Colour coding changes: Abvance Biotech, Richard, GSK, Lilly, MedImmune, Novartis
Impact of production process:
Likely (define cause of potential impact or reason), or
Unlikely
Likelihood of comparability definitions (assumption that individual assay within the same organisation is reproducible):
High - possible for direct comparison of absolute values across assay clusters which observe the same phenomenons e.g. DSF & DSC
Medium - possible for direct comparison of absolute or relative (wrt appropriate controls) values between individual assays
Low - possible to compare individual assays only using RAG flags
| Parameter measured | Assay name | Assay abbreviation | Assay description (link to primary citation) | Purpose | Tech platform/aka (e.g. NanoPro) | Used by <company> | Stage | Measure of (units) | Data needed for BASECASE | Impact of production process | Likelihood of comparability - inc. assay clusters which observe the same phenomenons | Additional information or comments |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| INTRINSIC MOLECULAR PROPERTIES | ||||||||||||
| Chemical stability | Capillary Isoelectric Focusing | cIEF | High throughout, capillary-based isoelectric immunoassay to determine charge profile and pI of molecule Michels et al., Anal. Chem 2012 | Measure of pI and charge heterogeneity - used in conjunction with LC-MS. | NanoPro Beckman Proteinsimple | GSK (check if used as control to determine pI) Lilly MedImmune (check) | Early selection Final profiling Endpoint data | pI | Capture buffer conditions Record pI of main isoform Record proportion of main isoform to acidic and basic species | Likely - Glycan composition will affect charge profile, therefore antibodies produced by different processes may show different profiles Low for antibodies (May be significant for other proteins in view of glycosylation impact) | High Should be comparable even with subtle differences in method and antibody formulation | Parameter of low value - can be predicted in silico reliably enough from sequence (published) |
| Chemical stability, Immunogenicity | Liquid chromatography-Mass spectrometry | LC-MS | Intact molecular mass and 'standard' glycosylation pattern and obvious fragmentation. | Abvance GSK Lilly MedImmune Novartis (check) | Final profiling Endpoint data | MW % intact species % of respective glycovariants | Capture buffer conditions | Likely - Glycan composition will affect 'standard' pattern | High | Parameter of low value - intact molecular mass and fragmentation can be assessed by SDS-PAGE. Ranking of molecules by LC-MS and CE-SDS comparable Re. Immunogenicity, need to define what a 'standard' glycosylation pattern is and how this is linked to immunogenicity | ||
| Structural stability | Differential Scanning Fluorimetry (including Thermofluor) (including Nano DSF) | DSF | DSF is a fast and economical method to determine the temperature at which a whole antibody or a Fab unfolds (Tm), i.e. the temperature at the mid-point of the unfolding transition. The Tm is measured by an increase in the fluorescence of a dye with affinity for hydrophobic parts of the protein, which become progressively more exposed as the antibody unfolds. The Tm is assigned using the first derivative of the raw fluorescence emission data. | Rank Fabs and antibodies according to their thermal stability as measured by the Tm. Screen for favourable buffer and/or additive conditions. | iQ5, CFX96 Real-Time Systems (Bio-Rad) Prometheus (Nanotemper) | GSK Lilly (rarely) MedImmune Abvance Novartis (Jain et al) | Early selection | Thermal stability (ºC) | Capture buffer conditions Record number of unfolding transitions Capture which dye is used as this can affect Tm (e.g SPYRO Orange) Capture conc of sample as will affect Tm Capture temp gradient used during run (e.g. A temperature ramp from 40 ºC to 95 ºC at a rate of 0.5 ºC/2 min) Capture whether Tm is for full IgG or Fab | Unlikely | High (assuming similar buffer conditions) | Ranking of molecules by DSF and DSC comparable |
| Structural stability | Differential Scanning Calorimetry | DSC | Rank Fabs and antibodies according to their thermal stability as measured by the Tm. Screen for favourable buffer and/or additive conditions. | Lilly MedImmune Novartis | Final profiling | Thermal stability (ºC) | Capture buffer conditions Record number of unfolding transitions Capture conc of sample as will affect Tm Capture temp gradient used during run (e.g. A temperature ramp from 40 ºC to 95 ºC at a rate of 0.5 ºC/2 min) Capture whether Tm is for full IgG or Fab | Unlikely | High | Ranking of molecules by DSF and DSC comparable | ||
| Structural stability, (self) Aggregation | Analytical Size-Exclusion Chromatography | aSEC | Antibodies are flowed through chromatography media consisting of Non-aggregated dimeric and monomeric antibodies elute as defined, well-separated peaks. Aggregated antibodies flow through the column faster and elute before the non-aggregated antibody peaks. Percentage aggregation can be worked out from the ratio of peak surface areas under the aggregated vs non-aggregated antibody peaks. | Verify structural integrity, oligomeric state, percentage aggregation of antibody samples. | GE Healthcare Agilent | GSK MedImmune Lilly Abvance Novartis (Jain et al) | Early selection Final profiling | Elution time, peak width, peak asymmetry relative to a well-behaved control antibody (% monomeric species, retention time per unit volume or time) | Capture buffer conditions Capture column type (e.g beads 25-45 µm in size - GE Superdex 200). Capture Flow rate Relative size (MW) | Unlikely | High | |
| Chemical stability (integrity), Structural stability (integrity) | Capillary electrophoresis | CE-SDS | Verify chemical and structural integrity, quantify fragmentation | Beckman Caliper | GSK Lilly MedImmune (check) Novartis | Final profiling Endpoint data | Approx. MW Impurities | Capture sample prep conditions: reduced, non-reduced, denaturation | Unlikely | High | Ranking of molecules by LC-MS and CE-SDS comparable | |
(self) Aggregation (Hydrophobicity) | Hydrophobic Interaction Chromatography | HIC | HIC separates antibodies on non-biological surfaces based on the distribution and nature of the antibody hydrophobic patches. This approach assumes that the increased retention of antibodies on hydrophobic columns at moderate to high salt concentrations (e.g. ammonium sulphate, ammonium phosphate) correlates with increased antibody hydrophobicity. Antibodies are eluted in a mobile phase without added salt, i.e. only in buffer. | Assess non-specific interactions. Rank antibodies according to their relative hydrophobicity | GE Healthcare Agilent | GSK Lilly MedImmune Abvance Novartis Jain et al | Early selection | Retention time (retention time per unit volume or time) | Capture buffer conditions Capture column type (e.g beads 25-45 µm in size - GE Superdex 200). Capture salt gradient | Unlikely | High should be comparable (relative retention time rather than absolute values) | Retention time used for ranking of molecules. |
| Structural stability, (self) Aggregation | Dynamic Light Scattering | DLS | Dynamic Light Scattering (DLS) measures the translational diffusion coefficients Dt of nanoparticles and colloids in solution by quantifying dynamic fluctuations in scattered light due to Brownian motion of the molecules in liquid. Sizes and size distributions, in turn, are calculated from the diffusion coefficients in terms of hydrodynamic radius rh or hydrodynamic diameter dh. Technique can be used with high concentration samples (cf SEC) | Measure hydrodynamic radius (characterise/quantify aggregation) | Wyatt | Lilly (rarely) MedImmune | Final profiling Endpoint data | hydrodynamic radius (nm) | Capture buffer conditions Sample concentration Molecule type | Unlikely | High should be comparable for same buffer conditions and sample concentrations | Ranking of molecules by DLS and AC-SINS comparable |
| (self) Aggregation | Self-interaction by light scattering | K-Diff | DLS experiment which measures the propensity for molecule to self-associate under particular conditions. K-diff = Association constant defined by diffusion. | Rank antibodies according to their propensity to self-assemble | DLS | Lilly MedImmune (check) Novartis | Final profiling | (mL/g) | Capture buffer conditions Capture concentration range K-diff | Unlikely | Probably High | |
| (self) Aggregation | Affinity-Capture Self-Interaction Nanoparticle Spectroscopy | AC-SINS | Sule et al., Mol. Pharmacuetics, 2013 High-throughput, low protein screen for self-association propensity, where antibodies are captured onto gold nanoparticles. Self-association is determined if the nanoparticles with captured antibody clump together resulting in a red-shift. | Assess propensity of antibody to self-associate / RSA at high concentrations | Gold nanoparticles Scanning plate reader | MedImmune Jain et al | Early selection | Plasmon wavelength redshift in nm | Capture buffer conditions Record redshift in nm, relative to non-capture / buffer only controls | Unlikely | Should be largely comparable if buffer conditions are the same | |
| (self) Aggregation | Multiangle laser light scattering | MALLS | MALLS describes a technique for measuring the light scattered by a sample into a plurality of angles. It is used for determining both the absolute molar mass and the average size of molecules in solution, by detecting how they scatter light. (wikipedia) | Determine symmetry or presence of multiple aSEC peaks and shape independent measure of MW | Abvance | Does not provide much additional information over aSEC experiment | ||||||
| Chemical stability, (Immunogenicity) | In silico predictions | Predict pI and PTM sites in CDR | GSK MedImmune Novartis | Early selection | Not of high interest for comparisons. Can be applied to all selected molecules easily if wished for. | |||||||
| (human, cyno, mouse-GSK, rat-Novartis) PK | FcRN interaction (SPR) | Binding to FcRn contributes to the long half-life of IgGs and albumine. Properties of antibodies to bind FcRn receptors is assessed for candidate ranking and risk minimizing. | Assessment of FcRN binding properties to identify risk of unfavorable PK. | Biacore EPIC | GSK Lilly Novartis | Final profiling Endpoint data | Apparent affinity at pH 6.0, 7.4 to FcRn (M) % Residual binding to FcRn observed | Capture buffer conditions Capture assay setup (ratio Ab:receptor, fitting model) and readout (ratio Ab:receptor)? | Unlikely | Medium/Low Comparability vs control, not absolute numbers as dependent on conditions used | Not necessarily predictive - unusual activity for FcRN signifies potential issue. However, in some cases no FcRN activity results in abnormal PK. | |
| (human, cyno) PK | Heparin Sulphate Chromatography | Rule out proteoglycan interactions affecting PK | HPLC (Agilent, Waters, Thermo, etc) | Lilly | Early selection Final profiling | Retention time (or salt concentration at elution), relative to control antibodies | Column details Chromatography conditions (buffer, salt gradient details) | Unlikely | High | Not standard use. Data comparable to heparin binding (in cells) | ||
| (human, cyno) PK | ELISA Baculovirus particles | BVP ELISA | ELISA on baculovirus particles and non coated ELISA plates under low stringency blocking conditions to assess nonspecific binding / polyspecificity of lead antibody candidates. Binding observed to either BVPs or plastic indicates that the antibody has a risk of faster nonspecific clearance in vivo. Hotzel et al., mAbs, 2012 | Polyspecificity / nonspecific binding / aberrant PK risk | ELISA | MedImmune Jain et al | Early selection Final profiling | Absorbance at 450nm if using HRP detection | Capture buffer conditions Capture concentration range of antibodies tested Record absorbance on both BVPs and non-coated wells Record binding ratio relative to controls (sticky and non-sticky) | Unlikely (providing there is high monomer purity) | Whether an antibody flags or not, should be comparable. Actual numbers may not be | Low use across industry - possibly used by Adimab and Genentech |
| (human, cyno) PK | Nonspecific HEK or CHO cell binding by flow cytometry (HyperCyte) or non-wash cytometry (Mirrorball) | HEK/CHO binding | Providing the target for an antibody panel is not expressed on a certain cell type (e.g. HEK293 or CHO), then assay systems using this cells are effective in determining the polyspecificity and therefore PK risk. | Polyspecificity / nonspecific binding / aberrant PK risk | Flow cytometry, Mirrorball (TPP Labtech) | MedImmune | Early selection Final profiling | Fluorescence counts | Capture assay conditions Capture concentration range of antibodies tested Record total fluorescence counts Record fluorescence ratio relative to controls (sticky and non-sticky) | Unlikely (providing there is high monomer purity) | Whether an antibody flags or not, should be comparable. Actual numbers may not be | |
| Viscosity | Viscosity | Assess viscosity of unformulated vs formulated antibody | Malvern Rheometer Malvern Viscosizer | GSK (not routine so limited data) Lilly Novartis MedImmune (check) | Endpoint data | Capture buffer conditions Capture concentration | Unlikely | Low High if measured in similar buffer conditions | Very condition dependent so we need to get a sense of what comparisons are possible | |||
ACCELERATED STABILITY (AS) PROPERTIES Stressed samples will be compared to non-stressed samples to determine changes due to accelerate stability conditions. Capture stress conditions (sample treatment) applied ahead of running each assay: stress stimulus, buffer conditions, storage temperature, sampling timepoints (incubation time) | ||||||||||||
| Chemical stability | Capillary Isoelectric Focusing | cIEF | High throughout, capillary-based isoelectric immunoassay to determine charge profile and pI of molecule Michels et al., Anal. Chem 2012 | Change in acidic and basic variants. Effect on binding (SPR) of changing pI. | NanoPro Beckman Proteinsimple | GSK Lilly MedImmune | Early selection Final profiling Endpoint data | Capture buffer conditions Record pI of main isoform Record proportion of main isoform to acidic and basic species Record change in main isoform upon stress to acidic or basic species | As above, except stressed samples will be compared to non-stressed samples to determine changes in charge profile due to AS conditions. | |||
| Chemical stability | PTM by Liquid chromatography-Mass spectrometry | LC-MS (PTM) | Description (citation) | Identification and approx. quantification of post-translational modifications by peptide mapping | ESI-MS | Lilly Novartis | Final profiling Endpoint data | Relative amount of modification | Capture buffer conditions Capture digest conditions (pH) | Medium | High | Detailed experiment carried out on AS samples |
| Structural stability, (self) Aggregation | Analytical Size-Exclusion Chromatography | aSEC (AS) | Antibodies are flowed through chromatography media consisting of Non-aggregated dimeric and monomeric antibodies elute as defined, well-separated peaks. Aggregated antibodies flow through the column faster and elute before the non-aggregated antibody peaks. Percentage aggregation can be worked out from the ratio of peak surface areas under the aggregated vs non-aggregated antibody peaks. (citation) | Verify structural integrity, oligomeric state, percentage aggregation of antibody samples. | GE Healthcare Agilent | GSK MedImmune Lilly Abvance Novartis (Jain et al) | Early selection Final profiling Endpoint data | Elution time, peak width, peak asymmetry relative to a well-behaved control antibody (% monomeric species, retention time per unit volume or time) (% high molecular formation per unit time) | Capture buffer conditions Capture antibody concentration Capture column type (e.g beads 25-45 µm in size - GE Superdex 200). Capture Flow rate Relative size (MW) | Unlikely | High | Same as aSEC but using stressed samples |
| Chemical stability (integrity), Structural stability (integrity) | Capillary electrophoresis | CE-SDS | Verify chemical and structural integrity, quantify fragmentation | Beckman Caliper | GSK Lilly MedImmune (check) Novartis | Final profiling Endpoint data | Approx. MW Impurities | Capture sample prep conditions: reduced, non-reduced, denaturation | Unlikely | High | Used in AS Cluster with LC-MS | |
| Structural stability, (self) Aggregation | Dynamic Light Scattering | DLS | Dynamic Light Scattering (DLS) measures the translational diffusion coefficients Dt of nanoparticles and colloids in solution by quantifying dynamic fluctuations in scattered light due to Brownian motion of the molecules in liquid. Sizes and size distributions, in turn, are calculated from the diffusion coefficients in terms of hydrodynamic radius rh or hydrodynamic diameter dh. Technique can be used with high concentration samples (cf SEC) | Measure hydrodynamic radius (characterise/quantify aggregation) | Wyatt | Abvance GSK | Final profiling Endpoint data | hydrodynamic radius (nm) | Capture buffer conditions Molecule type Sample concentration | Unlikely | High - should be comparable for same buffer conditions and sample concentrations | Cluster with AC-SINS |
| Chemical stability | Cation Exchange Chromatography | CEC (IEC) | Description (citation) | Change in acidic and basic variants. | Novartis (GSK) |
| ||||||
| Chemical stability | Capillary Zone Electrophoresis | CZE | Description (citation) | Change in acidic and basic variants. | Novartis | Final profiling |
| |||||
| Structural stability, (self) Aggregation | Analytical Ultra Centrifugation | AUC | Low throughput method used to characterise and quantify aggregation (especially at high concentrations). Two types - sedimentation velocity or sedimentation equilibrium. Sedimentation velocity uses centrifugal force to measure sedimentation coefficient, diffusion coefficient and mass of sample. | Characterise and quantify aggregation | Beckman Coulter XLI-1 | GSK Lilly Abvance | Final profiling | mass, distribution, sedimentation coefficient | Capture buffer conditions Molecule type concentration rotor speed | should be comparable for same buffer conditions | ||
| Chemical stability, Structural stability | Surface Plasmon Resonance | SPR | SPR measures changes in refractive index at a gold sensor chip surface which are proportional to changes in mass. It is used to measure binding kinetics and affinities of molecules. SPR can be used in conjunction with other techniques to look at the consequence of chemical or structural changes on the affinity of an antibody (an intrinsic property) to either to its antigen of interest or Fc receptors e.g. FcRn. SPR can also be used to measure the active concentration of a molecule (i.e. the proportion able to bind). | Measure changes in active concentration or kinetics. Change in target binding upon stress | Biacore ProteOn | GSK Novartis |
Endpoint data | Ka (1/Ms), Kd (1/s), KD (M), active concentration (%) | Capture buffer conditions chip/surface used assay format (i.e. ligand and analyte) concentrations | technique measures impact of chemical/structural changes Low (define) | should be comparable if 1:1 assay format used High (define) | Probably not of high interest for comparison and prediction |
| Capillary Gel Electrophoresis | CGE | GSK | Endpoint data | Capture buffer conditions | ||||||||
| PEG-induced precipitation | Lilly | Final profiling | Capture buffer conditions | |||||||||
| Reversed phase HPLC | GSK | Endpoint data | Capture buffer conditions | |||||||||
| Rheometry | Turbidity | Novartis | Endpoint data | Capture buffer conditions | ||||||||