Screen early - fail early: Rapidanalytical screens for protein formulation

Paul Dalby

Dept. Biochemical Engineering, UCL

EPSRC Centre for Innovative Manufacturing

Creating manufacturing innovations so as to deliver affordable nextgeneration advanced therapies to the UK healthcare system

Biopharmaceutical Lifecycle Optimisation

2

2

Optimisemanufacturability

and lifecycle costs

Complexmacromolecular

candidates

Healthcaredelivery to

patient

Affordableadvancedtherapies

Project Partners. Industry Associations: ABPI, BIA, HealthTech & Medicines KTN. Companies: Aegis Analytical, Avacta, BioPharm Services, BTG Plc,Eli Lilly, Francis BioPharma Consulting, Fujifilm Diosynth Biotechnologies, GE Healthcare, GlaxoSmithKline, Lonza Biologics, MedImmune, Merck & Co,MSD Biologics, Novo Nordisk, Novozymes Biopharma, Office of Health Economics (OHE), Pfizer, Syntaxin, The Automation Partnership (TAP), UCBPharma. Government-related: Health Protection Agency, NIBSC, NHS QIPP.

EPSRC Centre for Innovative Manufacturing

• Collaboration initiated by an academic core:-

UCL – Biochemical Engineering (Lead)

Chemical Engineering

Health Economics

Centre Team & Consortium

3

Health Economics

LSoP – Formulation Engineering

ICL – Chemical Engineering

• Supported by a group of:-

25 industrial users including SMEs,

7 NGOs / Industry Associations

• Extended by a network of:-

23 national & international academics

When does formulation start?

Fermentation

Secondaryoperation

RecoveryBiomaterialparameters

Operationalparameters

Purification

Formulation

Product

Paradigms to improve “formulatability”

1. “Screen early: fail early”

• Find "troublemaker" proteins earlier

• Provide appropriate “stress tests" for formulation engineers

2. “Better by design”2. “Better by design”

• Develop robustly manufacturable protein scaffolds

• Establish predictive protein design evaluation tools

• Understand protein aggregation better

Formulation development

Design of Experiment (DoE) - Stress and analyse

Protein variants [Protein] pH

[Buffer] & type Ionic strength GRAS excipients

Design space

“Optimum” formulation

High T, low t? Freeze-thaw? Tm or Tagg? Agitation?

Shelf-life study

High end biophysics

Biophysical analysis of proteins

•Size•Structure content•Composition•Folding extent

•Chemical modifications•Local structure content•Shape and conformation•Folding dynamics

•Atomic structure•Bond formation•Molecular interactions•Structural dynamics

N N*

CDFluorescence spectroscopyDLS / particle imagingSize exclusion chromatographyAnalytical Ultracentrifugation

Low resolution Medium resolution High resolution

LCMSSmall angle X-ray scattering

NMR

•Folding extent •Folding dynamics •Structural dynamics

Challenges and possible solutions

• Too little material available at early bioprocess development stages

• Design space is very large for new entities

• Formulations are at high concentration (10-200 mg/ml)

• Many biophysical analyses use 0.1-2ml, 0.01-1 mg/ml

• Forced degradation is not the same as unforced degradation• Forced degradation is not the same as unforced degradation- which degradation species are a problem?

• Predict outside of measurement range

• Create lower volume analytics

• Higher throughput with high accuracy and high sensitivity

• Improve predictability of shelf-life and degradation pathways

Lyophilization Cycle

-50

-40

-30

-20

-10

0

10

20

30

Te

mp

era

ture

(ce

lciu

s)

0

100

200

300

400

500

0 120 240 360 480 600 720 840 960 1080 1200

Time (mins)

Pre

ss

ure

(mT

orr

)Shelf TemperatureVacuum

Freeze-drying in microplates

Rate of sublimation in microplate Thermal imaging of microplate freezing

1 2 3 4 5 6 7 8 9 10 11 12A

B

C

D

E

F

G

H

Rate of sublimation in microplate

135-145

125-135

115-125

105-115

95-105

85-95

Thermal imaging of microplate freezing

Grant Y, Matejtschuk P, Dalby PA (2009) Biotech. Bioeng. 104:957-964. Rapid optimisation of protein freeze-drying formulationsusing ultra scale-down and factorial design of experiment in microplates.

Freeze-drying optimisation - GCSF

0.1 ug/ml GCSF, pH 7

- stimulates white blood cell production- improves recovery post-chemotherapy- assayed by cell count after growth stimulation

Scott Grant & Paul Matejtschuk (NIBSC)

TrehaloseHSA

MannitolSucrose

HSAMannitol

Tween 20

0.33% Tween

3% HSA

Freeze-drying optimisation - GCSF

• Took two weeks and used 370 ng GCSF!

SucroseTween 20Arginine

Phenylalanine

Tween 20100% Activityretention

Scott Grant & Paul Matejtschuk (NIBSC)

GCSF forced degradation studies

0

50

100

-1 1 3 5 7

No. Days Incubated

Re

lati

ve

Ac

tiv

ity Initial comparison to

unformulated GCSF

♦ lyophilized formulation■ unlyophilized formulation▲ PBS

Resuspended lyophilisates in vials: after 6 month storage at various T

-7.5

-7

-6.5

-6

-5.5

-5

-4.5

-4

-3.5

-3

0.003 0.0032 0.0034 0.0036

1/T (K-1)

ln(k

/T)

0

20

40

60

80

100

120

-100 -50 0 50 100

T (degC)

%G

CS

Fac

tiv

ity

reta

ine

d

Resuspended lyophilisates in vials: after 6 month storage at various T

Scott Grant & Paul Matejtschuk (NIBSC)

Equilibrium denaturation:

• Capillary DSC (eg. MicroCal)

• Autotitrating fluorimeter/circular dichroism

Some thermostability methods

X

Flu

ore

sc

en

ce

• ANS/SYPRO binding (eg. ThermoFluor / DSF)

Equilibrium exchange kinetics:

• MALDI-TOF• NMR H/D exchange

Early microplate-based screens

Protein precipitation / solubility

Wild typeY440AD381A

Protein stability

Ahmad SS, Dalby PA (2011) Biotechnol. Bioeng. 108:322-332.Thermodynamic parameters for salt-induced reversible proteinprecipitation from automated microscale experiments.

Aucamp, J. P. (2005) Biotech Bioeng. 89 599-607

Aucamp, J. P. (2008) Biotech Bioeng. 99, 1303-1310

280 nm 340 nm

Denaturantin

autotitrator

Effect of excipients on G1/2

0

1

2

3

4

5

6

Wate

r

La

cto

se

So

rbito

l

Su

cro

se

Tre

ha

lose

K2

PH

4

Na

Cl

TM

AO

Tw

een

20

Tw

een

80

Arg

inin

e

Eth

an

ol

Gly

ce

rol

K2

PH

410

0

Na

Cl

100

De

xtra

n

Eth

yle

ne

Gly

col

Gly

cin

e

Pe

g4

00

Pe

g2

000

Excipient

Mid

-po

pin

t(M

)

1.0

1.5

2.0

2.5

3.0

3.5

2 4 6 8 10 12pH

Mid

po

int

[Gd

mC

l](M

)

Serial dilution

Literature

RNaseA in 50 mM formate pH3, MOPS pH7 or glycine pH9.9Literature values from Pace et al 1990 Biochemistry 29:2564-2572

Protein stability in microfluidics

100mm I.D.

10mL/min

Gaudet M, Remtulla N, Jackson SE, Main ERG, Bracewell DG, Aeppli G, Dalby PA (2010) Protein Science. 19: 1544-1554.Protein denaturation and protein:drug interactions from intrinsic protein fluorescence measurements at the nanolitre scale.

• 266 nm, 5 mW, 1 kHz pulsed UV laser

• fused-silica glass micro-capillary (ID=100 mm, OD=300 mm)

• Emission filtered by 320-400 nm dichroic mirror

• Measurement volume of 1.5 nL minimum

PMTMeasurements

Signal processing

PDMeasurements

ProcessedDataset

• Beam splitter and photodiodemeasurement used as a reference

Measurement stability and cleaning

Sample flow removes optical bleaching

Acquiring the fluorescence measurement

1.5

2.0

2.5

3.0

Re

lativ

eF

luo

resce

nce

sample enters optics

• Signal is stable (photobleaching avoided)• Measurement is rapid (50-100ms to average 50-100 pulses)

0.0

0.5

1.0

0 0.2 0.4 0.6 0.8 1

Time (s)

Re

lativ

eF

luo

resce

nce

each datum is averaged from 50-100 laser pulses

Sensitivity of fluorescence measurement

• Signal response is linear: 0.15 mM to 1.5 mM (0.01-100 mg/ml) BSA

Matthieu Gaudet

Comparison to microplates

Method Limit ofdetection(mg/ml)

Minimum[Protein]

(uM)

Volume

(L)

Number ofProteins

Microplate

BSA 0.005 0.076 2.6x10-4 1.16 x 1013

RNaseA 0.0001 0.007 2.6x10-4 1.14 x 1012

Microfluidics

BSA 0.01 0.15 1.5x10-9 1.4 x 108

• 2x greater minimum concentration required• 80,000x less protein required• concentration range: 0.15 uM to 1.5 mM (0.01-100 mg/ml)

Matthieu Gaudet

Nanolitre stability & ligand affinity screening

100mm I.D.

10mL/min

[GdnHCl] (M)

0 1 2 3 4 5 6

Norm

alis

ed

fluo

resce

nce

0.0

0.2

0.4

0.6

0.8

1.0

+ Rapamycin

No Rapamycin

CFKBP=10uM; Crap= 15uM

open : capillary

full: microplate

Gaudet M, Remtulla N, Jackson SE, Main ERG, Bracewell DG, Aeppli G, Dalby PA (2010) Protein Science. 19: 1544-1554.

• 80,000x less protein than 96-well• 0.15 uM to 1.5 mM (0.01-100 mg/ml)• accurate ΔGD-N, C1/2, Kd

Optical control of heat in microfluidics

Acknowledgements

PhD/PDRA:Stability Jean Aucamp, Julio Martinez-Torres, Michael RosePrecipitation Shahina AmhadFreeze-drying Scott GrantMicrofluidics Matthieu Gaudet, Sagar Dodderi, Samir Aoudjane

Collaborators:Microfluidics Gabriel Aeppli (LCN)

Dan Bracewell (UCL)FKBP-12 Ewan Main (QMUL)FKBP-12 Ewan Main (QMUL)

Sophie Jackson (Cambridge)Freeze-drying Paul Matejtschuk (NIBSC/ HPA)

Funding & material donors:EPSRC (IMRC and EngD)BBSRC (BRIC, FOF and studentships)British Council (BC)Association of Commonwealth Universities (ACU)NIBSC (Health Protection Agency)