Sensors in Bioprocess ControlHarry Lam

Department of Manufacturing SciencesGenentech, Inc.

February 28, 2000

Genentech, Inc.

Harvard’s LawUnder the most rigorously controlled

conditions of pressure, temperature, volume, humidity, and other variables, the organism

will do as it darn well pleases

Translation …You put the organisms into the tanks and pray

Genentech, Inc.

Genentech, Inc.

Presentation Outline

The needs for process controlIndustrial bioprocessesCell and its environmentMeasurementsSensor requirementsExamplesConclusion and future directionsAcknowledgments

Genentech, Inc.

The Need for Process Control

To maintain consistent process performance (productivity, quality) throughout the development cycle (R&D to Manufacturing).

Genentech, Inc.

Consistent Process Performance

R&DExperiment to ExperimentScale to Scale

ManufacturingThaw to ThawRun to RunCampaign to CampaignPlant to Plant

Genentech, Inc.

Sensors for Bioprocess Control

Measurements for process or system analysis

Search for underlying functional relationshipsIn depth analysis of the interaction of the organisms with their environment

Provide capabilities for process controlSetting up and maintaining the optimum environmental conditions for growth and/or formation of product

Large-scale versus laboratory bioreactors

Genentech, Inc.

Generic Process Flow Diagram for Protein Pharmaceutical

Inoculum Preparation

Medium Preparation

Raw Materials

Inoculum Fermentor

Production Fermentor CellsOxygen

Crude Product

Product ConcentrationPurification

Purified Product

Purified Product Formulation Sterile

FiltrationFilling & Freeze Drying

Final Product

Viral Inactivation

Genentech, Inc.

Harvest

Harvest

PRODUCTION STAGE non-selective serum-free

Production MediumMCB WCB

PREQUALIFIED CELL AGE LIMIT (Master Cell Bank Thaw to Harvest)

Culture Fluid to Recovery

Culture Fluid to Recovery

Medium Exchange or Direct Transfer

Medium Exchange or Direct Transfer

TYPICAL LARGE-SCALE CELL CULTURE PROCESS

INOCULUM TRAIN non-selective

Inoculum Medium

SEED TRAIN

selective Seed Train Medium

with 2% FBS

(continuous passaging with methotrexate)

Nutrient Feeds

Nutrient Feeds

Wash Medium

Wash Medium

Genentech, Inc.

Scaling Up From 2-L to 12,000-L

Genentech, Inc.

Cell and Its EnvironmentCells are isolated from complex multicellular organisms

Homeostasis is maintained in these organisms by many specialized organs and tissues working synergistically.

Therefore mammalian cells do not have the capacity to maintain homeostasis by themselves

Cell culture is not a natural environment for mammalian cells.

Genentech, Inc.

Genentech, Inc.

Measurements

BiologicalChemicalPhysical

Genentech, Inc.

Biological Measurements

Cell DensityCell ViabilityCell SizeMorphologyCellular assaysMolecular/Genetic Assays

Genentech, Inc.

Chemical Measurements

MediaCarbohydrates - e.g. glucose, galactoseComplex medium – protein hydrolysates, yeast extracts, etc.Amino acidsSaltsLipids - Linoleic AcidHormones, growth factors (serum, insulin)VitaminsTrace elements - e.g. metals (Fe, Mn, etc.)AntifoamF-68AntibioticsMethotrexate

Genentech, Inc.

Chemical Measurements (cont.)

Product ConcentrationQuality

By-productsOrganic acids – acetate, lactate, etcProteinsAmmonia

Chemical environmentpHDissolved gases (dissolved O2, pO2, pCO2)Osmolality

Off-gasO2 (OUR)CO2 (CER)

Genentech, Inc.

Physical Measurements

TemperatureAgitationPressureLevel (volume)WeightBroth densityViscosity

Genentech, Inc.

Important Criteria for Sensors

Reliability, Accuracy, ReproducibilityLong-term Stability SpecificityResponse timeDynamic behavior Ability to be repeatedly cleaned and sterilizedEase of operationEase of maintenanceSizeCost

Genentech, Inc.

Modes of Bioprocess Monitoring

Callis JB, Illman DL, Kowalski BR. Process Analytical Anal Chem 1987;59:624A-637A.

Genentech, Inc.

Sampling Considerations

Representative sampleSample sizeSterility requirementsUtilities considerationsDisposal considerationsLiquid versus vapor samplesSample preservation

Genentech, Inc.

Requirements for in Situ Sensors

Fully cleanable and sterilizableGood thermal stability and compressive strength, no temperature hysteresis

Long-term stability and accuracyFast responseNo flow dependenceNo interference

Air bubbles (O2 and CO2) or by microbesComplex media

No foulingLow maintenanceSmall size

Genentech, Inc.

Sterile Sampling Designs

Sample locationSample withdrawal positionMethod of connectionfluid velocity profileContainment considerations

Sampler and container designMaterials of constructionProcess and sample variables

TemperaturePressureSlurry/two phasesviscosity

Genentech, Inc.

Sampling Designs (Cont.)

Sterilization optionsPurging considerationsVenting considerationsContainmentDesign options

Genentech, Inc.

Cell Density and Viability Measurements

Need to know the viable and nonviable cell density to evaluate growth rate, death rate and specific productivityDirect - Count the cells

Hemocytometer with Trypan Bluestaining - total viable and total nonviable cellsCoulter counting - total cell number

Indirect - Measure a factor which correlates with cell numberPacked Cell Volume (PCV) - correlation with cell numberOptical Density (OD)Dry weightTotal DNATotal proteinCellular enzymatic activityCellular metabolic activity

Genentech, Inc.

CHO Cell Biomass Estimation via Oxygen Uptake Rate Measurement

0

2

4

6

8

10

12

0 10 20 30 40Viable Cell Density(1E5 vc/ml)

Oxy

gen

Upt

ake

Rat

e (%

sat/m

in) QO2 = 0.30 ± 0.03 mmol O2/E9 vc-hr

Genentech, Inc.

Cellular Assays

Measure physiological or expressed parameters and will be useful in measuring “How” cell lines differ.

Bromodeoxyuridine (BrdU) cell cycleFluorescent methotrexate (F-Mtx) bindingSpecific productivityCell tracer

Genentech, Inc.

Cell cycle analysis with anti-BrdU and flow cytometry

S Phase

G1 G2

BrdU-FITC

BrdU

-FIT

CNeg.

Pos.

Dual parameterhistogram

Single parameterhistogram

Y

Y

G2/M2n DNA

G11n DNA

S-phase1->2n DNA

YY

Y YY

Y Y Y

YY

YY Y

Anti-BrdUMAb

BrdUIncorporation

Genentech, Inc.

Measurement of intracellular product

Cells are•Formaldehyde-fixed

•Detergent permeabilized

•Product is detected with FITC conjugated F(ab’)2

Neg control

Low producingcells

Normally producingcells

Genentech, Inc.

Rapid Loss of Intracellular Product Over Time

7 days

15 days 57 days

C lineB lineA line

Genentech, Inc.

Cell Culture pH Controlglucose, amino acids, vitamins, O2, …

----> Cells, CO2, Lactate, NH3, H2O, Product ...

Control ObjectivesMaintains desired pH with minimum osmolality riseEnsure consistent pH profiles from run to run

Typical Means of ControlUse acid source (CO2 gas) and base source (Na2CO3)Apply gap/deadband controller (± 0.03 pH units from setpoint)

Genentech, Inc.

Control of Dissolved Gases in Cell CulturesDuring aerobic growth, cells require O2 and produce CO2Cells sensitive to extremes of dissolved gas concentrations

Hypoxia (<1% of air saturation?)Hyperoxia (>100% of air saturation?)CO2 required for synthesis/energy metabolism reactionsExcess CO2 can inhibit respiration reactions, change intracellular pHMinimum required levels unknownCO2 levels may influence product characteristicsControl of culture pH, dO2, dCO2, pressure all inter-related

Genentech, Inc.

Theoretical Nutrient Depletion Times in E.coli Fermentation

OxygenAt 30% of air saturation, [O2] ~ 0.075 mMWith OUR = 5 mmoles/L-minDepletion Time = 1.8 seconds

GlucoseTo avoid acetate formation, [glucose] ~ Ks, glucKs, gluc ~ 20 µM = 3.6 mg/LDuring growth, glucose uptake rate ~ 2 mmoles/L-minDepletion Time ≤ 0.6 seconds

Typical mixing times = 12 to 50 seconds

Genentech, Inc.

Galactosylation Control via Controlled Nutrient Feeding

NH3 can influence glycosylation efficiency by increasing intracellular pH

Reduce waste product accumulation (i.e., NH3) by controlled feeding of nutrients which generate NH3

Use on-line biomass estimation to generate feed profile

Process Controller

Nutrient Uptake Model:

Bioreactor

Nutrient Feed

In-situ viable biomass probe

(capacitance, OUR, etc.) Mass Balance

Genentech, Inc.

Nutrient Feed Rate Based on Estimated Viable Cell Population

0

2

4

6

8

10

12

14

16

0 50 100 150 200Run Time (h)

Cap

acita

nce

(pF)

0

200

400

600

800

Total Fed (mL)

CapacitanceTotal Fed

2L Bioreactor

Genentech, Inc.

Controlled Nutrient Feed Results in Improved Galactosylation

0123456789

10

0 50 100 150 200Run Time (h)

Glu

tam

ine

(mM

) Up-Front GlnFed-Batch GlnFed-Batch Gln

0123456789

10

0 50 100 150 200

Run Time (h)

Am

mon

ium

(mM

)

(2L Bioreactors)

Genentech, Inc.

Controlled Nutrient Feed Results in Improved Galactosylation

0.30

0.40

0.50

0.60

0.70

0 50 100 150 200Run Time (h)

Gal

Con

tent

(mol

/HC

)

Up-Front GlnFed-Batch GlnFed-Batch Gln

0 50 100 150 200Run Time (h)

MA

b Ti

ter (

mg/

L)

Up-Front GlnFed-Batch GlnFed-Batch Gln

* No medium galactose present for these experiments *

(2L Bioreactors)

Genentech, Inc.

Immediately Applicable But Still Missing...

On-line in situ glucose (and other critical metabolites) sensorsOn-line viability measurement for cell cultureSimpler (for manufacturing use) population assessment toolsSoft sensorsAnalysis tools for retrospective modeling, troubleshooting and optimization

Genentech, Inc.

Conclusion and Future Directions

Improved sensor technology can provide the basis for new control strategiesIncrease our ability to run more consistent and more productive bioprocessesThe challenge is to develop sensors which can be readily implemented for more effective process control

Genentech, Inc.

Acknowledgments

John FrenzCynthia HoyTom IhrigBob KissJim Swartz