Anti-TNF alpha antibodies which selectively inhibit TNF alpha signalling through the p55R

Abstract

The present invention provides anti-TNFα antibodies which selectively inhibit TNFα signalling through the p55R. In particular the present invention provides anti-TNFα antibodies which selectively inhibit TNFα signalling through the p55R relative to the p75R.

Claims

1. An anti-TNFα antibody or a fragment thereof that selectively binds TNFα, each comprising a heavy chain variable domain and a light chain variable domain, wherein the heavy chain variable domain comprises the sequence given in SEQ ID NO:9 for CDR-H1, the sequence given in SEQ ID NO:10 or SEQ ID NO:21 for CDR-H2 and the sequence given in SEQ ID NO:11 for CDR-H3 and wherein the light chain variable domain comprises the sequence given in SEQ ID NO:12 for CDR-L1, the sequence given in SEQ ID NO:13 for CDR-L2 and the sequence given in SEQ ID NO:14 for CDR-L3.

2. The anti-TNFα antibody or fragment thereof according to claim 1 comprising (a) a heavy chain comprising SEQ ID NO:6 or SEQ ID NO:20 and (b) a light chain comprising SEQ ID NO:8.

3. The antibody according to claim 1, wherein the antibody or fragment thereof is a CDR-grafted antibody.

4. The antibody according to claim 1.

5. The antibody according to claim 2.

6. The fragment according to claim 1 wherein the fragment is an Fab, Fab′, F(ab′).sub.2, or scFv fragment.

7. The antibody or fragment thereof according to claim 1 wherein the antibody or fragment thereof is conjugated to one or more effector molecule(s).

8. The antibody according to claim 7.

9. A pharmaceutical composition comprising an anti-TNFα antibody according to claim 1 and a pharmaceutically acceptable carrier.

10. A pharmaceutical composition comprising the fragment according to claim 1 and a pharmaceutically acceptable carrier.

Description

(1) The present invention is further described by way of illustration only in the following examples, which refer to the accompanying Figures, in which:

(2) FIG. 1: p55TNFR and p75TNFR binding inhibition assay showing the effect of different anti-TNFα antibodies on the binding of TNFα to the p55R and the p75R.

(3) FIG. 2: shows the expression cassette cloned into the large NotI and XhoI restricted fragment of pBluescript® II SK(+) to generate the bioassay receptor shuttle vector.

(4) FIG. 3: shows a titration of the TNFα induced luciferase response from the p75/CD28-TCR zeta bioassay receptor.

(5) FIG. 4: shows the effect of antibody ‘462’, infliximab and adalimumab on p55R signalling.

(6) FIG. 5: shows the effect of antibody ‘463’ on p55R signalling.

(7) FIG. 6: shows the effect of antibody ‘462’, infliximab and adalimumab on p75R signalling.

(8) FIG. 7: shows the effect of antibody ‘463’ on p75R signalling.

EXAMPLES

Example 1. Isolation of a Panel of Anti-TNFα Antibodies

(9) Rats were immunised with soluble human recombinant TNFα. 4×5 ug at 3-4 week intervals initially in complete Freund's adjuvant by the sub-cutaneous route.

(10) Spleen cells from one rat were then seeded into 40 microtitre plates at a cell density that ensures that any detected TNFα binding antibody is clonal. The cells were then cultured in T cell conditioned media (3%) and EL-4 cells (5×10.sup.4/well) for seven days. Seven days later supernatants from these plates were screened by ELISA for anti-TNFα antibodies using human TNFα (50 ng/ml) captured by a sheep polyclonal coated onto immunoplates. The supernatants from positive wells were then further tested in the L929 bioassay and the p55 and p75 receptor specific protein assays described below.

Example 2. TNFα Receptor Binding Inhibition Assays

(11) L929 Assay

(12) L929 cells (a mouse fibroblast cell line) that express the mouse p55TNFα receptor but not the p75TNFα receptor were used to assay for anti-TNFα antibodies that block binding of TNFα to this receptor. These cells are killed by human TNFα if sensitised with a protein synthesis inhibitor.

(13) Cells were grown in standard tissue culture medium and seeded into 96 well tissue culture plates the day before being required in the assay. The culture medium was removed and test supernatants were added to individual wells. Human recombinant TNFα was then added to each well at 200-400 pg/ml in the presence of 1 μg/ml (final concentration) actinomycin D. The plates were then incubated overnight at 37° C.

(14) On the following day the plates were washed gently in PBS and the cells fixed in methanol. They were then stained with 1% Crystal violet (live cells remain attached to the plates and take up the dye). Excess stain was washed off and the remaining stained cells solubilised in 30% acetic acid. The plates were then read at 570/405 nM.

(15) Wells containing antibodies that block binding of TNFα to the mouse p55TNFR protect the cells from TNFα mediated cytotoxicity and show an enhanced signal compared with negative/control wells.

(16) Positive wells were further tested in the p55R and p75R assays.

(17) p55TNFR and p75TNFR Binding Inhibition Assay

(18) Standard ELISA plates were coated with a sheep anti-human TNFα polyclonal antibody diluted 1/10,000. The plates were then blocked with PBS+1% BSA. Human TNFα was then added to each well at 25-50 ng/ml. After 1 hour unbound TNFα was washed off. Supernatants containing anti-TNFα antibodies were then added to replicate wells. In addition to one well of each replicate was added either human p55TNFR-Human Fc fusion protein or Human p75TNFR-Human Fc fusion protein. These were incubated for 1 hour and then washed to remove unbound receptor. Following this step an anti-Human IgG Fc peroxidase conjugated polyclonal antibody (Stratech Scientific) was added at 1/2000 dilution. The plates were left for 1 hr and then washed to remove unbound conjugate. TMB substrate was then added to each well, and the colour allowed to develop. Wells where the anti-TNFα antibodies have blocked binding of the receptor(s) can therefore be visualised.

(19) FIG. 1 shows the percentage inhibition of TNFα binding to the p55TNFR and p75TNFR by four different anti-TNF antibodies. Antibody ‘3D6’ inhibited binding of TNFα to the p55TNFR by 49.3% but only inhibited binding of TNFα to the p75TNFR by 14.6%. In contrast, antibody 22H3 for example inhibited binding of TNFα to the p55R and the p75R by 78.9 and 71.9% respectively. Antibody 3D6 therefore selectively blocks binding of TNFα to the p55R.

Example 3. Isolation of Further Selective Antibodies

(20) Using the same rat population as Example 1 cultured B cells were screened to identify TNFα selective antibodies.

(21) Human TNFα (Strathman Biotech GmbH) was biotinylated with a 10 fold molar excess of Sulfo-NHS-LC-LC-biotin (Pierce) for 1 hour at room temperature following the manufacturers' protocol. 5 μg of biotinylated TNFα was mixed with 50 μl of 9.95 micron superavidin coated microspheres (Bangs Beads) for 1 hour at room temperature in a volume of 500 μl (mix for 1×384-well plate). Beads were then washed 5 times in PEG block (1% PEG/0.1% tween/PBS) to remove unbound TNFα. TNFα-coated beads were then resuspended in approx. 4 ml PEG block and 10 μl added to each well of a 384-well plate. 10 μl of B cell culture containing rat antibody and 10 μl of goat-anti-rat IgG Fc gamma specific-Cy5 conjugate at 1:1666 dilution were added to the well containing beads. Plates were incubated at room temperature in the dark for 1 hour and then read on an Applied Biosystems 8200 machine. Applied Biosystems software was used to identify positive wells.

(22) Approximately 1400 B cell culture plates were screened, which is approximately 140000 wells which represents approximately 2×10.sup.8 B cells.

(23) Of those screened 2500 wells contained antibodies which bound TNFα.

(24) These were further screened for the ability to selectively block TNFα signalling through the p55R relative to the p75R using the assays described in Example 4. Antibodies were first screened for blocking p55R signalling and those that blocked signalling were then tested for the ability to block TNFα signalling through the p75R. Those antibodies which selectively blocked signalling through the p55R were isolated using the homogeneous fluorescence assay described in WO2004/051268 and heavy and light chain variable region genes were cloned via reverse transcription PCR from single rat B cells. Variable regions were expressed in recombinant IgG format to confirm binding and activity in signalling assays by sub-cloning into expression vectors containing the human antibody constant region genes (human kappa light chain and gamma-4 heavy chain in which the serine at position 241 has been changed to proline as described in Angal et al., Molecular Immunology, 1993, 30(1), 105-108) and a rat/human chimeric antibody expressed transiently in CHO cells. Transfections of CHO cells were performed using the lipofectamine procedure according to manufacturer's instructions (InVitrogen, catalogue No. 18324).

(25) Two antibody sequences were obtained and these were termed, ‘462’ and ‘463’. The V-region sequences of ‘462’ are given in SEQ ID NOS:1, 2, 3 and 4. The variable region sequences without the leader sequences are provided in SEQ ID NOS: 5, 6, 7 and 8. The V-region sequences of ‘463’ are given in SEQ ID NOS: 15, 16, 17 and 18. The variable region sequences without the leader sequences are provided in SEQ ID NOS: 19, 20, 7 and 8.

Example 4. TNFR Signalling Assays

(26) 4.1 p55R Signalling Assay

(27) p55 NFkB Luciferase Assay

(28) A549-ES-Luc cells were used for this reporter gene assay. A549 cells are an epithelial lung cell carcinoma that express the p55 TNF receptor and have been stably transfected with a vector comprising the E-selectin promoter (contains 3×NFkB binding sites) linked to the luciferase gene and a selectable marker for stable cell line generation.

(29) A549-ES-Luc were grown in the following media:

(30) RPMI 1640 (Phenol Free)

(31) +10% FCS

(32) +2 Mm Glutamine

(33) +1 mg/ml G418 (Life Tech, 50 mg/ml stock)

(34) A549-ES-luc cells were plated out into white opaque 96-well plates (Perkin Elmer) using a cell suspension of 1.5×10.sup.5 cells/ml; 100 μl/well=15,000 cells/well. The cells were allowed to adhere overnight at 37° C./5% CO.sub.2. The following day the media was aspirated and replaced with 100 μl of antibody in assay medium that has been pre-incubated for 30 minutes with human TNFα at 3 ng/ml final concentration. The cells were incubated for 5 hours at 37° C./5% CO.sub.2. Luciferase expression was then assayed using a luciferase reporter gene assay kit (LucLite from Perkin Elmer). The plate was then read in a luminescence plate reader, the LJL Analyst.

(35) 4.2 p75 Signalling Assay

(36) Jurkat cells that have been stably transfected with a vector containing a cassette coding for the p75R extra-cellular domain linked to the intra-cellular signalling regions of CD28 and TCR zeta was used to assay for p75 signalling. Within the same vector there are 5 binding sites for NFκB with a minimal E-selectin promoter region, this drives expression of the reporter gene luciferase, and a selectable marker for stable cell line generation. Stimulation of the p75 bioassay receptor with its ligand, human TNFα, leads via the CD28/zeta regions of the bioassay receptor, to the initiation of a signalling cascade within the cell. The signalling cascade induces NFκB activation and allows transcription of the luciferase reporter gene. Activation levels can then be measured in a luciferase assay. Antibodies that can block this activation will prevent expression of luciferase.

(37) Construction of Receptor Expression Cloning Cassette and Shuttle Vector.

(38) Intermediate shuttle vector containing the entire expression cassette necessary for the expression of the Bioassay receptor was used. This vector includes the cloning cassette devised in pBluescript SK+ (Stratagene) described previously (Finney et al., J. Immunol. 2004 172: 104). 5′ to this cloning cassette is the HCMV promoter, and the SV40 polyadenylation signal is 3′ to this cloning cassette. The cloning cassette consists of an extracellular domain (ECD) binding component, a transmembrane component and a signalling region component, and facilitates easy exchange of each individual component. Combining the following DNA fragments generated the shuttle vector: A) The vector backbone of pBluescript II SK(−) (Stratagene) on a NotI to XhoI fragment B) The cloning cassette described previously on a HindIII to EcoRI fragment C) The HCMV promoter on a NotI to HindIII fragment D) The SV40 polyadenylation signal on an EcoRI to XhoI fragment. The generation of this shuttle vector is shown in FIG. 2.

(39) Construction of Binding, Transmembrane and Signalling Component Fragments

(40) Human p75 TNFα receptor extracellular domain binding component HindIII to NarI fragment.

(41) A fragment comprising the leader sequence and extracellular domain residues 1 to 257 (GenBank ref:NM 001066) of the human p75 TNF-α receptor was PCR cloned using oligos 4023 (SEQ ID NO:22) and 4024 (SEQ ID NO:23) from plasmid pORF9-hTNFRSF1B (Invivogen). Oligo 4023 introduces a 5′ HindIII site and Kosak sequence. Oligo 4024 introduces a 3′ NarI site. The PCR product was then digested with restriction enzymes HindIII and NarI.

(42) Human CD28 transmembrane and signalling region and human TCR zeta signalling region component NarI to EcoRI fragment.

(43) A fragment comprising residues 135 to 202 of human CD28 transmembrane and signalling region and residues 31 to 142 of human TCR zeta intracellular region was digested from a plasmid previously described (Finney et al., J. Immunol. 2004 172: 104) with restriction enzymes NarI and EcoRI.

(44) Construction of Bioassay Receptor Reporter Gene Vectors

(45) The full length expression cassette for the Bioassay receptor was generated by combining the binding, transmembrane and signalling components described above in the shuttle vector described above. This was then subcloned into the reporter gene vector pNifty2-Luc(Invivogen). This vector contains a Luciferase reporter gene under control of a NF-kB inducible promoter and the selectable marker Zeocin™ for selection in both E. coli and mammalian cells. The Bioassay receptor expression cassette was removed from the shuttle vector on a NotI to NotI fragment and cloned into the NotI site of pNifty2-Luc.

(46) Generation of Stable Bioassay Receptor Reporter Gene Cell Lines

(47) Plasmid DNA of the vector was transfected into the human T cell leukaemia cell line, Jurkat E6.1 using the Amaxa Nucleofector device according to the manufacturers instructions (Amaxa Biosystems). Stable cell lines were then generated by culture in Zeocin™ at a concentration of 200 μg/ml.

(48) Analysis of Anti-Human TNFα Antibody using a p75/CD28-TCR Zeta Bioassay Receptor.

(49) A stable cell line expressing a bioassay receptor that comprises the human p75 TNFα receptor extracellular domain binding component, human CD28 transmembrane and signalling region, and human TCR zeta signalling region components was generated as described above. To these cells, a titration of human TNFα was added and the amount of Luciferase produced determined 4 hours later with a Luclite assay kit (Promega) according to the supplier's instructions. The TNFα induced Luciferase response from the p75/CD28-TCR zeta Bioassay receptor is shown in FIG. 3. A concentration of TNFα was selected from this titration and used to assess the ability of an anti-TNFα antibody to block Luciferase production via the p75/CD28-TCR zeta Bioassay receptor.

(50) Assay Media:

(51) 500 ml DMEM (phenol free)

(52) +10% Foetal Calf Serum

(53) +2 mM Glutamine

(54) +1 ml Normacin

(55) +200 μg/ml zeocin

(56) +1% enhancer solution, protease inhibitor

(57) Jurkat cells were plated out into white opaque 96-well plates using a cell suspension of 2×10.sup.6 cells/ml. Antibodies were then added to the plate in the desired titration scale. The plate was incubated for 30 minutes at 37° C. and 10 μl of human TNFα ligand added to each well at a concentration of 30 ng/ml to give a final concentration of 3 ng/ml human TNFα in each well. The plate was incubated for 4 hours at 37° C. Luciferase expression was then assayed using a luciferase reporter gene assay kit (Luclite 1000 kit, Perkin-Elmer).

(58) Results

(59) The effect of antibody ‘462’ and the commercially available anti-TNFα antibodies Adalimumab and Infliximab on Luciferase production in the p55R signalling assay is shown in FIG. 4. It is clear that all three antibodies inhibit TNFα signalling through the p55R. FIG. 5 shows that antibody ‘463’ also inhibits TNFα signalling through the p55R.

(60) The effect of antibody ‘462’ and the commercially available anti-TNFα antibodies Adalimumab and Infliximab on Luciferase production in the p75R signalling assay is shown in FIG. 6. It is clear that only Adalimumab and Infliximab inhibit TNFα signalling through the p75R while antibody ‘462’ leaves TNFα signalling through the p75R largely unaffected. FIG. 7 shows that antibody ‘463’ also leaves TNFα signalling through the p75R largely unaffected. Both antibodies ‘462’ and ‘463’ were significantly less potent in the p75R signalling assay than in the p55R signalling assay.

(61) Antibodies ‘462’ and ‘463’ therefore selectively inhibit TNFα signalling through the p55R.