried out at a constant voltage of 150V for 1 h in SDS running buffer (25 mM Tris, 192 mM glycine, 1% SDS, pH 8.3).

(119) Western Blot:

(120) Protein samples (30 μg) were fractionated by SDS-PAGE as described above and transferred to nitrocellulose membranes. Membranes were then blocked and immunoblotted with antibodies overnight at 4° C. After washing three times in PBS-Tween the membranes were then incubated for 1 h at room temperature with horseradish peroxidase (HRP)-conjugated secondary antibodies. The results were visualized via a chemiluminescent Pierce ECL Substrate Western blot detection system (Thermo Scientific, USA) and exposure to autoradiography film (Kodak XAR film).

(121) The primary antibodies used for detection were goat anti-human IgG-HRP (GenScript Cat No. A00166) and goat anti-human kappa-HRP (SouterhnBiotech Cat No. 2060-05).

(122) The result for anti-HER3 antibody clone 10D1 ([1] of Example 2.2) is shown in FIG. 13. 10D1 was easily expressed, purified and processed at high concentrations.

Example 3: Biophysical Characterisation

3.1 Analysis of Cell Surface Antigen-Binding by Flow Cytometry

(123) Wildtype HEK293 cells (which do not express high levels of HER3) and HEK293 cells transfected with vector encoding human HER3 (i.e. HEK 293 HER O/E cells) were incubated with 20 μg/ml of anti-HER3 antibody or isotype control antibody at 4° C. for 1.5 hr. The anti-HER3 antibody clone LJM716 (described e.g. in Garner et al., Cancer Res (2013) 73: 6024-6035) was included in the analysis as a positive control.

(124) The cells were washed thrice with FACS buffer (PBS with 5 mM EDTA and 0.5% BSA) and resuspended in FITC-conjugated anti-FC antibody (Invitrogen, USA) for 40 min at 2-8° C. Cells were washed again and resuspended in 200 μL of FACS flow buffer (PBS with 5 mM EDTA) for flow cytometric analysis using MACSQuant 10 (Miltenyi Biotec, Germany). After acquisition, all raw data were analyzed using Flowlogic software. Cells were gated using forward and side scatter profile percentage of positive cells was determined for native and overexpressing cell populations.

(125) The results are shown in FIGS. 1 to 4 and 30 to 32. The anti-HER3 antibodies were shown to bind to human HER3 with high specificity. 10D1 and LJM716 were shown to bind to human HER3-expressing cells to a similar extent.

3.2 ELISAs for Determining Antibody Specificity and Cross-Reactivity

(126) ELISAs were used to determine the binding specificity of the antibodies. The antibodies were analysed for binding to human HER3 polypeptide, as well as respective mouse, rat and monkey homologues of HER3 (Sino Biological Inc., China). The antibodies were also analysed for their ability to bind to human EGFR and human HER2 (Sino Biological Inc., China).

(127) ELISAs were carried out according to standard protocols. Briefly, 96-well plates (Nunc, Denmark) were coated with 0.1 μg/ml of target polypeptide in phosphate-buffered saline (PBS) for 16 h at 4° C. After blocking for 1 h with 1% BSA in Tris buffer saline (TBS) at room temperature, the anti-HER3 antibody was serially diluted with the highest concentration being 10 μg/ml, and added to the plate. Post 1 h incubation at room temperature, plates were washed three times with TBS containing 0.05% Tween 20 (TBS-T) and were then incubated with a HRP-conjugated anti-His antibody (Life Technologies, Inc., USA) for 1 h at room temperature. After washing, plates were developed with colorimetric detection substrate 3,3′,5,5′-tetramethylbenzidine (Turbo-TMB; Pierce, USA) for 10 min. The reaction was stopped with 2M H.sub.2SO.sub.4, and OD was measured at 450 nM.

(128) The results of the ELISAs are shown in FIGS. 5 to 7 and FIG. 33.

(129) Anti-HER3 antibody clone 10D1 was found not to bind to human HER2 or human EGFR even at high concentrations of the antibody (FIG. 5A). Anti-HER3 antibody clone 10D1 was also found to display substantial cross-reactivity with mouse HER3, rat HER3 and cynomolgus macaque HER3 (FIG. 5B).

(130) Anti-HER3 antibody clone 4-35-B2 was found to bind to human HER2 and human EGFR (FIG. 6A). Anti-HER3 antibody clone 4-35-B2 also displayed substantial cross-reactivity with mouse HER3, rat HER3 and cynomolgus macaque HER3 (FIG. 6B).

(131) Anti-HER3 antibody clone 4-35-B4 was found to bind to human HER2 and human EGFR (FIG. 7A). Anti-HER3 antibody clone 4-35-B4 also displayed substantial cross-reactivity with mouse HER3, rat HER3 and cynomolgus macaque HER3 (FIG. 7B).

(132) All of the 10D1 variants were demonstrated to bind to human HER3 (FIGS. 33A and 33B).

3.3 Global Affinity Study Using Octet QK384 System

(133) The anti-HER3 antibody clones in IgG1 format were analysed for binding affinity to human HER3.

(134) Bio-Layer Interferometry (BLI) experiments were performed using the Octet QK384 system (ForteBio). anti-Human IgG Capture (AHC) Octet sensor tips (Pall ForteBio, USA) were used to anti-HER3 antibodies (25 nM). All measurements were performed at 25° C. with agitation at 1000 rpm. Kinetic measurements for antigen binding were performed by loading His-tagged human HER3 antigens at different concentrations for 120 s, followed by a 120 s dissociation time by transferring the biosensors into assay buffer containing wells. Sensograms were referenced for buffer effects and then fitted using the Octet QK384 user software (Pall ForteBio, USA). Kinetic responses were subjected to a global fitting using a one site binding model to obtain values for association (K.sub.on), dissociation (K.sub.off) rate constants and the equilibrium dissociation constant (K.sub.D). Only curves that could be reliably fitted with the software (R.sup.2>0.90) were included in the analysis.

(135) A representative sensorgram for the analysis of clone 10D1 is shown in FIG. 8. Clone 10D1 was found to bind to human HER3 with an affinity of K.sub.D=9.58 nM.

(136) The humanised/optimized 10D1 variants bind to human HER3 with very high affinity. Representative sensorgrams are shown in FIGS. 36A to 36M.

(137) The affinities determined for 10D1 clone variants are shown below:

(138) TABLE-US-00005 Antibody Affinity clone (KD) 10D1_c89 72.6 pM 10D1_c90 <1 pM 10D1_c91 176 pM 10D1_11B 0.41 nM 10D1_c85o 17.3 nM 10D1_c87 <1 pM 10D1_c93 <1 pM 10D1_c76 <1 pM 10D1_c77 1.93 nM 10D1_c78 <1 pM 10D1_c75 <1 pM 10D1_c85 7.58 nM 10D1_c85o1 18.2 nM

(139) Clone 4-35-B2 was found to bind to human HER3 with an affinity of K.sub.D=80.9 nM (FIG. 9), and clone 4-35-B4 was found to bind to human HER3 with an affinity of K.sub.D=50.3 nM (FIG. 10).

3.4 Analysis of Thermostability by Differential Scanning Fluorimetry

(140) Briefly, triplicate reaction mixes of antibodies at 0.2 mg/mL and SYPRO Orange dye (ThermoFisher) were prepared in 25 μL of PBS, transferred to wells of MicroAmp Optical 96-Well Reaction Plates (ThermoFisher), and sealed with MicroAmp Optical Adhesive Film (ThermoFisher). Melting curves were run in a 7500 fast Real-Time PCR system (Applied Biosystems) selecting TAMRA as reporter and ROX as passive reference. The thermal profile included an initial step of 2 min at 25° C. and a final step of 2 min at 99° C., with a ramp rate of 1.2%. The first derivative of the raw data was plotted as a function of temperature to obtain the derivative melting curves. Melting temperatures (Tm) of the antibodies were extracted from the peaks of the derivative curves.

(141) The first derivative of the raw data obtained for Differential Scanning Fluorimetry analysis of the thermostability of antibody clone 10D1 is shown in FIG. 11. Three different samples of the antibody were analysed. The Tm was determined to be 70.3° C.

(142) The analysis was also performed for the 10D1 variant clones and LJM716. The first derivative of the raw data and the determined Tms are shown in FIGS. 35A to 35C.

3.5 Analysis of Anti-HER3 Antibody 10D1 Epitope

(143) Anti-HER3 antibody 10D1 was analysed to determine whether it competes with anti-HER3 antibodies MM-121 and/or LJM-716 for binding to HER3. The epitope for MM-121 has been mapped in domain I of HER3; it blocks the NRG ligand binding site. The epitope for LJM-716 has been mapped to conformational epitope distributed across domains II and IV, and it locks HER3 in an inactive conformation.

(144) Bio-Layer Interferometry (BLI) experiments were performed using the Octet QK384 system (ForteBio). anti-Penta-HIS (HIS1K) coated biosensor tips (ForteBio, USA) were used to capture His-tagged human HER3 (75 nM; 300 s). Binding by saturating antibody (400 nm; 600 s) was detected, followed by a dissociation step (120 s), followed by detection of binding with competing antibody (300 nM; 300 s), followed by a dissociation step (120 s). The variable region of MM-121 antibody was cloned in the PDZ vector having human IgG2 and IgKappa Fc backbone. The variable region of LJM-716 antibody was cloned in the PDZ vector having human IgG1 and IgKappa Fc backbone.

(145) The results of the analysis are shown in FIGS. 14A and 14B. Anti-HER3 antibody was found not to compete with MM-121 and/or LJM-716 for binding to HER3.

(146) 10D1 was found to bind a distinct and topologically distant epitope of HER3 than MM-121 and/or LJM-716.

(147) The epitope for 10D1 was mapped using overlapping 15-mer amino acids to cover the entire HER3 extracellular domain. Each unique 15-mer was elongated by a GS linker at C and N-terminals, conjugated to a unique well in 384 well plates, and the plates were incubated with 0.1, 1, 10 and 100 ug/ml of 10D1 antibody for 16 hrs at 4° C. The plates were washed and then incubated for 1 hr at 20° C. with POD-conjugated goat anti-human IgG. Finally POD substrate solution was added to the wells for 20 min. before binding was assessed by measurement of chemiluminescence at 425 nm using a LI-COR Odyssey Imaging System, and quantification and analysis was performed using the PepSlide Analyzer software package. The experiment was performed in duplicate.

(148) The 10D1 epitope was found not to be directly located at a β-hairpin structure of the HER3 dimerisation arm located at domain II, but instead at a dimerisation interface N-terminal to the β-hairpin.

(149) The site of HER3 to which 10D1 and 10D1-derived clones was determined to bind corresponds to positions 218 to 235 of the amino acid sequence of human HER3 (as shown e.g. in SEQ ID NO:1); the amino acid sequence for this region of HER3 is shown in SEQ ID NO:229. Within this region, two consensus binding site motifs were identified, and are shown in SEQ ID NOs:230 and 231.

(150) Binding to this location of HER3 acts to impede HER family heterodimerisation and consequent downstream signalling pathways (see Example 4). Binding is ligand (NRG) independent. The 10D1 binding site is solvent accessible in both the open and closed HER3 conformations, is not conserved between HER3 and other HER family members, and is 100% conserved between human, mouse, rat and monkey HER3 orthologs.

Example 4: Functional Characterisation

4.1 Inhibition of Dimerisation of HER2 and HER3

(151) The anti-HER3 antibodies were analysed for their ability to inhibit heterodimerisation of HER3 and HER2.

(152) Briefly, 96-well plates (Nunc, Denmark) were coated with 0.1 μg/ml His-tagged HER2 protein in PBS for 16 h at 4° C. After blocking for 1 h with 1% BSA in PBS at room temperature, recombinant biotinylated human HER3 protein was added in the presence of different concentrations of anti-HER3 antibody clone 10D1, and plates were incubated for 1 h at room temperature. Plates were subsequently washed three times, and then incubated with HRP-conjugated secondary antibody for 1 h at room temperature. After washing, plates were developed with colorimetric detection substrate 3,3′,5,5′-tetramethylbenzidine (Turbo-TMB; Pierce, USA) for 10 min. The reaction was stopped with 2M H.sub.2SO.sub.4, and OD was measured at 450 nM.

(153) The results are shown in FIG. 15. Anti-HER3 antibody clone 10D1 was found to inhibit interaction between HER2 and HER3 in a dose-dependent fashion.

(154) In further experiments, inhibition of HER2:HER3 dimerisation was analysed

(155) In further experiments, inhibition of HER2:HER3 dimerisation was evaluated using the PathHunter Pertuzumab Bioassay Kit (DiscoverX) according to the manufacturer's instructions.

(156) Briefly, HER2 and HER3 overexpressing U2OS cells were thawed using 1 ml of pre-warmed CP5 media and 5,000 cells were seeded per well and cultured at 37° C. in 5% CO.sub.2 atmosphere for 4 hr. Cells were then treated with an 8-point serial dilution of 10D1F.FcA or Pertuzumab, starting from 25 μg/ml.

(157) After 4 hr incubation, 30 ng/ml of heregulin-β2 was added to each well and the cells were incubated for a further 16 hr. 10 μL of PathHunter bioassay detection reagent 1 was added to wells, and incubated for 15 min at room temperature in the dark. This was followed by addition of 40 μL PathHunter bioassay detection reagent 2, and incubation for 60 min at room temperature in the dark. Plates were then read using Synergy4 Biotek with 1 second delay.

(158) The results are shown in FIG. 65. 10D1F.FcA was found inhibit HER2:HER3 dimerisation with greater efficiency than pertuzumab, as reflected by its lower IC50.

4.2 Identification of Cancer Cell Lines for Analysis

(159) The inventors characterised expression of EGFR protein family members by cancer cell lines to identify appropriate cells to investigate inhibition of HER3.

(160) FIG. 16A shows mRNA expression data EGFR family members and ligands by N87, SNU16, HT29, FaDu, A549, HCC95, OvCAR8 and AHCN cells according to the Cancer Cell Line Encyclopaedia (CCLE; Barretina et al., Nature (2012) 483: 603-607 and The Cancer Cell Line Encyclopedia Consortium & The Genomics of Drug Sensitivity in Cancer Consortium, Nature (2015) 528: 84-87). FIG. 16A also shows protein expression data for EGFR, HER2 and HER3 as determined by FlowLogic.

(161) Cell lines used in the experiments were purchased from ATCC and cultured as recommended. Briefly, cell lines maintained in the indicated cell culture medium, supplemented with 10% FBS and 1% Pen/Strep. Cells were cultured at 37° C., in 5% CO.sub.2 incubators. Cultured cells were plated at the appropriate seeding density in a 96 well plate: HT29, HCC95, FADU and OvCar8 cells were seeded at 2000 cells/well, NCI-N87 cells were seeded at 5000 cells/well, SNU-16, ACHN and cells were seeded at 1500 cells/well, and A549 cells were seeded at 1200 cells/well.

(162) FIG. 16B shows surface expression of EGFR, HER2 and HER3 as determined by flow cytometry. Briefly, 500,000 cells were stained in staining buffer containing 0.5% BSA and 2 mM EDTA with primary antibodies (20 μg/ml) for 1.5 h at 4° C. The secondary Antibody used was anti-human Alexafluor488 at 10 μg/ml for 20 min at 4° C.

4.3 Inhibition of HER3-Mediated Signalling

(163) Anti-HER3 antibody 10D1 was analysed for its ability to inhibit HER-3 mediated signalling in vitro.

(164) Briefly, N87 and FaDu cells were seeded in wells of a 6 well plate with 10% serum at 37° C., 5% CO.sub.2. After 16 hrs, cells were starved by culture overnight in 1% FBS cell culture medium (to reduce signalling elicited by growth factors in the serum). On the following day cells were treated with 50 μg/ml anti-HER3 antibody 10D1 for 4 hrs, followed by 15 min stimulation with NRG (100 ng/ml). Proteins were then extracted, quantified using standard Bradford protein assay, fractionated by SDS-PAGE, and transferred to nitrocellulose membranes. The membranes were then blocked and immunoblotted with the following antibodies overnight at 4° C. anti-pHER3, anti-pAKT, pan anti-HER3, pan anti-AKT and anti-beta-actin. The blots were visualized via Bio-Rad Clarity Western ECL substrate, and bands were quantified using densiometric analysis; data were normalized to beta actin controls.

(165) The results are shown in FIG. 17. Anti-HER3 antibody 10D1 was found to inhibit HER3 phosphorylation and downstream signalling.

(166) In further experiments the inventors investigated the intracellular signalling pathways affected by anti-HER3 antibody-mediated inhibition of HER3.

(167) FaDu cells were seeded in wells of a 6 well plate with 10% serum at 37° C., 5% CO.sub.2. After 16 hrs, cells were starved by culture overnight in 1% FBS cell culture medium. On the following day cells were treated with 50 μg/ml anti-HER3 antibody 10D1 for 4 hrs, followed by 15 min stimulation with NRG (100 ng/ml). Proteins were then extracted, quantified using standard Bradford protein assay, and incubated overnight with pre-blocked Phosphoprotein Antibody Array membrane (Ray Biotech) at 4° C. The membrane was then washed with washing buffer and incubated with detection antibody cocktail for 2 hrs at room temperature, followed by washing and incubation with HRP-Conjugated anti-IgG. After 2 hrs the membrane was washed and probed using the kit detection buffer. Images were captured with Syngene Gbox imaging system, the intensity of each dot/phosphoprotein was measured and percent inhibition was calculated by comparison with intensity measured for cells treated in the same way in the absence of the antibody.

(168) The results are shown in FIG. 18. Anti-HER3 antibody 10D1 was found to inhibit PI3K/AKT/mTOR and MAPK signalling.

(169) In further experiments the inventors investigated the effect of treatment with anti-HER3 antibody 10D1 on proliferation of HER3-expressing cells.

(170) Briefly, N87 and FaDu cells were treated with serially diluted concentrations of anti-HER3 antibody 10D1, starting from 100 μg/ml with a 9-point half log dilution. Cell proliferation was measuring using the CCK-8 proliferation assay (Dojindo, Japan) after a period of 5 days, in accordance with the manufacturer's instructions. Briefly 1×CCK-8 solution was added to each well followed by incubation for 2 h at 37° C. The OD was then measured at 450 nm.

(171) FIGS. 19A and 19B shows the percent cell confluence relative to untreated control cells (data points are averages of three replicates).

(172) Anti-HER3 antibody 10D1 displayed dose-dependent inhibition of cell proliferation by N87 and FaDu cells.

Example 5: Analysis In Vivo

5.1 Pharmacokinetic Analysis

(173) Female NCr nude mice approximately 6-8 weeks old were housed under specific pathogen-free conditions and treated in compliance with the Institutional Animal Care and Use Committee (IACUC) guidelines.

(174) 500 μg anti-HER3 antibody was administered and blood was obtained from 3 mice by cardiac puncture at baseline (−2 hr), 0.5 hr, 6 hr, 24 hr, 96 hr, 168 hr and 336 hr after administration. Antibody in the serum was quantified by ELISA.

(175) The results are shown in FIG. 50. Anti-HER3 antibody clone 10D1 was found to have a half-life of 16.3 days in NCr nude mice.

5.2 Safety Immunotoxicity

(176) Anti-HER3 antibody clone 10D1 was analysed in silico for safety and immunogenicity using IMGT DomainGapAlign (Ehrenmann et al., Nucleic Acids Res., 38, D301-307 (2010)) and IEDB deimmunization (Dhanda et al., Immunology. (2018) 153(1):118-132) tools.

(177) Anti-HER3 antibody clone 10D1 had numbers of potential immunogenic peptides few enough to be considered safe, and did not possess any other properties that could cause potential developability issues.

(178) The Table of FIG. 37 provides an overview of the properties of the 10D1 variant clones relevant to safety and developability.

(179) Mice treated with anti-HER3 antibodies in the experiments described in Example 5.3 were monitored for changes in weight and gross necroscopy. No differences were detected in these mice as compared to mice treated with vehicle only.

(180) Hemotoxicity was investigated in an experiment in which 6-8 week old female BALB/c mice (20-25 g) were injected intraperitoneally with a single dose of 1000 μg anti-HER3 10D1 antibody or an equal volume of PBS. Blood samples were obtained at 96 hours post injection and analysed for numbers of different types of white blood cells by flow cytometry and electrolyte indices for NA.sup.+, K.sup.+, and Cl.sup.−.

(181) FIGS. 51A and 51B show that the numbers of the different cell types and electrolyte indices were found to be within the Charles River reference range (3 mice) and did not differ significantly from the PBS-treated group (3 mice). Left bars represent vehicle, right bars represent 10D1 treatment, end points of the Charles River reference range indicated with dotted lines. No differences in clinical signs, gross necroscopy or weight were detected between the different groups.

(182) Mice were also analysed for correlates hepatotoxicity, nephrotoxicity and pancreatic toxicity at 96 hours post injection. The levels of alanine aminotransferase (ALT), aspartate transaminase (AST), blood urea nitrogen (BUN), creatinine (CREA), alkaline phosphatase (ALP), glucose (GLU), calcium (CAL), total bilirubin (BIL), total protein (TPR) and albumin (ALB) detected following administration of a single dose of 1000 μg anti-HER3 antibody were found to be within the Charles River reference range and do not differ significantly from the levels of these markers in the PBS-treated group. These are shown in FIGS. 51C to 51F. Left bars represent vehicle, right bars represent 10D1 treatment, end points of the Charles River reference range indicated with dotted lines. 10D1 treatment has no effect on the kidney, liver or pancreatic indices and thus does not affect normal kidney, liver or pancreatic functions.

5.3 Analysis of Efficacy to Treat Cancer In Vivo

(183) Female NCr nude mice approximately 6-8 weeks old were purchased from InVivos (Singapore). Animals were housed under specific pathogen-free conditions and were treated in compliance with the Institutional Animal Care and Use Committee (IACUC) guidelines.

(184) Cell lines used included N87 cells (gastric cancer), FaDu cells (head and neck cancer), OvCAR8 cells (ovarian cancer), SNU16 cells (gastric cancer), HT29 cells (colorectal cancer), A549 cells (lung cancer), HCC95 cells (lung cancer) and AHCN cells (kidney cancer).

(185) Tumor volumes were measured 3 times a week using a digital caliper and calculated using the formula [L×W2/2]. Study End point was considered to have been reaches once the tumors of the control arm measured >1.5 cm in length.

5.3.1 N87 Model

(186) FIG. 20 shows the results obtained in an experiment wherein the anti-cancer effect of anti-HER3 antibody 10D1 ([1] of Example 2.2) was investigated in a N87 cell-line derived mouse gastric carcinoma model. The model was established by subcutaneous injection of 1×10.sup.6 N87 cells into the right flank (n=6 mice per treatment group).

(187) 10D1 was administered IP, biweekly at 500 μg per dose (for a total of 10 doses); a control treatment group received an equal volume of PBS.

(188) Anti-HER3 antibody clone 10D1 was found to be highly potent in this model, and capable of inhibiting tumor growth by ˜76%.

(189) FIG. 21 shows the results obtained in a similar experiment in wherein anti-HER3 antibody clone 4-35-B4 was administered weekly IP at a dose of 11 mg/kg (total of 4 doses). Anti-HER3 antibody clone 4-35-B4 was similarly found to be highly potent in this model, and capable of inhibiting tumor growth by ˜60%.

5.3.2 SNU16 Model

(190) FIG. 22 shows the results obtained in an experiment wherein the anti-cancer effect of anti-HER3 antibody 10D1 ([1] of Example 2.2) was investigated in a SNU16 cell-line derived mouse gastric carcinoma model. The model was established by subcutaneous injection of 1×10.sup.6 SNU16 cells into the right flank (n=6 mice per treatment group).

(191) 10D1 was administered IP, biweekly at 500 μg per dose (for a total of 9 doses); a control treatment group received an equal volume of PBS.

(192) Anti-HER3 antibody clone 10D1 was found to be highly potent in this model, and capable of inhibiting tumor growth by ˜68%.

5.3.3 FaDu Model

(193) FIG. 23 shows the results obtained in an experiment wherein the anti-cancer effect of anti-HER3 antibody 10D1 ([1] of Example 2.2) was investigated in a FaDu cell-line derived mouse model of head and neck squamous cell carcinoma. The model was established by subcutaneous injection of 1×10.sup.6 FaDu cells into the right flanks of female NPG mice (NOD scid gamma phenotype; n=6 mice per treatment group).

(194) 10D1 was administered IP, weekly at 500 μg per dose (for a total of 4 doses). Control treatment groups received an equal volume of PBS, or the same dose of an isotype control antibody.

(195) Anti-HER3 antibody clone 10D1 was found to be highly potent in this model, and capable of inhibiting tumor growth by ˜85%.

(196) FIG. 24 shows the results obtained in an experiment wherein the anti-cancer effect of anti-HER3 antibody 10D1 ([1] of Example 2.2) was investigated in a FaDu cell-line derived mouse model of head and neck squamous cell carcinoma. The model was established by subcutaneous injection of 1×10.sup.6 FaDu cells into the right flanks of female NCr nude mice (n=6 mice per treatment group).

(197) 10D1 was administered IP, biweekly at 500 μg per dose (for a total of 8 doses); a control treatment group received an equal volume of PBS.

(198) Anti-HER3 antibody clone 10D1 was found to be highly potent in this model, and capable of inhibiting tumor growth by ˜86%.

5.3.4 OvCAR8 Model

(199) FIG. 25 shows the results obtained in an experiment wherein the anti-cancer effect of anti-HER3 antibody 10D1 ([1] of Example 2.2) was investigated in an OvCAR8 cell-line derived mouse model of ovarian carcinoma. The model was established by subcutaneous injection of 1×10.sup.6 OvCAR8 cells into the right flanks of female NCr nude mice (n=6 mice per treatment group).

(200) 10D1 was administered IP, biweekly at 500 μg per dose (for a total of 9 doses); a control treatment group received an equal volume of PBS.

(201) Anti-HER3 antibody clone 10D1 was found to be highly potent in this model, and capable of inhibiting tumor growth by ˜74%.

5.3.5 HCC-95 Model

(202) FIG. 26 shows the results obtained in an experiment wherein the anti-cancer effect of anti-HER3 antibody 10D1 ([1] of Example 2.2) was investigated in a HCC-95 cell-line derived mouse model of squamous cell lung carcinoma. The model was established by subcutaneous injection of 1×10.sup.6 HCC-95 cells into the right flanks of female NCr nude mice (n=6 mice per treatment group).

(203) 10D1 was administered IP, biweekly at 500 μg per dose (for a total of 4 doses); a control treatment group received an equal volume of PBS.

(204) Anti-HER3 antibody clone 10D1 was found to be highly potent in this model, and capable of inhibiting tumor growth by ˜90%.

5.3.6 A549 Model

(205) FIG. 27 shows the results obtained in an experiment wherein the anti-cancer effect of anti-HER3 antibody 10D1 ([1] of Example 2.2) was investigated in an A549 cell-line derived mouse model of lung adenocarcinoma. The model was established by subcutaneous injection of 1×10.sup.6 A549 cells into the right flanks of female NCr nude mice (n=6 mice per treatment group).

(206) 10D1 was administered IP, biweekly at 500 μg per dose (for a total of 10 doses); a control treatment group received an equal volume of PBS.

(207) Anti-HER3 antibody clone 10D1 was found to be highly potent in this model, and capable of inhibiting tumor growth by ˜91%.

(208) FIG. 28 shows the results obtained in a similar experiment in wherein the anti-cancer effect of anti-HER3 antibody clone 4-35-B2 was investigated in an A549 cell-line derived model established by injection of 1×10.sup.6 A549 cells into the right flanks of female NPG mice (NOD scid gamma phenotype). Anti-HER3 antibody clone 4-35-B2 was administered IP weekly at a dose of 500 μg per dose (total of 4 doses). A control treatment group received an equal volume of PBS (6 mice per treatment group).

(209) Anti-HER3 antibody clone 4-35-B2 was similarly found to be highly potent in this model, and capable of inhibiting tumor growth by ˜63%.

5.3.6 ACHN Model

(210) FIG. 29 shows the results obtained in an experiment wherein the anti-cancer effect of anti-HER3 antibody 10D1 ([1] of Example 2.2) was investigated in an ACHN cell-line derived mouse model of renal cell carcinoma. The model was established by subcutaneous injection of 1×10.sup.6 ACHN cells into the right flanks of female NCr nude mice (n=6 mice per treatment group).

(211) 10D1 was administered IP, biweekly at 500 μg per dose (for a total of 7 doses); a control treatment group received an equal volume of PBS.

(212) Anti-HER3 antibody clone 10D1 was found to be highly potent in this model, and capable of inhibiting tumor growth by ˜61%.

5.4 Treatment of Gastric Carcinoma

(213) First in Human

(214) Patients with HER2+ advanced gastric cancer who have failed or cannot receive trastuzumab are treated by intravenous injection of anti-HER3 antibody selected from: 10D1, 10D1_c75, 10D1_c76, 10D1_c77, 10D1_c78v1, 10D1_c78v2, 10D1_11B, 10D1_c85v1, 10D1_c85v2, 10D1_c85o1, 10D1_c85o2, 10D1_c87, 10D1_c89, 10D1_c90, 10D1_c91, 10D1_c92 and 10D1_c93, at a dose calculated in accordance with safety-adjusted ‘Minimal Anticipated Biological Effect Level’ (MABEL) approach. Patients are monitored for 28 days post-administration.

(215) The patients are then evaluated according to the Common Terminology Criteria for Adverse Events (CTCAE), to determine the safety and tolerability of the treatment, and to determine the pharmacokinetics of the molecules.

(216) Treatment with the anti-HER3 antibodies is found to be safe and tolerable.

(217) Dose Escalation—Monotherapy

(218) 12-48 patients with HER2+ advanced gastric cancer who have failed or cannot receive trastuzumab are treated by intravenous injection of anti-HER3 antibody selected from: 10D1, 10D1_c75, 10D1_c76, 10D1_c77, 10D1_c78v1, 10D1_c78v2, 10D1_11B, 10D1_c85v1, 10D1_c85v2, 10D1_c85o1, 10D1_c85o2, 10D1_c87, 10D1_c89, 10D1_c90, 10D1_c91, 10D1_c92 and 10D1_c93 (e.g. 10D1_c89, 10D1_c90 or 10D1_c91; e.g. 10D1_c89), in accordance with a 3+3 model based escalation with overdose control (EWOC) dose escalation.

(219) The patients are then evaluated according to the Common Terminology Criteria for Adverse Events (CTCAE) to determine the safety and tolerability of the treatment, and the pharmacokinetics of the molecules and efficacy of the treatment is evaluated. The maximum tolerated dose (MTD) and maximum administered dose (MAD) are also determined.

(220) Dose Escalation—Combination Therapy

(221) 9-18 patients with HER2+ advanced gastric cancer who have failed or trastuzumab are treated by intravenous injection of anti-HER3 antibody selected from: 10D1, 10D1_c75, 10D1_c76, 10D1_c77, 10D1_c78v1, 10D1_c78v2, 10D1_11B, 10D1_c85v1, 10D1_c85v2, 10D1_c85o1, 10D1_c85o2, 10D1_c87, 10D1_c89, 10D1_c90, 10D1_c91, 10D1_c92 and 10D1_c93 (e.g. 10D1_c89, 10D1_c90 or 10D1_c91; e.g. 10D1_c89) in combination with trastuzumab, in accordance with a 3+3 model based escalation with anti-PD-L1 antibody (3 mg/kg).

(222) The patients are then evaluated according to the Common Terminology Criteria for Adverse Events (CTCAE) to determine the safety and tolerability of the treatment, and the pharmacokinetics of the molecules and efficacy of the treatment is evaluated.

(223) Dose Expansion

(224) Patients with HER2+ advanced gastric cancer who have recently failed trastuzumab, and whose tumours have been well-characterised genetically and histologically are treated with anti-HER3 antibody selected from: 10D1, 10D1_c75, 10D1_c76, 10D1_c77, 10D1_c78v1, 10D1_c78v2, 10D1_11B, 10D1_c85v1, 10D1_c85v2, 10D1_c85o1, 10D1_c85o2, 10D1_c87, 10D1_c89, 10D1_c90, 10D1_c91, 10D1_c92, 10D1_c93 (e.g. 10D1_c89, 10D1_c90 or 10D1_c91; e.g. 10D1_c89) in combination with trastuzumab, cisplatin, and either 5-FU or capecitabine

(225) The anti-HER3 antibodies are found to be safe and tolerable, to be able to reduce the number/proportion of cancer cells, reduce tumor cell marker expression, increase progression-free survival and increase overall survival.

Example 6: Affinity Matured and Humanised Clones

(226) Humanization of the variable regions of the parental mouse antibody 10D1P was done by CDR grafting. Human framework sequences for grafting were identified by blasting the parental amino acid sequence against the human V domain database and the genes with highest identity to the parental sequence were selected. Upon grafting the mouse CDRs into the selected human frameworks, residues in canonical positions of the framework were back mutated to the parental mouse sequence to preserve antigen binding. A total of 9 humanized variants of 10D1P were designed.

(227) Affinity against human HER3 was increased by two rounds of affinity maturation using yeast display. In the first round, a mixed library of the 9 designed variants was constructed by random mutagenesis and screened by flow cytometry using biotinylated antigen. In the second round, one heavy chain and one light chain clones isolated in the first round were used as template to generate and screen a second library. A total of 10 humanized and affinity matured clones were isolated.

(228) Potential liabilities (immunogenicity, glycosylation sites, exposed reactive residues, aggregation potential) in the variable regions of the designed and isolated humanized variants of 10D1P was assessed using in silico prediction tools. The sequences were deimmunised using IEDB deimmunisation tool. The final sequence of 10D1F was selected among the optimized variants based on its developability characteristics as well as in vitro physicochemical and functional properties.

(229) Clone 10D1F comprises VH of SEQ ID NO:36 and VL of SEQ ID NO:83. 10D1F displays 89.9% homology with human heavy chain and 85.3% homology with human light chain.

(230) The antigen-binding molecule comprising 10D1F variable regions and human IgG1 constant regions, and which is comprised of the polypeptides of SEQ ID NOs: 206 and 207, is designated 10D1F.FcA (also sometimes referred to herein as “10D1F.A” or “anti-HER3 clone 10D1_c89 IgG1”—see e.g. [16] of Example 2.2).

Example 7: Fc Engineering

(231) 10D1 and 10D1 variants were engineered to comprise mutations in CH2 and/or CH3 regions to increase the potency of the antibodies, e.g. optimise Fc effector function, enhance antibody-dependent cellular cytotoxicity (ADCC) and/or antibody-dependent cellular phagocytosis (ADCP), and improve half-life.

(232) The Fc regions of clones 10D1 and 10D1F.FcA were modified to include modifications ‘GASDALIE’ (G236A, S239D, A330L, 1332E) and ‘LCKC’ (L242C, K334C) in the CH2 region. The GASDALIE substitutions were found to increase affinity for the FcγRIIa (GA) and FcγRIIIa (SDALIE) receptors and enhance ADCP and NK-mediated ADCC (see Example 8.8), whilst decreasing affinity for C1q (AL) and reducing CDC. The LCKC substitutions were found to increase thermal stability of the Fc region by creating a new intramolecular disulphide bridge.

(233) The modified version of 10D1F.FcA heavy chain polypeptide comprising the GASDALIE and LCKC mutations is shown in SEQ ID NO:225. The antigen-binding molecule comprised of the polypeptides of SEQ ID NOs: 225 and 207 is designated 10D1F.FcB (also sometimes referred to herein as “10D1F.B”).

(234) A modified version of 10D1 comprising the GASDALIE and LCKC substitutions in CH2 region was prepared and its ability to bind Fc receptor FcγRIIIa was analysed by Bio-Layer Interferometry. The sequence for 10D1 VH-CH1-CH2-CH3 comprising substitutions GASDALIE and LCKC corresponding to G236A, S239D, A330L, I332E and L242C, K334C is shown in SEQ ID NO:227.

(235) Briefly, anti-Penta-HIS (HIS1K) coated biosensor tips (Pall ForteBio, USA) were used to capture His-tagged FcγRIIIa (V158) (270 nM) for 120 s. All measurements were performed at 25° C. with agitation at 1000 rpm. Association kinetic measurements for antigen binding were performed by incubating anti-HER3 antibodies at different concentrations (500 nM to 15.6 nM) for 60 s, followed by a 120 s dissociation time by transferring the biosensors into assay buffer (pH 7.2) containing wells. Sensograms were referenced for buffer effects and then fitted using the Octet QK384 user software (Pall ForteBio, USA). Kinetic responses were subjected to a global fitting using a one site binding model to obtain values for association (K.sub.on), dissociation (K.sub.off) rate constants and the equilibrium dissociation constant (K.sub.D). Only curves that could be reliably fitted with the software (R.sup.2>0.90) were included in the analysis.

(236) The thermostability of the variant was also analysed by Differential Scanning Fluorimetry analysis as described in Example 3.4.

(237) FIGS. 38A and 38B show the BLI analysis and thermostability analysis, respectively, for 10D1 comprising GASDALIE and LCKC Fc substitutions. The Fc engineered 10D1 variant showed significantly improved binding to FcγRIIIa (˜9 fold increase in affinity) compared to non-Fc-engineered 10D1 (see FIG. 39A) with thermal stability maintained above 60° C.

(238) A construct for 10D1 comprising the GASD substitutions in CH2 region was also prepared; a sequence of 10D1 VH-CH1-CH2-CH3 comprising substitutions corresponding to G236A and S239D is shown in SEQ ID NO:228.

(239) The affinity of anti-HER3 antibody clone 10D1 ([1] of Example 2.2) and the GASD variant thereof were analysed by Bio-Layer Interferometry for affinity of binding to FcγRIIIa. BLI was performed as described above.

(240) FIGS. 39A and 39B show representative sensorgrams, K.sub.on, K.sub.off and K.sub.D values. As expected, the 10D1 GASD variant (39B) displayed dramatically increased affinity for FcγRIIIa compared to 10D1 (39A).

(241) The thermostability of the 10D1 GASD variant was also analysed by Differential Scanning Fluorimetry analysis as described in Example 3.4. The results are shown in FIG. 40.

(242) Further 10D1F Fc Variant

(243) Another antibody variant was created comprising an N297Q substitution in the CH2 region. A representative sequence for 10D1F VH-CH1-CH2-CH3 comprising the N297Q substitution is shown in SEQ ID NO:226. This ‘silent form’ prevents both N-linked glycosylation of the Fc region and Fc binding to Fcγ receptors and is used as a negative control.

(244) FIGS. 41A and 41B show the binding affinity of 10D1F hIgG1 Fc variants 10D1F.FcA and 10D1F.FcB to human and mouse Fc receptors, determined as described above. 10D1F.FcB was found to show significantly improved binding to human and mouse Fcγ and FcRn receptors compared to non-modified 10D1F.FcA or commercially available antibodies. ND=K.sub.D Not Determined due to low binding affinity.

Example 8: Characterisation of Humanised and Modified Clones

8.1 Analysis of Cell Surface Antigen-Binding by Flow Cytometry

(245) Wildtype (WT) HEK293 cells (which do not express high levels of HER3) and HEK293 cells transfected with vector encoding human HER3 (i.e. HEK293 HER O/E cells) were incubated with 10 μg/ml of humanised anti-HER3 antibody 10D1F.FcA (10D1F), anti-HER3 antibody 10D1 (10D1P) or isotype control antibody at 4° C. for 1.5 hr. The anti-HER3 antibody clone LJM716 (described e.g. in Garner et al., Cancer Res (2013) 73: 6024-6035, and Example 3.5) was included in the analysis as a positive control.

(246) The cells were washed with buffer (PBS with 2 mM EDTA and 0.5% BSA) and resuspended in FITC-conjugated anti-FC antibody (Invitrogen, USA) at 10 μg/ml for 20 min at 4° C. Cells were washed again and resuspended in 200 μL of FACS flow buffer (PBS with 5 mM EDTA) for flow cytometric analysis using MACSQuant 10 (Miltenyi Biotec, Germany). Unstained WT and transfected HEK293 cells were included in the analysis as negative controls. After acquisition, all raw data were analyzed using Flowlogic software. Cells were gated using forward and side scatter profile percentage of positive cells was determined for native and overexpressing cell populations.

(247) The results are shown in FIGS. 42A and 42B. Anti-HER3 antibody 10D1F.FcA was shown to bind to human HER3 with high specificity (42A). 10D1F.FcA, 10D1P and LJM716 were shown to bind to human HER3-expressing cells to a similar extent (42B).

8.2 ELISAs for Determining Antibody Specificity and Cross-Reactivity

(248) ELISAs were used to confirm the binding specificity of the 10D1F.FcA antibody. The antibodies were analysed for their ability to bind to human HER3 polypeptide as well as human HER1 (EGFR) and human HER2 (Sino Biological Inc., China). Human IgG isotype and an irrelevant antigen were included as negative controls.

(249) ELISAs were carried out according to standard protocols. Plates were coated with 0.1 μg/ml of target polypeptide in phosphate-buffered saline (PBS) for 16 h at 4° C. After blocking for 1 h with 1% BSA in Tris buffer saline (TBS) at room temperature, the anti-HER3 antibody was serially diluted with the highest concentration being 10 μg/ml, and added to the plate. Post 1 h incubation at room temperature, plates were washed three times with TBS containing 0.05% Tween 20 (TBS-T) and were then incubated with a HRP-conjugated anti-His antibody (Life Technologies, Inc., USA) for 1 h at room temperature. After washing, plates were developed with colorimetric detection substrate 3,3′,5,5′-tetramethylbenzidine (Turbo-TMB; Pierce, USA) for 10 min. The reaction was stopped with 2M H2504, and OD was measured at 450 nM.

(250) The results are shown in FIG. 43. Anti-HER3 antibody 10D1F.FcA was found not to bind to human HER2 or human HER1 (EGFR) even at high concentrations of the antibody.

(251) The ability of 10D1F.FcA to bind HER4 was analysed using flow cytometry. Wildtype (WT) HEK293 cells (which do not express high levels of HER4) and HEK293 cells transfected with vector encoding human HER4 (i.e. HEK293 HER O/E cells) were incubated with 10 μg/ml of anti-HER3 antibody 10D1F.FcA (10D1F) or isotype control antibody (negative control) at 4° C. for 1.5 hr. The anti-HER3 antibody clones LJM716 (described e.g. in Garner et al., Cancer Res (2013) 73: 6024-6035) and MM-121 (seribantumab), as described in Example 3.5, were included in the analysis as positive control. Also included was a commercial anti-HER4 antibody (Novus, Cat: FAB11311P). Unstained HEK293 cells were included in the analysis as negative controls.

(252) HEK293 cells were incubated with 10 μg/ml of each antibody for 1 hour at 4° C. Flow cytometry was performed as described above. Cells were contacted with FITC-conjugated anti-FC antibody (Invitrogen, USA) at for 30 min at 4° C.

(253) The results are shown in FIG. 44. Anti-HER3 antibody 10D1F.FcA was found not to bind to cell-surface expressed HER4.

(254) In addition, antibody 10D1F.FcA was analysed for its ability to bind to HER3 polypeptide homologues from mouse, rat and monkey (Sino Biological Inc., China). M. musculus, R. norvegicus and M. cynomolgus HER3 homologues share 91.1, 91.0 and 98.9% sequence identity respectively with human HER3 and the HER3 signalling pathways are conserved between the four species.

(255) ELISAs were performed as above.

(256) The results are shown in FIG. 45. 10D1F.FcA antibody was found to bind with high affinity to HER3 cyno, mouse, rat and human orthologs, thus displaying substantial cross-reactivity between species.

8.3 Global Affinity Study Using Octet QK384 System

(257) The anti-HER3 antibody clones 10D1F.FcA and 10D1F.FcB were analysed for binding affinity to human HER3.

(258) Bio-Layer Interferometry (BLI) experiments were performed using the Octet QK384 system (ForteBio). Antibodies (25 nM) were coated onto anti-Human IgG Capture (AHC) Octet sensor tips (Pall ForteBio, USA). Binding was detected using titrated HIS-tagged human HER3 in steps of baseline (60 s), loading (120 s), baseline2 (60 s), association (120 s), dissociation (FcA 120 s, FcB 600 s) and regeneration (15 s). Antigen concentrations are shown in the table in FIGS. 46A and 46B. Sensorgrams were analysed as described in Example 3.3. Values were obtained for association (K.sub.on), dissociation (K.sub.off) rate constants and the equilibrium dissociation constant (K.sub.D).

(259) Representative sensorgrams for the analysis of clones 10D1F.FcA and 10D1F.FcB are shown in FIGS. 46A and 46B. 10D1F.FcA binds to human HER3 with a high affinity of K.sub.D=72.6 pM (46A). 10D1F.FcB binds to human HER3 with a high affinity of K.sub.D=22.2 pM (46B).

8.4 Analysis of Thermostability by Differential Scanning Fluorimetry

(260) Differential Scanning Fluorimetry was performed for antibodies 10D1F.FcA and 10D1F.FcB as described in Example 3.4.

(261) The first derivative of the raw data obtained for Differential Scanning Fluorimetry analysis of the thermostability of antibody clone 10D1F.FcA is shown in FIG. 47A. Three different samples of the antibody were analysed and the Tm was determined to be 70.0° C.

(262) The first derivative of the raw data obtained for Differential Scanning Fluorimetry analysis of the thermostability of antibody clone 10D1F.FcB is shown in FIG. 47B. Three different samples of the antibody were analysed and the Tm was determined to be 62.7° C.

8.5 Antibody Purity Analysis

(263) The purity of antibodies 10D1F.FcA and 10D1F.FcB was analysed by size exclusion chromatography (SEC). 150 μg of 10D1F.FcA in 500 μl PBS pH 7.2 or 150 μg of 10D1F.FcB in 500 μl PBS pH 7.45 was injected on a Superdex 200 10/30 GL column in PBS running buffer at a flow rate of 0.75 min/ml or 0.5 min/ml, respectively, at room temperature and the A280 of flow through was recorded.

(264) The results are shown in FIGS. 48A (10D1F.FcA) and 48B (10D1F.FcB).

8.6 Analysis of Anti-HER3 Antibody 10D1F.FcA Epitope

(265) Anti-HER3 antibody 10D1F.FcA was analysed to determine whether it competes with anti-HER3 antibodies M-05-74 or M-08-11 (Roche) for binding to HER3. Epitopes of M-05-74 and M-08-11 were both mapped to the β-hairpin structure of the HER3 dimerisation arm located at domain II. M-08-11 does not bind to HER4 whereas M-05-74 recognises the HER4 dimerisation arm. Binding of M-05-74 and M-08-11 to HER3 is ligand (NRG) independent.

(266) BLI experiments were performed as described in Example 3.5 with one alteration: 400 nM of competing antibodies were used. The variable regions of M-05-74 and M-08-11 antibodies were cloned in the PDZ vector having human IgG1 and IgKappa Fc backbone.

(267) The results of the analysis are shown in FIGS. 49A and 49B. Anti-HER3 10D1F.FcA antibody was found not to compete with M-05-74 or M-08-11 for binding to HER3. 10D1F.FcA was found to bind a distinct and topologically distant epitope of HER3 compared to M-05-74 and M-08-11. Binding of 10D1F.FcA to HER3 is ligand (NRG) independent.

(268) Conclusions: Binding of 10D1F.FcA to human HER3 can be achieved in a ligand-independent manner. 10D1F.FcA binding epitope is distinct and topologically distant from that of M-05-74 and M-08-11.

8.7 Inhibition of Dimerisation of HER2-HER3 and EGFR-HER3

(269) Anti-HER3 antibody 10D1F.FcA was analysed for its ability to inhibit heterodimerisation of HER3 and HER2.

(270) Plate-based ELISA dimerisation assays were carried out according to standard protocols. Plates were coated with 1 μg/ml HER2-Fc protein. After blocking and washing, the plate was incubated with different concentrations of candidate antibodies 10D1F.FcA, MM-121, LJM716, Pertuzumab, Roche M05, Roche M08 or isotype control and constant HER3 His 2 μg/ml and NRG 0.1 μg/ml for 1 hour. Plates were then washed and incubated for 1 hour with secondary anti-HIS HRP antibody. Plates were washed, treated with TMB for 10 mins and the reaction was stopped using 2M H2504 stop solution. The absorbance was read at 450 nm.

(271) The results are shown in FIG. 52. Anti-HER3 antibody clone 10D1F.FcA was found to directly inhibit interaction between HER2 and HER3 in a dose dependent fashion.

(272) In another assay, inhibition of dimerization was detected using PathHunter® Pertuzumab Bioassay Kit (DiscoverX, San Francisco, USA). HER2 and HER3 overexpressing U2OS cells were thawed using 1 ml of pre-warmed CP5 media and 5K cells were seeded per at 37° C., 5% CO.sub.2 for 4 hrs. Cells were treated with serially diluted concentrations of 10D1F.FcA, Seribantumab, or Pertuzumab starting from 25 μg/ml with an 8-point serial dilution. After 4 hrs incubation, 30 ng/ml of heregulin-82 was added to each well and the plates were further incubated for 16 hrs. Following incubation, 10 μL PathHunter bioassay detection reagent 1 was added and incubated for 15 mins at room temperature in the dark, followed by addition of 40 μL PathHunter bioassay detection reagent 2 which was then incubated for 60 mins at room temperature in the dark. Plates were read using Synergy4 Biotek with 1 second delay.

(273) 10D1F.FcA was found to have an EC.sub.50 value of 3.715e-11 for inhibition of HER2-HER3 heterodimerisation. In the same assay, the comparative EC.sub.50 value for Seribantumab/MM-121 was found to be 6.788e-10 and the comparative EC.sub.50 value for Pertuzumab was found to be 2.481e-10.

(274) Anti-HER3 antibody 10D1F.FcA was analysed for its ability to inhibit heterodimerisation of EGFR and HER3.

(275) Plate-based ELISA dimerisation assays were performed according to standard protocols. Plate was coated with 1 μg/ml human EGFR-His. After blocking and washing, the plate was incubated with different concentrations of candidate antibodies 10D1F.FcA, MM-121, LJM716, Pertuzumab, or isotype control with constant HER3-biotin 4 μg/ml and NRG 0.1 μg/ml for 1 hour. Plates were then washed and incubated for 1 hour with secondary anti-avidin HRP antibody. Plates were washed, treated with TMB for 10 mins and the reaction was stopped using 2M H.sub.2SO.sub.4 stop solution. The absorbance was read at 450 nm.

(276) The results are shown in FIG. 53. Anti-HER3 antibody clone 10D1F.FcA was found to directly inhibit interaction between EGFR and HER3 in a dose dependent fashion.

8.8 Analysis of Ability to Induce ADCC

(277) Anti-HER3 antibody clones 10D1F.FcA and 10D1F.FcB were analysed for their ability to induce antibody-dependent cell-mediated cytotoxicity (ADCC).

(278) Target cells (HEK293 overexpressing HER3) were plated in U-bottom 96-well plates at a density of 20,000 cells/well. Cells were treated with a dilution series (50,000 ng/ml-0.18 ng/ml) of one of 10D1F.FcA, 10D1F.FcB, 10D1F.FcA_N297Q (silent form), LJM-716, Seribantumab (MM-121) or left untreated, and incubated at 37° C. and 5% CO.sub.2 for 30 min. Effector cells (Human Natural Killer Cell Line No-GFP-CD16.NK-92; 176V) were added to the plate containing target cells at a density of 60,000 cells/well.

(279) The following controls were included: Target cell maximal LDH release (target cells only), spontaneous release (target cells and effector cells without antibody), background (culture media only). Plates were spun down and incubated at 37° C. and 5% CO.sub.2 for 21 hrs.

(280) LDH release assay (Pierce LDH Cytotoxicity Assay Kit): before the assay, 10 μl of Lysis Buffer (10×) were added to target cell maximal LDH release controls and incubated at 37° C. and 5% CO.sub.2 for 20 min. After incubation, plates were spun down and 50 μL of the supernatant were transferred to clear flat-bottom 96-well plates. Reactions were started by addition of 50 μl of LDH assay mix containing substrate to the supernatants and incubated at 37° C. for 30 min. Reactions were stopped by addition of 50 μl of stop solution and absorbance was recorded at 490 nm and 680 nm with a BioTek Synergy HT microplate reader.

(281) For data analysis, absorbance from test samples was corrected to background and spontaneous release from target cells and effector cells. Percent cytotoxicity of test samples was calculated relative to target cell maximal LDH release controls and plot as a function of antibody concentration.

(282) The results are shown in FIG. 54. Anti-HER3 antibody 10D1F.FcB was found to induce potent ADCC activity against HER3 overexpressing cells in a dose dependent fashion.

8.9 Inhibition of HER3-Mediated Signalling

(283) Anti-HER3 antibody 10D1F.FcA was analysed for its ability to inhibit HER-3 mediated signalling in vitro in cancer cell lines.

(284) N87, FaDu or OvCAR8 cells were seeded in wells in a 6 well plate with 10% serum overnight at 37° C. with 5% CO.sub.2. Cells were starved with 0.2% FBS culture medium for 16 hrs, and were then treated for 0.5 hours with different antibodies at IC50 corresponding to the cell line. Antibodies tested were: 10D1F.FcA (10D1), Seribantumab (SBT), Elegemtumab (LJM), Pertuzumab (PTM), Cetuximab (CTX), and Trastuzumab (TZ).

(285) Before harvesting, cells were stimulated with 100 ng/ml of NRG1. Protein extracted from cell lines were quantified using standard Bradford protein assay. Protein samples (50 μg) were fractionated by SDS-PAGE and transferred to nitrocellulose membrane. Membranes were then blocked and immunoblotted with the indicated antibodies. The results were visualized via Bio-Rad Clarity Western ECL substrate. The blots were quantified using densiometric analysis and data was normalised to beta actin.

(286) The results are shown in FIGS. 55A to 55C. Anti-HER3 antibody 10D1F.FcA was found to inhibit HER3 phosphorylation and downstream signalling in N87 (55A), FaDu (55B), OvCar8 (55C) and A549 (55D) cell lines.

(287) For the experiments using N87 cells, A549 cells, OvCar8 cells and FaDu cells total RNA was extracted at 16 hrs post antibody treatment was analysed to determine pathway activation based on the level of expression of key signal transduction pathway proteins by gene set enrichment analysis. The results of the analysis are shown in FIGS. 63A to 63D. 10D1F.FcA was the most effective inhibitor of downstream signalling.

(288) In further experiments using A549 cells, in vitro phosphorylation assays were performed as above except that the cells were treated for 0.5 hours or 4 hours with the different antibodies. The results are shown in FIG. 64.

8.10 Analysis of Binding to HER3 Provided in ‘Open’ and ‘Closed’ Conformations

(289) In further experiments, anti-HER3 antibodies were analysed in order to determine the affinity of binding to human HER3 in the context of a human HER3: human NRG1 complex (that is, HER3 provided in the ligand-bound, ‘open’ conformation), as compared to the affinity of binding to human HER3 in the absence of NRG1 complex (that is, HER3 provided in the unbound, ‘closed’ conformation).

(290) Binding of anti-HER3 antibodies to human HER3 was evaluated by BLI. Briefly, anti-Human IgG Capture (AHC) sensors (ForteBio) were loaded with anti-HER3 IgG antibodies (25 nM). Kinetic measurements were performed in absence or presence of NRG1. NRG1 was used at 1:1 molar ratio with HER3 wherein the complex was allowed to form at RT for 2 h. His-tagged human HER3 or HER3-NRG1 complexes were loaded to antibody coated AHC sensors at different concentrations for 120 s, followed by a 120 s dissociation time. All measurements were performed at 25° C. with agitation at 1000 rpm. Sensorgrams were referenced for buffer effects and then analyzed using the Octet QK384-software (ForteBio). Kinetic responses were globally fitted using a one-site binding model to obtain values for association (K.sub.on), dissociation (K.sub.off) rate constants and the equilibrium dissociation constant (K.sub.D).

(291) The following anti-HER3 antibodies were analysed in the experiments: 10D1F.A ([16] of Example 2.2), MM-121 and LJM-716.

(292) The results are shown in FIGS. 78A to 78F. 10D1F.A was found to bind to human HER3 with sub-picomolar affinity in the presence or absence of NRG1 (FIGS. 78A and 78B). MM-121 bound to human HER3 with nanomolar in the presence or absence of NRG1 (FIGS. 78C and 78D). LJM-716 bound to human HER3 with subpicomolar affinity in the absence of NRG1 (FIG. 78E). However, in the presence of NRG1, binding to HER3 was dramatically reduced (FIG. 78F).

8.11 Conclusions

(293) Taken together, the results identify 10D1F as an anti-HER3 antibody which binds to an epitope of HER3 providing it with the unique combination of properties that (i) it competitively inhibits heterodimerisation of HER3 with EGFR or HER2 (see Example 8.7 and FIGS. 52 and 53), and (ii) binds HER3 similarly well in the presence or absence of NRG1 (FIGS. 78A and 78B).

Example 9: Analysis of Humanised and Modified Clones In Vitro and In Vivo

9.1 Pharmacokinetic Analysis

(294) Mice

(295) Female NCr nude mice approximately 6-8 weeks old were housed under specific pathogen-free conditions and treated in compliance with the Institutional Animal Care and Use Committee (IACUC) guidelines.

(296) 500 μg anti-HER3 antibody 10D1F.FcA or 10D1F.FcB was administered and blood was obtained from 4 mice by cardiac puncture at baseline (−2 hr), 6 hr, 24 hr, 96 hr, 168 hr and 336 hr after administration. Antibody in the serum was quantified by ELISA.

(297) The parameters for the pharmacokinetic analysis were derived from a non-compartmental model: maximum concentration (C.sub.max), AUC (0-336 hr), AUC (0-infinity), Half-life (t.sub.1/2), Clearance (CL), Volume of distribution at steady state (V.sub.ss).

(298) The results are shown in FIGS. 56A and 56B. Anti-HER3 antibody clone 10D1F.FcA was found to have a half-life of 253 hours (56A) and anti-HER3 antibody clone 10D1F.FcB was found to have a half-life of 273 hours (56B) in NCr nude mice.

(299) Rats

(300) 10D1F variants were analysed to determine a single dose pharmacokinetic profile in female Sprague Dawley rats with mean weight of 320 g.

(301) Antibody clones 10D1F.FcA and 10D1F.FcB were administered in a single dose of 4 mg (˜10 mg/kg), 10 mg (˜25 mg/kg), 40 mg (˜100 mg/kg) or 100 mg (˜250 mg/kg) via tail vein slow i.v. injection. Vehicle was administered as a negative control. Blood from 2 rats per treatment was obtained at baseline (−24 hr), 6 hr, 24 hr, 96 hr, 168 hr and 336 hr after administration. Antibody in the serum was quantified by ELISA.

(302) The parameters for the pharmacokinetic analysis were derived from a non-compartmental model: maximum concentration (C.sub.max), AUC (0-336 hr), AUC (0-infinity), Half-life (t.sub.1/2), Clearance (CL), Volume of distribution at steady state (V.sub.d).

(303) The results are shown in FIGS. 57A (10 mg/kg), 57B (25 mg/kg), 57C (100 mg/kg) and 57D (250 mg/kg).

9.2 Safety Immunotoxicity

(304) The toxicological effects of 10D1F.FcA and 10D1F.FcB were analysed.

(305) Mice

(306) 6-8 week old female BALB/c mice (20-25 g) were injected intraperitoneally with a single dose of either 10D1F.FcA and 10D1F.FcB at one of doses: 200 ug (˜10 mg/kg), 500 ug (˜25 mg/kg), 2 mg (˜100 mg/kg), or 5 mg (˜250 mg/kg), or an equal volume of PBS. 3 mice were injected with each treatment, 4 mice were injected with PBS control. Blood samples were obtained at 96 hours post injection and analysed for RBC indices (total RBC count, haematocrit, haemoglobin, platelet count, mean corpuscular volume, mean corpuscular haemoglobin, mean corpuscular haemoglobin concentration) and WBC indices (total WBC count, lymphocyte count, neutrophil count, monocyte count). Analysis was performed using a HM5 Hematology Analyser.

(307) The results are shown in FIGS. 58A, 58B (RBC indices) and 58C (WBC indices). Anti-HER3 antibodies 10D1F.FcA and 10D1F.FcB had no effect on the RBC indices, but were found to have an effect on the WBC indices at higher doses (10D1F.FcA 250 mg/kg, 10D1F.FcB 100 mg/kg, 10D1F.FcB 250 mg/kg).

(308) Hepatotoxicity, nephrotoxicity and pancreatic toxicity were also analysed 96 hours post injection. 10D1F.FcA and 10D1F.FcB had no effect on the levels of alanine aminotransferase, alkaline phosphatase, albumin, total protein (liver indices; FIG. 58D), creatine, blood urea nitrogen, glucose or amylase (kidney and pancreatic indices; FIG. 58E). Nor did 10D1F.FcA and 10D1F.FcB have an effect on electrolyte indices sodium, potassium, calcium or phosphate (FIG. 58F).

(309) Mice treated with 10D1F.FcA or 10D1F.FcB showed no abnormalities after 96 hours in weight, behaviour, skin condition, oral examination, stool and urine examination or eye examination. Enlarged spleen (splenomegaly) approximately 1.5 times the normal size was observed in mice treated with higher doses: 10D1F.FcA 250 mg/kg, 10D1F.FcB 100 mg/kg, 10D1F.FcB 250 mg/kg.

(310) A further study was performed in the BALB/c mice to assess the toxicological effects of repeat doses of 500 μg (˜25 mg/kg) 10D1F.FcA or 10D1F.FcB. Antibody was administered once a week for four weeks. Blood was obtained 28 days after the first administration. There was no effect observed on RBC, liver, kidney, pancreatic or electrolyte indices for either antibody, no signs of clinical abnormalities and no differences detected in gross necroscopy. Total WBC count, lymphocyte count and neutrophil count was observed to be decreased in mice treated with 10D1F.FcA or 10D1F.FcB but this was not considered to be toxic.

(311) In another study, BALB/c mice were administered with a single dose of 10D1F.FcA or an equal volume of PBS (vehicle control), and analysed after 96 hours (vehicle) or 336 hours (10D1F.FcA). Representative results are shown in FIGS. 69A to 69C.

(312) Rats

(313) 6-8 week old female Sprague Dawley rats (400-450 g) were injected intraperitoneally with a single dose of either 10D1F.FcA or 10D1F.FcB antibody at one of doses: 4 mg (˜10 mg/kg), 10 mg (˜25 mg/kg), 40 mg (˜100 mg/kg), 100 mg (˜250 mg/kg). Blood was obtained at −24 hours, 6 hours, 24 hours, 96 hours, 168 hours and 336 hours. Up to 366 hours post injection there was no effect on RBC indices, no toxic effect on WBC indices, and no effect on liver, kidney, pancreatic or electrolyte indices. There were no signs of clinical abnormalities and no differences detected in gross necroscopy.

(314) Representative results obtained from rats administered with 250 mg/kg 10D1F.FcA are shown in FIGS. 70A to 70C.

(315) The absence of toxicity signals in rodent toxicology models indicates superior clinical safety for 10D1 and variants.

9.3 Analysis of Efficacy to Treat Cancer In Vitro

(316) Anti-HER3 antibody 10D1F.FcA was analysed for its ability to inhibit tumour growth in vitro in a number of tumour models: N87 cells (gastric cancer), HCC95 cells (lung cancer), FaDu cells (head and neck cancer), SNU-16 cells (gastric cancer), A549 cells (lung cancer), OvCar8 cells (ovarian cancer), ACHN cells (kidney cancer) and HT29 cells (colorectal cancer). 10D1F.FcA efficacy was compared to other anti-HER3 antibodies seribantumab (MM-121) and LJM-716, and other EGFR family therapies cetuximab, trastuzumab and pertuzumab.

(317) Cells were treated with serially diluted concentrations of therapeutic antibodies, starting from 1500 ug/ml with a 9-point dilution. Cell viability was measured using CCK-8 cell proliferation assay, 3-5 days post treatment. The percentage of cell inhibition shown is relative to cells treated with only buffer (PBS). Data points indicates average of three replicates.

(318) The results are shown in FIG. 59A to 59D. Anti-HER3 antibody 10D1F.FcA demonstrates superior in vitro tumour inhibition in multiple tumour models compared to other HER3 antibodies (59A & 59B) and EGFR family therapies (59C & 59D).

(319) FIGS. 77A and 77B show the ability of different anti-ErbB antibodies to inhibit proliferation of different cancer cell lines in vitro, at the C.sub.max concentration they achieve in mice administered IP with 25 mg/kg of the relevant antibody. 10D1F.FcA displays outstanding ability to inhibit growth of a wide variety of different cancer cell types.

9.4 Analysis of Efficacy to Treat Cancer In Vivo

(320) Anti-HER3 antibody clones 10D1F.FcA and 10D1F.FcB were assessed for their effect on tumour growth in in vivo cancer models.

9.4.1 A549 Model

(321) Tumour cells were inserted subcutaneously into the right flanks of female NCr nude mice. Antibodies (25 mg/kg 10D1F.FcA, 10D1F.FcB, Cetuximab, LJM-716 or MM-121; n=6 for each treatment) or vehicle (n=8) were administered biweekly for six weeks.

(322) The results are shown in FIG. 60. Anti-HER3 antibody clones 10D1F.FcA and 10D1F.FcB both displayed potent efficacy in the A549 model of lung carcinoma. 10D1F.FcB was found to be particularly potent and made tumours regress.

9.4.2 FaDu Model

(323) Tumour cells were inserted with matrigel subcutaneously into the right flanks of female NCr nude mice. Antibodies (10 and 25 mg/kg 10D1F.FcA and 10D1F.FcB, or 25 mg/kg of Cetuximab, Trastuzumab, Pertuzumab, LJM-716 or MM-121; n=6 for each treatment) or vehicle (n=6) were administered once a week for six weeks.

(324) The results are shown in FIG. 61. Anti-HER3 antibody clones 10D1F.FcA and 10D1F.FcB were both found to be effective to prevent tumour growth in the FaDu model of head and neck cancer.

9.4.3 OvCar8 Model

(325) Tumour cells were inserted with matrigel subcutaneously into the right flanks of female NCr nude mice. Antibodies (10 and 25 mg/kg 10D1F.FcA, or 25 mg/kg Cetuximab, LJM-716 or MM-121; n=6) or vehicle (n=6) were administered once a week for six weeks.

(326) The results are shown in FIG. 62. Anti-HER3 antibody clone 10D1F.FcA was found to be effective at reducing tumour volume at higher dose.

9.4.4 N87 Model

(327) Tumour cells are inserted with matrigel subcutaneously into the right flanks of female NCr nude mice. Antibodies (25 mg/kg 10D1F.FcA, or 50 mg/kg of Trastuzumab, LJM-716 or MM-121; n=6 for each treatment) or vehicle (n=6) were administered biweekly for six weeks.

(328) The results are shown in FIG. 74. Anti-HER3 antibody 10D1F.FcA and was found to be effective to prevent tumour growth in the N87 model of gastric cancer.

Example 10: Analysis of Inhibition of Proliferation of BRAFV600E Mutant Thyroid Cancer Cell Lines

(329) The following cell lines were investigated:

(330) TABLE-US-00006 Cell Line Type of Cancer Mutation SW1736 Anaplastic thyroid cancer BRAF V600E BHT101 Anaplastic thyroid cancer BRAF V600E BCPAP Papillary thyroid cancer BRAF V600E and p53 mutation

(331) The cells were investigated for surface expression of EGFR family members by flow cytometry. Briefly, 300,000 cells were incubated with 20 μg/ml of 10D1F.FcA, cetuximab or trastuzumab for 1 hr at 4° C. Alexafluor 488-conjugated anti-human antibody was used at 10 μg/ml as a secondary antibody (40 min at 4° C.).

(332) The results are shown in FIG. 66A to 66C. SW1736, BHT101 and BCPAP cells were shown to express EGFR, HER2 and HER3.

(333) The inventors investigated the ability of different HER3-binding antibodies to inhibit in vitro proliferation of different thyroid cancer cell lines harbouring the V600E BRAF mutation.

(334) Briefly, cells of the different cell lines were seeded at a density of 1.5×10.sup.5 cells/well, and treated the next day with a 10 point serial dilution starting from 1000 μg/ml of 10D1F.FcA, seribantumab, LJM-716, pertuzumab or isotype control antibody. After 3 days, proliferation was measured using a CCK-8 cell proliferation assay. Percent inhibition of proliferation was calculated relative to cells treated with an equal volume of PBS instead of antibodies.

(335) The results are shown in FIGS. 67A to 67C. 10D1F.FcA was found to be more effective at inhibiting proliferation of cell lines harbouring BRAF V600E mutation than any other of the anti-HER3 antibodies analysed.

(336) In further experiments, the ability of a combination of 10D1F.FcA and vemurafenib to inhibit in vitro proliferation of different thyroid cancer cell lines harbouring the V600E BRAF mutation was investigated.

(337) Cells were seeded at a density of 1.5×10.sup.5 cells/well, and treated the next day with a 10 point serial dilution starting from 1000 μg/ml of 10D1F.FcA or isotype control antibody, in the presence or absence of 200 nM vemurafenib. After 3 days, proliferation was measured using a CCK-8 cell proliferation assay. Percent inhibition of proliferation was calculated relative to cells treated with an equal volume of PBS instead of antibodies.

(338) The results are shown in FIGS. 68A to 68C. 10D1F.FcA was found to enhance the ability of vemurafenib to inhibit proliferation of SW1736 and BHT101 cells, which are susceptible to vemurafenib. 10D1F.FcA was also found to be a potent inhibitor of proliferation of vemurafenib-resistant BCPAP cells.

Example 11: Analysis of Inhibition of HER3-Mediated Signalling In Vivo

(339) The inventors investigated the ability of 10D1F.FcA to inhibit HER3-mediated signalling in vivo. 1×10.sup.6 FaDu or OvCar8 cells were introduced subcutaneously into NCr nude mice, to establish ectopic xenograft tumors.

(340) Once tumors had reached a volume of greater than 100 mm.sup.3, mice were with treated by biweekly intraperitoneal injection of 10D1F.FcA at a dose of 25 mg/kg, or an equal volume of vehicle (control). After 4 weeks, tumors were harvested. Protein extracts were prepared from the tumors and quantified via Bradford assay, 50 μg samples were fractionated by SDS-PAGE, and analysed by western blot using antibodies in order to determine in vivo phosphorylation of HER3 and AKT, as described in Example 4.3.

(341) The results are shown in FIG. 71. 10D1F.FcA was found to inhibit phosphorylation of HER3 and AKT in tumor cells in vivo.

Example 12: Analysis of Internalisation of Anti-HER3 Antibodies

(342) The inventors investigated internalisation of anti-HER3 antibodies by HER3-expressing cells.

(343) Briefly, 100,000 HEK293 cells engineered to express HER3, HCC95, N87 or OVCAR8 cells were seeded in wells of 96-well tissue culture plates and cultured overnight at 37° C. in 5% CO.sub.2. Cells were then treated with 120 nM of 10D1F.FcA, LJM-716, seribantumab or trastuzumab, and 360 nM of pHrodo iFL Green reagent, and incubated at 37° C. in 5% CO.sub.2. The cells in culture were imaged every 30 min for 24 hours, in 4 different fields of each well. The maximum signal intensity in the FITC channel of each field was quantified at 24 hours.

(344) The results are shown in FIG. 72. Modest to moderate internalization of LJM-716 and seribantumab was observed in OvCar8 cells, whereas modest internalization of trastuzumab was observed in N87 cells.

(345) No significant internalization of 10D1F.FcA was observed in HCC95, N87, or OvCar8 cells.

(346) As expected, significant internalization of 10D1F.FcA, LJM-716, and seribantumab was observed in HEK293 cells overexpressing HER3.

(347) In further experiments, antibody internalisation was investigated by flow cytometry.

(348) N87 cells were seeded in wells of 96-well tissue culture plates at a density of 50,000 cells/well, and allowed to adhere overnight (37° C., 5% CO.sub.2). 10D1F.FcA or trastuzumab were mixed with labelling reagent, and the labelled complexes were added to cells. Samples were harvested at 0 min, 10 min, 30 min, 1 hour, 2 hour and 4.5 hour time points, by aspiration of cell culture medium, washing with PBS and treatment with accutase. Accutase activity was neutralised, and cells were resuspended in FACs buffer and analysed by flow cytometry.

(349) The results are shown in FIGS. 73A and 73B. The cells displayed minimal internalisation of 10D1F.FcA. By contrast, substantial internalisation of anti-HER2 antibody trastuzumab was observed.

Example 13: Use of HER3-Binding Antibodies in Immunohistochemistry

(350) Anti-HER3 antibody 10D1F in mIgG2a format was evaluated for its ability to be used in immunohistochemistry for the detection of human HER3 protein.

(351) Processing of sections was performed using Bond reagents (Leica Biosystems). Arrays of commercially-available frozen tissue sections were obtained. Slides were dried in a desiccator for 10 min and then subjected to the following treatments, with water washes and/or TBS-T rinses between steps: (i) fixation by treatment with 100% acetone for 10 min at room temperature; (ii) endogenous peroxidase blocking by treatment with 3% (v/v) H.sub.2O.sub.2 for 15 min at room temperature; (iii) blocking by treatment with 10% goat serum for 30 min at room temperature, (iv) incubation with 10D1F-mIgG2a at 1:250 dilution of a 6.2 mg/ml solution overnight at 4° C., (v) incubation with HRP-polymer conjugated goat anti-mouse antibody for 30 min at room temperature, and (vi) development with Bond Mixed DAB Refine for 5 min at room temperature, followed by rinsing with deionised water and 1× Bond Wash to stop the reaction.

(352) Slides were then dehydrated, mounted in synthetic mounting media and scanned with high resolution.

(353) The results are shown in FIGS. 75A and 75B. 10D1F preferentially stained malignant human tissue sections, with low cross-reactivity to normal tissue.

(354) In further experiments, an A549 xenograft tumor was harvested in cold PBS, embedded in OCT cryoembedding medium, frozen in dry-ice and stored at −80° C. 10 μm sections were obtained using a cryostat.

(355) Slides were dried in a desiccator for 10 min and then subjected to the following treatments, with water washes and/or TBS-T rinses between steps: (i) fixation by treatment with 100% acetone for 10 min at room temperature; (ii) endogenous peroxidase blocking by treatment with 3% (v/v) H.sub.2O.sub.2 for 15 min at room temperature; (iii) blocking by treatment with 10% goat serum for 30 min at room temperature, (iv) incubation with 10D1F.FcA at 1:50 dilution of 8.8 mg/ml solution, or with 1:200 dilution of Sino Biological rabbit anti-HER3 (Cat. No. 10201-T24) overnight at 4° C., (v) incubation with Invitrogen F(ab′)2-Goat anti-Human IgG (H+L) HRP (A24470) (1:500), or HRP-polymer conjugated goat anti-rabbit antibody for 30 min at room temperature, and (vi) development with Bond Mixed DAB Refine for 5 min at room temperature, followed by rinsing with deionised water and 1× Bond Wash to stop the reaction.

(356) Slides were then counterstained with haematoxylin, dehydrated, mounted in synthetic mounting media and scanned with high resolution.

(357) The results are shown in FIG. 76. 10D1F.FcA displayed specific membrane and cytoplasmic staining of A549 tumor xenograft cryosections.

Example 14: Therapeutic Efficacy of HER3-Binding Antibodies in a Patient-Derived Xenograft Model of Ovarian Cancer Comprising CLU-NRG1 Fusion

(358) The inventors investigated the therapeutic efficacy of 10D1F in a patient-derived xenograft model of ovarian cancer comprising CLU-NRG1 fusion.

(359) Female BALB/c nude mice approximately 5-7 weeks old were housed under specific pathogen-free conditions and were treated in compliance with the Institutional Animal Care and Use Committee (IACUC) guidelines.

(360) A model of ovarian cancer comprising CLU-NRG1 fusion was established by inoculating mice subcutaneously at the right flank with four pellets of a patient-derived xenograft designated OV6308 (Crown Bioscience Inc.). OV6308 is an ovarian high-grade serous adenocarcinoma derived from a 51-year female patient, and comprises the CLU-NRG1 gene fusion.

(361) Tumor volumes were measured 3 times a week using a digital caliper and calculated using the formula [L×W2/2]. Study End point was considered to have been reached once the tumors of the control arm measured >1.5 cm in length.

(362) Mice were administered biweekly by intraperitoneal injection with (i) 25 mg/kg of 10D1F.FcA (i.e. [16] of Example 2.2), or (ii) 25 mg/kg of hIgG1 isotype control antibody (n=10 for each treatment group).

(363) The results are shown in FIG. 79. Treatment with 10D1F.FcA was found to be extremely potent, inhibiting tumor growth by 111% relative to the isotype control-treated group.