Synthetic Single Domain Library

Abstract

The present invention relates to the identification of a fully humanized single domain antibody scaffold as well as its use in generating synthetic single domain antibodies. The invention further relates to antigen-binding proteins comprising said single domain antibody scaffold and their use in therapy.

Claims

1. A method of making a synthetic single domain antibody library, said method comprising: i. introducing a diversity of nucleic acids encoding CDR1, CDR2, and CDR3, between the respective framework coding regions of a synthetic single domain antibody to generate nucleic acids encoding a diversity of synthetic single domain antibodies with the same synthetic single domain antibody scaffold amino acid sequence, wherein said synthetic single domain antibody scaffold comprises the following amino acid residues: FRW2-V4, FRW2-G11, FRW2-L12, FRW2-W14, FRW1-V5, FRW1-E6, FRW1-L11, FRW3-V35, FRW4-R2, FRW4-L7.

2. The method according to claim 1, wherein said synthetic single domain antibody scaffold further comprises at least one of the following amino acid residues: FRW1-P14, FRW3-S17, FRW3-R29, FRW3-A30, FRW2-S16, FRW3-K18, FRW3-V21, FRW3-Y22, FRW3-L23, FRW3-S27.

3. The method according to claim 1, wherein said synthetic single domain antibody scaffold comprises the following framework regions consisting of FRW1 of SEQ ID NO: I, FRW2 of SEQ ID NO:2, FRW3 of SEQ ID NO: 3 and FRW4 of SEQ ID NO:4, or functional variant framework regions with no more than 1, 2 or 3 conservative amino acid substitutions within each framework region with the proviso that said synthetic single domain antibody scaffold contains at least one of the amino acid residues consisting of FRW2-V5, FRW3-V21 and FRW4-R2.

4. The method according to claim 1, wherein the amino acid residues of the synthetic CDR1 and CDR2 are determined by the following rules: at CDR1 position 1: Y, R, S, T, F, G, A, or D; at CDR1 position 2: Y, S, F, G or T; at CDR1 position 3: Y, S, F, or W; at CDR1 position 4: Y, R, S, T, F, G, A, W, D, E, K or N; at CDR1 position 5: S, T, F, G, A, W, D, E, N, I, H, R, Q, or L; at CDR1 position 6: S, T, Y, D, or E; at CDR1 position 7: S, T, G, A, D, E, N, I, or V; at CDR2 position 1: R, S, F, G, A, W, D, E, or Y; at CDR2 position 2: S, T, F, G, A, W, D, E, N, H, R, Q, L or Y; at CDR2 position 3: S, T, F, G, A, W, D, E, N, H, Q, P; at CDR2 position 4: G, S, T, N, or D; at CDR2 position 5: S, T, F, G, A, Y, D, E, N, I, H, R, Q, L, P, V, W, K or M; at CDR2 position 6: S, T, F, G, A, Y, D, E, N, I, H, R, Q, L, P, V, W, or K; at CDR2 position 7: S, T, F, G, A, Y, D, E, N, I, H, R, Q, L, P, or V; and wherein CDR3 amino acid sequence comprises between 9 and 18 amino acids randomly selected among one or more of the following amino acids: S, T, F, G, A, Y, D, E, N, I, H, R, Q, L, P, V, W, K, M.

5. A synthetic single domain antibody library obtainable by the method of claim 1.

6. The synthetic single domain antibody library of claim 5, comprising at least 1.Math.10.sup.9 distinct antibody coding sequences.

7-8. (canceled)

9. An antigen-binding protein, comprising a synthetic single domain antibody of the following formula: FRW1-CDR1-FRW2-CDR2-FRW3-CDR3-FRW4, the framework regions consisting of FRW1 of SEQ ID NO: I, FRW2 of SEQ ID NO:2, FRW3 of SEQ ID NO: 3 and FRW4 of SEQ ID NO:4, or functional variant framework regions with no more than 1, 2 or 3 conservative amino acid substitutions within each framework region with the proviso that said synthetic single domain antibody scaffold contains at least one of the amino acid residues consisting of FRW2-V5, FRW3-V21 and FRW4-R2; optionally wherein the framework regions are derived from VHH framework regions FRW1, FRW2, FRW3, and FRW4 of Lama species.

10. The antigen-binding protein of claim 9, wherein said synthetic single domain antibody has one or more of the following functional properties: i. it can be expressed as soluble single domain antibody in E. coli periplasm, ii. it can be expressed as soluble intrabodies in E. coli cytosol, iii. it does not aggregate when expressed in mammalian cell lines as fluorescent protein fusions.

11. The antigen-binding protein of claim 9, wherein the amino acid residues of the synthetic CDR1 and CDR2 are: at CDR1 position 1: Y, R, S, T, F, G, A, or D; at CDR1 position 2: Y, S, F, G, or T; at CDR1 position 3: Y, S, F, or W; at CDR1 position 5: S, T, F, G, A, W, D, E, N, I, H, R, Q, or L; at CDR1 position 6: S, T, Y, D, or E; at CDR1 position 7: S, T, G, A, D, E, N, I, or V; at CDR2 position 1: R, S, F, G, A, W, D, E, or Y; at CDR2 position 2: S, T, F, G, A, W, D, E, N, H, R, Q, L or Y; at CDR2 position 3: S, T, F, G, A, W, D, E, N, H, Q, P; at CDR2 position 4: G, S, T, N, or D; at CDR2 position 5: S, T, F, G, A, Y, D, E, N, I, H, R, Q, L, P, V, W, K or M; at CDR2 position 6: S, T, F, G, A, Y, D, E, N, I, H, R, Q, L, P, V, W, or K; at CDR2 position 7: S, T, F, G, A, Y, D, E, N, I, H, R, Q, L, P, or V; and wherein CDR3 amino acid sequence comprises between 9 and 18 amino acids selected among one or more of the following amino acids: S, T, F, G, A, Y, D, E, N, I, H, R, Q, L, P, V, W, K, M.

12. The antigen-binding protein of claim 9, which further comprises an F-box domain for targeting a protein to the proteasome.

13. An isolated nucleic acid that encodes an antigen-binding protein of claim 9.

14. The isolated nucleic acid of claim 13 comprising the following nucleic acid sequences SED ID NO:5, SEQ ID NO:6, SEQ ID NO:7, SEQ ID NO:8 encoding respectively framework regions FRW1, FRW2, FRW3 and FRW4 of SEQ ID NO:1-4.

15. A method of producing the antigen binding protein of claim 6 in a recombinant host cell comprising: i. culturing a host cell comprising a nucleic acid encoding the antigen-binding protein under appropriate conditions for the production of the antigen-binding protein, and ii. isolating said antigen binding protein.

Description

FIGURES LEGENDS

[0137] FIG. 1. (A). Affinity determination of nanobodies to recombinant protein via surface plasmon resonance spectroscopy. Single cycle kinetics analysis was performed on immobilized FGFR4 through covalent amine binding on the dextran based sensor chip. The analytes F8 and mCh were injected in 5 different concentrations followed by a dissociation phase. A final dissociation step was added after the last injection step to determine Koff rates for the KD calculations. The black curves represent the measured data and red curves show the fit analysis (heterogeneous ligand model) performed with the BIAevaluation software.

EXAMPLES

[0138] Validation of the Fully Humanized s2dAb Scaffold

[0139] Validation of the scaffold was done using CDR grafting. CDRs from VHH antibodies were inserted into the fully humanized single domain scaffold as herein described. Antibodies targeting various antigens (GFP, mCherry, alpha-tubulin, MUC18) were inserted and the resulting sdAbs used to check that these fully human sdAbs behave as their parental VHH counterparts in terms of antigen detection, display at the phage surface, expression in the bacteria periplasm, expression in the bacteria cytosol, expression in mammalian cell cytosol. Despite the absence of camelid-specific amino acids thought to be essential for stability and solidity, this showed that this fully human sdAb enable efficient production and stability in reducing environment.

Building a Synthetic Phage Display Library Based on the Design Disclosed Herein

[0140] Several ways exist to build synthetic and diverse library and we used here an oligonucleotide-based approach (provided by Twist Bioscience). Synthetic genes based on the design describe here were ordered and inserted in a modified pHEN2 plasmid bearing 3 myc tags. A library of 1.6 10.sup.9 clones, the Gimli-1 library, was constructed.

Use of Fully Humanized sdAb

[0141] Examples of fully humanized sdAb (resulting from CDR grafting or from selection from the Gimli-1 library) were tested to validate the use of the design disclosed here for various antibody-based applications like immunostaining, inhibition of signal transduction or cell targeting, including CAR-T cell development. Fully human sdAb were used as monomeric soluble forms, displayed on phages, fused to Fc domains or fused to CAR-T scaffolds.

Phage Display Selection of GFP-Specific Nanobodies

[0142] A screening in native conditions was performed using the GFP protein as a target. Four non redundant clones out of 80 analysed detected GFP-Rabb, by immunofluorescence. Importantly, we observed that these antibodies were usable as intrabodies against recombinant GFP expressed in Hela cells (see for example the anti_GFP_Gimli_D8 of SEQ ID NO:10).

Phage Display Selection of Tubulin Nanobodies

[0143] A screening was performed in native condition (Nizak, 2005, see supra) using biotinylated tubulin (Cytoskeleton) as a target. After three rounds of selection, 80 clones were screened at random by immunofluorescence on HeLa cells fixed with methanol. 71 recombinant Ab stained the endogenous tubulin (34 unique sequences) (see for example the anti_tubulin_Gimli_B1 of SEQ ID NO:11).

Phage Display Selection of FGFR4-Specific Nanobodies

[0144] Identification for antibodies targeting the cell surface of cancer cells were exemplified by screening for FGFR4-targeting sdAb. The screening of FGFR4-binding nanobodies was performed using the fully humanized sdAb library Gimli-1. We performed a phage display selection with three rounds of biopanning against recombinant FGFR4. In order to verify the binding specificity for FGFR4, we used FGFR4 knocked-out cells RMS cells (from M. Bernasconi, University of Zurich), and tested 80 phage clones the screening for their binding to Rh4 FGFR4 wildtype cells (Rh4-FR4 wt) and Rh4 FGFR4 knockout cells (Rh4-FR4ko). Flow cytometry analysis revealed 55 phage clones from Gimli-1 library binding to the Rh4-FR4 wt cells only. Sanger sequencing of the 55 phage clones revealed that 28 unique nanobodies from the Gimli-1 library were obtained. Next, phage clones from the Gimli library (i.e. Gimli-1: A4, F8, F11, H2) that showed the best binding to Rh4-FR4 wt by flow cytometry were expressed recombinantly. As negative control, we expressed an anti-mCherry nanobody (mCh). Recombinant nanobodies of approximately 17 kDa were engineered to be expressed with a C-terminal Myc/6×His-tag and an additional cysteine for maleimide coupling. 6×His-tag purification and size exclusion chromatography resulted in proteins of high purity, with yields in the range of 3-16 mg per liter of bacterial culture.

[0145] Selected nanobodies bind to FGFR4-expressing cells Validation of the binding of recombinant nanobodies to cell-surface expressed FGFR4 was performed with Rh4-FR4 wt and Rh4-FRK4ko cells by flow cytometry. A FITC-labeled anti-6×His-tag antibody was used to detect surface-bound nanobodies. Three of the recombinant nanobodies tested displayed no significant binding to Rh4-FR4 wt cells (A4, F11, H2, data not shown) whereas recombinant nanobody F8 (SEQ ID NO:9) showed a specific binding to Rh4-FR4 wt cells and no binding to Rh4-FR4ko cells. As expected, the anti-mCherry negative control nanobody did not bind to Rh4-FR4 wt nor to Rh4-FR4ko cells. Median fluorescence intensities (MFIs) of the the FGFR4 binder incubated with Rh4-FR4 wt cells were in the range of 400, but anti-mCherry negative control, or the anti-6×His-tag antibody only displayed MFI of 200, similar to the binding to Rh4-FR4ko cells.

Nanobodies High Affinity Binding to FGFR4

[0146] To determine the binding affinity of the nanobody to FGFR4, we performed surface plasmon resonance (SPR) spectroscopy with recombinant FGFR4. As already mentioned above, FGFR1 and FGFR2 are expressed on Rh4-FR4ko cells and flow cytometry analysis indicated no binding of the nanobody to the cells. To further confirm FGFR4-specificity, we included also affinity measurements with recombinant FGFR1, FGFR2 and FGFR3. Nanobodies F8, and mCh were injected in five different concentrations on a FGFR coated chip (Suppl. table 1). Except for the negative control mCh, calculated K.sub.D values for FGFR4 binding were in the nano- and picomolar range (FIG. 1; Table 1). Affinity parameters could not be fitted with a 1:1 binding model and best fits were obtained with the heterogeneous ligand model of the BIA evaluation software resulting in two K.sub.D values for each candidate. Measurements of the affinities to the receptor family isoforms FGFR1 and FGFR3 showed as expected no binding of the analytes. The SPR data confirmed the strong binding F8 to FGFR4 and further suggests that F8 has a strict FGFR4 specificity.

TABLE-US-00002 TABLE 2 Surface plasmon resonance spectroscopic determination of nanobody binding affinities to FGFR4. Measured data was fitted with the heterogeneous ligand model and revealed association-and dissociation constants (k.sub.on and k.sub.off) used for calculating affinities in terms of dissociation equilibrium constants K.sub.D (k.sub.off/k.sub.on). The maximal analyte binding signal Rmax is indicated in RU for both determined K.sub.D and resembles their fraction within the amount of total bound nanobodies. Nanobody k.sub.on1(1/M*s) k.sub.off1(1/s) K.sub.D1(M) k.sub.on2(1/M*s) k.sub.off2(1/s) K.sub.D2(M) R.sub.max1(RU) R.sub.max2(RU) F8 5.45E+04 1.04E−06 1.91E−11 1.35E+06 5.57E−03 4.14E−09 83.0 86.4 mCh 2.60E+03 5.11E−03 1.96E−06 2.32E+03 5.05E−03 2.18E−06 20.7 20.7

Materials and Methods

CDR Grafting

[0147] In silico design was done so that CDRs of VHH binding to known targets, for example mCherry (but also GFP, Tubulin or MUC18), were grafted in the scaffold disclosed herein. Synthetic genes were ordered and cloned into pHEN2-derivated plasmid for expression in E. coli and phage display and in fusion to a fluorescent protein for expression in mammalian cytosol.

Soluble Expression in E. coli Periplasm

[0148] Single domain antibody fragments can be subcloned in a pHEN2 derivated bacterial periplasm expression vector and expressed downstream of the pelB secretion sequence. Freshly transform colonies can be grown in Terrific Broth medium supplemented with 1% glucose and 100 μg/ml ampicillin antibiotic until A600=0.6-0.8 was reached. The expression of antibody fragment tagged with 6 His can be then induced with 500 μM isopropyl P-D-thiogalactopyranoside for 16 h at 16° C. or 4 h at 28° C. then span down. After centrifugation, the cell pellets can be incubated in Tris-EDTA-Sucrose osmotic shock buffer and centrifuged again. The cell lysates can be cleared and loaded onto an IMAC resin affinity column for poly Histidine tag. The eluted fractions are dialyzed, and the purity of the protein analyzed typically by SDS-PAGE.

Soluble Expression of Intrabodies in E. coli Cytosol

[0149] Single domain antibody fragments can be subcloned in a bacterial expression vector under the control of a T7 promoter. The plasmid constructs can be transformed into E. coli BL21(DE3) cells. Single colonies can be grown in LB medium supplemented with 1% glucose and 100 μg/ml ampicillin antibiotic until A600=0.6-0.8 was reached. Antibody fragment expression can then be induced with 500 μM isopropyl β-D-thiogalactopyranoside for 16 h at 16° C. and then be span down. After centrifugation, the cell pellets are lysed and centrifuged again. The cell lysates are cleared and loaded typically onto an IMAC resin affinity column for poly Histidine tag. The eluted fraction is dialyzed, and the purity of the protein analyzed typically by SDS-PAGE.

Aggregation Assays in Mammalian Cell Expression System

Functional Expression as Intracellular Antibodies in Eukaryote Cells

[0150] Single domain antibody fragments can be subcloned into a mammalian expression vector in order to express it as a fusion with a fluorescent protein and typically under the control of a CMV promoter. Mammalian cell lines are transfected and fluorescence in the cells is observed 24 h or 48 h after transfection.

Cell Lines

[0151] The cell lines Rh4 (kindly provided by Peter Houghton, Research Institute at Nationwide Children's Hospital, Columbus, Ohio), Rh30, HEK293 ft HEK293T (purchased from ATCC, LGC Promochem) were maintained in DMEM supplemented with 10% PBS (both Sigma-Aldrich), 2 mM L-glutamine and 100 U/ml penicillin/streptomycin (both Thermo Fisher Scientific) at 37° C. in 5% CO.sub.2. RMS cell lines were tested and authenticated by cell line typing analysis (STR profiling) in 2014/2015 and positively matched.sup.48. All cell lines tested negative for mycoplasma.

Phage Display Selection

[0152] Screening for against soluble proteins was performed with biotinylated targets or SBP-tagged targets (e.g. extracellular FGFR4—G&P Biosciences) in native condition (as described in Nizak, C., Moutel, S., Goud, B. & Perez, F. Methods Enzymol. 403, 135-153 (2005)) the herein disclosed single domain antibody library composed of 1.6×10.sup.9 fully humanized hs2dAb. Briefly, biotinylated antigens or SBP-antigen are diluted to obtain a 10-20 nM (1.5 mL final) and confirm efficient recovery on 50 μL streptavidin-coated magnetic beads (Dynal). As a reference, a solution of 10 nM of a 100-kDa protein represents 1 μg protein/mL (hence per round of selection). One can then compare fractions of bound and unbound samples by Western blot using streptavidin-HRP or anti-AviTag antibodies. For screening, the adequate amount of biotinylated antigen coated beads is incubated for 2 h with the phage library (10.sup.13 phages diluted in 1 mL of PBS+0.1% Tween 20+2% non-fat milk) Phages were previously adsorbed on empty streptavidin-coated magnetic beads (to remove nonspecific binders). Phage bound to streptavidin-coated beads are recovered on a magnet. 10 times (round 1) or 20 times (round 2 and 3) washes are carried out using PBS+Tween 0.1% on a magnet. Bound phages are eluted using triethylamine (TEA, 100 mM) and eluted phages are neutralized using 1M Tris pH 7.4. Elution are done twice on beads. Eluted phages are then used to infect E. coli (TG1). Note that usually for round 2 and round 3, only 10.sup.12 phages were used as input.

Protein Expression and Purification

[0153] Periplasmic expression of nanobodies was performed in E. coli MC1061 harboring the pSB_init vector enabling protein production with a C-terminal cysteine and 6×His-tag. A 20 ml overnight pre-culture grown in Terrific Broth medium (25 μg/ml Chloramphenicol) was diluted in 2000 ml fresh medium and grown at 37° C. for 2 h. The temperature was then reduced to 25° C. and after 1 h protein expression was induced with 0.02% L-arabinose. The bacterial culture was grown overnight at 25° C. and cells were harvested by centrifugation (12000 g, 15 min) Periplasmic protein extraction was performed with the osmotic shock method. The cells were resuspended with 50 ml lysis buffer 1 (50 mM Tris/HCl, pH 8.0, 20% sucrose, 0.5 mM EDTA, 5 μg/ml lysozyme, 2 mM DTT) and incubated for 30 min on ice. After the addition of ice-cold lysis buffer 2 (PBS, pH 7.5, 1 mM MgCl.sub.2, 2 mM DTT) the cell debris were harvested by centrifugation (3800 g, 15 min) and the protein containing supernatant was supplemented with a final concentration of 10 mM imidazole. 10 ml of Co.sup.2+-beads slurry (HisPur Cobalt Resin, Thermo Fisher Scientific) were washed with wash buffer (PBS, pH 7.5, 30 mM imidazole, 2 mM DTT) and the supernatant was added to the beads. After an incubation of 1 h at 4° C. the beads were washed with 20 ml wash buffer and bound protein was eluted with 20 ml elution buffer (PBS, pH 7.5, 300 mM imidazole, 2 mM DTT). Prior size exclusion chromatography (SEC), the protein elution was dialyzed overnight into PBS, pH 7.5, 2 mM DTT and concentrated via spin filter centrifugation (Amicon Ultra 15, 3 kDa, Merck Millipore).

Flow Cytometry

[0154] Binding validation of selected phages, recombinant nanobodies was performed on Rh4-FR4 wt and Rh4-FR4ko cells. Specificity of selected phage clones binding to FGFR4 was determined by flow cytometry in 96-well plates (Becton Dickinson). Cell surface staining of Rh4-FR4 wt or Rh4-FR4ko cells was performed on ice in PBS supplemented with 1% FBS. 80 μL phages+20 μL PBS/1% milk were incubated on 1×10.sup.5 cells for 1 h on ice. After 2 washes in PBS, phage binding was detected by a 1:250 dilution of anti-M13 antibody (27-9420-01; GE healthcare) for 1 h on ice followed by a 1:400 dilution of A488-conjugated anti-Mouse antibody (715-545-151; Jackson ImmunoResearch, Europe Ltd) for 45 min. Samples were analyzed after two washes by flow cytometry on a MACSQuant cytometer (Miltenyi) and results were analyzed with FlowJo software (BD Biosciences, France). Phages displaying anti-mCherry nanobodies were used as negative control.sup.24 and as positive control we used an anti-FGFR4 antibody (BT53, kindly provided by J. Khan lab, NCI, Bethesda, Md.). For binding test of recombinant nanobodies, cells were detached with Accutase (Stemcell Technologies) and washed with PBS. All following steps were performed on ice: 4×10.sup.5 cells were incubated with nanobody concentrations of 30 μg/ml for 1 h, washed once with PBS and incubated for an additional 30 min with anti His-tag FITC labeled antibody (LS-057341, LSBioscience, diluted 1:10). The cells were washed once more with PBS and analyzed. The cells were washed twice with PBS and detached with Accutase. All flow cytometry measurements were performed with Fortessa flow cytometer (BD Biosciences) and the data were analyzed using FlowJo™ 10.4.1 software.

Western Blotting

[0155] SDS-PAGE samples were separated on 4-12% NuPAGE Bis-Tris gels (Thermo Fisher Scientific) and blotted on Trans-Blot Turbo Transfer Blot membranes (Biorad). After blocking the membranes with blocking buffer (5% milk/TBST) for 1 h at room temperature, the primary antibody was added at a 1:1000 dilution and incubated overnight at 4°. The secondary HRP-conjugated antibody was diluted 1:10′000 in blocking buffer and added to the washed membrane for 1 h at room temperature. Chemiluminescence was detected after incubation with Amersham™ ECL™ detection reagent (GE Healthcare) or SuperSignal™ West Femto Maximum Sensitivity Substrate (ThermoFisher Scientific) in a ChemiDoc™ Touch Imaging system (BioRad).

Surface Plasmon Resonance Spectroscopy

[0156] Single cycle kinetics analysis was performed with the BIAcore T200 instrument (GE Healthcare) on CMD200M sensor chips (XanTec bioanalytics GmbH) activated with a mixture of 300 mM NHS (N-hydroxysuccinimide) and 50 mM EDC (N-ethyl-N′-(dimethylaminopropyl) carbodiimide). Recombinant FGFR1, FGFR2, FGFR3 and FGFR4 (G&P Biosciences) were immobilized on the activated biosensors (800 to 12′000 RU; 1 RU=1 pg/mm.sup.2) followed by a blocking step with 1M ethanolamine. One flow channel per chip was used as a reference to provide background corrections. The nanobodies were injected at 5 different concentrations followed by a dissociation phase. Koff-rates were determined from a final dissociation step after the last injection. The measurements with FGFR4 were performed for each nanobody on freshly immobilized protein due to strong binding and incomplete dissociation from the surface Immobilization flow rate was 5 μl/min and binding studies were performed at 30 μl/min. Binding parameters were determined with the heterogeneous ligand model fit of the BIAevaluation software. The black curves represent the measured data and red curves show the performed fit analysis.

Sequences of Interest

[0157]

TABLE-US-00003 SEQ ID FRW1 EVQLVESGGGLVQPGGSLRLSCAASG NO: 1 SEQ ID FRW2 MGWVRQAPGKGLEWVSAIS NO: 2 SEQ ID FRW3 YYADSVKGRFTISRDNSKNTVYLQMNSLRAEDTAVYYCA NO: 3 SEQ ID FRW4 YRGQGTLVTVSS NO: 4 SEQ ID FRW1 (NA) gaagtgcagctggtggagtccgggggaggactggtgcagccgggggggtcattgcgac NO: 5 tgagctgcgccgcatccggg SEQ ID FRW2 (NA) atgggctgggttcgtcaggcccctggcaaggggctggagtgggtttccgccatctcc NO: 6 SEQ ID FRW3 (NA) tattacgctgacagcgtaaagggaagatttacaattagccgggataactccaaaaacacgg NO: 7 tctatctccagatgaacagcctcagggccgaggacactgcagtgtattactgtgca SEQ ID FRW4 (NA) tatcgtggacaggggacgctggtaactgtgagtagc NO: 8 SEQ ID Anti-FGFR4 EVQLVESGGGLVQPGGSLRLSCAASGTGYALDDMGWVRQAPGKGLEWVSA NO: 9 Gimli_F8 ISDDESMADYYADSVKGRFTISRDNSKNTVYLQMNSLRAEDTAVYYCASYKEY KYQSGHHYFAYRGQGTLVTVSS SEQ ID anti_GFP_ EVQLVESGGGLVQPGGSLRLSCAASGRFYGWYVMGWV NO: 10 Gimli_D8 RQAPGKGLEWVSAISDQPGTEYYYADSVKGRFTISRDNS KNTVYLQMNSLRAEDTAVYYCAHQKMHYERMYRGQGT LVTVSS SEQ ID anti_tubulin_ EVQLVESGGGLVQPGGSLRLSCAASGFTSERYIMGWVRQ NO: 11 Gimli_B1 APGKGLEWVSAISRRSNYKPYYADSVKGRFTISRDNSKN TVYLQMNSLRAEDTAVYYCALQHRYTQEMDQQHREYR GQGTLVTVSS