Multispecific antibody constructs
11345760 · 2022-05-31
Assignee
Inventors
- Pallavi Bhatta (Slough, GB)
- Emma Dave (Slough, GB)
- Sam Philip Heywood (Slough, GB)
- David Paul Humphreys (Slough, GB)
Cpc classification
C07K2317/64
CHEMISTRY; METALLURGY
C07K2317/94
CHEMISTRY; METALLURGY
International classification
Abstract
The present disclosure relates to a multi-specific antibody molecule comprising or consisting of: a) a polypeptide chain of formula (I): VH—CH.sub.1—X—V.sub.1; and b) a polypeptide chain of formula (II): VL-CL-Y—V.sub.2 and pharmaceutical formulations comprising, for example for use in treatment. The disclosure also provides polynucleotide sequences encoding said multispecific antibody molecules, vectors comprising the polynucleotides and host cells comprising said vectors and/or polynucleotide sequences. There is a provided a method of expressing a multispecific antibody molecule of the present disclosure from a host cell.
Claims
1. A multi-specific antibody molecule comprising a) a polypeptide chain of formula (I):
V.sub.H—CH.sub.1—X—V.sub.1; and b) a polypeptide chain of formula (II):
V.sub.L—C.sub.L—Y—V.sub.2; wherein: V.sub.H represents a heavy chain variable domain; CH.sub.1 represents a domain of a heavy chain constant region; X represents a bond or linker; Y represents a bond or linker; V.sub.1 represents a dsscFv; V.sub.L represents a light chain variable domain; C.sub.L represents a domain from a light chain constant region; V.sub.2 represents a dsscFv.
2. The multi-specific antibody molecule according to claim 1, wherein X is a linker.
3. The multi-specific antibody molecule according to claim 1, wherein Y is a linker.
4. The multi-specific antibody molecule according to claim 1, wherein the light chain variable domain of V.sub.1 is attached to X.
5. The multi-specific antibody molecule according to claim 1, wherein the light chain variable domain of V.sub.2 is attached to Y.
6. The multi-specific antibody molecule according to claim 1, wherein the light chain and heavy chain variable domains of V.sub.1 and/or the light chain and heavy chain variable domains of V.sub.2 are linked by a disulfide bond between two engineered cysteine residues.
7. The multi-specific antibody molecule according to claim 1, wherein the heavy chain and light chain variable domains of V.sub.1 and/or V.sub.2 are linked by a disulfide bond between two engineered cysteine residues, wherein the position of the pair of engineered cysteine residues is selected from the group comprising or consisting of: V.sub.H37 and V.sub.L95, V.sub.H44 and V.sub.L100, V.sub.H44 and V.sub.L105, V.sub.H45 and V.sub.L87, V.sub.H100 and V.sub.L50, V.sub.H100b and V.sub.L49, V.sub.H98 and V.sub.L46, V.sub.H101 and V.sub.L46, V.sub.H105 and V.sub.L43 and V.sub.H106 and V.sub.L57, wherein the V.sub.H and V.sub.L values are independently selected within a given V.sub.1 or V.sub.2, wherein the numbering is according to Kabat.
8. The multi-specific antibody molecule according to claim 7, wherein the position of the pair of engineered cysteine residues is V.sub.H44 and V.sub.L100, wherein the numbering is according to Kabat.
9. The multi-specific antibody molecule according to claim 1, wherein X is a peptide linker.
10. The multi-specific antibody molecule according to claim 1, wherein Y is a peptide linker.
11. The multi-specific antibody molecule according to claim 1 which is tri-specific.
12. The tri-specific antibody molecule according to claim 11, wherein VH/VL, V1 and V2 are each binding domains that bind to different antigens.
13. A pharmaceutical composition comprising a multi-specific antibody molecule according to claim 1 and at least one excipient.
14. The multi-specific antibody molecule according to claim 9, wherein X is a peptide linker selected from SEQ ID NO: 1, 2, 69, or 70.
15. The multi-specific antibody molecule according to claim 10, wherein Y is a peptide linker selected from SEQ ID NO: 1, 2, 69, or 70.
16. A formulation comprising the multi-specific antibody molecule of claim 1, wherein at least 98% of the multi-specific antibody molecule in the formulation is present in monomeric form after storage for 28 days at 4° C.
17. The multi-specific antibody of claim 1, wherein V.sub.1 and V.sub.2 are specific for two different antigens.
18. A multi-specific antibody molecule consisting of: c) a polypeptide chain of formula (I):
V.sub.H—CH.sub.1—X—V.sub.1; and d) a polypeptide chain of formula (II):
V.sub.L—C.sub.L—Y—V.sub.2; wherein: V.sub.H represents a heavy chain variable domain; CH.sub.1 represents a domain of a heavy chain constant region; X represents a bond or linker; Y represents a bond or linker; V.sub.1 represents a dsscFv; V.sub.L represents a light chain variable domain that binds the first antigen; C.sub.L represents a domain from a light chain constant region; V.sub.2 represents a dsscFv; and wherein V1 and V2 are specific for two different antigens.
19. A formulation comprising a multi-specific antibody molecule comprising: e) a polypeptide chain of formula (I):
V.sub.H—CH.sub.1—X—V.sub.1; and f) a polypeptide chain of formula (II):
V.sub.L—C.sub.L—Y—V.sub.2; wherein: V.sub.H represents a heavy chain variable domain; CH.sub.1 represents a domain of a heavy chain constant region; X represents a bond or linker; Y represents a bond or linker; V.sub.1 represents a dsscFv; V.sub.L represents a light chain variable domain; C.sub.L represents a domain from a light chain constant region; V.sub.2 represents a dsscFv; and wherein at least 98% of the multi-specific antibody molecule in the formulation is present in monomeric form after storage for 28 days at 4° C.
20. The multi-specific antibody molecule of claim 1, wherein only one of V.sub.H/V.sub.L, V.sub.1 or V.sub.2 has specificity for a serum carrier protein.
21. The multi-specific antibody molecule of claim 20, wherein the serum carrier protein is human serum albumin and wherein said V.sub.H/V.sub.L or V.sub.1 or V.sub.2 having specificity for a serum carrier protein forms an albumin binding site.
22. The multi-specific antibody molecule of claim 21, wherein the albumin binding site comprises SEQ ID NO: 71 for CDRH1, SEQ ID NO: 72 for CDRH2, SEQ ID NO: 73 for CDRH3, SEQ ID NO: 74 for CDRL1, SEQ ID NO: 75 for CDRL2 and SEQ ID NO: 76 for CDRL3; or a heavy chain variable domain selected from SEQ ID NO: 77 and SEQ ID NO: 78 and a light chain variable domain selected from SEQ ID NO: 79 and SEQ ID NO: 80.
23. The multi-specific antibody molecule of claim 22, wherein V.sub.1 comprises the albumin binding site.
24. The multi-specific antibody molecule of claim 22, wherein V.sub.2 comprises the albumin binding site.
25. The multi-specific antibody molecule of claim 1, wherein the heavy chain variable domain of V.sub.1 is attached to X.
26. The multi-specific antibody molecule of claim 1, wherein the heavy chain variable domain of V.sub.2 is attached to Y.
Description
BRIEF DESCRIPTION OF THE FIGURES
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
EXAMPLES
(11) Antibody fragments to ANTIGEN 1 are labelled #1
(12) Antibody fragments to ANTIGEN 2 are labelled #2
(13) Antibody fragments to ANTIGEN 3 are labelled #3
(14) Antibody fragments to ANTIGEN 4 are labelled #4
Example 1
Fab-2×dsscFv
(15) Construction of Plasmids for Expression in Mammalian Cells
(16) Plasmids for the expression of Fab #2-(HC)dsscFv #3,(LC)dsscFv #4 and Fab #2-(LC)dsscFv #3,(HC)dsscFv #4 (see
(17) Gene fragments encoding scFv #4 and dsscFv #4 (vHvL and vLvH orientation) were manufactured chemically and fused to Fab #2 as detailed above to generate:
(18) TABLE-US-00004 Plasmid e1: Light#2-(SGGGGSGGGGS [SEQ ID NO: 2])-vL#4- (GGGGSGGGGSGGGGSGGGGS [SEQ ID NO 68])-vH#4; Plasmid f1: Light#2-(SGGGGSGGGGS [SEQ ID NO: 2])-dsvL#4- (GGGGSGGGGSGGGGSGGGGS [SEQ ID NO: 68])-dsvH#4; Plasmid e2: Light#2-(SGGGGSGGGGS [SEQ ID NO: 2])-vH#4- (GGGGSGGGGSGGGGSGGGGS [SEQ ID NO: 68])-vL#4; Plasmid f2: Light#2-(SGGGGSGGGGS [SEQ ID NO: 2])-dsvH#4- (GGGGSGGGGSGGGGSGGGGS [SEQ ID NO: 68])-dsvL#4; Plasmid g1 Heavy#2-(SGGGGTGGGGS [SEQ ID NO: 1])-vL#4- (GGGGSGGGGSGGGGSGGGGS [SEQ ID NO: 68])-vH#4; Plasmid h1: Heavy#2-(SGGGGTGGGGS [SEQ ID NO: 1])-dsvL#4- (GGGGSGGGGSGGGGSGGGGS [SEQ ID NO: 68])-dsvH#4; Plasmid g2: Heavy#2-(SGGGGTGGGGS [SEQ ID NO: 1])-vH#4- (GGGGSGGGGSGGGGSGGGGS [SEQ ID NO: 68])-vL#4; and Plasmid h2: Heavy#2-(SGGGGTGGGGS [SEQ ID NO: 1])-dsvH#4- (GGGGSGGGGSGGGGSGGGGS [SEQ ID NO: 68])-dsvL#4. pND1 plasmid (Fab#2 Heavy-(SGGGGTGGGGS SEQ ID NO: 1)-dsvH#3- (GGGGSGGGGSGGGGSGGGGS SEQ ID NO: 68)-dsvL#3) [Plasmid i] was already available.
(19) The gene fragment encoding dsHLscFv #3 was excised from pND1 and fused to light chain #2 as detailed above to generate:
(20) TABLE-US-00005 Light#2-(SGGGGSGGGGS SEQ ID NO: 2)-dsvH#3- (GGGGSGGGGSGGGGSGGGGS SEQ ID NO: 68)- dsvL#3 [Plasmid j].
(21) All Fab fusion formats were cloned into mammalian expression vectors under the control of the HCMV-MIE promoter and SV40E polyA sequence.
(22) HEK293 Expression of Fab-1×dsscFv-1×scFv and Fab-2×dsscFv Formats
(23) HEK293 cells were transfected with the relevant plasmids using Invitrogen's 293fectin transfection reagent according to the manufacturer's instructions. Plasmids were mixed as shown in Table 4 to express the different constructs:
(24) TABLE-US-00006 TABLE 4 Antibody Construct Plasmids used Fab#2-(HC)dsHLscFv#3, (LC)HLscFv#4 1. Plasmid e2 2. Plasmid i Fab#2-(HC)dsHLscFv#3, (LC)dsHLscFv#4 1. Plasmid f2 2. Plasmid i Fab#2-(HC)dsHLscFv#3, (LC)LHscFv#4 1. Plasmid e1 2. Plasmid i Fab#2-(HC)dsHLscFv#3, (LC)dsLHscFv#4 1. Plasmid f1 2. Plasmid i Fab#2-(LC)dsHLscFv#3, (HC)HLscFv#4 1. Plasmid g2 2. Plasmid j Fab#2-(LC)dsHLscFv#3, (HC)dsHLscFv#4 1. Plasmid h2 2. Plasmid j Fab#2-(LC)dsHLscFv#3, (HC)LHscFv#4 1. Plasmid g1 2. Plasmid j Fab#2-(LC)dsHLscFv#3, (HC)dsLHscFv#4 1. Plasmid h1 2. Plasmid j
(25) The ratio of the plasmids used for the transfections was 1:1. A total of 50 μg plasmid DNA was incubated with 125 μl 293fectin+4.25 ml Optimem media for 20 mins at RT. The mixture was then added to 50 ml HEK293 cells in suspension at 1×10.sup.6 cells/ml and incubated with shaking at 37° C. Supernatants were harvested on day 10 by centrifugation at 1500 g to remove cells and the supernatant was passed through a 0.22 μm filter. Expression level was determined by Protein-G HPLC.
(26) Table 5 shows the results of the Protein-G HPLC assay. As can be seen, the levels of expression of all the constructs were comparable to each other, covering a range of 11-23 μg/ml. Those denoted with an asterisk (*) were expressed in a separate transfection, therefore the absolute expression levels of LC-scFv #3,HC-scFv #4 cannot be compared with HC-scFv #3,LC-scFv #4.
(27) TABLE-US-00007 TABLE 5 Antibody Construct Expression level (μg/ml) Fab#2(HC)dsHLscFv#3, (LC)HLscFv#4 23 Fab#2(HC)dsHLscFv#3, (LC)dsHLscFv#4 18 Fab#2(HC)dsHLscFv#3, (LC)LHscFv#4 21 Fab#2(HC)dsHLscFv#3, (LC)dsLHscFv#4 20 Fab#2(LC)dsHLscFv#3, (HC)HLscFv#4 13* Fab#2(LC)dsHLscFv#3, (HC)dsHLscFv#4 12* Fab#2(LC)dsHLscFv#3, (HC)LHscFv#4 12* Fab#2(LC)dsHLscFv#3, (HC)dsLHscFv#4 11*
Protein-G Purification of HEK293 Expressed Fab-1×dsscFv-1×scFv and Fab-2×dsscFv Formats
(28) The ˜50 ml HEK293 supernatants were concentrated ˜25 fold to ˜2 ml using 10 kDa molecular weight cut off centrifugation concentrators. The concentrated supernatants were applied to a 1 ml HiTrap Protein-G FF column (GE Healthcare) equilibrated in 20 mM phosphate, 40 mM NaCl pH 7.4. The column was washed with 20 mM phosphate, 40 mM NaCl pH 7.4 and the bound material was eluted with 0.1M glycine/HCl pH 2.7. The elution peak was collected and pH adjusted to ˜pH 7.0 with 2M Tris/HCl pH 8.5. The pH adjusted elutions were concentrated and buffer exchanged into PBS pH 7.4 using 10 kDa molecular weight cut off centrifugation concentrators.
(29) SDS-PAGE Analysis of Protein-G Purified, HEK293 Expressed Fab-1×dsscFv-1×scFv and Fab-2×dsscFv Formats
(30) Samples (2 μg) were diluted with PBS to a volume of 9.75 μl to which 3.75 μl 4×LDS sample buffer and 1.5 μl 100 mM N-ethylmaleimide (non-reduced samples) or 1.5 μl 10× NuPAGE reducing agent (reduced samples) was added. The samples were vortexed, incubated at 70° C. for 10 minutes, cooled and centrifuged at 12500 rpm for 30 seconds. The prepared samples were loaded onto a 4-20% acrylamide Tris/Glycine SDS gel and run for ˜100 minutes at 125V, constant voltage. SeeBluePlus2 (Life Technologies) molecular weight ladder was used. The gels were stained with Instant Blue protein stain (Expedeon) and destained with distilled water.
(31) The expected band sizes after reducing and non-reducing SDS-PAGE are indicated in Table 6.
(32) TABLE-US-00008 TABLE 6 Expected band sizes after SDS-PAGE (kDa) (H = heavy chain, L = light chain, +/− reducing agent) −Red +Red −Red +Red Fab-1xdsscFv, ~100 H~50 Fab-2xdsscFv ~100 H~50 1x scFv L~50 L~50
(33) For all proteins, the non-reducing gel was expected to show a band at ˜100 kDa, whilst the reducing gels were expected to show a doublet at ˜50 kDa with equal staining in the both bands.
(34) For all Fab-1×dsscFv-1×scFv and Fab-2×dsscFv proteins, the reducing SDS-PAGE gels showed banding patterns which indicated that the constructs were being expressed correctly in terms of both migration position and staining intensity with a doublet at ˜50 kDa (
(35) G3000 SE-HPLC Analysis of Protein-G Purified, HEK293 Expressed Fab-1×dsscFv-1×scFv and Fab-2×dsscFv Formats
(36) 10 μg purified protein samples (100 μl of 0.1 mg/ml stock diluted in PBS) were injected onto a TSK Gel G3000SWXL, 7.8×300 mm, column (Tosoh) 3 days post-purification and developed with an isocratic gradient of 200 mM phosphate pH 7.0 at 1 ml/min. Signal detection was by absorbance at 280 nm.
(37) The results are shown in
(38) Although all the samples have similar monomer levels in the assay, those samples that contain a scFv lacking the Fv disulphide are in a dynamic equilibrium. This means the % monomer measured in the assay for these samples is dependent on the concentration of the sample, more concentrated samples give higher % monomer and less concentrated samples give lower % monomer. In contrast the samples where both scFv contain a disulphide are stable and are not in dynamic equilibrium. Therefore the % monomer measured in the assay does not change with changes in the sample concentration.
(39) As the monomers and multimers in the Fab-2×dsscFv formats are stable and not in a dymamic equilibrium the samples can be easily further purified to increase the % monomer. The purified monomeric samples will also remain monomeric even when the concentration of the construct in a given formulation is increased, thereby making the constructs very suitable for use in pharmaceutical preparations. In contrast, Fab-2×scFv can after purification as monomer be subject to intermolecular dynamic domain exchange between the vL and vH of the scFv domains. In addition therefore to increased risk of formation of dimer, trimer, higher order structures and aggregate such molecules are difficult to observe since non-aggregated forms can resolve to monomer after dilution steps which are used during analytical methods. Hence, Fab-2×dsscFv provide additional clarity during analysis of pharmaceutical compositions.
(40) CHOS Expression of Fab-1×dsscFv-1×scFv and Fab-2×dsscFv Formats
(41) CHOS cells were transfected with the relevant plasmids by electroporation methods at 1 L scale. Plasmids were mixed as shown in Table 7 to express the protein. Cultures were grown in CD-CHO medium supplemented with 2 mM GlutaMAX and incubated at 37° C. with 8% CO.sub.2 at 140 rpm for 24 h and subsequently incubated for a further 13 days at 32° C. At day 4 post-transfection, sodium butyrate at a final concentration of 3 mM was added to the culture. On day 14 post-transfection, culture supernatants were harvested by centrifugation and 0.22 μm filter sterilized. Expression titres were measured by Protein G HPLC (Table 8).
(42) TABLE-US-00009 TABLE 7 Antibody Construct Plasmids used Fab#2-(LC)dsHLscFv#3, (HC)HLscFv#4 3. Plasmid g2 1. Plasmid j Fab#2-(LC)dsHLscFv#3, 3. Plasmid h2 1. Plasmid j (HC)dsHLscFv#4 Fab#2-(LC)dsHLscFv#3, (HC)LHscFv#4 3. Plasmid g1 1. Plasmid j Fab#2-(LC)dsHLscFv#3, 3. Plasmid h1 1. Plasmid j (HC)dsLHscFv#4
(43) TABLE-US-00010 TABLE 8 Antibody Construct Expression level (μg/ml) Fab#2-(LC)dsHLscFv#3, (HC)HLscFv#4 15 Fab#2-(LC)dsHLscFv#3, (HC)dsHLscFv#4 17 Fab#2-(LC)dsHLscFv#3, (HC)LHscFv#4 18 Fab#2-(LC)dsHLscFv#3, (HC)dsLHscFv#4 16
Protein-G Purification of CHO Expressed Fab-1×dsscFv-1×scFv and Fab-2×dsscFv Formats
(44) The 1 L CHO supernatants were concentrated ˜25 fold to ˜40 ml using a 10 kDa molecular weight cut off Amicon stirred cell. The concentrated supernatants were loaded onto an AKTA Express Purification system with a Protein G column and PBS as the running buffer. The bound material was eluted with 0.1 M glycine/HCl pH 2.7 and pH adjusted to ˜pH 7.0 with 2M Tris/HCl pH 8.5. The eluted material was then concentrated using 10 kDa molecular weight concentrators and the resulting concentrate was loaded onto a Superdex column (GE Healthcare) for gel filtration with PBS as the running buffer. Individual elution peaks were collected and analyzed by size exclusion HPLC to determine the monomeric fraction. The final monomeric protein was concentrated to 5 mg/ml in PBS and stored at 4° C.
(45) SDS-PAGE Analysis of Protein-G Purified, CHOS Expressed Monomeric Fraction of Fab-1×dsscFv-1×scFv and Fab-2×dsscFv Formats
(46) Samples (2 μg) were diluted with PBS to a volume of 9.75 μl to which 3.75 μl 4×LDS sample buffer and 1.5 μl 100 mM N-ethylmaleimide (non-reduced samples) or 1.5 μl 10× NuPAGE reducing agent (reduced samples) was added. The samples were vortexed, incubated at 70° C. for 10 minutes, cooled and centrifuged at 12500 rpm for 30 seconds. The prepared samples were loaded onto a 4-20% acrylamide Tris/Glycine SDS gel and run for ˜100 minutes at 125V, constant voltage. Mark12 (Life Technologies) molecular weight ladder was used. The gels were stained with Instant Blue protein stain (Expedeon) and destained with distilled water.
(47) The expected band sizes after reducing and non-reducing SDS-PAGE are indicated in Table 9.
(48) TABLE-US-00011 TABLE 9 Expected band sizes after SDS-PAGE (kDa) (H = heavy chain, L = light chain, +/− reducing agent) −Red +Red −Red +Red Fab-1xdsscFv, ~100 H~50 Fab-2xdsscFv ~100 H~50 1x scFv L~50 L~50
(49) For all proteins, the non-reducing gel was expected to show a band at ˜100 kDa, whilst the reducing gels were expected to show a doublet at ˜50 kDa with equal staining in the both bands.
(50) For all Fab-1×dsscFv-1×scFv and Fab-2×dsscFv proteins, the reducing SDS-PAGE gels showed banding patterns which indicated that the constructs were monomeric and expressed correctly in terms of both migration position and staining intensity with a doublet at ˜50 kDa (
(51) G3000 SE-HPLC Time-Course Analysis of Monomeric Fab-1×dsscFv-1×scFv and Fab-2×dsscFv Formats
(52) A 5 mg/ml of purified monomer of the antibody formats was stored at 4° C. in PBS and analyzed on day 4, day 14 and day 28 after purification. Samples were diluted in PBS to 10 μg and injected onto a TSK Gel G3000SWXL, 7.8×300 mm, column (Tosoh) and developed with an isocratic gradient of 200 mM phosphate pH 7.0 at 1 ml/min. Signal detection was by absorbance at 280 nm. The results are shown in
Example 2
Fab-2×dsscFv (Bivalent with Two of the Same dsscFv)
(53) Construction of Plasmids for Expression in Mammalian Cells
(54) Plasmids for the expression of Fab #2-(HC)dsscFv #3,(LC)dsscFv #3 (see
(55) pND1 plasmid (Fab #2 Heavy-(SGGGGTGGGGS [SEQ ID NO: 1)-dsvH #3-(GGGGSGGGGSGGGGSGGGGS [SEQ ID NO: 68])-dsvL #3) [Plasmid i] was already available. The gene fragment encoding dsHLscFv #3 was excised from pND1 and fused to light chain #2 as detailed above to generate: Light #2-(SGGGGSGGGGS [SEQ ID NO: 2])-dsvH #3-(GGGGSGGGGSGGGGSGGGGS [SEQ ID NO: 68])-dsvL #3 [Plasmid j].
(56) The Fab fusion formats were cloned into mammalian expression vectors under the control of the HCMV-MIE promoter and SV40E polyA sequence.
(57) HEK293 Expression of Fab-2×dsscFv #3
(58) HEK293 cells were transfected with the relevant plasmids using Invitrogen's 293fectin transfection reagent according to the manufacturer's instructions. Plasmids were mixed as shown in Table 10 to express the protein:
(59) TABLE-US-00012 TABLE 10 Antibody Construct Plasmids used Fab#2-(HC)dsHLscFv#3, (LC)dsHLscFv#3 1. Plasmid j 2. Plasmid i
(60) The ratio of the plasmids used for the transfections was 1:1. A total of 50 μg plasmid DNA was incubated with 125 μl 293fectin+4.25 ml Optimem media for 20 mins at RT. The mixture was then added to 50 ml HEK293 cells in suspension at 1×10.sup.6 cells/ml and incubated with shaking at 37° C. Supernatants were harvested on day 10 by centrifugation at 1500 g to remove cells and the supernatant was passed through a 0.22 μm filter. Expression level was determined by Protein-G HPLC. Table 11 shows the results of the Protein-G HPLC assay. The level of expression was 20 μg/ml.
(61) TABLE-US-00013 TABLE 11 Antibody Construct Expression level (μg/ml) Fab#2(HC)dsHLscFv#3, (LC)dsHLscFv#3 20
Purification of HEK293 Expressed Fab-2×dsscFv #3
(62) The ˜50 ml HEK293 supernatants were concentrated ˜25 fold to ˜2 ml using 30 kDa molecular weight cut off centrifugation concentrators. The concentrated supernatants were purified and further concentrated and buffer exchanged into PBS pH 7.4 using 30 kDa molecular weight cut off centrifugation concentrators.
(63) SDS-PAGE Analysis of Protein-G Purified, HEK293 Expressed Fab-2×dsscFv #3
(64) Samples (2 μg) were diluted with PBS to a volume of 9.75 μl to which 3.75 μl 4×LDS sample buffer and 1.5 μl 100 mM N-ethylmaleimide (non-reduced samples) or 1.5 μl 10× NuPAGE reducing agent (reduced samples) was added. The samples were vortexed, incubated at 70° C. for 10 minutes, cooled and centrifuged at 12500 rpm for 30 seconds. The prepared samples were loaded onto a 4-20% acrylamide Tris/Glycine SDS gel and run for ˜100 minutes at 125V, constant voltage. SeeBluePlus2 (Life Technologies) molecular weight ladder was used. The gels were stained with Instant Blue protein stain (Expedeon) and destained with distilled water. The expected band sizes after reducing and non-reducing SDS-PAGE are indicated in Table 12.
(65) TABLE-US-00014 TABLE 12 Expected band sizes after SDS-PAGE (kDa) (H = heavy chain, L = light chain, +/− reducing agent) −Red +Red Fab-2xdsscFv#3 ~100 H~50 L~50
(66) The non-reducing gel was expected to show a band at ˜100 kDa, whilst the reducing gel was expected to show a doublet at ˜50 kDa with equal staining in both bands.
(67) The reducing SDS-PAGE gels showed banding patterns which indicated that the constructs were being expressed correctly in terms of both migration position and staining intensity with a doublet at ˜50 kDa (
(68) S200 SE-HPLC Analysis of Purified, HEK293 Expressed Fab-2×dsscFv #3
(69) 10 μg purified protein samples (100 μl of 0.1 mg/ml stock diluted in PBS) were injected onto a Superdex 200 10/300 GL Tricorn column (GE Healthcare) 3 days post-purification and developed with an isocratic gradient of PBS pH 7.4 at 1 ml/min, with continuous detection by absorbance at 280 nm. The results are shown in
Example 3
Fab-(HC)dsscFv-(LC)dsFv
(70) Construction of Fab #2-(HC)dsscFv #3-(LC)dsscFv #1 and Fab #2-(HC)dsscFv #3-(LC)dsFv #1 (LC-vL Linked) Plasmids for Expression in Mammalian Cells
(71) The Fab #2 fusion proteins for the expression of Fab #2-(HC)dsscFv #3-(LC)dsscFv #1 and Fab #2-(HC)dsscFv #3-(LC)dsFv #1 (LC-vL linked) (see
(72) TABLE-US-00015 Light#2-(SGGGGSGGGGSGGGGS [SEQ ID NO: 69])- dsscFv#1 [Plasmid b] and Light#2-(SGGGGSGGGGSGGGGS [SEQ ID NO: 69])- dsvL#1 [Plasmid c];
and by fusing HLdsscFv #3 (dsvH-4×G4S-dsvL) to the C-terminus of the, γ1 isotype human gamma-1 CH1 constant region of the heavy chain #2 using the flexible linker SGGGGSGGGGTGGGGS (SEQ ID NO: 70) to generate:
(73) TABLE-US-00016 Heavy#2-(SGGGGSGGGGTGGGGS [SEQ ID NO: 70])- dsscFv#3 [Plasmid a].
(74) Point mutations (heavy chain G44C and light chain G100C) were introduced into the DNA sequences at selected residues in the framework region of vL #3, vH #3, vL #1 and vH #1 to create an interchain disulphide bond between the heavy and light chains of the Fv #3 and Fv #1. Heavy and light chain Fab-fusion genes were cloned into mammalian expression vectors under the control of the HCMV-MIE promoter and SV40E polyA sequence. Genes encoding:
(75) TABLE-US-00017 Heavy#2-(SGGGGSGGGGTGGGGS [SEQ ID NO: 70])- dsscFv#3 [Plasmid a] and dsvH#1 free domain [Plasmid d]
were manufactured chemically and individually cloned into mammalian expression vectors under the control of the HCMV-MIE promoter and SV40E polyA sequence.
(76) The genes encoding dsscFv #1 and dsvL #1 were manufactured chemically and fused to the C-terminal of light chain #2 to create:
(77) TABLE-US-00018 Light#2-(SGGGGSGGGGSGGGGS [SEQ ID NO: 69])- dsscFv#1 [Plasmid b] and Light#2-(SGGGGSGGGGSGGGGS [SEQ ID NO: 69])- dsvL#1 [Plasmid c]
and the entire constructs were cloned into a mammalian expression vector under the control of the HCMV-MIE promoter and SV40E polyA sequence.
HEK293 Expression of:
(78) Fab #2-(HC)dsscFv #3-(LC)dsscFv #1 and
(79) Fab #2-(HC)dsscFv #3-(LC)dsFv #1 (LC-vL linked)
(80) HEK293 cells were transfected with the relevant plasmids using Invitrogen's 293fectin transfection reagent according to the manufacturer's instructions. Plasmids were mixed as follows to express the different constructs as shown in Table 13 below:
(81) TABLE-US-00019 TABLE 13 Antibody Construct Plasmids used Fab#2-(HC)dsscFv#3-(LC)dsscFv#1 1. Plasmid a 2. Plasmid b Fab#2-(HC)dsscFv#3-(LC)dsFv#1 (LC-vL linked) 1. Plasmid a 2. Plasmid c 3. Plasmid d
(82) For the 2 plasmid combination, the ratio of the plasmids used for the transfections was 1:1, whereas for the 3 plasmid combinations the ratio was 1:1:1. A total of 50 μg plasmid DNA was incubated with 125 μl 293fectin+4.25 ml Optimem media for 20 mins at RT. The mixture was then added to 50 ml HEK293 cells in suspension at 1×10.sup.6 cells/ml and incubated with shaking at 37° C. Supernatants were harvested on day 10 by centrifugation at 1500 g to remove cells and the supernatant was passed through a 0.22 μm filter. Expression level was determined by Protein-G HPLC.
(83) The results are shown in Table 14. As can be seen, the levels of expression of both constructs were comparable to each other (16-18 μg/ml). There have been reports in the literature that the expression of Fv regions that lack either a linker between the vL and vH or a dimerisation motif to bring the vL and vH together have substantially lower expression levels than linked Fvs. Surprisingly, this is not observed in this data where there was no significant difference observed between the expression level of each type of construct.
(84) TABLE-US-00020 TABLE 14 Expression level Construct (μg/ml) Fab#2-(HC)dsscFv#3-(LC)dsscFv#1 18 Fab#2-(HC)dsscFv#3-(LC)dsFv#1 (LC-vL linked) 16
Protein-G Purification of HEK293 Expressed Fab #2-(HC)dsscFv #3-(LC)dsscFv #1 and Fab #2-(HC)dsscFv #3-(LC)dsFv #1 (LC-vL Linked)
(85) The ˜50 ml HEK293 supernatants were concentrated ˜25 fold to ˜2 ml using 10 kDa molecular weight cut off centrifugation concentrators. The concentrated supernatants were applied to a 1 ml HiTrap Protein-G FF column (GE Healthcare) equilibrated in 20 mM phosphate, 40 mM NaCl pH 7.4. The column was washed with 20 mM phosphate, 40 mM NaCl pH 7.4 and the bound material was eluted with 0.1M glycine/HCl pH 2.7. The elution peak was collected and pH adjusted to ˜pH 7.0 with 2M Tris/HCl pH 8.5. The pH adjusted elutions were concentrated and buffer exchanged into PBS pH 7.4 using 10 kDa molecular weight cut off centrifugation concentrators.
(86) SDS-PAGE Analysis of Protein-G Purified, HEK293 Expressed Fab #2-(HC)dsscFv #3-(LC) dsscFv #1 and Fab #2-(HC)dsscFv #3-(LC)dsFv #1 (LC-vL Linked)
(87) Samples (2 μg) were diluted with PBS to a volume of 9.75 μl to which 3.75 μl 4×LDS sample buffer and 1.5 μl 100 mM N-ethylmaleimide (non-reduced samples) or 1.5 μl 10× NuPAGE reducing agent (reduced samples) was added. The samples were vortexed, incubated at 70° C. for 10 minutes, cooled and centrifuged at 12500 rpm for 30 seconds. The prepared samples were loaded onto a 4-20% acrylamide Tris/Glycine SDS gel and run for ˜100 minutes at 125V, constant voltage. SeeBluePlus2 (Life Technologies) molecular weight ladder was used. The gels were stained with Instant Blue protein stain (Expedeon) and destained with distilled water.
(88) The results are shown in
(89) The reducing SDS-PAGE gels showed banding patterns which indicated that the constructs were being expressed correctly in terms of both migration position and staining intensity. Non-reducing SDS-PAGE gels showed significant banding patterns >200 kDa for Fab-2×dsscFv that were consistent with multimerisation (
(90) G3000 SE-HPLC Analysis of Protein-G Purified, HEK293 Expressed, Fab #2-(HC)dsscFv #3-(LC)dsscFv #1 and Fab #2-(HC)dsscFv #3-(LC)dsFv #1 (LC-vL Linked)
(91) 10 μg purified protein samples (100 μl of 0.1 mg/ml stock diluted in PBS) were injected onto a TSK Gel G3000SWXL, 7.8×300 mm, column (Tosoh) 3 days post-purification and developed with an isocratic gradient of 200 mM phosphate pH 7.0 at 1 ml/min. Signal detection was by absorbance at 280 nm.
(92) The results are shown in
(93) dsscFv #1 is known to be particularly prone to multimerisation. Given that multimers are physically joined by the dsscFv linker (the vL #1 or vH #1 is paired with vH #1 or vL #1 from a different polypeptide chain), by replacing dsscFv #1 with dsFv #1 (which does not comprise a scFv linker), there was a significant increase in the percentage monomer obtained.
(94) The skilled person will thus appreciate that, depending on the propensity of the dsscFv to multimerise, in some instances it would be advantageous to use a dsFv instead of a dsscFv.
Example 4
Biacore Affinity and Demonstration of Simultaneous Binding of Antigen Targets
(95) The binding affinities and kinetic parameters for the interactions of Fab #2-(HC)dsHLscFv #3 (LC)dsHLscFv #4 were determined by surface plasmon resonance (SPR) conducted on a BIAcore T100 or a BIAcore 3000 using CM5 sensor chips (GE Healthcare Bio-Sciences AB) and HBS-EP (10 mM HEPES (pH 7.4), 150 mM NaCl, 3 mM EDTA, 0.05% v/v surfactant P20) running buffer. All experiments were performed at 25° C. The antibody samples were captured to the sensor chip surface using either a human F(ab′).sub.2-specific goat Fab (Jackson ImmunoResearch) or an in-house generated anti human CH.sub.1 monoclonal antibody. Covalent immobilization of the capture antibody was achieved by standard amine coupling chemistry to a level of 6000-7000 response units (RU). Antigen #2, #3 or #4 was titrated separately over the captured antibody. Each assay cycle consisted of firstly capturing the antibody sample using a 1 min injection, before an association phase consisting of a 3 min injection of antigen, after which dissociation was monitored for 10 min. After each cycle, the capture surface was regenerated with 2×1 min injections of 40 mM HCl followed by 30 s of 5 mM NaOH. The flow rates used were 10 μl min.sup.−1 for capture, 30 μl min.sup.−1 for association and dissociation phases, and 10 μl min.sup.−1 for regeneration. Kinetic parameters were determined by simultaneous global-fitting of the resulting sensorgrams to a standard 1:1 binding model using BIAcore T100 Evaluation software v2.0.1 or BIAcore 3000 BIAEvaluation v3.2. The results are shown in Table 15. The antibody showed expected affinities within the pM-nM range for the antigens tested.
(96) TABLE-US-00021 TABLE 15 Analyte ka (1/Ms) kd (1/s) KD (M) KD (pM) Antigen #2 5.06E+06 3.80E−05 7.51E−12 7.51 Antigen #3 5.54E+06 6.43E−05 1.16E−11 11.6 Antigen #4 7.75E+04 1.35E−04 1.74E−09 1740
(97) The potential for of Fab #2-(HC)dsHLscFv #3 (LC)dsHLscFv #4 to bind simultaneously to all 3 antigens was assessed by capturing the tri-specific antibody to the sensor chip via immobilized anti-human IgG-F(ab′).sub.2. Each antigen or a mixed solution of antigen #2, #3 and #4 was titrated over the captured antibody in 3 min injections. The binding responses observed for the independent injections and combined responses are shown in Table 16. The binding response for the combined antigen #2/#3/#4 solution was equivalent to the sum of the responses of the independent injections. This confirms that of Fab #2-(HC)dsHLscFv #3 (LC)dsHLscFv #4 is capable of simultaneous binding to all 3 antigens tested.
(98) TABLE-US-00022 TABLE 16 Analyte Binding (RU) Antigen #2 58 Antigen #3 38 Antigen #4 47 Antigens #2 + #3 + #4 130 (143)
Example 5
Comparison of Monomeric Yield Between Fab-2×scFv and Fab-2×dsscFv Formats
(99) EXpiHEK cells were transfected with the relevant plasmids by electroporation methods at 50 ml scale. Plasmids were mixed as shown in Table 17 to express the protein. Cultures were grown in ExpiHEK expression medium and incubated at 37° C. with 8% CO.sub.2 at 120 rpm for 16-18 h, prior to addition of enhancer 1 and 2. The cultures were subsequently incubated for a further 4 days at 37° C. Culture supernatants were harvested by centrifugation and 0.22 μm filter sterilized. Expression titres were measured by Protein G HPLC (Table 18).
(100) TABLE-US-00023 TABLE 17 Antibody Construct Plasmids used Fab#4-(LC)HLscFv#5, (HC)HLscFv#6 1. Plasmid k1 2. Plasmid l1 Fab#4-(LC)dsHLscFv#5, (HC)dsHLscFv#6 1. Plasmid k2 2. Plasmid l2 Fab#4-(LC)HLscFv#7, (HC)HLscFv#8 1. Plasmid m1 2. Plasmid n1 Fab#4-(LC)dsHLscFv#7, (HC)dsHLscFv#8 1. Plasmid m2 2. Plasmid n2 Fab#4-(LC) HLscFv#9, (HC)LHscFv#10 1. Plasmid o1 2. Plasmid p1 Fab#4-(LC)dsHLscFv#9, (HC)dsLHscFv#10 1. Plasmid o2 2. Plasmid p2 Fab#4-(LC)HLscFv#7, (HC)LHscFv#10 4. Plasmid q1 5. Plasmid r1 Fab#4-(LC)dsHLscFv#7, (HC)dsLHscFv#10 1. Plasmid q2 2. Plasmid r2
(101) TABLE-US-00024 TABLE 18 Antibody Construct Expression (μg/ml) Fab#4-(LC)HLscFv#5, (HC)HLscFv#6 62 Fab#4-(LC)dsHLscFv#5, (HC)dsHLscFv#6 31 Fab#4-(LC)HLscFv#7, (HC)HLscFv#8 195 Fab#4-(LC)dsHLscFv#7, (HC)dsHLscFv#8 34 Fab#4-(LC) HLscFv#9, (HC)LHscFv#10 298 Fab#4-(LC)dsHLscFv#9, (HC)dsLHscFv#10 55 Fab#4-(LC)HLscFv#7, (HC)LHscFv#10 226 Fab#4-(LC)dsHLscFv#7, (HC)dsLHscFv#10 58
Protein-G Purification of EXPiHEK Expressed Fab-2×scFvFv and Fab-2×dsscFv Formats
(102) Supernatants were concentrated ˜25 fold to ˜2 ml using a 10 kDa molecular weight cut off concentrators. The concentrated supernatants were purified by protein G HPLC using phosphate buffer pH 7.4. The bound material was eluted with 0.1 M glycine/HCl pH 2.7 and pH adjusted to ˜pH 7.0 with 2M Tris/HCl pH 8.5. The eluted material was concentrated using 10 kDa molecular weight concentrators and buffer exchanged into PBS. The purified protein was concentrated to ˜4-5 mg/ml in PBS and stored at 4° C.
(103) SDS-PAGE Analysis of Protein-G Purified, EXPiHEK Expressed Fab-2×scFv and Fab-2×dsscFv Formats
(104) Samples (2 μg) were diluted with PBS to a volume of 9.75 μl to which 3.75 μl 4×LDS sample buffer and 1.5 μl 100 mM N-ethylmaleimide (non-reduced samples) or 1.5 μl 10× NuPAGE reducing agent (reduced samples) was added. The samples were vortexed, incubated at 70° C. for 10 minutes, cooled and centrifuged at 12500 rpm for 30 seconds. The prepared samples were loaded onto a 4-20% acrylamide Tris/Glycine SDS gel and run for ˜100 minutes at 125V, constant voltage. Seeblue2 (Life Technologies) molecular weight ladder was used. The gels were stained with Instant Blue protein stain (Expedeon) and destained with distilled water.
(105) The expected band sizes after reducing and non-reducing SDS-PAGE are indicated in Table 19.
(106) TABLE-US-00025 TABLE 19 Expected band sizes after SDS-PAGE (kDa) (H = heavy chain, L = light chain, +/− reducing agent) −Red +Red −Red +Red Fab-2xscFv ~100 H~50 L~50 Fab-2xdsscFv ~100 H~50 L~50
(107) For all proteins, the non-reducing gel was expected to show a band at ˜100 kDa, whilst the reducing gels were expected to show a doublet at ˜50 kDa with equal staining in the both bands.
(108) For all Fab-2×scFv and Fab-2×dsscFv proteins, the reducing SDS-PAGE gels showed banding patterns which indicated that the constructs expressed correctly in terms of both migration position and staining intensity with a doublet at ˜50 kDa (
(109) G3000 SE-HPLC Analysis of Fab-2×scFv and Fab-2×dsscFv Formats
(110) Purified antibody proteins at ˜5 mg/ml were stored at 4° C. in PBS for 24 h prior to analysis. Samples equivalent to a concentration of 25 μg were injected onto a TSK Gel G3000SWXL, 7.8×300 mm, column (Tosoh) and developed with an isocratic gradient of 200 mM phosphate pH 7.0 at 1 ml/min. Signal detection was by absorbance at 280 nm. The results are shown in Table 20. Analysis indicated that all Fab-2×dsscFv proteins were >90% monomeric. On the other hand, a higher occurrence of multimerisation was detected for all Fab-2×scFv. It is clear that formats that contain scFvs with an Fv disulphide are more monomeric compared to scFvs lacking the Fv disulphide. This indicates a preference for using formats with scFvs that contain Fv disulphides in terms of selecting therapeutic molecules with stable qualities and also attributes that would be advantageous in a manufacturing process.
(111) TABLE-US-00026 TABLE 20 Format % monomer Fab#4-(LC)HLscFv#5, (HC)HLscFv#6 51.3 Fab#4-(LC)dsHLscFv#5, (HC)dsHLscFv#6 96.8 Fab#4-(LC)HLscFv#7, (HC)HLscFv#8 80.8 Fab#4-(LC)dsHLscFv#7, (HC)dsHLscFv#8 98.8 Fab#4-(LC) HLscFv#9, (HC)LHscFv#10 84.9 Fab#4-(LC)dsHLscFv#9, (HC)dsLHscFv#10 94.3 Fab#4-(LC)HLscFv#7, (HC)LHscFv#10 53.2 Fab#4-(LC)dsHLscFv#7, (HC)dsLHscFv#10 90.8