Novel instrumentation to gain insight into the aggregation state of proteins by Taylor Dispersion Analysis: a comparison

Transcription

1 Novel instrumentation to gain insight into the aggregation state of proteins by Taylor Dispersion Analysis: a comparison with Dynamic Light Scattering Wendy Louise Hulse *, Rob Forbes, David Goodall, Alex Chapman, Brian Powell, Matthew Hancock * University of Bradford, UK June 11 th 2009

2

3 Introduction Project context and overview Alpha laboratory system Taylor Dispersion Analysis (TDA) Results and comparison with DLS Coupling ActiPix with CE system for automated analysis Results for automated system Conclusions

4 Biopharmaceuticals Context and opportunities Pharma companies typically devote 25% of revenue and 80% of total fixed assets to manufacturing operations. For a company with $20 billion in annual revenue, a 1% reduction in manufacturing cost translates into savings of $50 million per year -The Economist, October 2004 In 2005, approximate combined annual sales of Erythropoietin, Insulin and Interferon were $20 billion.

5 Immunogenicity: A key issue Largely dependent on aggregation state? Question: Can we produce formulations more cheaply with less aggregates and a lower tendency to aggregate on storage? In-vial On-storage

6 Project overview Key aim The ultimate goal is for intelligent switching of process line flow to avoid unnecessary loss of main-component material when rejecting byproduct dimers and aggregates. This will provide significant savings in production costs of highly valuable molecular biopharmaceuticals.

7 Paraytec s innovation - Actipix Uses a capillary as combined sample vessel and lens Unique signal referencing system using Active Pixel Sensor First ever dedicated UV area imager Uses UV for both quantification and sizing - two key measurements with a single instrument Readily coupled to separation stage to quantify and size all molecules in mixtures DLS and MALS just provide size measurements

8 Novel analytical instrumentation to detect aggregation in biopharmaceutical processing ( 1M+ project TSB{DTI} and EPSRC funded) Avecia Biologics Ltd Eli Lilly and Company Ltd Lonza Biologics Plc End Users 6 Consortium members Paraytec Intertek ASG Bradford Software development Hardware development Application development ActiPix D Pittcon silver award

9 Alpha laboratory instrument The ActiPix detector monitors broadening of a band of a therapeutic protein or small molecule injected into a stream of buffer solution. The buffer is driven by a syringe pump through a fused-silica capillary.

10 Principles The band is imaged at two points, the first on entry to and the second on exit from a loop in the capillary. The hydrodynamic radius is calculated from the measured differences between peak times (first moments) and variances (second moments) at the two windows. Applicable to proteins, small molecules and aggregates.

11 Sizing Fluid In ActiPix Benefits No other way to offer both sizing and quantification Difficult and expensive to measure using other techniques Measure size post LC Easily integrated with other instrumentation

12 Determination of hydrodynamic radius Band broadening due to Taylor dispersion is calculated from absorbance versus time data using the peak centre times at the first and second window, t 1 and t 2 respectively, and the corresponding standard deviations, τ 1 and τ 2 (band broadening). These values are used to calculate the hydrodynamic radius, R h, using the equation h where k B is the Boltzmann constant, T the absolute temperature, η the viscosity of the solution, and r the capillary radius.

13 Typical conditions mg/ml solutions 10 nl injection volume 5-10 min analysis time 214 nm filter Flowrate 0.5 µl/min Capillary size 75:200 µm ID:OD

14 Example: Ovalbumin Injection volume, 10 nl; concentration 10 mg/ml; buffer, PBS ph 7.4

15 Data analysis R = 4k T h ( τ τ ) B 2 1 ( ) πη r t t 2 1 Data analysis carried out automatically in ActiPix software and result for R h and diffusion coefficient reported

16 Results Caffeine Caffeine used as a small molecule reference standard; (Results from Paraytec) Literature value 0.32 nm; Leist and Hui, J. Phys. Chem

17 Results reference proteins Protein Radius (nm) ± SD (n = 30) Literature value BSA (A G) 1 mg/ml 3.5 ± nm Brownsey et al., Biophys. J BSA (A G) 10 mg/ml 3.8 ± 0.4 BSA (A2058-5G) 1 mg/ml 3.5 ± 0.6 BSA (A2058-5G) 10 mg/ml 3.7 ± 0.4 Ovalbumin 1 mg/ml 2.9 ± nm Ovalbumin 10 mg/ml 3.1 ± 0.5 Croguennec et al., J. Coll. Int. Sci Lysozyme 10 mg/ml 1.5 ± nm Bonincontro et al., Coll. Surf. 2001

18 Comparison with DLS: unfiltered samples Protein TDA radius ± SD (nm) (n = 30) DLS radius (z average) ± SD (nm) (n = 30) BSA (A G) 1 mg/ml 3.5 ± ± 95 BSA (A G) 10 mg/ml 3.8 ± ± 0.4 BSA (A2058-5G) 1 mg/ml 3.5 ± ± 38.6 BSA (A2058-5G) 10 mg/ml 3.7 ± ± 6.8 Ovalbumin 1 mg/ml 2.9 ± ± 275 Ovalbumin 10 mg/ml 3.1 ± ± 21.0 Lysozyme 10 mg/ml 1.5 ± ± 264 Note: DLS very susceptible to presence of large particles (scattering α sixth power of radius!)

19 Results IgG 4 Protein TDA radius ± SD (nm) DLS radius (z average) Literature value (n = 10) ± SD (nm) (n = 30) IgG ± ± Fujita et al., Dia. Care 1999 IgG 4 concentration 35 mg/ml; buffer, PBS ph 7.4 All measurements on IgG 4 sample as supplied in formulation solution Precision : RSDs are typically 3% for TDA, 1% for DLS Accuracy: radius from TDA accords well with literature value; radius from DLS Is systematically higher (possibly due to high concentration).

20 Coupling to CE system for automation

21 ActiPix detector head in Agilent 3D CE system

22 Analytical conditions Sample concentration mg/ml Run time 30 min (10 injections & data analysis) Detection wavelength 214 nm Capillary size 75:200 µm ID:OD, 127 cm Injection volume 60 nl (50 mbar, 18 s) Flowrate during run 3-4 µl/min (1000 mbar)

23 IgG: repeatability with automated system Similar results with standard proteins (BSA, Lysozyme and Ovalbumin) all show RSD less than 3%

24 Close up view: repeatability series Ability to determine aggregate levels?

25 Automated system - Results Sample Hydrodynamic radius ± SD (nm) (n = 10) Caffeine (1000 ppm) ± (RSD 2%) IgG (35 mg/ml) run ± 0.15 (RSD 2.6%) IgG (35 mg/ml) run ± 0.22 (RSD 3.8%) IgG (35 mg/ml) run ± 0.16 (RSD 2.9%)

26 Next steps? Polyacrylamide internally coated capillary to potentially eliminate sticking Analysis of partner proteins with known aggregate content Application to a wider range of proteins and small molecules (please feel free to send us your samples for analysis!) Modification of alpha instrument with novel injector valve (eliminate need of expensive automated injector?)

27 Conclusions Hydrodynamic radii using Taylor Dispersion Analysis (TDA) with ActiPix detector show good correlation with literature values Comparison with DLS shows that TDA has far less susceptibility to bias from large particles Repeatability optimised by coupling with a CE instrument. Benefits include automated run sequences and consistently low RSD (1-3%) The ability to determine aggregate levels using this technology is being explored

28 TSB (DTI) EPSRC Paraytec Intertek ASG Avecia Biologics Acknowledgements Eli Lilly and Co. Lonza Biologics PLC

29 Thank you for your attention