TNF MUTEINS AND USES THEREOF
20220387557 · 2022-12-08
Inventors
- Nicola R. SIBSON (Oxford (Oxfordshire), GB)
- Daniel C. ANTHONY (Oxford (Oxfordshire), GB)
- Sandra J. CAMPBELL (Oxford (Oxfordshire), GB)
Cpc classification
A61K39/3955
HUMAN NECESSITIES
A61K31/704
HUMAN NECESSITIES
A61K31/711
HUMAN NECESSITIES
A61K31/506
HUMAN NECESSITIES
A61K31/198
HUMAN NECESSITIES
A61K38/191
HUMAN NECESSITIES
A61P35/00
HUMAN NECESSITIES
International classification
A61K31/198
HUMAN NECESSITIES
A61K31/506
HUMAN NECESSITIES
A61K31/704
HUMAN NECESSITIES
A61K39/395
HUMAN NECESSITIES
A61P35/00
HUMAN NECESSITIES
Abstract
The present invention relates to tumour necrosis factor (TNF) muteins with improved properties, and in particular to TNF muteins which are agonists of, and bind selectively to, tumour necrosis factor receptor 1 (TNFR1). Compositions comprising said TNF muteins, which may additionally comprise appropriate anticancer agents or imaging agents are provided. The use of the muteins of the invention in methods of treating or detecting a tumour are also provided. The invention also provides nucleic acids (e.g. vectors) encoding the TNF muteins and host cells comprising said nucleic acids.
Claims
1. A tumour necrosis factor (TNF) mutein which comprises at least 4 amino acid mutations compared to a wild-type TNF sequence, wherein said mutations comprise: (a) a substitution of the residue at the position equivalent to position 84 of SEQ ID NO. 1; (b) a substitution of the residue at the position equivalent to position 85 of SEQ ID NO. 1; (c) a substitution of the residue at the position equivalent to position 88 of SEQ ID NO. 1; and (d) a substitution of the residue at the position equivalent to position 89 of SEQ ID NO. 1, wherein the mutein is an agonist of tumour necrosis factor receptor 1 (TNFR1) and binds selectively to TNFR1.
2. The mutein of claim 1, wherein the mutein further comprises a substitution of the residue at the position equivalent to position 86 of SEQ ID NO. 1.
3. The mutein of claim 1, wherein the mutein comprises an amino acid sequence from the motif [S/T]-[A/G/S/T/E/H/Q/D/I/L/M]-[S/T]-[Y/V/H]-[S/V/R/N/D/E/I/L/M/T]-[G/D/Y/L/P/E/A/V/I/M/F/W] at positions equivalent to positions 84 to 89 of SEQ ID NO: 1.
4. The mutein of claim 1, wherein the mutein comprises an amino acid sequence from the motif [S/T]-[A/G/S/T/E/H/Q]-[S/T]-[Y/V/H]-[S/V/R/N/D/E]-[G/D/Y/L/P] at positions equivalent to positions 84 to 89 of SEQ ID NO: 1.
5. The mutein of claim 1, wherein the mutein comprises an amino acid sequence from the motif [S/T]-[A/G/S/T/E]-[S/T]-[Y/V]-[S/V/R/N]-[G/D/Y/L/P] at positions equivalent to positions 84 to 89 of SEQ ID NO: 1.
6. The mutein of claim 1, wherein the mutein comprises an amino acid sequence from the motif [S/T]-[G/S/T/E]-T-[Y/V]-[S/V/R/N]-[G/D/Y/L/P] at positions equivalent to positions 84 to 89 of SEQ ID NO: 1.
7. The mutein of claim 1, wherein the mutein comprises an amino acid sequence from the motif [S/T]-[G/S/T]-T-Y-[S/V/R/N]-[D/Y/L/P] at positions equivalent to positions 84 to 89 of SEQ ID NO: 1.
8. The mutein of claim 1, wherein the mutein comprises an amino acid sequence from the motif [S/T]-[G/S]-T-Y-[S/V/N]-[D/Y/P] at positions equivalent to positions 84 to 89 of SEQ ID NO: 1.
9. The mutein of claim 1, wherein the mutein comprises the amino acid sequence selected from any one of SEQ ID NOs: 12-20 at positions equivalent to positions 84 to 89 of SEQ ID NO: 1.
10. The mutein of claim 1, wherein the mutein comprises an amino acid sequence having at least 70% sequence identity to any one of SEQ ID NOs: 2 to 10.
11. The mutein of claim 1, wherein the mutein comprises an amino acid sequence selected from any one of SEQ ID NOs: 2 to 10.
12. A nucleic acid molecule comprising a nucleotide sequence which encodes a TNF mutein of claim 1.
13. A vector comprising the nucleic acid molecule of claim 12.
14. A cell comprising the nucleic acid molecule of claim 12 or comprising a vector comprising the nucleic acid molecule of claim 12.
15. A process for producing or expressing the TNF mutein of claim 1 comprising the steps of: a) transforming or transfecting a host cell with a vector comprising a nucleic acid molecule including a nucleotide sequence which encodes the TNF mutein; b) culturing the host cell under conditions which allow the expression of the TNF mutein; and optionally c) isolating the TNF mutein.
16. A pharmaceutical composition comprising a TNF mutein as defined in claim 1 and one or more pharmaceutically acceptable carriers, diluents and/or excipients, optionally further comprising an anticancer agent or a signal generating agent.
17. The pharmaceutical composition of claim 16, wherein said anticancer agent is selected from a chemotherapeutic agent, an oncolytic virus and an exosome containing a therapeutic nucleic acid molecule, optionally wherein said chemotherapeutic agent is selected from lapatinib, doxorubicin, trastuzumab, melphalan and paclitaxel or said signal generating molecule is selected from a gadolinium-based compound and an iron oxide contrast agent.
18. A TNF mutein of claim 1 or a pharmaceutical composition thereof for use in therapy.
19. A TNF mutein according to claim 1 for use in permeabilising the vasculature of a tumour in a patient for treating, detecting or diagnosing said tumour wherein said TNF mutein is formulated for systemic administration to said patient.
20. A TNF mutein according to claim 1 for use in permeabilising the vasculature of a tumour in a patient for: (i) treating said tumour, wherein said TNF mutein is formulated for systemic administration with an anticancer agent, or is intended for use with an anticancer agent; or (ii) detecting or diagnosing said tumour, wherein said TNF mutein is formulated for systemic administration with a signal generating agent, or is intended for use with a signal generating agent.
21. The TNF mutein for use according to claim 19, wherein the tumour is: (i) a CNS tumour; (ii) a metastasis; and/or (iii) a metastasis in the CNS, liver, bone or breast, preferably in the CNS.
22. The TNF mutein for use according to claim 19, wherein the tumour is less than 20 mm in diameter, preferably less than 5 mm in diameter and/or wherein the patient is a human.
23. The TNF mutein for use according to claim 20, wherein: (i) said anticancer agent is selected from a chemotherapeutic agent, an oncolytic virus and an exosome containing a therapeutic nucleic acid molecule, optionally, wherein said chemotherapeutic agent is selected from lapatinib, doxorubicin, trastuzumab, melphalan and paclitaxel; or (ii) said signal generating molecule is selected from a gadolinium-based compound and an iron oxide contrast agent.
24. A TNF mutein as defined in claim 1 and an anticancer agent selected from the group consisting of a chemotherapeutic agent, an oncolytic virus and an exosome containing a therapeutic nucleic acid molecule, optionally, wherein said chemotherapeutic agent is selected from the group consisting of lapatinib, doxorubicin, trastuzumab, melphalan and paclitaxel as a combined preparation for simultaneous, separate or sequential use in treating a tumour, wherein said TNF mutein is formulated for systemic administration to said patient.
25. A TNF mutein as defined in claim 1 and a signal generating agent selected from the group consisting of a gadolinium-based compound and an iron oxide contrast agent as a combined preparation for simultaneous, separate or sequential use in detecting the presence or absence of a tumour in a patient, wherein said TNF mutein is formulated for systemic administration to said patient.
Description
[0268]
[0269]
[0270]
[0271]
[0272]
[0273]
[0274]
[0275]
[0276]
EXAMPLES
[0277] The inventors designed and implemented a screening process to identify TNF muteins with improved properties as discussed herein. In particular, phage display libraries were designed and selection criteria were developed to determine whether TNF muteins with the desired properties could be obtained.
Example 1: Phage Vector Construction and Testing of hTNF, hTNF R32W S86T and mTNF for Binding to hTNFR1, hTNFR2, mTNFR1 and mTNFR2
[0278] Human (hTNF) and mouse TNF (mTNF) sequences were codon optimised for expression in E. coli and synthesised with flanking restriction enzyme sites for cloning into phagemid vector pANT65 (Abzena). hTNF R32W S86T was generated using Quikchange mutagenesis of hTNF (Agilent Technologies, Santa Clara, USA) prior to cloning into pANT65. All three sequences were cloned into pANT65 using the restriction enzymes Ncol I and Not I, allowing for display of TNF on the phage surface as a gene III fusion protein. This plasmid also contains a C-terminal Flag tag for detection of the expressed protein as well as a His6 tag for purification. The cloned TNF constructs were transformed into E. coli (TG1) and all constructs were confirmed by sequencing.
[0279] Phage were recovered by infecting E. coli with either monovalent M13K07 helper phage (New England BioLabs, Hitchin, UK) or multivalent hyper phage (M13 K07ΔpIII helper phage) (Progen, Heidelberg, Germany). Phage displaying hTNF, hTNF R32W S86T, mTNF and an irrelevant scFv were prepared and tested for binding to Fc tagged hTNFR1, hTNFR2, mTNFR1 and mTNFR2 (Sino Biological, Beijing, P. R. China).
[0280] hTNF phage bound to hTNFR1, hTNFR2 and mTNFR1 but not mTNFR2, although a small degree of binding to mTNFR2 when presented on hyper phage with increased avidity. As expected, hTNF R32W S86T bound to hTNFR1 (although significantly weaker than WT hTNF) and did not show binding to hTNFR2, mTNFR1 or mTNFR2. mTNF showed binding to all four receptors. In all cases, the binding observed was specific as no binding was observed with the irrelevant scFv phage. TNF constructs recovered using hyper phage bound at lower titres than TNF constructs recovered using M13 helper phage as a result of presentation of an increased number of TNF molecules on the phage surface.
Example 2: Library Design and Construction
[0281] Library Design
[0282] Library 1 specifically targeted the A84 to T89 loop which forms part of the receptor binding site. A wobble between the wild type residue Ser and Thr was allowed at position 86 while for all other positions within the loop all possible 20 amino acids were allowed. In addition, a wobble between Arg and Trp was allowed at position 32 to allow inclusion of the wild type Arg amino acid.
[0283] Library 2 targeted two additional areas. The loop containing amino acids L29-R32 were randomised with L29 and R31 permitted to be any amino acid while a wobble between R and W was allowed at position 32. In the region of F144-S147, limited diversity was allowed in position F144 while for A145, E146 and S147 complete randomization to all possible 20 amino acids was allowed. In order to probe both wild type and hTNF S86T sequences, a wobble between wild type Ser and Thr was also allowed at position 86.
[0284] Library Construction
[0285] To construct the designed libraries, PCRs were performed on a hTNF template containing stop codons. The purpose of this step was to reduce the likelihood of hTNF and hTNF R32W S86T being produced and dominating selections (as is occasionally observed during affinity maturation), such that only randomized proteins generated by PCR can form in a phagemid vector. Stop codons were introduced into hTNF cloned into the unrelated plasmid, pJ201, using Quikchange mutagenesis at R32 and S86. Degenerate primers were designed to bind to these regions thus removing the stop codon when amplification occurs.
[0286] For library 1a and 1 b, randomization was carried out in three stages. Initially, using a stop codon template covering amino acid 86, the 3′ portion of the TNF containing amino acids A84-T89 was amplified with the randomized 5′ library primer and a 3′ primer specific for the 3′ end of TNF and containing a Not I restriction site. Next, in two separate PCR reactions the 5′ portion of TNF containing either R32 (WT, library 1a) or W32 (mutein, library 1b) was amplified with a 5′ primer containing Nco I and a 3′ primer that was complementary to a portion of the randomized primer. Lastly, the full-length libraries were constructed by annealing of the amplified 5′ and 3′ fragments and re-amplification with primers that included two restriction sites (Nco I and Not I) for sub cloning of the fragment. During construction, libraries 1a and 1b were kept separate.
[0287] For library 2a and 2b, randomization was also carried in three stages. Initially, using a stop codon template covering amino acid 32, the 5′ portion of TNF containing amino acids L29-R32 was amplified with the randomized 3′ library primer and a 5′ primer specific for the 5′ end of TNF and containing a Nco I restriction site. Next, in two separate PCRs the 3′ portion of TNF containing either S86 (WT, library 2a) or T86 (mutein, library 2b) was amplified with a 5′ primer that was complementary to a portion of the primer used to randomize amino acids L29-R32 and a 3′ primer that introduced randomization to amino acids F143-G146. Lastly, the full-length libraries were constructed by annealing of the amplified 5′ and 3′ fragments and re-amplification with primers that included two restriction sites (Nco I and Not I) for sub cloning of the fragment. During construction, libraries 2a and 2b were kept separate.
[0288] Following amplification, purified DNA for all four libraries was digested using Nco I and Not I and ligated into the similarly cut phagemid vector (pANT65). Ligated DNA was precipitated, resuspended in nuclease-free water, transformed by electroporation into freshly prepared electrocompetent TG1 E. coli cells and plated on LBCG (2% glucose) agar plates. The following day, colonies were counted, plates scraped and glycerol stocks prepared. Libraries were electroporated multiple times to sufficiently cover the theoretical library diversity. The observed library size of each of the libraries is shown in Table 3, and a total coverage of 4.9-fold was obtained. Individual colonies from each of the libraries were sequenced to confirm that the appropriate region had been mutated.
TABLE-US-00003 TABLE 3 Theoretical and observed TNF variant phage library sizes Library Library Library Library 1a 1b 2a 2b Theoretical Library 1.3 × 10.sup.8 5.4 × 10.sup.8 Size (DNA) Theoretical 6.70 × 10.sup.7 6.70 × 10.sup.7 2.7 × 10.sup.8 .sup. 2.7 × 10.sup.8 individual library size Actual library size 1.11 × 10.sup.9 4.95 × 10.sup.8 1.11 × 10.sup.9 5.68 × 10.sup.8 Library coverage 16.4× 7.4× 4.11× .sup. 2.1× Combined 6.7 × 10.sup.8 theoretical library size Total Library 3.27 × 10.sup.9 (4.9×) Coverage
[0289] Bacteria from each of the four libraries were inoculated into 100 ml 2xTYCG cultures using inoculum to cover ≥10× the initial library transformation size. The cultures were grown to mid-log phase (OD600 nm≈z 0.6) and the total number of cells estimated (based on an OD600 nm of 1≈z 5×108 cells/ml).
[0290] Monovalent M13K07 helper phage was added at a multiplicity of infection of 10 and incubated for 1 hour at 37° C., then centrifuged, resuspended in 2xTYCK media and grown overnight at 25° C. The following day, phage were harvested by recovering the culture supernatant by centrifugation followed by precipitation using 3/10th volume of chilled 20% PEG/2.5 M NaCl. After an incubation period of three hours, precipitated phage were recovered by centrifugation and the pellet resuspended in 1× PBS pH 7.4. The resuspended pellet was re-centrifuged to remove any cellular debris, following which the supernatant was re-precipitated using 3/10th volume of chilled 20% PEG/2.5 M NaCl. After a short incubation of 20 minutes to one hour (according to the round of selection), precipitated phage was recovered by centrifugation and the pellet resuspended in 1× PBS pH 7.4, following which a final centrifugation to remove any cellular debris was conducted. The precipitated phage was stored at 4° C. Phage were titred by incubating serial dilutions of phage with log phase TG1 E. coli cells for one hour before plating on LBCG (2% glucose) agar plates. The following day, colonies were counted and titres determined.
[0291] Prior to the first round of selection Libraries 1a and 1b were combined at a ratio 1:1, resulting in Library 1; and Libraries 2a and 2b were combined at a ratio 1:1 to generate Library 2.
Example 3: Affinity Improved Phage Selections
[0292] hTNFR1 was used throughout for positive selections, whereas mTNFR1 was included from round 2 to increase the possibility of selecting for variants with cross reactivity. hTNFR2 was used for deselections at a final concentration of 10 nM to reduce the possibility of selecting variants with hTNFR2 cross reactivity. Deselection against hTNFR2 was performed for the first three rounds of selections. The fourth round of selection was performed solely to enrich for positive binders to hTNFR1 as the round three output titres were relatively large (≈107).
[0293] For the selections, each of the libraries containing a known number of phage was pre-blocked with MPBS following which phage were incubated with decreasing concentrations of hTNFR1 or mTNFR1 antigen for up to three hours in the first round of selection to one hour in the following rounds. Following incubation, Protein A Dyna beads (ThermoFisher, Loughborough, UK) (pre-blocked as above) were added to each selection to capture the Fc-tagged receptors and incubated for 10 minutes. Receptor-phage complexes were washed using increasing numbers of washes with PBST at each successive round of selection followed by two 1× PBS washes, with complexes captured using a magnet between each step. Phage were eluted from the beads by the addition of 50 mM HCl following which the solution was neutralised by the addition of 1 M Tris-HCl pH 9.0. An overview of the selection cascade is shown in
Example 4: High Throughput Screening of Selection Outputs
[0294] As part of the characterisation of selection outputs, it was necessary to establish a high throughput screen. Cultures were tested initially as both phage and as periplasmic extracts, however, during the course of assay optimisation for screening, in contrast to expression as phage, it was observed that the hTNF R32W S86T mutein expressed poorly in the periplasmic extract when compared to WT hTNF, leading to a low signal in the binding ELISA.
[0295] As the mutein TNF R32W S86T did not express well in the periplasm (but was required as an assay comparator), a phage binding ELISA was selected as the primary high throughput screen. Phage produced from individual colonies from different rounds of selection were screened in a single point hTNFR1, hTNFR2 and mTNFR1 receptor binding ELISA. hTNF, hTNF R32W S86T, mTNF and an irrelevant scFv were included on each assay plate for comparison.
[0296] Individual colonies were picked into 500 μl 2xTYCG (2% glucose) media and grown by shaking at 37° C. for 5 hours. Monovalent M13K07 helper phage was added at a multiplicity of infection of 10 and incubated for 1 hour at 37° C. Following this, cultures were centrifuged, resuspended in 2xTYCK media and grown overnight at 25° C. The following day cultures were blocked with an equal volume of 2× MPBS and incubated for 1 hour at room temperature before being pelleted by centrifugation.
[0297] Nunc Immuno MaxiSorp 96 well flat bottom microtiter plates were coated with hTNFR1, mTNR1 and hTNFR2 at 0.5 μg/ml, 100 μl/well overnight. Plates were washed and blocked for 1 hour at room temperature with 3% MPBS following which 100 μl of blocked phage were then added. After incubation with phage, plates were washed with PBST, and binding of phage to hTNFR1, hTNFR2 and mTNFR1 was detected with α-M13-HRP conjugate (diluted 1:5000 in MPBS) (GE Healthcare, Little Chalfont, UK) and TMB substrate (Invitrogen, Loughborough, UK). The reaction was stopped with 1 M HCl, absorbance read at 450 nm on a Dynex Technologies MRX TC II plate reader and the binding data plotted.
[0298] Clones which bound both hTNFR1 and mTNFR1 but not hTNFR2 were classified as hits. 283 hits from the two libraries were identified from ˜1700 phage that were analysed. The library and selection strategy used to obtain the 283 hits is shown in Table 4.
TABLE-US-00004 TABLE 4 Summary of the number of clones screened and the number of hits from both libraries following screening in the phage and periplasmic format (Number of hits/Number of clones screened) # peri hits/ Selection Round # phase hits/# # peris Library 1 2 3 4 phage screened screened 1 hTNFR2- Fc hTNFR2-Fc hTNFR2- Fc hTNFR1-Fc 94/360 24/360 2 deselection (10 nM) (10 nM) selection 0/80 No phage (10 nM) hTNFR1-Fc hTNFR1-Fc (5 pM) ELISA HITS hTNFR1-Fc selection selection screened selection (1 nM) (50 pM) (10 nM) deselection deselection 1 hTNFR2-Fc hTNFR2- Fc 91/248 26/248 2 (10 nM) (10 nM) 10/400 5/400 hTNFR1-Fc mTNFR1-Fc selection selection (1 nM) (10 nM) deselection deselection 1 hTNFR2-Fc hTNFR2- Fc 82/248 17/248 2 (10 nM) (10 nM) 6/408 7/408 mTNFR1-Fc hTNFR1-Fc selection selection (10 nM) (1 nM) deselection deselection Total number of hits 283 80 (Library 1 = 267) (Library 1 = 65) Library 2 = 16 (Library 2 = 15)
Example 5: Screening of TNF Variants
[0299] Periplasmic ELISA Screening of Variants
[0300] To further characterise the clones tested in the primary phage binding screen, a second round of screening was undertaken. As mentioned above, the hTNF R32W S86T mutein expressed poorly when tested as a periplasmic extract. Soluble expression is a major consideration for manufacturing, and so a secondary screen was undertaken to investigate the binding of leads to hTNFR1 when expressed in the periplasm in order to allow identification of leads with potentially improved manufacturability. Periplasmic extracts of colonies from different rounds of selection were screened in a single point assay for their ability to bind hTNFR1. For comparison purposes, hTNF, mTNF, hTNF R32W S86T and irrelevant scFv were included on each assay plate.
[0301] TNF variants were expressed and tested as crude periplasmic extracts. Individual colonies were picked into 1 ml 2xTYCG (0.1% glucose) media and grown by shaking at 37° C. for 5 hours. Cultures were induced by adding IPTG to a final concentration of 1 mM and then grown overnight, with shaking, at 30° C. The following day, cultures were centrifuged and the supernatant discarded. Bacterial pellets were resuspended in TES buffer pH 7.4 and incubated on ice for 30 minutes. The plate was then centrifuged and the protein-containing supernatant transferred to a fresh plate for assay.
[0302] Nunc Immuno MaxiSorp 96 well flat bottom microtiter plates were coated with hTNFR1 at 0.5 μg/ml, 100 μl/well overnight. Plates were washed and blocked for one hour at room temperature with 3% MPBS. Periplasmic extracts were added to the blocked plates (100 μL per well) and incubated for one hour at room temperature. Plates were subsequently washed with PBST and 100 μl per well anti-FLAG antibody (Clone M2, Sigma, Gillingham, UK) diluted 1:10000 in BTBS was added. After one hour incubation, plates were washed with TBST, and the binding of clones to hTNFR1 was detected with an anti-mouse-HRP antibody (Sigma, Gillingham, UK) and TMB substrate (Invitrogen, Loughborough, UK). The reaction was stopped with 1 M HCl, absorbance, read at 450 nm on a Dynex Technologies MRX TC II plate reader and the binding data plotted.
[0303] Greater than ˜1600 clones from both libraries were analysed as part of the manufacturability assessment, with 80 clones identified as hits. In the majority of cases, clones identified in the periplasmic ELISA were also identified in the phage ELISA. Thus, the 80 periplasmic ELISA hits represent a subset of the 283 phage ELISA positive clones. A summary of the rounds of selection as well as the number of clones identifies following the phage ELISA or periplasmic ELISA are shown in Table 4 above.
[0304] Sequencing of Cherry Picked Variants
[0305] Following each round of selection, hits were sequenced. The top 80 clones which gave the highest signal in the hTNFR1 periplasmic binding ELISA were picked into a single 96-well cherry plate (row A to H, column 1 to 10) to allow direct comparison in the same assay together with hTNF, mTNF, hTNF R32W S86T and irrelevant scFv which were picked in triplicate in columns 11 and 12. The sequences of the 80 cherry plate clones were analysed and it was observed that in particular positions, there was a preference for particular amino acids. For example, in Library 1, there was a strong preference for serine or threonine at position 84. Similarly, in Library 2, there was a strong preference for acidic residues (aspartate or glutamate) at position 146.
[0306] Periplasmic Screening of Cherry Picked Variants
[0307] The clones picked into the single cherry picked plate were retested for binding to hTNFR1 using the periplasmic ELISA described above and further characterised for binding to hTNFR2 and mTNFR1 also coated at 0.5 μg/ml.
[0308] The cytotoxic activity of variants on the cherry picked plate was also measured in a cellular cytotoxicity assay using the two mouse cell lines L929 and WEHI164 and the human cell line HEp2. Cytotoxicity was assessed in a homogeneous luminescent cell viability assay in which the number of viable cells in culture is determined based on quantitation of the ATP present (ATP signals the presence of metabolically active cells). The cell lines used are described in Table 5.
TABLE-US-00005 TABLE 5 A summary of the cell lines used TNFR1 TNFR2 Cell Species Source Cat. No. expression expression HEp2 Human Sigma 86030501-1VL Y Y WEHI164 Mouse Sigma 87022501-1VL Y Y L929 Mouse Sigma 85011425-1VL Y N
[0309] As previously, hTNF, hTNF R32W S86R, mTNF as well as the irrelevant scFv were included as controls in the assays as well as a PBS control (representing 0% killing). For each assay, cells were dispensed in a final volume of 50 μl into each well of a 96-well white-walled tissue culture plate (Corning, Amsterdam, NLD). To avoid edge effects, outer wells contained growth media only. After incubation of cells with purified protein for the indicated time the plate was equilibrated at room temperature for 10 minutes and then developed by the addition of 50 μl of Cell TiterGlo® reagent (Promega, Madison, USA) to each well prior to reading in a FluoStar Optima plate reader (BMG Labtech). % killing was calculated using the following equation:
% killing=100−(Luminescence/Untreated Cells Luminescence×100)
[0310] 2×10.sup.5 HEp2 cells were seeded in a total volume of 50 μl of cell growth medium (EMEM, 10% FBS, L-glutamine, NEAAs, Pen/Strep) in a 96 well plate and incubated overnight in a humidified cell culture incubator (37° C., 5% CO.sub.2). The following day media was removed, and the cells sensitized for three hours by the addition of 50 μl of HEp2 assay media (growth medium containing 100 μg/ml cycloheximide (Abcam, Cambridge, UK) and the antibiotic ciprofloxacin (Glentham Life Sciences Ltd., Corsham, UK). A dilution plate was prepared containing periplasmic extracts diluted 1 in 12 in HEp2 assay media, and 50 μl was transferred directly onto the HEp2 cells, following sensitization to give a final volume of 100 μl. Plates were developed using Cell TiterGlo® reagent and read after 48 hours.
[0311] 5×10.sup.4 WEHI164 cells were seeded in 96-well plates in a total volume of 50 μl of WEHI164 cell growth medium (DMEM, 10% FBS, L-glutamine, Pen/Strep) containing 10 μg/ml actinomycin-D and 1:500 of ciprofloxacin. A dilution plate was prepared containing periplasmic extracts diluted 1 in 12 in WEHI164 cell growth medium containing 10 μg/ml actinomycin-D and 1:500 of ciprofloxacin, and 50 μl transferred directly onto the WEHI164 cells to give a final volume of 100 μl. Plates were developed using Cell TiterGlo® reagent and read after 24 hours.
[0312] 2×10.sup.4 L929 cells were seeded in 96-well plates in a total volume of 50 μl of L929 growth media (DMEM, 10% FBS, L-glutamine, Pen/Strep) containing 1:500 of ciprofloxacin. A dilution plate was prepared containing periplasmic extracts diluted 1 in 12 in L929 growth media (DMEM, 10% FBS, L-glutamine, Pen/Strep) containing 1:500 of ciprofloxaci and 50 μl transferred directly onto the L929 cells to give a final volume of 100 μl. Plates were developed using Cell TiterGlo® reagent and read after 48 hours.
[0313] Using the data from the cherry picked plate periplasmic ELISA together with the cellular cytotoxic assay, a top panel of 33 clones was selected.
[0314] All 33 variants bound to hTNFR1 better than hTNF R32W S86T. A large proportion of variants did not bind to hTNFR2, however, a number of variants did bind to hTNFR2 although the signal was lower than that observed for binding by hTNF. A range of binding to mTNFR1 was observed for the variants. No obvious binding to hTNFR1, hTNFR2 and mTNFR1 was observed for the Irrelevant scFv.
[0315] The cellular cytotoxicity data confirmed that hTNF and mTNF showed significant killing in all three cell lines tested. All 33 variants appeared to show killing in the L929 and WEHI164 mouse cell lines as well as the HEp2 human cell lines. For the majority of clones tested and in all three cell lines tested the cytotoxicity observed for the variants was greater than that observed for hTNF R32W S86T. The Irrelevant scFv showed minimal cytotoxicity in the assays.
[0316] The high throughput screening of phage and periplasmic extracts does not take into account differences in expression between the variants and the controls meaning that any differences observed could be as a result of differences in the amounts of each protein present. To further increase the accuracy of the characterisation of the variants with regard to purity and quantitation the 33 leads were purified and tested for hTNFR1, hTNFR2, mTNFR1 and mTNFR2 binding by Biacore.
Example 6: Expression and Purification of Lead TNF Variants
[0317] The initial screening phases described above were performed using variants produced in amber suppressor TG1 cells. These cells are suited for phage display as they allow production of phage as well as soluble protein depending on the conditions used. However, TG1 cells are not typically used for protein expression and purification and as a result an alternative E. coli bacterial strain (BL21(DE3)) was selected for production.
[0318] pANT65 plasmids encoding the lead 33 TNF variants together with the controls hTNF, hTNF R32W S86R and mTNF were co-transformed into Z-competent BL21(DE3) together with the pACYC GroEL/ES plasmid (Abzena) which encodes for expression of the chaperonins GroEL/ES. GroEL/ES has been shown previously to enhance folding of proteins and as a result increase soluble protein production. A pANT65 plasmid encoding an unrelated irrelevant scFv was also co-transformed to be used as a negative control in the cell based assays (37 proteins in total).
[0319] Transformed bacteria were plated out on LB agar plates containing 75 μg/mL carbenicillin and 20 μg/mL chloramphenicol and incubated overnight at 37° C. The next day, a single colony was inoculated in a 50 mL Greiner tube containing 5 mL of 2xTYCC and grown overnight at 30° C. and with shaking at 175 rpm. The following day, 1 ml of the overnight culture was transferred to 50 ml of 2TYCC in a 250 ml Erlenmeyer flask and grown at 30° C. and with shaking at 175 rpm until an OD600 nm of 0.5 was reached. 10 mL of this culture was transferred to 3 L baffled Erlenmeyer flask containing 350 mL of 2TYCC and grown at 30° C. and with shaking at 175 rpm. At an OD600 nm of 0.5, IPTG was added to a final concentration of 0.2 mM and the culture grown overnight at 30° C.
[0320] The next day, the bacteria were harvested by centrifugation for 20 minutes at 3000× g and the resulting pellet resuspended in lysis buffer (20 mM Tris pH 7.4, 200 mM MgCl.sub.2, 5% glycerol). The bacteria were lysed on ice by sonicating four times for 30 seconds using a Misonix XL2020 sonicator following which the bacterial debris was pelleted by centrifugation for 1 hour at 16800× g at 4° C. After centrifugation, the supernatants were transferred to a fresh 50 mL Greiner tube and a 1 M imidazole stock added to a final concentration of 20 mM imidazole. To each tube 0.25 mL of a 50% slurry of Ni-NTA beads (Qiagen, Hilden, Germany) that had been washed 3 times in lysis buffer containing 20 mM imidazole was added. The Greiner tubes were incubated on a rotating wheel at 4° C. for 30 minutes. After rotating, the samples were loaded on 5 mL disposable columns (ThermoFisher, Loughborough, UK). The beads were then washed with ˜50 mL of wash buffer (PBS containing 20 mM imidazole) following which the columns were spun for 30 seconds at 500× g to remove the residual wash buffer. Protein samples were eluted by adding 500 μl of PBS containing 200 mM imidazole, incubating for 2 minutes followed by centrifugation at 500×g for 1 minute. The elution procedure was repeated once more with 250 μl of elution buffer. After elution, fractions containing protein were buffer exchanged into 20 mM Tris pH 9.0. At this stage, it was noted that the proteins following Ni-NTA purification contained a number of impurities and so required additional purification. Proteins were further purified using a 1 mL HiTrap Q HP anion exchange column (GE Healthcare, Little Chalfont, UK) using a linear salt gradient of 0 mM to 500 mM NaCl 20 mM Tris pH 9.0 over 20 column volumes. 0.5 mL fractions were collected, analysed on SDS-PAGE and fractions containing the protein of interest pooled. The protein of interest typically eluted between 350 and 450 mM salt. Fractions were pooled and proteins were buffer exchanged into PBS pH 7.2, filter sterilised and quantified by OD280 nm using an extinction coefficient based on the predicted amino acid sequence. ˜1 μg of each reduced protein was analysed by SDS-PAGE. For each protein purified, a single major band corresponding to the profile of TNF or irrelevant scFv was observed following two rounds of purification. 29 variants expressed satisfactorily to continue analysis, however, the expression of four variants was poor and as a result were not taken forward. The molecular weights and yields of the 10 lead purified variants (following single cycle kinetic analysis described in Example 7) are shown in Table 6.
TABLE-US-00006 TABLE 6 Yields of hTNF, hTNF R32W S86T, and lead hTNF variants following purification by Ni-NTA affinity chromatography and HiTrapQ anion exchange chromatography from a 400 mL starting culture A.sub.280 at Yield Protein M.sub.w(kDa) 1 mg/mL (μg) hTNF 19.7 1.17 114.6 hTNF R32W 19.8 1.44 71.2 S86T mTNF 19.7 1.25 84.3 B2 19.76 1.17 85.2 B4 19.7 1.09 268.7 B5 19.8 1.17 541.4 B6 19.8 1.24 41.2 C4 19.7 1.09 851.9 C8 19.8 1.09 314.3 C9 19.7 1.17 304 F3 19.9 1.16 120.5 F7 19.7 1.09 166.4 G4 19.8 1.16 83.8
[0321] Single Cycle Kinetic Analysis of TNF Variant Binding to hTNFR1, hTNFR2, mTNFR1 and mTNFR2
[0322] The ability of the 29 purified lead muteins to bind to different TNF receptors and compared to hTNF, mTNF and hTNF R32W S86T was determined using single cycle Biacore analysis.
[0323] Single cycle kinetic analysis was performed using a Biacore T200 (serial no. 1909913) running Biacore T200 Evaluation Software V2.0.1 (Uppsala, Sweden). All experiments were run at 25° C. with HBS-P+ running buffer (pH 7.4) (GE Healthcare, Little Chalfont, UK).
[0324] Due to the limit in the number of available flow cells for analysis (three Fc per analysis, Fc2, Fc3 and Fc4, with Fc1 always being used as blank reference for non-specific subtraction), receptor binding was assessed in two separate overlapping runs using either: hTNFR1 (Fc2), hTNFR2 (Fc3) and mTNFR1 (Fc4), or; hTNFR1 (Fc2), mTNFR1 (Fc3) and mTNFR2 (Fc4). Receptors were diluted in running buffer to a concentration of ˜4 μg/mL and at the start of each cycle loaded onto Fc2, Fc3 and Fc4 of a Protein A chip (GE Healthcare, Little Chalfont, UK) at a flow rate of 10 μL/min to give an RU of 40-100. The surface was then allowed to stabilise. Single cycle kinetic data was obtained at a flow rate of 45 μL/min. A three point three-fold dilution range from 1.67 nM to 15 nM TNF, without regeneration between each concentration, was used. The association phase for the three injections of increasing concentrations of TNF was monitored for 240 seconds and a single dissociation phase was measured for 2700 seconds following the last injection of TNF. Regeneration of the Protein A surface was conducted using 10 mM glycine-HCl pH 1.5. The signal from the reference channel Fc1 (no receptor) was subtracted from that of Fc2, Fc3 and Fc4 to correct for differences in non-specific binding to a reference surface.
[0325] Controls were used to ensure the integrity of the receptors during the Biacore run. The relative binding to all four receptors for the three controls and the 29 variants tested is shown in
[0326] As receptor affinity may not translate directly to biological function, 10 variants (G4, B5, B6, F3, F7, B2, B4, C4, C8 and C9, corresponding to SEQ ID NOs: 2-11) were selected for further testing in cellular cytotoxicity assays. These variants all bound with high affinity to hTNFR1 but showed a range of binding kinetics to mTNFR1 and either no or significantly reduced binding to hTNFR2. Table 7 shows affinity values against different receptors.
TABLE-US-00007 TABLE 7 Single-cycle affinity data for binding of the 10 lead variants to hTNFR1, hTNFR2, mTNFR1 and mTNFR2. Affinity (K.sub.D) Analyte hTNFR1 hTNFR2 mTNFR1 mTNFR2 B2 1.64 × 10.sup.−11 3.15 × 10.sup.−10 2.56 × 10.sup.−11 — B4 1.95 × 10.sup.−11 — 1.32 × 10.sup.−10 — B5 1.99 × 10.sup.−11 — 1.28 × 10.sup.−10 — B6 1.69 × 10.sup.−11 — 3.04 × 10.sup.−10 — C4 1.29 × 10.sup.−11 — 3.23 × 10.sup.−11 — C8 1.88 × 10.sup.−10 — 3.63 × 10.sup.−10 — C9 2.48 × 10.sup.−11 +/− 6.07 × 10.sup.−10 — F3 6.35 × 10.sup.−10 — 7.35 × 10.sup.−9 +/− F7 4.24 × 10.sup.−11 — 8.27 × 10.sup.−11 — G4 1.71 × 10.sup.−11 — 4.07 × 10.sup.−11 — +/−, a small amount of potential binding observed
Cellular Cytotoxicity Assays Using Purified Proteins
[0327] The cytotoxic activity of the 10 lead purified proteins was tested in duplicate in a cellular cytotoxicity assay using the two mouse cell lines L929 and WEHI164 as well as the human cell line HEp2 as described above. Purified hTNF, hTNF R32W S86R, mTNF as well as the irrelevant scFv were included as controls in the assays as well as a PBS control (0% killing). For each assay, cells in logarithmic growth phase were dispensed in a final volume of 50 μl into each well of a 96-well white-walled tissue culture plate in the appropriate growth media as described previously. Outer wells contained growth media only in order to avoid edge effects. A six point, four-fold dilution of purified protein (from 100 ng/ml to 0.097 ng/ml final concentration) was prepared in the appropriate growth media and added to the cells. After incubation for the indicated time the plate was equilibrated at room temperature for 10 minutes and then developed by the addition of 50 μl of Cell TiterGlo® reagent to each well prior to reading in a FluoStar Optima plate reader. % killing was calculated using the following equation:
% killing=100−(Luminescence/Untreated Cells Luminescence×100)
[0328] All 10 variants tested showed cytotoxicity in WEHI164, L929 and HEp2 cells. hTNF R32W S86R showed some activity in WEHI164, L929 and HEp2 cells. In most cases all 10 variants performed better than hTNF R32W S86R (Table 8).
TABLE-US-00008 TABLE 8 Comparison of lead variants with hTNF and hTNF R32W S86T in cytotoxicity assays using various cell lines as targets. Assays in which variants performed as well or better than hTNF are highlighted in bold Variant WEHI164 L929 HEp2 B2 ≥ hTNF ≥ hTNF ≥ hTNF* B4 > hTNF R32W S86T > hTNF R32W S86T ≥ hTNF R32W S86T B5 > hTNF R32W S86T > hTNF R32W S86T ≥ hTNF R32W S86T B6 > hTNF R32W S86T > hTNF R32W S86T ≥ hTNF C4 > hTNF R32W S86T > hTNF R32W S86T ≈ hTNF C8 > hTNF R32W S86T > hTNF R32W S86T > hTNF R32W S86T C9 > human TNF RWST ≈ hTNF R32W S86T ≥ hTNF F3 > hTNF R32W S86T ≈ hTNF R32W S86T ≈ hTNF F7 > hTNF R32W S86T > hTNF R32W S86T ≥ hTNF R32W S86T G4 ≥ hTNF ≥ hTNF ≥ hTNF
[0329] Both mTNF and hTNF were more active than hTNF R32W S86R in all three cell types. As expected, the irrelevant scFv showed no activity in any of the assays. Variant G4 (SEQ ID NO: 2) appeared to have similar activity to hTNF in all three cell types. Variant B2 appeared to have a similar activity to hTNF when using the mouse cell lines, WEHI164 and L929, and was more active than hTNF in the human HEp2 cells.
[0330] Multi Cycle Kinetic Analysis of Variant Binding to hTNFR1, hTNFR2, mTNFR1 and mTNFR2
[0331] Multi cycle kinetics analysis of binding to hTNFR1, hTNFR2, mTNFR1 and mTNFR2 was performed on the 10 lead variants using a Biacore T200 (serial no. 1909913) instrument running Biacore T200 Evaluation Software V3.0.1 (Uppsala, Sweden) to characterise further the binding to these receptors.
[0332] As with the single cycle kinetics, receptor binding was assessed in two separate experiments. In the first set of experiments binding to hTNFR1 (Fc2), hTNFR2 (Fc3) and mTNFR1 (Fc4) was assessed. In the second set of experiments the extent of mTNF2 (Fc4) binding was assessed using hTNFR1 (Fc2) and hTNFR2 (Fc3) as controls. All receptors were diluted in running buffer to a concentration of ˜4 μg/mL and at the start of each cycle loaded onto Fc2, Fc3 and Fc4 of a Protein A CM5 chip and captured at a flow rate of 10 μL/min to give an RU of ˜40-100. The surface was then allowed to stabilise. All kinetic data was obtained using a flow rate of 75 μl/min to minimise any potential mass transfer effects. For hTNFR1, hTNFR2 and mTNFR1 kinetic analysis, a six point, two-fold dilution range was selected from 40 to 1.25 nM TNF. The association phase of TNF was monitored for 280 seconds and the dissociation phase was measured for 2100 seconds. Due to lack of binding or short dissociation kinetics observed during single cycle analysis while assessing mTNFR2 binding the association phase of TNF remained at 280 seconds, however the dissociation time was reduced to 300 seconds and the lowest concentration of receptor omitted resulting in a five point, two-fold dilution range from 40 to 2.5 nM TNF when assessing mTNTR2 binding. Regeneration of the Protein A surface was conducted using two injections of 10 mM glycine-HCL pH 1.5 at the end of each cycle. The signal from the reference channel Fc1 was subtracted from that of Fc2, Fc3 and Fc4 to correct for differences in non-specific binding to a reference surface, and a global Rmax parameter was used in the 1-to-1 binding model.
[0333] hTNF was shown to bind with high affinity to hTNFR1, hTNFR2 and mTNFR1, while hTNF R32W S86R showed binding to hTNFR1 but not to hTNFR2 and mTNFR1. Unlike hTNF R32W S86R, all 10 variants showed binding to both hTNFR1 and mTNFR1. Some binding to hTNFR2 was observed for B2 (SEQ ID NO: 7) although the off rate is significantly faster than that for hTNFR1 and mTNFR1. A small amount of potential binding to hTNFR2 was also observed for B4 (SEQ ID NO: 8), C8 (SEQ ID NO: 10) and C9 (SEQ ID NO: 11) although the binding and signal was not robust enough to accurately calculate kinetic data.
[0334] mTNF was shown to bind with high affinity to mTNFR2 while hTNF and hTNF R32W S86R showed very little if any binding to mTNFR2. Binding to hTNFR1 and hTNFR2 which was used to determine the integrity on the TNF proteins was consistent with previously obtained data. Consistent with the data discussed above no binding to mTNFR2 was observed for the 10 variants tested. Marginal binding to hTNFR2 was observed for B2, C8 and C9, consistent with the previous multicycle experiment. The kinetics parameters calculated for the 10 lead variants are shown in Table 9.
TABLE-US-00009 TABLE 9 Multi-cycle affinity data for binding of hTNF, hTNF R32W S86T and the 10 lead variants to hTNFR1, hTNFR2, mTNFR1 and mTNFR2. Analyte hTNFR1 hTNFR2 mTNFR1 mTNFR2 hTNF 1.58 × 10.sup.−11 3.53 × 10.sup.−11 6.67 × 10.sup.−11 — hTNF R32W 3.54 × 10.sup.−9 — — — S86T B2 2.16 × 10.sup.−11 4.07 × 10.sup.−10 1.77 × 10.sup.−10 — B4 1.10 × 10.sup.−10 —* 1.53 × 10.sup.−10 — B5 5.91 × 10.sup.−11 — 2.89 × 10.sup.−10 — B6 3.39 × 10.sup.−10 — 5.90 × 10.sup.−10 — C4 3.06 × 10.sup.−11 — 2.08 × 10.sup.−10 — C8 2.61 × 10.sup.−10 4.13 × 10.sup.−9** 1.12 × 10.sup.−10 — C9 7.17 × 10.sup.−11 2.89 × 10.sup.−7** 1.13 × 10.sup.−9 — F3 4.62 × 10.sup.−10 — 5.08 × 10.sup.−9 — F7 4.65 × 10.sup.−11 — 2.08 × 10.sup.−10 — G4 8.72 × 10.sup.−12 — 1.40 × 10.sup.−10 — *A residual amount of binding above background potentially observed but no KD determined. **A residual amount of binding is observed but 1 to 1 model is a poor fit
Example 8: CD4+ T Cell Epitope Avoidance
[0335] The amino acid sequences of the 10 purified variants were analysed using iTope™ technology for in silico analysis of peptide binding to human MHC class II alleles (Perry et al 2008), and using the TCED™ of known protein sequence-related T cell epitopes (Bryson et al 2010).
[0336] The iTope™ software predicts favourable interactions between amino acid side chains of a peptide and specific binding pockets (in particular pocket positions; p1, p4, p6, p7 and p9) within the open-ended binding grooves of 34 human MHC class II alleles. These alleles represent the most common HLA-DR alleles found world-wide with no weighting attributed to those found most prevalently in any particular ethnic population. Twenty of the alleles contain the ‘open’ p1 configuration and 14 contain the ‘closed’ configuration where glycine at position 83 is replaced by a valine. The location of key binding residues is achieved by the in silico generation of 9mer peptides that overlap by one amino acid spanning the test protein sequence. However, all predictive methods for MHC class II binding inherently over-predict the number of T cell epitopes since they do not allow for other important processes during antigen presentation such as protein/peptide processing, recognition by the T cell receptor or T cell tolerance to the peptide. The TCED™ contains the sequences of all the peptides previously screened in EpiScreen™ T cell epitope mapping assays. The TOED™ is used to search any test sequence against a large (>10,000 peptides) database of peptides derived from unrelated protein and antibody sequences which have been tested in EpiScreen™ T cell epitope mapping assays.
[0337] All potential binding peptides present in common within hTNF have been excluded from the analysis based on the assumption that hTNF is tolerated and thus non-immunogenic. The C-terminus of each TNF protein contains a His6 tag for purification and a FLAG tag for detection. The C-terminus is associated with four iTope™ Promiscuous High Epitopes and one iTope™ Promiscuous Moderate Epitope. The C-terminus has also been excluded from the iTope™ analysis of the lead muteins. iTope™ in silico analysis of the variants is summarised in Table 10.
TABLE-US-00010 TABLE 10 iTope ™ in silico analysis of hTNF, hTNF R32W S86T and the 10 lead variants. Summary of high and moderate affinity MHC class II binding peptides which showed promiscuous binding, based on the assumption that hTNF is tolerated and thus non-immunogenic. iTope ™ iTope ™ Promiscuous Promiscuous Moderate Variant Library High Epitopes Epitopes hTNF — 0 0 hTNF R32W — 0 4 S86T B2 1 0 1 B4 1 1 1 B5 1 0 1 B6 1 0 1 C4 1 1 1 C8 1 0 1 C9 2 0 2 F3 1 1 2 F7 1 1 4 G4 1 0 1
[0338] hTNF R32W S86T contains four promiscuous moderate epitopes when hTNF is set as the non-immunogenic reference. In variants B4, C4, F3 and F7 one promiscuous high epitope has been introduced which is not observed in the hTNF R32W S86T sequence. Variant B4 contains a R6H mutation and variants B5 and F7 contain a R2H mutation which is outside of the regions targeted during library design. Variant F3 contains a I155V mutation which does not introduce any promiscuous epitopes. Variant F7 also contains a A134V change which introduces a promiscuous low epitope. These changes were likely introduced by PCR during library construction.
Example 9: 1 Year Stability Study
[0339] The stability of the top 6 of the lead variants (B2, B5, B6, F3, F7 and G4) was assessed following storage at −20° C. and −80° C. for 3 months and 1 year.
[0340] At day zero (prior to freezing) and following thawing on ice after 3 months and 1 year the following assays were performed: [0341] SDS-PAGE [0342] OD280 nm [0343] Single cycle Biacore for binding to hTNFR1, hTNFR2 and mTNFR1
[0344] Lead variants were expressed and purified by Ni-NTA and SEC and frozen at ˜1 mg/ml in PBS in glass vials baked at 210° C. for one hour (ThermoFisher Scientific, cat. no. 2-CV). 2 vials for each condition (−20° C. and −80° C.) and each time point (3 months and 1 year) were frozen down (Day zero).
[0345] Tables 11 and 12 show that no significant (+/−10%) changes in protein concentration were observed following storage for 12 months at either −20° C. or −80° C. Similarly, SDS-PAGE analysis showed that bands of the expected size were observed and no significant changes in protein integrity were observed following storage for 12 months at either −20° C. or −80° C.
TABLE-US-00011 TABLE 11 Protein concentration of TNF variants measured using OD280 nm at various time points OD.sub.280 −20° C. −80° C. Variant t = 0 12 months 12 months B2 1.08 1.16 1.14 B5 1.13 1.19 1.14 B6 1.23 1.25 1.25 F3 1.14 1.14 1.2 F7 1.02 1.09 1.07 G4 1.42 1.51 1.44
TABLE-US-00012 TABLE 12 Relative protein concentration of TNF variants measured using OD280 nm after 12 months Relative absorbance (t = 0 OD280/t = 12 month OD280) −20° C. −80° C. Variant t = 0 12 months 12 months B2 1.00 0.93 0.95 B5 1.00 0.95 0.99 B6 1.00 0.99 0.98 F3 1.00 1.00 0.95 F7 1.00 0.94 0.95 G4 1.00 0.94 0.99
[0346] A Biacore single cycle kinetic screen (as described above) of the TNF variants demonstrated that binding kinetics prior to freezing are similar to those observed following storage for 12 months at either −20° C. or −80° C. As shown in Table 13, storage under freezing conditions did not alter the binding kinetics with respect to hTNFR2 and there was no evidence to suggest that storage in the tested conditions had a detrimental effect on the integrity or binding of B2, B5, F7 and G4. For B6 there appears to be a small loss of binding to hTNFR1 following storage at −20° C. or −80° C. for a year. For F3 there appears to be a small loss of binding to mTNFR1 following storage at −20° C. or −80° C. for a year.
TABLE-US-00013 TABLE 13 K.sub.d of TNF muteins for various receptors following storage at −20° C. or −80° C. for a year. Storage Variant condition hTNFR1 K.sub.D hTNFR2 K.sub.D mTNFR1 K.sub.D B2 0 1.75 × 10.sup.−11 1.98 × 10.sup.−10 6.24 × 10.sup.−11 1 year −20° C. 3.90 × 10.sup.−11 2.10 × 10.sup.−10 8.51 × 10.sup.−11 1 year −80° C. 4.07 × 10.sup.−11 3.49 × 10.sup.−10 8.64 × 10.sup.−11 B5 0 3.55 × 10.sup.−11 — 7.71 × 10.sup.−11 1 year −20° C. 4.93 × 10.sup.−11 — 1.15 × 10.sup.−10 1 year −80° C. 4.46 × 10.sup.−11 9.77 × 10.sup.−11 B6 0 9.69 × 10.sup.−11 — 6.70 × 10.sup.−11 1 year −20° C. 5.26 × 10.sup.−11 — 7.11 × 10.sup.−10 1 year −80° C. 6.67 × 10.sup.−11 4.74 × 10.sup.−11 F3 0 1.20 × 10.sup.−11 — 1.10 × 10.sup.−10 1 year −20° C. 2.71 × 10.sup.−11 — 5.27 × 10.sup.−10 1 year −80° C. 2.30 × 10.sup.−11 3.98 × 10.sup.−10 F7 0 2.20 × 10.sup.−11 — 1.16 × 10.sup.−10 1 year −20° C. 2.62 × 10.sup.−11 — 1.11 × 10.sup.−10 1 year −80° C. 3.53 × 10.sup.−11 1.65 × 10.sup.−10 G4 0 1.73 × 10.sup.−11 — 7.00 × 10.sup.−11 1 year −20° C. 3.31 × 10.sup.−11 — 6.92 × 10.sup.−11 1 year −80° C. 3.99 × 10.sup.−11 4.58 × 10.sup.−11
Example 10: Assessment of TNF Mutein Half-Life in Plasma
[0347] The amount of TNF or TNF mutein detectable in plasma was measured over a period of 60 minutes. The half-life for hTNF was determined to be 5.6 minutes. The TNF muteins showed half-lives within the range of the wild-type hTNF, with G4 (SEQ ID NO: 2) having the closest half-life to wild-type. The results are set out in Table 14.
TABLE-US-00014 TABLE 14 Half-life of hTNF and lead muteins in plasma Protein Half-life in plasma (mins) hTNF 5.6 B2 4.1 B5 20.5 B6 5.2 F3 7.25 F7 10 G4 5.8
Example 11: Safety Assessment of Mutein G4 (SEQ ID NO: 2) on Administration to Mice
[0348] The heart, lung, liver, brain, spleen and left kidney were examined from 16 mice from four groups—treated with saline, mouse TNF, human TNF and Mutein G4.
[0349] Histological slides were prepared and stained with H&E. Findings in the tissue were scored using a non-linear semi-quantitative grading system from 0 to 5 where 0=no significant change and 5=whole organ or tissue affected. For some lesions grading is not appropriate and they are scored P for present if seen.
[0350] There were no clear differences between groups. Two of the saline treated animals had slightly less extramedullary haematopoiesis (EMH) in the spleen which reflected in the slightly lower group mean spleen weight. However, the appearance of the spleen and amount of EMH is variable and all animals fell within the range of appearance expected in most mice strains. The spleens show no morphological changes which would be expected in immune-deficient animals despite the reduced WBC count seen in the haematology profiles.
Example 12: Dose Response and Window of Permeabilisation Studies of TNF Muteins (mutTNF) in a Brain Metastasis Model
[0351] Materials and Methods
[0352] All experiments were performed in accordance with the ARRIVE Guidelines and Guidelines for the Welfare and Use of Animals in Cancer Research. For in vivo experiments, 4T1-GFP cells were cultured as described previously (Connell et al. J Natl Cancer Inst. 2013; 105(21):1634-1643; Serres et al., Proc Natl Acad Sci USA. 2012; 109(17):6674-6679; and Soto et al., Neuro Oncol. 2014; 16(4):540-551) Female BALB/c mice (Charles River Kent, UK) were anaesthetised with isoflurane and injected via the left cardiac ventricle, under ultrasound guidance (Vevo 3100 Imaging System; Fujifilm VisualSonics), with 1×105 4T1-GFP cells in 100 μl of sterile phosphate buffered saline.
[0353] To determine the dose-response of permeabilisation, at 13 days post-metastasis induction, mice were injected intravenously with 5, 16.7, 50 or 150 μg/kg of the G4 mutein (mutTNF) in 100 μl saline, or saline alone (n=3-4 per group). Mice were perfusion-fixed 2 h later, and BBB permeability assessed using either horseradish peroxidase (HRP) histochemistry, or IgG immunohistochemistry (150 μg/kg mutTNF dose omitted). For HRP histochemistry, mice were injected intravenously with 100 μl type II HRP (300 units; SigmaAldrich, Dorset, UK) 30 min prior to transcardial perfusion-fixation with Karnovsky's fixative. For IgG assessment of BBB permeabilisation, mice were transcardially perfusion-fixed with PLPlight fixative.
[0354] To determine the window of permeabilisation, at 10 days post-metastasis induction, mice were injected via a tail vein with 50 μg/kg mutTNF G4 in 100 μl saline, or saline only, and perfusion-fixed, as above, for HRP histology 2, 12, 24 or 72 h later (n=6 per group).
[0355] HRP and IgG Histochemistry to Assess BBB Permeability
[0356] For HRP assessment of BBB permeability, alternate sections were stained with Hanker-Yates, as described previously, (Connell et al., supra) and cresyl violet (SigmaAldrich, Dorset, UK). Each metastasis >50 μm in diameter on cresyl violet sections was assigned as either positive or negative for HRP staining on the adjacent Hanker-Yates stained sections. For IgG assessment of permeabilisation, all brain sections were immunostained for IgG, and each metastasis counted as either positive or negative for IgG staining.
[0357] Magnetic Resonance Imaging Assessment of BBB Permeability
[0358] BALB/c mice injected intracardially with 4T1-GFP cells underwent magnetic resonance imaging (MRI) 13 days after metastasis induction, as described previously (Connell et al., supra). Pre- and post-Gd-DTPA T1-weighted images were acquired before and 2 h after mutTNF G4 injection (5, 16.7 or 50 ug/kg) or saline alone (n=6 per group). Pre-Gd-DTPA images were subtracted from post-Gd-DTPA images and a mask created for all voxels with signal intensity greater than mean+2SD for normal brain. This mask was applied to the post-Gd images, which were then cross-referenced with the corresponding histological sections to confirm metastasis presence and percentage signal change in metastasis-containing ROIs calculated. For full details of image acquisition and analysis.
[0359] Results
[0360] Dose Response and Window of BBB Permeabilisation
[0361] HRP-positive metastases, indicating BBB breakdown, were evident in all mice treated with the mutTNF G4, with the number increasing in a dose-dependent manner. No breakdown was evident in non-tumour bearing normal brain tissue. Some natural BBB breakdown was evident at metastatic sites in the saline group, as expected at this time-point (day 13), particularly for brain metastases >400 μm diameter. Nonetheless, the numbers of HRP-positive metastases as a percentage of total metastases (47.8±3.3%, 55.1±8.0% and 64.2±7.6%) at the 3 highest doses of mutTNF G4 (16.7, 50 and 150 μg/kg, respectively) were significantly greater than in the saline group (30.9±4.8%;
[0362] The above findings were confirmed through staining for endogenous serum IgG, which is normally excluded from the brain by an intact BBB. In this case, the percentages of IgG-positive metastases (57.0±8.1%, 50.9±4.9% and 41.2±1.4%; at 5, 16.7 and 50 μg/kg, respectively) were significantly greater than in the saline group (26.6±2.3%) for all mutTNF doses (
[0363] For the window study, although BBB breakdown was evident at all time-points after mutTNF administration, the percentage of HRP-positive metastases decreased over time. Thus, significant breakdown was still evident in mutTNF mice compared to saline treated mice at 24 h, but not 72 h, post-treatment (
[0364] Gd-DTPA Enhancement in Brain Metastases
[0365] Extravasation and accumulation of Gd-DTPA enables BBB breakdown to be detected in vivo as hyperintense areas on T1-weighted MR images. Post-Gd-DTPA T1-weighted images indicated a small number of metastases exhibiting natural BBB breakdown prior to mutTNF treatment, and these were excluded from the analysis.
[0366] At 2 h after mutTNF treatment, areas of hyperintensity were evident on post-Gd-DTPA images compared to pre-Gd-DTPA images, which corresponded spatially with sites of metastases. The majority of animals in the saline group did not show Gd-DTPA contrast enhancement; where occasional hyperintensities were evident, these showed lower contrast and a more spatially restricted profile than in the mutTNF-treated animals. This low level of breakdown likely reflects natural breakdown at metastatic sites that was missed in the pre-treatment scans owing to their lower resolution. The ratio of signal intensity at sites shown to contain metastases vs. equivalent non-tumour bearing regions was significantly greater at all mutTNF concentrations compared to the saline-injected group (
Example 13: BBB Permeabilisation of Endothelial Cell Monolayers
[0367] Materials and Methods
[0368] Paracellular Permeability Assay in hCMEC/D3 Cells
[0369] Human brain microvascular endothelial hCMEC/D3 cells were cultured as described previously (Lopez-Ramirez et al. J Immunol. 2016; 189(19):3130-3139). Permeabilising activity of the 10 lead mutTNFs on monolayers of hCMEC/D3 cells was assessed using a FITC-labeled 70 kDa dextran tracer (SigmaAldrich, Dorset, UK) assay. Confluent monolayers on permeable polyester transwell inserts were treated with mutTNF, wild-type hTNF or vehicle for 24 h. Subsequently, paracellular flux of FITC-dextran across the insert was evaluated and permeability coefficients calculated.
[0370] Results
[0371] All mutTNFs induced significant permeability in vitro (