Biological products
10407513 · 2019-09-10
Assignee
Inventors
Cpc classification
A61P1/04
HUMAN NECESSITIES
A61P29/00
HUMAN NECESSITIES
C07K2317/569
CHEMISTRY; METALLURGY
C07K2317/92
CHEMISTRY; METALLURGY
A61P1/16
HUMAN NECESSITIES
C07K2317/62
CHEMISTRY; METALLURGY
A61P35/00
HUMAN NECESSITIES
C07K2317/24
CHEMISTRY; METALLURGY
C07K16/28
CHEMISTRY; METALLURGY
C07K16/2878
CHEMISTRY; METALLURGY
International classification
A61K39/00
HUMAN NECESSITIES
C07K16/24
CHEMISTRY; METALLURGY
Abstract
A multivalent antibody fusion protein which comprises an immunoglobulin moiety, for example a Fab or Fab fragment, with a first specificity for an antigen of interest, and further comprises two single domain antibodies (dAb) with specificity for a second antigen of interest, wherein the two single domain antibodies are linked by a disulfide bond. There is also provided particular dual specificity antibody fusion proteins comprising a Fab or Fab fragment and one or more single domain antibodies which may be stabilized by a disulfide bond therebetween.
Claims
1. A divalent antibody fusion protein, comprising: (i) an immunoglobulin moiety with a specificity for an antigen of interest, wherein the immunoglobulin moiety is a Fab or Fab fragment having a heavy chain and a light chain; and (ii) two single domain antibodies (dAb), a VH dAb and a VL dAb, that together bind to the same human serum albumin, wherein the VH dAB comprises complementary determining region (CDR)H1 having the amino acid sequence set forth in SEQ ID NO:56, CDR-H2 having the amino acid sequence set forth in SEQ ID NO:57, and CDR-H3 having the amino acid sequence set forth in SEQ ID NO:58; wherein the VL dAB comprises CDR-L1 having the amino acid sequence set forth in SEQ ID NO:59, CDR-L2 having the amino acid sequence set forth in SEQ ID NO:60, and CDR-L3 having the amino acid sequence set forth in SEQ ID NO:61; wherein (i) the VH dAb is connected directly or via a linker to the C-terminus of the Fab or Fab heavy chain and the VL dAb is connected directly or via a linker to the C-terminus of the Fab or Fab light chain, or (ii) the VH dAb is connected directly or via a linker to the C-terminus of the Fab or Fab light chain and the VL dAb is connected directly or via a linker to the Fab or Fab heavy chain; and wherein the VH dAb and the VL dAb are linked by a disulfide bond between two engineered cysteine residues at positions VH44 and VL100.
2. The divalent antibody fusion protein of claim 1, wherein the VH dAb and the VL dAb are humanised.
3. The divalent antibody fusion protein of claim 1, wherein the Fab or Fab is fully human or humanised.
4. The divalent antibody fusion protein of claim 1, wherein the VH dAb is connected to the C-terminus of the Fab or Fab heavy chain directly or via a linker and the VL dAb is connected to the C-terminus of the Fab or Fab light chain directly or via a linker.
5. The divalent antibody fusion protein of claim 1, wherein the VH dAb is connected to the C-terminus of the Fab or Fab heavy chain and the VL dAb is connected to the C-terminus of the Fab or Fab light chain via a linker having the amino acid sequence set forth in any one of SEQ ID NOs:13 or 45.
6. The divalent antibody fusion protein of claim 1, wherein the VH dAb is fused to the C-terminus of the heavy chain constant region (CH1) of the Fab or Fab and the VL dAb is fused to the C-terminus of the light chain constant region of the Fab or Fab or wherein the VL dAb antibody is fused to the C-terminus of the heavy chain constant region (CH1) of the Fab or Fab and the VH dAb antibody is fused to the C-terminus of the light chain constant region of the Fab or Fab.
7. A pharmaceutical composition comprising the divalent antibody fusion protein of claim 1.
8. The divalent antibody fusion protein of claim 1, wherein the divalent antibody fusion protein contains one CL and one CH1 domain, wherein the CL and the CH1 domains are in the Fab or Fab fragment.
Description
LIST OF FIGURES
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
(15)
(16)
(17)
(18)
(19)
(20)
(21)
(22)
(23)
(24)
(25)
KEY
(26) 645Fv equates to didAbL1 and H1 (the linker used for each dAB will be the same unless indicated otherwise). 648Fv equates to didAbL2 and H2 (the linker used for each dAB will be the same unless indicated otherwise). 645dsFv equates to didAbL1 and H1 (the linker used for each dAB will be the same unless indicated otherwise) wherein L1 and H1 are stabilised by a disulfide bond. 648dsFv equates to didAbL2 and H2 (the linker used for each dAB will be the same unless indicated otherwise) wherein L2 and H3 are stabilised by a disulfide bond. Fab are Fabs which lack the interchain cysteine bond (ie between CH and CL or CK)
EXPERIMENTAL
(27) Abbreviations: unless the context indicates otherwise m as a pre-fix is intended to refer to murine.
(28) Unless the context indicates otherwise h as a pre-fix is intended to refer to human. Fab A, Fab B, Fab C and Fab D components may be provided below in different formats.
Example 1: Production of a Dab Specific for Human Serum Albumin
(29) An in-frame DNA encoded transcription unit encoding a dAb with specificity for human serum albumin was produced using recombinant DNA technology.
(30) Where desired an in-frame DNA encoded transcription unit encoding a dAb with specificity for a recruitment protein can be produced using recombinant DNA technology.
Example 2: Production of Antibody Fragment
(31) For fusion of a dAb to the C-terminus of the light chain, DNA was synthesised encoding a human kappa light chain constant region (with the Km3 allotype of the kappa constant region), a peptide linker and a dAb and cloned as a SacI-PvuII restriction fragment into the UCB-Celltech in-house expression vector pTTOD(Fab) (a derivative of pTTO-1, described in Popplewell et al., Methods Mol. Biol. 2005; 308: 17-30) which contains DNA encoding the human gamma-1 CH1 constant region. This gave rise to a dicistronic gene arrangement consisting of the gene for the humanised light chain fused via a linker to a dAb followed by the gene for the humanised heavy chain Fab fragment, both under the control of the tac promoter. Also encoded is a unique BspE1 site upstream of the Gly4Ser linker, or an AscI site upstream of the Ala-Pro-rich linker.
(32) For fusion of a dAb the C-terminus of the heavy chain, DNA was synthesised encoding a human CH1 fragment (of the 1 isotype) followed by a linker encoding sequence and a dAb. This was subcloned as an ApaI-EcoRI restriction fragment into the UCB-Celltech in-house expression vector pTTOD(Fab) (a derivative of pTTO-1, described in Popplewell et al., above) which contains DNA encoding the human gamma-1 CH1 constant region. This gave rise to a dicistronic gene arrangement consisting of the gene for the humanised light chain a non-coding intergenic sequence and followed by a heavy chain fused via a linker to a dAb, both under the control of the tac promoter. The recombinant expression plasmid was transformed into the E. coli strain W3110 in which expression is induced by addition of IPTG. Expression experiments were performed at small scale initially (5 ml culture volumes) with addition of 200 uM IPTG at OD (600 nm) of approx. 0.5, cells were harvested 2 hours post induction and extracted overnight at 30 C. in Tris/EDTA. Clarified extracts were used for affinity analysis by Biacore. Constructs giving promising expression yields and activities were selected for fermentation.
Methods Applicable to the Following Examples
(33) In the following examples the antibody chain to which the dAb is fused is denoted either as CK or LC for the cKappa light chain and as CHI or HC for the heavy chain constant domain, CH1.
(34) Construction of FabA-dAb Fusion Plasmids for Expression in E. coli
(35) Fab-dAb fusion proteins were constructed by fusing dAbL3 or dAbH4 to the C-terminus of the constant region of either the light or heavy chain of FabA. A flexible (SGGGGSE (SEQ ID NO:1)) or a rigid (G(APAPA).sub.2 (SEQ ID NO: 34)) linker was used to link the dAb to the cKappa region (SEQ ID NO:75) whereas the linker DKTHTS (SEQ ID NO:2) was used to link the dAb to the CHI region (SEQ ID NO:76). The DNA sequence coding for the constant region-dAb fusion was manufactured synthetically as fragments to enable sub-cloning into the FabA sequence of the in-house pTTOD vector.
(36) Light chain-dAb fusions were constructed by sub-cloning the SacI-ApaI fragment of the synthesized genes, encoding a C-terminal cKappa fused to either dAbL3 or dAbH4 via either a (SGGGGSE (SEQ ID NO:1)) or a rigid (G(APAPA).sub.2 (SEQ ID NO: 34)) linker, into the corresponding sites of a plasmid capable of expressing FabA. Heavy chain-dAb fusions were constructed by sub-cloning the ApaI-EcoRI fragment of the synthesised genes, encoding a C-terminal CHI fused to either dAbL3 or dAbH4 via a DKTHTS linker, into the corresponding sites of a plasmid capable of expressing FabA.
(37) Fab A is derived from an IL-1 beta binding antibody, the heavy and light chain sequences of which are provided in SEQ ID NOs:74 and 75 respectively as shown in
(38) FabA comprises the same light chain sequence (SEQ ID NO:75) and a truncated heavy chain sequence which terminates at the interchain cysteine (SEQ ID NO:77). dAbL3 and dAbH4 are light and heavy chain domain antibodies respectively which bind human serum albumin.
(39) Construction of FabA-didAb Fusion Plasmids for Expression in E. coli
(40) FabA-didAb with dAbL3 or dAbH4 on both light and heavy chains were constructed by sub-cloning the ApaI-EcoRI fragment coding for CH1-dAb fusions into the existing Fab-dAb plasmids where the dAb is fused to the light chain via the flexible linker.
(41) Construction of FabB-dAb Fusion Plasmids for Expression in Mammalian Cells
(42) The FabB-dAbs, FabB-dAbHI (CH1-G.sub.4Sx2), FabB-dAbH2 (CH1-G.sub.4Sx2), FabB-dAbL1 (CH1-G.sub.4Sx2), FabB-dAbL2 (CH1-G.sub.4Sx2) were all assembled by PCR then cloned into a mammalian expression vector under the control of the HCMV-MIE promoter and SV40E polyA sequence. These were paired with a similar vector containing the FabB light chain for expression in mammalian cells (see below).
(43) FabB is derived from an antibody which bids a cell surface co-stimulatory molecule. dAbH1, dAbH2, dAbL1 and dAbL2 were obtained as described in Example 3.
(44) Mammalian Expression of FabB-DAbs and didAbs
(45) HEK293 cells were transfected with the heavy and light chain plasmids using Invitrogen's 293fectin transfection reagent according to the manufacturer's instructions. Briefly, 2 g heavy chain plasmid+2 g light chain plasmid was incubated with 10 l 293fectin+340 l Optimem media for 20 mins at RT. The mixture was then added to 510.sup.6 HEK293 cells in suspension and incubated for 4 days with shaking at 37 C.
(46) Biacore
(47) Binding affinities and kinetic parameters for the interactions of Fab-dAb constructs were determined by surface plasmon resonance (SPR) conducted on a Biacore T100 using CM5 sensor chips and HBS-EP (10 mM HEPES (pH7.4), 150 mM NaCl, 3 mM EDTA, 0.05% v/v surfactant P20) running buffer. Fab-dAb samples were captured to the sensor chip surface using either a human F(ab).sub.2-specific goat Fab (Jackson ImmunoResearch, 109-006-097) or an in-house generated anti human CH1 monoclonal antibody. Covalent immobilisation of the capture antibody was achieved by standard amine coupling chemistry.
(48) Each assay cycle consisted of firstly capturing the Fab-dAb using a 1 min injection, before an association phase consisting of a 3 min injection of antigen, after which dissociation was monitored for 5 min. After each cycle, the capture surface was regenerated with 21 min injections of 40 mM HCl followed by 30s of 5 mM NaOH. The flow rates used were 10 l/min for capture, 30 l/min for association and dissociation phases, and 10 l/min for regeneration.
(49) For kinetic assays, a titration of antigen (for human serum albumin typically 62.5 nM-2 M, for IL-11.25-40 nM) was performed, a blank flow-cell was used for reference subtraction and buffer-blank injections were included to subtract instrument noise and drift.
(50) Kinetic parameters were determined by simultaneous global-fitting of the resulting sensorgrams to a standard 1:1 binding model using Biacore T100 Evaluation software.
(51) Fab-dAb Purification from E. coli
(52) Periplasmic Extraction
(53) E. coli pellets containing the Fab-dAbs within the periplasm were re-suspended in original culture volume with 100 mM Tris/HCl, 10 mM EDTA pH 7.4. These suspensions were then incubated at 4 C. for 16 hours at 250 rpm. The re-suspended pellets were centrifuged at 10000g for 1 hour at 4 C. The supernatants were removed and 0.45 m filtered.
(54) Protein-G Capture
(55) The Fab-dAbs were captured from the filtered supernatant by Protein-G chromatography. Briefly the supernatants were applied, with a 20 minute residence time, to a GammaBind Plus Sepharose (GE Healthcare) column equilibrated in 20 mM phosphate, 150 mM NaCl pH7.1. The column was washed with 20 mM phosphate, 150 mM NaCl pH7.1 and the bound material eluted with 0.1M glycine/HCl pH2.8. The elution peak was collected and pH adjusted to pH5 with 1M sodium acetate. The pH adjusted elutions were concentrated and diafiltered into 50 mM sodium acetate pH4.5 using a 10 k MWCO membrane.
(56) Ion Exchange
(57) The Fab-dAbs were further purified by cation exchange chromatography at pH4.5 with a NaCl elution gradient. Briefly the diafiltered Protein-G eluates were applied to a Source15S (GE Healthcare) column equilibrated in 50 mM sodium acetate pH4.5. The column was washed with 50 mM sodium acetate pH4.5 and the bound material eluted with a 20 column volume linear gradient from 0 to 1M NaCl in 50 mM sodium acetate pH4.5. Third column volume fractions were collected through out the gradient. The fractions were analysed by A280 and SDS-PAGE and relevant fractions pooled.
(58) Gel Filtration
(59) If required the Fab-dAbs were further purified by gel filtration. Briefly the FabA-dAbL3 (CK-SG.sub.4SE) pooled ion exchange elution fractions were applied to a Superdex200 (GE Healthcare) column equilibrated in 50 mM sodium acetate, 125 mM NaCl pH 5.0 and eluted with an isocratic gradient of 50 mM sodium acetate, 125 mM NaCl pH 5.0. 1/120 column volume fractions were collected through out the gradient. The fractions were analysed by A280 and SDS-PAGE and relevant fractions pooled. For Fab-dAbs that did not undergo gel filtration, the pooled ion exchange elution fractions were concentrated and diafiltered into 50 mM sodium acetate, 125 mM NaCl pH 5.0 using a 10 k MWCO membrane.
(60) SDS-PAGE
(61) Samples were diluted with water where required and then to 10 l was added 10 L 2 sample running buffer. For non-reduced samples, 24, of 100 mM NEM was added at this point, for reduced samples 24 of 10 reducing agent was added. The sample were vortexed, incubated at 85 C. for 5 mins, cooled and centrifuged at 12500 rpm for 30 secs. The prepared samples were loaded on to a 4-20% acrylamine Tris/Glycine SDS gel and run for 100 mins at 125V. The gels were either transferred onto PVDF membranes for Western blotting or stained with Coomassie Blue protein stain.
(62) Western Blotting
(63) Gels were transferred to PVDF membranes in 12 mM Tris, 96 mM glycine 018.3 for 16 hours at 150 mA. The PVDF membrane was block for 1 hr with 2% Marvel in PBS+0.1% Tween20 (Blocking buffer)
(64) Anti-Light Chain
(65) HRP-rabbit anti-human kappa light chains, 1/5000 dilution in blocking buffer for 1 hr.
(66) Anti-Heavy Chain
(67) mouse anti-human heavy chain, 1/7000 dilution in blocking buffer for 1 hr. Followed by HRP-goat anti-mouse, 1/2000 dilution in blocking buffer for 1 hr.
(68) Anti-His Tag
(69) rabbit anti-His6, 1/1000 dilution in blocking buffer for 1 hr. Followed by HRP-goat anti-rabbit IgG, 1/1000 dilution in blocking buffer for 1 hr.
(70) All blots were washed 6 times with 100 ml PBS+0.1% Tween20 for 10 minutes per wash. The blots were developed with either ECL reagent for 1 min before being exposed to Amersham Hyperfilm, or metal enhanced DAB reagent for 20-30 minutes followed by water.
(71) High Temperature Reverse Phase HPLC
(72) Samples (2 g) were analysed on a 2.1 mm C8 Poroshell column at 80 C., with a flow rate of 2 ml/min and a gradient of 18-38% B over 4 mins. A=0.1% TFA in H.sub.2O B=0.065% TFA in 80:20 IPA:MeOH. Detection is by absorption at 214 nm.
(73) ELISA
(74) The yields of Fab-dAb were measured using a sandwich ELISA. Briefly, the Fab-dAb was captured with an anti-CHI antibody then revealed with an anti-kappa-HRP.
(75) FACS
(76) Samples (mFabD-didAb's) were incubated with 5 g/ml FITC (fluorescein isothiocyanate) labelled HSA for 45 min. The sample/HSA-FITC incubations were then added to activated mouse CD4+ T-cells and incubated for a further 45 min. The cells were washed with PBS and the cell associated fluorescence measured by FACS (fluorescence activated cell sorting).
Example 3
(77) Generating Anti-Albumin Antibodies
(78) lop rabbits were immunised with recombinant chromapure human serum albumin (purchased from Jackson). Rabbits received 3 immunisations of 100 ug HSA protein subcutaneously, the first immunisation in complete Freunds adjuvant and subsequent immunisations in incomplete Freunds. Antibodies 1 and 2, 646, 647, and 649 which bind human, mouse and rat serum albumin were isolated using the methods described in WO04/051268. Genes for the heavy chain variable domain (VH) and light chain variable domain (VL) of antibodies 1 and 2 were isolated and sequenced following cloning via reverse transcription PCR.
(79) The light chain grafted sequences were sub-cloned into the rabbit light chain expression vector pVRbcK which contains the DNA encoding the rabbit C-Kappa constant region. The heavy chain grafted sequences were sub-cloned into the rabbit heavy chain expression vector pVRbHFab, which contains the DNA encoding the rabbit Fab heavy chain constant region. Plasmids were co-transfected into CHO cells and the antibodies produced screened for albumin binding and affinity (Table 1). Transfections of CHO cells were performed using the Lipofectamine 2000 procedure according to manufacturer's instructions (InVitrogen, catalogue No. 11668).
(80) Generating Humanised Domain Antibodies dAbL1, dAbH1, dAbL2 and dAbH2
(81) Humanised VL and VH regions were designed using human V-region acceptor frameworks and donor residues in the framework regions. One grafted VL region (L1 (SEQ ID NO:53) and L2 (SEQ ID NO:55)) and one VH region (H1 (SEQ ID NO:52) and H2 (SEQ ID NO:54)) were designed for each of antibodies 1 and 2 respectively and genes were built by oligonucleotide assembly and PCR mutagenesis. The grafted domain antibodies and their CDRs are shown in
(82) TABLE-US-00008 TABLE 1 Affinities of anti-albumin antibodies as rabbit Fab as humanised IgG HSA MurineSA HumanSA MurineSA RatSA nM nM nM nM nM Antibody 1 0.31 2.6 0.82 2.9 7.9 (Antibody 645) Antibody 2 0.33 12 0.13 23 54 (Antibody 648) Antibody 646 0.14 1.6 0.57 1.7 4.5 Antibody 647 0.60 3.6 1.3 26 10 Antibody 649 0.54 13 0.32 17 44
Example 4: Analysis of FabB-dAbs Expressed in Mammalian Cells
(83) FabB-dAb constructs were produced as described in the methods and the supernatants from the transfected HEK293 cells containing the FabB-dAbs were tested directly in BIAcore.
(84) Kinetic analysis was conducted to assess the interaction of HSA with FabB-dAb constructs. These consisted of either dAbL1, dAbH2 or dAbL3 fused to the C-terminus of CHI of FabB (See
(85) TABLE-US-00009 TABLE 2 Construct k.sub.a (10.sup.4 M.sup.1s.sup.1) k.sub.d (10.sup.3 s.sup.1) K.sub.D (10.sup.9M) FabB-dAbL1 (CH1- 1.91 0.74 2.18 1.21 170 78 G.sub.4Sx2) FabB-dAbH2 (CH1- 2.66 0.39 29 4.76 1074 42 G.sub.4Sx2) FabB-dAbL3 (CH1- 2.63 0.39 9.87 1.63 392 119 G.sub.4Sx2)
(86) Affinity and kinetic parameters determined for the binding of HSA to FabBs fused to dAbL1, dAbH2 or dAbL3. The data shown are mean values SEM. (For FabB-dAbL1 and FabB-dAbH2 n=4. For FabB-dAbL3 n=2).
(87) SDS-PAGE and western blotting of the FabB-dAb proteins confirmed that the FabB-dAbs produced were of the expected size.
Example 5: Analysis of FabB-didAbs Expressed in Mammalian Cells
(88) FabB-didAb constructs were produced as described in the methods and the supernatants from the transfected HEK293 cells containing the didAbs tested directly in BIAcore.
(89) Further analysis was performed using didAb constructs in which single dAbs were fused to both heavy and light C-termini of Fab. Constructs in which the didAb was derived from a natural heavy and light variable domain pairing showed a marked improvement in affinity compared to the single dAb alone (table 2 and 3). The didAb fusion consisting of two identical dAbL1s showed no improvement in affinity over that seen for the single dAbL1 (data not shown).
(90) TABLE-US-00010 TABLE 3 k.sub.a Construct (10.sup.4 M.sup.1s.sup.1) k.sub.d (10.sup.3 s.sup.1) K.sub.D (10.sup.9M) FabB-didAb, 1.78 0.16 9 -dAbL1 (CK-G.sub.4Sx2) & dAbH1 (CH1-G.sub.4Sx2) FabB-didAb, 0.54 0.21 39 -dAbL2 (CK-G.sub.4Sx2) & dAbH2 (CH1-G.sub.4Sx2)
(91) Affinity and kinetic parameters determined for the binding of HSA to FabBs fused to both dAbL1 & dAbH1 or dAbL2 & dAbH2.
(92) SDS-PAGE of the FabB-didAb proteins confirmed that the FabB-didAbs expressed well and were of the expected size (See
Example 6
(93) Analysis of Purified FabA-dAbs
(94) Plasmids for expression of the Fab-dAbs, FabA-dAbL3 (CK-SG.sub.4SE) FabA-dAbL3 (CK-G[APAPA].sub.2) in E. coli were constructed as described in the methods. The Fab-dAbs were expressed into the periplasm of the E. coli and purified to homogeneity as described in the methods. The purity of the Fab-dAbs were assessed by high temperature reverse phase HPLC, SDS-PAGE and Western blotting. The Fab-dAbs were also assessed for antigen binding by Biacore.
(95) High Temperature Reverse Phase HPLC
(96) High temperature reverse phase HPLC as performed as described in the methods gave quantitative analysis of all species contained in FabA-dAbL3 (CK-SG.sub.4SE) and FabA-dAbL3 (CK-G[APAPA].sub.2). The percentage of each species present is shown in table 4.
(97) TABLE-US-00011 TABLE 4 Quantification of species present in Fab-dAb batches FabA-dAbL3 FabA-dAbL3 Species (CK-SG.sub.4SE) (CK-G[APAPA].sub.2) Peak 1 0.6% 1.8% Peak 2 0.6% 0.0% Peak 3 1.0% 0.3% Peak 4 0.9% 0.8% Fab-dAb peak 85.5% 92.9% Di Fab-dAb peak 11.5% 4.2%
SDS-PAGE
(98) Fab-dAb samples were prepared under non-reduced and reduced conditions and run on a gel as described in the methods. The gel was Coomassie stained. The banding profile of both Fab-dAb samples, FabA-dAbL3 (CK-SG.sub.4SE) and FabA-dAbL3 (CK-G[APAPA].sub.2), corresponds well to the profile observed by high temperature reverse phase HPLC (
(99) Western Blot
(100) Fab-dAb samples were subjected to non-reduced SDS-PAGE followed by western blot analysis with anti-light chain and anti-heavy chain antibodies as described in the methods. This confirmed that the dAb was on the light chain of the Fab and that the heavy chain was unmodified in both samples (
(101) Biacore
(102) Kinetic analysis by SPR as described in the methods was used to assess the binding of human serum albumin to FabA-dAbL3 (CK-SG.sub.4SE) and FabA-dAbL3 (CK-G[APAPA].sub.2). The results in table 5 demonstrate that both constructs are able to bind human serum albumin with a similar affinity (K.sub.D) of approximately 1 M.
(103) TABLE-US-00012 TABLE 5 Construct k.sub.a (10.sup.4 M.sup.1s.sup.1) k.sub.d (10.sup.2 s.sup.1) K.sub.D (10.sup.9M) FabA-dAbL3 3.44 1.42 411 (CK-SG.sub.4SE) FabA-dAbL3 9.61 2.85 296 (CK-G[APAPA].sub.2)
(104) Further kinetic analysis demonstrated that all the fusion constructs retained the interaction characteristics of the original FabA towards IL-1, table 6, with only minor differences seen in the kinetic and affinity parameters.
(105) TABLE-US-00013 TABLE 6 Construct k.sub.a (10.sup.5 M.sup.1s.sup.1) k.sub.d (10.sup.5 s.sup.1) K.sub.D (10.sup.12M) FabA-dAbL3 1.90 4.21 221 (CK-SG.sub.4SE) FabA-dAbL3 2.17 3.99 184 (CK-G[APAPA].sub.2) FabA 2.02 6.46 320
(106) The potential for each construct to bind simultaneously to both human serum albumin and the IL-1 antigen was assessed by capturing each construct to the sensor chip surface, before performing either separate 3 min injections of 5 M human serum albumin or 100 nM IL-1, or a mixed solution of both 5 M human serum albumin and 100 nM IL-1. For each Fab-dAb construct the response seen for the combined HSA/IL-1 solution was almost identical to the sum of the responses of the independent injections, see table 7. This shows that the Fab-dAbs are capable of simultaneous binding to both IL-1 and human serum albumin, and that binding of either IL-1 or human serum albumin does not inhibit the interaction of the other. The original FabA bound only to IL-1, with negligible binding to human serum albumin.
(107) TABLE-US-00014 TABLE 7 Construct Analyte Binding (RU) FabA-dAbL3 (CK-SG.sub.4SE) HSA + IL-1 37.6 HSA 13.2 (37.9) IL-1 24.7 FabA-dAbL3 (CK-G[APAPA].sub.2) HSA + IL-1 61.9 HSA 30.7 (63.6) IL-1 32.9 FabA HSA + IL-1 30.3 HSA 1.3 (30.0) IL-1 28.7
(108) The table above shows the binding response (RU) seen for each construct after separate injections of HSA or IL-1, or injection of premixed HSA and IL-1. In each case the final concentration was 5 M for HSA and 100 nM for IL-1. The sum of the individual HSA and IL-1 responses is shown in parentheses.
Example 7 FabA DidAbs
(109) Expression of FabA-didAbs in E. coli
(110) FabA-dAbs and FabA-didAb fusions terminating with a C-terminal histidine tag (HIS6 tag) were expressed in Escherichia coli. After periplasmic extraction, dAb fusion proteins were purified via the C-terminal His6 tag. Fab expression was analysed by Western blotting of a non-reduced gel with anti-CHI and anti-cKappa antibodies. FabA-dAb and FabA-didAb were expressed as full-length proteins and were shown to react to both antibody detection reagents.
(111) Analysis of FabA-didAbs Expressed in E. coli
(112) Further analysis was conducted to characterise the binding of HSA to FabA constructs to which one or more dAbs were fused. Binding assays were performed on a variety of constructs in which dAbL3 or dAbH4 fused to either the light or heavy chain of the FabA (see Table 8 for details of the constructs and summary of the binding data). Although constructs carrying only dAbH4, on either the light or heavy chain, were seen to bind HSA with comparatively poor affinity (9 M and 3 M respectively), higher affinity binding was observed for constructs carrying dAbL3, either as a single fusion (on either light or heavy chain) or partnered with a second dAb (dAbL3 or dAbH4) on the opposing chain.
(113) TABLE-US-00015 TABLE 8 k.sub.a k.sub.a K.sub.D Construct (10.sup.4 M.sup.1s.sup.1) (10.sup.3 s.sup.1) (10.sup.9M) FabA nb FabA-dAbL3 (LC-SG4SE) 4.46 16.2 363 FabA-dAbH4 (LC SG4SE) 9142 FabA-dAbL3 (HC-DKTHTS) 8.24 15.4 187 FabA-dAbH4 (HC-DKTHTS) 2866 FabA-didAb, 3.00 15.1 502 -dAbL3 (LC-SG4SE) & -dAbL3 (HC-DKTHTS) FabA-didAb, 4.36 16.3 373 -dAbL3 (LC-SG4SE) & -dAbH4 (HC-DKTHTS)
(114) Affinity and kinetic parameters determined for the binding of HSA to FabAs carrying dAbL3 or dAbH4 on either light chain (LC) or heavy chain (HC) or both as indicated. No binding (nb) of HSA to the original FabA was detected. The interaction kinetics for the binding of HSA to the FabA with (dAbH4 on HC) or (dAbH4 on LC), were too rapid to determine, therefore affinity (KD) was determined from steady-state binding.
Example 8
(115) Expression and Purification of FabB-didAbs
(116) Mammalian Expression
(117) Prior to transfection CHO-XE cells were washed in Earls Balanced Salts Solution (EBSS), pelleted and resuspended in EBSS at 210.sup.8 cells/ml. Heavy and light chain plasmids were added to the cells at a total concentration of 400 ug. Optimised electrical parameters for 800 l cells/DNA mix on the in-house electroporator were used for transfection. Transfected cells were directly transferred to 1 L CD-CHO media supplied with glutamax, HT and antimycotic antibiotic solution. Cells were incubated, shaking at 37 C. for 24 hours and then shifted to 32 C. Sodium Butyrate 3 mM was added on day 4. Supernatants were harvested on day 14 by centrifugation at 1500g to remove cells. Expression levels were determined by ELISA.
(118) Mammalian Expression Supernatant Concentration
(119) The mammalian supernatants containing 55 g/ml of FabB-didAb as assessed by ELISA were concentrated from 1.8 L to 200 ml using a Minisette concentrator fitted with a 10 kDa molecular weight cut off polyethersulphone (PES) membrane.
(120) Protein-G Purification
(121) The concentrated supernatants were applied to a GammaBind Plus Sepharose (GE Healthcare) column equilibrated in 20 mM phosphate, 150 mM NaCl pH7.1. The column was washed with 20 mM phosphate, 150 mM NaCl pH7.1 and the bound material eluted with 0.1M glycine/HCl pH2.7. The elution peak was collected and pH adjusted to pH7 with 2M Tris/HCl pH8.8. The pH adjusted elutions were concentrated to 1 mg/ml and diafiltered into 20 mM phosphate, 150 mM NaCl pH7.1 using a 10 kD molecular weight cut off PES membrane.
(122) SDS-PAGE
(123) Samples were diluted with water where required and then to 26 l was added 10 L 4LDS sample running buffer. For non-reduced samples, 4 L of 100 mM NEM was added and for reduced samples 4 L of 10 reducing agent was added. The samples were vortexed, incubated at 85 C. for 5 mins, cooled and centrifuged at 12500 rpm for 30 secs. The prepared samples were loaded on to a 4-20% acrylamine Tris/Glycine SDS gel and run for 110 mins at 125V. The gels were stained with Coomassie Blue protein stain.
(124) ELISA
(125) The yields of Fab-didAb were measured using a sandwich ELISA. Briefly, the Fab-didAb was captured with an anti-CHI antibody then revealed with an anti-kappa-HRP.
(126) SDS-PAGE
(127) FabB and FabB-didAb samples were prepared under non-reduced and reduced conditions and separated on a gel and stained as described in the methods. See
Example 9
(128) Thermofluor Thermal Stability Assay on FabB-Fv
(129) Samples (1 l of sample at 1 mg/ml, 8 l of PBS and 1 l of 30 stock of Sypro orange fluorescent dye) were run in quadruplicate in 384 well plates. The plate is heated from 20-99 C. using a 7900HT fast real-time PCR system and the fluorescence (excitation at 490 nm, emission at 530 nm) measured. The results are shown in Table D and
(130) TABLE-US-00016 TABLE 9 Tm C. (Fab) Tm C. (Fv) FabB-didAb, 81.9 0.6 68.5 0.5 -dAbL1(CK-G.sub.4Sx2) & -dAbL1(CH1-G.sub.4Sx2) FabB-didAb, 82.4 0.2 70.6 0.8 -dAbL2(CK-G.sub.4Sx2) & -dAbL2(CH1-G.sub.4Sx2)
Example 10
(131) Aggregation Stability Assay of FabB-Fv
(132) Samples at 1 mg/ml in PBS were incubated at 25 C. with vortexing at 1400 rpm. The absorbance is measured at 595 nm. This absorbance is due to light scattered by particles and can be correlated with sample aggregation. Both FabB-645Fv (G.sub.4Sx2) and FabB-648Fv (G.sub.4Sx2) are as resistant to aggregation as FabB alone. They are all more resistant to aggregation than the IgG control. (
Example 11
(133) pH Dependency of Fab-Fv Binding to HSA
(134) Binding affinities for the interactions of Fab-Fv constructs with HSA were determined as described in the methods except that the running buffers at pH5.0, 5.5, 6.0 and 7.0 were created by mixing 40 mM citric acid, 150 mM NaCl, 3 mM EDTA, 0.05% v/v surfactant P20 and 80 mM disodium hydrogen phosphate, 150 mM NaCl, 3 mM EDTA, 0.05% v/v surfactant P20 to give the desired pH.
(135) The affinity of FabB-645Fv (G.sub.4Sx2) for HSA is unaffected by pH from 7.4 (standard assay pH) to 5.0. The affinity of FabB-648Fv (G.sub.4Sx2) for HSA is affected by pH and there is approximately a 10 fold loss in affinity between pH7.4 and pH5.0.
(136) TABLE-US-00017 TABLE 10 K.sub.D (10.sup.9M) pH7.0 pH6.0 pH5.5 pH5.0 FabB-645Fv (G.sub.4Sx2) 13.3 12.5 10.7 7.1 FabB-648Fv (G.sub.4Sx2) 3.3 11.1 24.1 47.8
Example 12
(137) In Vivo Murine PK of FabB-Fv
(138) The pharmacokinetics of FabB-645Fv (G.sub.4Sx2) and FabB-648Fv (G.sub.4Sx2) in male BALB/c mouse were determined following a single administration at 10 mg/kg either subcutaneously (sc) or intravenously (iv). Six mice were dosed for each construct and route of administration. Serial blood samples (30 L) were collected from the tail vein at the following time points: 1, 4, 8, 24, 48, 72, 102 and 168 hours following subcutaneous administration and 30 minutes, 1, 8, 24, 48, 72, 96 and 168 hours following intravenous administration. The collected blood was dispensed into a Sarstedt microvette CB300Z with clot activator for serum separation, and left at room temperature for at least 20 minutes. The microvette was then centrifuged at 20 C. at 10,000 rpm for 5 minutes. Serum was removed and stored frozen prior to analysis. The concentration of FabB-645Fv (G.sub.4Sx2) or FabB-648Fv (G.sub.4Sx2) in serum samples was assessed by ELISA. Briefly Nunc Maxisorb Immunomodule Plates were coated with hOX40-Fc in PBS and blocked with 1% BSA in PBS. Serum samples and standards were diluted in 1% BSA in PBS and applied to the plate for 1 hour. The plate was washed with PBS and the revealing antibody of goat anti-human kappa HRP conjugate applied in 1% BSA in PBS for 1 hour. The plate was washed and then developed with TMB substrate followed by stopping with 2.5M sulphuric acid. The absorbance at 630 nm wash measured and the concentrations determined from the standard curve.
(139) Both FabB-645Fv (G.sub.4Sx2) and FabB-648Fv (G.sub.4Sx2) have extended half-life in plasma,
Example 13
(140) In Vivo Efficacy Study of FabB-Fv
(141) A study to investigate if FabB-645Fv and FabB-648Fv are efficacious in vivo was undertaken. Briefly this involved steady state dosing in HuSCID mice and the read out was the prevention of T cell engraftment.
(142) CB17 SCID mice were dosed with a loading dose subcutaneously on day 2 of 2.475 mg/kg FabB-645Fv or FabB-648Fv or FabB-PEG40k or PBS. On every subsequent day up to and including day 10 they were dosed with a maintenance dose subcutaneously of 0.75 mg/kg FabB-645Fv or FabB-648Fv or FabB-PEG40k or PBS. Each dosing group consisted of 9-10 mice. On day 1 all the mice were treated with 0.87 mg/mouse of rat anti-murine TM-1 antibody to abrogate natural killer cell activity. On day 0 all the mice received an inter peritoneal injection of 810.sup.6 human peripheral blood mononuclear cells. On day 14 the mice are sacrificed and the blood, spleen and a peritoneal lavage were taken. The samples were analysed by FACS for CD4.sup.+ and CD8.sup.+ T cells. The data sets were analysed by one way Anova with Dunnett's post test comparison. All the test constructs FabB-645Fv, FabB-648Fv and FabB-PEG40k were equally efficacious in all the compartments tested, i.e. blood peritoneum and spleen.
Example 14
(143) FabB-645Fv mutations to chance the affinity of 645Fv for albumin
(144) Point mutations were introduced into selected residues in the CDRs of the heavy chain of the 645Fv portion of FabB-645dsFv (S3xG.sub.4S) by mutagenic PCR. For example 150A is a replacement of Ile 50 with Ala. The various mutations are given in Table 11 below. The affinity of the Fab-645Fv mutants for human albumin was assessed by BIAcore as described in the methods. All the mutations had either unchanged or reduced affinity for human albumin.
(145) TABLE-US-00018 TABLE 11 Fv heavy mutation Albumin ka (1/Ms) kd (1/s) KD (nM) I50A HSA 3.12E+04 1.90E03 60.9 T56A HSA 4.65E+04 3.78E04 8.12 T95A HSA 2.81E+04 2.64E03 94.0 V96A HSA 2.81E+04 6.42E04 22.9 P97A HSA 4.60E+04 1.26E02 275 G98A HSA 4.73E+04 2.71E04 5.73 Y99A HSA 4.71E+04 4.79E04 10.2 S100A HSA 3.94E+04 1.44E03 36.6 T100aA HSA 3.60E+05 1.86E02 51.6 Y100cA HSA 1.23E+04 1.07E03 87.0 I50A and T95A HSA 2.12E+04 9.94E03 468 I50A and G98A HSA 1.79E+04 6.96E03 389 I50A and Y99A HSA >3500 T56A and T95A HSA 2.84E+04 8.57E04 30.1 T56A and G98A HSA 2.40E+04 3.68E03 153 T56A and Y99A HSA 2.24E+04 1.49E02 664
Example 15
(146) 1-5 Gly4Ser Linker Length Between Fab and Fv
(147) Construction of FabB-645Fv Fusion Plasmids for Expression in Mammalian Cells
(148) The FabB-645Fv's with either a SGGGGS (SEQ ID NO: 224), SGGGGSGGGGS (SEQ ID NO: 225), SGGGGSGGGGSGGGGS (SEQ ID NO: 226), SGGGGSGGGGSGGGGSGGGGS (SEQ ID NO: 227) or SGGGGSGGGGSGGGGSGGGGSGGGGS (SEQ ID NO: 228) linker between the C-termini of the Fab and the N-termini of the Fv were assembled by PCR then cloned into a mammalian expression vectors under the control of the HCMV-MIE promoter and SV40E polyA sequence. The relevant heavy and light chain plasmids were paired for expression in mammalian cells.
(149) Mammalian Expression of FabB-645Fv (1-5G.sub.4S)
(150) HEK293 cells were transfected with the heavy and light chain plasmids using Invitrogen's 293fectin transfection reagent according to the manufacturer's instructions. Briefly, 24 g heavy chain plasmid+24 g light chain plasmid was incubated with 120 l 293fectin+4080 l Optimem media for 20 mins at RT. The mixture was then added to 6010.sup.6 HEK293 cells in 60 mL suspension and incubated for 4 days with shaking at 37 C. All the constructs were equally well expressed.
(151) Protein-G Purification
(152) The mammalian expression suspensions were clarified by centrifugation and the supernatants were concentrated to 1.8 mL using 10 kDa molecular weight cut off centrifugation concentrators. The concentrated supernatants were centrifuged at 16000g for 10 min to remove any precipitate and then 1.5 mL was loaded onto 1 ml HiTrap Protein-G columns (GE Healthcare) at 1 ml/min. The columns were washed with 20 mM phosphate, 40 mM NaCl pH7.4 and bound material eluted with 0.1M glycine/HCl pH2.7. The elution peak (2 mL) was collected and pH adjusted to pH5 with 250 L of 1M sodium acetate. The pH adjusted elutions were diafiltered into 20 mM phosphate, 150 mM NaCl pH7.1 using 10 kDa molecular weight cut off centrifugation concentrators and concentrated to 250 L. All the constructs had similar purification profiles and the final concentrations were 0.5-1.1 mg/ml.
(153) Affinity of FabB-645Fv (1-5G.sub.4S) for Albumin
(154) The affinities of the purified FabB-645Fv (1-5G.sub.4S) constructs for human and mouse albumin were determined as described in the Methods. The different linker lengths of the Fv of 1 to 5Gly4Ser between the C-termini of the Fab and the N-termini of the Fv had no affect on the affinity of the 645Fv for either human or mouse albumin.
(155) TABLE-US-00019 TABLE 12 Albumin KD (nM) Albumin KD (nM) FabB-645Fv (1xG.sub.4S) Human 8.77 Mouse 2.18 FabB-645Fv (2xG.sub.4S) Human 6.72 Mouse 8.01 FabB-645Fv (3xG.sub.4S) Human 9.87 Mouse 8.92 FabB-645Fv (4xG.sub.4S) Human 7.90 Mouse 7.24 FabB-645Fv (5xG.sub.4S) Human 3.90 Mouse 6.09
SDS-PAGE Analysis of Purified FabB-645Fv (1-5G.sub.4S)
(156) FabB-645Fv (1-5G.sub.4S) samples were prepared under non-reduced and reduced conditions and separated on a gel and stained as described in the methods. See
(157) Size Exclusion Analysis of Purified FabB-645Fv (1-5G.sub.4S)
(158) FabB-645Fv (1-5G.sub.4S) samples were analysed for size on a Superdex200 10/300GL Tricorn column (GE Healthcare) developed with an isocratic gradient of 20 mM phosphate 150 mM NaCl pH7.4 at 1 ml/min.
(159) A linker length between the C-termini of the Fab and the N-termini of the Fv of either 1G.sub.4S or 2G.sub.4S reduces the amount of monomer FabB-645Fv whilst increasing the amount of dimer and higher multimers. The amount of monomer is least for the 1G.sub.4S linker length. A linker length between the C-termini of the Fab and the N-termini of the Fv of either 3G.sub.4S, 4xa.sub.4S or 5G.sub.4S increased the amount of monomer FabB-645Fv whilst decreasing the amount of dimer and higher multimers with the levels being similar for all three linker lengths.
(160) TABLE-US-00020 TABLE 13 Monomer Dimer High Multimers FabB-645Fv (1xG.sub.4S) 5% 47% 48% FabB-645Fv (2xG.sub.4S) 27% 38% 36% FabB-645Fv (3xG.sub.4S) 51% 32% 17% FabB-645Fv (4xG.sub.4S) 55% 30% 15% FabB-645Fv (5xG.sub.4S) 51% 31% 18%
Thermofluor Thermal Stability Analysis of Purified FabB-645Fv (1-5G.sub.4S)
(161) Samples (1 l of sample at 1 mg/ml, 8 l of PBS and 1 l of 30 stock of Sypro orange fluorescent dye) were run in quadruplicate in 384 well plates. The plate is heated from 20-99 C. using a 7900HT fast real-time PCR system and the fluorescence (excitation at 490 nm, emission at 530 nm) measured. The results are shown in Table 14 and
(162) TABLE-US-00021 TABLE 14 Tm C. (Fab) Tm C. (Fv) FabB-645Fv (1xG.sub.4S) 82.8 0.6 67.4 0.4 FabB-645Fv (2xG.sub.4S) 83.4 0.3 68.7 0.3 FabB-645Fv (3xG.sub.4S) 83.4 0.3 69.5 0.6 FabB-645Fv (4xG.sub.4S) 83.8 0.3 71.3 1.0 FabB-645Fv (5xG.sub.4S) 83.8 0.4 72.0 0.7
Example 16
(163) Disulphide Stabilisation of the Fv in a Fab-Fv
(164) FabB-645dsFv (2G.sub.4S), FabB-648dsFv (2G.sub.4S), FabB-645dsFv (2G.sub.4S) and FabB-648dsFv (2G.sub.4S) Fusion Plasmids for Expression in Mammalian Cells
(165) Point mutations were introduced into the FabB-645Fv (2G.sub.4S) and FabB-648Fv (2G.sub.4S) DNA sequences at selected residues in the framework region of both the heavy chain and the light chain of the Fv by mutagenic PCR. The mutations introduced to create an interchain disulphide bond between the heavy and light chains of the Fv were heavy chain G44C and light chain G100C. As well as adding the cysteins to create the interchain disulphide bond in the Fv, the natural interchain disulphide between the heavy chain and light chain of the Fab was removed by mutagenic PCR by changing the cysteines to serines. Fvs that contain an interchain disulphide bond were termed dsFv, Fabs that lack an interchain disulphide bond were termed FabA. The DNA for all these constructs was then cloned into a mammalian expression vectors under the control of the HCMV-MIE promoter and SV40E polyA sequence. The relevant heavy and light chain plasmids were paired for expression in mammalian cells.
(166) Mammalian Expression of FabB-645dsFv (2G.sub.4S), FabB-648dsFv (2G.sub.4S), FabB-645dsFv (2G.sub.4S) and FabB-648dsFv (2G.sub.4S)
(167) HEK293 cells were transfected with the heavy and light chain plasmids using Invitrogen's 293fectin transfection reagent according to the manufacturer's instructions. Briefly, 24 g heavy chain plasmid+24 g light chain plasmid was incubated with 120 l 1293fectin+4080 l Optimem media for 20 mins at RT. The mixture was then added to 6010.sup.6 HEK293 cells in 60 mL suspension and incubated for 4 days with shaking at 37 C. All the constructs were equally well expressed.
(168) Protein-G Purification of FabB-645dsFv (2G.sub.4S), FabB-648dsFv (2G.sub.4S), FabB-645dsFv (2G.sub.4S) and FabB-648dsFv (2G.sub.4S)
(169) The mammalian expression suspensions were clarified by centrifugation and the supernatants were concentrated to 1.8 mL using 10 kDa molecular weight cut off centrifugation concentrators. The concentrated supernatants were centrifuged at 16000g for 10 min to remove any precipitate and then 1.5 mL was loaded onto 1 ml HiTrap Protein-G columns (GE Healthcare) at 1 ml/min. The columns were washed with 20 mM phosphate, 40 mM NaCl pH7.4 and bound material eluted with 0.1 M glycine/HCl pH2.7. The elution peak (2 mL) was collected and pH adjusted to pH5 with 250 L of 1M sodium acetate. The pH adjusted elutions were diafiltered into 20 mM phosphate, 150 mM NaCl pH7.1 using 10 kDa molecular weight cut off centrifugation concentrators and concentrated to 250 L. All the constructs had similar purification profiles and the final concentrations were 0.5-0.8 mg/ml.
(170) Affinity of FabB-645dsFv (2G.sub.4S), FabB-648dsFv (2G.sub.4S), FabB-645dsFv (2G.sub.4S) and FabB-648dsFv (2G.sub.4S) for Albumin
(171) The affinities of the purified FabB-645dsFv (2G.sub.4S), FabB-648dsFv (2G.sub.4S) FabB-645dsFv (2G.sub.4S), FabB-648dsFv (2G.sub.4S) constructs for human and mouse albumin were determined as described in the Methods. The disulphide stabilisation of the Fv had no affect or slightly increased the affinity of the Fv for both human or mouse albumin.
(172) TABLE-US-00022 TABLE 15 Albumin KD (nM) Albumin KD (nM) FabB-645Fv (2xG.sub.4S) Human 17.5 Mouse 24.7 FabB-645dsFv (2xG.sub.4S) Human 12.6 Mouse 14.0 FabB-645dsFv (2xG.sub.4S) Human 8.3 Mouse 12.2 FabB-648Fv (2xG.sub.4S) Human 9.4 Mouse 42.4 FabB-648dsFv (2xG.sub.4S) Human 3.1 Mouse 59.6 FabB-648dsFv (2xG.sub.4S) Human 8.3 Mouse 59.8
SDS-PAGE Analysis of Purified FabB-645dsFv (2G.sub.4S), FabB-648dsFv (2G.sub.4S), FabB-645dsFv (2G.sub.4S) and FabB-648dsFv (2G.sub.4S)
(173) Purified FabB-645dsFv (2G.sub.4S), FabB-648dsFv (2xa.sub.4S) FabB-645dsFv (2xa.sub.4S), FabB-648dsFv (2G.sub.4S) samples were prepared under non-reduced and reduced conditions and separated on a gel and stained as described in the methods. See
(174) Size Exclusion Analysis of Purified FabB-645dsFv (2G.sub.4S), FabB-648dsFv (2G.sub.4S), FabB-645dsFv (2G.sub.4S) and FabB-648dsFv (2G.sub.4S)
(175) Purified FabB-645dsFv (2xa.sub.4S), FabB-648dsFv (2G.sub.4S) FabB-645dsFv (2xa.sub.4S), FabB-648dsFv (2G.sub.4S) samples were analysed for size on a Superdex200 10/300GL Tricorn column (GE Healthcare) developed with an isocratic gradient of 20 mM phosphate 150 mM NaCl pH7.4 at 1 ml/min.
(176) The introduction of an interchain disulphide bond into the Fv of either a 645Fv or 648Fv increased the amount of monomer Fab-Fv species compared with the Fab-Fv in which the Fv did not have an inter chain disulphide. The removal of the natural interchain disulphide bond from the Fab part of a Fab-Fv had only a small effect on the amount of monomer species present.
(177) TABLE-US-00023 TABLE 16 Monomer Dimer High Multimers FabB-645Fv (2xG.sub.4S) 26% 38% 35% FabB-645dsFv (2xG.sub.4S) 43% 21% 37% FabB-645dsFv (2xG.sub.4S) 40% 25% 34% FabB-648dsFv (2xG.sub.4S) 50% 26% 24% FabB-648dsFv (2xG.sub.4S) 55% 24% 20%
Thermofluor Thermal Stability Analysis of Purified FabB-645dsFv (2G.sub.4S), FabB-648dsFv (2G.sub.4S), FabB-645dsFv (2G.sub.4S) and FabB-648dsFv (2G.sub.4S)
(178) Samples (1 l of sample at 1 mg/ml, 8 l of PBS and 1 l of 30 stock of Sypro orange fluorescent dye) were run in quadruplicate in 384 well plates. The plate is heated from 20-99 C. using a 79001-IT fast real-time PCR system and the fluorescence (excitation at 490 nm, emission at 530 nm) measured.
(179) The introduction of an interchain disulphide bond into the Fv part of a Fab-Fv of either a 645Fv or 648Fv increased the thermal stability of the Fv compared with the Fab-Fv in which the Fv did not have an inter chain disulphide. The removal of the natural interchain disulphide bond from the Fab part of a Fab-Fv decreased the thermal stability of the Fab part of the Fab-Fv
(180) TABLE-US-00024 TABLE 17 Tm C. (Fab) Tm C. (Fv) FabB-645Fv (2xG.sub.4S) 81.9 0.6 68.5 0.5 FabB-645dsFv (2xG.sub.4S) 83.6 0.3 71.6 0.3 FabB-645dsFv (2xG.sub.4S) 79.5 0.1 70.8 0.6 FabB-648Fv (2xG.sub.4S) 82.4 0.2 70.6 0.8 FabB-648dsFv (2xG.sub.4S) 82.8 0.3 75.0 0.6 FabB-648dsFv (2xG.sub.4S) n.d. 73.6 0.8 n.d. = not determined. The analysis software was unable to resolve this inflection point.
(181) Biacore Method for FabD
(182) Binding affinities and kinetic parameters for the interactions of Fab-dAb and Fab-didAb constructs were determined by surface plasmon resonance (SPR) conducted on a Biacore T100 using CM5 sensor chips and HBS-EP (10 mM HEPES (pH7.4), 150 mM NaCl, 3 mM EDTA, 0.05% v/v surfactant P20) running buffer. Human Fab samples were captured to the sensor chip surface using either a human F(ab).sub.2-specific goat Fab (Jackson ImmunoResearch, 109-006-097) or an in-house generated anti human CH1 monoclonal antibody. Murine Fab samples were captured using a murine F(ab)2-specific goat Fab (Jackson ImmunoResearch, 115-006-072). Covalent immobilisation of the capture antibody was achieved by standard amine coupling chemistry.
(183) Each assay cycle consisted of firstly capturing the Fab-dAb or Fab-didAb construct using a 1 min injection, before an association phase consisting of a 3 min injection of antigen, after which dissociation was monitored for 5 min. After each cycle, the capture surface was regenerated with 21 min injections of 40 mM HCl followed by 30 s of 5 mM NaOH. The flow rates used were 10 l/min for capture, 30 l/min for association and dissociation phases, and 10 l/min for regeneration.
(184) For kinetic assays, a titration of antigen (for human or mouse serum albumin typically 62.5 nM-4M, for IL-1 1.25-40 nM, for cell surface receptor D 20-1.25 nM) was performed, a blank flow-cell was used for reference subtraction and buffer-blank injections were included to subtract instrument noise and drift.
(185) Kinetic parameters were determined by simultaneous global-fitting of the resulting sensorgrams to a standard 1:1 binding model using Biacore T100 Evaluation software.
(186) In order to test for simultaneous binding, 3 min injections of either separate 5 M HSA or 100 nM IL-1, or a mixed solution of 5 M HSA and 100 nM IL-1 were injected over the captured Fab-dAb. Simultaneous binding of albumin and cell surface receptor D was assessed in the same manner using final concentrations of 2 M HSA or MSA and 20 nM murine cell surface receptor D.
Example 17
(187) Mammalian Expression of mFabC-mdidAbs and mFabD-mdidAbs
(188) HEK293 cells were transfected with the heavy and light chain plasmids using Invitrogen's 293fectin transfection reagent according to the manufacturer's instructions. Briefly, 2 g heavy chain plasmid+2 g light chain plasmid was incubated with 10 l 293fectin+3400 Optimem media for 20 mins at RT. The mixture was then added to 510.sup.6 HEK293 cells in suspension and incubated for 6 days with shaking at 37 C.
(189) ELISA
(190) The yields of mFab-mdidAb were measured using a sandwich ELISA. Briefly, the mFab-mdidAb was captured with an anti-CH1 antibody then revealed with an anti-kappa-HRP.
(191) TABLE-US-00025 TABLE 18 ELISA expression (ug/mL) mFabD-mdidAb, -dAbL1(CK-G.sub.4Sx2) & 44 -dAbH1(CH1-G.sub.4Sx2) mFabD-mdidAb, -dAbL2(CK-G.sub.4Sx2) & 35 -dAbH2(CH1-G.sub.4Sx2) mFabC-mdidAb, -dAbL1(CK-G.sub.4Sx2) & 11 -dAbH1(CH1-G.sub.4Sx2) mFabC-mdidAb, -dAbL2(CK-G.sub.4Sx2) & 14 -dAbH2(CH1-G.sub.4Sx2)
Example 18
(192) Further kinetic analysis was conducted to assess the interactions of serum albumin and human OX40 to the purified FabB-didAb, -dAbL1(CK-G4Sx2) & -dAbH1(CH1-G4Sx2) and FabB-didAb, -dAbL2(CK-G4Sx2) & -dAbH2(CH1-G4Sx2) fusions (Table 19). Both FabB-didAb, -dAbL1(CK-G4Sx2) & -dAbH1(CH1-G4Sx2) and FabB-didAb, -dAbL2(CK-G4Sx2) & -dAbH2(CH1-G4Sx2) retained the affinity for human OX40 of the original FabB (Table 20).
(193) The potential for the FabB-didAb, -dAbL1(CK-G4Sx2) & -dAbH1(CH1-G4Sx2) and FabB-didAb, -dAbL2(CK-G4Sx2) & -dAbH2(CH1-G4Sx2) constructs to bind simultaneously to both human or mouse serum albumin and human OX40 was assessed by capturing each Fab-didAb construct to the sensor chip surface, before performing either separate 3 min injections of 2 M albumin (human or mouse) or 50 nM human OX40, or a mixed solution of both 2 M albumin and 50 nM OX40. HSA binding was seen for both Fab-didAb constructs. For each Fab-didAb construct the response seen for the combined albumin/OX40 solution was almost identical to the sum of the responses of the independent injections (summarised in table 21). This shows that the Fab-didAbs are capable of simultaneous binding to both OX40 and serum albumin. The original FabB bound only OX40, with no significant binding to either human or mouse albumin.
(194) TABLE-US-00026 TABLE 19 Albu- k.sub.a k.sub.d K.sub.D Construct min (10.sup.4M.sup.1s.sup.1) (10.sup.5s.sup.1) (10.sup.9M) FabB-didAb, HSA 1.65 2.06 12.5 -dAbL1(CK-G4Sx2) & -dAbH1(CH1-G4Sx2 FabB-didAb, HSA 1.80 1.24 6.92 -dAbL2(CK-G4Sx2) & -dAbH2(CH1-G4Sx2 FabB-didAb, MSA 1.83 1.82 9.94 -dAbL1(CK-G4Sx2) & -dAbH1(CH1-G4Sx2 FabB-didAb, MSA nd nd -dAbL2(CK-G4Sx2) & -dAbH2(CH1-G4Sx2
(195) Affinity and kinetic parameters determined for HSA and MSA binding to Fab-didAb fusions.
(196) TABLE-US-00027 TABLE 20 k.sub.a k.sub.d K.sub.D Construct (10.sup.5M.sup.1s.sup.1) (10.sup.5s.sup.1) (10.sup.12M) FabB 2.92 22.6 775 FabB-didAb, 3.58 8.54 238 -dAbL1(CK-G4Sx2) & -dAbH1(CH1-G4Sx2 FabB-didAb, 3.27 13.6 415 -dAbL2(CK-G4Sx2) & -dAbH2(CH1-G4Sx2
(197) Affinity and kinetic parameters for hOX40-Fc binding to FabB and FabB-didAb fusions.
(198) TABLE-US-00028 TABLE 21 Construct Analyte Binding (RU) FabB HSA 2.5 MSA 2.5 OX40 89.5 HSA + 90.1 (92) OX40 MSA + 86.5 (87) OX40 FabB-didAb, HSA 109.1 -dAbL1(CK-G4Sx2) & -dAbH1(CH1-G4Sx2 MSA 93.3 OX40 73.7 HSA + 186.1 (182.8) OX40 MSA + 170.3 (167) OX40 FabB-didAb, HSA 50.9 -dAbL2(CK-G4Sx2) & -dAbH2(CH1-G4Sx2 MSA 2.4 OX40 52.9 HSA + 104.2 (103.8) OX40 MSA + 54.9 (55.3) OX40
(199) The table above shows the binding response (RU) seen for each construct after separate injections of HSA or MSA or hOX40-Fc, or injection of premixed albumin and hOX40-Fc. In each case the final concentration was 2 M albumin HSA and 50 nM hOX40-Fc. The sum of the individual albumin and hOX40-Fc responses is shown in parentheses.
Example 19
(200) Further kinetic analysis was conducted to assess the interactions of serum albumin and murine cell surface receptor D to mFabD-mdidAb, -mdAbL1(CK-G.sub.4Sx2) & mdAbH1(CH1-G.sub.4Sx2) and mFabD-mdidAb, -mdAbL2(CK-G.sub.4Sx2) & mdAbH2(CH1-G.sub.4Sx2) (Table 22). Both mFabD-mdidAbs showed relatively high affinity binding to HSA (K.sub.D=2.78 nM and 8.97 nM respectively). mFabD-mdidAb, -mdAbL2(CK-G.sub.4Sx2) & mdAbH2(CH1-G.sub.4Sx2) also bound MSA with a similar affinity (K.sub.D=22 nM), however no binding to MSA was seen for mFabD-mdidAb, -mdAbL1(CK-G.sub.4Sx2) & mdAbH1(CH1-G.sub.4Sx2). Both mFabD-mdidAbs retained the affinity for murine cell surface receptor D of the original mFabD (Table 23).
(201) The potential for mFabD-mdidAb, -mdAbL1(CK-G.sub.4Sx2) & mdAbH1(CH1-G.sub.4Sx2) and mFabD-mdidAb, -mdAbL2(CK-G.sub.4Sx2) & mdAbH2(CH1-G.sub.4Sx2) to bind simultaneously to both human or mouse serum albumin and murine cell surface receptor D was assessed by capturing each mFab-mdidAb construct to the sensor chip surface, before performing either separate 3 min injections of 2 M albumin (human or mouse) or 20 nM murine cell surface receptor D, or a mixed solution of both 2 M albumin and 20 nM cell surface receptor D. Again HSA binding was seen for both mFab-mdidAb constructs whereas only mFabD-mdidAb, -mdAbL2(CK-G.sub.4Sx2) & mdAbH2(CH1-G.sub.4Sx2) bound MSA. For each mFab-mdidAb construct the response seen for the combined albumin/cell surface receptor D solution was almost identical to the sum of the responses of the independent injections (summarised in table 24). This shows that the mFab-mdidAbs are capable of simultaneous binding to both cell surface receptor D and serum albumin. The original mFabD bound only cell surface receptor D, with no significant binding to either human or mouse albumin.
(202) TABLE-US-00029 TABLE 22 Albu- k.sub.a k.sub.d K.sub.D Construct min (10.sup.4M.sup.1s.sup.1) (10.sup.5s.sup.1) (10.sup.9M) mFabD-mdidAb, HSA 1.01 2.82 2.78 -mdAbL1(CK-G.sub.4Sx2) & mdAbH1(CH1-G.sub.4Sx2) mFabD-mdidAb, HSA 1.19 10.69 8.97 -mdAbL2(CK-G.sub.4Sx2) & mdAbH2(CH1-G.sub.4Sx2) mFabD-mdidAb, MSA -mdAbL1(CK-G.sub.4Sx2) & mdAbH1(CH1-G.sub.4Sx2) mFabD-mdidAb, MSA 1.03 22.73 22.06 -mdAbL2(CK-G.sub.4Sx2) & mdAbH2(CH1-G.sub.4Sx2)
(203) Affinity and kinetic parameters determined for HSA and MSA binding to mFabD-mdidAb, -mdAbL1(CK-G4Sx2) & mdAbH1(CH1-G4Sx2) and mFabD-mdidAb, -mdAbL2(CK-G4Sx2) & mdAbH2(CH1-G4Sx2).
(204) TABLE-US-00030 TABLE 23 k.sub.a k.sub.d K.sub.D Construct (10.sup.5M.sup.1s.sup.1) (10.sup.5s.sup.1) (10.sup.12M) mFabD 1.98 2.50 126 mFabD-mdidAb, 2.01 4.67 233 -mdAbL1(CK-G.sub.4Sx2) & mdAbH1(CH1-G.sub.4Sx2) mFabD-mdidAb, 3.62 6.36 176 -mdAbL2(CK-G.sub.4Sx2) & mdAbH2(CH1-G.sub.4Sx2)
(205) Affinity and kinetic parameters for murine cell surface receptor D-Fc binding to mFabD, mFabD-mdidAb, -mdAbL1 (CK-G.sub.4Sx2) & mdAbH1 (CH1-G.sub.4Sx2) and mFabD-mdidAb, -mdAbL2(CK-G.sub.4Sx2) & mdAbH2(CH1-G.sub.4Sx2).
(206) TABLE-US-00031 TABLE 24 Construct Analyte Binding (RU) mFabD receptor D 61.3 HSA 0.9 MSA 1.1 receptor D + 62.9 (62.2) HSA receptor D + 59.2 (60.2) MSA mFabD-mdidAb, receptor D 39.8 -mdAbL1(CK-G.sub.4Sx2) & HSA 59.9 mdAbH1(CH1-G.sub.4Sx2) MSA 0.6 receptor D + 101.2 (99.7) HSA receptor D + 39.9 (39.2) MSA mFabD-mdidAb, receptor D 42.6 -mdAbL2(CK-G.sub.4Sx2) & HSA 61.9 mdAbH2(CH1-G.sub.4Sx2) MSA 43.5 receptor D + 105.3 (104.5) HSA receptor D + 86.3 (86.1) MSA
(207) The table above shows the binding response (RU) seen for each construct after separate injections of HSA or MSA or murine cell surface receptor D-Fc, or injection of premixed albumin and murine cell surface receptor D-Fc. In each case the final concentration was 2 M albumin HSA and 20 nM murine cell surface receptor D-Fc. The sum of the individual albumin and murine cell surface receptor D-Fc responses is shown in parentheses.
Example 20
(208) Further analysis was conducted to assess the simultaneous interaction of mFabD-mdidAb, -mdAbL1 (CK-G.sub.4Sx2) & mdAbH1 (CH1-G.sub.4Sx2) or mFabD-mdidAb, -mdAbL2 (CK-G.sub.4Sx2) & mdAbH2 (CH1-G.sub.4Sx2) with serum albumin and murine cell surface receptor D expressed on the cell surface. Both mFabD-mdidAbs were capable of binding FITC labelled HSA and cell surface receptor X expressed on the cell surface of activated murine T-cells simultaneously (