Large-Scale Mesenchymal Stem/Stromal Cell Expansion: A Visualization Tool for Bioprocess Comparison

Abstract

Large-scale and cost-effective cell expansion processes are a prerequisite for the clinical and commercial translation of cell-based therapies. A large variety of cell expansion processes are described in literature, utilizing different cell types, culture vessels, and medium formulations. Consequently there are no straightforward means for the comparison or benchmarking of these processes in terms of efficiency, scale, or costs. The purpose of this study was to systematically review the available mesenchymal stromal cell (MSC) expansion literature and develop an interactive visualization tool for comparing the expansion processes. By using this computational tool, process data could be concentrated, standardized, and analyzed to facilitate a more general understanding of the parameters that define a cell culture process, and in the future allow rational selection or design of these bioprocesses. Additionally, a set of bioprocess metrics were defined that assured the comparability between different processes. Currently, the literature-based data repository holds 73 individual cell expansion processes on seven different types of human MSCs in five different types of culture vessels. The visualization tool allowed benchmarking of these processes against each other, serving as a reference point for cell expansion process efficiency.

Introduction

The number of clinical trials for cell-based therapies has been constantly increasing over the last years. Approximately 400 of these trials use human mesenchymal stromal cells (MSCs) as a therapeutic cell source, and many other (stem) cell types are considered for clinical translation.^1,2 At the same time, the potential benefits that can be brought by cell-based therapy also seems to be recognized by industry since more and more cell-based companies arise.^3,4 It is expected that most cell therapy applications will require between 10⁶ and 10⁹ MSCs for a single dose.^5,6 Although the relatively low number of Phase III clinical trials and limited commercialization has been partially attributed to manufacturing challenges associated with manual flask-based cell culture, the flask-based method is still the gold standard. Manual flask-based cell culture is labor intensive, requires sequential “open events” that are prone to contamination, and does not allow online monitoring and control of the microenvironment of the cells. At the same time, the flask-based strategy is limiting the attainable cell density levels and the scale of operation due to the inherent design of the flasks with a low volume to culture surface ratio.

The translation of clinically promising cell therapies to commercially successful products will require the conversion from manual cell culture processes to controlled bioprocesses that are able to guarantee the production of cell-based therapies with manageable cost of goods (COGs) and robust in vivo performance.^7–10 Bioreactors have been shown to provide an improved cell culture environment by controlling nutrient refreshment and waste removal rates¹¹ while reducing complexity involved in bioprocessing.¹² From the translational perspective, bioreactors have been employed to provide efficiency in terms of cost, yield, and scale. A growing body of literature examines the use of bioreactor systems with various designs for the expansion of human MSCs and reflects on the concerns regarding large-scale MSC production, indicated by an increasing number of recent review articles.^13–15

Existing literature and regulatory perspectives¹⁶ discuss a very broad number of parameters that all affect bioprocess variability and efficiency at the same time. Their effects should therefore be taken into account when analyzing and designing bioprocesses for the expansion of MSCs. These parameters can be classified as follows:

(i) Cell type and donor: There is a large variation of sources of adult progenitor cells, with the most common being bone marrow (BM),¹⁷ umbilical cord (UC),¹⁸ adipose,¹⁹ synovium,²⁰ and periosteum.²¹ In addition, interdonor variability within the same cell type is also a factor that will affect translation to large scale.²² Even from the very beginning of the process, the number of MSCs obtained from a biopsy is variable. For example, 1 mL of human BM provides ∼1000 MSCs.²³ Whereas for adipose-derived MSCs for 0.5–2.0 × 10⁶ cells/g of adipose tissue, the percentages of MSCs range from 1% to 10%.²⁴ Again here, there is a large interdonor variability regarding cell yields from biopsies that affects translation, especially for autologous cell therapies.

(ii) Type of culture vessel: Not only in the (multilayered) flask-based expansion, where the brand of culture plastic itself might already induce variability in cell yields,²⁵ but also the large diversity of culture systems that has been reported to support MSC expansion contributes to the variety in bioprocess efficiency and variability. For instance microcarrier-based stirred tank reactors,²⁶ hollow fiber,²⁷ wave bags,²⁸ and multiplate bioreactors²⁹ have been successfully employed to generate large-scale batches of MSCs.

(iii) Raw materials and reagents: There is a broad variety of media formulations for cell expansion that differ for example in protein source or glucose concentration. Since nondefined sources of protein such as fetal bovine serum (FBS) and human platelet lysate (HPL) introduce batch-dependent process variability and present a non-negligible risk for pathogen transmission, considerable effort goes to the development of defined xeno-free medium formulations that support more efficient and less variable cell growth.^30,31 For a review on serum-free media formulations see Gottipamula et al.³²

(iv) Process parameters: The operating conditions, for instance initial cell seeding density, media refreshment strategy, vessel geometry, type of impeller, mixing intensity and stress exposure, perfusion rates and dissolved oxygen tension, vary for almost every process.³³ The impact of the microenvironment in light of the large-scale MSC expansion is reviewed in Ma et al.³⁴

To add to this complex landscape, inconsistent metrics have been used to define bioprocess efficiency making comparison across research studies a rather challenging task. Therefore, there is need for a critical and systematic analysis of existing information to allow comparability, process benchmarking, and improved understanding of bioprocess efficiency, to ultimately allow a more rational design of these bioprocesses.

The first aim of this work was to analyze the performance of different expansion processes for MSCs. An exhaustive literature study was performed, currently resulting in a database of 73 individual cell expansion processes in five different types of culture vessels (tissue culture flasks, hollow fiber bioreactors, microcarrier-based bioreactors, multiplate bioreactor, and fixed bed bioreactors), seven different types of MSCs, and a wide range of media compositions. The scale of the processes in terms of final cell numbers ranged between 7.5 × 10⁶ and 1.1 × 10¹⁰ cells. Interactive visual process performance maps were created where the scale, expansion efficiency, cell type, culture method, load on downstream processing, medium formulation, and population doubling (PD) time could be explored. Finally, based on scale-up studies carried out in our research group, we provide a cross-system cost comparison using a specific adult progenitor cell type namely, human periosteum-derived cells (hPDCs, an MSC-like cell type that holds promise for skeletal regeneration and repair strategies) reaching yields of 3.5 × 10⁸ cells per run.

Methods

Database construction

A database of individual cell expansion processes was assembled based on a literature search for articles that have a detailed description of one or multiple cell culture processes. This database therefore not necessarily represents the current situation in industry. If multiple processes were described in one article, for example comparing different microcarrier densities or protein sources, all described processes were included. The first criterion for a process to be included was that the article described processes for human mesenchymal stem/stromal cells (MSCs), as claimed by the authors. The second criterion was that the scale-up of the process was taken into consideration during data collection, with the aim of only including fairly optimized processes. However, scale-up was not necessarily the main goal of the article. Third, precise information on critical process parameters, that is, cell type, starting cell number at seeding, culture time, final cell yield, culture surface or microcarrier concentration, medium composition, and exact type of culture vessel had to be included in the article, either in text format, table, or clear graphs. Unfortunately certain MSC expansion articles that were found failed to mention some of these critical parameters and were therefore not included.

Finally, a cell quality read-out was required, either in terms of the International Society for Cellular Therapy (ISCT) criteria or a potency assay,³⁵ as long as it evaluated the influence of the process on relevant quality attributes of the specific cells. Although extremely important, (parts of the) processes that described passage 0 (P0) results were excluded, as the precise number of MSCs initially present within the biopsy is unknown. Most processes were situated between passage 1 and passage 6 (on the use of passage number as a metric, see below). Additional parameters that were collected in the database for their potential effect on the therapeutic effectiveness of the cells or indirect relation to the efficiency or cost of production, but that are therefore not explicitly used further in this work, are: brand of culture vessel, specific culture surface and coating, working volumes and refreshment rates, oxygen concentrations, maintenance of cell function or potency (yes/no), and bead-to-bead transfers in microcarrier culture (yes/no).

Metrics for process comparison

To allow a fair comparison between all the different processes in different culture vessels, specific metrics were calculated from the collected data of which the authors believe that they can provide objective comparison over multiple systems and cell types. For example, passage number is still often used to define the extent of cell growth. However, in contrast to the number of PDs, this metric does not describe the proliferative precedent of the cultured cells. Similarly, in two-dimensional (2D) culture, the seeding and harvesting densities are generally stated as cells per cm² of available culture surface. As the exact culture surface is generally less easy to quantify for different microcarrier types, this metric cannot be used for comparing all different culture vessels. One metric that was therefore used often in this work is Expansion Factor or Expansion fold (EF) as this allowed us to estimate how efficient the generation of cells during a certain process step was, irrespective of the available culture surface or the vessel volume. \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} & Expansion \ factor \ or \ expansion \ fold \\ & = {\frac {number \ of \ cells \ in \ vessel \ at \ harvest \ \times \ downstream \ process \ efficiency} {number \ of \ cells \ seeded \ \times \ seeding \ efficiency}} \end{align*} \end{document} \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} \approx {\frac {final \ cell \ yield}{number \ of \ cells \ seeded}} \end{align*} \end{document}

As the downstream process efficiency (e.g., taking into account the amount of cells lost during harvest or volume reduction processes, before the actual count of the cell yield) is generally not separately quantified in literature, the enumerator was simplified to the reported final cell yield. Estimates for seeding efficiency are more often mentioned for microcarrier processes, but are generally not available for flask-based or hollow fiber processes. Therefore, also the seeding efficiency was not taken into account for the EF calculation in this work. If an EF was provided in the article that took seeding efficiency into account, it was recalculated to not include the cells lost at seeding, therefore resulting in a simple “cells out” over “cells in” ratio.

When using the EF metric as a measure for expansion efficiency, starting from the total initial number of cells seeded is more accurate from a process efficiency point of view compared with using the number of cells that are effectively seeded. Taking into account the exponential growth of cells, losing even small amounts during seeding quickly becomes expensive as these lost cells do not contribute to the final yield.

Reversely, to calculate the effective PDs and population doubling times (PDT), it is necessary to quantify and take into account the seeding efficiency as otherwise the apparent PDT seems longer than the effective PDT. Note that, in this study, the apparent PDT was calculated based on the initial and final cell number (as seeding efficiency and downstream process efficiency is generally not quantified in all articles) and, therefore, also includes lag phase and possible stationary phase at the end of the culture time: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} Number \ & of \ population \ doublings \ per \ passage \\ & = PD = log_2 ( EF ) \end{align*} \end{document} \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} Population \ doubling \ time = PDT = \frac { culture \ time } { PD } \end{align*} \end{document}

The final volumetric cell density was calculated as the total number of cells harvested divided by the total medium volume (working volume) in the culture vessel. This metric was taken as a proxy for the load on downstream processes as it was considered that in relative terms larger cell numbers in smaller volumes are less cumbersome during volume reduction and possible purification steps.

For the classification in low and high supplemented protein concentrations an arbitrary cutoff point at 5% was chosen. A medium composition with more than 5% FBS or HPL was considered to be high FBS and high HPL, respectively. Medium compositions with concentrations equal or lower than 5% FBS or HPL were considered to be low FBS and low HPL, respectively.

Cost calculations

Cost calculations were based on the cell expansion processes required for large preclinical studies performed in the host laboratory (Prometheus, KUL) for the treatment of large bone defects with hPDCs. The calculated cost included labor costs for the operators at KU Leuven, disposable costs at list price, and reagent cost at list price for the expansion of 20 million hPDCs to 350 million hPDCs. Other costs such as laboratory space, depreciation costs, and utilities were not included. All processes were performed in Dulbecco's Modified Eagle Medium (DMEM) high glucose supplemented with 10% irradiated FBS. The hollow fibers of the Terumo BCT bioreactor were coated with human fibronectin, whereas no fibronectin was used in the other culture vessels.

Visualizations

The visualization tool was created with JavaScript and Google Charts and the visuals are generated automatically from the database that is hosted online. In this way the tool can be easily updated with new processes when new publications become available. The process performance maps shown in this article are static representations of the interactive visualization tool. In the online version, the viewer is able to select processes per vessel type or per cell type. The authors would like to encourage the reader to open the following link (mtm.kuleuven.be/prometheus/processmap) in their browser to experience the interactive dataset visualization.

Results and Discussion

Based on data acquired through a literature search and subsequent data extraction, 73 individual bioprocesses were included in the visualization tool (Table 1). The database included seven different MSC types (Fig. 1A), five different types of culture vessel (Fig. 1B), and many different medium compositions that are roughly categorized based on protein source in Figure 1C, but described in more detail in Table 1. Although this data is based on research articles, the distribution of cell types used is comparable to the cell types used in recent clinical trials (e.g., 62% BM MSC and 15.4% adipose-derived stem cell [ADSC] as collected in,¹ and 56% BM MSC and 12% ADSC as collected in Sharma et al.³⁶).

FIG. 1.

Overview of (A) cell types, (B) types of culture vessel, and (C) general source of protein supplement used for MSC expansion that are included in the database. MSC, mesenchymal stromal cell; BM, bone marrow; ADSC, adipose-derived stem cell; hPDC, human periosteum-derived cell, UC, umbilical cord; FBS, fetal bovine serum; HPL, human platelet lysate. Color images available online at www.liebertpub.com/teb

Table 1.

Mesenchymal Stromal Cell Bioprocesses Used for Obtaining Data for the Visualization Tool

Reference	Type of culture vessel	Cell type	Expansion factor	Final cell yield ( × 10 ⁶ )	Medium composition	Population doubling time (h)	Growth surface or coating
⁶¹	Hollow fiber	BM MSC	31.4	660	DMEM, 5% platelet lysate, 2.1 U/mL heparin, 2 mM GlutaMAX, 10 mM N-acetylcysteine	29	Fibronectin
⁶²	Fixed bed	BM MSC	9.2	92	α-MEM, 10% FBS, Pen-Strep	67	Fibra-Cel Disks
⁶³	Microcarrier	BM MSC	8.4	21	MesenPRO, 2% FBS	63	CultiSpher-S (Gelatin)
⁶³	Microcarrier	BM MSC	7.5	7.5	α-MEM, 5% FBS, 2 mM l-glutamine, 0.2 mM l-ascorbic acid 2-phosphate, 1 ng/mL FGF 2, 10 nM dexamethasone, Pen-Strep	74	Cytodex 1
⁶⁴	Microcarrier	BM MSC	11.1	41.7	IMDM, 10% FBS, 10 ng/mL FGF-2, 2 mM l-glutamine, pen-strep	48	CultiSpher-G + Sigmacote
⁶⁵	Microcarrier	Placental MSC	20	27	20% FBS, DMEM low glucose	50	Cytodex 3 + Sigmacote
⁶⁶	Microcarrier	ADSC	2.8	11.2	StemPro MSC SFM xeno-free	226	CultiSpher-S (Gelatin) + FBS
	Microcarrier	BM MSC	4	16	StemPro MSC SFM xeno-free	168	CultiSpher-S (Gelatin) + FBS
⁴⁷	Microcarrier	BM MSC	14.8	74	MesenCultTM-XF culture	86	Synthemax^® II
³¹	Microcarrier	BM MSC	2.8	8.58	10% FBS, DMEM low glucose, 2 mM UltraGlutamine	95	P-102L (SoloHill) + FBS containing growth medium
	Microcarrier	BM MSC	10	30.1	PRIME-XV	43	P-102L (SoloHill) + Fibronectin
⁶⁷	Microcarrier	BM MSC	6.5	420	DMEM, 10% FBS, 2 mM UltraGlutamine	80	P-102L (SoloHill) + FBS containing growth medium
	Microcarrier	BM MSC	4.3	15	DMEM, 10% FBS, 2 mM UltraGlutamine	116	P-102L (SoloHill) + FBS containing growth medium
²⁶	Microcarrier	BM MSC	6.5	110	StemPro MSC SFM xeno-free medium	62	Nonporous plastic microcarriers (SoloHill) + CELLstart CTS (life tech)
	Microcarrier	ADSC	2.85	45	StemPro MSC SFM xeno-free medium	111	Nonporous plastic microcarriers (SoloHill) + CELLstart CTS (life tech)
	Microcarrier	BM MSC	6.5	104	StemPro MSC SFM xeno-free medium	62	Nonporous plastic microcarriers (SoloHill) + CELLstart CTS (life tech)
	Microcarrier	BM MSC	7.5	120	StemPro MSC SFM xeno-free medium	58	Nonporous plastic microcarriers (SoloHill) + CELLstart CTS (Life Tech)
	Microcarrier	BM MSC	18.51	200	StemPro MSC SFM xeno-free medium	63	Nonporous plastic microcarriers (SoloHill) + CELLstart CTS (Life Tech)
⁶⁸	Microcarrier	BM MSC	12	660	MesenCult–XF	80	Synthemax II
	Microcarrier	BM MSC	12	75	MesenCult–XF	80	Synthemax II
⁶⁹	Microcarrier	Fetal MSC	13.6	68	DMEM, 10% FBS, GlutaMAX	76	Cytodex 3
	Microcarrier	Fetal MSC	13	65	DMEM, 10% FBS, GlutaMAX	78	Cytodex 1
	Microcarrier	fetal MSC	7.6	38	DMEM, 10% FBS, GlutaMAX	98	CultiSpher GL
	Microcarrier	Fetal MSC	13.6	68	DMEM, 10% FBS, GlutaMAX	76	HyQspheres
	Microcarrier	Fetal MSC	12	600	DMEM, 10% FBS, GlutaMAX	54	Cytodex 3
⁷⁰	Microcarrier	BM MSC	3.9	24.1	α-MEM, 15% FBS, Pen-Strep, 2 mM glutamine	86	Cytodex 3
⁷¹	Fixed bed	hMSC-TERT	16	41.6	EMEM, 10% FBS	34	Nonporous borosilicate glass spheres
	Fixed bed	hMSC-TERT	13.6	105	EMEM, 10% FBS	45	Nonporous borosilicate glass spheres
	Fixed bed	hMSC-TERT	11.2	615	EMEM, 10% FBS	48	Nonporous borosilicate glass spheres
⁷²	Fixed bed	Hpdc	7.8	1.57	DMEM, 10% FBS, GlutaMAX, Pen-Strep	105	Ti6Al4V + FBS containing growth medium
⁷³	Fixed bed	UC MSC	38.7	348	DMEM, 20% FBS	41	Polycarbonate
⁷⁴	Microcarrier	ADSC	43	10.8	Serum-reduced (5%) Lonza	31	ProNectin F-coated SoloHill
	Microcarrier	ADSC	21.7	220	Serum-reduced (5%) Lonza	38	ProNectin F-coated SoloHill
	Microcarrier	ADSC	6.5	100	Serum-reduced (5%) Lonza	62	ProNectin F-coated SoloHill
	Microcarrier	ADSC	3.7	70	Serum-reduced (5%) Lonza	89	ProNectin F-coated SoloHill
	Microcarrier	ADSC	35.4	530	Serum-reduced (5%) Lonza	35	ProNectin F-coated SoloHill
	Microcarrier	ADSC	58	29.2	Serum-reduced (5%) Lonza	22	ProNectin F-coated SoloHill
	Microcarrier	ADSC	53	39.9	Serum-reduced (5%) Lonza	21	ProNectin F-coated SoloHill
	Microcarrier	ADSC	58	58.4	Serum-reduced (5%) Lonza	29	ProNectin F-coated SoloHill
	Microcarrier	ADSC	40	61.1	Serum-reduced (5%) Lonza	32	ProNectin F-coated SoloHill
⁷⁵	Microcarrier	hMSC-TERT	7.3	700	DMEM, 10% FBS	50	RapidCell microcarrier
⁷⁶	Microcarrier	BM MSC	71.8	1440	Serum-reduced medium	22	Na
⁷⁷	Microcarrier	BM MSC	35.3	1060	Serum-reduced medium	42	Na
²⁹	Multiplate bioreactor	hPDC	3.9	636	DMEM, 10% FBS, GlutaMAX, Pen-Strep	84	Polystyrene
⁴¹	Hollow fiber	hPDC	16.4	371	DMEM, 10% FBS, GlutaMAX, Pen-Strep	53	Fibronectin
⁷⁸	Multilayer flask	BM MSC	3.9	117	Cellgro stemgro hMSC medium	61	HYPERStack 12
⁷⁹	Multilayer flask	BM MSC	5.2	45	MSCGM™ medium	50	HYPERFlask
	Multilayer flask	BM MSC	2.8	86	MSCGM™ medium	79	HYPERStack 12
	Multilayer flask	BM MSC	1.7	154	MSCGM™ medium	155	HYPERStack 36
	Multilayer flask	BM MSC	3.9	25	MSCGM™ medium	60	2-Cell stack
	Multilayer flask	BM MSC	1.7	55	MSCGM™ medium	150	10-Cell stack
⁸⁰	Hollow fiber	BM MSC	7.6	155	α-MEM, 10% FBS	39	Fibronectin
	Hollow fiber	BM MSC	8.2	118	α-MEM, 10% HPL, 2 IU/mL heparin	50	HPL
²⁷	Hollow fiber	BM MSC	10.1	45.9	DMEM low glucose, 10% HPL, 1% N(2)-l-alanyl-l-glutamine	77	Fibronectin
⁸¹	Hollow fiber	BM MSC	18.8	154	DMEM, 10% FBS, GlutaMAX, Pen-Strep	40	Fibronectin
	Hollow fiber	BM MSC	7.5	150	DMEM, 10% FBS, GlutaMAX, Pen-Strep	59	Fibronectin
⁸²	Hollow fiber	BM MSC	6.7	133	α-MEM, 10% FBS, GlutaMAX	61	Fibronectin
⁸³	Multilayer flask	BM MSC	150	15	α-MEM, 10% FBS, 2 mM L-glutamine, Pen-Strep	50	CF-4
⁸⁴	Multilayer flask	BM MSC	27	516	DMEM low glucose, 5% HPL, 0.1 mM gentamicin, 1000 IU heparin	35	5-CellSTACK
	Multilayer flask	UC MSC	79	751	DMEM low glucose, 5% HPL, 0.1 mM gentamicin, 1000 IU heparin	19	5-CellSTACK
	Multilayer flask	UC MSC	71	114	DMEM low glucose, 5% HPL, 0.1 mM gentamicin, 1000 IU heparin	27	5-CellSTACK
⁸⁵	Multilayer flask	BM MSC	1453	436	α-MEM, 10% HPL, 2 mM L-glutamine, 2 U/mL heparin, 25 mM HEPES, Pen-Strep	30	CF-4
	Multilayer flask	BM MSC	249	74.9	α-MEM, 10% FBS, 2 mM l-glutamine, 25 mM HEPES, pen-strep	39	CF-4
⁸⁶	Multilayer flask	UC MSC	900	270	α-MEM, 10% HPL, 2 mM l-glutamine, 2 U/mL heparin, 25 mM HEPES, Pen-Strep	34	CF-4
	Multilayer flask	UC MSC	220	66	α-MEM, 10% FBS, 2 mM l-glutamine, 25 mM HEPES, Pen-Strep	43	CF-4
⁸⁷	Multilayer flask	BM MSC	10.4	259	α-MEM, 8% HPL, 1 IU Na-heparin	59	2-Chamber CellStack
	Multilayer flask	BM MSC	9.2	156	α-MEM, 8% HPL, 1 IU Na-heparin	38	2-Chamber CellStack
	Multilayer flask	BM MSC	7.2	970	α-MEM, 10% HPL, 1 IU Na-heparin	68	2-Chamber CellStack
	Multilayer flask	BM MSC	11.1	125	α-MEM, 10% HPL, 1 IU Na-heparin	37	2-Chamber CellStack
⁸⁸	Microcarrier	BM MSC	5	24	MesenCult-XF medium	24	Synthemax II
	Microcarrier	BM MSC	6.0	48	Stemgro hMSC medium	20	Synthemax II
⁸⁹	Microcarrier	UC MSC	5.8	42	DMEM low glucose, 10% HPL, 1% GlutaMAX, 1000 U/mL heparin, 1 Pen-Strep	29	SoloHill Plastic Plus
⁹⁰	Microcarrier	Fetal MSC	15.0	1200	α-MEM, 10% FBS, Pen-Strep	35	Cytodex 3

MSC, mesenchymal stem/stromal cell; BM, bone marrow; ADSC, adipose-derived stem/stromal cell; hPDC, human periosteum-derived cell, UC, umbilical cord; FBS, fetal bovine serum; HPL, human platelet lysate; FGF, fibroblast growth factor; DMEM, Dulbecco's Modified Eagle Medium; EMEM, Eagle's Minimum Essential Medium; α-MEM, Eagle's Minimum Essential Medium-Alpha Modification; IMDM, Iscove's Modified Dulbecco's Medium.

Visualizing process performance

To present a complete overview of all the processes, an interactive tool was created, where process performance maps can be plotted allowing a visual analysis of MSC culture expansion across studies. The interactive version of the visualization tool can be accessed online at the following link (mtm.kuleuven.be/prometheus/processmap). For all process performance maps presented in this work, the Y-axis represents the EF for a single passage as an approximation of the expansion efficiency. The scale of each process is reflected by the X-axis that is indicating the final cell yield of the passage. Figures 2 –4 provide additional information regarding cell density (cells/mL) at harvest represented by the size of the circles, while different colors are indicative of the specific MSC type.

FIG. 2.

Capture of an interactive process performance map on 73 different cell expansion processes. The y-scale indicates the expansion factor, the x-scale indicates the total cell yield, the color of the dots indicates the specific MSC type. The size of the circle indicates final cell density (cells/mL) before the cell harvest step. Color images available online at www.liebertpub.com/teb

FIG. 3.

Selection of the microcarrier-based MSC expansion processes present in the database. The y-scale indicates the expansion factor, the x-scale indicates the total cell yield, the color of the dots indicates the specific MSC type. The size of the circle indicates final cell density (cells/mL) before the cell harvest step. Color images available online at www.liebertpub.com/teb

FIG. 4.

Selection of the multilayer flask-based MSC expansion processes present in the database. The y-scale indicates the expansion factor, the x-scale indicates the total cell yield, the color of the dots indicates the specific MSC type. The size of the circle indicates final cell density (cells/mL) before the cell harvest step. Color images available online at www.liebertpub.com/teb

An interesting observation is that there seems to be a barrier on the process scale around a batch size greater than 1 × 10⁹ cells that is not often exceeded in the scientific MSC expansion literature. This could indicate a technological barrier (i.e., size of bioreactor or load on downstream processes) or a current lack of market demand. For certain cell types and medium combinations, this might also imply a biological barrier as certain MSC types undergo senescence after an extensive amount of PDs.³⁷

With the interactive visualization tool, it is possible to categorize the process performance maps, for example based on the type of culture vessel. In the case of microcarrier-based suspension culture (Fig. 3) a linear correlation on the log–log plot (therefore, actually a y = ax^b correlation) between the final cell yield and the expansion efficiency can be witnessed. This is surprising considering the wide range of process parameters that could affect the EF and final yield in a microcarrier-based process, for example, the type of microcarrier, microcarrier concentrations, and energy dissipation rate. Contrary, in the case of multilayered tissue flasks, where there are much less process parameters that need to be optimized, a more scattered picture is observed in Figure 4. The multilayer flask-based processes seem to show both some of the most efficient processes as well as some of the least efficient processes for a certain scale, however, average values are close to those observed in the microcarrier case. In recent work,³⁸ the relative performance of MSCs from different donors was reported to be the same for cell culture flasks and during experiments with the same cells on microcarriers.

In an alternative visualization of the process performance map (Fig. 5), it was chosen to use the same axis for scale and expansion efficiency (x and y, respectively), but while the first version focuses on the technical side of the process by providing information on the type of bioreactor and load on downstream processes, in this study, it was chosen to show more the biological side of the process. Therefore, in this case, the color of the circle represents the amount and source of protein supplemented to the medium (i.e., HPL, FBS, or serum-free medium) and the size of the circle represents the apparent PD time of the cells. The interactive format allows to sort the processes per cell type. It can be seen that the size of the circles was generally larger at the bottom of the map, indicating a longer PD time and, therefore, suggesting that the lower EF of these processes could be attributed to a slower growth rate of the cells. Interestingly, both high and low protein content medium supplements were able to support very efficient cell growth. HPL-supplemented media result in relatively small circles, indicating fast PD times, and were, therefore, generally located at an average or higher EF for a certain process scale. This indicates a more efficient cell growth in HPL-supplemented medium, which is often concluded from comparative studies between FBS- and HPL-containing media.³⁰ Notwithstanding the clear benefits of serum-free cultures, from a process efficiency point of view, certain serum-free processes reach an average EF for a certain scale, whereas most of them currently resulted in a less efficient expansion process and, therefore, seemingly required further developments to accommodate every cell type–culture vessel combination.³⁹

FIG. 5.

Capture of an interactive process performance map on 73 different cell expansion processes. The y-scale indicates the expansion factor, the x-scale indicates the total cell yield, the color of the dots indicates the source of protein supplements in the medium. The size of the circle indicates the population doubling time of the cells (larger bubbles indicate longer population doubling times). Color images available online at www.liebertpub.com/teb

Another interesting observation that could be made from Figures 2 and 5 together was that the five best-in-class processes that were able to show EFs above 100, all originate from the same research laboratory, where a notable lower cell seeding density was used in multilayered flasks (30–40 cells/cm², compared with the average over the whole database of ±3900 cells/cm²) in combination with relatively high concentrations of supplemented protein (10% HPL or FBS), and this is for two different cell types (BM MSC and UC MSC). It is regularly stated in literature that lower seeding densities promote faster PDs, therefore, in turn allowing larger and more efficient EFs.⁴⁰

Considering cost-effectiveness during bioprocess design

COGs and cost breakdown structures are generally not mentioned in articles on stem cell expansion and could, therefore, not be included in the database. However, there exists a need for more cost-effective production methods that yield large numbers of high-quality cells at a commercially viable price in order for cell therapies to be affordable for healthcare systems and reimbursement. Therefore, a cost calculation from multiple standardized large-scale expansion processes as performed multiple times in the host laboratory for a forthcoming clinical trial was included to provide a COGs reference value for the different types of processes described in the database. The calculation specifically compared the expansion of 20 million hPDCs to 350 million cells in high-glucose DMEM supplemented with 10% irradiated FBS in T175 tissue culture flasks, a hollow fiber bioreactor (Terumo BCT Quantum^® Cell Expansion System⁴¹), a multiplate bioreactor (Pall Integrity Xpansion²⁹), and in a spinner flask with CultiSpher-S microcarriers (unpublished results). Figure 6 illustrates the total cost per million of cells cultured in these vessels for the specific preclinical expansion process in the Prometheus Lab. The reagent, disposable, and labor axis provide the cost breakdown of the total cost. The average EF axis represents the EF that was obtained per vessel for the hPDCs, and is included since this significantly influences process efficiency and cost. For the final axis, the number of built-in sensors was used as a measure for the monitoring capability of the vessel, since these features are increasing the COGs, but facilitate the clinical translation.

FIG. 6.

Cost breakdown structure of a standardized large-scale cell expansion process of 20 million hPDCs to 350 million cells in high-glucose Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% irradiated FBS in T175 tissue culture flasks, a hollow fiber bioreactor (Terumo BCT Quantum^® Cell Expansion system⁴⁰ a multiplate bioreactor (Pall Integrity Xpansion²⁹), and in a spinner flask with CultiSpher-S microcarriers (unpublished data). Color images available online at www.liebertpub.com/teb

Since these processes were carried out by the same operators, with a common end-goal, and with the same cell type and medium formulation, the authors believe these cost calculations form a fair comparison between multiple systems regarding COGs and cost distribution specifically for the scale of operations at Prometheus KU Leuven. These estimations could represent a manufacturing scenario for early stages of an advanced therapy medicinal product (ATMP) development and potentially up to clinical trial phase I.

It should be emphasized that these data were generated in a university setting and, therefore, they could potentially differ from an industrial setting regarding labor costs as well as laboratory practice. Moreover, even though the cultured cells were expanded for preclinical trials, the process was not yet entirely compliant to current good manufacturing practices (GMP), which would lead to increased costs, mainly for the manual culture processes, once the setting is switched. Even so, it is expected that the trends between the different culture processes will remain relatively similar. It can be seen that closed-system bioreactors entail high COGs, mainly due to the current high cost of the disposables (expected to eventually drop as the market grows), whereas the labor costs are reduced due to partial automation of the process. At this scale, the tissue flask-based strategy is still competitive, while further scale-up will favor microcarrier-based production.

An ongoing discussion exists in the bioprocessing field regarding the design of the most cost-effective bioprocess strategies for autologous and allogeneic cell production on the basis of “scale-out” or “scale-up.” While scale-out is simply doing more of the same (parallelization of processes with the same dimensions), scale-up envisions a direct increase in volume or surface of the culture. In the case of an allogeneic therapy, it seems that a scale-up strategy in which a transition from 2D (flask-based) culture toward microcarrier-based stirred tanks is the most attractive option, and there are bioprocess economic models that support this strategy.⁶

The economics of allogeneic expansion for pluripotent stem cells (PSCs) (induced PSCs and embryonic stem cells) have been recently modeled and described both for the upstream⁴² as well as for the subsequent downstream operations.⁴³ However, this is not yet clearly addressed in the case of autologous large-scale expansion, where most likely a generic solution might not exist and optimal solutions might be case specific, based on inherent cell or donor properties as well as practical limitations (e.g., cells obtained from biopsies vs. cells required for therapy).⁸ The use of such deterministic models for the autologous case would, therefore, be a challenging task due to the inherent donor-related variability and uncertainty involved in these processes.

Integrated bioprocess design: downstream processes

MSC expansion (the upstream process) has gained considerable attention, addressing to a certain extent the scalability and GMP considerations. However, we are still far from whole bioprocessing design. Downstream processing is only recently gaining attention as a result of the increasing volumes and batches of cells produced at the upstream stages. For example, the dynamic harvest of MSCs from microcarriers in suspension reactors was recently investigated providing scalable methodologies,⁴⁴ while in a follow up study the authors linked this process to the subsequent cryopreservation step.^31,44 Dynamic harvest of single cells from fixed bed bioreactors was also recently described for the recovery of hPDCs that retained their regenerative potential in vivo⁴⁵ and the recovered viable cell fraction per step during the harvest and volume reduction process in a multiplate bioreactor was described in Lambrechts et al.²⁹ Downstream operations are gaining importance for stem cell bioprocessing, and initial discussion on separation techniques reviewed by Diogo et al.⁴⁶ are now translated increasingly in research results. An increasing number of bioprocesses and methods for the clarification and volume reduction of MSC suspensions using membranes and tangential^47,48 or dead-end⁴⁹ filtration have been recently described. Moreover, the use of expanded bed chromatography for the washing of MSC suspension resulted in improved efficiency.⁵⁰ This shows the rapid evolution of the field in reaching a pipeline of unit operations for (autologous) MSC manufacturing, customized per application, and from patient to patient. By developing similar data-based visualization tools for downstream operations, as was done here for the upstream process, a whole bioprocess design approach could be further exploited.

Rational data-based bioprocess design

As illustrated before by the large variety in the process maps, there is growing awareness on the need for standardization given the specialized and complex components involved in cell therapy research and development.⁵¹ The lack of standardization, or at least some degree of harmonization, hampers a swift transition from the development phase that is often based on trial and error, to the translational stage that requires robust and cost-effective processes.⁵² For instance, the adoption of improved standards for stem cells entering clinic⁵³ was recently suggested. Regarding input cell material, recent literature has highlighted the need for standardized MSC lines as calibration tool⁵⁴ or reference material,⁵⁵ whereas a systematic data-based approach and centralized manufacturing facilities able to conduct systematic comparability studies were advocated by McKenna et al.,⁵⁶ highlighting the invasiveness of the immortalization of the standardized MSC lines to the initial cell properties. Moreover, to address the complex regulatory landscape, a cell therapy regulatory toolkit (online regulatory resource) was introduced for new ATMPs entering clinical trials for the EU and United States of America.⁵⁷

It is clear that a certain degree of standardization would be helpful to move the field forward, however many cell-based therapies (in particular autologous therapies) will require personalized approaches, where flexibility is required to allow customization per patient or per therapy. Based on this work, we propose and illustrate the potential of an in-silico data-based approach, where data related to process performance and efficiency for MSC expansion could be concentrated, standardized, and analyzed based on objective quantitative criteria. This could help to obtain insights and conclusions regarding the translation of cell therapy and contribute toward efficient MSC bioprocessing and manufacturing. A similar data-based initiative was started recently by the Food and Drug Administration (FDA) to monitor manufacturing information of modified T-cells for cancer immunotherapies (CAR T) (http://raps.org/Regulatory-Focus/News/2016/03/16/24549/FDA-Proposes-New-Databases-to-Monitor-CAR-T-Cell-Safety-Across-INDs).

The ultimate aim is to rationalize MSC production, that is, standardizing wherever possible, but allowing calculated flexibility where needed. However, it speaks for itself that process efficiency should not be pursued at the cost of biological functionality of the cells, since a high expansion efficiency is not necessarily linked to therapeutic success. It would be extremely interesting to add a measure for the therapeutic potential of the cultured cells in this type of databases, however, currently there are not often objectively quantifiable biomarkers available for the in vivo potential of stromal (or stem) cells, either due to a lack of the proper sensors and assays, or due to a limited understanding of the mechanism of action. Even if, for a specific case, a potency assay is available, it is closely linked to the intended clinical application of the cells, therefore, preventing a fair comparison across multiple processes.

Future challenges

The authors will attempt to keep the current database updated when new relevant publications become available. Access to the database is available upon request for collaboration on further data analysis or the addition of new parameters. As more and more data are collected in the database, more in depth data mining techniques could allow for the extraction of data-driven strategies for bioprocess improvements. From the bioprocess design point of view, the incorporation of time-series data⁵⁸ or process parameters that quantitatively describe the dynamic culture environment⁵⁹ could provide much more insightful information on the dynamics as well as the robustness of MSC bioprocesses. For example, for stirred vessels, parameters such as energy dissipation rate, shear stress, or oxygen transfer characteristics could provide readouts for shear stress and mass transport properties. Unfortunately, there are very few studies reporting on these readouts for dynamic MSC expansion. For instance Nienow et al.³⁸ evaluated agitation conditions across a number of stirred reactors widely used for microcarrier-based MSC expansion. By providing this kind of information, an identification of optimal bioprocess operating conditions could be also achieved.

From the cell quality attributes point of view, although ISCT criteria provide minimum cell identification criteria,³⁵ processes should also be linked to potency assays of increased sensitivity and functionality that could however be application specific. For instance, the performance of expanded cells in an in vivo setting is usually omitted and only seldom linked to its bioprocess history. There is still a need for predictive potency assays that correlate with in vivo activity, to ensure product comparability during manufacturing changes.⁶⁰ For most of the studies included in this article, the link with the impact of process conditions on MSC in vivo performance has been largely ignored, and only a handful of studies evaluated this.

Conclusion

The steady increase in MSC production scales demonstrates the continuous maturation of the field. However, there are considerable challenges to be faced for the successful transition from early preclinical to late commercial stage manufacturing. A major factor contributing to this challenge is that there is no typical, one-size-fits-all manufacturing solution. Therefore, we present an interactive visualization tool that provides an integrated perspective on MSC expansion that is able to increase the understanding of scale-up and commercialization of cell production processes.

Footnotes

Acknowledgments

T.L. is supported by the KU Leuven Concerted Research Actions (GOA/13/016). M.S. is supported by a PhD grant of the Agency for Innovation by Science and Technology (IWT/111457). I.P. is funded by Fonds Wetenschappelijk Onderzoek (FWO fellowship, project No 12O7916N).

Disclosure Statement

No competing financial interests exist.

References

Heathman

T.R.J.

, Nienow

A.W.

, McCall

M.J.

, Coopman

, Kara

, and Hewitt

C.J.

The translation of cell-based therapies: clinical landscape and manufacturing challenges. Regen Med, 10, 49, 2015.

Trounson

, and McDonald

Stem cell therapies in clinical trials: progress and challenges. Cell Stem Cell, 17, 11, 2015.

Mason

, and Culme-Seymour

Cell therapy commercialization—transformative therapies, economic activity and reducing public healthcare costs. J Tissue Eng Regen Med, 8, 17, 2014.

Culme-Seymour

E.J.

, Edwards-Parton

, Carmen

, Folkerts

, Smith

, and Mason

A new cell therapy sector arising from the convergence of cell and gene therapy. Cytotherapy, 15, S52, 2013.

Jung

, Panchalingam

K.M.

, Wuerth

R.D.

, Rosenberg

, and Behie

L.A.

Large-scale production of human mesenchymal stem cells for clinical applications. Biotechnol Appl Biochem, 59, 106, 2012.

Simaria

A.S.

, Hassan

, Varadaraju

, Rowley

, Warren

, Vanek

, et al. Allogeneic cell therapy bioprocess economics and optimization: single-use cell expansion technologies. Biotechnol Bioeng, 111, 69, 2014.

Galipeau

The mesenchymal stromal cells dilemma-does a negative phase III trial of random donor mesenchymal stromal cells in steroid-resistant graft-versus-host disease represent a death knell or a bump in the road?. Cytotherapy, 15, 2, 2013.

Hourd

, Ginty

, Chandra

, and Williams

D.J.

Manufacturing models permitting roll out/scale out of clinically led autologous cell therapies: regulatory and scientific challenges for comparability. Cytotherapy, 16, 1033, 2014.

Kaiser

A.D.

, Assenmacher

, Schroder

, Meyer

, Orentas

, Bethke

, et al. Towards a commercial process for the manufacture of genetically modified T cells for therapy. Cancer Gene Ther, 22, 72, 2015.

10.

Salmikangas

, Menezes-Ferreira

, Reischl

, Tsiftsoglou

, Kyselovic

, Borg

J.J.

, et al. Manufacturing, characterization and control of cell-based medicinal products: challenging paradigms toward commercial use. Regen Med, 10, 65, 2015.

11.

Csaszar

, Chen

, Caldwell

, Chan

, and Zandstra

P.W.

Real-time monitoring and control of soluble signaling factors enables enhanced progenitor cell outputs from human cord blood stem cell cultures. Biotechnol Bioeng, 111, 1258, 2013.

12.

Leijten

, Chai

Y.C.

, Papantoniou

, Geris

, Schrooten

, and Luyten

F.P.

Cell based advanced therapeutic medicinal products for bone repair: keep it simple?. Adv Drug Deliv Rev, 30, 2015.

13.

dos Santos

F.F.

, Andrade

P.Z.

, da Silva

C.L.

, and Cabral

J.M.S.

Bioreactor design for clinical-grade expansion of stem cells. Biotechnol J, 8, 644, 2013.

14.

Kumar

, and Starly

Large scale industrialized cell expansion: producing the critical raw material for biofabrication processes. Biofabrication, 7, 044103, 2015.

15.

Cierpka

, Elseberg

C.L.

, Niss

, Kassem

, Salzig

, and Czermak

hMSC production in disposable bioreactors with regards to GMP and PAT. Chem Ing Tech, 85, 67, 2013.

16.

Mendicino

, Bailey

A.M.

, Wonnacott

, Puri

R.K.

, and Bauer

S.R.

MSC-based product characterization for clinical trials: an FDA perspective. Cell Stem Cell, 14, 141, 2014.

17.

Bianco

, Robey

P.G.

, and Simmons

P.J.

Mesenchymal stem cells: revisiting history, concepts, and assays. Cell Stem Cell, 2, 313, 2008.

18.

Harris

D.T.

, and Rogers

Umbilical cord blood: a unique source of pluripotent stem cells for regenerative medicine. Curr Stem Cell Res Ther, 2, 301, 2007.

19.

Ogura

, Wakao

, Kuroda

, Tsuchiyama

, Bagheri

, Heneidi

, et al. Human adipose tissue possesses a unique population of pluripotent stem cells with nontumorigenic and low telomerase activities: potential implications in regenerative medicine. Stem Cells Dev, 23, 717, 2014.

20.

De Bari

, Dell'Accio

, Tylzanowski

, and Luyten

F.P.

Multipotent mesenchymal stem cells from adult human synovial membrane. Arthritis Rheum, 44, 1928, 2001.

21.

De Bari

, Dell'Accio

, and Luyten

F.P.

Human periosteum-derived cells maintain phenotypic stability and chondrogenic potential throughout expansion regardless of donor age. Arthritis Rheum, 44, 85, 2001.

22.

Heathman

T.R.J.

, Rafiq

Q.A.

, Chan

A.K.C.

, Coopman

, Nienow

A.W.

, Kara

, et al. Characterisation of human mesenchymal stem cells from multiple donors and the implications for large scale bioprocess development. Biochem Eng J, 108, 14, 2016.

23.

Fennema

E.M.

, Renard

A.J.S.

, Leusink

, van Blitterswijk

C.A.

, and de Boer

The effect of bone marrow aspiration strategy on the yield and quality of human mesenchymal stem cells. Acta Orthop, 80, 618, 2009.

24.

Oedayrajsingh-Varma

M.J.

, van Ham

S.M.

, Knippenberg

, Helder

M.N.

, Klein-Nulend

, Schouten

T.E.

, et al. Adipose tissue-derived mesenchymal stem cell yield and growth characteristics are affected by the tissue-harvesting procedure. Cytotherapy, 8, 166, 2006.

25.

Sotiropoulou

P.A.

, Perez

S.A.

, Salagianni

, Baxevanis

C.N.

, and Papamichail

Characterization of the optimal culture conditions for clinical scale production of human mesenchymal stem cells. Stem Cells, 24, 462, 2006.

26.

Dos Santos

, Campbell

, Fernandes-Platzgummer

, Andrade

P.Z.

, Gimble

J.M.

, Wen

, et al. A xenogeneic-free bioreactor system for the clinical-scale expansion of human mesenchymal stem/stromal cells. Biotechnol Bioeng, 111, 1116, 2014.

27.

Nold

, Brendel

, Neubauer

, Bein

, and Hackstein

Good manufacturing practice-compliant animal-free expansion of human bone marrow derived mesenchymal stroma cells in a closed hollow-fiber-based bioreactor. Biochem Biophys Res Commun, 430, 325, 2013.

28.

Timmins

N.E.

, Palfreyman

, Marturana

, Dietmair

, Luikenga

, Lopez

, et al. Clinical scale ex vivo manufacture of neutrophils from hematopoietic progenitor cells. Biotechnol Bioeng, 104, 832, 2009.

29.

Lambrechts

, Papantoniou

, Viazzi

, Bovy

, Schrooten

, Luyten

F.P.

, et al. Evaluation of a monitored multiplate bioreactor for large-scale expansion of human periosteum derived stem cells for bone tissue engineering applications. Biochem Eng J, 108, 58, 2015.

30.

Tan

K.Y.

, Teo

K.L.

, Lim

J.F.

, Chen

A.K.

, Reuveny

, and Oh

S.K.

Serum-free media formulations are cell line-specific and require optimization for microcarrier culture. Cytotherapy, 17, 1152, 2015.

31.

Heathman

T.R.J.

, Glyn

V.A.M.

, Picken

, Rafiq

Q.A.

, Coopman

, Nienow

A.W.

, et al. Expansion, harvest and cryopreservation of human mesenchymal stem cells in a serum-free microcarrier process. Biotechnol Bioeng, 112, 1696, 2015.

32.

Gottipamula

, Muttigi

M.S.

, Kolkundkar

, and Seetharam

R.N.

Serum-free media for the production of human mesenchymal stromal cells: a review. Cell Prolif, 46, 608, 2013.

33.

Sart

, Agathos

S.N.

, and Li

Engineering stem cell fate with biochemical and biomechanical properties of microcarriers. Biotechnol Prog, 29, 1354, 2013.

34.

, Tsai

A.C.

, and Liu

Biomanufacturing of human mesenchymal stem cells in cell therapy: influence of microenvironment on scalable expansion in bioreactors. Biochem Eng J [Epub ahead of print]; DOI: 10.1016/j.bej.2015.07.014, 2015.

35.

Dominici

, Le Blanc

, Mueller

, Slaper-Cortenbach

, Marini

, Krause

, et al. Minimal criteria for defining multipotent mesenchymal stromal cells. The International Society for Cellular Therapy position statement. Cytotherapy, 8, 315, 2006.

36.

Sharma

R.R.

, Pollock

, Hubel

, and McKenna

Mesenchymal stem or stromal cells: a review of clinical applications and manufacturing practices. Transfusion, 54, 1418, 2014.

37.

Larson

B.L.

, Ylostalo

, Lee

R.H.

, Gregory

, and Prockop

D.J.

Sox11 Is expressed in early progenitor human multipotent stromal cells and decreases with extensive expansion of the cells. Tissue Eng Part A, 16, 3385, 2010.

38.

Nienow

A.W.

, Hewitt

C.J.

, Heathman

T.R.J.

, Glyn

V.A.M.

, Fonte

G.N.

, Hanga

M.P.

, et al. Agitation conditions for the culture and detachment of hMSCs from microcarriers in multiple bioreactor platforms. Biochem Eng J, 108, 24, 2016.

39.

Tan

K.Y.

, Reuveny

, and Oh

S.K.

Recent advances in serum-free microcarrier expansion of mesenchymal stromal cells: parameters to be optimized. Biochem Biophys Res Commun, 473, 769, 2015.

40.

Balint

, Richardson

S.M.

, and Cartmell

S.H.

Low-density subculture: a technical note on the importance of avoiding cell-to-cell contact during mesenchymal stromal cell expansion. J Tissue Eng Regen Med, 9, 1200, 2015.

41.

Lambrechts

, Papantoniou

, Rice

, Schrooten

, Luyten

F.P.

, and Aerts

J.M

. Large-scale autologous stem cell expansion in monitored hollow fibre bioreactors: evaluation of donor-to-donor variability. Cytotherapy [Epub ahead of print]; DOI: 10.1016/j.jcyt.2016.05.013, 2016.

42.

Jenkins

M.J.

, and Farid

S.S.

Human pluripotent stem cell-derived products: advances towards robust, scalable and cost-effective manufacturing strategies. Biotechnol J, 10, 83, 2015.

43.

Hassan

, Simaria

A.S.

, Varadaraju

, Gupta

, Warren

, and Farid

S.S.

Allogeneic cell therapy bioprocess economics and optimization: downstream processing decisions. Regen Med, 10, 591, 2015.

44.

Nienow

A.W.

, Rafiq

Q.A.

, Coopman

, and Hewitt

C.J.

A potentially scalable method for the harvesting of hMSCs from microcarriers. Biochem Eng J, 85, 79, 2014.

45.

Sonnaert

, Luyten

F.P.

, Schrooten

, and Papantoniou

Bioreactor-based online recovery of human progenitor cells with uncompromised regenerative potential: a bone tissue engineering perspective. PloS One, 10, e0136875, 2015.

46.

Diogo

M.M.

, da Silva

C.L.

, and Cabral

J.M.S.

Separation technologies for stem cell bioprocessing. Biotechnol Bioeng, 109, 2699, 2012.

47.

Cunha

, Aguiar

, Silva

M.M.

, Silva

R.J.S.

, Sousa

M.F.Q.

, Pineda

, et al. Exploring continuous and integrated strategies for the up- and downstream processing of human mesenchymal stem cells. J Biotechnol, 213, 97, 2015.

48.

Cunha

, Peixoto

, Silva

M.M.

, Carrondo

M.J.T.

, Serra

, and Alves

P.M.

Filtration methodologies for the clarification and concentration of human mesenchymal stem cells. J Membr Sci, 478, 117, 2015.

49.

Tostoes

, Dodgson

, Mason

, and Veraitch

A novel filtration device for point of care preparation of cellular therapies. Cytotherapy, 17, S26, 2015.

50.

Cunha

, Silva

R.J.

, Aguiar

, Serra

, Daicic

, Maloisel

J.L.

, et al. Improving washing strategies of human mesenchymal stem cells using negative mode expanded bed chromatography. J Chromatogr A, 1429, 292, 2016.

51.

Bravery

C.A.

, and French

Reference materials for cellular therapeutics. Cytotherapy, 16, 1187, 2014.

52.

Kinzebach

, and Bieback

Expansion of mesenchymal stem/stromal cells under xenogenic-free culture conditions. Adv Biochem Eng Biotechnol, 129, 33, 2013.

53.

Dolgin

Better standards sought for range of stem cells entering the clinic. Nat Med, 20, 797, 2014.

54.

Deans

Towards the creation of a standard MSC line as a calibration tool. Cytotherapy, 17, 1167, 2015.

55.

Viswanathan

, Keating

, Deans

, Hematti

, Prockop

, Stroncek

D.F.

, et al. Soliciting strategies for developing cell-based reference materials to advance mesenchymal stromal cell research and clinical translation. Stem Cells Dev, 23, 1157, 2014.

56.

McKenna

, Matthay

M.A.

, and Pati

Correspondence to: soliciting strategies for developing cell-based reference materials to advance mesenchymal stem/stromal cell research and clinical translation. Stem Cells Dev, 23, 1717, 2014.

57.

Culme-Seymour

E.J.

, Davies

J.L.

, Hitchcock

, Mason

, Carpenter

M.K.

, and Mason

Cell therapy regulatory toolkit: an online regulatory resource. Regen Med, 10, 531, 2015.

58.

Viazzi

, Lambrechts

, Schrooten

, Papantoniou

, and Aerts

J.M.

Real-time characterisation of the harvesting process for adherent mesenchymal stem cell cultures based on on-line imaging and model-based monitoring. Biosyst Eng, 138, 104, 2015.

59.

Lambrechts

, Papantoniou

, Sonnaert

, Schrooten

, and Aerts

J.M.

Model-based cell number quantification using online single-oxygen sensor data for tissue engineering perfusion bioreactors. Biotechnol Bioeng, 111, 1982, 2014.

60.

Bravery

C.A.

, Carmen

, Fong

, Oprea

, Hoogendoorn

K.H.

, Woda

, et al. Potency assay development for cellular therapy products: an ISCT review of the requirements and experiences in the industry. Cytotherapy, 15, 9, 2013.

61.

Hanley

P.J.

, Mei

Z.Y.

, Durett

A.G.

, Cabreira-Harrison

M.D.

, Klis

, Li

, et al. Efficient manufacturing of therapeutic mesenchymal stromal cells with the use of the quantum cell expansion system. Cytotherapy, 16, 1048, 2014.

62.

Tsai

A.C.

, Liu

, and Ma

Expansion of human mesenchymal stem cells in fibrous bed bioreactor. Biochem Eng J, 108, 51, 2015.

63.

Eibes

, dos Santos

, Andrade

P.Z.

, Boura

J.S.

, Abecasis

M.M.A.

, da Silva

C.L.

, et al. Maximizing the ex vivo expansion of human mesenchymal stem cells using a microcarrier-based stirred culture system. J Biotechnol, 146, 194, 2010.

64.

Sun

L.Y.

, Hsieh

D.K.

, Syu

W.S.

, Li

Y.S.

, Chiu

H.T.

, and Chiou

T.W.

Cell proliferation of human bone marrow mesenchymal stem cells on biodegradable microcarriers enhances in vitro differentiation potential. Cell Prolif, 43, 445, 2010.

65.

Hewitt

C.J.

, Lee

, Nienow

A.W.

, Thomas

R.J.

, Smith

, and Thomas

C.R.

Expansion of human mesenchymal stem cells on microcarriers. Biotechnol Lett, 33, 2325, 2011.

66.

dos Santos

, Andrade

P.Z.

, Abecasis

M.M.

, Gimble

J.M.

, Chase

L.G.

, Campbell

A.M.

, et al. Toward a clinical-grade expansion of mesenchymal stem cells from human sources: a microcarrier-based culture system under xeno-free conditions. Tissue Eng Part C Methods, 17, 1201, 2011.

67.

Rafiq

Q.A.

, Brosnan

K.M.

, Coopman

, Nienow

A.W.

, and Hewitt

C.J.

Culture of human mesenchymal stem cells on microcarriers in a 5 l stirred-tank bioreactor. Biotechnol Lett, 35, 1233, 2013.

68.

Sousa

M.F.Q.

, Silva

M.M.

, Roldao

, Alves

P.M.

, Serra

, Giroux

, et al. Production of oncolytic adenovirus and human mesenchymal stem cells in a single-use, vertical-wheel bioreactor system: impact of bioreactor design on performance of microcarrier-based cell culture processes. Biotechnol Prog, 31, 1600, 2015.

69.

Goh

T.K.

, Zhang

Z.Y.

, Chen

A.K.

, Reuveny

, Choolani

, Chan

J.K.

, et al. Microcarrier culture for efficient expansion and osteogenic differentiation of human fetal mesenchymal stem cells. BioRes Open Access, 2, 84, 2013.

70.

Caruso

S.R.

, Orellana

M.D.

, Mizukami

, Fernandes

T.R.

, Fontes

A.M.

, Suazo

C.A.

, et al. Growth and functional harvesting of human mesenchymal stromal cells cultured on a microcarrier-based system. Biotechnol Prog, 30, 889, 2014.

71.

Weber

, Freimark

, Portner

, Pino-Grace

, Pohl

, Wallrapp

, et al. Expansion of human mesenchymal stem cells in a fixed-bed bioreactor system based on non-porous glass carrier—part A: inoculation, cultivation, and cell harvest procedures. Int J Artif Organs, 33, 512, 2010.

72.

Sonnaert

, Papantoniou

, Luyten

F.P.

, and Schrooten

Perfusion bioreactor expansion and dynamic harvest preserve in vivo bone forming capacity of osteoprogenitor cells. J Tissue Eng Regen Med, 8, 48, 2014.

73.

Reichardt

, Polchow

, Shakibaei

, Henrich

, Hetzer

, and Lueders

Large scale expansion of human umbilical cord cells in a rotating bed system bioreactor for cardiovascular tissue engineering applications. Open Biomed Eng J, 7, 50, 2013.

74.

Schirmaier

, Jossen

, Kaiser

S.C.

, Jungerkes

, Brill

, Safavi-Nab

, et al. Scale-up of adipose tissue-derived mesenchymal stem cell production in stirred single-use bioreactors under low-serum conditions. Eng Life Sci, 14, 292, 2014.

75.

Elseberg

C.L.

, Leber

, Salzig

, Wallrapp

, Kassem

, Kraume

, et al. Microcarrier-based expansion process for hMSCs with high vitality and undifferentiated characteristics. Int J Artif Organs, 35, 93, 2012.

76.

Kaiser

S.C.

Characterization and of single-use bioreactors and biopharmaceutical production processes using computational fluid dynamics Doctoral Degree Fakultät III—Prozesswissenschaften, Technische Universität Berlin, 2015.

77.

Jossen

, Kaiser

S.C.

, Schirmaier

, Herrmann

, Tappe

, Eibl

, et al. Modification and qualification of a stirred single-use bioreactor for the improved expansion of human mesenchymal stem cells at benchtop scale. Pharmaceut Bioprocess, 2, 311, 2014.

78.

Pardo

A.M.P.

, Sherman

H.A.

, Strathearn

K.E.

, and Rothenberg

M.E.

Xeno-free expansion of functional hMSCs using corning^® HYPER technology and corning cellgro^® stemgro^® hMSC medium, 2012. Available at: http://csmedia2.corning.com/LifeSciences/media/pdf/ISCT_Sept2013_Poster.pdf (accessed August 9, 2016).

79.

Pardo

A.M.P.

, and Rothenberg

M.E.

Expansion of human mesenchymal stem cells using corning^® hyper technology cell culture vessels. Technical report from the Corning development group, 2012. http://csmedia2.corning.com/LifeSciences/media/pdf/snappshots_208_Expansn_of_hMSC_Using_Corning_HYPER_Technology%20CC_Vessels.pdf

80.

Rojewski

M.T.

, Fekete

, Baila

, Nguyen

, Furst

, Antwiler

, et al. GMP-compliant isolation and expansion of bone marrow-derived MSCs in the closed, automated device quantum cell expansion system. Cell Transplant, 22, 1981, 2013.

81.

Lechanteur

, Baila

, Janssens

M.E.

, Giet

, Briquet

, Baudoux

, et al. Large-scale clinical expansion of mesenchymal stem cells in the GMP-compliant, closed automated quantum^® cell expansion system: comparison with expansion in traditional T-flasks. J Stem Cell Res Ther, 4, 1, 2014.

82.

Jones

, Varella-Garcia

, Skokan

, Bryce

, Schowinsky

, Peters

, et al. Genetic stability of bone marrow-derived human mesenchymal stromal cells in the quantum system. Cytotherapy, 15, 1323, 2013.

83.

Bartmann

, Rohde

, Schallmoser

, Purstner

, Lanzer

, Linkesch

, et al. Two steps to functional mesenchymal stromal cells for clinical application. Transfusion, 47, 1426, 2007.

84.

Capelli

, Pedrini

, Valgardsdottir

, Da Roit

, Golay

, and Introna

Clinical grade expansion of MSCs. Immunol Lett, 168, 222, 2015.

85.

Schallmoser

, Bartmann

, Rohde

, Reinisch

, Kashofer

, Stadelmeyer

, et al. Human platelet lysate can replace fetal bovine serum for clinical-scale expansion of functional mesenchymal stromal cells. Transfusion, 47, 1436, 2007.

86.

Reinisch

, Bartmann

, Rohde

, Schallmoser

, Bjelic-Radisic

, Lanzer

, et al. Humanized system to propagate cord blood-derived multipotent mesenchymal stromal cells for clinical application. Regen Med, 2, 371, 2007.

87.

Fekete

, Rojewski

M.T.

, Furst

, Kreja

, Ignatius

, Dausend

, et al. GMP-compliant isolation and large-scale expansion of bone marrow-derived MSC. PloS One, 7, e43255, 2012.

88.

Hervy

, Weber

J.L.

, Pecheul

, Dolley-Sonneville

, Henry

, Zhou

, et al. Long term expansion of bone marrow-derived hMSCs on novel synthetic microcarriers in xeno-free, defined conditions. PloS One, 9, e92120, 2014.

89.

Petry

F.S.J.

, Leber

, Salzig

, Czermak

, and Weiss

M.L.

Manufacturing of human umbilical cord mesenchymal stromal cells on microcarriers in a dynamic system for clinical use. Stem Cells Int, 2016, 4834616, 2016.

90.

Chen

A.K.L.

, Chew

Y.K.

, Tan

H.Y.

, Reuveny

, and Oh

S.K.W.

Increasing efficiency of human mesenchymal stromal cell culture by optimization of microcarrier concentration and design of medium feed. Cytotherapy, 17, 163, 2015.