A Quality-by-design Approach to Upstream Bioprocess

Online PAT measurement and ... cell viability, substrates concentrations, ... Cell Culture CHO cells were cultivated in a glucose-...

0 downloads 54 Views 848KB Size
Manufacturing

A Quality-by-design Approach to Upstream Bioprocess Interrogation and Intensification Efficient biopharmaceutical process development relies on the quality-bydesign (QbD) paradigm. QbD is a scientific, risk-based proactive approach to drug development that aims to have a full understanding of how the process and product are related. This knowledge is gained by process analytical technology (PAT). In this case study the Applied Process Company (APC) integrated external PAT and an APC-developed controller with a parallel bioreactor system. Online PAT measurement and control of the identified critical process parameters led to greater understanding and the streamlined optimisation of a mammalian cell bioprocess. The study exemplifies the value of flexible bioreactor systems which allow ease of integration of third-party technology. Introduction Biopharmaceuticals represent a significant and growing sector of the pharmaceutical industry. The global biopharmaceutical industry is currently worth over 107 billion euros, according to research conducted by BioPlan Associates1. However, a lack of understanding of the process and product interdependencies potentially results in manufacturing challenges. Optimisation of protein production based on empirical experimentation and quality testing of the end product often results in bioprocesses operating under suboptimal conditions and hence in extended cycle times, excessive raw material and utility requirements, and elevated numbers of process or product failures, which together culminate in a high cost of manufacturing. To ensure consistent drug quality and reduce batch-to-batch variability, pharmaceutical companies and regulatory authorities aim to replace the “quality-by-testing” with a “quality-bydesign” strategy. In a QbD approach, critical quality attributes of the end product are defined and subsequently a production process is developed which aims at meeting those attributes. This involves identification of a design space, a range of experimental conditions within which variations in process parameters do not affect product quality. 78 INTERNATIONAL PHARMACEUTICAL INDUSTRY

IPI_December2015.indd 78

Prerequisite is the identification of critical process parameters by PAT that in the course of protein manufacturing are then monitored and, if needed, controlled. The PAT required to support a QbD product may be simple or complex, depending on the nature of the product and process used for its manufacture. Generally speaking, mammalian cell culture processes and the products which they produce are complex and hence, there are many challenges to implementing a PAT/QbD strategy. However, the potential benefits coupled with regulatory pressures are driving a strong interest in and an increased level of adoption in the biopharmaceutical industry. Biological processes are known to have a much higher level of variability than their chemical counterparts. PAT and QbD offer an avenue to better understand and control this variability, which has implications for the process economics, control and final product quality. Traditionally, temperature, pH, and dissolved oxygen are the main parameters measured and controlled online due to the availability of traditional, robust sensors. However, as technology advances, other process parameters such as cell density, cell viability, substrates concentrations, product and by-product concentration, dissolved carbon dioxide and biomarkers can now be measured and analysed in real time in an automated manner, although not yet routinely. The availability of realtime process information is of particular value in addressing the variability and unpredictability of animal cell cultures as it opens up the possibility of implementing advanced control strategies capable of directly impacting critical process parameters and critical quality attributes.

APC delivers chemical and bioprocess engineering solutions and technologies to enable streamlined development, optimisation, and supply of new and existing chemical and biological entities. APC’s process development technology platform (BioACHIEVE™) combines PAT technology, multivariate data analysis, process modelling, and advanced control strategies2. This article describes the successful integration of external PAT with a parallel bioreactor system using an object linking and embedding for process control (OPC) communication protocol. The aim was to optimise the performance of a Chinese hamster ovary (CHO) mammalian cell bioprocess. In a previous study the glucose concentration was identified as a critical process parameter3. The objective of the following case study was to improve process performance by optimising the glucose feed profile. By moving from the traditional bolus fed-batch to a continuous feeding fed-batch strategy, nutrient depletion should be prevented and a stable macro-environment for the cells should be established. Material and Methods Cell Culture CHO cells were cultivated in a glucosefree formulation of EX-CELL® CHO DHFRmedium (Sigma-Aldrich® Co., LLC, USA) supplemented with 20 mM glucose, 4 mM glutamine, 1 µM methotrexate, 0.1% (v/v) Pluronic® F-68 (SigmaAldrich Co., LLC, USA) and 10 mL/L penicillin-streptomycin (Sigma-Aldrich Co., LCC, USA). Cells were cultivated for 7 to 9 days at 37°C using a four-fold DASGIP® Parallel Bioreactor System for

Fig. 1: Eppendorf DASGIP Parallel Bioreactor System for cell culture. Winter 2015 Volume 7 Issue 4

23/12/2015 15:14:29

Manufacturing

80 INTERNATIONAL PHARMACEUTICAL INDUSTRY

IPI_December2015.indd 80

with an Eppendorf DASGIP Parallel Bioreactor System. The glucose concentration in the CHO mammalian cell culture was determined online using Raman spectroscopy coupled with chemometric partial least squared modelling. The offline results determined via an enzymatic assay were highly comparable to the results obtained online by Raman spectroscopy (Figures 3A, 3B). Overall, the continuous feeding fed-batch bioprocess resulted in an increase in peak viable cell density and the integral of the viable cell density (Figures 3C, 3D) which is directly related to increased titre. A Glucose concentration (mM)

30

Online

25

Offline 20 15 10 5 0 20

40

60

80

100

120

140

Time (h)

30

Glucose concentration (mM)

B

Online

25

Offline 20 15 10 5 0 0

20

40

60

80

100

120

140

Time (h)

100

Cell density (cells/mL)

7x106

90

6x106

80

5x106

70 60

4x106

50 3x106 2x10

Viability (%)

C

40 30

6

20

1x106

10

0

0 0

20

40

60

80

100

120

140

Time (h)

100

7x106

90

6x106

80

5x106

70 60

4x106

50 3x106

Viability (%)

D Cell density (cells/mL)

close to the setpoint, an APCdeveloped model Eppendorf Eppendorf predictive controller DASGIP MP8 DASware control Multi Pump Module (MPC) algorithm Adding glucose was integrated via Calculation of feed rate bidirectional OPC Feedback communication. Loop The essence of Eppendorf MPC is to optimise APC-developed DASGIP Model Predictive Controller Bioreactor forecasts of process Algorithm behaviour. This Quantification of forecasting is Raman measurement glucose concentration Kaiser Optical Systems accomplished with RamanRxn2 Multi-channel Raman analyzer a process model, and, therefore, Fig. 2: Integration of external PAT with an Eppendorf DASGIP the model is an Parallel Bioreactor System. essential element of an MPC controller. cell culture equipped with 2.5 L (working Based on online volume) DASGIP® Benchtop Bioreactors readings of the identified critical process (Eppendorf AG, Germany) (Figure 1). parameters, the model is used to Cultures were agitated at 120 rpm. The calculate optimal feed rates for set-point pH was controlled at 7.2 using 1 M maintenance. MPC facilitates the ability NaOH and CO2 gas, respectively. Air to proactively counteract deviations saturation was set to 50 % and controlled from the set-point and to simultaneously using air and oxygen supplied via a control multiple process parameters, sparger. and is hence especially suited to ensure constant culture conditions in multivariate Offline Measurement of Process bioprocesses. Parameters The MPC actuated a DASGIP MP8 multiCell density and cell viability were pump module to execute continuous determined offline using a Cedex® HiRes glucose feeding. system (Roche Diagnostics® GmbH, Germany). Bolus Fed-batch and Continuous FedOffline measurements of glucose batch Glucose Feeding concentrations were performed using The glucose concentration was adjusted an enzymatic assay kit (Megazyme® using a feed medium consisting of International Ireland Ltd). 653.6 mM glucose, 58.8 mM glutamine and 58.8 g/L soy protein hydrolysate Integration of External PAT and In-house dissolved in glucose-free EX-CELL CHO Developed Controllers DHFR- medium. Figure 2 illustrates the integration of external PAT and an in-house developed Two fed-batch feeding strategies were controller with a parallel bioreactor investigated in the study. The first was a system. Online measurements of glucose fed-batch culture manually fed with bolus concentration were performed by Raman additions, at 24 h intervals, the volume spectroscopy. A RamanRxn2 Multi- of which was proportional to offline channel Raman analyser (Kaiser Optical integral viable cell density measurements Systems. Inc., USA) was integrated with for the previous 24 h interval. The second the bioreactor system using DASware regime was a continuous feed, the rate analyze software. DASware analyze of which was determined and adjusted utilises an OPC communication protocol, automatically using a model predictive an industry standard which facilitates the controller and Raman-determined communication of devices from different glucose values to keep the glucose manufacturers. To translate the Raman concentration in the bioreactor at a setspectra into concentration information, point of 11 mM throughout the culture. chemometric partial least squared calibration models were developed using Results and Discussion the SIMCA® multivariate data analysis Using the DASware analyze software package (MKS Umetrics AB, Sweden). a RamanRxn2 Multi-channel Raman analyser and an APC-developed MPC To maintain the glucose concentration algorithm were successfully integrated Flowrate control

40 30

2x106

20

1x106

10

0

0 0

20

40

60

80

100

120

140

Time (h)

Fig. 3: Glucose set-point control via a model predictive controller algorithm resulted in an increase in cell densities. A, B: Glucose concentration was measured online and offline, respectively. (A) Glucose concentration was adjusted via a bolus fed-batch feeding strategy. (B) The glucose set-point was maintained in a continuous feeding fed-batch strategy using MPC. C, D: Total cell densities (TCD), viable cell densities (VCD), and viability were determined for the bolus fed-batch culture (C) and the continuous feeding fed-batch culture (D). Winter 2015 Volume 7 Issue 4

23/12/2015 15:14:32

DEC_IPT_105x297+3_05-2015 17.06.15 13:19 Seite 1

Manufacturing Conclusion The parallel bioreactor system used was a suitable host for APC’s BIOACHIEVE process development technology platform, which APC applies to upstream bioprocesses. The DASGIP system facilitated the ease of integration of external PAT and transfer of this information to and from APC-developed controllers. This allowed optimisation of the process performance based on greater process understanding, ultimately delivering improvements in cell growth. This case study exemplifies the value of increased process understanding and control in biopharmaceutical manufacturing. Advancements in sensor technology and data analysis facilitate the ability to control the identified critical process parameters to their respective target levels and thus optimise the critical quality attribute design space. To apply the QbD approach to upstream bioprocess development, bioreactor systems which allow smooth integration of external sensors and controllers are crucial.

Are your plants ready

References 1. BioPlan Associates Inc. 11th Annual Report and Survey of Biopharmaceutical Manufacturing Capacity and Production. ISBN 978-1-934106-24-2 (2014) 2. APC Ltd. Bioprocess development and optimization: A technology enabled approach to process development. APC White Paper (2013) 3. Craven, S. A quality-by-design approach to upstream bioprocess development. The Engineers Journal (2014)

to swiftly respond to changing markets and production needs, reliably provide top quality drugs while meeting regulatory compliance?

Stephen Craven, APC Ltd, Dublin, Ireland Dr Stephen Craven earned his PhD from the School of Chemical and Bioprocess Engineering, University College Dublin, where he developed bioprocess models for mammalian cell fermentations and also developed and applied advanced control strategies to PAT-enabled bioprocesses. Dr Craven currently works as the life sciences team leader within APC

We can help you to have them. Contact us.

Ltd, Dublin. Email: [email protected]

Ulrike Becken, Eppendorf AG Bioprocess Center, Juelich, Germany Dr Ulrike Becken works as a Scientific Communication Manager at the Eppendorf AG Bioprocess Center in Juelich, Germany. Amongst others she is responsible for authoring application notes in cooperation with customers. Dr Becken earned her PhD from the Rheinische-FriedrichWilhelms University, Bonn, in the field of cell biology. Email: [email protected] 81 INTERNATIONAL PHARMACEUTICAL INDUSTRY

IPI_December2015.indd 81

Winter 2015 Volume 7 Issue 4

23/12/2015 15:14:33