ANTI-MUC1 ANTIBODY-DRUG CONJUGATE

Abstract

The present invention pertains to novel antibody drug conjugates (ADC) comprising anti-MUC1 antibody. In particular, said ADC showed significant anti-tumor efficacy.

Claims

1. A conjugate comprising an antibody conjugated to a cytotoxic agent, wherein the antibody is capable of binding to MUC1, and wherein the antibody comprises (i) a heavy chain variable region comprising the complementarity-determining regions (CDRs) CDR-H1 having the amino acid sequence of SEQ ID NO: 1, CDR-H2 having the amino acid sequence of SEQ ID NO: 2 and CDR-H3 having the amino acid sequence of SEQ ID NO: 3, and (ii) a light chain variable region comprising the complementarity-determining regions (CDRs) CDR-L1 having the amino acid sequence of SEQ ID NO: 4, CDR-L2 having the amino acid sequence of SEQ ID NO: 5 and CDR-L3 having the amino acid sequence of SEQ ID NO: 6.

2. The conjugate according to claim 1, wherein the amino acid at position 8 of SEQ ID NO: 2 is selected from the group consisting of glutamine, histidine, tryptophan, tyrosine, lysine and arginine, or wherein the CDR-H2 has the amino acid sequence of SEQ ID NO: 7.

3. The conjugate according to claim 1, wherein the heavy chain variable region of the antibody has the amino acid sequence of SEQ ID NO: 9.

4. The conjugate according to claim 1, wherein the heavy chain variable region of the antibody has the amino acid sequence of SEQ ID NO: 10.

5. The conjugate according to claim 1, wherein the light chain variable region of the antibody has the amino acid sequence of SEQ ID NO: 12.

6. The conjugate according to claim 1, wherein the heavy chain variable region of the antibody has the amino acid sequence of SEQ ID NO: 10 and the light chain variable region of the antibody has the amino acid sequence of SEQ ID NO: 12.

7. The conjugate according to claim 1, wherein the antibody comprises an Fc region and is an IgG1-, IgG2-, or IgG4-type antibody.

8. The conjugate according to claim 7, wherein the antibody comprises a glycosylation pattern having one or more of the following characteristics: (i) a detectable amount of glycans carrying a bisecting GlcNAc residue; (ii) a relative amount of glycans carrying at least one galactose residue of at least 25% of the total amount of glycans attached to the Fc glycosylation sites of the antibody in a composition.

9. The conjugate according to claim 1, wherein the heavy chain of the antibody has the amino acid sequence of SEQ ID NO: 15, and the light chain of the antibody has the amino acid sequence of SEQ ID NO: 16.

10. The conjugate according to claim 1, wherein the heavy chain of the antibody has the amino acid sequence of SEQ ID NO: 22 or a variant thereof in which one amino acid has been removed from the C-terminus, and the light chain of the antibody has the amino acid sequence of SEQ ID NO: 16.

11. The conjugate according to claim 1, wherein the CDR-H2 has the amino acid sequence of SEQ ID NO: 8.

12. The conjugate according to claim 1, wherein the heavy chain variable region of the antibody has the amino acid sequence of SEQ ID NO: 11.

13. The conjugate according to claim 12, wherein the light chain variable region of the antibody has the amino acid sequence of SEQ ID NO: 12 or an amino acid sequence which is at least 90% identical to the amino acid sequence of SEQ ID NO: 12.

14. The conjugate according to claim 1, wherein the antibody has the activity of being internalized into MUC1-expressing cells through binding to MUC-1.

15. The conjugate according to claim 1, wherein the cytotoxic agent is an anti-tumor agent.

16. The conjugate according to claim 1, wherein the cytotoxic agent is a chemotherapeutic agent.

17. The conjugate according to claim 16, wherein the chemotherapeutic agent is selected from the group consisting of a microtubule inhibitor, a topoisomerase I inhibitor, a DNA damaging agent, a DNA alkylating agent and a DNA minor groove binder.

18. The conjugate according to claim 17, wherein the chemotherapeutic agent is a topoisomerase I inhibitor.

19. The conjugate according to claim 18, wherein the topoisomerase I inhibitor is an antitumor compound represented by the following formula: ##STR00024##

20. The conjugate according to any claim 1, wherein the antibody is conjugated to the cytotoxic agent via a linker having any structure selected from the group consisting of the following formulas (a) to (f):
-(Succinimid-3-yl-N)—CH.sub.2CH.sub.2—C(═O)-GGFG-NH—CH.sub.2CH.sub.2CH.sub.2—C(═O)—,  (a)
-(Succinimid-3-yl-N)—CH.sub.2CH.sub.2CH.sub.2CH.sub.2CH.sub.2—C(═O)-GGFG-NH—CH.sub.2CH.sub.2CH.sub.2—C(═O)—,  (b)
-(Succinimid-3-yl-N)—CH.sub.2CH.sub.2CH.sub.2CH.sub.2CH.sub.2—C(═O)-GGFG-NH—CH.sub.2—O—CH.sub.2—C(═O)—,  (c)
-(Succinimid-3-yl-N)—CH.sub.2CH.sub.2CH.sub.2CH.sub.2CH.sub.2—C(═O)-GGFG-NH—CH.sub.2CH.sub.2—O—CH.sub.2—C(═O)—,  (d)
-(Succinimid-3-yl-N)—CH.sub.2CH.sub.2—C(═O)—NH—CH.sub.2CH.sub.2O—CH.sub.2CH.sub.2O—CH.sub.2CH.sub.2—C(═O)-GGFG-NH—CH.sub.2CH.sub.2CH.sub.2—C(═O)—, and  (e)
-(Succinimid-3-yl-N)—CH.sub.2CH.sub.2—C(═O)—NH—CH.sub.2CH.sub.2O—CH.sub.2CH.sub.2O—CH.sub.2CH.sub.2O—CH.sub.2CH.sub.2O—CH.sub.2CH.sub.2—C(═O)-GGFG-NH—CH.sub.2CH.sub.2CH.sub.2—C(═O)—, wherein the antibody is connected to the terminus of -(Succinimid-3-yl-N), the cytotoxic agent is connected to the carbonyl group in the rightmost of formulas (a) to (f) with the nitrogen atom of the amino group at position 1 as a connecting position, GGFG represents an amino acid sequence consisting of glycine-glycine-phenylalanine-glycine linked through peptide bonds, and -(Succinimid-3-yl-N)— has a structure represented by the following formula: ##STR00025## which is connected to the antibody at position 3 thereof and is connected to a methylene group in the linker structure containing this structure on the nitrogen atom at position 1.

21. The conjugate according to claim 1, wherein the linker is represented by any formula selected from the group consisting of the following formulas (a) to (c):
-(Succinimid-3-yl-N)—CH.sub.2CH.sub.2CH.sub.2CH.sub.2CH.sub.2—C(═O)-GGFG-NH—CH.sub.2—O—CH.sub.2—C(═O)—,  (a)
-(Succinimid-3-yl-N)—CH.sub.2CH.sub.2CH.sub.2CH.sub.2CH.sub.2—C(═O)-GGFG-NH—CH.sub.2CH.sub.2—O—CH.sub.2—C(═O)—, and  (b)
-(Succinimid-3-yl-N)—CH.sub.2CH.sub.2—C(═O)—NH—CH.sub.2CH.sub.2O—CH.sub.2CH.sub.2O—CH.sub.2CH.sub.2—C(═O)-GGFG-NH—CH.sub.2CH.sub.2CH.sub.2—C(═O)—.  (c)

22. The conjugate according to claim 1, wherein the linker is:
-(Succinimid-3-yl-N)—CH.sub.2CH.sub.2CH.sub.2CH.sub.2CH.sub.2—C(═O)-GGFG-NH—CH.sub.2—O—CH.sub.2—C(═O)—.

23. The conjugate according to claim 1, wherein the antibody is conjugated to a drug linker represented by the following formula (wherein asterisk* represents the point of connection to the antibody) by a thioether bond: ##STR00026##

24. The conjugate according to claim 1, which is represented by the following formula; ##STR00027## wherein AB represents the antibody, y represents an average number of units of the drug-linker structure conjugated to the antibody per itself, and the antibody is conjugated to a drug linker represented by the above formula by a thioether bond.

25. The conjugate according to claim 23, wherein the antibody comprises any one of the following combinations a) to d) of a heavy chain variable region and a light chain variable region, or heavy chain and light chain: (a) the heavy chain variable region has the amino acid sequence of SEQ ID NO: 10 and the light chain variable region has the amino acid sequence of SEQ ID NO: 12, (b) the heavy chain variable region has the amino acid sequence of SEQ ID NO: 11 and the light chain variable region has the amino acid sequence of SEQ ID NO: 12, and (c) the heavy chain has the amino acid sequence of SEQ ID NO: 22 and the light chain has the amino acid sequence of SEQ ID NO: 16.

26. The conjugate according to claim 1, wherein the antibody comprises one or more modifications selected from the group consisting of defucosylation, reduced fucose, N-linked glycosylation, O-linked glycosylation, N-terminal processing, C-terminal processing, deamidation, isomerization of aspartic acid, oxidation of methionine, the substitutions of two leucine (L) residues to alanine (A) at position 234 and 235 of the heavy chain (LALA), amidation of a proline residue, and a deletion or lack of one or two amino acids at the carboxyl terminus.

27. A conjugate, which is represented by any one of the following formulas: ##STR00028## wherein AB represents the antibody, the antibody comprises the heavy chain variable region having the amino acid sequence of SEQ ID NO: 10 and the light chain variable region having the amino acid sequence of SEQ ID NO: 12, y represents an average number of units of the drug-linker structure conjugated to the antibody per itself, and the antibody is conjugated to a drug linker represented by the above formula by a thioether bond; or ##STR00029## wherein AB represents the antibody, the antibody comprises the heavy chain variable region having the amino acid sequence of SEQ ID NO: 11 and the light chain variable region having the amino acid sequence of SEQ ID NO: 12, y represents an average number of units of the drug-linker structure conjugated to the antibody per itself, the antibody is conjugated to a drug linker represented by the above formula by a thioether bond.

28. A conjugate, which is represented by the following formula, ##STR00030## wherein AB represents the antibody, the antibody comprises the heavy chain having the amino acid sequence of SEQ ID NO: 22 or a variant thereof in which one amino acid has been removed from the C-terminus and the light chain having the amino acid sequence of SEQ ID NO: 16, y represents an average number of units of the drug-linker structure conjugated to the antibody per itself, and the antibody is conjugated to a drug linker represented by the above formula by a thioether bond.

29. A pharmaceutical composition comprising the conjugate of claim 28.

30. A pharmaceutical composition comprising a conjugate comprising an antibody conjugated to a cytotoxic agent, wherein the antibody is capable of binding to MUC1, and wherein the antibody comprises (i) a heavy chain variable region comprising the complementarity-determining regions (CDRs) CDR-H1 having the amino acid sequence of SEQ ID NO: 1, CDR-H2 having the amino acid sequence of SEQ ID NO: 2 and CDR-H3 having the amino acid sequence of SEQ ID NO: 3, and (ii) a light chain variable region comprising the complementarity-determining regions (CDRs) CDR-L1 having the amino acid sequence of SEQ ID NO: 4, CDR-L2 having the amino acid sequence of SEQ ID NO: 5 and CDR-L3 having the amino acid sequence of SEQ ID NO: 6.

Description

FIGURES

[0408] FIG. 1 shows ELISA binding curves of the anti-MUC1 antibodies to different MUC1 peptides. (A) shows antigen binding of PankoMab N54Q (PM-N54Q) lacking Fab glycosylation and PankoMab comprising Fab glycosylation (PM) to the MUC1 peptide comprising the epitope sequence PDTR. The threonine of the MUC1 peptide is glycosylated with Tn, sTn, TF or sTF. (B) shows binding of PankoMab and PM-N54Q to the MUC1 peptide comprising the epitope sequence variant PESR. The serine of the MUC1 peptide is glycosylated with Tn. (C) shows binding of PM-N54Q to the MUC1 peptide comprising the epitope sequence PDTR. The threonine of the MUC1 peptide is glycosylated with Tn or not glycosylated. (D) shows binding of several N54X variants to Tn-PDTR MUC1 peptide compared to PankoMab comprising Fab glycosylation diluted from cell culture supernatant of transiently transfected cells. (E) shows binding curves of three purified N54X variants without Fab glycosylation in comparison to PankoMab with Fab glycosylation on Tn-PDTR, TF-PDTR and non-glycosylated PDTR MUC1 peptide. (F) shows binding of two framework variants of PM-N54Q to Tn-PDTR MUC1 peptide compared to PankoMab with Fab glycosylation. For framework variant mf-a nine amino acids are mutated in the VH and three in the VL framework, for mf-b also nine amino acids are mutated in the VH and four in the VL framework.

[0409] FIG. 2 shows surface plasmon resonance (Biacore) binding of the anti-MUC1 antibodies PM and PM-N54Q to a glycosylated PDTR-MUC1 peptide. The maximal binding signal of different concentrations of PM-N54Q and PankoMab are plotted against the antibody concentration.

[0410] FIG. 3 shows results of Fluorescence Proximity Sensing on DRX instrument. Association and dissociation curves are shown. (A) PM with Fab glycosylation compared to (B) PM-N54Q without Fab glycosylation

[0411] FIG. 4 shows an SDS acrylamide gel of an electrophoretic separation of PM-N54Q and PankoMab under non-reducing (left) and reducing (right) conditions. Lane 1: PM-N54Q after capture step; lane 2: PM-N54Q after polishing step; lane 3: PankoMab after capture step; lane 4: PankoMab after polishing step; lane 5: molecular weight marker.

[0412] FIG. 5 shows the Coomassie blue stained gel of an isoelectric focusing assay with PM-N54Q lacking Fab glycosylation and PankoMab being Fab-glycosylated. Lane 1: PankoMab with Fab glycosylation; lane 2: PM-N54Q without Fab glycosylation.

[0413] FIG. 6 shows anti-MUC1 antibody binding to Fcγ receptor IIIa. Increasing concentrations of the antibody PM-N54Q or PankoMab displace rabbit-anti-mouse coupled acceptor beads from FcγRIIIa loaded donor beads, thereby reducing the detected chemiluminescence. In FIG. 6A low-fucosylated antibodies and in FIG. 6B high-fucosylated antibodies were applied into the assay.

[0414] FIG. 7 shows binding of the anti-MUC1 antibodies PM-N54Q, PM-N54D and PM with Fab glycosylation to the tumor cell lines (A) CaOV-3 and (B) HSC-4 as analyzed by flow cytometry.

[0415] FIG. 8 shows cytotoxic activity of A) control hIgG-ADC, naked PankoMab and PankoMab-ADC against cancer cell lines MDA-MB-468 with expression of TA-MUC1 proteins, B) control hIgG-ADC, naked PankoMab and PankoMab-ADC against cancer cell lines HCT-15 without expression of TA-MUC1 proteins, C) control hIgG-ADC, naked PankoMab, PankoMab-ADC, naked PM-N54Q and PM-N54Q-ADC against cancer cell lines NCI-H441 with expression of TA-MUC1 proteins, D) control hIgG-ADC, naked PankoMab PankoMab-ADC, naked PM-N54Q and PM-N54Q-ADC against cancer cell lines HPAC with expression of TA-MUC1 proteins. The cells were treated with each compound for 6 days and cell viability (%) was calculated by ATP assay. Data represent the mean±SD (N=3).

[0416] FIG. 9 shows antitumor efficacy of control hIgG-ADC, naked PankoMab and PankoMab-ADC against MDA-MB-468-bearing nude mice. Control hIgG-ADC, naked PankoMab and PankoMab-ADC at doses of 3 mg/kg or vehicle (acetate buffer solution) was single dose administered intravenously into MDA-MB-468-bearing nude mice (N=6/group). Data of estimated tumor volume represent the mean±SEM. The arrow shows the timing of administration. Estimated tumor volumes at 21 days after the administration of the PankoMab-ADC were compared with that of control hIgG-ADC or that of naked PankoMab treated group by Student t-test. ***P<0.001.

[0417] FIG. 10 shows antitumor efficacy of control hIgG-ADC, naked PankoMab and PankoMab-ADC against HCC70-bearing nude mice. Control hIgG-ADC, naked PankoMab and PankoMab-ADC at doses of 10 mg/kg or vehicle (acetate buffer solution) was single dose administered intravenously into HCC70-bearing nude mice (N=6/group). Data of estimated tumor volume represent the mean±SEM. The arrow shows the timing of administration. Estimated tumor volumes at 21 days after the administration of the PankoMab-ADC were compared with that of the control hIgG-ADC or naked PankoMab treated group by Student t-test. ***P<0.001.

[0418] FIG. 11 shows antitumor efficacy of control hIgG-ADC, naked PM-N54Q, PankoMab-ADC and PM-N54Q-ADC against HPAC-bearing nude mice. Control hIgG-ADC, naked PM-N54Q, PankoMab-ADC and PM-N54Q-ADC at doses of 10 mg/kg or vehicle (acetate buffer solution) was single dose administered intravenously into HPAC-bearing nude mice (N=6/group). Data of estimated tumor volume represent the mean±SEM. The arrow shows the timing of administration. Estimated tumor volumes at 21 days after the administration of PankoMab-ADC and PM-N54Q-ADC were compared with that of the control hIgG-ADC treated group by Dunnett's test. ***P<0.001.

[0419] FIG. 12 shows antitumor efficacy of control hIgG-ADC, naked PankoMab, naked PM-N54Q, PankoMab-ADC and PM-N54Q-ADC against NCI-H441-bearing nude mice. Naked PankoMab, naked PM-N54Q at doses of 10 mg/kg, control hIgG-ADC, PankoMab-ADC and PM-N54Q-ADC at doses of 3 mg/kg or vehicle (acetate buffer solution) was single dose administered intravenously into NCI-H441-bearing nude mice (N=6/group). Data of estimated tumor volume represent the mean±SEM. The arrow shows the timing of administration. Estimated tumor volumes at 31 days after the administration of PankoMab-ADC and PM-N54Q-ADC were compared with that of the control hIgG-ADC treated group by Dunnett's test. ***P<0.001.

[0420] FIG. 13 shows antitumor efficacy of control hIgG-ADC, naked PankoMab, naked PM-N54Q, PankoMab-ADC and PM-N54Q-ADC against OVCAR-5-bearing nude mice. Control hIgG-ADC, naked PankoMab, naked PM-N54Q, PankoMab-ADC and PM-N54Q-ADC at doses of 10 mg/kg or vehicle (acetate buffer solution) was single dose administered intravenously into OVCAR-5-bearing nude mice (N=6/group). Data of estimated tumor volume represent the mean±SEM. The arrow shows the timing of administration. Estimated tumor volumes at 14 days after the administration of PankoMab-ADC and PM-N54Q-ADC were compared with that of the control hIgG-ADC treated group by Dunnett's test. Estimated tumor volumes at 14 days after the administration of PM-N54Q-ADC were compared with that of PankoMab-ADC treated group by Student t-test. ***P<0.001.

[0421] FIG. 14 shows antitumor efficacy of control hIgG-ADC, PankoMab-ADC and PM-N54Q-ADC against HCT-15-bearing nude mice. Control hIgG-ADC, PankoMab-ADC and PM-N54Q-ADC at doses of 10 mg/kg or vehicle (acetate buffer solution) was single dose administered intravenously into HCT-15-bearing nude mice (N=6/group). Data of estimated tumor volume represent the mean±SEM. The arrow shows the timing of administration. Estimated tumor volumes at 21 days after the administration of PankoMab-ADC and PM-N54Q-ADC were compared with that of control hIgG-ADC treated group by Dunnett's test. ***P<0.001.

[0422] FIG. 15 shows the amino acid sequence of the heavy chain of the humanized antibody PM N54Q (SEQ ID No: 15, wherein the amino acid at position 57 is Gln, namely SEQ ID No: 22).

[0423] FIG. 16 shows the amino acid sequence of the light chain of the humanized antibodies PM N54Q and PankoMab (SEQ ID No: 16).

[0424] FIG. 17 shows the amino acid sequence of the heavy chain of the humanized antibody PankoMab (SEQ ID No: 19).

[0425] FIG. 18 shows the amino acid sequence of the heavy chain of chimeric antibody PM N54Q (SEQ ID No: 20, wherein the amino acid at position 76 is Gln, namely SEQ ID No: 23).

[0426] FIG. 19 shows the amino acid sequence of the light chain of chimeric antibody PM N54Q (SEQ ID No: 21).

EXAMPLES

Example 1

[0427] 1. Production of Anti-MUC1 Antibodies

[0428] The nucleic acid sequence of the heavy chain of humanized PankoMab antibody (see, e.g., WO 2011/012309) was modified by mutating the codon for Asn54 according to the Kabat/Eu numbering system (amino acid position 57 in SEQ ID NO: 11) into the codon for any amino acid except Asn, especially for Gln.

[0429] 1) Production of the Anti-MUC1 Antibodies in a Human Myeloid Leukemia Derived Cell Line

[0430] Vectors comprising the coding sequences of the γ1-type heavy chain and the κ-type light chain of the mutated antibodies were transfected into the human myeloid leukemia derived cell line NM-H9D8 (DSM ACC2806). The different αMUC1-antibodies comprising the N54X mutation (PankoMab N54X/PM-N54X, wherein X is any amino acid except N/Asn) or amino acid mutations in the framework sequences of the VH and VL were expressed in the obtained clones, producing the constructs with a human glycosylation pattern. The concentration of the αMUC1-antibodies in the supernatant was determined by Octet measurement using Protein A coated pins or were quantified by UV280 absorbance after purification by protein A chromatography. The binding characteristics of the different αMUC1-antibodies were determined by Antigen-ELISA (see example 2), and selected purified antibodies were also analyzed by Scatchard analysis (see example 3), by Biacore (see example 4a), by DRX.sup.2 switchSENSE® Technology (see example 4b), or by flow cytometry (example 7).

[0431] In addition, PM-N54Q and non-mutated PankoMab with Fab-glycosylation were also expressed in the human myeloid leukemia derived cell line NM-H9D8-E6Q12 (DSM ACC2856) expressing antibody with reduced fucose. Together with the same antibodies expressed in NM-H9D8, these antibodies were purified and analyzed in example 6 for their binding behavior to Fc gamma receptor III A.

[0432] 2) Production of the Anti-MUC1 Antibody in CHO Cell Line

[0433] PM-N54Q encoding sequences (nucleotide sequence of heavy chain of PM-N54Q represented by SEQ ID NO: 17 and nucleotide sequence of light chain of PM-N54Q) represented by SEQ ID NO: 18) which was synthesized by GeneArt™ of ThermoFisher scientific were cloned into expression vectors and resulting plasmids were electro-transfected into CHO cells. Pooled cells grown under selection pressure were applied to manufacture PM-N54Q mutant antibody with general procedures. Anti-MUC 1 antibody (PM-N54Q) produced in CHO cell line was used for Example 8 and 9.

[0434] 2. PankoMab-ADC, N54Q-ADC and DXd

[0435] PankoMab-GEX, which refers to a humanized, anti-TA-MUC1 monoclonal antibody comprising a glycosylation site in CDR2-H2 (Fab glycosylation), and PM-N54Q, which refers to a humanized anti-TA-MUC1 monoclonal lacking Fab glycosylation (Example 1-1), PankoMab-ADC and PM-N54Q-ADC were produced by a known method such as WO 2014/057687 and WO 2015/115091. PankoMab-GEX antibody comprises a heavy chain comprising SEQ ID NO: 19 and a light chain comprising SEQ ID NO: 16, thus the PankoMab-GEX antibody being linked to a drug-linker of Formula 2. PM-N54Q antibody mentioned above comprises a heavy chain comprising SEQ ID NO: 15 and a light chain comprising SEQ ID NO: 16, thus the PM-N54Q antibody being linked to a drug-linker of Formula 2.

##STR00022##

[0436] Such PankoMab-ADC and PM-N54Q-ADC structures show the following Formula 5 (y: The number of conjugated drug molecules per antibody molecule is from 4 to 8, namely, average number of conjugated drug molecules (y) per antibody: approximately 8), AB represents PankoMab or PM-N54Q.

##STR00023##

[0437] Control hIgG-ADC was composed of a humanized IgG1 isotype control monoclonal antibody, not binding to mammalian cells, and the same drug-linker as PnakoMab-ADC and PM-N54Q-ADC. ADC payload (DXd) was produced by a known method such as WO 2014/057687 and WO 2015/115091.

Example 2: Antigen ELISA

[0438] The antigen binding characteristics of PankoMab N54X, wherein the N-glycosylation site in the Fab part is knocked out, was compared to PankoMab having an N-glycosylation site in its Fab part.

[0439] Binding characteristics of the Fab-deglycosylated version of the MUC1-specific antibody PankoMab (PM-N54Q) compared to the (glycosylated) PankoMab-GEX® were analyzed using differently glycosylated and the non-glycosylated MUC1-derived tandem repeat peptides in ELISA studies. In principle, both antibodies show the same gradation by means of binding to glycosylated PDTR peptides (APPAHGVTSAPD-T(X)-RPAPGSTAPPAHGVTSA) (SEQ ID NO: 24) with different glycosylations at T: Strongest binding was observed to the PDTR peptide carrying a Galß1-3GalNAc.sub.alpha (TF) followed by sialylated TF and GalNAc.sub.alpha (Tn) O-glycosylation. Binding to sialylated GalNAc.sub.alpha (sTn) O-glycosylation was significantly lower. As PankoMab-GEX®, PM-N54Q showed only little binding affinity to non-glycosylated MUC1 PDTR peptide indicating adequate tumor specificity (FIG. 1C).

[0440] However, in comparison to PankoMab-GEX® four-fold higher binding was found for PM-N54Q in the TA-MUC1 antigen ELISA using the biotinylated glycopeptide carrying a GalNAc.sub.alpha (Tn) O-glycan. PM-N54Q binds about seven-fold better to the same MUC1 peptide when glycosylated with sialylated GalNAc.sub.alpha (sTn). The binding to Galß1-3GalNAc.sub.alpha (TF) and sialylated TF (sTF) at the threonine of the PDTR-sequence (FIG. 1A) was two-fold better for PM-N54Q.

[0441] Both antibodies show strongly diminished binding to the MUC1 peptide variant APPAHGVTSAPE-S(Tn)-RPAPGSTAPPAHGVTSA (SEQ ID NO: 25) with Tn glycosylation at the serine compared to that at PDT(Tn)R-peptide. However, also here the Fab-deglycosylated PM-N54Q binds significantly stronger than PankoMab-GEX® (FIG. 1B).

[0442] Different other Fab-deglycosylated PM-N54X variants were compared to PankoMab having an N-glycosylation in its Fab part. First, all variants were compared directly from the supernatant, without purification. The concentration was determined by Octet. All PM-N54X variants bound better than Fab-glycosylated PM. In addition, a clear trend depending on the chemical properties of the amino acid side chain was visible.

[0443] Carboxylic acid groups at the side chain showed the lowest binding enhancement. Best binding was observed for amino acids with one or two nitrogens (as primary or secondary amines) (FIG. 1D).

[0444] In addition, selected Fab-deglycosylated variants (PM-N54H, -W, and -Q) were purified by Protein A chromatography and analyzed on ELISA (FIG. 1E). The improvement of binding to TF-MUC1 peptide is about 5- to 8-fold and to Tn-MUC1 peptide about 2- to 3-fold compared to PankoMab with Fab-glycosylation, respectively.

[0445] Furthermore, two different framework variants of the PM-N54Q were analyzed for the binding to the Tn-glycosylated PDTR-MUC1 peptide in ELISA (see FIG. 1F). The framework variant mf-a carries nine amino acid mutations in the VH and three in the VL framework; the variant mf-b carries also nine amino acid mutations in the VH and four in the VL framework. Both mutated variants show similar binding compared to the PM-N54Q antibody.

Example 3: Saturation Binding Analyses of Anti-MUC1 Antibodies to MCF-7 and ZR-75-1 Cells

[0446] Two factors are especially critical for the therapeutic suitability of an antibody: the affinity and number of binding sites of an antibody on tumor cells.

[0447] Binding of the Fab-deglycosylated version of the MUC1-specific antibody PankoMab (PM-N54Q) on TA-MUC-1 positive human tumor cell lines was evaluated using radiolabeled antibodies by saturated binding analysis on the human mamma carcinoma cell lines ZR-75-1 and MCF-7 in comparison to Fab-glycosylated PankoMab-GEX®. The antibodies were chelated with a 12-fold molar excess of p-SCN-Benzyl-DTPA in 50 mM sodium carbonate, 150 mM NaCl, pH 8.7, for 2 h at 37° C., followed by over-night incubation at 2-8° C. Free chelator was removed over desalting columns and dead-end filtration (50 kDa cut-off, 6× buffer exchange to PBS). The chelated antibodies were radiolabeled with carrier-free .sup.111In (2 μCi/μg antibody) for 1 h at 37° C. in 6 mM phosphate, 1.6 mM KCl, 80 mM NaCl, 0.2 M Na-acetate, 0.1 M HCl. The preparations were neutralized by addition of 8-9 fold volume of 10× concentrated PBS. About 1/50 volume of fetal bovine serum were added to the neutralized labelled antibody preparation. Per cell binding approach 1*10.sup.6 cells were used. Several concentrations of labelled antibodies were added to the pelleted cells (30-1000 ng/200 μL in 1% BSA/PBS). The resuspended cell-antibody mixtures were measured in a gamma-counter and incubated 1 h at 4° C. Cells with bound antibodies were separated by centrifugation and washed with 1% BSA/PBS for another hour at 4° C. The cell pellet was then measured for bound .sup.111In-labelled antibody in a gamma counter. Evaluation was performed by “one-site specific ka” in GraphPad Prism. The obtained data are summarized in Table 1. The data show the high affinity and very high number of binding sites of PM-N54Q on these tumor cells. The binding was more than 2.5-fold higher than for PankoMab-GEX® and also the number of binding sites was slightly increased.

TABLE-US-00001 TABLE 1 Association constant and antigen binding sites on MUC1.sup.+ tumor cells K.sub.ass[1/M] ZR-75-1 MCF-7 PM w Fab glyc. 1.2 × 10.sup.7 3 × 10.sup.7 PM N54Q w/o Fab glyc. 3.4 × 10.sup.7 7.8 × 10.sup.7 Binding sites ZR-75-1 MCF-7 PM w Fab glyc. 20 × 10.sup.5 0.6 × 10.sup.5 PM N54Q w/o Fab glyr. 30 × 10.sup.5 0.9 × 10.sup.5

Example 4a: Surface Plasmon Resonance (BiaCore) Analysis

[0448] Binding of the Fab-deglycosylated version of the MUC1-specific antibody PankoMab (PM-N54Q) on TA-MUC-1 derived glycosylated peptide was evaluated by surface plasmon resonance analysis (Biacore). A streptavidin sensor chip was coated with biotinylated TA-MUC1 peptide (Tn glycosylated or not glycosylated). PankoMab and PM-N54Q were diluted sequentially 1:3 from 3,600 to 4.9 nM in HPS-EP. The dilutions were injected at 50 μL/min. Maximal binding of each concentration was determined as response units (RU), respectively, and evaluated with GraphPad Prism using “one-site specific binding”. FIG. 2 shows the obtained binding curves with PM-N54Q compared to PankoMab-GEX®. Affinities (K.sub.D) of 388 nM and 652 nM were calculated for PM-N54Q and PankoMab-GEX®, respectively. Therefore, in this experimental setting a nearly two-fold increase in affinity was detectable.

Example 4b: Fluorescence Proximity Sensing (by DRX.SUP.2., Dynamic Biosensors)

[0449] A new method to determine binding constants and affinity is the fluorescence proximity sensing using single stranded DNA (96mer) spotted on a chip and complementary DNA coupled to a ligand. In the present study streptavidin was used as a ligand to capture biotinylated TA-MUC1 peptides. Binding of PankoMab to the peptides resulted in a fluorescence change. On- and off-rates can be calculated during association and dissociation. Due to a higher sensitivity faster interactions can be monitored compared to surface plasmon resonance. This results in binding kinetics different from SPR but more comparable to the “gold standard” method KinExA, measured in a liquid system.

[0450] PankoMab and PM-N54Q were diluted from 300 nM in 1:9 steps to 3.67 nM in PE140 buffer and applied to the chip-bound peptides. Binding curves were evaluated by mono-exponential global fit (instrument software). Binding curves of PM and PM-N54Q are exemplarily shown in FIGS. 3A and B. Calculated affinities of PankoMab variants are shown in Table 2:

TABLE-US-00002 TABLE 2 Dissociation constants of PankoMab variants to antigen peptide PankoMab variant K.sub.D PM with Fab glycosylation 4.1 nM PM-N54D 1.9 nM PM-N54Q 1.6 nM PM-N54H 0.6 nM

Example 5: Biochemical Characterization

[0451] Non-reducing and reducing SDS-PAGE is used to analyze purity and identity of an antibody. The band pattern in non-reducing gels shows the major band at about 160 kDa and methodical artefacts of heavy and light chains and combinations thereof (˜25, 50-55, 75, 110, 135 kDa). Reducing gels show distinct light and heavy chain bands at and 50-55 kDa. Due to lack of the Fab glycosylation PM-N54Q has a smaller heavy chain, as expected (see FIG. 4, right).

[0452] The charge profile is clearly different, as shown by isoelectric focusing (IEF; see FIG. 5). The Fab glycosylation is considerably sialylated, whereas the Fc glycosylation is only minimally sialylated. Thus PankoMab-GEX® has more charged isoforms than PM-N54Q, reflecting its higher level of negatively charged sialic acids in the Fab part.

Example 6: Fcγ Receptor Binding

[0453] FcγR binding assays for FcγRIIIa (CD16a) are based on the AlphaScreen® technology of PerkinElmer. The AlphaScreen® platform relies on simple bead-based technology of PerkinElmer and is a more efficient alternative to traditional ELISA since no washing steps are necessary.

[0454] For the receptor binding assays, His-tagged FcγRIIIa (Glycotope GmbH) is captured by Ni-chelate donor beads. Anti-MUC1 antibodies and rabbit-anti-mouse coupled acceptor beads compete for binding to FcγR. In case of interaction of FcγR with rabbit-anti-mouse-bound acceptor beads, donor and acceptor beads come into close proximity which leads, upon laser excitation at 680 nm, to light emission. A maximum signal is achieved (signal.sub.max) without a competitor. In case of competition, where a test antibody binds to FcγR, the signal.sub.max is reduced in a concentration-dependent manner. Chemiluminescence was quantified by measurement at 520-620 nm (AlphaScreen® method) using an EnSpire 2300 multilabel reader (PerkinElmer). All results were expressed as the mean±standard deviation of duplicate samples. The data were evaluated and calculated using non-linear curve fitting (sigmoidal dose-response variable slope) with GraphPad Prism 5 software. As a result, a concentration dependent sigmoidal curve was obtained, which is defined by top-plateau, bottom-plateau, slope and EC.sub.50.

[0455] As shown in FIGS. 6A and B, the FcγRIIIa binding affinity was comparable for PankoMab N54Q and PankoMab whereby in Figure A low-fucosylated antibodies and in Figure B high-fucosylated antibodies were applied into the assay. Hence, removal of the Fab glycosylation did not affect receptor interaction of the antibody.

Example 7: Binding to Cellular TA-MUC1

[0456] N54Q and N54D were transiently expressed and purified by protein A chromatography. Binding of the two variants to cell surface TA-MUC1 was compared to PM with Fab glycosylation using two different carcinoma cell lines. The tongue squamous cell carcinoma line HSC-4 expresses TA-MUC1 to a medium degree and the ovarian carcinoma cell line CaOV-3 to a high degree. Tumor cells were incubated with antibodies in serial dilutions and bound antibodies were detected using a Phycoerythrin-conjugated goat anti-human IgG (heavy and light chain) antibody. A human IgG control was included to control for background staining. Binding was analyzed by flow cytometry.

[0457] The analyzed constructs PM, PM-N54Q and PM-N54D show strong and specific binding to the TA-MUC1 expressing HSC-4 and CaOV-3 cells compared to a human IgG1 control (FIG. 7). The binding of PM-N54D to the TA-MUC1.sup.high CaOV-3 cells was comparable to PM with Fab glycosylation while PM-N54Q showed a slightly better binding (FIG. 7A). Using HSC-4 carcinoma cells that express TA-MUC1 at an intermediate level, the variant PM-N54Q was clearly superior in binding to cellular TA-MUC1 compared to PM while PM-N54D showed an inferior binding compared to PM with Fab glycosylation (FIG. 7B).

Example 8: Evaluation of In Vitro Efficacy of PankoMab-ADC and PM-N54Q-ADC

[0458] 8.1 Cell Lines

[0459] The human breast cancer cell line MDA-MB-468, the human pancreatic cancer cell line HPAC, and the human lung cancer cell line NCI-H441 were used as TA-MUC1 medium to high-expressing cells. The human colorectal cancer cell line HCT-15 was used as TA-MUC1 negative cells. These cell lines were purchased from ATCC. Each cell line was cultured in accordance with an instruction manual. Expression level of TA-MUC1 on each cancer cell line was confirmed by flow cytometry.

[0460] 8.2 Evaluation of In Vitro Efficacy of PankoMab-ADC

[0461] MDA-MB-468 suspension was prepared to have a concentration of 1.25×10.sup.4 cells/mL by using culture medium, and added to each well of a black clear bottom 96-well plate at 80 uL/well (1000 cells/well). For blank wells, the medium alone was added to the wells at 80 uL/well (N=3). All cells were incubated overnight in the appropriate condition for MDA-MB-468.

[0462] HCT-15 suspension was prepared to have a concentration of 3.1×10.sup.3 cells/mL by using culture medium, suspension was added to each well of a black clear bottom 96-well plate at 80 uL/well (250 cells/well). For blank wells, the medium alone was added to the wells at 80 uL/well (N=3). All cells were incubated overnight in the appropriate condition for HCT-15.

[0463] On the next day, each naked PankoMab, control hIgG-ADC, and PankoMab-ADC was 3-fold serially diluted with the each culture medium from 500 nM to 0.2 nM. Twenty microliters of these diluted solutions were added to the appropriate wells (final concentration: 100 nM to 0.04 nM). For blank wells and untreated wells, 20 uL of the each culture medium alone was added to the wells. All plates were incubated for 6 days in the appropriate condition for each cell line.

[0464] After the incubations, the amount of ATP in each well was measured by using a CellTiter-Glo Luminescent Cell Viability Assay (Promega). Luminescence was measured by a multilabel counter (ARVO X3, PerkinElmer Japan Co., Ltd.). This assay was performed in triplicate.

[0465] The cell viability of each sample was calculated by the following equation:

Cell viability (%)=100×(T−B)/(C−B)

[0466] T: the luminescence intensity of the test well

[0467] C: mean luminescence intensity of untreated wells

[0468] B: mean luminescence intensity of blank wells

[0469] 8.3 Comparison of In Vitro Efficacy Between PankoMab-ADC and PM-N54Q-ADC

[0470] HPAC suspension was prepared to have a concentration of 1.25×10.sup.4 cells/mL by using culture medium, and added to each well of a black clear bottom 96-well plate at 80 uL/well (1000 cells/well). For blank wells, the medium alone was added to the wells at 80 uL/well (N=3). All cells were incubated overnight in the appropriate condition for H PAC.

[0471] NCI-H441 suspension was prepared to have a concentration of 1.25×10.sup.4 cells/mL by using culture medium, and added to each well of a black clear bottom 96-well plate at 80 uL/well (1000 cells/well). For blank wells, the medium alone was added to the wells at 80 uL/well (N=3). All cells were incubated overnight in the appropriate condition for NCI-H441.

[0472] On the next day, each naked PankoMab, naked PM-N54Q, hIgG-ADC, PankoMab-ADC, and PM-N54Q-ADC was 3-fold serially diluted with the each culture medium from 500 nM to 0.2 nM. Twenty microliters of these diluted solutions were added to the appropriate wells (final concentration: 100 nM to 0.04 nM). For blank wells and untreated wells, 20 uL of the each culture medium alone was added to the wells. All plates were incubated for 6 days in the appropriate condition for each cell line.

[0473] After the incubations, the amount of ATP in each well was measured by using a CellTiter-Glo Luminescent Cell Viability Assay (Promega). Luminescence was measured by a multilabel counter (ARVO X3, PerkinElmer Japan Co., Ltd.). This assay was performed in triplicate.

[0474] The cell viability of each sample was calculated by the following equation:

Cell viability (%)=100×(T−B)/(C−B)

[0475] T: the luminescence intensity of the test well

[0476] C: mean luminescence intensity of untreated wells

[0477] B: mean luminescence intensity of blank wells

[0478] Potency ratio of cytotoxic activity of PankoMab-ADC vs PM-N54Q-ADC against HPAC and NCI-H441, and their 95% CIs were calculated as post-hoc analysis using a 3-parameter logistic parallel-line analysis (common slope) by using EXSUS ver.8.1 (CAC Croit, Tolyo, Japan) based on SAS release 9.4 (SAS Institute Japan, Tokyo, Japan) (Emax: 100, Emin: estimate). The difference in the potency of cytotoxic activity was considered to be significant if the 95% CI of potency ratio excluded 1.

Example 9: Evaluation of In Vivo Efficacy of PankoMab-ADC and PM-N54Q-ADC

[0479] 9.1 Cell Lines

[0480] The human breast cancer cell line MDA-MB-468 and HCC70, the human pancreatic cancer cell line HPAC, and the human lung cancer cell line NCI-H441 were used as TA-MUC1 medium to high-expressing tumor cells. The human colorectal cancer cell line HCT-15 was used as TA-MUC1 negative tumor cells. These cell lines were purchased from ATCC. The human ovarian cancer cell line OVCAR-5 was purchased from National Cancer Institute and used as TA-MUC1 low-expressing tumor cells. Each cell line was cultured in accordance with an instruction manual. Expression level of TA-MUC1 on each cancer cell line was confirmed by flow cytometry and IHC staining.

[0481] 9.2 Evaluation of In Vivo Efficacy of PankoMab-ADC

[0482] MDA-MB-468 cells were suspended in Matrigel (BD), and 1×10.sup.7 cells were subcutaneously transplanted to the right side of the body of each female nude mice (Day 0), and the mice were randomly grouped on Day 20 (N=6). After grouping, each naked PankoMab, control hIgG-ADC, or PankoMab-ADC solution was single dose administered intravenously at a dose of 3 mg/kg. A vehicle (acetate buffer solution) administration group was established as a control group. After administration, the tumor length and width of each mouse were measured with the digital caliper twice a week for 21 days.

[0483] HCC70 cells were suspended in physiological saline (Otsuka Pharmaceutical Factory, Inc.) and 1×10.sup.7 cells were subcutaneously transplanted to the right side of the body of each female nude mice (Day 0), and the mice were randomly grouped on Day 19 (N=6). After grouping, each naked PankoMab, control hIgG-ADC, or PankoMab-ADC solution was single dose administered intravenously at a dose of 10 mg/kg. A vehicle (acetate buffer solution) administration group was established as a control group. After administration, the tumor length and width of each mouse were measured with the digital caliper twice a week for 21 days.

[0484] The estimated tumor volume of each mouse was calculated by the following equation:

Estimated tumor volume (mm.sup.3)=½×length (mm)×width(mm).sup.2

[0485] The tumor growth inhibition (TGI, %) of each group on the last measurement day of vehicle treated groups was also calculated according to the following equation, and rounded to an integer.

TGI (%)=(1−T/C)×100

[0486] T: the mean estimated tumor volume (mm.sup.3) of the naked PankoMab, control hIgG-ADC, or PankoMab-ADC

[0487] C: the mean estimated tumor volume (mm.sup.3) of the vehicle treated group

[0488] In order to evaluate the anti-tumor efficacy of PankoMab-ADC, tumor volumes of each mouse on the last measurement day of PankoMab-ADC treated groups (MDA-MB-468: Day 41, HCC70: Day 40) were compared with that of the control hIgG-ADC treated groups or that of naked PankoMab treated group by Student t-test. All statistical analyses were performed as post-hoc analysis using SAS System Release 9.2 (SAS Institute Inc.). A P value of less than 0.05 was considered to be statistically significant.

[0489] 9.3 Comparison of In Vivo Efficacy of PankoMab-ADC and PM-N54Q-ADC

[0490] HPAC cells were suspended in physiological saline (Otsuka Pharmaceutical Factory, Inc.) and 3×10.sup.6 cells were subcutaneously transplanted to the right side of the body of each female nude mice (Day 0), and the mice were randomly grouped on Day 11 (N=6). After grouping, each naked PM-N54Q, control hIgG-ADC, PankoMab-ADC, or PM-N54Q-ADC solution was single dose administered intravenously at a dose of 10 mg/kg. A vehicle (acetate buffer solution) administration group was established as a control group. After administration, the tumor length and width of each mouse were measured with the digital caliper twice a week for 21 days.

[0491] NCI-H441 cells were suspended in Matrigel (BD), and 5×10.sup.6 cells were subcutaneously transplanted to the right side of the body of each female nude mice (Day 0), and the mice were randomly grouped on Day 7 (N=6). After grouping, each naked PankoMab or naked PM-N54Q solutions was single dose administered intravenously at a dose of 10 mg/kg, and control hIgG-ADC, PankoMab-ADC, or PM-N54Q-ADC solution was single dose administered intravenously at a dose of 3 mg/kg. A vehicle (acetate buffer solution) administration group was established as a control group. After administration, the tumor length and width of each mouse were measured with the digital caliper twice a week for 31 days.

[0492] OVCAR-5 cells were suspended in physiological saline (Otsuka Pharmaceutical Factory, Inc.) and 5×10.sup.6 cells were subcutaneously transplanted to the right side of the body of each female nude mice (Day 0), and the mice were randomly grouped on Day 12 (N=6). After grouping, naked PankoMab, naked PM-N54Q, control hIgG-ADC, PankoMab-ADC, or PM-N54Q-ADC solution was single dose administered intravenously at a dose of 10 mg/kg. A vehicle (acetate buffer solution) administration group was established as a control group. After administration, the tumor length and width of each mouse were measured with the digital caliper twice a week for 21 days.

[0493] HCT-15 cells were suspended in physiological saline (Otsuka Pharmaceutical Factory, Inc.) and 5×10.sup.6 cells were subcutaneously transplanted to the right side of the body of each female nude mice (Day 0), and the mice were randomly grouped on Day 10 (N=6). After grouping, control hIgG-ADC, PankoMab-ADC, or PM-N54Q-ADC solution was single dose administered intravenously at a dose of 10 mg/kg. A vehicle (acetate buffer solution) administration group was established as a control group. After administration, the tumor length and width of each mouse were measured with the digital caliper twice a week for 21 days.

[0494] Tumor volume of each mouse was calculated by the following equation:

Estimated tumor volume (mm.sup.3)=½×length (mm)×width (mm).sup.2

[0495] The tumor growth inhibition (TGI, %) of each mouse on the last measurement day of vehicle treated groups or the last day that all groups are alive was also calculated according to the following equation, and rounded to an integer.

TGI (%)=(1−T/C)×100

[0496] T: the mean estimated tumor volume (mm.sup.3) of the naked PankoMab, naked PM-N54Q, control hIgG-ADC, PankoMab-ADC or PM-N54Q-ADC

[0497] C: the mean estimated tumor volume (mm.sup.3) of the vehicle treated group

[0498] In order to evaluate the anti-tumor efficacy of each compound against HPAC, NCI-H441, OVCAR-5, and HCT-15-bearing mice, tumor volumes of each mouse on the last measurement day of control hIgG-ADC treated groups (H PAC: Day 32, NCI-H441: Day 38, HCT-15: Day 32) or the last day that all groups are alive (OVCAR-5: Day 26) were compared with that of the control hIgG-ADC treated groups by Dunnett's test. In addition, tumor volumes of OVCAR-5-bearing nude mice on Day 33 were compared by Student's t-test between PankoMab-ADC and PM-N54Q-ADC treated groups. All statistical analyses were performed as post-hoc analysis using SAS System Release 9.2 (SAS Institute Inc.). A P value of less than 0.05 was considered to be statistically significant.

Example 10: Results

[0499] 10.1 Cytotoxic Activity of PankoMab-ADC Against TA-MUC1 Positive Cancer Cell Lines and Negative Cells In Vitro

[0500] To investigate whether PankoMab-ADC shows the target-dependent and drug-dependent cytotoxic activity against human cancer cell lines, in vitro efficacy of naked PankoMab, control hIgG-ADC, and PankoMab-ADC against the human breast cancer cells MDA-MB-468 (TA-MUC1 positive) and the human colorectal cancer cells HCT-15 (TA-MUC1 negative) was evaluated. As shown in FIG. 8 naked PankoMab and hIgG-ADC showed little activity against each cell line (IC.sub.50>100 nM). Under these conditions, PankoMab-ADC exhibited dose-dependent cytotoxic activity against TA-MUC1 positive cells MDA-MB-468 (FIG. 8A, IC.sub.50<10 nM). But it didn't show the activity against TA-MUC1 negative cells HCT-15 (FIG. 8B, IC.sub.50>100 nM). Based on these results, it was concluded that PankoMab-ADC shows target-dependent and drug-dependent cytotoxicactivity against TA-MUC1 positive cancer cell lines in vitro.

[0501] 10.2 Comparison of the In Vitro Cytotoxic Activity Between PankoMab-ADC and PM-N54Q-ADC Against TA-MUC1 Positive Cells In Vitro

[0502] To investigate whether improvement of antigen binding affinity may contribute to enhancement of cytotoxic activity, in vitro efficacy of PankoMab-ADC and PM-N54Q-ADC against the human pancreatic cancer cell line HPAC and the human lung cancer cell line NCI-H441 was evaluated. The cytotoxic activity of PM-N54Q-ADC against them was more than 1.5-fold potent than that of PankoMab-ADC (FIG. 8C and FIG. 8D). The potency ratio of PM-N54Q-ADC to PankoMab-ADC against HPAC was 1.917 (1.611−2.280, 95% CI), and that against NCI-H441 was 1.663 (1.495 to 1.849, 95% CI at EC50). These data demonstrated that cytotoxic activity of PM-N54Q-ADC is significantly more potent than that of PankoMab-ADC. These results suggest that improvement of antigen binding affinity of PankoMab-ADC may contribute to significant enhancement of cell killing activity.

[0503] 10.3 Anti-Tumor Efficacy of PankoMab-ADC Against TA-MUC1 Positive Tumor

[0504] To investigate whether PankoMab-ADC shows not only in vitro but also in vivo efficacy, anti-tumor efficacy of naked PankoMab, control hIgG-ADC, and PankoMab-ADC against MDA-MB-468-bearing mice was evaluated. As shown in FIG. 9, naked PankoMab and control IgG-ADC (3 mg/kg, single administration) didn't show anti-tumor efficacy (both of TGIs were −18% on Day 41). By contrast, PankoMab-ADC (3 mg/kg, single administration) remarkably inhibited the tumor growth (TGI was 97% on Day 41). Moreover, it showed significant anti-tumor efficacy compared to control hIgG-ADC and naked PankoMab (both of P<0.001 on Day 41). In terms of body weight change, any body weight loss caused by drug treatment was not observed in all drug-treatment groups.

[0505] Anti-tumor efficacy of naked PankoMab, control hIgG-ADC, and PankoMab-ADC against HCC70-bearing mice was also evaluated. As shown in FIG. 10, naked PankoMab and control hIgG-ADC (10 mg/kg, single administration) showed weak anti-tumor efficacy against these xenograft models (TGI was 10% and 29% on Day 40, respectively). By contrast, PankoMab-ADC (10 mg/kg, single administration) remarkably inhibited the tumor growth (TGI was 95% on Day 40). Moreover, it showed statistically significant anti-tumor efficacy compared to control hIgG-ADC (both of P<0.001 on Day 40). In terms of body weight change, any body weight loss caused by drug treatment was not observed in all drug-treatment groups. These results suggest that PankoMab-ADC has strong anti-tumor efficacy and it showed target-dependent and drug-dependent anti-tumor efficacy against various TA-MUC1 positive xenograft models.

[0506] 10.4 Comparison of the Anti-Tumor Efficacy Between PankoMab-ADC and PM-N54Q-ADC Against TA-MUC1 Positive Tumor In Vivo

[0507] To investigate whether PM-N54Q-ADC has equal to or greater anti-tumor efficacy against TA-MUC1 positive tumor cells than PankoMab-ADC, anti-tumor efficacy of PankoMab-ADC and PM-N54Q-ADC against various types of TA-MUC1 positive tumor cells was compared.

[0508] At first, we evaluated the in vivo efficacy against HPAC and NCI-H441 tumor cells with medium to high TA-MUC1 expression. As shown in FIG. 11, naked PM-N54Q and control hIgG-ADC (10 mg/kg, single administration) showed weak anti-tumor efficacy against HPAC-bearing mice (TGI was 27% and 18% on Day 32, respectively). By contrast, PankoMab-ADC and PM-N54Q-ADC (10 mg/kg, single administration) remarkably inhibited the tumor growth (both of TGIs were 93% on Day 32). Moreover, PankoMab-ADC and PM-N54Q-ADC (10 mg/kg, single administration) showed statistically significant anti-tumor efficacy compared to control hIgG-ADC (both of P<0.001 on Day 32). In terms of body weight change, any body weight loss caused by drug treatment was not observed in all drug-treatment groups.

[0509] As shown in FIG. 12, naked PankoMab and naked PM-N54Q (10 mg/kg, single administration) showed weak anti-tumor efficacy against NCI-H441-bearing mice (TGI was 8% and 12% on Day 38, respectively). Although control hIgG-ADC (3 mg/kg, single administration) treated group showed anti-tumor efficacy for two weeks after administration, tumor regrowth was observed after day 21 (TGI was 71% on Day 38). By contrast, PankoMab-ADC and PM-N54Q-ADC (3 mg/kg, single administration) remarkably inhibited the tumor growth (both of TGI was 99% on Day 38). Moreover, PankoMab-ADC and PM-N54Q-ADC showed statistically significant anti-tumor efficacy compared to control hIgG-ADC (P<0.001 on Day 38, respectively). In terms of body weight change, any body weight loss caused by drug treatment was not observed in all drug-treatment groups.

[0510] Next, we evaluated the in vivo efficacy against OVCAR-5 tumor cells in which TA-MUC1 low expression. As shown in FIG. 13, naked PankoMab, PM-N54Q and control hIgG-ADC (10 mg/kg, single administration) showed little anti-tumor efficacy against OVCAR5-bearing mice (TGI was 1%, 11% and 3% on Day 26, respectively). In this model, anti-tumor efficacy of PankoMab-ADC (10 mg/kg, single administration) was limited (TGI was 37% on Day 26), but PM-N54Q-ADC (10 mg/kg, single administration) showed strong anti-tumor efficacy (TGI was 73% on Day 26). Moreover, PankoMab-ADC and PM-N54Q-ADC showed statistically significant anti-tumor efficacy compared to control hIgG-ADC (P=0.01 and P<0.001 on Day 26, respectively). In addition, PM-N54Q-ADC showed statistically significant anti-tumor efficacy compared to PankoMab-ADC (P<0.001 on Day 26). In terms of body weight change, any body weight loss caused by drug treatment was not observed in all drug-treatment groups.

[0511] Finally, we evaluated the in vivo efficacy against HCT-15 tumor cells which TA-MUC1 negative.

[0512] As shown in FIG. 14, naked PankoMab and PM-N54Q (10 mg/kg, single administration) showed little anti-tumor efficacy against this model (TGI was 7%, 4% on Day 32, respectively). Moreover, PankoMab-ADC, PM-N54Q-ADC and control hIgG-ADC also showed little anti-tumor efficacy against this model (TGI was 15%, 22%, and 26% on Day 32, respectively).

[0513] Based on these results, it was concluded that the anti-tumor efficacy of PankoMab-ADC and PM-N54Q-ADC is target-dependent and drug-dependent. And, improvement of antigen binding affinity may contribute to enhancement of anti-tumor efficacy against TA-MUC1 positive tumor cells.

[0514] Identification of the Deposited Biological Material

[0515] The cell lines DSM ACC 2806, DSM ACC 2807 and DSM ACC 2856 were deposited at the DSMZ—Deutsche Sammlung von Mikroorganismen and Zellkulturen GmbH, Inhoffenstraße 7B, 38124 Braunschweig (DE) by Glycotope GmbH, Robert-Rössle-Str. 10, 13125 Berlin (DE) on the dates indicated in the following table.

TABLE-US-00003 Name of the Accession Date of Cell Line Number Depositor Deposition NM-H9D8 DSM ACC 2806 Glycotope GmbH Sep. 15, 2006 NM-H9D8-E6 DSM ACC 2807 Glycotope GmbH Oct. 5, 2006 NM-H9D8-E6Q12 DSM ACC 2856 Glycotope GmbH Aug. 8, 2007