Antibodies against phosphorylcholine

Abstract

The present invention relates to an antibody or antibody fragment capable of binding to phosphorylcholine and/or a phosphorylcholine conjugate, wherein the antibody or antibody fragment comprises a variable heavy chain (VH) domain and/or a variable light chain (VL) domain, and wherein—(a) the VH domain comprises an amino acid sequence that includes one, two or three complementarity determining regions (CDRs) selected from the group consisting of: a CDR1 sequence comprising an amino acid sequence having at least 25%, 50%, 75% or 100% sequence identity to the sequence of SEQ ID NO: 17; a CDR2 sequence comprising an amino acid sequence having at least 5%, 11%, 17%, 23%, 29%, 35%, 47%, 52%, 58%, 64%, 70%, 76%, 82%, 94% or 100% sequence identity to the sequence of SEQ ID NO: 18; and a CDR3 sequence comprising an amino acid sequence having at least 4%, 9%, 13%, 18%, 22%, 27%, 31%, 36%, 40%, 45%, 50%, 54%, 59%, 63%, 68%, 72%, 77%, 81%, 86%, 90%, 95% or 100% sequence identity to the sequence of SEQ ID NO: 19, 20, 21 or 22; and/or (b) the VL domain comprises an amino acid sequence that includes one, two or three complementarity determining regions (CDRs) selected from the group consisting of: a CDR4 sequence comprising an amino acid sequence having at least 5%, 11%, 17%, 23%, 29%, 35%, 47%, 52%, 58%, 64%, 70%, 76%, 82%, 94% or 100% sequence identity to the sequence of SEQ ID NO: 23 or 24; a CDR5 sequence comprising an amino acid sequence having at least 14%, 28%, 42%, 57%, 71%, 85% or 100% sequence identity to the sequence of SEQ ID NO: 25; a CDR6 sequence comprising an amino acid sequence having at least 11%, 22%, 33%, 44%, 55%, 66%, 77%, 88% or 100% sequence identity to the sequence of SEQ ID NO: 26.

Claims

1. An antibody or antibody fragment capable of binding to phosphorylcholine and/or a phosphorylcholine conjugate, wherein the antibody or antibody fragment comprises a variable heavy chain (VH) domain and/or a variable light chain (VL) domain, and wherein: (a) the VH domain comprises an amino acid sequence that comprises a complementarity determining region (CDR)1 having SEQ ID NO: 17, a CDR2 having SEQ ID NO: 18 and a CDR3 having SEQ ID NO: 19; and (b) the VL domain comprises an amino acid sequence that comprises a complementarity determining region (CDR)4 having SEQ ID NO: 23, a CDR5 having SEQ ID NO: 25 and a CDR6 having SEQ ID NO: 26.

2. The antibody or antibody fragment according to claim 1, wherein: the VH domain comprises an amino acid sequence having at least 80%, 85%, 90%, 95% or 100% sequence identity SEQ ID NO:1; and the VL domain comprises an amino acid sequence having at least 80%, 85%, 90%, 95% or 100% sequence identity SEQ ID NO: 2.

3. The antibody or antibody fragment according to claim 1, wherein the VH domain, the VL domain, or preferably both of the VH and VL domains, comprise an amino acid sequence having 100% sequence identity to the, or one or more (such as all) of SEQ ID NO: 1 and SEQ ID NO: 2, respectively.

4. The antibody or antibody fragment according to claim 1, wherein the VH domain, the VL domain, or both of the VH and VL domains, comprise an amino acid sequence having less than 100%, but at least 80%, 85%, 90%, 95%, sequence identity to the, or one or more (such as all) of SEQ ID NO: 1 and SEQ ID NO: 2, respectively.

5. The antibody or antibody fragment according to claim 1, and wherein the ability of the antibody or antibody fragment to bind to phosphorylcholine and/or a phosphorylcholine conjugate is equivalent to (that is, at least 80%, 85%, 90% or 95%, of), or greater than, the ability of a corresponding antibody or antibody fragment, wherein the VH domain and the VL domain of the corresponding antibody or antibody fragment each comprise an antigen-binding sequence comprising an amino acid sequence having 100% sequence identity to the, or SEQ ID NO: 1 and SEQ ID NO: 2, respectively.

6. The antibody or antibody fragment according to claim 1, wherein the VH domain and the VL domain are present in a linear polypeptide sequence.

7. The antibody or antibody fragment according to claim 1, wherein the VH domain and the VL domain are each present in a separate polypeptide sequence, and preferably wherein the separate polypeptide sequence are directly or indirectly bound together (such as by one or more disulphide bonds between the separate polypeptide sequence).

8. The antibody according to claim 1, wherein the antibody is a monoclonal antibody.

9. The antibody fragment according to claim 1, wherein the antibody fragment is a single chain antibody, Fv, scFv, Fab, F(ab′)2, Fab′, scFv-Fc fragment, diabody, or any such fragment that has been stabilized such as by PEGylation.

10. The antibody or antibody fragment according to claim 1, which is a human or humanized antibody or antibody fragment, such as a human or humanized monoclonal antibody.

11. The antibody or antibody fragment according to claim 1, which is capable of binding to a phosphorylcholine conjugate.

12. The antibody or antibody fragment according to claim 11, wherein the phosphorylcholine conjugate is a phosphorylcholine moiety linked to a carrier, optionally via a spacer, and preferably the antibody or antibody fragment binds specifically to the phosphorylcholine moiety in the phosphorylcholine conjugate.

13. A pharmaceutical composition comprising or consisting essentially of an antibody or an antibody fragment according to claim 1, and a pharmaceutically acceptable carrier or excipient.

14. A nucleic acid sequence encoding an antibody or an antibody fragment according to claim 1.

15. A vector or plasmid comprising the nucleic acid sequence of claim 14.

16. A host cell comprising the nucleic acid sequence of claim 14.

17. The host cell of claim 16, wherein the cell is a prokaryotic cell, such as an Escherichia coli cell, or a eukaryotic cell, such as animal, plant, or fungal cell.

18. The host cell of claim 16, which expresses the nucleic acid sequence to produce an antibody or an antibody fragment.

19. A method of producing an antibody or an antibody fragment comprising culturing a host cell according to claim 18, and recovering therefrom said antibody or antibody fragment.

Description

DESCRIPTION OF THE DRAWINGS

(1) FIG. 1. Estimates of Binding Affinity from an Equilibrium Binding Analysis by Biacore.

(2) (.diamond-solid.) M99-B05 (lot W21573) (Kd=160±32 nM), (◯) M99-B05 (lot W22595) (Kd=148±8 nM). The panel compares these two different preparations of the antibody.

(3) FIG. 2. Purified IgGs Binding to PC-BSA as Measured by ELISA.

(4) (.circle-solid.) M4-G02 (EC.sub.50=0.14 nM), (◯) M73-G03 (EC.sub.50=0.91 nM), (Δ) M99-B05 (EC.sub.50=0.11 nM). The data were fit to a 4 parameter logistic equation with a global B.sub.max to obtain EC50 value estimates.

(5) FIG. 3. Inhibition of CD45 Positive Leukocyte Influx into Medial in Femoral Artery Cuffed Mice.

(6) Transgenic male ApoE*3 Leiden mice were fed a high-cholesterol and high-fat diet containing 1% cholesterol and 0.05% cholate to induce hypercholesterolemia. After three weeks of the high fat diet, mice were anesthetized and the femoral artery was dissected from its surroundings and loosely sheathed with a non-constrictive polyethylene cuff (Portex, 0.40 mm inner diameter, 0.80 mm outer diameter and 2.0 mm length). Mice were treated with either 10 mg/kg recombinant anti-PC IgG antibodies dissolved in PBS, 10 mg/kg anti-streptavidin A2 IgG antibodies dissolved in PBS or PBS only through IP injection on day 0. Mice were sacrificed three days after surgery and cuffed femoral arteries were harvested and paraffin-embedded. Serial cross-sections (5 μm) were taken from the entire length of the cuffed femoral artery segment for histochemical analysis. *p<0.01, n=15.

(7) FIG. 4. Inhibition of Intimal Thickening in Femoral Artery Cuffed Mice.

(8) Transgenic male ApoE*3 Leiden mice were fed a high-cholesterol and high-fat diet containing 1% cholesterol and 0.05% cholate to induce hypercholesterolemia. After three weeks of the high fat diet, mice were anesthetized and the femoral artery was dissected from its surroundings and loosely sheathed with a non-constrictive polyethylene cuff (Portex, 0.40 mm inner diameter, 0.80 mm outer diameter and 2.0 mm length). Mice were treated with either 10 mg/kg recombinant anti-PC IgG antibodies dissolved in PBS, 10 mg/kg anti-streptavidin A2 IgG antibodies dissolved in PBS or PBS only through IP injection on day 0, 3, 7, and 10 after surgery. Mice were sacrificed 14 days after surgery and cuffed femoral arteries were harvested and paraffin-embedded. Serial cross-sections (5 μm) were taken from the entire length of the cuffed femoral artery segment for histochemical analysis.

(9) A. Comparison of the intimal area (indicated by the arrow) in the 3 panels indicates that the antibodies M99-B05 reduced the intimal thickening that was observed 14 days after cuff-induced vascular injury.

(10) B. Intimal thickening in (μm).sup.2, n=10, *p<0.05

(11) FIG. 5. PC Binding Activity of M99-B05 Mutants Measured Using ELISA

(12) (.circle-solid.) M99-B05 (EC.sub.50=0.28 nM), (◯) X19-A01 (EC.sub.50=0.42 nM), (.Math.) X19-A03 (EC.sub.50=0.54 nM), (Δ) X19-A05 (EC.sub.50=0.52 nM), (.square-solid.) X19-A07 (EC.sub.50=0.62 nM), (□) X19-A09 (EC.sub.50=0.58 nM), (.diamond-solid.) X19-A11 (EC.sub.50=0.97 nM), (⋄) X19-C01(EC.sub.50=1.4 nM).

(13) FIG. 6. Immunohistochemistry Staining of Frozen Human Atherosclerotic Lesion Tissue with an Anti-Phosphorylcholine Antibody.

(14) Human atherosclerotic lesion tissue, along with a normal tissue control was obtained commercially from Biochain Human frozen tissues. The tissue was incubated with 0.1 μg/mL biotinylated M99-B05 anti-phosphorylcholine IgG overnight at 4° C. Antibody binding to tissue was visualized following the addition of streptavidin-horse radish peroxidase (HRP) and HRP substrate. The presence of antibody binding is show by the color that is generated from the HRP substrate. No binding was observed with an isotype control (data not shown).

(15) FIG. 7. Inhibition of Intimal Thickening in Femoral Artery Cuffed Mice.

(16) Transgenic male ApoE*3 Leiden mice were fed a high-cholesterol and high-fat diet containing 1% cholesterol and 0.05% cholate to induce hypercholesterolemia. After three weeks of the high fat diet, mice were anesthetized and the femoral artery was dissected from its surroundings and loosely sheathed with a non-constrictive polyethylene cuff (Portex, 0.40 mm inner diameter, 0.80 mm outer diameter and 2.0 mm length). Mice were treated with either the indicated antibody and amount dissolved in PBS by IP injection on day 0, 3, 7, and 10 after surgery. Mice were sacrificed 14 days after surgery and cuffed femoral arteries were harvested and paraffin-embedded. Serial cross-sections (5 μm) were taken from the entire length of the cuffed femoral artery segment for histochemical analysis and the intimal thickening in (μm).sup.2 calculated, n=10, *p<0.05.

EXAMPLES

(17) The following examples are included to further illustrate various aspects of the invention. It should be appreciated by those of skill in the art that the techniques disclosed in the examples which follow represent techniques and/or compositions discovered by the inventor to function well in the practice of the invention, and thus can be considered to constitute preferred modes for its practice. However, those of skill in the art should, in light of the present disclosure, appreciate that many changes can be made in the specific embodiments which are disclosed and still obtain a like or similar result without departing from the spirit and scope of the invention.

(18) Screening of Phage Display Antibody Library

(19) A phage display selection and screening campaign to identify human antibodies that bind PC and neutralize the pro-inflammatory activity of PC that becomes exposed on oxLDL or apoptotic endothelial cells in cardiovascular disease was performed.

(20) The selection of anti-PC antibodies was directed using PC conjugated to bovine serum albumin (BSA) and alternated between rounds with PC conjugated to ferritin.

(21) The phage display selection output was screened as individual phage for binding to PC-BSA by ELISA and the hits were DNA sequenced to identify the exact number of unique antibodies; all of which were recombinantly converted to IgG. In total, after performing selections on two different phage display libraries 41 fully human IgGs we identified and produced. These antibodies were identified after screening a total of 10,660 different phage clones by ELISA, from which there were 1,511 ELISA positive hits.

(22) An ELISA hit was defined as have a signal on immobilized target (i.e. PC-BSA) that was at least 3-fold greater than the background signal (streptavidin-coated plate).

(23) After sequencing the 1,511 ELISA positives and converting the antibodies from Fab fragments displayed on phage to fully human IgGs, 56 different antibody sequences that bind PC, 26 from the first phagemid library and 30 from the second phage library were recovered.

(24) IgG Reformatting, Expression and Purification

(25) Here we describe the results of recovery of 40 of the 56 antibodies after recombinant reformatting from Fab displayed on phage to full length IgG.

(26) DNA for each IgG was prepared and transfected into human kidney 293T cells to transiently generate IgG after a 10 day media harvest. The IgGs used for in vitro studies were purified using protein A Sepharose (MabSelect) and buffer exchanged into PBS.

(27) IgGs intended for in vivo testing were purified by protein A Sepharose, followed by cation ion exchange (Poros HS) with gradient elution. IgG antibodies intended for in vivo testing were buffer exchanged into Antibody Formulation Buffer (0.1 M citrate-phosphate, 50 mM NaCl, 0.01% Tween-80, 2% Trehalose, pH 6.0). Antibody concentrations were determined on purified samples by absorbance at 280 nm (1 mg/mL=1.4 O.D.).

(28) In Vitro Assays

(29) The 40 IgGs were tested in a battery of in vitro tests to identify the antibodies with the desired properties. Table 1 summarizes binding properties for a selection of fully human IgG Anti-Phosphorylcholine antibodies.

(30) The second column (column A) in Table 1 shows the ELISA signal obtained using only 15.6 ng/mL IgG added to PC-BSA immobilized on a 96 well plate surface. Antibodies with ELISA signals>1 are expected to be higher affinity antibodies.

(31) The third column (column B) in Table 1 shows the signal obtained when the antibodies were injected over aminophenyl phosphorylcholine covalently immobilized on a biosensor chip and binding was detected by surface plasmon resonance using a Biacore 3000 instrument. The higher the Biacore signal, the more binding was observed.

(32) The fourth column (column C) in Table 1 shows the results of test to determine specificity of the antibodies towards phosphorylcholine, by testing for binding to covalently immobilized aminophenol, which is the linker used to covalently couple phosphorylcholine to BSA or the biosensor chip. Several of the antibodies bind the linker molecule as well as, or better than, aminophenyl phosphorylcholine. These antibodies are not likely to be effective therapeutic anti-phosphorylcholine antibodies.

(33) The fifth column (column D) in Table 1 summarizes the results of testing the ability of the antibodies to inhibit the uptake of oxLDL by macrophages, which is an early event in cardiovascular inflammation and leads to the formation of foam cells. The macrophage uptake was monitored by flow cytometry using fluorescently modified oxLDL in the presence or absence of 80 μg/mL tested antibody. In each experiment, 100 μg/mL of affinity purified IgM anti-PC polyclonal antibodies was used as a positive control. The amount of oxLDL taken up in the presence of the tested monoclonal antibodies, as monitored by fluorescence, was divided by the fluorescence observed in the presence of the polyclonal antibody, and then multiplied by 100. Thus, a value below 100 indicate that the antibody in a concentration of 80 μg/mL was more effective in inhibiting oxLDL uptake than the polyclonal anti-PC extracted from human serum in a concentration of 100 μg/mL. A value above 100 similarly indicate that the antibody was less effective than the polyclonal anti-PC.

(34) It was observed that several of the antibodies inhibited the uptake similarly, or better than, the polyclonal anti-PC control. In addition, it was observed that several antibodies stimulated macrophage uptake of oxLDL, a property that excludes these antibodies from lead selection.

(35) The last column (column E) of Table 1 shows ELISA data obtained by adding the IgGs to wells of a 96 well plate that contain either oxLDL or native LDL. The ratio of the ELISA signal observed for binding to oxLDL divided by that observed with LDL is listed in Table 1 for each tested antibody. It is evident that certain antibodies are better binders of oxLDL as compared to LDL.

(36) TABLE-US-00036 TABLE 1 Summary of Binding Properties for Fully Human IgG Anti-PC Antibodies Sample ID A B C D E M0004-B02 1.24 366.4 38.6 233.3 6.7 M0004-C02 0.11 44.8 0.2 93 1.2 M0004-G02 1.23 1028.5 15.7 nd 8.4 M0007-H10 0.49 415.8 2.7 105 0.6 M0009-A06 0.48 912.1 2.5 80.5 2.8 M0011-F05 1.56 4473.6 155.6 547.5 10.3 M0024-B01 0.26 nd nd nd 11.1 M0026-H05 0.03 1.6 17.8 73.7 1.4 M0027-H05 0.03 −3.3 1.4 79.3 1.1 M0028-H05 0.03 1.8 5 86 0.6 M0029-H05 0.08 nd nd 370 0.9 M0030-H05 0.02 19.1 32.8 nd nd M0031-H05 0.03 −4.1 0.2 81 1 M0034-G12 0.84 462.3 14.6 78 nd M0035-E11 0.14 41.5 2.1 68 0.5 M0039-H05 2.73 −6.4 2.1 80.4 0.7 M0042-G07 nd −2.9 2.3 93.7 0.8 M0043-D09 1.24 172.7 2.1 1310 16.8 M0050-H09 0.22 279.1 7 71.5 nd M0073-G03 0.18 46.3 19.9 51.1 1.2 M0077-A11 0.26 836.3 1.3 78.4 0.7 M0086-F02 0.99 1.4 12.6 315 nd M0086-H01 0.41 51.2 4.9 85 1 M0086-H11 1 −1.1 0.9 74 nd M0097-B04 0.22 109.5 −0.5 98 1.3 M0097-B05 1.01 699.6 −3.2 80 1.1 M0099-B05 1.03 5219.3 23.3 71 1.5 M0099-D11 0.03 170.7 8.6 560 2.1 M0100-A01 1.53 7532.8 3934.7 nd 1.1 M0102-E11 0.02 1.6 −1.3 83 nd M0108-H03 nd 532.7 4.5 nd 1.1 M0126-A04 0.03 34.2 −8 nd 2.8 M0126-F10 nd 32.9 −8.3 nd nd M0126-H08 0.03 114.3 566.1 98 nd M0127-A09 0.03 18.2 −8.7 160 1.6 M0127-B07 0.05 16.3 −7 67 nd M0127-E06 nd 21.9 −4.2 nd nd M0127-E07 nd 15.4 −6.2 nd 1.8 M0127-F01 0.02 9.6 3.6 77 nd X0009-A01 0.23 198.1 2 95 1.5 Full Column Headings: A) Binding to PC conjugated to BSA by ELISA at 15.6 ng/ml Ab (OD) B) Binding to aminophenyl PC by Biacore (RU) C) Binding to aminophenol linker by Biacore (RU) D) Percent oxLDL Uptake by Macro-phages in presence of 80 μg/ml Ab (a) E) Binding to oxLDL versus LDL by ELISA (oxLDL signal/LDL signal) (b)
a) OxLDL Uptake by Macrophages

(37) The uptake of Dil-labelled (1,1′-dioctadecyl-3,3,3″,3″-tetramethylindocarbocyanine perchlorate) Cu-oxidized LDL (oxLDL, Intracel Corp, US) was investigated in macrophages that were derived from human THP-1 monocytes (ATCC, US). Differentiation was induced by incubation with 100 nM PMA (Sigma-Aldrich) in RPMI and 10% FCS for 24 h, after which medium was replaced and cells left for another 48 hours. Cells were then incubated with antibodies as indicated at 37° C. for 50-60 min. Thereafter, 20 μg/ml oxLDL was added and incubation continued for 5 hours. At the end of the incubation period, cells were washed two times with ice-cold PBS/0.2% BSA and once with PBS. The cells were harvested in PBS containing 2% PFA. For data acquisition and analysis, FACS Calibur with Cell Quest software was used. For each sample, a minimum of 10.000 cells were analyzed.

(38) b) OxLDL ELISA.

(39) hLDL (Kalen Biomedical #770200-4), oxLDL (Kalen Biomedical #770252-7) (as these data are not shown) were coated at a concentration of 10 μg/ml and a volume of 100 μl/well on an ELISA plate (Immulon 2HB) overnight at 4° C. Plates were blocked with a 1% BSA solution (300 μl/well) for 2 hours at room temperature. After washing, the plate was incubated with the indicated antibodies (100 μl/well; 25-100 nM) for 1 hour at room temperature. AP-conjugated goat anti-human secondary antibody (ThermoScientific #31316) at a 1:5000 dilution was added to the washed plate at 100 μl/well and incubated for 1 hour at room temperature. Detection reagent (ThermoScientific #37621) was added (100 μl/well) and the plate was immediately read in kinetic mode at 405 nm with the temperature at 30° C. Results are shown as OD.sub.oxLDL/OD.sub.LDL.

(40) Analysis of Anti-PC IgG Affinity to PC by SPR.

(41) The IgGs were screened for binding to PC using the Biacore surface plasmon resonance (SPR) biosensor. Aminophenyl phosphorylcholine (Biosearch Technologies) was coupled through the free amine group to one flow cell of a CM5 chip to a density of 120 RU. The aminophenol linker was coupled to another flow cell of the same CM5 chip to a density of approximately 120 RU. PC-KLH and PC-BSA were also coupled to separate flow cells of a CM5 chip.

(42) Using these surfaces with PC immobilized in different contexts, the antibodies were injected at 100 nM at 50 μL/min and binding sensorgrams were obtained. The affinity of M99-B05 was investigated by flowing different concentrations of antibody over the surface at 50 μL/min. Towards this immobilized antigen the antibodies display a fast on rate and a fast off rate, which prevented us from obtaining reliable k.sub.on and k.sub.off estimates from the kinetic sensorgrams.

(43) The observed signal for each antibody concentration near the end of the injection was plotted versus the antibody concentration and fit the data to a standard hyperbolic equilibrium binding equation (FIG. 1). As seen in FIG. 1, M99-B05 appears to bind aminophenyl PC with an apparent Kd of approximately 150 nM. Both tested preparations having equivalent Kd values but the Rmax (the maximum response) differs. The apparent Kd values observed for each antibody on this surface may or may not represent the affinity observed on more physiological substrates.

(44) ELISA Screening of Purified Anti-PC IgGs

(45) The purified IgGs were also screened for binding to PC using an ELISA with PC-BSA. This data was fitted to provide estimated EC50 values (FIG. 2).

(46) Inhibition of oxLDL Induced MCP-1 Release from Monocytes

(47) Several of the antibodies were tested for their ability to block the release of the chemokine MCP-1 from monocytes in response to stimulation with oxLDL. As shown in Table 2, M99-B05 was very effective in blocking oxLDL-induced MCP-1 release. This antibody potently inhibited MCP-1 release with an IC.sub.50 in the nM range.

(48) MCP-1 is a potent pro-inflammatory chemokine that promotes the influx of leukocytes at the site of an atherosclerotic lesion (Reape and Groot. 1999. Chemokines and atherosclerosis. Atherosclerosis 147, 213-225). Control IgG anti streptavidin A2 as negative control showed no inhibition of oxLDL induced MCP-1 release from monocytes (data not shown).

(49) TABLE-US-00037 TABLE 2 Anti-PC inhibition of oxLDL-induced MCP-1 secretion from human monocytes. IC.sub.50 of M99-B05 Donor 1 1.8 ± 0.74 nM Donor 2  1.3 ± 0.7 nM

(50) Moncytes were isolated from human blood and stimulated with 2 μg/mL copper-oxidized oxLDL in the presence or absence of 10 μM to 40 nM anti-PC IgG. MCP-1 levels in the cell media were quantified using a commercially available MCP-1 specific ELISA kit

(51) The antibody (M99-B05) was also shown to bind human atherosclerotic lesion tissue (FIG. 6).

(52) In Vivo Assays

(53) Based on a combination of favorable in vitro binding properties and functionality in in vitro assays antibodies M4-G2, M73-G03, and M99-B05 were further tested in an in vivo model of coronary inflammation.

(54) This mouse model measured inflammatory cell influx into the sub-endothelial tissue (i.e. the media) in response to vascular injury induced by placing a restrictive cuff around the exposed femoral artery (FIG. 3). It is evident from FIG. 3 that M99-B05 reduced leukocyte influx into the sub-endothelial layer, the reduction seen being statistically significant. By contrast, and despite their favorable in vitro binding properties and functionality in in vitro assays, neither of M4-G2 or M73-G03 showed any statistically significant reduction compared to the control antibody (the anti-strepavidin A2 IgG termed “HuIgG1a-A2”).

(55) The very distinctive effect of M99-B05 in this assay, compared to M4-G2 and M73-G03, could not have been predicted and was a surprise to the inventors. This demonstrates that in vivo efficacy of anti-PC antibodies may not be predictable from positive in vitro data.

(56) Consequently, M99-B05 was tested in a vascular restenosis model in mice, in which injury was again induced by positioning a cuff around the femoral artery but was allowed to progress for 14 days instead of 3 days. The amount of stenosis, observed as a thickening of the vessel neotima in the affected arteries, was then analyzed by histochemistry (FIG. 4). From FIG. 4 it is evident that M99-B05 significantly inhibited vessel wall thickening after cuff-induced vascular injury. This further demonstrates that M99-B05 is highly effective in vivo.

(57) Construction of Germline and Stability Mutants

(58) An amino acid sequence analysis of the variable domains of both the heavy and the light chains of the M99-B05 antibody, identified amino acid substitutions to construct with the intention of reducing potential immunogenicity and avoiding susceptible amino acid modification that may occur during antibody expression and purification.

(59) The following tables show the alignment of the amino acid sequence of the variable domain with its most closely related germline antibody sequence using the Kabat database. Also highlighted in the tables are the amino acid substitutions that were made in the antibody to make it closer to germline, in addition to mutants that removed potential deamidation sites, an HCDR3 disulfide bond, all of which may raise concerns for manufacturability (so called “Stability Mutants”).

(60) Mutants of M99-805

(61) The X19-A01 mutant has the same heavy and light amino acid sequences as wild type M99-B05, except that first amino acid of the light chain in M99-B05 (a glutamine) is deleted in X19-A01 to better match the germline sequence.

(62) The sequence of the X19-A03 mutant encodes the fully germlined antibody, relative to the VH3-23, JH3 heavy chain and VK4-B3, JK1 light chain germline sequences, without an inserted phenylalanine (F) in framework 1 of the heavy variable region (HV-FR1), plus amino acid substitutions (stability mutations) to potentially reduce protein amino acid modification during expression and purification. The M99-B05 antibody was found to have a deleted F amino acid at the tail of HV-FR1 relative to the germline sequence. Inserting the F at this position makes the antibody closer to the germline sequence and possibly less likely to be immunogenic. The X19-A03 mutant was constructed to contain all other germline substitutions except the F insertion in the event that this insertion affected PC binding. The stability mutants contain a G to A mutation in the HV-CDR3 that was performed to disrupt a potential deamidation site (NG) and an N to Q substitution in LV-CDR1 to remove another potential deamidation site.

(63) The sequence of the X19-A05 mutant contains all the germline substitutions, including the inserted F in HV-FR1, and the stability mutations. The X19-A05 antibody is the only mutant antibody generated in this example that contains all the germline substitutions and stability mutations.

(64) The X19-A11 mutant has the same sequence as X19-A01 but has two C to S substitutions in the HV-CDR3 to remove the disulfide that is expected to be formed in this region.

(65) The X19-001 is germlined, without the F insertion, and with stability mutants with the C to S substitutions to remove the disulfide. The comparable antibody (pre-disulfide removal) is X19-A03.

(66) TABLE-US-00038 TABLE 3 Heavy chain sequence optimization of M99-B05. M99-B05 X19-A01 X19-A03 X19-A05 X19-A07 X19-A09 X19-A11 X19-C01 **************************** ******************** M99-B05 0 X19-A01 X19-A03 X19-A05 X19-A07 X19-A09 X19-A11 X19-C01 ************************************************* M99-B05 SEQ ID NO: 3 X19-A01 SEQ ID NO: 5 X19-A03 0 SEQ ID NO: 7 X19-A05 SEQ ID NO: 1 X19-A07 SEQ ID NO: 9 X19-A09 SEQ ID NO: 11 X19-A11 SEQ ID NO: 13 X19-C01 SEQ ID NO: 15 ******.**.*.*************** *****

(67) Germlined sequence mutations are shown in bold. Residue mutations that may alleviate possible manufacturing issues are underscored. CDR regions are .

(68) TABLE-US-00039 TABLE 4 Light chain sequence optimization of M99-B05 M99-B05 X19-A01 X19-A03 X19-A05 X19-A07 0 X19-A09 X19-A11 X19-C01  ** ****************************:*************.****** M99-B05 X19-A01 X19-A03 X19-A05 X19-A07 X19-A09 X19-A11 0 X19-C01 **:***************************.*******:************* M99-B05 GQGTKVEIK SEQ ID NO: 4 X19-A01 GQGTKVEIK SEQ ID NO: 6 X19-A03 GQGTKVEIK SEQ ID NO: 8 X19-A05 GQGTKVEIK SEQ ID NO: 2 X19-A07 GQGTKVEIK SEQ ID NO: 10 X19-A09 GQGTKVEIK SEQ ID NO: 12 X19-A11 GQGTKVEIK SEQ ID NO: 14 X19-C01 GQGTKVEIK SEQ ID NO: 16 *********

(69) Germlined sequence mutations are shown in bold. Residue mutations that may alleviate possible manufacturing issues are underscored. CDR regions are .

(70) For the avoidance of doubt, in the event of any inadvertent disparity between the presentation of sequences within this application, the sequences provided for the VH and VL domains and the various CDR sequences in Tables 3 and 4 are the definitive sequences.

(71) PC Binding of the Mutants of M99-B05.

(72) PC binding of the mutants of M99-B05 was assessed by ELISA (FIG. 5). From the ELISA data (FIG. 5) it is evident that many of the mutations in M99-B05 did not significantly affect the binding to PC. However, replacing the cysteine residues in the Hv-CDR3 with serine (X19-A11 and X19-001) did reduce the affinity. These cysteine residues are expected to form a disulfide and are present in the germline antibody sequence encoded by VK4-B3. The differences in the observed binding signals may be attributed to differences in the amount of active antibody in each preparation and/or slight errors in the concentration measurements. The antibody mutant of M99-B05 that contained the maximum number of permissive sequence optimized substitutions was X19-A05. This antibody contained all the stability and germline substitutions but retains the Hv-CDR3 disulfide.

(73) Comparison of the In Vivo Effect of M99-B05 and X19-A05

(74) M99-B05 and X19-A05 were tested in the vascular restenosis model in mice, in which injury was again induced by positioning a cuff around the femoral artery but was allowed to progress for 14 days instead of 3 days. The amount of stenosis, observed as a thickening of the vessel neotima in the affected arteries, was then analyzed by histochemistry and the intimal thickening calculated (FIG. 7). From FIG. 7 it is evident that X19-A05 significantly inhibited vessel wall thickening after cuff-induced vascular injury to a similar extent as compared to M99-B05. The effect of X19-A05 also showed a clear dose-response relation.