Table of Contents
Cover
Title Page
Copyright
Preface
List of Contributors
Chapter 1: Challenges for Bioreactor Design and Operation
1.1 Introduction
1.2 Biotechnology Milestones with Implications on Bioreactor Design
1.3 General Features of Bioreactor Design
1.4 Recent Trends in Designing and Operating Bioreactors
1.5 The Systems Biology Approach
1.6 Using Conceptual Design Methodology
1.7 An Outlook on Challenges for Bioreactor Design and Operation
References
Chapter 2: Design and Operation of Microbioreactor Systems for Screening and Process Development
2.1 Introduction
2.2 Key Engineering Parameters and Properties in Microbioreactor Design and Operation
2.3 Design of Novel Stirred and Bubble Aerated Microbioreactors
2.4 Robotics for Microbioreactors
2.5 Fed-Batch and Continuous Operation of Microbioreactors
2.6 Monitoring and Control of Microbioreactors
2.7 Conclusion
References
Chapter 3: Bioreactors on a Chip
3.1 Introduction
3.2 Advantages of Microsystems
3.3 Scaling Down the Bioreactor to the Microfluidic Format
3.4 Microfabrication Methods for Bioreactors-On-A-Chip
3.5 Fabrication Materials
3.6 Integrated Sensors for Key Bioreactor Parameters
3.7 Model Organisms Applied to BRoCs
3.8 Applications of Microfluidic Bioreactor Chip
3.9 Scale Up
3.10 Conclusion
References
Chapter 4: Scalable Manufacture for Cell Therapy Needs
4.1 Introduction
4.2 Requirements for Cell Therapy
4.3 Stem Cell Types and Products
4.4 Paradigms in Cell Therapy Manufacture
4.5 Cell Therapy Manufacturing Platforms
4.6 Microcarriers and Stirred-Tank Bioreactors
4.7 Future Trends for Microcarrier Culture
4.8 Preservation of Cell Therapy Products
4.9 Conclusions
References
Chapter 5: Artificial Liver Bioreactor Design
5.1 Need for Innovative Liver Therapies
5.2 Requirements to Liver Support Systems
5.3 Bioreactor Technologies Used in Clinical Trials
5.4 Optimization of Bioartificial Liver Bioreactor Designs
5.5 Improvement of Cell Biology in Bioartificial Livers
5.6 Bioreactors Enabling Cell Production for Transplantation
5.7 Cell Sources for Bioartificial Liver Bioreactors
5.8 Outlook
References
Chapter 6: Bioreactors for Expansion of Pluripotent Stem Cells and Their Differentiation to Cardiac Cells
6.1 Introduction
6.2 Culture Technologies for Pluripotent Stem Cell Expansion
6.3 3D Suspension Culture
6.4 Autologous Versus Allogeneic Cell Therapies: Practical and Economic Considerations for hPSC Processing
6.5 Upscaling hPSC Cardiomyogenic Differentiation in Bioreactors
6.6 Conclusion
References
Chapter 7: Culturing Entrapped Stem Cells in Continuous Bioreactors
7.1 Introduction
7.2 Materials Used in Stem Cell Entrapment
7.3 Synthetic Materials
7.4 Natural Materials
7.5 Manufacturing and Regulatory Constraints
7.6 Mass Transfer in the Entrapment Material
7.7 Continuous Bioreactors for Entrapped Stem Cell Culture
7.8 Future Perspectives
References
Chapter 8: Coping with Physiological Stress During Recombinant Protein Production by Bioreactor Design and Operation
8.1 Major Physiological Stress Factors in Recombinant Protein Production Processes
8.2 Monitoring Physiological Stress and Metabolic Load as a Tool for Bioprocess Design and Optimization
8.3 Design and Operation Strategies to Minimize/Overcome Problems Associated with Physiological Stress and Metabolic Load
8.4 Bioreactor Design Considerations to Minimize Shear Stress
Acknowledgments
References
Chapter 9: Design, Applications, and Development of Single-Use Bioreactors
9.1 Introduction
9.2 Design Challenges of Single-Use Bioreactors
9.3 Cell Culture Application
9.4 Microbial Application of Single-Use Bioreactors
9.5 Outlook
References
Chapter 10: Computational Fluid Dynamics for Bioreactor Design
10.1 Introduction
10.2 Multiphase Flows
10.3 Turbulent Flow
10.4 CFD Simulations
10.5 Case Studies for Application of CFD in Modeling of Bioreactors
Summary
References
Chapter 11: Scale-Up and Scale-Down Methodologies for Bioreactors
11.1 Introduction
11.2 Bioprocess Scale-Down Approaches
11.3 Characterization of the Large Scale
11.4 Computational Methods to Describe the Large Scale
11.5 Scale-Down Experiments and Physiological Responses
11.6 Outlook
References
Chapter 12: Integration of Bioreactors with Downstream Steps
12.1 Introduction
12.2 Improvements in Cell-Culture
12.3 Interactions with Centrifugation Steps
12.4 Interactions with Filtration Steps
12.5 Interactions with Chromatographic Steps
12.6 Integrated Processes
12.7 Integrated Models
12.8 Conclusions
References
Chapter 13: Multivariate Modeling for Bioreactor Monitoring and Control
13.1 Introduction
13.2 Analytical Measurement Methods for Bioreactor Monitoring
13.3 Multivariate Modeling Approaches
13.4 Case Studies
13.5 Conclusions
Acknowledgments
References
Chapter 14: Soft Sensor Design for Bioreactor Monitoring and Control
14.1 Introduction
14.2 The Process Analytical Technology Perspective on Soft Sensors
14.3 Conceptual Design of Soft Sensors for Bioreactors
14.4 “Hardware Sensor” Alternatives
14.5 The Modeling Part of Soft Sensors
14.6 Strategy for Using Soft Sensors
14.7 Applications of Soft Sensors in Bioreactors
14.8 Concluding Remarks and Outlook
References
Chapter 15: Design-of-Experiments for Development and Optimization of Bioreactor Media
15.1 Introduction
15.2 Fundamentals of Design-of-Experiments Methodology
15.3 Optimization of Culture Media by Design-of-Experiments
15.4 Conclusions and Outlook
References
Chapter 16: Operator Training Simulators for Bioreactors
16.1 Introduction
16.2 Simulators in the Process Industry
16.3 Training Simulators
16.4 Requirements on Training Simulators
16.5 Architecture of Training Simulators
16.6 Tools and Development Strategies
16.7 Process Models and Simulation Technology
16.8 Training Simulator Examples
16.9 Concluding Remarks
References
Index
End User License Agreement
Pages
xv
xvi
xvii
xviii
xix
xx
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
Guide
Cover
Table of Contents
Preface
Begin Reading
List of Illustrations
Chapter 1: Challenges for Bioreactor Design and Operation
Figure 1.1 (a) An old fermentation plant from the late nineteenth century. (b) A modern fermentation plant one century later. The gap in time between the plants reveals that some of the design features have undergone changes, while others are unchanged: the bioreactors are cylindrical vessels, the containment of the broth and concern about contamination were in former days less, piping are essential, many vessels are using the available plant space, and few plant operators are close to the process.
Figure 1.2 Twelve examples of bioreactor designs: (a) stirred-tank reactor, (b) bubble reactor, (c) airlift reactor, (d) loop reactor, (e) reactor with immobilized cells, (f) fluidized reactor with recycling of cells, (g) solid-phase tray reactor, (h) rotary drum bioreactor, (i) agitated-tank reactor with movable impeller, (j) continuous screw bioreactor, (k) hollow-fiber reactor, and (l) wave bioreactor
Figure 1.3 An emerging new trend is the replacement of old stainless steel fermenter with single-use wave-bioreactors. It is a striking example of how smart designs based on fabrication technology use compatible low-cost materials and new conceptual thinking lead to a leap in design
Figure 1.4 The CDIO concept: the process of developing a new product or production system is considered as a consecutive activity spanning from conceiving the product and production concept, designing the product or production system, implementing it into full-scale production, and finally operating it continuously for regular production. It is advocated that The CDIO is applicable to all industrial development work and should, therefore, be the framework for all engineering activity – from training and education till operating a process
Figure 1.5 Framework for the development of a new product and its production depicted from three critical hurdles: the transfer of the product concept into a bioprocess, the transcendence of development from natural science to engineering, and the interaction between screening, scale-up, and full-scale manufacture. In all, the bioreactor has a key role
Figure 1.6 A systems biology approach for bioreactor process development based on a three-round procedures.
Figure 1.7 Conceptual design principle according to Ulrich and Eppinger sequential design concept where product alternatives are screened versus customer needs.
Figure 1.8 The biomechatronic methodology according to Mandenius and Björkman where conceptual methods and tools are combined in a complementary methodology.
Figure 1.9 Hubka–Eder map of a bioreactor for the microbial production of a recombinant protein. Overall Hubka–Eder map showing the transformation process and the systems and subsystems involved for performing the transformation.
Figure 1.10 A zoom-in of the Hubka–Eder map in Figure 1.9 showing the
and
systems and the interaction of subsystems.
Figure 1.11 (a) Functional elements derived from the systems in the Hubka–Eder map. In a real design, the number of elements may exceed 100. (b) The elements are combined in order to envisage various configurations. The four examples shown can be configured in a variety of other permutations.
Figure 1.12 The CDIO concept as defined in Figure 1.4, here adapted to bioreactor design and operation with several of the topics and challenges addressed in this book.
Chapter 2: Design and Operation of Microbioreactor Systems for Screening and Process Development
Figure 2.1 Operating and physicochemical parameters influencing the reactor performance of shaken microbioreactor systems. Shake-flask and shaken-microtiter plates are illustrated schematically. Rotation of the bulk liquid inside the vessel is due to rotating centrifugal forces during the shaking process. The following parameters directly influence the performance of shaken microbioreactors: shaking frequency
n
, filling volume
V
L
, reactor volume
V
R
, vessel diameter
d
, shaking diameter
d
0
, viscosity of liquid
η
, oxygen solubility
, mass transfer resistance through the cover, mass transfer resistance at the gas–liquid interface, surface properties of the reactor wall, inner vessel design, temperature
T
, pressure
p
. Depending on the reactor size and properties of the bioreactor wall material, surface tension
σ
of the liquid has an influence on liquid distribution.
Figure 2.2 Liquid distribution during shaking in shaken bioreactors at in-phase operating conditions. The liquid exhibits a characteristic shape and rotates inside the vessel around the inner radius
r
i
with angular speed
ω
. At in-phase conditions, the liquid is oriented in the direction of the centrifugal force
F
c
. A liquid film is formed on the reactor wall, which is not covered by the bulk liquid. Gas/liquid mass transfer occurs through the surface of the bulk liquid as well as through the liquid film.
Figure 2.3 Influence of the viscosity on mixing time in shake flasks. With increasing viscosity, the mixing time in shake flasks sharply increases at phase numbers of
Ph
≤ 1.26. Liquid sticking to the flask bottom during shaking indicates out-of-phase operating conditions. At in-phase operating conditions, the liquid follows the rotational movement of the shaking table. The following operating conditions were used: 250 ml flask,
n
= 200 rpm,
V
L
= 25 ml,
d
0
= 50 mm.
Figure 2.4 Influence of the reactor surface properties on the maximum oxygen transfer capacity in shaken bioreactors. Maximum oxygen transfer capacity (OTR
max
) was determined in hydrophilic and hydrophobic shake flask. Experimental condition: 1 M sulfite system, 250 ml flask,
d
0
= 50 mm,
V
L
= 26 ml.
Figure 2.5 Schematic representation of a fed-batch microtiter plate and dialysis shake flask. (a) Dialysis shake flask. A substrate reservoir is installed inside the flask. Through a membrane, feed solution diffuses into the bulk liquid. (b) Fed-batch MTP. Two wells are connected through a hydrogel channel in the bottom of the plate. Substrate diffuses from substrate well to culture well, depending on gel properties, channel geometry, and substrate feed concentration.
Figure 2.6 X-D diagram of a continuous cultivation of
C. glutamicum
in the CosBios bioreactor. At dilution rate higher than 0.4 h
−1
, oxygen limitation occurs. Cultivation conditions: 10 g l
−1
glucose in the feed stock,
n
= 275 rpm, resulting in a filling volume of 27 ml,
d
0
= 50 mm,
q
= 2 vvm,
T
= 30°C.
Figure 2.7 Prototype of an optical online measuring device based on the BioLector technology. Up to four microtiter plates can be operated in parallel. The wells are monitored by scattered light or fluorescence measurements subsequently. The device is placed in a climate chamber with constant temperature. (a) Shaking table, (b) microtiter plates, (c)
x–y
drive to move the fiber-optic cable, (d) fiber-optic cable below the microtiter plates.
Figure 2.8 Influence of fluorescent proteins on online monitoring in microtiter plates.
In vitro
characterization and correction of the influence of YFP fluorescent protein on the optical DOT signal is shown. The DOT signal is corrected for the cultivation of
E. coli
expressing YFP by using fluorescence-dependent calibration curves. The measured original DOT signal shows final values of 300%. By using an appropriate calibration curve, the signal is corrected. Cultivation conditions: 48-well FlowerPlate with optodes for DOT and pH measurement,
V
L
= 800 µl,
n
= 1100 rpm,
d
0
= 3 mm,
T
= 37°C, Wilms-MOPS medium with 20 g l
−1
glucose, induction with 0.1 mM IPTG after 6 h.
Figure 2.9 RAMOS measurement of
H. polymorpha
RB11 FMD-GFP cultivation in 96-deepwell microtiter plate. Oxygen transfer rate was determined online in three different microtiter plates with different filling volumes. Cultivation parameters: Syn6-MES medium,
c
Glucose
= 20 g l
−1
,
n
= 350 rpm,
d
0
= 50 mm,
T
= 37°C.
Chapter 3: Bioreactors on a Chip
Figure 3.1 (a) Multiphase laminar flow patterning (from [31]) (b) Mixing of two laminar flows by diffusion. (c) A linear concentration generator with eight different concentrations at the outputs
Figure 3.2 Schematics of the photolithography (a–c) and soft lithography (d–f) procedures. (a) SU-8 is spin-coated and prebaked on a bare wafer. (b) With a transparency photomask (black), UV light is exposed on the SU-8. (c) Exposed SU-8 is then baked after exposure and developed to define channel patterns. (d) PDMS mixed solution is poured on the wafer and cured. (e) Cured PDMS is then peeled from the wafer. (f) The device is trimmed, punched, and autoclaved ready for assembly
Figure 3.3 A set of two micromachined Delrin holders with short Teflon microtubes. There are frits with filters at both sides of the microtubes. The diameter of the tubes is 150 µm
Figure 3.4 (a) A schematic of a chemostat. It is a continuous-flow reactor, keeping the amount of medium inside the reactor constant. (b) False-color photograph of the continuous culture device with device components. B and P are individual on-chip blocking valves
Figure 3.5 An
ex vivo
tumor spheroid model. Microscopic images of monolayers and spheroids of human mesothelioma cell line NCI-H226. 2D, monolayers; 3D, spheroids. Scale: 400 µm. Microarray analysis revealed that 142 probe sets were differentially expressed between tumor spheroids and monolayers
Figure 3.6 The principle of a 3D cell culture chip. (a) A pillar array 200 µm with a 20 µm gap between the pillars retains the cells. On both sides of the pillar array, a medium is perfused (from [27]). (b) Scanning electron micrograph of pillar array. (c) Microfluidic culture unit design. The microfluidic unit consisted of three parts: a 150 µm wide by 440 µm long cell culture area (blue), a microfluidic perfusion barrier (gray), and a medium-flow channel (red). Cells were introduced from the top port and localized into the cell culture area. The perfusion barrier consisted of a grid of channels 5 µm wide and 2 µm long, serving to prevent cells from passing through, while enabling nutrient exchange from the flow channel on the opposite side. Inset shows scanning electron micrograph of the perfusion channels. Scale bar represents 5 µm. (d) HepG2/C3A human hepatoma cell growth inside the microfluidic cell culture array. Scale bar represents 100 µm
Figure 3.7 A human-on-a-chip with four-cell chambers used to culture multiple cell types with a base medium and cell-specific microenvironments created by gelatin microspheres loaded with soluble factors
Chapter 4: Scalable Manufacture for Cell Therapy Needs
Figure 4.1 Overview of approximate number of Prochymal® doses required for increasing numbers of patients from Phase III clinical trials to post-market approval (example scale).
Figure 4.2 An overview of the two main manufacturing paradigms for cell therapy; process flow diagrams for autologous (patient specific) manufacture and allogeneic (off the shelf) manufacture.
Figure 4.3 (a) A scale-out expansion platform for stem cells; an automated robotic tissue culture flask handler, the CompacT SelecT, (b) Overview of the main internal components within the CompacT SelecT including robotic arm, stripette holder, decapper and flask holder, (c) The barcoded T-flask infeed and outfeed ports for traceability and monitoring, (d) Inlet, waste, and reservoir pumps and automated Cedex mammalian cell counter.
Figure 4.4 (a) An overview of the 5 l stirred-tank bioreactor configuration used in our laboratories. The microcarriers (with cells attached) are kept in suspension with a downward pumping three-blade 45° pitch-blade impeller (
D
/
T
∼ 0.44) agitated at the minimum speed required to keep the microcarriers in suspension, (b) The high-throughput, automated ambr15™ microbioreactor that allows for the parallel culture of 24 microbioreactor vessels (15 ml working volume), which we have recently demonstrated is amenable for hMSC microcarrier culture.
Figure 4.5 Human mesenchymal stem cells cultured on microcarriers in spinner flasks and subsequently harvested and characterized to determine cell quality with respect to retention of differentiation capacity postharvest and identity. (a) Phase-contrast image of hMSCs attached to Plastic P102-L microcarriers. (b) Fluorescent image of identical microcarriers in the previous image depicting viable cells (green fluorescent calcein-AM) and nonviable cells (red-fluorescent ethidium homodimer-1). (c) Successful detachment of hMSCs from Plastic P102-L microcarriers. (d) Alkaline phosphatase and Von Kossa staining demonstrating osteogenic differentiation. (e) Alcian Blue staining demonstrating chondrogenic differentiation. (f) Oil Red O staining demonstrating adipogenic differentiation. (g) Multiparameter flow cytometry demonstrating cell identity with the dual gating of CD73 (+), CD90 (+), CD105(+), CD 34(−), and HLA-DR (−) for hMSCs postharvest from microcarrier culture.
Chapter 5: Artificial Liver Bioreactor Design
Figure 5.1 Potential clinical applications of bioreactors for cell maintenance to support the liver function in extracorporeal devices or for production of cells to be transplanted into the injured organ.
Figure 5.2 (a) Four-compartment hollow-fiber bioreactors for clinical use in extracorporeal liver support systems (left) or in downscaled versions for laboratory research (right, foreground); (b) structure of the capillary network with independent capillary systems for countercurrent medium/plasma perfusion (blue and red) and oxygenation (yellow); (c) detection of cytochrome 2C9 (red) and the transporter protein multidrug resistance protein 2 (MRP2, green); (d) reorganization of hepatocytes (CK 18, green) and nonparenchymal cells (vimentin, red) between the artificial capillaries in the bioreactor.
Figure 5.3 Bioreactor perfusion circuit with tubing for pump-driven medium recirculation through the bioreactor (red: inflow tube sections, blue: outflow tube sections) and for fresh medium inlet (red) as well as used medium outlet (blue). The bioreactor disposes of two medium perfusion capillary systems (M
1
in and out, M
2
in and out), which are countercurrent perfused to enhance mass exchange. Cells are inoculated via a tube line (CC in) branching from the recirculation circuit. An electronically controlled gas mix unit provides defined flow rates and concentrations of gases in the supplied gas mixture (gray lines, gas in and gas out). Automated pH control is possible by the integration of pH sensors into the perfusion circuit. POF, plastic optical fiber, MC, microcontroller.
Figure 5.4 (a) One layer of the capillary membrane network from an 8 ml bioreactor before inoculation of mouse embryonic stem cells (upper left) and after 6 days of culture (lower right); aggregates are visible between the fibers. (b) Toluidine blue staining of cells obtained from an 800 ml bioreactor after day 3. Inset: SSEA-1 immunoreactivity (green) on day 3 in the 800 ml bioreactor culture (blue: DAPI staining of cell nuclei). (c) Glucose consumption and lactate production in the 800 and 8 ml bioreactors versus glucose consumption in 2D control dishes over 3 days. Values are given as consumption or production rates per hour per bioreactor/culture flask. Thus, the extremely different scales of the bioreactors and cell numbers are reflected by these medium parameters
Chapter 6: Bioreactors for Expansion of Pluripotent Stem Cells and Their Differentiation to Cardiac Cells
Figure 6.1 Process strategies for pluripotent stem cell-based heart repair. hPSCs – human pluripotent stem cells; ECs – endothelial cells; MSCs – mesenchymal stem cells; PCs – pericytes; ECM – extracellular matrix.
Figure 6.2 Culture method-dependent morphology of hPSCs. Representative hPSC images of conventional 2D feeder culture (a) and feeder-free culture using chemically defined medium (b). For aggregate generation under fully defined conditions in stirred bioreactors, single-cell hPSCs (c) are inoculated. Formation of spherical aggregates is exemplarily shown over time (d–f) at 1, 3, and 7 days after inoculation, respectively. The 3D aggregates generated in a stirred bioreactor maintain high differentiation potential into beating cardiomyocytes (g, h). Differentiation was performed under fully defined conditions applying small-molecule Wnt pathway modulators. Bright-field image (g) and corresponding NKX2.5-GFP expression (h) on day 7 of differentiation using an HES3-NKX2.5
w/GFP
reporter cell line [31]. Cryosectioned aggregates were stained against the structural cardiomyocyte marker α-actinin (i).
Figure 6.3 Schematic of stirred instrumented bioreactor technology. The bioreactor is equipped with an eight-blade pitched impeller for magnet-coupled overhead drive stirring, a temperature sensor, pH- and DO-electrodes, as well as a sampling port and a liquid-free exhaust gas condenser. Overlay gassing is performed utilizing a flow-controlled gas mixing system. Perfusion is technically established via peristaltic pumps, a retention system placed at the waste medium stream, which is connected to a waste bottle, as well as of a feed line connected to a fresh medium reservoir.
Figure 6.4 Culture strategies for hPSC processing in stirred-tank bioreactors. (a) In batch-fed processes the entire culture volume is manually exchanged, whereas in perfusion processes the culture medium is continuously replaced by application of an automatic pump device combined with a cell retention system to maintain cells inside the bioreactor. The culture volume is thereby kept constant via applying equal feed and outflow rates. (b) In advanced hPSC bioprocesses, perfused bioreactors would be equipped with an additional external loop for a process-integrated aggregate filtration, online metabolite analysis, as well as an subsequent online cell and aggregate monitoring to enable sophisticated process control.
Figure 6.5 Process schemas for autologous and allogeneic cell production.
Chapter 7: Culturing Entrapped Stem Cells in Continuous Bioreactors
Figure 7.1 Process flow diagram for a mass transfer-based process design.
Figure 7.2 The effect of autocrine factor washout in the differentiation of hMSCs. This Figure depicts the findings in [76], where the authors demonstrate that the convective flow of culture medium through the hMSC-containing scaffold removes ECM components (Collagen I, Fibronectin and Vitronectin) and the bFGF soluble factor that it is bound to the ECM. When these cell constructs were cultivated in parallel flow the diffusional mass transfer allowed for an accumulation of the ECM and the bound Fgf-2 molecules which in turn maintained the undifferentiated phenotype of the cultured hMSCs.
Chapter 8: Coping with Physiological Stress During Recombinant Protein Production by Bioreactor Design and Operation
Figure 8.1 Environmental and intrinsic stress factors affecting recombinant protein production.
Figure 8.2 Aeration spargers and impellers commonly used for laboratory scale microbial and cell culture applications.
Chapter 9: Design, Applications, and Development of Single-Use Bioreactors
Figure 9.1 Membrane device for cell-free sampling.
Figure 9.2 Scales commercially available for different single-use bioreactor concepts. (a) Stirred single-use systems are available from the milliliter scale for parallel cultivation up to the cubic meter scale for production. The lab scale system can be operated with a flexible volume. At larger scale, operation at varying liquid volumes during operation affects the power input and is restricted by the position of the stirrers. (b) Orbital-shaken bioreactors can be scaled-up very easily from the microwell plate scale to the production scale. Film formation along the wall is crucial for the achievement of a certain gas mass transfer; optimal filling levels have to be maintained. (c) Wave-mixed reactors require more space when scaled-up. Therefore, a scale larger than a few hundred liters is not conducted so far. Since the bags can be easily modified in size without changing drastically the power input even during operation, each bioreactor is applicable in a comparably broad volume range.
Figure 9.3 Removable bodies creating a narrow channel in the CELL-tainer SUB, ensuring a scalability from the milliliter to the liter range without the necessity to change the bag.
Figure 9.4 Lab-scale wave-mixed single-use bioreactor.
Figure 9.5 Stirred-tank single-use bioreactor.
Figure 9.6 Cell culture expansion process including the application of perfusion and wave-mixed single-use bioreactors. (a) Steps necessary involving different scales of wave-mixed single-use bioreactors. (b) Steps when using the scalability of one single-use bioreactor with devices that allow a noninvasive modulation of the cultivation volume such as channel blocks.
Figure 9.7 Cultivation performance of a stepwise expansion process when channel blocks were applied [69].
Figure 9.8 Comparison of the growth performance of an
Escherichia coli
BL 21 cultivation [11] (fed-batch phase) performed in a stirred-tank reactor with a working volume of 2 l (unfilled squares), in a wave-mixed CELL-tainer CT 20 with a working volume of 12 l (gray diamonds), and in a wave-mixed CELL-tainer CT 200 with a working volume of 120 l (black triangles). The exponential feed rate was increased by a factor of 0.25 h
−1
t. After the induction of the expression of a maltogenic amylase at
t
fed-batch
= 19 h, the feed rate was kept constant. No oxygen blending was applied.
Chapter 10: Computational Fluid Dynamics for Bioreactor Design
Figure 10.1 Schematic of an aerated bioreactor.
Figure 10.2 (a) Velocity vector variation at different impeller speeds of (1) 100 RPM, (2) 150 RPM, and (3) 200 RPM. Three impellers out of which one is a Rushton turbine while the other two are 3-blade propeller-type impellers are mounted on shaft. The reactor capacity is of 67 l.(b) Gas phase volume fraction at gas inlet flow rate of (1) 1 LPM and (2) 4 LPM for 67 l capacity bioreactor.
Figure 10.3 Front view of meshed 67 l capacity bioreactor where tetrahedral and unstructured meshing used. Bioreactor equipped with three impellers consisting of one Rushton turbine and two 3-blade propeller-type impellers with a central shaft of diameter 0.025 m.
Figure 10.4 Grid independence test: plot of variation of
k
L
a
with grid size
Figure 10.5
k
L
a
Variation in reactor along
x
= 0.0625 plane with PBE model
Figure 10.6 Contour plot for mass transfer coefficient
k
L
a
(s
−1
) for the case: (a) with–without the PBE, (b) 9 classes PBE, and (c) 13 classes PBE
Figure 10.7 Bioreactor with Rushton turbine with 180 RPM impeller speed, gas flow rate 0.00164 m
3
s
−1
, Froude number = 0.31 and experimental gas hold-up 3%: (a) bubble diameter variation (mm) and (b) gas volume fraction variation in the section of tank
Figure 10.8 Circulations produced in velocity illustrating the cavity formation
Figure 10.9 Comparison of mean bubble diameter (mm) in Case 4 (
) with experimental measurements of (
) at a height (in m) of (a) 0.125, (b) 0.25, (c) 0.435, and (d) 0.805 in the tank stirred by Rushton turbine
Chapter 11: Scale-Up and Scale-Down Methodologies for Bioreactors
Figure 11.1 Anticipated distribution of substrate, dissolved oxygen, pH value, and dissolved carbon dioxide in a top-fed industrial-scale high-cell-density cultivation process. Usually, the substrate nearly depletes in the middle and bottom parts. The altered metabolic activity of cells in the different regions due to varying substrate availability leads to an opposite dissolved oxygen concentration gradient. The pH value is affected by large mixing times of its controlling agents. The dissolved oxygen concentration is elevated in zones of high metabolic activity and in the bottom part due to hydrostatic pressure.
Figure 11.2 Schematic presentation of different scale-down simulators. (a–d) One-reactor systems. (a) tubular closed-loop air-lift reactor [3], (b) single-loop bioreactor design by Gschwend
et al.
[4], (c) reactor with internal disks for increased mixing times [5], (d) cyclically changing feeding approaches. (e–g) More compartment reactors with (e) two stirred-tank reactors, (f) with a simple plug-flow reactor, (g) with a plug-flow reactor that contains static mixers and can be aerated, (h) three-compartment reactor system with two plug-flow reactors for simulation of different zones at the same time [6]. F = substrate feed.
Figure 11.3 (a) Schematic and (b) photographic presentation of a three-compartment scale-down bioreactor. It comprises a stirred-tank reactor TECHFORS-S with a working volume of 10 l (Infors AG, Switzerland) and two plug-flow reactor modules (total liquid volume of 1.4 l). The residence time can be set by peristaltic pumps at the bottom part between 30 s and 1.5 min. The plug flow reactors comprise four static mixer modules. They are connected with Tri-Clamps. Five triple ports for monitoring pH, DO, and for sampling are located between each static mixer. Aeration is conducted via a sparger from the bottom (if needed). All parts are steam-sterilizable.
Figure 11.4 The methodology of an iterative scale-up and scale-down approach – knowledge is gained from the large scale (left) and mimicked in the small scale. Lack of resolution and measurement methods is (partly) replaced by model approaches for the description of the conditions in large scale.
Figure 11.5 Electrodes as applied in multisensor devices; upper left corner: Hamilton ARC sensors for the optical measurement of dissolved oxygen, redox potential, pH value (both glass electrodes), and temperature; lower left corner: microsensors of the Kurt-Schwabe-Institute for temperature measurement (Pt1000), redox (platinum electrode), pH value, and reference (both glass electrodes); diameter: maximum 4 mm, length: 40 mm; right: waterproof, movable built-in sensor heads with a diameter of 76 mm (top) and 45 mm (bottom).
Chapter 12: Integration of Bioreactors with Downstream Steps
Figure 12.1 Interactions between the cell culture step and other steps in a typical bioprocess. Class A pertains to the bioreactor, and Class B to all the steps downstream of the bioreactor. Class A includes the usual nonproduct-based as well as product-based metrics for bioreactor efficiency. Class B describes all the metrics describing impacts of the bioreactor step on steps downstream of it, whether a single step (unit operation) or multiple steps are affected. Any interaction between parameters of different classes represents a trade-off: either within the bioreactor, or between the bioreactor step and the steps downstream.
Figure 12.2 Global sensitivities of operational parameters for typical centrifugation of mammalian and yeast culture broth [10]. The mammalian culture results are on the right, and the yeast results (referred to as “high cell density”) are on the left of the diagram. The operational parameters are:
m
, the mean particle size;
μ
, the broth viscosity; Δ
ρ
, the density difference between the particles and the liquid; and
Q
, the flow rate. The mammalian centrifugation is dominated by
m
, but the yeast centrifugation is dominated by
Q
.
Figure 12.3 Supernatant volumes collected as a function of time when cell-culture broth was exposed to a variety of ion-exchange resins [4]. The number in parentheses next to each resin is the average particle size of the resin beads. The Si-PEI provided the most efficient collection of supernatant volume, and was chosen for further study (see text for more details).
Figure 12.4 Addition of Si-PEI ion-exchange resin to reduce pool cell density and host-cell protein (HCP) content while recovering the product [4]. Increasing the amount of resin, through 10%, decreased the pool cell density and HCP while maintaining product recovery (see text for more details).
Figure 12.5 Comparison of centrifugation and depth filtration for 5000 l bioreactor [14] (see text for more details).
Figure 12.6 Two-dimensional gel electrophoresis pictures [21] of an intermediate process sample from the apolipoprotein purification process: (a) total protein stain; (b)
E. coli
HCP Western blot of the same gel. Total protein loading was 400 µg (see text for more details).
Figure 12.7 Schematic workflow for model analysis [26] (see text for more details).
Chapter 13: Multivariate Modeling for Bioreactor Monitoring and Control
Figure 13.1 Overview of publication rate in bioprocess modeling, monitoring, and control area between 1975 and 2014, based on a search of Web of Science.
Figure 13.2 Example of an NIR fingerprint from bioprocess monitoring.
Figure 13.3 An example of gas uptake measurements from high-throughput monitoring experiments
Figure 13.4 Visual representation of a PCA transformation of process measurements.
Figure 13.5 (a) Scores plot of sensor RESPONSE parameter for all sensors, (b) scores plot for sensors RESPONSE parameter after the removal of insignificant and highly correlated variables.
Figure 13.6 (a) Scores plot for batch 1A response parameter only, (b) scores plot for batch 1B response parameter only.
Figure 13.7 (a) Scores plot for batch 2A response parameter only, (b) scores plot for batch 2B response parameter only.
Figure 13.8 PCA loading plot for all sensor responses
Figure 13.9 Amino acid concentration predictions for (a) Run 5 and (b) Run 13. The details of each model in terms of input and output data, model structure and the RMSE values of each model for each validation run are provided in Table 13.1.
Figure 13.10 Glycoform predictions for (a) Run 5 and (b) Run 13. The details of each model in terms of input and output data, model structure, and the RMSE values of each model for each validation run are provided in Table 13.2.
Chapter 14: Soft Sensor Design for Bioreactor Monitoring and Control
Figure 14.1 The soft sensor principle as defined in Refs. [1, 2]. The Figure shows one hardware sensor and one estimator. In reality, these can be multiplied.
Figure 14.2 The information flow in a soft sensor with examples from typical bioreactor applications.
Figure 14.3 Alternative soft sensor configurations : (a) a bioreactor unit with media inlet and outlet and gas vent equipped with three online sensor sensors (e.g., 1, CO
2
gas analysis; 2, cell density sensor; 3, dissolved oxygen electrode); (b) sensors placed inline in the bioreactor outlet (e.g., 1, HPLC for metabolites; 2, immunosensor for a recombinant protein; 3, NIR sensor for biomass); (c) sensors placed
in situ
and inline (e.g., 1, mass spectrometer for volatile metabolites; 2, capacitive sensor for cell viability; 3, HPLC for substrate); (d) bioreactor sequence where a product is formed (e.g., 1, gas sensor volatile product; 2, cell density sensor; 3, substrate sensor); (e) separation unit where component in an input flow is separated (e.g., 1, cell density sensor; 2,3, protein sensors); (f) a column separation where the outlet stream is monitored (e.g., 1, UV sensor, 2, conductivity sensor, 3, immunosensor); (g) a downstream unit where a product is dried (e.g., 1, moisture, 2, temperature, 3, weight).
Figure 14.4 A general bioreactor soft sensor setup showing typical standard sensor signals supplemented with additional hardware sensor signals.
Figure 14.5 (a) A soft sensor setup where an online fluorescence probe predicts concentrations of glucose, biomass, and ethanol using an MVDA algorithm in yeast cultivation. (b) Four examples of predictions of analytes from the multiwavelength fluorescence spectrometer. Time profiles of (○) measured and (—) predicted (I) ethanol and (II) consumed glucose using one TRAD PLS model for the whole cultivation. (III) and (IV) show the corresponding values using segmented models
Figure 14.9 (a) Configuration of the experimental setup of a soft sensor with an electronic nose and NIR spectroscopy. (b) A detailed flow chart of the system. (c) Trajectory score plots generated from the first principal component of preselected NIR and EN signals for the scaled-down
V. cholerae
cultivations (▪ and •, calibration data sets; ▴, validation data set). The shaded area represents the range of values, in which the culture is in its “normal” state
Figure 14.6 Soft sensor configuration where the temperature signals from the cooling jacket of the reactor to estimate the heat production of the cell. This information is used for deriving the specific growth rate and then used that for controlling the dosage of the feed. (a) Experimental setup of the soft sensor. (b) Induced fed-batch cultivation producing recombinant GFP showing in upper panel controlled feed rate (), metabolic heat production (----), and
μ
metabol
(—), and in lower panel biomass from heat (·····), optical density (▪), and capacitance measurements (—). The production of GFP (○) is induced at 12 h
Figure 14.7 (a) Soft sensor configuration that estimates the biomass concentration from calculating the addition of titrand. (b) The
μ
NH3
sensor signal (thick continuous line) from an
E. coli
fed-batch cultivation on minimal medium compared with μ calculated on dcw data (filled circle). Also shown are the measured data for
μ
NH3
calculation, that is, the NH
3
vessel balance data () and the cell dry weights (○). The exponential glucose feed was started at 12.6 h and was constant from 16.5 h
Figure 14.8 (a) Configuration approach where feeding is controlled around the specific production rate of mixed acid side-production. (b) Soft sensor control strategies of the feeding of the fed-batch cultivation were applied using the sum of specific production rates of the metabolites acetate, lactate, ethanol, and formate. (c) Control of the sum of MAF metabolite concentrations in the fed-batch GFP cultivation. Feeding starts at 5 h and induction is carried out at 9–10 h using 0.03 g l
−1
IPTG. Cultivations with identical control settings (data for biomass concentration from NIR measurements in (b) is not available)
Chapter 15: Design-of-Experiments for Development and Optimization of Bioreactor Media
Figure 15.1 Factors useful in DoE for optimizing media versus critical output parameters (termed responses in the DoE methodology). Note that not all factors are added in the initial culture media, some are added during processing. Also, note that other critical factors for the outcome of the process and that cross-interact with the media components can favorably be included in the DoE procedure.
Figure 15.2 (a) The Figure shows how a quasi-optimum is achieved by varying one variable at a time. When keeping variable
X
1
constant, five experiments varying variable
X
2
are performed. Then, starting from the optimum (the center point), variable
X
1
is varied in another five experiments. A correct optimum is never reached as there is a dependency between variable
X
1
and
X
2
. (b) By simultaneous variations of variable
X
1
and
X
2
, and analyzing the result in experimental design software, the direction of the true optimum can be found.
Figure 15.3 (a) Experimental design performed as screening of important variables applied in a three-variable case. In addition, replicate experiments at the center point are recommended (not shown in the figure). (b) Two types of central composite designs: the central composite face-centered design (CCF, left) and the central composite circumscribed design (CCC, right) in a three-variable case including triplicate experiments at the center point.
Figure 15.4 The sequential working procedure of DoE and RSM when optimizing growth and production media in bioreactors: In the figure, also media components and factors added during processing are included. DoE procedures for addition of components are, however, seldom reported in literature while combinations of media factors with other physical factors controlled in the bioreactor are common.
Figure 15.5 (a) Typical representations of optimization in a DoE study. A contour plot showing selected factors on
x
and
y
axes and one response represented as topographic bars. (b) A response surface plot showing two factors and a response in a 3D graph representation.
Figure 15.6 A three-variable design including two qualitative variables at three and four discrete levels, respectively, as well as a third quantitative variable.
Figure 15.7 A multi-bioreactor setup (Greta System, Belach Bioteknik AB, Stockholm, Sweden) suited for factor studies.
Figure 15.8 Optimization for the production of the hepatoma cell line C3A. (a) Media factors (HGH, hepatocyte growth factor, oncostatin M; FGF-4, fibroblast growth factor; EGF, epidermal growth factor, nicotinamide, dexamethasone, human serum albumin) and responses (urea rate, glucose rate, lactate rate, lactate dehydrogenase rate) in the hepatocyte DoE model. Response surface graphs for optimization of formation of (b) urea, (c) lactate dehydrogenase, and (d) lactate and at varying oncostatin M and hepatocyte growth factor a constant FGF4 concentration of 20 µg ml
−1
Figure 15.9 (a) Application of the two-step strategy using the Plackett–Burman design in a first screening followed by factor selection and further adjustment using the MinRes IV design. (b) MTT responses (optical density) of mESC cultured without or with feeder cells (MEF) in 14 different media composed using a MinRes IV design. Values (means ± SD) were normalized to the reference medium no. 14. The unpaired
t
-test was performed to compare each group with the reference medium no. 14
Chapter 16: Operator Training Simulators for Bioreactors
Figure 16.1 Changing operator tasks during the last century
Figure 16.2 Structure of training simulators.
Figure 16.3 Modeling levels for training simulators – numbers refer to the model types pointed out in the text
Figure 16.4 Modeling cycle for training simulator modeling.
Figure 16.5 Virtual pretraining with an operator training simulator for recombinant protein production with
E. coli
in a fed-batch bioreactor can support the training effectiveness in a subsequent real cultivation experiment
Figure 16.6 P&I diagram of a bioethanol plant training simulator with two bioreactors (B-8; B-17), a microfiltration unit (F-1) and a rectification column (C-1)
Figure 16.7 Graphical user interfaces on the monitoring and control station for the bioethanol plant simulator (bioreactor, cross-flow-filtration, distillation column)
Figure 16.8 The bioethanol plant training simulator is automated using GRAFCETS as sequential function charts (SFC), realizing well defined operational sequences. Left: P&ID distillation column; right: GRAFCET for distillation operation (from Kuntzsch [2] with kind permission).