Velocity factor

Abstract

The current invention is directed to the velocity factor. Based on the velocity factor antibodies can be classified, i.e. antibodies can be characterized on their binding properties as e.g. entropic or enthalpic antigen binder. A velocity factor based classification does not require detailed thermodynamic determinations and/or calculations. The velocity factor is the ratio of the antigen-antibody complex association rate constants ka determined at 37 C. and 13 C. As only two experimental determinations are required to calculate the velocity factor this is a fast and high-throughput suited method.

Claims

1. A method for obtaining a humanized antibody comprising: a) providing a parent antibody, b) providing a set of humanized forms of the parent antibody, c) measuring the association rate constant for each humanized antibody form to its antigen at 37 C. and at 13 C., d) calculating the velocity factor for the provided antibodies, e) comparing the velocity factor of the humanized forms with the velocity factor of the parent antibody, and f) selecting a humanized form of the parent antibody as the humanized antibody with a velocity factor that is less than twice the velocity factor of the parent antibody and thereby obtaining the humanized antibody; wherein the velocity factor of an antibody is a ratio of its Ka(37 C.)/Ka(13 C.).

2. The method according to claim 1, wherein the association rate constant is measured using a surface plasmon resonance assay.

3. The method according to claim 1, wherein the selection of the antibody is further selected by determining Sass.

4. The method according to claim 2, wherein for the surface plasmon resonance assay, the antigen is immobilized on the surface plasmon resonance chip.

5. The method according to claim 3, wherein the antibody is selected with a Sass of less than 200 J/mol*K.

6. The method according to claim 5, wherein the Sass is less than 125% of the Sass of the parent antibody.

7. The method according to claim 6, wherein the Sass is less than 110% of the Sass of the parent antibody.

8. The method according to claim 1, wherein the parent antibody is produced by a single hybridoma or B-cell.

9. The method according to claim 6, wherein the velocity factor is less than 1.25.

10. The method according to claim 6, wherein the velocity factor is less than 1.10.

Description

DESCRIPTION OF THE FIGURES

(1) FIG. 1 Illustration of the Binding Late (BL) and Stability Late (SL) data of exemplary antibodies.

(2) FIG. 2 Binding Late/Complex Stability plot of 549 hybridoma primary cultures: the encircled data spot shows sufficient antigen response signal and 100% complex stability, whereas the enframed data spot shows no sufficient antigen response.

(3) FIG. 3 Secondary antibody response of the <IgGFCM>R antibody capture system versus the analyte monoclonal antibody at 25 nM, 50 nM, 75 nM and 100 nM and at increasing temperatures.

(4) FIG. 4 a) Exemplary concentration-dependent sensogram of the temperature-dependent antibody-antigen interaction of antibody M D1.1. The kinetics were measured in HBS-EP pH 7.4 at 25 C., 3 min. association time, 5 min. dissociation time, fitting according to Langmuir model;

(5) b) Exemplary concentration-dependent sensogram of the temperature-dependent antibody-antigen interaction of antibody M 9.3.1. The kinetics were measured in HBS-EP pH 7.4 at 25 C., 3 min. association time, 15 min. dissociation time, fitting according to a Langmuir 1.1. model.

(6) FIG. 5 Calculation of thermodynamic parameters according to the linear equations of (a) van't Hoff, (b) Eyring and (c) Arrhenius. Exemplary plots shown for antibody M D1.1 are shown for the van't Hoff calculation in 5a and for the Eyring calculation in 5b (association) and 5c (dissociation).

(7) FIG. 6 a) Double logarithmic plot of the temperature-dependent characteristics of 34 exemplary antibodies;

(8) b) Double logarithmic plot of the temperature-dependent characteristics of three exemplary antibodies: filled circlesantibody with increasing affinity with increasing temperature, open circlesantibody with constant affinity with increasing temperature, squaresantibody with decreasing affinity with increasing temperature.

(9) FIG. 7 Rate map with double logarithmical plotting of the antigen dissociation constant kd [1/s] versus the antigen association rate constant ka [1/Ms]; isometric lines indicate areas of constant affinity; the temperature-dependent kinetics of two humanized antibodies, antibody 1 (filled triangles) and antibody 2 (hollow stars) are plotted in the temperature gradient 13 C. to 37 C. in steps of +4 C., beginning at slower association rates at the left of the graphic.

(10) FIG. 8 Data of Table 1. Sass is plotted versus VF. The data was fitted with an exponential association model according to the equation y=y0+A1*(1exp(x/t1))+A2*(1exp(x/t2)); Y0=185.05789+/36.83135; A1=399.74088+/36.60886; t1=41.93317+/9.834; A2=198.50729+/29.933; t2=3.13298+/1.32572; R.sup.2=0.97575. The dashed line at VF=10 indicates, that most of the suitable interactions show VF<10.

(11) FIG. 9 Segmentation of the plot of FIG. 8 in four corridors by short dotted lines. The segments are numbered 1, 2, 3 and 4.

EXAMPLE 1

(12) Immunization of Mice

(13) Balb/c mice 8-12 weeks old were subjected to intraperitoneal immunization with 100 g of the antigen formulated as a KLH (keyhole limpet haemocyanine) fusion in complete Freud's adjuvant.

(14) The immunization was performed 4 times: initial boost, 6 weeks, 10 weeks and 14 weeks after the initial boost. The second and third immunization was done using incomplete Freud's adjuvant. The final boost was done i.V. using 100 g antigen three days before the hybridoma fusion took place. The production of hybridoma primary cultures was done according to Khler and Milstein (Kohler, G., et al., Nature 256 (1975) 495-497). The hybridomas were isolated in 96-well micro titer plates (MTPs) by limited dilution and were screened for antigen binding by ELISA methods according to the manufacturer's manual. Primary hybridoma cell cultures, which showed a positive color formation upon antigen binding in ELISA, were transferred into the kinetic screening process.

EXAMPLE 2

(15) Preparation of the CM5 Sensor Chip

(16) The BIAcore A100 system under the control of the Software V.1.1 was prepared like follows: A BIAcore CM5 sensor (series S) was mounted into the system and was hydrodynamically addressed according to the manufacturer's recommendations.

(17) In case of analyzing a murine antibody, the polyclonal rabbit anti-IgG antibody <IgGFCM>R (Jackson ImmunoResearch Laboratories Inc.) was immobilized on the flow cells via EDC/NHS chemistry according to the manufacturer's instructions.

(18) In case of using human chimeric or fully humanized antibodies, the polyclonal goat antibody pAb<h-lgG, Fcg-Frag>G-IgG(IS) (Jackson ImmunoResearch Laboratories Inc.) was immobilized on all flow cells via EDC/NHS chemistry according to the manufacturer's instructions.

(19) In case of using murine IgG Fab or Fab2 fragments, the polyclonal goat antibody <MFab>G-IgG(IS) (Bethyl L. Cat. No. A90-100A-5 v. 9.8.2000) was immobilized on all flow cells via EDC/NHS chemistry according to the manufacturer's instructions.

(20) In case of using human or humanized IgG Fab or Fab2 fragments, the polyclonal goat antibody <huFab2>G-IgG (Jackson Immuno Research Laboratories Inc.) was immobilized on all flow cells via EDC/NHS chemistry according to the manufacturer's instructions.

EXAMPLE 3

(21) Kinetic Screening of Primary Hybridoma Culture Supernatants

(22) Hybridoma culture supernatants from different immunization campaigns conducted according to Example 1 were processed as outlined below.

(23) The spots 2 and 4 of a sensor chip obtained according to Example 2 were used as a reference (1-2, 5-4). In order to capture antibody on the sensor surface hybridoma culture supernatants were diluted 1:5 with running buffer HBS-EP (10 mM HEPES pH 7.4, 150 mM NaCl, 1 mM EDTA, 0.05% P20, BIAcore) and were injected at 30 l/min. for 1 min. Subsequently, the respective antigen was injected at 30 l/min. for 2 to 3 min. association time. The dissociation phase was monitored for 5 to 15 min. Finally the surface was regenerated with a 2 min. injection of 100 mM phosphoric acid.

(24) The sensor was preconditioned by repeated cycles of antibody capturing and regeneration.

(25) For the selection of primary hybridomas the following procedure was used: A Binding Late (BL) reference point was set shortly before the antigen's injection ended. A Stability Late (SL) reference point was set shortly before the end of the complex dissociation phase. The BL and SL data were graphically visualized (FIG. 1). The data was used to calculate the antigen complex stability using formula (XIV):
(1[BL(RU)SL(RU)/BL(RU)])(XIV)
(see FIG. 2). E.g. the encircled data spots show sufficient antigen response signal and 100% complex stability, whereas the enframed data spot shows no sufficient antigen response.

(26) Thus, the top 10% hybridomas according to antigen response signal and complex stability have been selected.

EXAMPLE 4

(27) Hybridoma Cloning and Antibody Production

(28) Antibody producing hybridoma primary cultures selected according to Example 3 were subcloned using the cell sorter FACSAria (Becton Dickinson) under the control software V4.1.2. The deposited single clones were incubated under suitable conditions for further proliferation in 24 well plates and were subsequently transferred to the thermodynamic screening process according to Example 5 after having determined the antibody concentration in solution using ELISA methods according to the instruction of the manufacturer.

EXAMPLE 5

(29) Thermodynamic Screening

(30) Secreted antibodies were characterized by a thermodynamic screening employing the determination of the temperature-dependent kinetics in order to determine the antigen-antibody complex thermostability and in order to calculate thermodynamic properties.

(31) A CM5 sensor series S was mounted into the BIAcore T100 System driven under the control software V1.1.1 and preconditioned by 1 min. injection at 100 l/min. of a mixture comprising 0.1% SDS, 50 mM NaOH, 10 mM HCl and 100 mM H.sub.3PO.sub.4.

(32) In case of analyzing a murine antibody, the polyclonal rabbit anti-IgG antibody <IgGFCgammaM>R (Jackson ImmunoResearch Laboratories Inc.) was immobilized on the flow cells via EDC/NHS chemistry according to the manufacturer's instructions.

(33) In case of using human chimeric or fully humanized antibodies, the polyclonal goat antibody pAb<h-lgG, Fcg-Frag>G-IgG(IS) (Jackson ImmunoResearch Laboratories Inc.) was immobilized on all flow cells via EDC/NHS chemistry according to the manufacturer's instructions.

(34) In case of using murine IgG Fab or Fab2 fragments, the polyclonal goat antibody <MFab> G-IgG(IS) (Bethyl L. Cat. No. A90-100A-5 v. 9.8.2000) was immobilized on all flow cells via EDC/NHS chemistry according to the manufacturer's instructions.

(35) In case of using human or humanized IgG Fab or Fab2 fragments, the polyclonal goat antibody <huFab2>G-IgG (Jackson Immuno Research Laboratories Inc.) was immobilized on all flow cells via EDC/NHS chemistry according to the manufacturer's instructions.

(36) The concentration values of the reference antibody were adjusted in order to achieve similar secondary antibody response levels at different temperatures.

(37) Kinetic measurements at different temperatures were performed at 20 l/min., the flow rate was 30 l/min., 50 l/min., 100 l/min., respectively. The sample injection of the antigen was done for 30 sec., 90 sec., 180 sec., respectively, or other suitable injection times in order to achieve ligand saturation or entry into the binding equilibrium during the complex association phase (see FIG. 4 a)). The dissociation rate was monitored first for up to 300 sec. and further for 15 min. (see FIG. 4 b)). The antigen injections were repeated in different concentration steps of at least five concentrations. As control one concentration step was analyzed twice to control the reproducibility of the assay. Flow cell 1 served as a reference. A buffer injection was used instead of an antigen injection to double reference the data by buffer signal subtraction. The capture system was regenerated using 100 mM H.sub.3PO.sub.4 by a 2 min. injection at 100 l/min. The regeneration procedure was optimized to guarantee quantitative surface regeneration also at 13 C., 17 C. and 21 C. At these temperatures the regeneration solution was injected three times whereas at 25 C., 29 C., 33 C. and 37 C. the regeneration solution was injected one time.

(38) The data obtained was evaluated according to a 1:1 binary Langmuir interaction model in order to calculate the association rate constant ka [1/Ms], the dissociation rate constant kd [1/s] and the resulting affinity constant K.sub.D [M] at different temperatures. Thermodynamic equilibrium data was calculated according to the linear form of the Van't Hoff equation. Transition State thermodynamics were calculated according to the Eyring and Arrhenius equations using e.g. the BIAcore T100 evaluation software V.1.1.1 or the program Origin 7SRI v. 7.0300.

EXAMPLE 6

(39) Example for the Effect of the Adjustment to Homogeneous RU.sub.MAX Values

(40) A BIAcore T100 device was mounted with a CM5 series-S BIAcore sensor, and was immobilized with 6000 RU <IgGFCM>R (Jackson ImmunoResearch Laboratories Inc., USA) on each flow cell according to the manufacturer's instructions. The non-optimized experiment used 40 nM capture antibody at 20 l/min., in HBS-EP buffer (0.05% P20). The sample buffer was the system buffer, supplemented with 1 mg/ml CMD (carboxymethyldextrane).

(41) The antigen was injected after the capturing of the secondary antibody in six concentration steps of 1.2 nM, 4 nM, 11 nM, 33 nM, 100 nM and 300 nM, whereby 11 nM were used as a double control and 0 nM were used as reference. The antigen was injected at 100 l/min. for 2 min. association and 5 min. dissociation, followed by a HBS-EP wash of 15 min. at 30 l/min. and a regeneration with 10 mM glycine pH 1.7 at 3 l/min. for 3 min. Concentration-dependent measurements were done at 4 C., 11 C., 18 C., 25 C., 32 C., and 40 C.

(42) The optimized system was used like described above, but with the exceptions that the antibody to be captured was injected for 3 min. association time at different concentration steps of 100 nM at 15 C., 80 nM at 20 C., 60 nM at 25 C., 50 nM at 30 C., 40 nM at 35 C. and 40 nM at 40 C.

(43) Finally kinetics and thermodynamics were determined using the BIAcore evaluation software.

EXAMPLE 7

(44) Determination and Calculation of the Velocity Factor

(45) A CM5 sensor series S was mounted into the BIAcore T100 System.

(46) In case of using murine antibodies, the polyclonal rabbit IgG antibody <IgGFCM>R (Jackson ImmunoResearch Laboratories Inc.) was immobilized on the flow cells via EDC/NHS chemistry according to the manufacturer's instructions.

(47) In case of using human chimeric or fully humanized antibodies, the polyclonal goat antibody pAb<h-lgG, Fcg-Frag>G-IgG(IS) (Jackson ImmunoResearch Laboratories Inc.) was immobilized on all flow cells via EDC/NHS chemistry according to the manufacturer's instructions.

(48) In case of using murine IgG Fab or Fab2 fragments, the polyclonal goat antibody <MFab>G-IgG(IS) (Bethyl L. Cat. No. A90-100A-5 v. 9.8.2000) was immobilized on all flow cells via EDC/NHS chemistry according to the manufacturer's instructions.

(49) In case of using human or humanized IgG Fab or Fab2 fragments, the polyclonal goat antibody <huFab2>G-IgG (Jackson Immuno Research Laboratories Inc.) was immobilized on all flow cells via EDC/NHS chemistry according to the manufacturer's instructions.

(50) The sample buffer was the system buffer supplemented with 1 mg/ml carboxymethyldextrane to reduce unspecific sensor matrix effects. Kinetic measurements in the temperature gradient 13 C. to 37 C. were performed at 100 l/min.

(51) The analyte injections of recombinant synthetic human full length antigen 1, recombinant human antigen 2 (10 kDa), recombinant human antigen 3 human FC-chimera (R&D Systems, 160 kDa), recombinant human antigen 4 (29 kDa), or recombinant human antigen 5 (72 kDa) were done for 180 sec. The dissociation rate was monitored for up to 900 sec. The antigen injections were repeated in different concentration steps of at least five concentrations. As a control one concentration step was analyzed twice to control the reproducibility of the assay. Flow cell 1 was used as reference. A blank buffer injection was used instead of an antigen injection to double reference the data by buffer signal subtraction.

(52) Prior to each assay, homogenous RU.sub.MAX values in the temperature range 13 C.-37 C. were adjusted by titration experiments with the respective antibodies to be presented on the sensor surface (see section in which the temperature-dependent titration experiments are described in detail). The kinetics were measured in the temperature gradient 13 C., 17 C., 21 C., 25 C., 29 C., 33 C. and 37 C. The capture systems were regenerated using a buffer wash with HBS-ET buffer at 30 l/min. for 15 sec., prior to regeneration with 10 mM glycine pH 1.5 at 30 l/min. for 15 sec., followed by a 1 min. injection and a 30 sec. injection of 10 mM glycine at pH 1.7.

(53) The data obtained was evaluated according to a 1:1 binary Langmuir interaction model in order to calculate the association rate constants ka [1/Ms], the dissociation rate constants kd [1/s] and the resulting affinity constants K.sub.D [M] at the respective temperatures. Thermodynamic equilibrium data was calculated according to the linear and non linear form of the Van't Hoff equation. Transition State thermodynamics were calculated according to the Eyring and Arrhenius equations using e.g. the BIAcore T100 evaluation software V.1.1.1. Graphic evaluation was done using Origin 7SRI v. 7.0300.

(54) The Velocity Factor (VF) was calculated as the quotient of the antigen complex association rates ka (1/Ms) at 37 C. and 13 C. The exponential association fitting curve y=y0+A1*(1exp(x/t1))+A2*(1exp(x/t2)) was used.

EXAMPLE 8

(55) High Throughput Velocity Factor Analysis

(56) A CM5 sensor series S was mounted into the BIAcore A100 System and the detection spots were hydrodynamically addressed according to the manufacturer's instructions.

(57) The polyclonal rabbit IgG antibody <IgGFCM>R (Jackson ImmunoResearch Laboratories Inc.) was immobilized at 4 kRU on the detection spots 1 and 5 in each flow cell. 800 RU <IgGFCM>R were immobilized on spots 2 and 4 in each flow cell. Coupling was done via EDC/NHS chemistry according to the manufacturer's instructions. The sample buffer was the system buffer supplemented with 1 mg/ml carboxymethyldextrane to reduce unspecific sensor matrix effects.

(58) The basic buffer system was PBS (phosphate buffered saline). The buffer was adjusted to four different pH conditions: pH 6.8, pH 7.0, pH 7.4 and pH 7.8, and four different KCl concentrations: 2.7 mM, 54 mM, 162 mM and 324 mM. Sixteen different sample buffer conditions were tested.

(59) The mAbs to be captured were injected at 10 l/min. for 1 min. at different concentration steps between 60 nM at 37 C. and 240 nM at 13 C. to ensure homogenous RU.sub.MAX values in the subsequent antigen interaction measurements.

(60) Kinetic measurements in the temperature gradient 13 C. to 37 C. were performed at 30 l/min. The analyte injections of recombinant human antigen 1-84 (9.4 kDa), were done for 180 sec. The dissociation rate was monitored for 600 sec. The antigen injections were repeated in two concentration steps at 60 nM and 240 nM. A blank buffer injection was used instead of an antigen injection to double reference the data by buffer signal subtraction.

(61) The capture system was regenerated using a 15 sec. buffer wash with HBS-ET buffer at 30 l/min., prior to regeneration with 10 mM glycine pH 1.5 at 30 l/min. for 90 sec., followed by 230 sec. injection of the same buffer.

(62) The kinetic data obtained was evaluated according to a 2 over 2 kinetic model, which uses two different ligand densities of the sensor matrix to calculate kinetics by just using two antigen concentrations. The association rate constants ka [1/Ms], the dissociation rate constants kd [1/s] and the resulting affinity constants KD [M] at the respective temperatures were calculated using the BIAcore A100 evaluation software 1.1.

(63) Thermodynamic equilibrium data was calculated from the kinetic data according to the linear form of the Van't Hoff equation. Transition State thermodynamics were calculated according to the Eyring equations using Excel. Graphic evaluation was done using Origin 7SRI v. 7.0300. The Velocity Factor (VF) for the association phase was calculated as the quotient of the antigen complex association rates ka (1/Ms) at 37 C. and 13 C. The Velocity Factor for the dissociation rates were calculated as the quotient of kd (1/s) at 13 C. and 37 C.