Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Analyzing Aggregates of Different Sizes and Types: SEC vs. Analytical Ultracentrifugation vs. Light Scattering

798 views

Published on

invited talks, WCBP and IBC Well Characterized Biologicals, 2004

Published in: Health & Medicine
  • How To Get Rid Of Acne, The amazing clear skin secret Of top models and celebrities  http://ishbv.com/buk028959/pdf
       Reply 
    Are you sure you want to  Yes  No
    Your message goes here
  • Can You Spare 1 Minute Per Day? If you have a few spare minutes per day, we have some very exciting news for you! ♣♣♣ https://tinyurl.com/1minweight
       Reply 
    Are you sure you want to  Yes  No
    Your message goes here
  • Be the first to like this

Analyzing Aggregates of Different Sizes and Types: SEC vs. Analytical Ultracentrifugation vs. Light Scattering

  1. 1. Analyzing Aggregates of Different Sizes and Types: SEC vs. Analytical Ultracentrifugation vs. Light Scattering John Philo Director of Biophysical Chemistry © copyright 2004, Alliance Protein Laboratories, Inc.
  2. 2. Outline Different types of oligomers and aggregates Some problems with SEC Applications and advantages/drawbacks of: 1. sedimentation velocity 2. sedimentation equilibrium 3. SEC with on-line classical light scattering 4. dynamic light scattering Some recommended applications for these biophysical methods
  3. 3. I can’t really cover this broad topic in 20 min so I’m going to omit some background and summaries but… 1. See poster P-30-W 2. An expanded presentation can be downloaded from the APL web site at http://www.ap-lab.com/further_reading.htm 3. Also see our web site [www.ap-lab.com] for more method background and applications
  4. 4. The word “aggregate” covers a wide spectrum of types and sizes of associated states 1. rapidly-reversible non-covalent small oligomers (dimer, trimer, tetramer…) 2. irreversible non-covalent oligomers 3. covalent oligomers (e.g. disulfides) 4. “large” aggregates (> 10-mer) could be reversible if non-covalent 5. “very large” aggregates (diameter ~50 nm to 3 µm) could be reversible if non-covalent 6. visible particulates probably irreversible
  5. 5. Whether aggregates are “irreversible” or “reversible” depends on the context solvent components salts, sugars, other excipients organic modifiers (alcohols, acetonitrile) pH temperature how long you wait
  6. 6. The kinetics of non-covalent association and dissociation (half-times) can vary from milliseconds to days metastable oligomers with lifetimes of hours to days occur frequently for an antibody example see J.M.R. Moore et al. (1999) Biochemistry 38: 13960-13967 it may take hours to days for a protein to re-equilibrate its association after a change in concentration, solvent conditions or temperature
  7. 7. Any given protein sample may contain multiple aggregate forms with widely-varying properties both covalent and non-covalent different types of non-covalent: both rapidly-reversible and irreversible both rapidly- and slowly-dissociating even the same size of oligomer may exist in multiple forms
  8. 8. We must be aware that any given analysis technique may not detect all the aggregate sizes or types that are present Separation methods (SEC, sedimentation velocity) typically will not resolve individual oligomer species for a system in rapidly- reversible association equilibrium for associations to oligomers larger than dimer, from theory multiple peaks may be seen even for infinitely-fast kinetics, but those peaks usually do not represent single oligomers what is detected may depend on the rates of association and dissociation compared to the speed of separation
  9. 9. The measurement technique itself may perturb the distribution of aggregates that was initially present dissociation or loss of aggregates can be caused by: 1. dilution 2.change of solvent conditions 3.adsorption to surfaces (e.g. column resin) 4.physical filtration (e.g. column frit) 5.physical disruption by shear forces creation of new aggregates can be caused by: 1. change of solvent conditions 2.surface or shear-induced denaturation 3.concentration on surface of column resin (IEX) ← SEC ← SEC ← SEC ← SEC ← SEC
  10. 10. SEC problems & issues with the industry workhorse
  11. 11. Some shortcomings of SEC 1. change in solvent may change the aggregate distribution 2. limited range of sizes; larger aggregates often unresolved 3. often does not detect rapidly-reversible non-covalent association 4. aggregates may be filtered out by column 5. elution position not reliable indicator of molecular mass 6. limited sensitivity; difficult to resolve and detect species at or below ~0.1%
  12. 12. SEC issue #1: solvent-induced changes Achieving good resolution and high recovery often requires extreme solvent conditions that can alter the distribution of non- covalent aggregates high ionic strength addition of organic modifiers (alcohol, acetonitrile) I have seen cases where the SEC elution buffer completely dissociated the non-covalent aggregates, and also cases where it induced substantial amounts of new aggregates.
  13. 13. 1. want to add organics or salts to improve resolution, recovery, peak symmetry 2. want to pre-dilute the samples with the elution buffer 3. want to validate the method by spiking with pure aggregate but you cannot a pure sample of non-covalent aggregates end up optimizing and validating for covalent aggregates “Good” chromatography is often in direct conflict with good relevance for measuring non-covalent aggregates! ← this can cause false conclusions! The things chromatographers are taught to do to achieve “good” and robust methods often exacerbate this problem!
  14. 14. 1. want to add organics or salts to improve recovery and resolution 2. want to pre-dilute the samples with the elution buffer 3. want to validate the method by spiking with pure aggregate but you cannot a pure sample of non-covalent aggregates end up optimizing and validating for covalent aggregates “Good” chromatography is often in direct conflict with good relevance for measuring non-covalent aggregates! ← this can cause false conclusions! The things chromatographers are taught to do to achieve “good” and robust methods often exacerbate this problem!
  15. 15. SEC issue #2: limited size range Often the largest aggregates are not resolved, and elute in a single peak or shoulder at or near the exclusion limit Therefore we often cannot tell whether the large aggregate species present in different formulations or different manufacturing lots are actually similar in size Thus SEC alone may not distinguish samples with significant differences (e.g. immunogenicity)
  16. 16. SEC issue #5: incorrect masses inaccuracy due to dependence on molecular shape and/or undesirable interactions with the column matrix many, many examples of wrong stoichiometry for native proteins absolute accuracy may be unimportant; often all we want is fraction main peak But is that “dimer” peak really a dimer, or is it a partially-unfolded monomer (which might be much more immunogenic)?
  17. 17. Analytical Ultracentrifugation (AUC) sedimentation equilibrium sedimentation velocity
  18. 18. Both AUC methods are sensitive to all types of aggregates, but… Sedimentation equilibrium is primarily a thermodynamic tool for studying rapidly- reversible self-association (equilibrium constants and stoichiometry) Sedimentation velocity is much more a separation method and is particularly useful for characterizing irreversible (covalent or non-covalent) and relatively stable reversible non-covalent aggregates
  19. 19. Both sedimentation methods are “first principle” methods based on fundamental physical laws theory is well understood true for dilute solutions; concentrations > 10 mg/ml become complex and difficult require no standard molecules for calibration calibration is based only on fundamental units of distance, time, and temperature
  20. 20. Sedimentation velocity
  21. 21. The fundamentals of sedimentation velocity 6.0 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 7.0 Radius (cm) -0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 Absorbance 6.0 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 7.0 Radius (cm) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Absorbance centrifugal force diffusion ← meniscus The sedimentation coefficient is determined from the boundary motion over time. It depends on both molecular weight and shape. cell base → friction ←regionofsolute depletion boundary
  22. 22. High resolution analysis of a highly stressed antibody sample resolves 6 aggregate peaks plus 2 fragments 0 2 4 6 8 10 12 14 16 18 20 22 24 0.0 0.2 0.4 0.6 0.8 1.0 heptamer,0.1% hexamer,0.4% pentamer1.4% tetramer5.3% trimer14.6% dimer30.6% main peak (monomer), 45.5% ?HLhalfmolecule,0.8% ?freelightchain,1.4% c(s),normalized(totalarea=1) sedimentation coefficient (Svedbergs)
  23. 23. Comparability of a monoclonal antibody; detecting aggregate peaks at levels below 0.05% 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 0 1 2 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 0.00 0.01 0.02 normalizedc(s) sedimentation coefficient (Svedbergs) 0.42% 0.10% 0.03% 0.30% lot 1 lot 2 X 100 0.95% 0.05% 0.07%
  24. 24. This interferon-β sample is 13.7% non-covalent aggregate; by SEC (in 30% acetonitrile + 0.2% TFA) it would be pure monomer 0 2 4 6 8 10 0 1 2 3 4 5 6 7 0 2 4 6 8 10 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 IFN-β in 5 mM glycine, pH 3, 86.3% main peak c(s) sedimentation coefficient (Svedbergs)
  25. 25. 0 1 2 3 0.0 0.5 1.0 1.5 0 8 16 24 32 40 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 no salt c(s) +50 mM NaCl c(s) +150 mM NaCl c(s) sedimentation coefficient (Svedbergs) 0 2 4 6 8 10 0.00 0.05 0.10 0.15 20X expanded 0 2 4 6 8 10 0.00 0.05 0.10 0.15 20X expanded Adding NaCl to interferon-β formulations leads to a broad distribution non-covalent aggregates out to ~100-mers
  26. 26. We must study a range of concentrations to check for reversible association; this is a monomer-dimer-tetramer system 0 1 2 3 4 5 6 7 8 9 10 0 2 4 stock 4-fold dilution normalizedc(s) s20,w (Svedbergs)
  27. 27. Strengths of sedimentation velocity 1. high resolution (generally better than SEC) 2. covers very large range of masses in a single experiment (much larger than SEC) 3. detects both covalent and non-covalent aggregates 4. generally can be done directly in formulation buffers 5. little dilution of sample (~25%) 6. strong theory; “first principles” method
  28. 28. Weaknesses of sedimentation velocity 1. low throughput; often 3-7 samples/day 2. equipment and data analysis not automated like HPLC; labor intensive 3. expensive equipment (~250-300 K$) 4. requires substantial training Sedimentation velocity can not replace SEC, but it is an excellent tool to test whether SEC is missing important features. It can also serve as a “gold standard” to help improve SEC methods.
  29. 29. Non-covalent self-association by sedimentation equilibrium
  30. 30. 6.30 6.35 6.40 6.45 6.50 6.55 6.60 0.0 0.2 0.4 0.6 0.8 1.0 cell base Absorbance radius (cm) diffusion buoyancysedimentation meniscus The fundamentals of sedimentation equilibrium The concentration distribution depends only on molecular weight, independent of shape! ← smaller sample size to reduce time to reach equilibrium →
  31. 31. Size-exclusion chromatography of a TNF homolog 6 8 10 12 14 16 18 20 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0absorbance(arb.units) elution volume (ml) ovalbumin TNF homolog, elutes exactly as expected for a 17 kDa monomer
  32. 32. Linearized plot of equilibrium data for the TNF homolog -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 -1.9 -1.8 -1.7 -1.6 -1.5 -1.4 -1.3 -1.2 -1.1 -1.0 -0.9 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 Ln(Absorbanceat230nm) (r2 - ro 2 ) / 2 theoretical slope for monomer theoretical slope for trimer
  33. 33. Apparent mass vs. concentration for the soluble extracellular domain of the atrial natriuretic peptide receptor (monomer mass 58 kDa) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20 0.22 Concentration (mg/ml) 60000 70000 80000 90000 100000 110000 ApparentMass(Da)
  34. 34. Global analysis of all 18 samples gives a good fit to a monomer-dimer association model with Kd = 520 +/- 20 nM (∆G = -8570 +/- 25 cal/mol) -1.0-0.9-0.8-0.7-0.6-0.5-0.4-0.3-0.2-0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 (R^2 - Ro^2) / 2 (cm^2) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 Absorbance
  35. 35. Strengths/weaknesses of sedimentation equilibrium Strengths 1. equilibrium method; all forms of association are detected 2. wide choice of solvent conditions 3. strong theory; “first principles” method Weaknesses 1. difficult to quantitate irreversible aggregates 2. low throughput (9-21 samples/day typical) 3. expensive equipment 4. difficult data analysis 5. requires highly trained personnel
  36. 36. How can we validate AUC methods for non-covalent aggregates given we can’t spike in pure species? In part these methods are validated by over 60 years of experience and tens of thousands of publications More significantly, we have a theory, and we can create test data sets in silico (including realistic noise levels) and submit them to the analysis software to evaluate the ability to quantitate various species
  37. 37. “Classical” or “static” light scattering used on-line with SEC
  38. 38. Typical setup for size-exclusion chromatography with on-line light scattering detection light scattering detector absorbance detector solvent pump injector refractive index detector size-exclusion column
  39. 39. Getting molecular mass from static light scattering: the basic idea the light scattering signal is proportional to the product c × M we measure c simultaneously with a UV or RI detector then the ratio of the scattering to concentration signals will be proportional to M masses obtained this way are absolute, and independent of conformation and elution position
  40. 40. Demonstrating that scattering is independent of elution position and molecular conformation: the ratio of LS to RI signals is the same even for an unfolded protein 18 20 22 24 26 28 signal(arbitraryunits) LS RI LS RI Native RNase Unfolded (reduced & carboxymethlyated) RNase Retention Time (min)
  41. 41. An example for an Fc-fusion protein: the aggregate signals are much stronger in 90° scattering than in the UV chromatogram 4 5 6 7 8 9 10 11 12 Elution Volume (ml) 0 1 2 3 4 5 6 UVSignal(volts) 90° light scattering signal (scaled) UV (280 nm)
  42. 42. “Oligomer hunting”: display the absolute molecular weight from LS in units of monomers 6.2 6.4 6.6 6.8 7.0 7.2 7.4 7.6 7.8 8.0 8.2 8.4 8.6 8.8 9.0 9.2 9.4 9.6 9.8 10.0 Elution Volume (ml) 0 1 2 3 4 5 6 7 8 9 10 relativeMWfromLS/UV relative MW (LS/UV) UV (arb units)
  43. 43. This antibody sample has traces of dimer and trimer 12 13 14 15 16 17 18 Elution Volume (ml) 0.0 1.0 2.0 3.0 4.0 5.0 relativemassfromLS/UV relative mass (LS/UV) UV (arb units)
  44. 44. A different lot contains more higher oligomers, and they are so sticky that even dimer is no longer resolved 11 12 13 14 15 16 17 18 Elution Volume (ml) 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 relativemassfromLS/UV relative mass (LS/UV) UV (arb units)
  45. 45. 0.0 5 1.0x10 5 2.0x10 5 3.0x10 5 4.0x10 12.0 13.0 14.0 15.0 16.0 MolecularWeight Volume (mL) 81221H_1 or extended dimer dimer monomer SEC/LS shows an antibody sample contains only monomer and dimer, but dimer elutes at more than one position
  46. 46. Strengths of SEC + classical LS 1. absolute molecular mass, independent of conformation or elution position 2. gives us at least an average mass for the “aggregate” fraction near the exclusion limit 3. helps tell us whether our chromatography is really working properly 4. strong theoretical background; “first principles” method 5. high throughput, low cost (less than the HPLC it is used with), fairly easy
  47. 47. Weaknesses of SEC + classical LS 1. it inherits all the problems of SEC (change in aggregate distribution from dilution, change in buffer, adsorption/filtration, etc.) 2. while it is very sensitive to high MW aggregates, quantitation of % by weight still relies on the concentration detector (RI or UV) 3. particles shed from columns may obscure the region near the exclusion limit 4. good signal/noise may require larger injection amounts than normally used
  48. 48. Batch-mode dynamic light scattering (DLS) also known as quasi-elastic light scattering (QELS) or photon correlation spectroscopy (PCS)
  49. 49. Often when proteins go bad they develop “snow” (a.k.a. “white amorphous material” [WAM] or “floaters”) • may only appear after many months • often nucleation-controlled reaction • often ≤ 0.01% of total protein
  50. 50. When this happens our valuable protein can only be used for… ⇒ Dynamic scattering is one of the few tools that may be able to detect the precursors that eventually form ‘snow’
  51. 51. Dynamic light scattering: the basic idea 1. In dynamic scattering we measure the fluctuations in scattering intensity 2. The time scale of those fluctuations depends on the diffusion coefficient of the macromolecule, which in turn depends on its size 3. Like classical LS, the intensity of scattered light is proportional to M, so the sensitivity to very large aggregates is very high
  52. 52. Typically the data are transformed into a distribution of hydrodynamic radius; this distribution shows 2 well-resolved peaks 2.39 nm, 86.6% of intensity 92.3 nm, 13.4% of intensity 0.012% of mass 99.988% of mass
  53. 53. A different lot has 3 aggregate peaks, but they still represent only ~0.3% by weight <0.0001% 99.7% of mass 0.3% 0.008%
  54. 54. DLS drawback 1: Poor quantitation of mass fractions Usually at best the reproducibility of mass fractions is only within a factor of 2 There is no universally-accepted standard algorithm; different methods can give quite divergent results
  55. 55. DLS drawback 2: Low Resolution Generally, to be resolved as a separate species, a second component must be > 2-fold different in Rh and thus > 8-fold in mass Thus small oligomers are not resolved, and this is generally a poor method for detecting them This limited resolution may simply not be good enough to tell you what you need to know
  56. 56. Although small aggregates are generally not resolved as separate species, they do shift the distribution to higher sizes 4 6 8 10 12 0 2 4 6 8 10 12 14 "good" sample "bad" sample, >10% aggregates %Intensity Hydrodynamic radius, nm
  57. 57. DLS drawback 3: “Blinded by the light” It can be difficult to detect the main component in the glare from large aggregates; if you lose the main peak, you can’t quantitate fractions 8.5% of intensity, 96% of mass as supplied after centrifugation
  58. 58. Strengths of DLS 1. high sensitivity to large aggregates that may be immunogenic and/or precursors to visible particulates 2. covers an enormous range of size in one analysis (range of mass > 109) 3. done at equilibrium; theoretically senses all forms of aggregates 4. batch mode no dilution no change of solvent conditions no loss of species to frit or column matrix
  59. 59. Despite its limitations, DLS can be quite useful for: 1. detecting large aggregates at levels below 0.01% 2. tracking down which process steps generate large aggregates 3. relative ranking of different formulations or processes (which is better) 4. accelerated stability analysis done in situ in the DLS instrument
  60. 60. Recommended applications/approaches 1. Use all these orthogonal approaches to test whether your SEC method is missing anything that is significant 2. Use sedimentation velocity and on-line static LS to help develop better SEC methods 3. Trace onset of damage during manufacturing using DLS 4. For products formulated at high protein concentrations, dilute into PBS and run sedimentation velocity to detect long-lived aggregates that may persist in vivo [“dilute and shoot” protocol]
  61. 61. With thanks to the people who made the proteins 1. Several clients who allow me to show data for “protein X” 2. Kunio Misono at Cleveland Clinic Foundation (sANPR) 3. former Amgen colleagues

×