METHOD FOR RESOLVING COMPLEX, MULTISTEP ANTIBODY INTERACTIONS

Abstract

Herein is reported a method for determining antibody-FcRn-interaction comprising the steps of immobilizing FcRn on a solid surface, which is suitable for surface plasmon resonance measurement, individually applying to the solid surface obtained in step a) solutions comprising the antibody at different concentrations and determining the association rate constant and the dissociation rate constant for each concentration, and determining with the rates obtained in step b) the KD-value of the antibody-FcRn-interaction, wherein the immobilized FcRn is monomeric FcRn, the monomeric FcRn is immobilized using functional (capture) groups that are directly attached to said solid surface, the solid surface is free of branched glucan, and the immobilization is at a pH value of from pH 7 to pH 8.

Claims

1. A method for determining antibody-FcRn-interaction comprising the steps of a) immobilizing FcRn on a solid surface, which is suitable for surface plasmon resonance measurement, b) individually applying to the solid surface obtained in step a) solutions comprising the antibody at different concentrations and determining the association rate constant and the dissociation rate constant for each concentration, c) determining with the rates obtained in step b) the K.sub.D-value of the antibody-FcRn-interaction, wherein the immobilized FcRn is monomeric FcRn, wherein the monomeric FcRn is immobilized using functional (capture) groups that are directly attached to said solid surface, wherein the solid surface is free of branched glucan, and wherein the immobilization of the FcRn is at a pH value of from pH 7 to pH 8.

2. The method according to claim 1, wherein the immobilization is at a pH value of about pH 7.4.

3. The method according to any one of claims 1 to 2, wherein the FcRn is immobilized at a density of 50-150 RU.

4. The method according to any one of claims 1 to 3, wherein the FcRn is a single chain FcRn (scFcRn).

5. The method according to claim 4, wherein the scFcRn is a fusion polypeptide of beta-2-microglobulin and human FcRn fusion polypeptide, which are conjugated to each other by a (GGGGS).sub.4-peptidic linker, and which comprises a C-terminal Avi-tag.

6. The method according to any one of claims 1 to 5, wherein the FcRn is immobilized using amine coupling or biotin/streptavidin coupling.

7. The method according to any one of claims 1 to 6, wherein the FcRn is immobilized at a density of about 50-150 pg/mm.sup.2 chip surface.

8. The method according to any one of claims 1 to 7, wherein the immobilization is with a solution comprising FcRn at a concentration of about 250 g/ml in 10 mM HEPES buffer at a pH value of pH 7.4.

9. The method according to any one of claims 1 to 8, wherein the solution of the antibody applied to the immobilized FcRn in step b) comprises 150 mM NaCl or 400 mM NaCl or 400 mM NaCl and 20% (w/w) ethylene glycol.

10. The method according to claim 9, wherein the solution of the antibody applied to the immobilized FcRn in step b) comprises either 10 mM MES, 150 or 400 mM NaCl, 0.05% P-20 and optionally 20% (w/w) ethylene glycol at pH value of pH 5.8, or comprises 10 mM HEPES, 150 mM or 400 mM NaCl, 0.05% P- and optionally 20% (w/w) ethylene glycol at a pH value of pH 7.4.

11. The method according to any one of claims 1 to 10, wherein the branched glucan is dextran.

12. The method according to any one of claims 1 to 11, wherein the Fab-FcRn interaction as well as the Fc-region-FcRn interaction are divided and visualized using a 2-/3-dimensional diagram, in which the stability (log kd, off-rate) is shown/corresponds to the x-axis and the recognition (log ka, on-rate) is shown/corresponds to the y-axis.

Description

DESCRIPTION OF THE FIGURES

[0228] FIG. 1 SPR sensorgram of different humanized/chimeric human IgG1 Fc-region comprising antibodies (exploratory and approved); sensorgrams were recorded under the same SPR conditions with the same concentration and same monomer concentration; the only difference is the antigen binding site.

[0229] FIG. 2 Three-dimensional diagram (with the stability (log kd) on the x-axis, the recognition (log ka) on the y-axis and intensity on the z-axis) of an isolated, Fab-less Fc-region-FcRn-interaction, i.e. a theoretical 1:1 interaction of an isolated antibody Fc-region with a single FcRn-binding site obtained by introducing the mutations I253A/H310A/H435A (numbering according to Kabat is used herein) in one Fc-region polypeptide.

[0230] FIG. 3 Two-dimensional diagram of the interaction of a full-length, monospecific anti-digoxygenin antibody with FcRn on an SPR solid surface.

[0231] FIG. 4 Sensorgram of a simple Fc-region-FcRn-interaction, i.e. a 1:1 interaction of an isolated, Fab-less antibody Fc-region with a single FcRn-binding site obtained by introducing the mutations I253A/H310A/H435A in one Fc-region polypeptide and maintaining the corresponding wild-type Fc-region polypeptide as the respective other Fc-region polypeptide at low FcRn immobilization levels with amine coupling.

[0232] FIG. 5 Two-dimensional diagram (with the stability (log kd) on the x-axis and the recognition (log ka) on the y-axis) of a simple Fc-region-FcRn-interaction, i.e. a 1:1 interaction of an isolated, Fab-less antibody Fc-region with a single FcRn-binding site obtained by introducing the mutations I253A/H310A/H435A in one Fc-region polypeptide and maintaining the corresponding wild-type Fc-region polypeptide as the respective other Fc-region polypeptide at low FcRn immobilization levels with amine coupling.

[0233] FIG. 6 Sensorgram of a complex IgG1 full length antibody-FcRn-interaction.

[0234] FIG. 7 Two-dimensional diagrams of a complex IgG1 full length antibody-FcRn-interaction at low FcRn immobilization levels using amine coupling.

[0235] FIG. 8 Two-dimensional diagrams of a complex IgG1 full length antibody-FcRn-interaction at high FcRn immobilization levels using biotin/avidin coupling.

[0236] FIG. 9 Effect of isolated Fab-FcRn-interaction of an anti-digoxygenin-Fabs added to the Fc-region with a single FcRn binding site (mutations I253A/H310A/H435A in one Fc-region polypeptide and wild-type in the respective other Fc-region polypeptide) determined at 150 mM sodium chloride.

[0237] FIG. 10 Effect of isolated Fab-FcRn-interaction of an anti-digoxygenin-Fabs added to the Fc-region with a single FcRn binding site (mutations I253A/H310A/H435A in one Fc-region polypeptide and wild-type in the respective other Fc-region polypeptide) determined at 400 mM sodium chloride.

[0238] FIG. 11 Effect of isolated Fab-FcRn-interaction of an anti-digoxygenin-Fabs added to the Fc-region with a single FcRn binding site (mutations I253A/H310A/H435A in one Fc-region polypeptide and wild-type in the respective other Fc-region polypeptide) determined at 400 mM sodium chloride and 20% (w/w) ethylene glycol (MW=62.07 g/mol).

[0239] FIG. 12 Effect of isolated Fab-FcRn-interaction of an anti-digoxygenin-Fabs added to a wild-type IgG1 Fc-region determined at 150 mM sodium chloride.

[0240] FIG. 13 Effect of isolated Fab-FcRn-interaction of an anti-digoxygenin-Fabs added to a wild-type IgG1 Fc-region determined at 400 mM sodium chloride.

[0241] FIG. 14 Effect of isolated Fab-FcRn-interaction of an anti-digoxygenin-Fabs added to a wild-type IgG1 Fc-region determined at 400 mM sodium chloride and 20% (w/w) ethylene glycol (MW=62.07 g/mol).

[0242] FIG. 15 Sketch depicting the different intramolecular Fab-FcRn and Fc-region-FcRn interactions.

[0243] FIG. 16 Sketch depicting intramolecular and intermolecular antibody FcRn-interactions.

[0244] FIG. 17 Sensorgram with separated interrelation of Fc- and Fab-FcRn binding strength can be separated.

[0245] FIG. 18 Two-dimensional diagram with separated interrelation of Fc- and Fab-FcRn binding strength.

[0246] FIG. 19 Biotin/avidin coupling density for immobilizing (sc)FcRn to the surface of the SPR chip.

[0247] FIG. 20 Amine coupling for immobilizing (sc)FcRn to the surface of the SPR chip for controlling the coating density down to low levels.

[0248] FIG. 21 Two-dimensional diagram showing the Fc- and Fab-FcRn binding strength of an anti-digoxygenin antibody with wild-type IgG1 Fc-region using a chip with about 1700 RU of (sc)FcRn captured by biotin/avidin coupling.

[0249] FIG. 22 Two-dimensional diagram showing the Fc- and Fab-FcRn binding strength of an anti-digoxygenin antibody with IgG1 Fc-region with symmetric M252Y/S254T/T256E mutations with about 1700 RU of (sc)FcRn captured by biotin/avidin coupling.

[0250] FIG. 23 Two-dimensional diagram showing the Fc- and Fab-FcRn binding strength of an anti-digoxygenin antibody with wild-type IgG1 Fc-region using a chip with about 80 RU of (sc)FcRn captured by biotin/avidin coupling.

[0251] FIG. 24 Two-dimensional diagram showing the Fc- and Fab-FcRn binding strength of an anti-digoxygenin antibody with IgG1 Fc-region with symmetric M252Y/S254T/T256E mutations with about 80 RU of (sc)FcRn captured by biotin/avidin coupling.

[0252] FIG. 25 Two-dimensional diagram showing the Fc- and Fab-FcRn binding strength of an anti-digoxygenin antibody with wild-type IgG1 Fc-region using a chip with about 80 RU of (sc)FcRn captured by amine coupling.

[0253] FIG. 26 Two-dimensional diagram showing the Fc- and Fab-FcRn binding strength of an anti-digoxygenin antibody with IgG1 Fc-region with symmetric M252Y/S254T/T256E mutations with about 80 RU of (sc)FcRn captured by amine coupling.

[0254] FIG. 27 Two-dimensional diagrams of the Fab-FcRn and Fc-region-FcRn interaction of the parent and five Fab-charge variants of the same parental anti-CD44 antibody.

[0255] FIG. 28 Two-dimensional diagrams of the Fab-FcRn and Fc-region-FcRn interaction of four Fc-region variants of the same parent anti-CD44 antibody (upper left diagram) are shown.

[0256] FIG. 29 Monitoring the effect of the introduction of a M252Y/S254T/T256E mutation, i.e. a single modification, in the antibody on the overall antibody-FcRn-interaction using a one-armed Fab-Fc-region fusion.

[0257] FIG. 30 Monitoring the effect of the introduction of a V308P/Y436H mutation, i.e. a single modification, in the antibody on the overall antibody-FcRn-interaction using a one-armed Fab-Fc-region fusion.

[0258] FIG. 31 Monitoring the effect of the introduction of an I253A/H310A/H435A mutation, i.e. a single modification, and the removal of the Fab (vs. mAb-2 on FIG. 29) in the antibody on the overall antibody-FcRn-interaction using a one-armed Fab-Fc-region fusion.

[0259] FIG. 32 Monitoring the effect of the introduction of a T307Q/N434A and further of a V308P/Y436H mutation, i.e. two single modifications, in the antibody on the overall antibody-FcRn-interaction using a one-armed Fab-Fc-region fusion.

[0260] FIG. 33 Delinearization of hydrophobic and charge driven antibody-FcRn interactions.

[0261] FIG. 34 Analysis of the different antibody-FcRn-interactions.

[0262] FIG. 35 Two-dimensional scheme showing the determination/weakening of the intramolecular avidity by the increase/addition of salt and the determination/weakening of the intermolecular avidity by solution density.

[0263] FIG. 36 Interaction spot pattern showing that matching the parental antibody interaction spot pattern is preferred over a shifted spot pattern for pharmacokinetic engineering of an antibody, e.g. when introducing the YTE mutations as the antibody might have a modified thermostability

[0264] FIG. 37 pH-dependent 2-dimensional diagram of the interaction of Ustekinumab with YTE mutation with FcRn.

[0265] FIG. 38 pH-dependent 2-dimensional diagram of the interaction of Briakinumab with FcRn.

[0266] FIG. 39 pH-dependent 2-dimensional diagram of the interaction of Briakinumab with YTE mutation with FcRn

[0267] FIG. 40 Light chain amino acid sequence alignment of Ustekinumab and Briakinumab; CDRs are marked with an asterisk (*).

[0268] FIG. 41 Heavy chain amino acid sequence alignment of Ustekinumab and Briakinumab; CDRs are marked with an asterisk (*).

[0269] FIG. 42 Sketches of the antibodies used in the examples.

MATERIALS

[0270] Information According to the Manufacturer:

[0271] Sensor chip C1 has a flat carboxymethylated surface. Provides the same functionality as Sensor chip CM5 but has no dextran matrix (the carboxyl groups are attached directly to the surface layer). The absence of a surface matrix makes Sensor chip C1 less hydrophilic than Sensor chip CM5. Experimental protocols follow the same principles for Sensor Chip C1 and Sensor chip CM5. The absence of a surface matrix will result in an immobilization yield that is approximately 10% of that obtained on Sensor chip CM5 under comparable conditions.

[0272] Amine coupling makes use of the N-terminus and c-amino groups of lysine residues of the ligand.

[0273] Immobilization Procedures

[0274] In general the immobilization procedure consists of three distinct parts: [0275] Activation: the priming of the sensor chip so it can form a covalent bond with another molecule [0276] Coupling: the injection of ligand so it forms covalent bonds with the sensor surface [0277] Deactivation: injection of a low molecular reactive group to quench the remaining active surface groups

[0278] Activation:

[0279] In case of a covalent amine binding chemistry on the dextran-based sensor chips, the carboxyl groups are activated with a mixture of NHS (N-hydroxy succinimide) and EDC (1-Ethyl-3-(3-dimethylaminopropyl)carbodiimide) to create N-Hydroxy-succinimide esters. By varying the activation time, more or fewer carboxyl groups are activated. In addition, the concentration of the NHS/EDC mixture can be varied to control the quantity of activated carboxyl groups. The quantity of activated groups determines how much ligand can bind to the sensor surface. The standard activation period for a CM5 sensor chip of BIACORE is 7 minutes with 0.05 M NHS/0.2 M EDC at a flow rate of 5 l/min.

[0280] Coupling:

[0281] The reactivity of the ligand at the chosen pH determines, how fast the ligand will bind to the activated surface. The rate of pre-concentration is directly related to the ligand concentration and pH of the immobilization solution. Ligand concentrations that are too high will give high ligand pre-concentration response but will also make it difficult to immobilize the proper amount of ligand. The relation between the amount of time the ligand is in contact with the activated surface and the amount of ligand bound is not linear as the sensor chip reaches saturation.

[0282] How much ligand to immobilize:

[0283] The amount of ligand to be immobilized depends on the application.

[0284] For specificity measurements, almost any ligand density will do as long as it gives a good signal.

[0285] Concentration measurements need the highest ligand density to facilitate mass transfer limitation. In a total mass transfer controlled experiment, binding will depend on the analyte concentration and not on the binding kinetics between the ligand and analyte.

[0286] Affinity ranking can be done with low to moderate density sensor chips. It is important that the analyte saturates the ligand within a proper time frame.

[0287] Kinetics should be done with the lowest ligand density that still gives a good response without being disturbed by secondary factors such as mass transfer or steric hindrance.

[0288] Low molecular mass binding should be done with high-density sensor chips to bind as much as possible of the analyte to gain proper signal.

[0289] In general, for kinetic measurements, a total analyte response of at most 100 RU, when the analyte is injected (1),(2) is desired (see mass transport). With this value in mind (Rmax), the amount of ligand (in response units) to be immobilized can be calculated with: Rmax response/Ligand response.

[0290] Deactivation:

[0291] The deactivation solution will block all remaining activated sites with an excess of reagent and because of its high ionic strength and high pH, the solution will wash away most of the electro-statically bound ligand. The amine coupling procedure is usually blocked with ethanolamine, but the use of BSA or casein is also possible. If high salt concentrations are detrimental to the ligand, the experimenter can just wait until all active sites are decayed back to carboxylic groups. The goal of the blocking is to remove the activated groups and make the surface as inert as possible.

[0292] In cases where positively charged analytes are being analyzed, the surface can be blocked with ethylenediamine to reduce the negative charge of the sensor surface and thus decrease the potential for non-specific binding.

REFERENCES

[0293] (1) Karlsson, R. et al Kinetic analysis of monoclonal antibody-antigen interactions with a new biosensor based analytical system. Journal of Immunological Methods 229-240; (1991). [0294] (2) Myszka, D. G. Survey of the 1998 optical biosensor literature. J. Mol. Recognit. 12: 390-408; (1999).

[0295] Amine Coupling

[0296] Amine coupling makes use of the N-terminus and c-amino groups of lysine residues of the ligand. The numbered points refer to the different stages in the immobilization procedure. [0297] 1) Baseline for the unmodified sensor chip surface with continuous flow (5 l/min). [0298] 2) 35 l injection of NHS/EDC to activate the surface by modification of the carboxymethyl groups to N-Hydroxy succinimide esters. [0299] 3) Baseline after activation. Activation of the surface has only a very slight effect on the SPR signal (100-200 RU). [0300] 4) Injection of ligand (10-200 g/ml) leads to electrostatic attraction and coupling to the surface matrix. At this point, the ligand solution is still in contact with the sensor surface, and response includes both immobilized and non-covalently bound ligand. The N-Hydroxy succinimide esters react spontaneously with the amines on the ligand to form covalent links (1). [0301] 5) Immobilized ligand before deactivation. The ligand has passed the sensor surface and most of the protein that is not covalently bound is eluted. [0302] 6) Deactivation of unreacted NHS-esters using 35 l of 1 M ethanolamine hydrochloride adjusted to pH 8.5 with NaOH. The increased SPR signal is due to a change in the bulk refractive index. The deactivation process also removes any remaining electrostatically bound ligand. [0303] 7) Point 7 minus Point 3 gives the amount of immobilized ligand after deactivation.

[0304] Amine coupling is the first choice with new molecules to couple. However acidic ligands (pI<3.5) are difficult to immobilize. Also when the free amine groups are in the biological active site, one of the other chemistries must be investigated.

REFERENCES

[0305] (1) Johnsson, B. et al Immobilization of proteins to a carboxymethyldextran-modified gold surface for biospecific interaction analysis in surface plasmon resonance sensors. Analytical Biochemistry 198: 268-277; (1991).

Example 1

[0306] General SPR-Method for Determining Fc-FcRn Interaction

[0307] All measurements were carried out at 25 C. using a BIAcore T200 instrument (GE Healthcare). Biotinylated single chain human FcRn was used for all interaction studies.

[0308] FcRn was immobilized in two different ways, a low density immobilization and a high density immobilization.

[0309] For the low density immobilization, FcRn was immobilized on a C1 chip using standard amine coupling. Therefore, the protein was diluted to a concentration of mg/ml with buffer (10 mM HEPES; pH 7.4) and injected for 60 sec. over the chip surface. The immobilization resulted in immobilization levels of around 80 RU.

[0310] For the high density immobilization, the FcRn was immobilized through biotin capturing. At first Neutravidin (ThermoScientific) was immobilized on a C1-chip using standard amine coupling. The Neutravidin was diluted in 10 mM Na-Acetate buffer at a pH of 4.5 to a concentration of 0.1 mg/ml and injected for 6 min. over the chip surface. The immobilization resulted in immobilization of around 1000 RU. Following the immobilization of Neutravidin, FcRn was captured by injecting the biotinylated protein for 5 min. over the chip with a concentration of 0.24 mg/ml. The capturing resulted in an immobilization level of around 1700 RU.

[0311] For the interaction measurements with different antibodies a buffer consisting of 10 mM MES at pH 5.8, 150 mM NaCl and 0.05% P-20 was used. The antibody-FcRn interactions were analyzed tested using single cycle or multi cycle kinetics and 2-fold or 3-fold dilution series. Recorded sensorgrams were double reference subtracted using a reference flow cell and blank injections. The resulting sensorgrams were evaluated using the TraceDrawer software (Ridgeview Instruments AB).