Isolation and Expansion of Placental Derived Mesenchymal Stromal Cells in a Packed Bed Bioreactor

ISOLATION AND EXPANSION OF PLACENTAL DERIVED MESENCHYMAL STROMAL CELLS IN A PACKED BED BIOREACTOR

Michael Joseph Russo Osiecki BEng (Chemical), BBiotech (Process Technology) Hons.

William B. Lott

Emad Kiriakous Izake

Submitted in fulfilment of the requirements for the degree of Doctor Philosophy

Institute of Health and Biomedical Innovation Science and Engineering Faculty Queensland University of Technology 2016

Keywords

Bioreactor, packed/fixed bed bioreactor, mesenchymal stromal/stem/progenitor cell, expansion device, cell culture expansion, stem cell isolation, placenta derived mesenchymal stromal cells, mathematical modelling, finite element model, 2D axisymmetric model, COMSOL, mass and momentum transport, conditioned medium, biomaterial testing, plasma surface modification.

Isolation and Expansion of Placental Derived Mesenchymal Stromal Cells in a Packed Bed Bioreactor i

Abstract

Large numbers of Mesenchymal stem/stromal cells (MSCs) are required to obtain clinically relevant doses to treat a number of diseases. To economically manufacture these MSCs, an automated bioreactor system will be required. This thesis describes the development of a scalable closed-system, packed bed bioreactor that is suitable for large-scale MSC expansion. The packed bed was formed from fused polystyrene pellets that were air plasma treated to endow them with a surface chemistry similar to traditional tissue culture plastic. The packed bed was encased within a gas permeable shell to decouple the oxygen supply from the bulk nutrient flow. This enabled a significant reduction in medium flow rates, thus potentially reducing shear and even facilitating single pass medium exchange. The system was optimised in a small-scale bioreactor format (total surface area approximately 160 cm2) with murine-derived green fluorescent protein expressing MSCs, and then scaled- up to a 2800 cm2 format which successfully expanded 1.12x108 murine MSCs. Theoretically this large scale bioreactor could support up to 5.6x107 human MSCs in the current format. To demonstrate the bioreactors’ ability to expand primary human cells, pre-isolated passage four placental derived MSCs (pMSCs) were expanded in the bioreactor. Our bioreactor achieved a 10-fold expansion of pMSCs after a one week culture while maintaining their mesodermal tri-linage differential potential. To demonstrate the efficiency of our bioreactor, pMSCs were isolated and expanded directly onto the bioreactor from digested placental tissue without requiring a pre- isolation step. These pMSCs also maintained their tri-lineage differentiation potential.

Although the biological function of MSCs was preserved throughout the expansion process, MSCs exhibit a reduced cell growth rate in bioreactor expansion devices compared to traditional flask cultures. To characterise and understand this phenomenon, a two dimensional axial-symmetric model was developed that accounts for the influence of glucose and oxygen (essential nutrients) and lactate (a waste product) on the cell growth rate, assuming Michaelis-Menten kinetics. The model also considered the oxygen diffusion through the gas permeable wall of the bioreactor, assuming a constant external oxygen concentration. The model suggested that our

ii Isolation and Expansion of Placental Derived Mesenchymal Stromal Cells in a Packed Bed Bioreactor

design, which allowed for oxygen diffusion through the bioreactor wall, would be superior to traditional packed bed bioreactor systems that require a higher medium perfusion rate to provide an adequate oxygen supply. The model demonstrated that this diffusion concept is scalable in bioreactor systems by adding a central gas permeable capillary to increase the oxygen supply as the format increases. Importantly, the model predicted that lactate accumulation would have a far greater effect on cell growth than the oxygen and glucose concentrations, especially in the small scale bioreactor. However, the base model did not accurately predict our experimental results. The model was then modified to account for an initial non-uniform cell distribution, to represent this a linear cell distribution was modelled in the bioreactor in both the radial and vertical directions. This adjustment significantly affected the total cell number, and the model more closely agreed with the experimental results. To better fit the data, the model was further adjusted to combine the non-uniform cell distribution with a cell growth lag phase due to the shock of the seeding process. This simulation output matched the experimental data very well. This suggests that the most important parameters for achieving optimal cell expansion in the proliferation of MSCs in our bioreactor design are efficient initial seeding, achieving as near homogeneous a cell distribution as possible, and the gentle handling of the cells during the seeding process.

In summation, our packed bed bioreactor design, which allows the decoupling of oxygen from the bulk nutrient flow through wall diffusion, can successfully isolate and expand placental derived MSCs. This design was shown to be scalable to a total growth area of 2800cm2. Mathematical modelling of this bioreactor design confirmed that this design is scalable and supported the hypothesis that a significantly lower flow rate can be used, due to the oxygen diffusion through the bioreactor wall, compared to a traditional perfusion packed bed bioreactor. The mathematical model also identified the importance of cell seeding distribution on the overall final cell numbers and growth rate, thereby explaining the significantly reduced cell growth rate observed in the bioreactor relative to traditional flask cultures.

Isolation and Expansion of Placental Derived Mesenchymal Stromal Cells in a Packed Bed Bioreactor iii

Table of Contents

Keywords ...... i Abstract ...... ii Table of Contents ...... iv List of Figures ...... vii List of Tables ...... xii List of Abbreviations ...... xiii Statement of Original Authorship ...... xv Acknowledgements ...... xvi Publications ...... xviii Chapter 1: Introduction and Literature Review ...... 1 1.1 Mesenchymal Stromal Cell Characterisation ...... 2 1.2 Mesenchymal Stromal Cell Mechanisms to Repair Damaged Tissue ...... 4 1.2.1 Mesenchymal Stromal Cell Differentiation and Transdifferentiation ...... 5 1.2.2 Mesenchymal Stromal Cells are Immunosuppressive, but not Immune Privileged...... 7 1.2.3 Mesenchymal Stromal Cells Secrete Soluble Factors that Promote Cell Survival, Proliferation and Immunomodulatory Effects ...... 9 1.2.4 Mesenchymal Stromal Cell Homing and Migration ...... 10 1.3 Clinical Applications of Mesenchymal Stromal Cells ...... 11 1.3.1 Cardiac Failure ...... 14 1.3.2 Stroke ...... 15 1.3.3 Neural and Spinal Cord Repair ...... 16 1.3.4 Crohn’s Disease and Multiple Sclerosis ...... 17 1.3.5 Graft Versus Host Disease ...... 18 1.3.6 Mesenchymal Stromal Cell role in Enhancing Wound Healing ...... 19 1.4 Ex Vivo Expansion of Mesenchymal Stromal Cells for Clinical Use ...... 21 1.4.1 Wave Bag Bioreactors ...... 25 1.4.2 Stirred Tank Bioreactors ...... 26 1.4.3 Hollow Fibre Bioreactors ...... 27 1.4.4 Packed Bed Bioreactors ...... 28 1.5 Summary and Thesis Structure ...... 29 Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall ...... 31 2.1 Methods ...... 34 2.1.1 Single Pass Small-Scale Bioreactor Design ...... 34 2.1.2 Recirculating Large-Scale Bioreactor Design ...... 36 2.1.3 Large Surface Area Scaffold ...... 36 2.1.4 Prevention of Bubble Formation in Bioreactor System ...... 37 2.1.5 Bioreactor Sterilisation...... 37 2.1.6 Air Plasma Treated Polystyrene Scaffold Characterisation ...... 37 2.1.7 Cell Culture ...... 38

iv Isolation and Expansion of Placental Derived Mesenchymal Stromal Cells in a Packed Bed Bioreactor

2.1.8 Comparison Cell Growth Rate of Glass Packing Material and Tissue Culture Plastic ...... 38 2.1.9 Comparison of Air Plasma Treated Polystyrene and Commercial Tissue Culture Plastic on Cell Growth Rate and Attachment...... 39 2.1.10 Packed Bed Bioreactor Seeding Method...... 39 2.1.11 Expansion of GFP-mMSCs in the Packed Bed Bioreactor ...... 40 2.1.12 Packed Bed Bioreactor Harvest and Harvest Efficiency ...... 40 2.1.13 Packed Bed Bioreactor Imaging ...... 40 2.1.14 Scale-up of GFP-mMSC in Bioreactor ...... 41 2.1.15 Statistics...... 42 2.2 Results ...... 43 2.2.1 Glass Bioreactor Development ...... 43 2.2.2 Plasma Treated Polystyrene Scaffold Characterisation ...... 44 2.2.3 GFP-mMSCs Expansion in Small-scale Packed Bed Bioreactor ...... 45 2.2.4 Scale-up of the Packed Bed Bioreactor ...... 49 2.3 Discussion ...... 50 Chapter 3: Bioreactor Application: Isolation and Expansion of Placental Derived Cells on a Packed Bed Bioreactor ...... 53 3.1 Methods ...... 54 3.1.1 Bioreactor set-up ...... 54 3.1.2 Cell Culture ...... 54 3.1.3 Placental and Bone Marrow Conditioned Medium ...... 56 3.1.4 HaCaT growth in response from MSC conditioned media ...... 57 3.1.5 Pre-Isolated pMSCs Expansion in the Packed Bed Bioreactor ...... 57 3.1.6 pMSC Isolation and Expansion in the Packed Bed Bioreactor ...... 57 3.1.7 Tri-Linage Mesodermal Differentiation ...... 58 3.1.8 Statistics...... 59 3.2 Results ...... 60 3.2.1 Difference in tri-lineage Differentiation Potential of bmMSCs and pMSCs ...... 60 3.2.2 Difference in pMSC and bmMSC conditioned media on HaCaT growth ...... 61 3.2.3 Human Placental MSCs Expansion in Small-Scale Packed Bed Bioreactor ...... 63 3.2.4 Isolation and Expansion of Placental MSCs Directly in a Small-Scale Packed Bed Bioreactor ...... 65 3.3 Discussion ...... 67 Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor . …………………………………………………………………….71 4.1 Development of the Mathematical Model ...... 74 4.1.1 Model Description ...... 74 4.1.2 Development of Model Equations ...... 76 4.1.3 Initial Conditions and Parameters ...... 81 4.1.4 Solution Method ...... 84 4.2 Results and Discussion ...... 84 4.2.1 Base Case- Decoupling the Oxygen Supply from the Bulk Medium is Superior to a Traditional Packed Bed Bioreactor ...... 84 4.2.2 Investigating the Effect of Cell Distribution on Growth ...... 91 4.2.3 Modelling a Batch Bioreactor ...... 98 4.2.4 Using the Model to Investigate Scaling Up ...... 100

Isolation and Expansion of Placental Derived Mesenchymal Stromal Cells in a Packed Bed Bioreactor v

Chapter 5: Summary of Findings and Future Directions ...... 109 References ...... 117 Appendices ...... 143

vi Isolation and Expansion of Placental Derived Mesenchymal Stromal Cells in a Packed Bed Bioreactor

List of Figures

Figure 1: Observed mechanisms that MSCs repair damaged tissue. Adapted from [33] ...... 5 Figure 2 Summary of MSCs’ interactions with immune cells. MSCs secrete many soluble factors that alter the function and proliferation of the immune cells and regulatory T cells. Adapted from [89,99,100]...... 8 Figure 3 Bioreactor geometry and mass transfer diagrams indicating mass transfer for various bioreactor geometries: (A) Media is mixed by a magnetic impeller. (B) Media inside a rotating wall bioreactor has the same angular velocity as the wall. (C) Bag cultures on a rocker generate a wave motion which mixes cell suspension and medium. (D) Plug flow is maintained throughout the bed to generate uniform perfusion. (E) Cells are generally grown on the extra capillary space either in suspension or in a gel scaffold; media is perfused in the intra capillary space to provide nutrients and oxygen through the semipermeable wall. Taken from [211]...... 23 Figure 4 The reaction between polystyrene and oxygen plasma on surfaces results in a variety of chemical groups and breaking of the chain back bone or opening of benzene rings (not shown) that results in a more hydrophilic surface. Adapted from [281]...... 32 Figure 5 (A) Process flow diagrams of the single pass small-scale 160cm2 bioreactor and (B) recirculating perfusion for the large-scale 2800 cm2 bioreactor. (C) Picture of 160 cm2 left and 2800 cm2 right bioreactor scaffolds made from 2.5 mm polystyrene pellets (scale bar is 10 mm). (D) 2800 cm2 bioreactor inside the incubator...... 35 Figure 6 (A) Comparison of cell density of mouse GFP-mMSC between glass and tissue culture plastic (TCP), mean + SD of n=4. (B & C) IVIS images of two separate glass packed bed bioreactors containing mouse GFP-mMSC stained with PI. Fluorescence intensity ranges from low (dark red) to high (yellow). The bright spot in C is due to autofluorescence of a large deposit of silicon glue. (D, E, F & G) Fluorescence microscopy images of GFP-mMSC growing on glass beads...... 44 Figure 7 (A) Surface composition of the air plasma treated polystyrene scaffold determined by XPS, commercial TCP based on literature value [265]. (B) Cell density of GFP-mMSC attachment after one and half hours, initially seeded at 2000 cells/cm2 in a 60 mm plasma treated petri dish or a T75 flask control (n=4, mean + SD). (C) The growth after 3 days of culture of GFP-mMSC seeded at 1000 cells/cm2 in a 60 mm plasma treated petri dish or a T75 flask control (n=6, mean + SD)...... 45 Figure 8 (A) The fold expansion of GFP-mMSCs seeded at 1000 cell/cm2 in four bioreactor (BR) under static conditions and, (B) under 5 mL/day perfusion, compared to four T175 flask controls (2D). (C) Cell harvest

Isolation and Expansion of Placental Derived Mesenchymal Stromal Cells in a Packed Bed Bioreactor vii

recovery and viability from the bioreactor undergoing 5mL/day perfusion...... 46 Figure 9 (A) Seeding efficiency of using the roller method compared to just injecting the cells and allowing them to attach under static conditions, (n=4, mean + SD, * p<0.01). (B) IVIS imaging of the fluorescent intensity of PI stained GFP-mMSCs in the bioreactor under static and (C) 5 mL/day perfusion conditions after roller seeding method. (D, E & F) IVIS images of static expanded GFP-mMSCs using the static seeding method. IVIS images are a red (low) / yellow (high) heat map of fluorescent intensity. (G, H, I & J) Fluorescent microscopy showed that the GFP-mMSCs attached to the scaffold after cell expansion using the roller seeding method (scale bar is 500 µm,)...... 48 Figure 10 (A) The fold expansion of GFP-mMSCs in a scaled-up bioreactor under 0.5 ml/min perfusion with T175 flask control (n=4, mean + SD, * p<0.01). (B) Glucose and lactate levels in the bioreactor (n=3, mean + SD). (C, D, E & F) IVIS imaging of the fluorescent intensity of PI stained GFP-mMSCs in the bioreactor...... 49 Figure 11 (A) Tri-lineage potential of pMSC and bmMSC was assessed by staining and visualising the lipid droplets (Oil Red O, adipogenic), bone nodules (Alizarin Red, osteogenic) and GAG in the pellet (Alcian blue, chondrogenic). (B) The levels of triglycerides were measured to indicate the degree of adipogenic differentiation. (C) The activity of ALP was measured to indicate the degree of osteogenic differentiation and (D) the chondrogenic differentiation was assessed by the level of GAG present in the pellet. *indicated p<0.01 with n=4 in all conditions...... 60 Figure 12 (A) HaCaT growth in response to MSC conditioned medium of different passage number after 4 days (n=4). (B) Average HaCaT density achieved for pMSCs and bmMSCs from Figure 12A removing the outliers of bmMSC p8 & 10 and pMSC p10 (n=4). * Significant (p<0.01) increase over XVIVO 15 negative control...... 62 Figure 13 (A) HaCaT growth in response to concentrated conditioned medium after 6 days. (B) HaCaT growth in response to cell aggregates after 6 days. * Significant (p<0.01) increase over XVIVO 15 negative control, # significant (p<0.01) increase in HaCaT growth over traditional conditioned medium...... 63 Figure 14 (A) pMSC expansion in the small-scale 160 cm2 bioreactor (BR) in static and (B) 5 mL/day perfusion (n=4 mean + SD), with the 2D controls (2D). (C) IVIS imaging of PI stained pMSCs under perfusion conditions. IVIS images are a red (low) / yellow (high) heat map of fluorescent intensity...... 64

viii Isolation and Expansion of Placental Derived Mesenchymal Stromal Cells in a Packed Bed Bioreactor

Figure 15 (A) Two week tri-lineage differentiation induction of bioreactor expanded pMSCs and 2D controls down the adipogenic (Oil Red O, 10x, scale bar is 100 µm), osteogenic (Alizarin Red, 5x, scale bar 500 µm) and chondrogenic (Alcian Blue, 10x, scale bar 500 µm) lineages. Quantification of (B) triglycerides (n=4, mean + SD), (C) ALP activity (n=4, mean + SD) and (D) GAG production (n=4, mean + SD). * indicates p<0.01...... 65 Figure 16 (A) Cell growth of pMSCs isolated directly from a placenta in the 160 cm2 packed bed bioreactor undergoing 5 mL/day perfusion (n=4, mean + SD). On day 13, the cells were redistributed back into the same flask or bioreactor. # indicating no 2D control result was measured as the flasks were harvested on day 18. (B) Two week tri- lineage differentiation induction of bioreactor expanded (BR) pMSCs and 2D controls down the adipogenic (Oil Red O, 10x, scale bar is 100 µm), osteogenic (Alizarin Red, 5x, scale bar 500 µm) and chondrogenic (Alcian Blue, 10x, scale bar 500 µm) lineages. Quantification of (C) triglycerides (n=4, mean + SD), (D) ALP activity (n=4, mean + SD) and (E) GAG (n=4, mean + SD) production compared to its matching 2D control. * indicates p<0.01...... 66 Figure 17 (A) Axis of symmetry diagram of perfusion bioreactor with a gas permeable wall. This allows oxygen to diffuse radially into the bioreactor. The radius of the PDMS cylinder is denoted by R2, the radius of the scaffold volume of the bioreactor is denoted by R1 and the length is denoted by L. In the equation development the radial coordinate was denoted by r and the longitudinal coordinate by z (measured from the inlet). (B) The experiments were also performed in a batch configuration where the medium was replaced every two days, however oxygen could still diffuse through the wall. (C) Traditional packed/fixed bed bioreactor where all the nutrient and oxygen are provided by medium perfusion alone. (D) Further enhancement of the design by the incorporation of a central capillary with outer wall radius Rw and lumen Rc in the centre of the reactor provides an additional oxygen source...... 75 Figure 18 Model results: heat map of (A) cell density, (B) glucose, (C) lactate and (D) oxygen concentration in the wall and in the vessel on day 7. The white arrow on the oxygen concentration heat map represents the direction of total flux of oxygen while the black arrow represents the direction of the diffusive flux...... 85 Figure 19 (A) The average cell density change as a result of increasing the wall thickness from 0.1 to 5 cm. (B) A bioreactor model average cell density without wall diffusion requires a higher medium flow rate compared to our bioreactor with wall diffusion (Base model). (C) The average cell densities achieved experimentally from T175 flask and bioreactor expanded cells were compared to base model output. Note that only final cell numbers were measured for the bioreactor expanded cells; the curve represents exponential approximations of the previous days based on the overall growth rate from the experiment. (D) Predicted average cell density at different flow rates in a gas permeable wall packed bed bioreactor...... 87

Isolation and Expansion of Placental Derived Mesenchymal Stromal Cells in a Packed Bed Bioreactor ix

Figure 20 Heat maps of (A-C) cell density distribution (D-F) lactate concentration (G-I) and oxygen concentration at 0.5, 1.5, 3 cm3/hr respectively...... 88 Figure 21 (A) The average cell density with different cell growth rate equations defined in Table 6. (B) The average cell density with the initial homogeneous cell distribution adjusted from 500 cell/cm2 to 1000 cell/cm2 (base model)...... 91 Figure 22 (A) Cell density heat map of linear cell distribution in the z direction at day 0 and with an inlet concentration of 2000 cell/cm2. (B) Cell 2 distribution at day 7 with Nin=2000 cell/cm with a linear cell distribution in the z direction. (C) Lactate concentration at day 7 with 2 a Nin=2000 cell/cm with a linear cell distribution in the z direction. (D) The initial cell density of linear cell distribution in the r direction with a centre concentration of 2000 cell/cm2. (E) Cell density of linear 2 cell distribution in the r direction at day 7 with Nin=2000 cell/cm . (F) Lactate concentration for linear cell distribution in the r direction with 2 Nin=2000 cell/cm at day 7...... 94 Figure 23 Heterogeneous cell distribution are modelled as a linear function in the z direction, setting the cell concentration at the inlet (z= 0 cm) between 0 to 2000 cell/cm2with (A) 0 hr, (B) 24 hr and (C) 48 hr cell quiescence. Note that the total initial cell number is the same in all conditions (158,800 cells)...... 96 Figure 24 Heterogeneous cell distribution modelled as a linear function in the r direction by setting the centre cell density (r=0) between 0 to 2000 cell/cm2 with 0 hr (F), 24 hr (G) and 48 hr (H) cell quiescence. Note that the total initial cell number is the same in all conditions (158,800 cells) ...... 97 Figure 25 Experimental and model average cell densities of a batch bioreactor with a full volume exchange every two days, exploring the effect of no medium exchange, cell death with every medium exchange, quiescence due to cell shock and heterogeneous cell distribution ...... 99

Figure 26 (A) Average cell density comparing scaled-up bioreactor R1=2.5 cm, R2= 3 cm and L= 12 cm with bioreactor design that includes wall diffusion, no wall diffusion, wall diffusion with additional central capillary providing oxygen. (B) Shear stress at different flow rates...... 101 Figure 27 Cell density heat map of bioreactor with (A) oxygen provided by diffusion through the wall, (B) no wall diffusion and (C) oxygen provided by both diffusion through the wall and central capillary at day 7 at a flow rate 15 ml/hr. Oxygen concentration with (D) oxygen provided by diffusion through the wall, (E) no wall diffusion and (F) oxygen provided by both diffusion through the wall and central capillary at day 7...... 104

Figure 28 Average cell density in the larger R1=2.5 and L=12 cm bioreactor with different pellet and flow rate for a bioreactor design that includes wall diffusion (A), no wall diffusion (B), wall diffusion with additional central capillary providing oxygen (C)...... 106

x Isolation and Expansion of Placental Derived Mesenchymal Stromal Cells in a Packed Bed Bioreactor

Figure 29 (A) Average cell density of the bioreactor design that provides oxygen by diffusion through the wall and central capillary using 0.5 mm pellets with different bioreactor geometries with the same reactor volume and surface area as a R1=2.5 and L=12 cm bioreactor with a flow rate of 30 ml/hr. (B) Average cell density of the bioreactor providing oxygen through the wall and central capillary at different flow rate rates. (C) Shear stresses of different pellet size and different flow rates. Shear stresses of different bioreactor geometries and flow rate (D)...... 107 Figure 30 KG1a cell growth after four days with different medium types. * indicates P<0.01...... 148 Figure 31 (A) KG1a day 4 cell concentration exposed to ethanol in RPMI 10% FBS or XVIVO 15. (B) KG1a day 4 cell concentration after being diluted with PBS in RPMI 10% FBS or XVIVO 15. (C) bmMSC cell density after 4 days growth in DMEM 10% FBS exposed to ethanol. (D) bmMSC cell density after 4 days growth in DMEM 10% FBS diluted with PBS...... 149 Figure 32 (A) KG1a growth after 4 days on two different TCE surfaces n=4. (B) KG1a growth after 4 days on a TCE surfaces that have been unwashed (Dry), washed and dry (washed), and washed and remained wet (wet), the cells were either in contact or grown in conditioned media n=4. (C) Human bmMSC grown in contact or in conditioned media on surfaces that have been unwashed (dry), rinsed briefly (wet) or washed by rigorous method (washed) n=4. * indicates P<0.01...... 150 Figure 33 KG1a growth after 4 days exposed to NOX surface n=4 (A). Human bmMSC growth after 4 days exposed to NOX surface n=4 (B). Human bmMSC growth after 4 days on different types NOX surfaces that have stronger antibacterial properties n=4 (C). Human bmMSC growing on NOX surface (D & E) and tissue culture plastic (F & G) stained with phalloidin to show actin filaments (red) and DAPI to stain the nucleus (blue). * indicates P<0.01...... 152 Figure 34 HaCaT cell density in response to pMSC and bmMSC for different conditioning times and medium types at (A) day 4 and (B) day 6. * indicates cell density significantly greater (p<0.01) over the DMEM negative control, # indicates significantly greater (p<0.01) over XVIVO 15 control and ** indicates a significate difference (p<0.01) between XVIVO 15 and DMEM media types...... 157 Figure 35 (A & B) HaCaT growth in response to pMSC and bmMSC conditioned medium from different donors and cell aggregates two different experimental replicates after 6 days. * Significant (p<0.01) increase over XVIVO 15 negative control...... 158

Isolation and Expansion of Placental Derived Mesenchymal Stromal Cells in a Packed Bed Bioreactor xi

List of Tables

Table 1: Summary of surface markers for defining and isolating MSC, compiled from [43,45-47] ...... 3 Table 2 Summary of some of the paracrine factors and extra cellular matrix proteins secreted by MSCs. Compiled from [89,101,103,104,110-115] ..... 10 Table 3 Summary of MSC dosages used in a selection of clinical trials ...... 12 Table 4 Advantages and disadvantages of bioreactor systems [31,209,212,218,219,224-232]...... 24 Table 5 Model parameters and values used in model ...... 83 Table 6 Cell growth rate equations adjusted to account for different limitations in nutrient concentrations and cell contact inhibition...... 90

xii Isolation and Expansion of Placental Derived Mesenchymal Stromal Cells in a Packed Bed Bioreactor

List of Abbreviations

2D Two dimensional

3D Three dimensional

ANG-1 Angiopoietin

bFGF Basic fibroblast growth factor

BMP Bone morphogenetic protein

BR Bioreactor

CM Conditioned medium

DMEM Dulbecco's modified Eagle medium

DTI Diffusion tensor imaging

EDTA Ethylenediaminetetraacetic acid

EGF Epidermal growth factor

FBS Foetal bovine serum

G-CSF Granulocyte colony stimulating factor

GFP Green fluorescent protein

GFP-mMSC Mouse green florescence protein labelled bone marrow derived mesenchymal stromal cells

GMP Good manufacturing practise

GVHD Graft versus host disease

HGF Hepatocyte growth factor

HSC Haemopoietic stem cells

IGF Insulin-like growth factor

IL1β interleukin-1β

IL-6 Interleukin 6

INF-γ Interferon-gamma

Isolation and Expansion of Placental Derived Mesenchymal Stromal Cells in a Packed Bed Bioreactor xiii

MCP-1 Monocyte chemotactic protein-1

MHC Major histocompatibility complex

MMP Matrix metalloproteinase

MRI Magnetic resonance imaging

MSC Mesenchymal stromal cell

bmMSC Bone marrow derived MSC

pMSC Placenta derived MSC

NK Natural killer cells

NO Nitric oxide

PBS Phosphate buffer saline

PDGF Platelet-derived growth factor

PDMS Polydimethylsiloxane

PFA Paraformaldehyde

PGE2 Prostaglandin E2

PI Propidium iodide

SDF-1 Stromal cell-derived factor-1

TCP Tissue culture plastic

TGF-β Transforming growth factor-β

TPO Thrombopoietin

TSG-6 Tumour necrosis factor-inducible gene 6 protein

VCAM-1 vascular cell adhesion molecule-1

VEGF-1 Vascular endothelial growth factor -1

VLA-4 very late antigen-4

xiv Isolation and Expansion of Placental Derived Mesenchymal Stromal Cells in a Packed Bed Bioreactor

Statement of Original Authorship

The work contained in this thesis has not been previously submitted to meet requirements for an award at this or any other higher education institution. To the best of my knowledge and belief, the thesis contains no material previously published or written by another person except where due reference is made.

QUT Verified Signature

Signature:

Date: 15/07/2016

Isolation and Expansion of Placental Derived Mesenchymal Stromal Cells in a Packed Bed Bioreactor xv

Acknowledgements

I like to dedicate this thesis to my grandfather Vince Russo who passed away in 2013. With his skills to fix and construct anything thing on his sugar cane farm, I think if he was born later he would have been excellent engineer. I would like to think some of his skills and ingenuity have trickled down.

I would like to thank my supervisor Dr. Bill Lott for supporting me and giving me the opportunity to complete my PhD. It was a difficult year with some very low points early on. Without his guidance and support, to give me confidence to change direction in my thesis for the better, I would not have completed my PhD. I am very lucky to have a true scientist as a supervisor. It was fantastic to be able to talk about other things outside my field of research and be encouraged to keep on asking question.

I would also like to thank Prof. Sean McElwain in for providing guidance in developing the model. Without his time and help I would be still battling away with this PhD and understanding some of the concepts. I would like to thank Thomas Michl and Prof. Hans Griesser from the University of South Australia for expertise in plasma surface modification and Paul Reynolds of Styron for the donation of the polystyrene pellets, without their collaboration I would not have the scaffold for my bioreactor. I would like to thank Josie Tarren from the Mater Hospital Pathology for her assistance in performing the glucose and lactic acid quantification. I also like to thank Dr. Emad Izake, for taking on my associate supervisor role at the last minute to help make the transition less stressful.

I want to thank my wonderful wife Lavania for supporting me on my PhD journey. Always understanding the time and commitment it takes to complete a PhD and always supporting to keep me going. It’s been a crazy time with highs and lows. I am sure witnessing me isolating MSCs from placenta will always haunt your dreams. But I’m sure our trip to America for the ISSCR meeting in Boston would counter act this awful image.

xvi Isolation and Expansion of Placental Derived Mesenchymal Stromal Cells in a Packed Bed Bioreactor

I would like to thank all my friends and colleagues: Dr. Betul Kul-Babur, Myfanway King, Dr. Parisa Ghanavi, Dr. Mahboubeh Kabiri and Eman Othman, that provided some fun into the lab and not just with brainstorming ideas and discussion about our work but debates social, political, religious and pop culture ideas.

Finally, I would like to thank my parents, Henry and Vera, my grandmother and my sister Mariangela. They always instilled in me the importance of scientific research and understanding. Always to be creative, honest, truthful and to be the best you can be there, for then there is nothing to be shy away from and people will come to you for your advice or help you.

Isolation and Expansion of Placental Derived Mesenchymal Stromal Cells in a Packed Bed Bioreactor xvii

Publications

First Author Published

Osiecki, M. J., Michl, T. D., Kul Babur, B., Kabiri, M., Atkinson, K., Lott, W. B., Griesser, H. J., and Doran, M. R. (2015) Packed Bed Bioreactor for the Isolation and Expansion of Placental-Derived Mesenchymal Stromal Cells, PLoS ONE 10, e0144941.

First Author Manuscript

Osiecki M. J., McElwain D.L.S, Lott W. B. (2016) Modelling Mesenchymal Stromal Cell Growth In a Packed Bed Bioreactor with a Gas Permeable Wall, submitted to PLoS Computational Biology

Other Published Works

Cook, M. M., Futrega, K., Osiecki, M., Kabiri, M., Kul, B., Rice, A., Atkinson, K., Brooke, G., and Doran, M. (2012) Micromarrows--three-dimensional coculture of hematopoietic stem cells and mesenchymal stromal cells, Tissue Eng Part C Methods 18, 319-328.

Michl, T. D., Coad, B. R., Doran, M., Osiecki, M., Kafshgari, M. H., Voelcker, N. H., Husler, A., Vasilev, K., and Griesser, H. J. (2015) Nitric oxide releasing plasma polymer coating with bacteriostatic properties and no cytotoxic side effects, Chemical communications (Cambridge, England) 51, 7058-7060.

xviii Isolation and Expansion of Placental Derived Mesenchymal Stromal Cells in a Packed Bed Bioreactor

Previous Work

Osiecki, M., Ghanavi, P., Atkinson, K., Nielsen, L. K., and Doran, M. R. (2010) The ascorbic acid paradox, Biochemical and Biophysical Research Communications 400, 466-470.

Isolation and Expansion of Placental Derived Mesenchymal Stromal Cells in a Packed Bed Bioreactor xix

Chapter 1: Introduction and Literature Review

Mesenchymal stem/stromal/progenitor cells (MSCs) are attractive cell type that can be used for cell based therapies that aim to treat a number of diseases or enhance innate regenerative processes [1-3]. They are one of the most commonly evaluated cell therapy candidates, with 612 clinical trials in tissue regeneration currently registered (clinicaltrials.org as of 9 September 2015). MSCs have the potential to home to sites of injury [4,5] and the capacity to produce a range of trophic factors that promote tissue repair and dampen inflammation [6-8]. Most importantly, MSCs appear to be non- immunogenic, making them a versatile donor tissue allowing allogeneic donors to be used [9-11]. These attributes may enable MSCs to be clinically useful in a diverse range of applications, which include improving cardiac function following myocardial infarction [12-15], treating non-healing and diabetic wounds [15,16], enhancing skeletal tissue repair [17-19], and treating inflammatory diseases such as Crohn’s disease and Graft Versus Host disease [20-22].

Bone marrow is the traditional source of MSCs and cells from this tissue are the most widely studied and best characterised [23]. However, MSCs can be isolated from other tissues, such as placenta, adipose tissue and umbilical cord and they’re becoming increasingly well characterized and may be more accessible than bone marrow-derived cells [24-26]. However, there are some differences in the phenotype and mesodermal differential potential depending on the source [24,26,27]. Regardless of the source, for clinical applications a large number of MSCs are required per dose and this may exceed 200 million cells [3,16,28,29]. MSC isolation and expansion is traditionally achieved using plastic tissue culture flasks and manual culture manipulation, increasing the number of flasks at each passage to achieve the required cell number [30]. However, the large-scale manual manipulation of tissue culture flasks is not ideal for commercial production, as the process is labour intensive, requires a clean room, carries a significant risk of contamination and has poor culture condition control [31]. MSC isolation and expansion will likely require an automated closed-system

Chapter 1: Introduction and Literature Review 1

bioreactor for commercial translation of these therapies to be routine and economically viable.

This thesis focuses on developing a novel packed bed bioreactor that can be used to isolate and expand human placental derived MSCs (pMSCs). The following literature review covers the basic biology of MSCs, clinical trials using MSCs as a cell based therapy highlighting the number of cells used to treat such conditions, the technology used to expand MSCs and what culture conditions are required to be maintained during the expansion.

1.1 MESENCHYMAL STROMAL CELL CHARACTERISATION

MSCs were first identified from bone marrow mononuclear cells that can differentiate down the osteogenic, adipogenic and chondrogenic lineages [32]. However, MSCs have been identified and isolated from most fetal, neonatal and adult tissue types including placenta, adipose tissue, umbilical cord blood, amniotic fluid, dental pulp, skeletal muscle, liver and bone [33-35]. This is attributed to the fact that MSCs are perivascular cells and therefore have similar surface markers and cell adhesion molecules even though they can be found in different tissues in the body [36- 39]. However, it should be stated that not all pericytes are MSCs [39].

In 2006, the International Society for Cellular Therapy delineated the minimum defining criteria as being adherence to tissue culture plastic (TCP), the expression of a cluster of differentiation markers CD105, CD73 and CD90, and the absence of CD45, CD34, CD14 or CD11b, CD79α or CD19 and HLA class II expression. In addition, the cells have to be able to undergo tri-lineage mesodermal differentiation [40]. However, selection using these surface markers and plastic adherent cells can result in isolation of a very heterogeneous population that is a mix of different committed progenitor cells and only a very small fraction of true self renewing cells maybe isolated [41]. An extensive list of additional surface markers that are associated with MSCs is found in Table 1. Through the isolation of MSCs using the additional markers

2 Chapter 1: Introduction and Literature Review

STRO-1, CD146, VCAM1, CD 271 and CD105 it was demonstrated that this more defined population of cells formed colony forming units in vitro and when transplanted into mice could generate a miniature bone organ with bone, cartilage, adipocytes and haematopoiesis-supporting stroma [36,42]. The issue with isolated MSCs using such markers is that these markers change over time and culture conditions, with some markers such as SSEA-4 being induced in in vitro manipulations and the use of fetal bovine serum [43]. The marker profile of MSCs is highly transient, meaning that there is currently no sole marker or group of markers which are truly specific for MSCs. Furthermore, it may not be important to isolate for true stemness markers for MSCs to achieve the desired clinical outcomes, but it may be beneficial to isolate for markers for a subtype of MSCs that have a desired function, such as the CD317 positive subpopulation that has a high immunomodulatory activity, which may allow for more predictable clinical outcomes [44].

Table 1: Summary of surface markers for defining and isolating MSC, compiled from [43,45-47]

Positive Markers Negative Markers CD9, CD10, CD13, CD29, CD44, CD49a, CD45, CD34, CD14, CD49b, CD49c, CD49e, CD51, CD54, CD58, CD11a, CD19, CD86, CD61, CD62L, CD71, CD73, CD90, CD102, CD80/CD40, CD15, CD104, CD105, CD106, CD119, CD120a, CD18, CD25, CD120b, CD121, CD123, CD124, CD126, CD31, CD49d, CD127, CD140a, CD146, CD166, CD271, CD50, CD62E, CD317, CCR1, CCR4, CCR7, CXCR5, CCR10, CD62P, CD117 VCAM-1, CD166, AL-CAM, ICAM-1, STRO- 1 (CD140b), SSEA-4, HER-2/erbB2 (CD340), frizzled-9 (CD349), W8B2, W3D5, W4A5, W5C4, W5C5, W7C6, 9A3, 58B1, F9-3C2F1, HEK-3D6.

Although MSCs can be found is almost all vascularised tissues and share similar marker profiles, there are distinct differences in their function and ability to differentiate, especially in vivo [26,43,48-50]. MSCs isolated from human muscle, which expressed some osteogenic features when exposed to bone morphogenetic proteins (BMPs) in vitro, did not form bone when transplanted in vivo [51]. Likewise, MSCs isolated from human dental pulp that were subsequently transplanted into

Chapter 1: Introduction and Literature Review 3

mouse conditions identical to those used for bone marrow MSCs formed dentin rather than bone [52]. Even the reported in vitro tri-lineage potential (bone, cartilage and fat) of MSCs correlates poorly with the observed in vivo differentiation capacity of MSCs [53,54]. Only 10% of the MSC population is able to form bone, stroma and marrow adipocytes in vivo, while the majority of the population are committed progenitors [53]. This evidence suggests that a varied class of clonogenic progenitors are found in different tissues that are endowed with tissue-specific potency, and that the reported multi-lineage potential in vitro could be in fact an artefact of MSC populations being a heterogeneous mix of lineage-committed progenitor cells [53].

Interestingly, MSCs are used successfully to treat a variety of pathologies in tissues that are not derived from cells belonging to the osteogenic, adipogenic and chondrogenic lineages, including acute myocardial ischemia, stroke, liver cirrhosis, amyotrophic lateral sclerosis, graft versus host disease, skin burns and Crohn’s disease [55]. Thus, the mechanism by which MSC therapies promote regeneration cannot involve direct differentiation into the repair tissue(s) alone, and must facilitate tissue regeneration via alternative mechanisms.

1.2 MESENCHYMAL STROMAL CELL MECHANISMS TO REPAIR DAMAGED TISSUE

When the perivascular location of MSCs was discovered, the hypothesis was proposed that the MSCs would support growth and repair of the surrounding tissue through the secretions of cytokines, paracrine and chemokines such as PDGF, ANG-1 and TGF-β. As such, makes MSCs act as the “injury drug store” of human body [8,36,56]. This makes MSCs an ideal candidate for clinical applications in the treatment of a variety of congenital and acquired diseases. The mechanisms proposed for MSCs as a cell-based therapy in tissue repair is thought to include (I) differentiation and transdifferentiation into multiple cell types to repair target tissue through autologous transplant, secretion of paracrine factors that promote cell (II) an immunomodulatory effect, (III) survival and proliferation, and (IV) through the ability to localise specifically to the injury site (Figure 1) [8,57,58]. However, an additional

4 Chapter 1: Introduction and Literature Review

mechanism has been recently proposed in which mitochondria from MSCs are transferred to cells having non-functional mitochondria, which has been hypothesise to help promote cardiac tissue repair after infarction [59-61]. This literature review here will primarily focus on the first three mechanisms.

III

I II

Figure 1: Observed mechanisms that MSCs repair damaged tissue. Adapted from [33]

1.2.1 Mesenchymal Stromal Cell Differentiation and Transdifferentiation

The most controversial MSC-associated tissue repair mechanism is the MSCs’ is their ability to differentiate and transdifferentiate (Figure 1 I) into the target repair tissue, as this have been observed to occur in vitro and in vivo animals. The cellular process of differentiation occurs when a less specialised immature cell becomes a more specialised cell type. However, cells belonging to a specific cell lineage can occasionally undergo abrupt changes in phenotype that exhibit phenotypical characteristics of a different cell lineage. This process is called transdifferentiation. MSCs are multipotent cells that can be induced to differentiate in vitro and in vivo into

Chapter 1: Introduction and Literature Review 5

mesenchymal tissues (bone, cartilage, tendon muscle, etc.) [40,62-64]. MSCs have been reported to trans-differentiate, both in vitro and in vivo in animals, into different cell lineages [65-67]. However, the process of MSCs’ transdifferentiation is induced in vitro in the presence of growth and transcription factors prior to implantation into animals [65-69]. Human MSCs isolated from Wharton’s jelly of the umbilical cord were induced to transform into islet-like cell clusters in vitro through a stepwise culture in neuron-conditioned media [67]. These islet-like cells secreted insulin and glucagon in response to physiological changes in serum glucose when implanted into a diabetic rat liver. Human amniotic membrane isolated mesenchymal cells can undergo cardiomyogenic transdifferentiation and are able to beat spontaneously through cardiomyogenic induction in vitro [70]. MSCs were transplanted into the infarcted myocardium of Wistar rats. The MSCs were reported to have transdifferentiated into cardiomyocytes in situ and survived for more than 4 weeks after the transplantation without requiring immunosuppressant agents. Xenograft of human MSCs from adipose or bone marrow applied to a skin wound site in mice and rats can contribute to wound repair by transdifferentiating into multiple skin cell types, such as keratinocytes and endothelial cells that form vasculature [5,71,72].

The contribution of differentiation and transdifferentiation to repair tissue remains controversial, as only a small number of donor-derived cells can be detected in vivo in repair sites in human studies, as little as 0.4% of the total cells injected are present after 60 days [56,73-75]. Additionally, there is little available evidence to support differentiation and transdifferentiation in humans in vivo [5,33,62]. This proposed repair mechanism appears to be limited to autologous donor sources of MSCs, as MSCs lose their immune privilege status when they differentiate [76-78]. In addition, due to the heterogeneous nature of isolated MSC populations, the reported transdifferentiation and differentiation could be due to contamination of different lineage progenitor cells [48,53]. Consequently, the beneficial effect of MSCs at the site of injury must rely on a mechanism outside of their differentiation capacity alone.

6 Chapter 1: Introduction and Literature Review

1.2.2 Mesenchymal Stromal Cells are Immunosuppressive, but not Immune Privileged.

The main mechanism that MSCs promote tissue repair is through the modulation of the immune system (Figure 1 II). MSCs secrete a number of immunosuppressive and anti-inflammatory cytokines, summarised in Table 2 and Figure 2 that can modulate the immune system. MSCs secrete transforming growth factor-β (TGF-β) and induce secretions of interleukin-10 (IL-10), both immunosuppressive, that inhibit T cell proliferation [79,80]. Interferon-gamma (INF-γ) secretion by MSCs suppresses the proliferation of CD4+ and CD8+ T lymphocytes, as well as natural killer (NK) cells, but it has no effect on the proliferation of B lymphocytes [81]. Interestingly, the suppressive activity of MSCs are not contact-dependent when MSCs are in the presence of IFN-γ produced by activated T cells and NK cells in inflammatory environments. Even activated B lymphocytes became susceptible to the suppressive activity of MSCs when IFN-γ was added exogenously [81]. In another recently described mechanism, MCP-1 secreted by MSCs recruit T cells for FasL-mediated apoptosis. These apoptotic T cells subsequently triggered macrophages to produce high levels of TGF-2, which in turn led to the up-regulation of CD4+ Treg cells and, ultimately, led to immune tolerance which could allow allogenic MSCs to persist longer [82]. However, this immune damping effect is a double edged sword, in response to inflammatory cytokines MSCs become highly immunosuppressive [83]. This becomes a problem in the cancer niche as MSCs have been shown to promote cancer cell growth and prevent the immune system from attacking the tumour and even enhance drug resistance [84-88]. However, no clinical trial using MSC base therapy has reported any tumour growth [86].

MSCs are reported to be immune privileged as they express low levels of major histocompatibility complex (MHC) class I and do not express MHC class II antigens or costimulatory molecules CD40, CD80, and CD86 [89,90]. MSCs’ lack of MHC antigen expression permits cells from allogeneic donor sources to be transplanted to unmatched recipients with reduced requirements for immune suppressants. This characteristic allows for the collection and banking of allogeneic MSCs, which can then be used for the treatment of tissue damage when needed. However, some studies

Chapter 1: Introduction and Literature Review 7

report rejection of allogeneic MSCs, raising questions regarding their immune privilege status [77,91]. This can be due to the increased expression of MHC antigens when the cells are undergoing differentiation or undergoing cell culture expansion [76- 78,92]. Although allogenic MSCs are not as immunogenic as unmatched fibroblasts and are reported to persist twice as long as fibroblasts [93], MSCs still activate the innate immune system. Through the activation of the complement and coagulation cascade following injection the MSCs are quickly killed, this is more notable when the MSCs have undergone in vitro expansion [75,94]. Confidence in the idea that MSCs are immune privileged has very much diminished since their immune privilege status appears largely dependent on the environment and they may just act like antigen- presenting cells that are able to promote inflammation in vivo [95-98].

Figure 2 Summary of MSCs’ interactions with immune cells. MSCs secrete many soluble factors that alter the function and proliferation of the immune cells and regulatory T cells. Adapted from [89,99,100].

8 Chapter 1: Introduction and Literature Review

1.2.3 Mesenchymal Stromal Cells Secrete Soluble Factors that Promote Cell Survival, Proliferation and Immunomodulatory Effects

Caplan et al. stated that MSCs are a site-regulated injury "drug store", because they migrate to the injury site and secrete soluble cytokines, chemokines and growth factors that promote cell survival, proliferation, angiogenesis and modulate the immune response (Figure 1 III) [8]. The paracrine factor theory is supported by the observation that the addition of a MSC conditioned medium, can in some cases, recapitulate the same tissue repair benefits of MSCs. Additionally, MSC conditioned media act as a chemoattractant for recruiting macrophages and endothelial cells into the wound [101]. Factors released by MSCs include VEGF-1, IGF-1, EGF, nitric oxide, hepatocyte growth factor (HGF), keratinocyte growth factor, angiopoietin-1, stromal derived factor-1, macrophage inflammatory protein-1a and -1b and erythropoietin [101-103]. Tumour necrosis factor-α (TNF-α)-stimulated gene 6 protein (TSG-6) is secreted by MSCs that acts as a negative feedback loop to attenuate the inflammatory cascade initiated through TNF-α secreted by resident macrophages [104-106]. A highly inflammatory environment during skin wound healing causes scar formation, MSCs’ secretions of TSG-6 in response to this environment prevents scar formation [106]. Additionally, hypoxia common in ischemic tissue enhances the production of several factors [102]. Hypoxic conditions also increase MSCs’ paracrine secretions in vitro [102]. Additionally, genetic engineering techniques are being explored to increase the secretions from MSCs and make the response to the injured environment more predictable [107]. Table 2 lists further paracrine factors and extra cellular matrix proteins secreted by MSCs that have been reported. These molecules secreted by MSCs are important for cell survival, proliferation, neovascularisation during tissue repair and wound healing, regulating cell migration, anti-inflammatory and immunomodulation [33,55,66,108]. It should be noted that the tissue repair abilities of MSCs in the absence of the immune system is very much reduced [109]. This suggests that MSCs’ ability to promote tissue repair may be by their immunomodulatory secretions reducing inflammation and promoting the natural repair mechanisms of the immune system.

Chapter 1: Introduction and Literature Review 9

Table 2 Summary of some of the paracrine factors and extra cellular matrix proteins secreted by MSCs. Compiled from [89,101,103,104,110-115]

Trophic Extra cellular Immunomodulatory Chemoattractant matrix VEGF Collagen type 1 PGE2 CCL2 (MCP-1) bFGF Fibronectin HLA-G5 CCL3 (MIP-1a) IL-6 HGF CCL4 (MIP-1b) MCP-1 TGF-b CCL5 (RANTES) HGF IDO (induced by IFN- CCL7 (MCP-3) γ) TGF-β TSG-6 CCL20 (MIP-3a) EGF Nitric Oxide CCL26 (eotaxin- 3) IGF IL-10 CX3CL1 (fractalkine) SDF-1 PGE2 CXCL1 (GROa) Angiopoietin-1 CXCL2 (GROb) Macrophage inflammatory protein CXCL5 (ENA- 78) Keratinocyte growth factor CXCL8 (IL-8) Erythropoietin CXCL10 (IP-10) PDGF CXCL11 (i-TAC) TPO CXCL12 (SDF-1)

1.2.4 Mesenchymal Stromal Cell Homing and Migration

MSCs’ therapeutic applications are acclaimed due to the ability of intravenously injected MSCs to home to sites of inflammation following injury (Figure 1 IV). In a model of multiple organ failure through irradiation of mice, green fluorescent protein (GFP)-labelled MSCs were infused and subsequently homed to numerous tissues based on the severity of the injury [116]. The migration depended on signals from growth factors, and also from chemokines secreted from injured cells and respondent immune cells [8,117]. Interestingly, the inflammatory cytokine TNF-α increased the sensitivity of MSCs to chemokines such as SDF-1 and IGF, which then induced MSCs to migrate to the site of inflammation [118]. In the migration of the MSCs to the injury site, the cells first adhere to the vascular endothelial cells mediated by very late antigen-4/ vascular cell adhesion molecule-1 (VLA-4/VCAM-1) and display coordinated rolling to cross the endothelial barrier to achieve trans-endothelial migration [119-121]. The MSCs’ ability to specifically home to injury sites allows a

10 Chapter 1: Introduction and Literature Review

diverse range of options to administer MSC treatment in a clinical setting. Therefore, MSCs do not require direct administration to the wound site, but can be infused into a patient’s blood stream. However, MSCs lose their homing abilities during in vitro expansion, even at early passage numbers [122]. This could explain why, in some cases, no homing was observed in vivo; typically less than 1% of the infused MSCs reach the target tissue [123,124]. Strategies to engineer improved homing of MSCs after expansion are being devised [125,126].

1.3 CLINICAL APPLICATIONS OF MESENCHYMAL STROMAL CELLS

The reported regenerative and immunomodulatory capacity of MSCs observed in vitro and in vivo in animals, is currently being explored in 612 clinical trials (clinicaltrials.gov). MSCs are being investigated to treat a wide variety of conditions, including acute myocardial ischemia, stroke, liver cirrhosis, graft versus host disease, burns, non-healing diabetic foot ulcers and Crohn's disease [55]. One key aspect of MSC based therapy is the cell dosages used in clinical trials. Table 3 highlights the cell dosages reported in clinical trials to treat different conditions. Another important aspect is the cell source, as many of the clinical trials use allogeneic MSCs to treat these conditions. This becomes important for the future manufacturing of MSCs as an expansion device such as a bioreactor will be necessary to reduce the cost of manufacturing to allow these therapies to become more affordable. This following section will review the clinical trial results of conditions that rely on MSCs ability to promote tissue regeneration through their ability to modulate the immune system and secrete paracrines.

Chapter 1: Introduction and Literature Review 11

Table 3 Summary of MSC dosages used in a selection of clinical trials

Condition Cell Cell source Delivery method Ref. Dosage Articular 1-1.5x106 Autologous bone Injected into the [127] cartilage marrow MSCs defect damage 1.46x107 Autologous Bone Injected into the [128] marrow MSCs defect with hyaluronic acid Severe 21.5 x 106 Allogeneic bone Intra-myocardial [129] ischaemic heart marrow MSCs injection failure Acute 6x1010 Autologous bone Direct injection [130] myocardial marrow MSCs into the ischemia myocardium Ischemic 2x108 Autologous Bone Transendocardial [131] cardiomyopathy marrow MSCs cell injection and mononuclear bone marrow cells Ischemic and 25,75, and Allogeneic bone Transendocardial [132] non-ischemic 150 x106 marrow MSCs cell injection heart failure Myocardial 0.5, 1.6 and Allogeneic MSCs Intravenous [133] infarction 5 x106 infusion cells/kg 1.8 to 2.2 x Allogeneic bone Intravenous [134] 108 cells marrow MSCs infusion Stroke two dosages Autologous bone Intravenous [135] of 5x107 marrow MSCs infusion 50-60 x106 Allogeneic bone Intravenous [136] cells marrow MSCs or infusion mononuclear cells

12 Chapter 1: Introduction and Literature Review

Severe 6x106 Autologous bone Intravenous [137] traumatic brain marrow MSCs infusion injury Hepatic 30-50x108 Autologous bone Intravenous [138] fibrogenesis marrow MSCs infusion Crohn's disease 1x106 Autologous bone Intravenous [139] cells/kg marrow MSCs infusion 2x106 Allogeneic bone Intravenous [140] cells/kg marrow MSCs infusion Graft versus 1.4x106 Autologous and Intravenous [141] host disease cells/kg allogeneic bone infusion marrow MSCs 1,5, or 10 Allogeneic bone Intravenous [142] x106 marrow MSCs infusion cells/kg (MultiStem) single dose or additional dose after two days Multiple 1-2 x106 Autologous bone Intravenous [143] Sclerosis cells/kg marrow MSCs infusion Radiation burn two dosages Autologous bone Injection around [29] of 1.68x108 marrow MSCs the wound site and 2.26x108 Ulcers and acute 1x 106 Autologous bone Sprayed on wound [144] wounds cells/cm2 marrow MSCs with fibrinogen wound area spray Diabetic ulcers 1x106 Autologous bone Injection around [145] and Buerger’s cell/cm2 marrow MSCs the wound site disease wound area

Chapter 1: Introduction and Literature Review 13

1.3.1 Cardiac Failure

MSCs through their paracrine secretions are reported to stimulate growth of new blood vessels and cardiomyocytes [146-149]. It is thought that this functionality of MSCs will help repair the damaged heart tissue after myocardial infarcts. In a randomised clinical trial of MSC therapy for acute myocardial ischemia, patients received either an intracoronary injection of 6x1010 autologous bone marrow MSCs or a control vehicle [130]. MSCs significantly increased the left ventricular ejection fraction at the three month follow-up following MSC transplantation compared to the control group. MSC transplantation was safe, with no deaths or malignant arrhythmias reported. A large Chinese study which injected MSCs derived from Wharton’s jelly showed significant cardiac function increase over the placebo control [150]. Similarly, a more recent phase II trial showed improvement in left ventricular end systolic volume, left ventricular ejection force and volume, and increase in myocardial mass compared to the placebo [129]. However, there was no significant difference in the New York Heart Association functional classification, 6 minute walking test and Kansas City cardiomyopathy questionnaire. This suggests that although there was a small anatomical improvement with MSC treatment overall there was not a significant improvement in the patients’ health. Another recent phase II clinical trial using immune selected allogenic MSCs at three different dosages (25, 75, or 150 x106 cells) reported similar results [132]. The trial found no difference between any cell dosages and placebo for survival probability, MACE-free probability and all-causes of mortality. Although they noted that the hospitalisation rate was reduced for the highest dosages (p=0.025). The above demonstrates that trans-endocardial injection of allogenic MSCs are safe for chronic heart failure patient however no significant or strong evidence has been shown of any clinical improvement.

Several other studies have explored the safety and efficacy of MSC transplants and evaluated different delivery methods, including direct injection into the affected regions of the heart, mobilisation via granulocyte colony stimulating factor (G-CSF), and intravenous infusion [134,151-156]. Cell dosages ranged from 107 to 1011 MSCs. Most studies showed some improvement in cardiac function with MSC transplant compared to control. However, they generally did not significantly improve the

14 Chapter 1: Introduction and Literature Review

patient’s overall health or function. Larger clinical trials are needed. So far the sample sizes have been small, there is a wide range of cell dosages, with differences in delivery without direct comparisons, and there are considerable differences in the timing of cell transplantation after the myocardial infarct. In addition, greater investigation is required to find the right phenotypic markers that characterise functional MSCs that can promote tissue repair. In the TAC-HFT randomised trial both MSCs (characterised by the criteria set out by the ISCT [40]) and bone marrow mononuclear cells (mononuclear cells from a Ficoll density gradient) had no difference in clinical outcomes between the two cell types. However the MSCs did decrease infarct size and increase 6 minute walking distance compared to placebo whereas the bone marrow mononuclear cells did not [131].

1.3.2 Stroke

It is thought through the secretion of VEGF and IGF-1 from MSCs would promote angiogenesis and activate the endogenous neural cells to further express more VEGF, EGF and bFGF to promote neural cell growth [8,157]. A clinical trial was performed that recruited patients for up to 2 years after having a stroke. The subjects were treated with an intravenous infusion of 50-60 million autologous bone marrow MSCs (isolated from plastic adherents) or mononuclear cells (isolated from a Ficoll density gradient). There was no difference in outcomes observed with functional MRI and DTI and clinical scores, only the Barthel Index (ability to perform activities of daily living) showed a statistically significant improvement for those given the mononuclear cells [136]. They observed that the MSCs that underwent expansion performed worse than the mononuclear cells. This is consistent with the observed loss in some phenotypic functions after long term culture of MSCs [158].

A longitudinal open-label, observer blinded clinical trial, studied autologous MSC transplants in stroke patients (two dosages of 5x107 cells two weeks apart transplanted through intravenous infusion) and followed the patients up to 5 years [135]. This study concluded that the procedure is safe with no adverse effects following MSC treatment, and it demonstrated an improvement in functional recovery

Chapter 1: Introduction and Literature Review 15

(p=0.046) and a decrease in mortality. However, the treatment group was small (N=16), this was not a randomised trial and the hazard ratio between the MSC group and the control group was insignificant. This brings into doubt the claim of a decrease in mortality and will require a larger sample size and randomisation of the treatment groups to remove any biases.

1.3.3 Neural and Spinal Cord Repair

Following a spinal cord injury there is an influx of inflammatory immune cells into the injury site, this is a necessary part of the tissue regeneration process. However, this can cause secondary damage to the tissue surrounding the original injury site that can continue for days and weeks following the injury and which can lead to cavitation and cyst formation [159,160]. It is thought that the immunomodulatory function of MSCs will help prevent this from happening and that the paracrine secretion from the MSCs will help the tissue regeneration process [161]. A nonrandomised clinical trial was conducted which injected, via lumbar puncture, autologous bone marrow MSCs into the cerebrospinal fluid, of 11 patients. 20 patients were in a control group who underwent a traditional treatment method [162]. None of the patients experienced adverse reactions, although headaches and tingling were reported by some of the patients who were treated with the MSCs, after 24-48 hours. Although 45% of the patients treated with the MSCs showed a marked recovery compared to only 15% of the control group, this was found to be not statistically significant. Other trials reported similar side effects. In one clinical trial where autologous bone marrow MSCs plus granulocyte-colony stimulating factor were injected into the area surrounding the spinal cord injury, fevers were reported [163]. While others studies reported fevers and myalgia [66]. Although all these trials show an improvement in clinical function, none of these results were statistically significant.

It is also thought that the immunomodulatory function of MSCs would help in acute neuro-inflammation after a traumatic brain injury. A phase I clinical trial was conducted, children were treated with an intravenous infusion of 6x106 cells/kg autologous bone marrow mononuclear cells in addition to the standard treatment

16 Chapter 1: Introduction and Literature Review

protocol for a traumatic brain injury. A control group who did not receive any cell therapy was also followed [137]. This trial showed a statistically significant increase in clinical performance in those children who were given cell therapy and less intensive treatment was required compared to the control group. Most significant was that the requirement for intracranial pressure monitoring was half the time of the control group. A trial using four dosages over intervals of 5-7 days of 1x107 allogeneic umbilical cord MSCs injected into the subarachnoid space via lumbar puncture and followed up 6 months after treatment showed an improvement in the functionality using the Fugl-Meyer Assessments and Functional Independence Measurement over the control group [164]. However, the control group received no medical or surgical treatment whatsoever, hence it is difficult to draw a conclusion. This makes the improved MSC injection result suspect. The control group must undergo a standard treatment to truly demonstrate if MSCs improve clinical outcomes. Longer follow up periods are required to see if these outcomes result in better recovery of the patients. Larger randomised clinical trials are needed to confirm that this is a viable strategy to improve treatment for traumatic brain injuries.

1.3.4 Crohn’s Disease and Multiple Sclerosis

The reported immunomodulatory and anti-inflammatory properties of MSCs make them ideal for treating inflammatory diseases such as Crohn's disease. Crohn’s disease is a chronic inflammatory disorder of the gastrointestinal tract. In a pilot study of eight patients, autologous bone marrow MSCs were infused intravenously at a dose of 1x106 cell/kg body mass. In five of the patients, the Crohn's disease activity index scores improved, with the clinical response observed in three patients at week 6 [139]. In another study, ten patients with fistulising Crohn's disease were treated with autologous bone marrow MSCs [165]. The MSCs were injected into the lumen and the wall of the fistula tracks. There was complete closure of the fistula track in seven of the ten patients after twelve months, with a reduction of Crohn's disease and perianal disease symptoms and healing of the rectal mucosa. In a more recent study, patients received between 8-10x106 cells/kg weekly for four weeks of allogeneic bone marrow MSCs with a primary end point, the Crohn’s Disease Activity Index was below 150 indicating remission [140]. Twelve patients had a clinical response, 50% of all patients

Chapter 1: Introduction and Literature Review 17

went into clinical remission and seven patients had endoscopic improvements. However, no placebo or control group was used in these studies which reduce their clinical significance.

Multiple sclerosis is an inflammatory disease that causes destruction of the myelin sheaths of the neurons. MSCs have the potential to treat multiple sclerosis through their immunomodulatory function by inhibiting pathogenic T and B cell responses and the secretion of neuroprotective and pro-oligodendrogenic molecules to promote tissue repair [166-168]. However, the most robust clinical trial that had a randomised study and had a placebo control group showed no difference between placebo and 1-2x106 cells/kg of bone marrow MSCs intravenously infused in all clinical scores of the number of gadolinium-enhancing lesions with no difference in the frequency of TH1 cells in the blood [143].

1.3.5 Graft Versus Host Disease

Graft versus host disease (GVHD) is a severe life threatening complication with allogeneic haemopoietic stem cell (HSC) transplants. The immunomodulatory aspects of MSCs have been used in the clinical situation to help reduce the incidence and severity of GVHD. In a Phase II clinical study, 1.4x106 MSC/kg body weight were transplanted into patients who had suffered steroid resistant severe and acute GVHD after they underwent a HSC transplant [141]. Both MSCs from autologous and allogeneic sources proved effective therapy for these patients with refractory chronic GVHD. Other small pilot studies have confirmed their benefit [169-171]. It has also been reported that HSC engraft faster when co-transplanted with MSC [141,172]. However, due to the high cost of the treatment, other methods will be used first to treat GVHD when more traditional steroid treatments fail. Even then it appears that MSC therapy is only effective in less severe and less intensively treated cases of GVHD [173]. However, more recent studies suggest that overall survival is not improved by MSC therapy compared to placebo over the long term [22]. It has also been reported that MSCs were ineffective in patients with prolonged GVHD with lower lymphocyte

18 Chapter 1: Introduction and Literature Review

counts [174]. This suggests that a functioning immune system is required for effective control of GVHD by MSCs.

1.3.6 Mesenchymal Stromal Cell role in Enhancing Wound Healing

There are a number of in vitro and animal studies that outline the mechanisms in which MSCs promotes wound healing [71,72,101,110,175,176]. The pro- inflammatory mediators interferon-γ (INF-γ), tumour necrosis factor-α (TNF-α) and interleukin-1β (IL-1β), which are released from the wound during the inflammation phase, firstly promote migration of MSCs to the wound site; secondly, they activate regulatory functions in MSCs to modulate the immune response thereby attenuating the immune response to injury through the secretion of TSG-6[104,115]. In addition, the inflammatory wound environment stimulates COX 2 activity in MSCs, upregulating prostaglandin E2 (PGE2) and transition in wound environment in favour of dermal regeneration [177]. The shift to anti-inflammatory cytokine expression promotes wound fibroblasts to upregulate matrix metallopoteinase (MMPs) expression and down regulates the expression of various collagen types [178,179]. This results in a less dense fibrotic granulation tissue in the wound bed. The secretions of basic fibroblast growth factor (bFGF), vascular endothelial growth factor-A (VEGF-A) and adrenomedullin from MSC’s promote proliferation of microvascular endothelial cells [180,181], vascular stability [182] and the development of a long lasting functional vascular network [183]. In addition, MSCs secrete a number of cytokines and growth factors that have antifibrotic properties, including hepatic growth factor (HGF), IL-10 adrenomedullin and MMP-3 [101,184-186]. These cytokines promote the turnover of the extra cellular matrix, keratinocytes proliferation and inhibition of myofibroblast differentiation [187-190]. Therefore, MSCs promote generation of well vascularised granulation tissue, enhance re-epithelialisation of the wound and attenuate the formation of fibrotic scar tissue.

A number of pilot studies have been conducted using MSCs to treat a variety of non-healing wounds in humans [16,29,144,145]. A Chilean man with a severe 10 cm diameter radiation burn was treated with MSCs after a conventional excision and skin

Chapter 1: Introduction and Literature Review 19

autograft and rotation flap had failed. A combined approach of numerical dosimetry- guided surgery with injection of 168x106 autologous expanded MSCs followed by a second 226x106 MSCs 39 days after the operation, to provide a source of trophic factors to promote tissue regeneration [29] resulted in a favourable clinical outcome. The lesion healed in 6 months, and no recurrence was observed during the 11 month follow-up.

In a second trial, MSCs were used to treat chronic non-healing wounds (diabetic ulcers). An autologous biograft composed of fibroblasts and 3x106 MSCs on a biodegradable collagen membrane was placed directly in the wound site, whilst MSCs were injected into the edges of the wound being covered with the biograft [16]. The wound steadily decreased in size, increased in vascularisation of the dermis, and increased in dermal thickness after 29 days. These in vivo data, coupled with in vitro skin equivalent models, suggest that MSCs could be beneficial in the treatment of chronic wounds [16]. Other trials that treated diabetic foot ulcers and critical limb ischemia with intramuscular injections of peripheral blood or bone marrow MSCs (cell dosage ranging 3x107 to 3x109 cells) reported similar results with improved blood flow, new vessel formation and rescue of the ischaemic limb [191-194].

A fibrinogen and thrombin gel spray was trialled to apply autologous bone marrow MSCs to treat acute surgical wounds after removal of skin cancer (n=5) and on chronic non-healing foot ulcers (n=8) [144]. A strong correlation was observed between the number of cells applied (greater than 1x 106 cells per cm2 wound area) and the subsequent decrease in the chronic wound size.

A study was conducted in India with 24 patients with diabetic foot ulcers or Buerger disease. Autologous bone marrow derived MSCs were injected into the wound site at a concentration of 1x106 cell/cm2 and a control group was treated with a standard wound dressing. The wound area reduced from 5 cm2 to 1.5 cm2 in patients with foot ulcers, and patients with Buerger showed a decrease in the diseased area from 7.3 cm2 to 2 cm2 after 12 weeks. There was no significant size reduction in the control group [145].

20 Chapter 1: Introduction and Literature Review

The clinical trials using MSCs to treat non-healing wounds generally show statistically significant results. This could partly be due to the topical application of the MSCs rather than by intravenous injection which is known to activate the innate immune system [11,75]. As the MSCs remain viable for longer in these wound sites and are able to function as an “injury drug store” widely reported in popular media [8]. However, many of these trials lack control groups or placebo controls. In addition, in all the trials there is great variability in the MSC functionality even when they have been characterised. With the increasing failure of clinical trials to demonstrate that MSC therapy is effective, it is hard to be uncritical of the science involved and many researches in the stem cell field have been starting to question the methods used in the MSC therapy field [56,195]. This criticism is even stronger with the clinical results in trials repairing mesodermal tissues of bone and cartilage when the addition of MSCs to scaffolds made little difference in the tissue repair, even when the scaffold or bone chips were sufficient to promote repair [196]. In addition, many of these clinical trials used intravenous infusions, it has been shown that the majority of the MSCs are embolised in the lungs which can cause endothelial damage [56,105,197]. As the true surface marker and phenotype that makes up a true MSC is yet to be discovered, the populations that are being used are very heterogeneous. In addition, donor tissue type and culture conditions have a significant impact on the function of MSCs [26,158,198- 203]. To add to the complexity of the problem the understanding of how MSCs respond to an inflamed environment is really unknown. It is worrying that there are reports that when MSCs respond to the highly inflamed cancer niche they result in supporting the growth of the cancer by damping the natural immune response [87,204,205]. This indicates that the minimum criteria for MSCs [40] requires updating in order to have better standardisation across all laboratories. In addition, better in vivo animal models are needed as rodent models are far more efficient in undergoing wound repair than humans [206-208].

1.4 EX VIVO EXPANSION OF MESENCHYMAL STROMAL CELLS FOR CLINICAL USE

The clinical results reveal very little evidence of the reported benefits of MSCs. This has created a heated debate regarding moving ahead with clinical trials on these

Chapter 1: Introduction and Literature Review 21

new cell therapies as it is considered that this is too far ahead of the current available science [56,195]. However, one observation is clear that the number of MSCs required to treat wounds and other clinical conditions is in the order of 2x108 cells per dose (see Table 3). Cell expansion is therefore required to enable MSCs to be used as cell-based therapy to treat such conditions, as only small numbers of MSCs can be isolated from bone marrow. Using traditional methods in tissue culture flasks is not viable because of the large surface area required to expand cells just to produce a single dose. This method is time intensive, requires skilled technicians to maintain the cell culture and needs a clean room for the clinical expansion, thus making this a very expensive therapy [89,209].

Assuming that two million MSCs can be manufactured in a T175 flask, a single cell dosage of 2x108 cells will require 100 T175 flasks, which is 1.75 m2 of surface area just for a single dose. However, it would be ideal to expand at least 10-100 cell dosages from a single donor source to develop an allogeneic bank of cells and have significant benefits with respect to product validation and consistency, and reducing production costs [210]. Surface area is the key limiting factor to achieve 100 cell dosages, as this will require 175 m2, which is 10,000 T175. This is not viable without an efficient automated bioreactor system. Even from a non-invasive source such as placenta, which weighs approximately 500 g, only enough MSCs can be isolated to provide cells for 2 dosages after one passage in culture. However, using a bioreactor system to expand placental derived MSCs over four passages is hypothetically enough to make 7000 dosages [210].

22 Chapter 1: Introduction and Literature Review

Figure 3 Bioreactor geometry and mass transfer diagrams indicating mass transfer for various bioreactor geometries: (A) Media is mixed by a magnetic impeller. (B) Media inside a rotating wall bioreactor has the same angular velocity as the wall. (C) Bag cultures on a rocker generate a wave motion which mixes cell suspension and medium. (D) Plug flow is maintained throughout the bed to generate uniform perfusion. (E) Cells are generally grown on the extra capillary space either in suspension or in a gel scaffold; media is perfused in the intra capillary space to provide nutrients and oxygen through the semipermeable wall. Taken from [211].

The expansion and differentiation of stromal cell populations under controlled conditions remain a major technical challenge. This is due to the complex kinetics of the heterogeneous nature of MSC populations, the transient nature of the subpopulations of interest, the lack of surface markers and multiple interactions between culture parameters; such as growth factor concentration, nutrient concentration, dissolved oxygen tension, cell–cell interactions or providing a large surface area for cell attachment and growth [212]. When the nutrient, glucose, is considered, high glucose medium (25 mM) results in a 40% reduction in the confluence compared to low glucose medium (5.5 mM) because high glucose has an effect on MSCs colony formation and differentiation [213-217]. This suggests that a low glucose concentration is required to maintain the stemness of the MSCs during in vitro expansion. However, supplying extra glucose through higher flow rates is limited by the fact that a low shear stress of 0.015 Pa has been reported to up-regulate the osteogenic pathways in human bone marrow MSCs [218-221]. This issue becomes

Chapter 1: Introduction and Literature Review 23

even more complex because of the limited solubility of oxygen (0.2 mM) in the medium. A tight balance is needed between providing adequate nutrients and oxygen to the MSCs while removing waste products such as lactate or lactic acid which affects the culture pH whilst maintaining a low shear stress on the cells [222]. This becomes increasingly difficult when it has been shown that hypoxic conditions are far better in maintaining stemness of MSCs [199,200,202,203,223]. Suppling oxygen at this low concentration by the medium perfusion alone would be very difficult to control [224].

Table 4 Advantages and disadvantages of bioreactor systems [31,209,212,218,219,224-232].

Bioreactor Type Advantages Disadvantages

Wave bag Gentle mixing, low shear stress, Expensive, more suited easy scale up to cell suspensions than microcarrier cultures.

Stirred tank Supports high density cell cultures, High shear, difficult to easy to scale up, can support harvest cells. differentiation and microcarriers, can act as a cell delivery system or cells can be encapsulated to protect from shear

Hollow fiber Low shear stress, easier control on Difficulty in scale up and the cell microenvironment heterogeneous environment due to oxygen and nutrient gradients.

Packed bed Lower shear than stirred tank, Difficult to harvest cells, bioreactor provides 3D microenvironment to shear stress can be an allow special organisation of cells, issue at high flow rates proliferation and differentiation for high density cultures. can be regulated, can be used in tissue formation

24 Chapter 1: Introduction and Literature Review

Bioreactors need to be tailor suited to meet the requirements of MSC culture and need to be optimised for the mass transport of oxygen, nutrients and waste elimination [227]. The large scale expansion of MSCs has the additional criteria of providing a large surface area for cell attachment. This could be done by using roller bottles or multiple plates such as CellFactory (Thermo Scientific Nunc) and CellSTACK (Corning). However, these approaches are only suitable for low dosages and small patient numbers [233]. In addition, roller bottles and multiple plates are very labour intensive and difficult to equip with online monitoring for process control, thus a bioreactor system providing large surface area for cell attachment is required. Figure 3 highlights bioreactor geometry that can provide large surface area for cell attachment and growth by either incorporating micro-carriers or making use of the bioreactor geometry and Table 4 highlights the advantages and disadvantages of these different bioreactor systems.

1.4.1 Wave Bag Bioreactors

The simplest designed bioreactor is a wave bag bioreactor, which simply consists of a gas permeable bag sitting on top of a rocker to gently induce wave and undertow in the semi-closed vessel. This bioreactor layout has some advantages, including purchasing of pre-sterilised parts, good heat transfer, aeration and oxygen monitoring. In addition, the bioreactor can be adapted to a batch, fed-batch, continuous perfusion configuration [234]. They also perform as well as stirred tank bioreactors without the high shear [235]. There are a number of commercially available wave bag bioreactors through GE Healthcare and Sartorius Stedim Biotech that are disposable and come pre-sterilised with the desired medium, making wave bag bioreactors ideal for production of cells under GMP. A wave bag bioreactor was used to isolate placental derived MSCs through an automated process, demonstrating a process to automate MSC isolation and expansion in one device is possible [210]. However, there are a few problems with this system; sampling, monitoring and controlling the system is not as simple as other bioreactor systems, and the cost of a single disposable bag is quite high [212]. Also, when growing adherent cells, it is difficult to get even seeding of cells onto the carriers leading to over confluence on some beads and some beads devoid of any cells.

Chapter 1: Introduction and Literature Review 25

1.4.2 Stirred Tank Bioreactors

The most widely used configuration is the stirred suspension bioreactor. This bioreactor can be made to operate in a batch, fed-batch or continuous configuration. Stirred suspension bioreactors are appealing for the large scale production of stem cells as concentrations of 106-107 cells/mL can be attained [236]. Cells can be cultured in a single cell suspension, as aggregates, or on microcarriers in order to attain high cell densities. The cell density can be increased significantly with the use of porous microcarriers which also protect the cells from the high shear stress as they are able to grow inside the pellets [237]. However, it is very difficult to harvest the cells from the pores making porous microcarriers only suitable for when the cells are not the product. Stirred suspension bioreactors have been used for a number of years with a wide range of models and components commercially available, facilitating this bioreactor use in a wide variety of applications [212]. MSCs have been cultured in aggregates as well as on microcarriers in stirred suspension bioreactors [238-240]. MSCs grown in 3D aggregates do not grow rapidly, however they do maintain stemness and differentiation capacity better than 2D cultures [238]. When growing MSCs, seeding is far more difficult with microcarriers compared to other cell types [241]. In addition, growth rate of MSCs on the microcarriers was generally less than MSCs grown on 2D tissue culture plastic and expansion of microcarriers in xeno-free culture media has been demonstrated [228,239-242]. There are many commercial bench top scale bioreactors available including Mobius® Cell Ready (Merck Millipore) and BIOSTAT® UniVessel (Sartorius Stedim Biotech) that have been used to expand MSCs [31,243- 246]

A similar concept to the stirred suspension bioreactor is the rotating wall bioreactor. Instead of an impeller agitating the medium, the wall of the bioreactor rotates to mix the media leaving the particles suspended in "free fall" [247]. Rotating wall bioreactors are reported to have a positive effect on osteogenic differentiation of MSCs [248,249]. However, due to the nature of the mechanics, the system is complex and not easily scalable.

26 Chapter 1: Introduction and Literature Review

One of the major problems with stirred suspension bioreactors is that, although homogenous conditions are achieved by stirring the media, it becomes increasingly difficult to maintain homogenous conditions as the bioreactor is scaled up, requiring an increase in the speed of the rotor. This can cause hydrodynamic shear stress due to mechanical agitation, which is harmful to the cells [250]. Due to hydrodynamic shear to keep the microcarriers in suspension, stirred suspension bioreactors are not ideal for expanding large number of cells due to heterogeneous culture conditions and micro- eddies induced by the rough external surface of macroporous carriers damaging the cells [224].

1.4.3 Hollow Fibre Bioreactors

Hollow fibre bioreactors offer an increase in surface area to volume ratio compared to other bioreactors. Hollow fibre bioreactors consist of two compartments: a vessel that normally contains cells (extracapillary space) and a semipermeable membrane capillary through which media is perfused (intracapillary space), see Figure 3F. These membranes generally have a low molecular weight cut off <10 kDa; this allows micronutrients such as glucose and oxygen to diffuse freely through the membrane, but larger molecules such as proteins cannot [251]. The hollow fibres provide nutrients and eliminate waste products from the cells in a manner which is analogous to blood vessels in vivo [252]. However, in tissue engineering applications the cells are maintained in a gel scaffold in the extra capillary space, which hinders perfusion leading to concentration gradients [253,254]. One of the major problems with growing cells in the extra-capillary space is recovery of the cell product after expansion if they have been grown in a scaffold. It has been qualitatively shown that MSCs can effectively grow in the extra-capillary space suspended in a gel. However the researcher could not remove the cells to get a quantitative measure of the expansion [68]. This result suggests that growing MSCs in the extra-capillary space is not an ideal way of using a hollow fibre bioreactor to expand cells for a clinical cell based therapy application.

Chapter 1: Introduction and Literature Review 27

1.4.4 Packed Bed Bioreactors

Packed bed bioreactors provide a very large surface area to volume ratio. This bioreactor is simply a vessel with immobilized scaffold arranged in a column where the cells are seeded and adhere to the packed bed. Media is perfused axially to provide nutrients, including oxygen. A packed bed bioreactor containing 500 μm diameter beads in a bioreactor volume of approximately 400 mL can provide approximately 3 m2 of surface area. By simply decreasing the bead size down to 100 μm (a common commercial microcarrier size) the surface area that this volume of bioreactor can provide is 14 m2, which equates to 8 cell dosages of 2x108 cells [236].

Packed bed bioreactors are widely used in the fermentation industry as they provide a low foaming and low shear environment, however they are not yet widely used in cell therapy [227]. However, these sorts of perfusion devices are more commonly used for tissue engineering purposes especially for bone [255]. Uniform flow (plug flow) must be maintained through the column to prevent unequal distribution of flow that will result in nutrient and oxygen gradients within the packed bed [211]. In addition, as flow is applied axially through the pack bed, there is a concentration gradient of nutrient and oxygen along the length of the reactor. No design has developed a solution of spatial concentration gradient within a packed bed bioreactor. Simply increasing the flow rate is not ideal, as it will induce a high hydrodynamic shear on the cells adhering to the packed bed. Haematopoietic progenitor cells have been expanded in packed bed bioreactors using a stromal cell support layer with limited success [256,257]. MSCs have been expanded on glass spheres in a packed bed, and all the necessary information to scale up such a packed bed bioreactor has been provided [218]. Another packed bed bioreactor using Fibra- Cell disk (Eppendorf) as the packing material was able to expand MSCs and produce 4.15 x 108 cells with no detriment on to the MSC’s characteristic [258]. However, an increased flow rate (>3x10-4 m/s) resulted in a decrease in the growth rate [218,226]. This greatly limits the size of the bioreactor due to oxygen concentration restraints. In addition, similar to microcarrier stirred tank bioreactors, cell harvest is difficult. However, the surface of these microcarrier or scaffolds plays a large role in modulating the phenotype of cells growing on these surfaces [238,259-263].

28 Chapter 1: Introduction and Literature Review

1.5 SUMMARY AND THESIS STRUCTURE

For MSCs based therapy to become affordable and routine, an automatable bioreactor system is required to produce the large number of cells used in a single dose for clinical applications. A packed bed bioreactor is required to achieve the large surface area to volume ratio to grow adherent cell type such as MSCs on. However, to provide adequate amounts of the limiting metabolite oxygen to the MSCs, a high flow rate of the medium is required. This higher velocity can cause shear stress on the cells, resulting in delamination and cell death [218]. This limits scale-up potential of traditional packed bed bioreactors as the flow rate is determined by oxygen concentration which is in limited supply. To counteract this, a packed bed bioreactor that incorporates a gas permeable wall made from PDMS that allows gas diffusion radially into the bioreactor was developed. This allows for a decoupling of oxygen from the bulk nutrient flow through the radial diffusion of oxygen. We hypothesize that this novel design allows for a lower median flow rate, as oxygen is no longer the limiting metabolite in the perfused medium. This allows for greater scale up potential and better control over the microenvironment as the flow rate can be optimised for other soluble factors and nutrients that are at a higher concentration than oxygen.

To explore and develop this bioreactor design, murine derived bone marrow MSCs were initially used. In Chapter 2, the design process is demonstrated by exploring the use of different scaffold materials of glass and plasma treated polystyrene for the bioreactor in a 1.5 cm diameter by 7.5 cm bioreactor providing 160 cm2 of surface area. Furthermore, to demonstrate the scale-up potential of this bioreactor design, a 5 cm diameter by 12 cm length scaffold providing 2830 cm2 of surface area was used to expand the murine bone marrow MSCs.

In Chapter 3, to demonstrate the application of the packed bed bioreactor design and to confirm that MSCs can maintain their phenotype during expansion in the bioreactor compared to tissue culture flasks, the packed bed bioreactor was used to isolate and expand MSCs from digested human placenta.

Chapter 1: Introduction and Literature Review 29

To further develop the bioreactor design and more importantly help explain the results observed in Chapters 2 & 3, in Chapter 4, a 2D axisymmetric finite element model was developed that included the effects of oxygen, glucose and lactate concentration on the cell growth rate. This model was solved using the finite element software COMSOL. For ease of the reader, this Chapter follows a different format from the other experimental Chapters with a combined results and discussion section to give better understanding of the thought processes involved for the model development.

In Chapter 5, the work of this thesis is summarised and the future directions of this work is outlined.

30 Chapter 1: Introduction and Literature Review

Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall

MSC populations will require in vitro expansion in order to generate clinically relevant cell numbers to be used as a cell based therapy. Many promising therapies require single or multiple doses of approximately 2 x 106 cells/kg [3]. For MSC-based therapies to become a routine and economically viable treatment approach, the most efficient and cost effective method for their large-scale manufacture will require an automated closed-system bioreactor.

Bioreactor designs used for MSC expansion include micro-carrier suspensions in spinner flasks, stirred tank reactors, and perfusion reactors, such as fixed/packed beds or hollow fibre bioreactors [68,226,231,258]. Simple microcarrier suspension cultures achieve a large surface area for adherent cell culture. However, there is no connectivity between individual microcarriers, and empty micro-carriers do not contribute to the total surface area available to the culture. As a result, some microcarriers rapidly reach confluence, whilst others remain empty; this requires frequent passaging to overcome localized space limitations [231,264]. Packed bed bioreactors potentially solve both problems by providing a continuous and connected surface to allow the cells to migrate between pellets.

The packed bed bioreactor described here contains scaffold that was constructed from air plasma modified polystyrene pellets to which promote cell attachment similar to commercial tissue culture polystyrene (TCP) [265]. This surface modification is required as many polymers used as a scaffold for cell growth are hydrophobic and exhibit low surface energy, which leads to inefficient cell attachment, spread and proliferation [263,266-268]. However, surface modification to improve the surface cell affinity can be achieved in a number of ways [269]. Wet chemistry methods are commonly used, such as sodium hydroxide treatment to hydrolyse the polymer surface [259,270,271] or alternatively, acids such as sulfuric, or chromic acids, to oxidise the

Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall 31

surface to increase the hydrophilicity [272,273]. However, these reactions are nonspecific, and these surface modifications are inconsistent and heterogeneous with respect to polymers with different molecular weight, crystallinity or tacticity [274,275]. The modifications condition are also harsh and can degrade the polymer, resulting in a loss of mechanical properties and the chemical reagents generate hazardous chemical waste [275,276]. Alternatively, peroxide oxidation, ozone oxidation and γ- and UV-radiation can be applied to add functional groups. However, this also leads to degradation of the polymer, and the desirable modifications are non- permanent [277-280].

H H H H H H H Hydrophobic C C C C C C C C C ~ ~ Polystyrene H H H H H H H

Oxygen plasma surface modification

OH H O O H HO O Hydrophilic Plasma C C C C C H C C C C ~ Modified Polystyrene H H H H H H OH

Figure 4 The reaction between polystyrene and oxygen plasma on surfaces results in a variety of chemical groups and breaking of the chain back bone or opening of benzene rings (not shown) that results in a more hydrophilic surface. Adapted from [281].

Plasma surface modification is a superior method to add functional groups to the surface to improve the hydrophilicity of a surface without changing the bulk properties of the material [269,282,283]. This method uses non-thermal plasma sustained in oxygen, air, nitrogen or ammonia without the need for damaging solvents [265,284]. Surface modification is regulated by the pressure, power, time, gas flow rate and the gas used [274,285]. Oxygen, nitrogen and ammonia plasma add oxygen and amine

32 Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall

functional groups, respectively, while carbon dioxide gas can introduce carboxyl groups [286]. Polymers such as polystyrene exposed to the corona discharge react to the highly energetic oxygen ions, which oxidise and graft to the surface polystyrene chains (Figure 4) that result in the surface becoming hydrophilic and negatively charged. These characteristics allow vitronectin and fibronectin (cell attachment proteins found in serum) to coat the surface and provide better surface properties for cell attachment [273,281]. In contrast, gases, like helium and argon discharges generate free radicals, which can later be cross-linked to oxygen-containing groups when exposing the surface to oxygen or air [269,283]. Plasma treatment provides a far more homogeneous surface modification without degradation of the polymer compared to wet chemistry methods. It is also able to modify scaffolds with small pore sizes and particles of 500 µm in size [269,274,287,288], which is important for the research described here. However, plasma surface modification is sensitive to contaminates that cause undesirable reactions, especially with the more exotic gases [274]. In addition, the multiple parameters that must be optimised for the plasma surface modification creates difficulty in reproducing the results at a laboratory scale [289]. In Appendix 1the methodology for testing plasma modified surfaces is demonstrated.

An additional issue in expansion devices is the mixing or perfusion of the medium to enable nutrient exchange and prevent concentration gradients. The shear stress arising from mixing or medium perfusion must be carefully modulated, as this can compromise MSC stemness characteristics during expansion [219,220]. As the maximum perfusion flow velocity cannot exceed 3 x 10-4 m/s without compromising the growth rate [218]. This greatly limits the scalability, as both soluble nutrients and oxygen must be supplied by medium perfusion alone.

The bioreactor design described here overcome these problem by incorporating a gas permeable polydimethylsiloxane (PDMS) shell, which decouples the bulk medium perfusion from the supply of oxygen. This allows a reduced perfusion flow rate or even a single pass medium supply circuit to be used. Bubble formation within the bioreactor caused by pressure drops and temperature changes across the bioreactor was prevented by pressurizing the waste reservoir to 2 PSI. The system was optimised

Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall 33

using an immortalized murine MSC population that expresses green fluorescent protein (GFP-mMSC). This allowed for easier imaging of the cells on the bioreactor using fluorescent microscopy and also reduced complications due to primary human cells such as donor specific changes. Data from earlier prototypes was also included to demonstrate the design process, moving from a glass bead packing material to the plasma treated polystyrene scaffold.

2.1 METHODS

2.1.1 Single Pass Small-Scale Bioreactor Design

This system contained a 1.5 cm diameter by 7.5 cm long scaffold providing a total surface area of 160 cm2, connected to a single pass circuit (Figure 5 A). This was made from either plasma treated 2.5 mm polystyrene pellets (Figure 5 C) or serum coated glass 3 mm beads. A 5 mm thick polydimethylsiloxane (PDMS, Dow Corning, MI, USA) tube was moulded to just fit the polystyrene scaffold with an additional 1 cm head space to function as a bubble trap. Perfused medium was driven by a syringe pump (New Era Pump Systems Inc., NE-1800, Farmingdale, NY, USA) that was maintained outside of the incubator. Medium was perfused first through a 30 cm length of 16 mm diameter silastic tubing (Cole-Parmer, IL, USA) to enable conditioning of the medium before it entered the bioreactor flow cell. The rest of the tubing set was constructed of 1.6 mm Masterflex PharMed BPT tubing (Cole-Parmer, IL, USA).

34 Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall

A: Small-scale single pass bioreactor Vent

Relief Valve Gas Regulator

0.4µm filter Air @ 2 PSI

160 cm2 Packed Bed Bioreactor Syringe Pump 5 mL/day

Media Reservoir Air Cylinder Cell Suspension Syringe Port

B: Large-scale recirculating bioreactor Vent

Relief Valve

Gas Regulator

Air @ 2 PSI 2800 cm2 Packed Bed 0.4µm filter Bioreactor

Air Cylinder Media Reservoir Cell Suspension Syringe Port

Peristaltic Pump 0.5 mL/min C D 2 PSI Air through 4µm Filter 2800 cm2 Bioreactor

Injection Port

Reservoir Split inlets

Figure 5 (A) Process flow diagrams of the single pass small-scale 160cm2 bioreactor and (B) recirculating perfusion for the large-scale 2800 cm2 bioreactor. (C) Picture of 160 cm2 left and 2800 cm2 right bioreactor scaffolds made from 2.5 mm polystyrene pellets (scale bar is 10 mm). (D) 2800 cm2 bioreactor inside the incubator.

Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall 35

2.1.2 Recirculating Large-Scale Bioreactor Design

A recirculating system was implemented as shown in the process flow diagram (Figure 5 B). The large-scale bioreactor was a 5 cm diameter by 12 cm scaffold, providing a total surface area of 2800 cm2, encased in a 5 mm thick PDMS tube. The peristaltic pump (Watson Marlow Alitea, 403V/VM 4, Stockholm, Sweden) was located outside the incubator. A single 3 mm diameter Masterflex PharMed BPT tubing passed through the peristaltic pump and was expanded into two 30 cm lengths of 1.6 mm diameter silastic tubing to allow conditioning of the medium for the inlet. This was then further expanded into the four 1.6 mm inlet holes into the bioreactor. Medium then left the bioreactor via the four 1.6 mm outlet holes, reducing firstly into two 30 cm lengths of 1.6 mm diameter silastic tubing and finally one 3 mm diameter Masterflex PharMed BPT tubing into the reservoir (Figure 5 D).

2.1.3 Large Surface Area Scaffold

Scaffolds were constructed from 2.5 mm in diameter by 3 mm length fused cylindrical polystyrene pellets (a generous gift from Paul Reynolds of Styron, PA, USA) or 3 mm serum coated glass beads. The polystyrene pellets were fused together by passing Acetone through a mould with a 1.5 cm diameter and 7.5 cm in length for the small-scale and a 5 cm diameter and 12 cm in length for the large-scale. The surface area of the scaffold was calculated by using the mass of the column and the volume estimate was confirmed by measuring the void volume (0.47).

Details of the plasma reactor used to treat the polystyrene scaffolds have been reported previously [290]. In brief, the plasma reactor was loaded with the untreated scaffold columns, sealed and pumped under vacuum to a base pressure of 7x10-3 mbar. Air was then introduced into the reactor at a flow rate of 4 cm3/min, while the chamber was rotating at approximately 10 revolutions per minute. Plasma was ignited at 50 W for a total period of 10 minutes.

36 Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall

2.1.4 Prevention of Bubble Formation in Bioreactor System

To prevent bubble formation due to pressure drop and temperature changes in the system, the system was pressurised with air at 2 PSI (13.79 kPa) using a low pressure two stage gas regulator (Gascon Systems, Thornbury, Vic, Australia) and a pressure safety valve set at 2 PSI (Generant, Butler, NJ, USA).

2.1.5 Bioreactor Sterilisation.

The bioreactors were sterilised overnight with a reactor volume (~6.2 ml for small scale and ~110 ml for large scale) of 70% ethanol. The ethanol was drained and the bioreactor was washed 2 reactor volumes (~ 12.4 ml for small scale and ~220 ml for large scale) of PBS (Gibco). The bioreactor was left overnight to allow the ethanol to leach out of the PDMS wall. This is because PDMS is known to absorbed small molecules [291]. The bioreactor was washed with one reactor volume of PBS and drained ready to be connected to the tubing circuit and injection of cells. The reservoir and tubing circuit was sterilised by autoclave. For the glass scaffold the bioreactor, reservoir and tubing circuit was autoclaved together.

2.1.6 Air Plasma Treated Polystyrene Scaffold Characterisation

Surface analysis was carried out using Axis Ultra DLD spectrometer (Kratos, Manchester UK.) using a monochromatic Al Kα X-ray source operating at 225 W, which corresponds to an energy of 1486.6 eV. The area of analysis was 0.3 x 0.7 mm and an internal flood gun was employed to minimize the charging of the samples. The survey spectra were collected at a dwell time of 55 ms with 160 eV pass energy at steps of 0.5 eV with three sweeps. The collected data were then analysed and processed using CasaXPS (ver.2.3.16 Casa Software Ltd. ®) utilizing Shirley baseline correction. To compensate for the charging effects, the C-C peak was offset to 284.8 eV in all spectra. Finally, all atomic percentages were mathematically rounded to one digit after the comma.

Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall 37

2.1.7 Cell Culture

Green fluorescence protein labelled mouse mesenchymal stromal cell line (GFP- mMSCs, generous gift from Mater Medical Research Institute (MMRI)) were originally isolated from the bone marrow of UBI-GFP/BL6 mice by MMRI in accordance with The University of Queensland Animal Ethics Committee, previously published and described in [292,293]. The GFP-mMSCs were maintained in DMEM

10% FBS and incubated in standard cell culture conditions (37°C and 5% CO2).

2.1.8 Comparison Cell Growth Rate of Glass Packing Material and Tissue Culture Plastic

Round glass cover slips with the surface area of 1.9 cm2 (the same surface area as a 24 well plate) were used. The cover slips were sterilised overnight in 70% ethanol before being rinsed with PBS (Gibco) and placed into 4 wells of a 24 well plate. The cover slips were coated in 200 µl of FBS overnight. A total of 8 plates were prepared, one for each time point (days 1 through 7) and one plate for attachment. Each plate utilised 4 wells containing the serum coated glass and 4 wells that were tissue culture plastic. The wells were seeded with 1000 cell/cm2 of green florescence protein labelled mouse mesenchymal stromal cells (GFP-mMSC) with 0.5 mL of DMEM 10 % FBS and incubated in standard cell culture conditions (37 ⁰C, 5 % CO2). At each time point, a plate was removed from the incubator, the medium was removed and the well was washed with PBS. The cells were then fixed with 100 µL of 4 % paraformaldehyde (PFA, Sigma) for 20 minutes, and the PFA was removed from the wells. The wells were then washed twice with 200 µL PBS before being stained with 100 µL of 300 nM DAPI hydrochloride (Invitrogen) for 15 minutes. The wells were washed twice again with 200 µL PBS before being observed under a florescent microscope (Nikon Eclipse, 4x objective lens). The numbers of cells per well were obtained by photographing the bottom of each well, and scanning the entire well surface. The nuclei were counted in each image by ImageJ (version 1.45), first by highlighting the nuclei by adjusting the threshold (adjust -> threshold), then by making a binary of the highlighted nuclei (process -> binary -> make binary). Any particles that were touching were automatically separated using the watershed algorithm (process ->

38 Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall

binary -> watershed). The nuclei were counted automatically using the “analyse particle” function, excluding particles smaller than 100 pixels2 (analyse -> analyse particles).

2.1.9 Comparison of Air Plasma Treated Polystyrene and Commercial Tissue Culture Plastic on Cell Growth Rate and Attachment.

Non-tissue culture treated polystyrene 60 mm Petri dishes (Sigma-Aldrich) were plasma treated in the same manner as the polystyrene scaffolds above. To assess cell attachment, the air plasma treated Petri dishes and T75 flask (Nunc) were seeded at 2000 cells/cm2 with GFP-mMSCs in DMEM 10% FBS. This cell density was used instead of 1000 cells/cm2, so that enough cells to provide an accurate count using flow cytometry method. The cells were permitted to attach for one and half hours in standard cell culture conditions. The adhered cells were detached and then counted on an FC 500 flow cytometer (Beckman and Coulter, USA) using flow cytometry counting beads (Beckman and Coulter, USA). To compare the growth rate, plasma treated Petri dishes and T75 flask were seeded with GFP-mMSCs at a density of 1000 cells/cm2 and incubated for 3 days in DMEM 10% FBS. The cells were then harvested and counted via flow cytometry as described above.

2.1.10 Packed Bed Bioreactor Seeding Method.

1000 cells/cm2 was seeded into the bioreactor suspended in DMEM 10% FBS. To distribute the cells evenly, the bioreactor was placed on a tube rocker roller set at 5

RPM for 10 minutes, followed by 5 minutes of rest in an incubator (37°C and 5% CO2) this process was repeated for 3 hr. After this process was completed the pump was switched on for perfusion experiments.

Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall 39

2.1.11 Expansion of GFP-mMSCs in the Packed Bed Bioreactor

The 160 cm2 bioreactor was seeded at 1000 cells/cm2 (25,685 cells/ml) of GFP- mMSCs suspended in DMEM 10% FBS as described in 2.1.10. As a 2D control, GFP- mMSCs were seeded at a density of 1000 cells/cm2 (5,833 cells/ml in 30 ml of medium) in T175 flasks (Nunc). The GFP-mMSCs were grown either under static conditions (no continuous perfusion with the media exchanged every two days) or perfusion (5 mL/day) in the bioreactor. The cell numbers were quantified every two days by replacing the medium with a 1:50 dilution of AlamarBlue in DMEM 10% FBS medium and incubated for three hours. The reacted AlamarBlue medium was replaced with fresh DMEM 10% FBS and the culture was continued. The fluorescence signal of the AlamarBlue in medium was measured in a plate reader (FLUOstar Omega, BMG Labtech, Germany) using excitation and emission filters of 544 and 590. To infer the cell numbers a cell titration of 0 to 16,000 cell/ml in a 96 well plate (n=8 wells) was analysed.

2.1.12 Packed Bed Bioreactor Harvest and Harvest Efficiency

The bioreactor was washed with one reactor volume (~6.2 ml for small scale, ~110 ml for large scale) of PBS and replaced with one reactor volume of 0.05%

Trypsin EDTA (Gibco). The Bioreactor was then incubated (5% CO2 and 37⁰C) for 15 minutes. The detached cells were removed by first draining the Trypsin solution and then pumping through three reactor volumes (~19 ml for small scale, ~330 ml for large scale) of DMEM 10% FBS and removing the remaining medium in the reactor. Live cell numbers were determined by an automated cell counter (Bio-Rad TC20, CA, USA) using Trypan Blue. This number was compared to the AlamarBlue result to estimate the percentage of live cells removed from the bioreactor.

2.1.13 Packed Bed Bioreactor Imaging

For direct imaging of the GFP-mMSCs attached to the 160 cm2 column at the end of culture, the column was removed and fluorescence microscopy (Nikon Eclipse

40 Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall

Ti-u with a Nikon Digital Sight Ds-Qimc camera, Japan) was used to image GFP expressing cells.

To image the whole bioreactor using the IVIS Imaging System 200 series (Caliper, PerkinElmer, MA, USA), the cells were fixed with 4% paraformaldehyde for 1 hour. The PFA was drained from the bioreactor and washed with two reactor volumes PBS. The scaffolds were stained with a reactor volume 10 µg/mL propidium iodide in PBS (PI, Invitrogen) for 10 minutes. As the IVIS detection filters are in the red spectrum so the GFP signal could not be detected. In addition, this process was needed for the detection using IVIS primary human MSCs. The bioreactor was then washed with a reactor volume PBS and stored on ice prior to imaging. As negative controls, bioreactors containing no cells underwent the same fixing and staining process. The bioreactors were imaged with an excitation filter of 520 nm and an emission filter of 620 nm. The thresholds were adjusted to remove background auto florescence.

2.1.14 Scale-up of GFP-mMSC in Bioreactor

The 2800 cm2 bioreactor was seeded at 1000 cells/cm2 (~25,455 cells/ml) with GFP-mMSCs suspended in DMEM 10% FBS using the injection port. The T175 flasks functioned as a 2D control. The same seeding procedure as per section 2.1.10 was followed. The reservoir was filled with 250 mL of DMEM 10% FBS to give a total medium volume of 360 mL in the circuit (bioreactor volume 110 mL). The medium was pumped in the recirculating circuit at 0.5 mL/min. A 1 mL medium sample was taken from the injection port each day. In addition, at the end of the culture period an additional 1mL sample was taken from the bulk medium in the reservoir. Glucose and lactate concentrations were quantified by the Mater Hospital Pathology laboratory. At the end of expansion, the cell numbers were quantified by the AlamarBlue method.

Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall 41

2.1.15 Statistics

Results are expressed as means and standard deviations of four biological replicates unless stated otherwise. Differences were determined by T-test using SPSS statistics (ver 17, SPSS Inc, Chicago) and values of p ≤ 0.01 were considered significant.

42 Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall

2.2 RESULTS

2.2.1 Glass Bioreactor Development

Glass was first chosen as a packing substrate, as it allowed for easier sterilisation. As PDMS, silicon tubing and glass are all autoclavable materials, the entire bioreactor and tubing circuit could be autoclaved before each bioreactor run, simplifying the sterilisation process. Historically, glass was commonly used for tissue culture before the advent of tissue culture plastic, so FBS coated glass was believed to be a suitable matrix for bioreactor design. To be certain, the growth rates of GFP-mMSCs on both glass and TCP were compared. There was no significant difference in cell number after a week’s expansion between the glass cover slip and traditional TCP (Figure 6 A). However, in a 3D scaffold configuration in the bioreactor there was a problem with achieving cell expansion. This was due to the cells delaminating from the surface so that the surface coverage of the serum coated glass beads became very patchy (Figure 6 D, E, F & G). This was so dramatic as to cause areas in the bioreactor to have no cells present and there was a very heterogeneous cell distribution overall (Figure 6 B & C). This demonstrated that unmodified serum coated glass was unsuitable as a scaffold material and the alternative of plasma modified polystyrene was explored.

Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall 43

A B C

30000 ) 2 Inlet 25000 Inlet 20000 15000 Glass 10000 TCP

5000 Cells Cells density(cell/cm 0 Outlet 0 2 4 6 Outlet Day

D E

F G

Figure 6 (A) Comparison of cell density of mouse GFP-mMSC between glass and tissue culture plastic (TCP), mean + SD of n=4. (B & C) IVIS images of two separate glass packed bed bioreactors containing mouse GFP-mMSC stained with PI. Fluorescence intensity ranges from low (dark red) to high (yellow). The bright spot in C is due to autofluorescence of a large deposit of silicon glue. (D, E, F & G) Fluorescence microscopy images of GFP-mMSC growing on glass beads.

2.2.2 Plasma Treated Polystyrene Scaffold Characterisation

The rotary air plasma treatment greatly increased the charged oxygen groups on the surface on the polystyrene pellets (Figure 7 A), providing a similar growth surface chemical composition to commercial TCP [265]. However, high amounts of silicon and oxygen contamination in the untreated polystyrene were detected. This was likely

44 Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall

because the PDMS mould used to make the scaffold is known to leach unreacted polymer which contains silicon and oxygen [294].

A B C )

) 25,000 2 Sample %C %N %O %Si 2,000 2 1,750 20,000 Untreated 87.46 - 9.29 3.25 1,500 cells/cm 1,250 * 15,000 Plasma 67.37 0.55 27.6 4.34 1,000 Treated 750 10,000

density( 500 Commercial 75.20 2.1 21.7 - 5,000

Cell Cell 250 TCP Cell density(cells/cm 0 0 Plasma Commercial Plasma Commercial Treated TCP Treated TCP Polystyrene Polystyrene

Figure 7 (A) Surface composition of the air plasma treated polystyrene scaffold determined by XPS, commercial TCP based on literature value [265]. (B) Cell density of GFP-mMSC attachment after one and half hours, initially seeded at 2000 cells/cm2 in a 60 mm plasma treated petri dish or a T75 flask control (n=4, mean + SD). (C) The growth after 3 days of culture of GFP-mMSC seeded at 1000 cells/cm2 in a 60 mm plasma treated petri dish or a T75 flask control (n=6, mean + SD).

The attachment and growth of GFP-mMSCs were compared, on our plasma treated surface in 2D and commercial TCP. Less than half of the 2000 cells/cm2 seeded attached to the plasma treated surface after one and half hours (Figure 7 B). However, after 3 days the cell number on the plasma treated surface was equivalent to commercial TCP (Figure 7 C). Following this observation, the seeding period for the bioreactor experiments was extended to 3 hours to allow more robust cell attachment.

2.2.3 GFP-mMSCs Expansion in Small-scale Packed Bed Bioreactor

The GFP-mMSCs growth rate in the bioreactor was significantly (p<0.01) less than the 2D controls under both static and perfusion conditions (Figure 8 A & B) with the doubling time of 30.19±0.6 hr and 26.48±0.49 hr respectively for static conditions and 27.18±0.8 hr and 24.59±0.53 hr respectively under perfusion conditions (5 mL/day). The doubling time was calculated by ln(2)/max growth rate assuming logarithmic cell growth. The cells were harvested from the bioreactor with an efficiency of 84%±11% with a cell viability of 71%±15% (Figure 8 C).

Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall 45

50 BR 1 40 BR 2 BR 3 30 BR 4 20 2D 1 Expansion 2D 2 10 2D 3 0 2D 4 0 2 4 6 Day B

70 BR 1 60 BR 2 50 BR 3 40 BR 4 30 2D 1 Expansion 20 2D 2 10 2D 3 0 2D 4 0 2 4 6 Day C % Recovery % Viable 100% 80% 60% 40% 20% 0% BR 1 BR 2 BR 3 BR 4

Figure 8 (A) The fold expansion of GFP-mMSCs seeded at 1000 cell/cm2 in four bioreactor (BR) under static conditions and, (B) under 5 mL/day perfusion, compared to four T175 flask controls (2D). (C) Cell harvest recovery and viability from the bioreactor undergoing 5mL/day perfusion.

46 Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall

The roller method of seeding greatly increased attachment efficiency with only 1.89% of the total injected cells not attaching to the scaffold after 24 hours (Figure 9 A). This is approximately 13% more cells attached to the scaffold using the roller method as compared with the static method of injecting the cells and allowing them to attach with no other manipulations. The cells formed a monolayer on the fused pellets and maintained a spindle-like morphology. In addition, the cells were observed to display normal MSC morphology and where grow across the connection between the beads (Figure 9 G, H, I, J). IVIS imaging used the fluorescent intensity of the PI stained cells to estimate the global distribution of cells within the bioreactor. Under both static and perfusion conditions, fluorescence intensity was homogeneous within the scaffold (Figure 9 B & C), suggesting that the cells could be homogeneously distributed in the scaffold using the roller seeding method when compared to the static method where there was a higher fluorescent intensity at the inlet (Figure 9 D, E & F).

Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall 47

A 20% B C 15% Outlet Inlet

10%

5% *

0% Roller Static Inlet Outlet

% % attachedNon cellsseedingafter method method D E F Outlet Outlet Outlet

Inlet Inlet Inlet

G H

I J

Figure 9 (A) Seeding efficiency of using the roller method compared to just injecting the cells and allowing them to attach under static conditions, (n=4, mean + SD, * p<0.01). (B) IVIS imaging of the fluorescent intensity of PI stained GFP-mMSCs in the bioreactor under static and (C) 5 mL/day perfusion conditions after roller seeding method. (D, E & F) IVIS images of static expanded GFP- mMSCs using the static seeding method. IVIS images are a red (low) / yellow (high) heat map of fluorescent intensity. (G, H, I & J) Fluorescent microscopy showed that the GFP-mMSCs attached to the scaffold after cell expansion using the roller seeding method (scale bar is 500 µm,).

48 Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall

2.2.4 Scale-up of the Packed Bed Bioreactor

GFP-mMSCs were used to demonstrate the scale up potential of a larger 250 mL vessel that provided 2800 cm2 of growth surface. The bioreactor was able to expand up to 1.12x108 GFP-mMSCs. However, cell growth was significantly slower in the bioreactor than in the 2D controls, with doubling times of 21.5 ± 0.1 hr and 20.8 ± 0.4 hr respectively (Figure 10 A). The medium perfusion rate was increased to 0.5 mL/min from the small-scale experiments to provide adequate glucose to the cells. The glucose levels were nearly depleted in the bioreactor and the reservoir by day 5 (Figure 10 B), suggesting either that the medium must be changed every three to four days or that a larger medium reservoir is required to prevent glucose concentration dropping by half. The IVIS imaging showed heterogeneous fluorescence intensity, implying that the seeding method is not as robust when scaled-up to the large-scale bioreactor compared to the small-scale bioreactor (Figure 10 C, D, E & F).

A B

70 12

60 10 8 50 * Glucose Reservoir 6 40 Lactate Reservoir 4

(mM) Glucose 30

Expansion 2 Lactate 20 0 0 2 4 6 10

Day Glucose and Lactate Concentration Concentration Lactate and Glucose 0 2D control Bioreactor C D E F

Figure 10 (A) The fold expansion of GFP-mMSCs in a scaled-up bioreactor under 0.5 ml/min perfusion with T175 flask control (n=4, mean + SD, * p<0.01). (B) Glucose and lactate levels in the bioreactor (n=3, mean + SD). (C, D, E & F) IVIS imaging of the fluorescent intensity of PI stained GFP-mMSCs in the bioreactor.

Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall 49

2.3 DISCUSSION

Packed or fixed bed bioreactors like many other bioreactor designs, rely either on the perfusion or mixing of the medium to provide the two limiting metabolites, glucose and oxygen. Due to the low oxygen solubility, the perfusion or mixing rates must be high enough to meet the oxygen demand of the expanding cells and to prevent depletion and significant concentration gradients [31,218]. Such high perfusion rates can potentially cause shear stress on the cells, which has been shown to reduce cell growth rate and induce differentiation. A shear stress of 0.015 Pa has been reported to up-regulate osteogenic pathways in human bmMSCs [218-221]. Our bioreactor employed a gas permeable wall to decouple the oxygen supply from the bulk medium perfusion, resulting in lower flow rates and putatively decreasing the shear stress to which the cells were exposed. In addition, the lower medium volume required for tissue culture would the reduce flow rate, representing a significant cost saving. Please refer to Chapter 4 and Appendix 11 for a thorough analysis of momentum transfer in the bioreactor that shows that the flow rate used in the experiments is below the 0.015 Pa threshold that induces differentiation.

The Glass packing material as was commonly used for tissue culture before the advent of tissue culture plastic (TCP), so glass was believed to be a suitable scaffold for bioreactor design. In 2D expansion experiments there is no difference in a glass surface and TCP GFP-mMSC growth rate (Figure 6A). However, the cells were easily delaminated from the surface which made glass very unsuitable, in addition, movement of the beads during the seeding process could damage and kill the cells (Figure 6 G, H, I & J). It was decided to move to a scaffold of fused polystyrene pellets that was plasma treated. An air plasma treated polystyrene surface provided a comparable surface to commercial grade tissue culture plastic (Figure 7 A), and the growth rate was comparable to commercial TCP (Figure 7 C). However, the overall growth rate of the GFP-mMSCs growing in the bioreactor in static and perfusion conditions was less than the traditional flask expanded cells (Figure 8 A & B). Although, medium perfusion enhanced the growth rate, which is consistent with literature observations [226,228,231,295]. Similarly, when the bioreactor was scaled up to a larger 5 cm diameter by 12 cm in height, there was a decrease in the growth

50 Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall

rate of the cells growing in the bioreactor compared to the tissue culture flask. However, difference between the two growth rates was less with a higher flow rate providing more nutrients in the bioreactor compared to the lower flow rate used in the small 160 cm2 bioreactor (Figure 10 A). Nevertheless, this result supports the hypothesis that suppling oxygen diffusion through the wall is scalable and can use a flow rate that is lower than what has been previously been published [226,228,231,295].

This impaired bioreactor growth rate relative to traditional tissue culture flasks could be attributed to surface chemistry and geometry of the scaffold or to the method used to seed the bioreactor. As the growth on the plasma modified petri dishes was the same as commercial TCP and the GFP-mMSC maintain their spindle like morphology on the bioreactor scaffold (Figure 9 G, H, I & J). This suggests that the impaired growth rate observed in the bioreactor could not be entirely attributed to chemical differences in surface composition. The inefficiencies in the cell seeding process and geometry can also strongly affect the growth rate. A noticeable uneven cell distribution was observed in the scale up bioreactor (Figure 10 C, D, E & F). The geometric features of the bioreactor scaffold surface could alter the contact inhibition characteristics of cell colonies, which would be likely to reduce the observed growth rate. However, due to the size of the pellets used in the scaffold being large, the surface would appear 2D at a cellular level. Which greatly reduces the contact inhibition effect due to curved surfaces and the 3D orientation of the scaffold during cell expansion. Despite this, the geometric features of the scaffold can affect the collision rate during seeding, defined as the rate at which cells contact the scaffold, would be greatly reduced on these 3D structures in dilute cell suspension [264]. Of the many bioreactor seeding methods reported in the literature, no method was completely efficient [218,296-298]. The best method achieved ~85% cell attachment with 85% uniformity [299,300]. The roller seeding method was fare more efficient with 98% of the cells attaching and the static method achieved an efficiency of 87%.

Chapter 4 explores this reduced growth in the packed bed bioreactor by modelling the bioreactor in a 2D axisymmetric model, which was developed to explore the effect of nutrient concentrations and cell seeding distribution on the overall cell

Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall 51

growth rate. Furthermore, the model was used to aid future scale up designs by modifying shape and demonstrated that by decreasing particle size this bioreactor design was able to produce greater than 2x108 cells at a lower medium flow rate than has been previously been reported [226,228,231,295].

The next Chapter discusses the use of this bioreactor to isolate and expand MSCs from human placenta to further demonstrate the flexibility of this bioreactor and its use as a potential single closed system isolation and expansion device for MSCs.

52 Chapter 2: Design and Development of Packed Bed Bioreactor with a Gas Permeable Wall

Chapter 3: Bioreactor Application: Isolation and Expansion of Placental Derived Cells on a Packed Bed Bioreactor

The most widely studied and best characterized MSCs are derived from bone marrow (bmMSCs) [23]. However, MSCs can be isolated from other tissues that may be more accessible than bone marrow-derived cells, these include placenta, adipose tissue and umbilical cord [24-26]. The placenta is a particularly attractive source of MSCs, as a single placenta (500-700 g tissue) is sufficient for manufacturing several thousand units of allogeneic MSCs once expanded [231]. Furthermore, placental tissue can be harvested aseptically during a Caesarean section with no added risk to the mother or child. Placental-derived MSCs (pMSCs) are reported to behave similarly to bone marrow derived MSCs (bmMSCs) in many respects [26,301]. However, pMSCs may be only partially immunogenic and less immunomodulatory relative to bmMSC and do appear to have a reduced differential potential down the oestrogenic linage [26,49].

MSCs have been reported to promote repair of non-healing wounds and ulcers [302]. MSCs secrete a number of soluble factors that promote cell growth, survival and have an immunomodulatory effect [8,101,303]. Nevertheless, tissue repair is not only enhanced by injecting the MSCs, but also supported by injecting the MSC conditioned media alone [101,110,175,238,303], despite the observation that the overall half-life of secreted proteins in the conditioned media is less than by injecting MSCs directly [304]. However, intravenously injected MSCs are mainly captured in the lungs and are also attacked by the innate immune system through the complement and coagulation cascade, preventing the MSCs from reaching their desired location [75,94]. This suggests that the topical application of MSCs conditioned media would be ideal method for wound repair.

It was demonstrated in Chapter 2 that decoupling the oxygen supply from the bulk medium flow by allowing oxygen diffusion through the wall could support GFP-

Chapter 3: Bioreactor Application: Isolation and Expansion of Placental Derived Cells on a Packed Bed Bioreactor 53

mMSC growth. However, as the growth rate was reduced in the bioreactor compared to traditional tissue culture flask, it was necessary to confirm if this had any effect on the MSCs’ ability to differentiate. Primary human placental derived MSCs were used to investigate if there were any detrimental effects in growing human MSCs in our bioreactor. In order to demonstrate the flexibility of the bioreactor, pMSCs from digested placenta were isolated directly onto the bioreactor. In addition to, exploring the effects isolating and expanding pMSCs in the packed bed bioreactor on the cell stemness, a comparison of the functional difference between placental and bone marrow derived MSCs in their ability to differentiate and to promote cell growth was investigated. By using the human keratinocyte cell line, HaCaT, growth response to conditioned medium of the MSCs from the two tissue sources at different passages, to examine the effect of long term culture on MSCs phenotype.

3.1 METHODS

3.1.1 Bioreactor set-up

The bioreactor was set up in a single pass circuit system described in 2.1.1 & 2.1.4, using the same plasma treated polystyrene scaffold described in 2.1.3.

3.1.2 Cell Culture

Placenta was obtained from full term elective Caesarean sections with written consent from the patients and ethics approval from the Queensland University of Technology’s Human Ethics Committee (1000000938). The method used to isolate pMSCs from the placenta has previously been reported [231,305], the resultant product was <1% CD45+ and >90% CD73+/CD105+. Briefly, 10 g of placental tissue cut into 5 mm cubes (pieces) was digested in a 50 mL falcon tube with DMEM low glucose (Invitrogen), 100 Units/mL Collagenase I (Worthington) and 2.5 µg/mL DNase I (Roche) that was incubated at 37 °C for 2 hours on a shaker (250 RPM). The cell supernatant was passed through a 70 µm cell strainer (BD Falcon) and was centrifuged at 400 g for 5 minutes. The supernatant was removed and the cell pellet was

54 Chapter 3: Bioreactor Application: Isolation and Expansion of Placental Derived Cells on a Packed Bed Bioreactor

resuspended in PBS, and mononuclear cells were then isolated using a Ficoll-Paque (GE Healthcare) density gradient at 400 g for 30 minutes. The mononuclear layer was collected and washed with PBS and finally the cells were resuspended in low glucose DMEM supplemented with 10% fetal bovine serum (Thermofisher), 100 units/mL penicillin, and 100 µg/mL streptomycin (Life Technologies) (referred to as DMEM 10% FBS from here on) and seeded into a single T175 flask. The cells were then incubated in 20% oxygen, 5% CO2 at 37 °C for 24 hours to allow adherent cells to attach. The flask was washed and fresh medium was added. The flask was placed into a hypoxic incubator (2% oxygen and 5% CO2). This process was standardised in our group for MSC isolation procedures [202]. The medium was replaced twice a week until the monolayer was 80-90% confluent before passaging. The pMSCs were harvested and cryogenically stored after passage one.

To isolate bone marrow MSCs, human bone marrow aspirates were obtained from volunteer donors. Written consent and ethical approval for these studies was granted by the Mater Health Services Human Research Ethics Committee (Ethics number: 1541A) and the Queensland University of Technology Human Research Ethics Committee in accordance with the Australian National Statement on Ethical Conduct in Human Research. Every 5 ml bone marrow aspirate was diluted to 35 ml with Phosphate buffer saline (PBS, Invitrogen). The mononuclear cells were isolated using Ficoll-Paque (GE Healthcare) density gradient at 400g for 20 minutes. The buffy layer was removed and washed in PBS and re-suspended in low glucose DMEM (Invitrogen) supplemented with 10% fetal bovine serum (Thermofisher), 100 units/mL penicillin, and 100 µg/mL streptomycin (Life Technologies), referred to as DMEM 10% FBS from here on, and seeded into a single T175 flask (Nunc). The cells were then incubated in 20% oxygen, 5% CO2 at 37 ⁰C for 24 hours to allow adherent cells to attach. The medium was then replaced with fresh medium and the flask was placed into a hypoxic incubator (2% oxygen and 5% CO2 at 37 ⁰C). The medium was replaced twice a week until the monolayer was 80-90% confluent before passaging.

HaCaT keratinocyte cell line were maintained in low glucose DMEM (Invitrogen) supplemented with 10 % FBS and 100 units/mL penicillin 100 µg/mL

Chapter 3: Bioreactor Application: Isolation and Expansion of Placental Derived Cells on a Packed Bed Bioreactor 55

streptomycin and incubated in standard cell culture conditions (37 ⁰C, 20 %O2 and 5 %CO2).

3.1.3 Placental and Bone Marrow Conditioned Medium

To make conditioned media, 10,000 cells/cm2 of pMSCs or bmMSCs were seeded onto a 6-well plate in 5 mL of DMEM 10 % FBS. The cells were left to attach for four hours, after which the media was removed and the wells were washed with 5 mL of PBS to remove any residual serum proteins. 5 mL of XVIVO 15 or DMEM (low glucose, Invitrogen) cell culture media supplemented with 100 units/mL penicillin 100µg/mL streptomycin (is referred to as DMEM) was added to the well. The cells were then incubated up to 48 hours in hypoxic conditions (2 % oxygen, 5 %

CO2) in a humid atmosphere at 37 ⁰C. The medium was removed and centrifuged at 500 G for 5 minutes to remove the cell debris and stored at -20⁰C.

For the cell aggregate condition medium, 60,000 cells where added to a 48 well plate with microwell technology developed in the Doran Lab, that contains 600 microwells/cm2 suspended in 0.5 ml XVIVO 15 [306,307]. After 48 hours incubation in hypoxic conditions, the media was removed, centrifuged at 500 G for 5 minutes to remove cell debris and stored at -20⁰C. For use in the growth experiments the conditioned media was diluted to 5 ml so that the cell concentration would be the same as the monolayer conditioned media.

The concentrated condition media followed the same procedure as the monolayer conditioned media in the 6 well plates above. After 48 hr of conditioning, the media was removed and centrifuged at 500g for 5 minutes. The media was then dialysed with a 12 kDa (Sigma) cut off for 48 hours changing the deionised water every 6 hours for the first 24 hours and every 12 hours for the last 48 hours. 5ml aliquots of the dialysed media were lyophilised (Christ Alpha 1-2 LDplus). The lyophilised media was stored at -20⁰C and resuspended in 2.5 ml XVIVO 15 for experiments to give a 2x concentration conditioned media.

56 Chapter 3: Bioreactor Application: Isolation and Expansion of Placental Derived Cells on a Packed Bed Bioreactor

3.1.4 HaCaT growth in response from MSC conditioned media

To gauge MSC conditioned medium capacity to enhance cell growth, HaCaTs were seeded at 1000 cells/cm2 in a 96 well plate, and the cells were left to attach for 4 hours in DMEM 10 % FBS. The medium was then removed, and the wells were washed three times with PBS to remove any residual serum proteins. 100 µL MSC conditioned medium, positive control of the 10% FBS supplement DMEM or XVIVO and a negative control of serum free XVIVO 15 or DMEM. The cell number was quantified on day 4 and 6 using AlamarBlue comparing the relative florescence to cell number standard on a plate reader (BMG labtech). The media was changed every 2 days. The HaCaTs were grown in normoxic conditions.

3.1.5 Pre-Isolated pMSCs Expansion in the Packed Bed Bioreactor

The 160 cm2 scaffold bioreactor was seeded through the injection port with 1000 cells/cm2 of passage four pMSCs suspended in DMEM 10% FBS, using the same seeding process, culture conditions and 2D controls as the expansion of the GFP- mMSCs in section 2.1.10 & 2.1.11. However, in the perfusion experiment the cell numbers were quantified by AlamarBlue at the end of the culture, in the method outlined in 2.1.11. The cells were then harvested (2.1.12) and characterised by tri- lineage mesodermal differentiation capacity (3.1.7). The cell distribution was observed by IVIS imaging described in 2.1.13.

3.1.6 pMSC Isolation and Expansion in the Packed Bed Bioreactor

The placental tissue was digested by the same method outlined in the pMSC isolation and cell culture section above. Instead of a density gradient, the filtered cell suspension from the 10 g of starting tissue was seeded into the bioreactor or into a T175 flask using the same seeding method outlined in the GFP-mMSC expansion (2.1.11). After 24 hours, the cell suspension was removed, the medium was replaced and the flow rate was adjusted in the bioreactor to 5 mL/day. Once the pMSC colonies

Chapter 3: Bioreactor Application: Isolation and Expansion of Placental Derived Cells on a Packed Bed Bioreactor 57

were observed (between day 10 and 14) in the T175 flasks, the cells were harvested in both the bioreactors and flasks and reseeded into the same bioreactor or flask to continue the expansion process. The cell number was monitored using the AlamarBlue method. The cells were harvested once confluence was achieved and characterised by tri-lineage differentiation capacity.

3.1.7 Tri-Linage Mesodermal Differentiation

Osteogenic differentiation was achieved by culturing 25000 cells/cm2 pMSCs in 24 well plate with 2 ml of induction medium containing high glucose DMEM (HGDMEM, Invitrogen) supplemented with 10% FBS, 10 mM β-glycerol phosphate (Sigma-Aldrich), 100 nM Dexamethasone (Sigma-Aldrich) and 50 µM L-ascorbic acid 2-phosphate (Sigma-Aldrich). The calcium deposits were visualised by first fixing the cells with 4% PFA and washing twice with 2ml of PBS before staining with 1ml 2% AlizarinRed S (Sigma-Aldrich) incubating for 5 minutes before washing 2ml of de-ionised water until it runs clear and the calcium deposits where observed under a microscope. Alkaline phosphatase (ALP) activity was quantified with p-nitrophenyl Phosphate (Sigma-Aldrich) following manufactures instructions, measuring the absorbance at 405 nm after a 30 minute incubation at room temperature protected from light.

Chondrogenesis was induced using cultures of 200,000 cell pellets grown in 1 ml HG DMEM supplemented with 110 μg/mL sodium pyruvate (Invitrogen), 10 ng/mL recombinant human Transforming Growth Factor β1 (TGF-β1, Peprotech), 100 nM dexamethasone, 200 μM ascorbic acid 2-phosphate, 40 μg/mL L-proline (Sigma- Aldrich) and 1% ITS-X (Invitrogen). Every two days 75% of the chondrogenic medium was changed, the collected medium was stored at -20°C for subsequent glycosaminoglycan (GAG) analysis. GAG was visualized by staining frozen sections of the cell pellets with 0.5% Alcian Blue (Sigma-Aldrich). GAG was then quantified by first digesting the pellets in 2 mg/ml papain (Sigma) solution and then staining with 1,9-Dimethyl-methylene blue zinc chloride double salt (Sigma-Alrich) and the

58 Chapter 3: Bioreactor Application: Isolation and Expansion of Placental Derived Cells on a Packed Bed Bioreactor

absorbance was measured at 590 nm and compared to a standard made from shark chondroitin sulphate (Sigma) to quantify the GAG.

Adipogenic differentiation was induced over 2 weeks with HGDMEM supplemented with 10% FBS, 10 µM insulin (Invitrogen), 1 µM dexamethasone, 200 µM indomethacin (Sigma-Aldrich) and 500 µM 3-isobutyl-1-methyl-xanthine (Sigma- Aldrich) in a 24 well plate seeded at 25000 cells/cm2. Lipid droplets were visualised by staining with 0.2% Oil Red O (Sigma-Alrich) of 4% PFA fixed cells. The lipids were quantified with an abiogenesis detection kit (Abcam) following the manufacturer’s protocol.

The DNA was quantified by PicoGreen dsDNA Reagent Kit (Invitrogen) following the manufacturer’s protocol.

3.1.8 Statistics

Refer to section 2.1.15.

Chapter 3: Bioreactor Application: Isolation and Expansion of Placental Derived Cells on a Packed Bed Bioreactor 59

3.2 RESULTS

3.2.1 Difference in tri-lineage Differentiation Potential of bmMSCs and pMSCs

A Adipogenic Osteogenic Chondrogenic

bmMSC

100 µm 500 µm 500 µm

pMSC 100 µm 500 µm 500 µm

B C D

18 12 6 16 ) 10 5

nM 14 12 8 4 10 * 6 3 8 * 6 4 2

4 Absorbace/DNA(ug)

2 (µg)/DNA GAG (µg) 1 Triglycerides Triglycerides per Well ( 2 * 0 0 0 pMSC D1 bmMSC D-11 pMSC D1 bmMSC D-11 pMSC D1 bmMSC D-11 Figure 11 (A) Tri-lineage potential of pMSC and bmMSC was assessed by staining and visualising the lipid droplets (Oil Red O, adipogenic), bone nodules (Alizarin Red, osteogenic) and GAG in the pellet (Alcian blue, chondrogenic). (B) The levels of triglycerides were measured to indicate the degree of adipogenic differentiation. (C) The activity of ALP was measured to indicate the degree of osteogenic differentiation and (D) the chondrogenic differentiation was assessed by the level of GAG present in the pellet. *indicated p<0.01 with n=4 in all conditions.

To understand the functional difference between bmMSCs and pMSCs, their capacity to undergo tri-lineage differentiation must be assessed. The two types of MSCs underwent 2 weeks of induction under adipogenic, osteogenic and chondrogenic conditions. By visualising the bone nodules and GAG expression by staining the lipid droplets a significant decrease in the differentiation capacity of pMSC compared to bmMSCs was easily observed (Figure 11 A). Under adipogenic induction conditions, bmMSCs produced significantly more lipid droplets, confirmed

60 Chapter 3: Bioreactor Application: Isolation and Expansion of Placental Derived Cells on a Packed Bed Bioreactor

by the near doubling of the triglycerides present in the well (Figure 11 B). More and larger bone nodules were observed in bmMSCs under osteogenic induction than the pMSCs, which was confirmed by a 10-fold increase in ALP expression levels in bmMSCs (Figure 11 C). The bmMSCs also produced a larger pellet with twice as much GAG produced (Figure 11 D), strongly suggesting that there is a difference between the biological functions of MSCs from these two tissue types. For the purposes of potential wound healing applications, it is necessary to determine if this biological difference also affects the ability of these two types of MSCs to induce HaCaT growth using the conditioned media.

3.2.2 Difference in pMSC and bmMSC conditioned media on HaCaT growth

To explore the effect of long term culture on pMSC and bmMSC and the difference in the two cell types’ ability to promote HaCaT growth conditioned medium made from increasing passage number was made. The conditioned medium was produced in hypoxic conditions as maintains the MSCs and improves the secretion of trophic factors to promote cell growth [199,201,203,238]. In Appendix 2 demonstrate the rational for the conditioning time for the medium and the use of the medium type. It was found that there was no link in passage number and MSCs’ ability to promote HaCaT growth (Figure 12A). In the case of pMSCs, all except passage 10 promoted HaCaT growth above XVIVO 15 medium. On the other hand, for bmMSC only passage 3 and 6 was effective in increasing HaCaT growth significantly over the negative control (XVIVO 15). When taking the average of the cell density achieved by the pMSCs and bmMSC there was no significance difference between the two tissue types in the ability to promote HaCaT growth. However, no consistency in the result was achievable on top of donor variations, which made it impossible to determine if there was a true difference between the MSC tissue source on the length of culture (Appendix 3).

Chapter 3: Bioreactor Application: Isolation and Expansion of Placental Derived Cells on a Packed Bed Bioreactor 61

A B

50,000 50,000

)

) 2 2 * * * 40,000 * 40,000 * 30,000 * * * 30,000 20,000 * 20,000

density(cells/cm 10,000 10,000

0 0

HaCaT density(cells/cm HaCaT

Figure 12 (A) HaCaT growth in response to MSC conditioned medium of different passage number after 4 days (n=4). (B) Average HaCaT density achieved for pMSCs and bmMSCs from Figure 12A removing the outliers of bmMSC p8 & 10 and pMSC p10 (n=4). * Significant (p<0.01) increase over XVIVO 15 negative control.

To see if any nutrient depletion was playing a role in the inconsistent results or improve the growth response of HaCaTs, the conditioned medium was concentrated by dialysis to remove the nutrients but still keep the proteins. The dialysate was then lyophilised and resuspended to a higher concentration. It was found that this process actually reduced the ability of the conditioned medium to promote growth when concentrated three times (Figure 13 A). The one times concentration conditioned medium did provide some benefit, when it was made for the bmMSC conditioned medium and only resuspended 1x, however this did not have a significant improvement on the unmodified conditioned medium. As aggregates have been shown to better maintain the stemness of MSCs [238,306-308], Conditioned medium from cell aggregates was explored. The MSC cell aggregates did improve HaCaT growth compared to negative control, however there due to inconsistency it was impossible to determine a significant improvement in aggregate conditioned medium to traditional conditioned medium ability to promote cell growth (Figure 13 B & Appendix 3).

62 Chapter 3: Bioreactor Application: Isolation and Expansion of Placental Derived Cells on a Packed Bed Bioreactor

A )

2 100,000

80,000 *

60,000 * 40,000 * 20,000

HaCaT (cells/cm HaCaT density * 0 *

160,000 ) 2 140,000 120,000 100,000 80,000 * * 60,000 # * 40,000 20,000 *

HaCaT density(cells/cm 0

Figure 13 (A) HaCaT growth in response to concentrated conditioned medium after 6 days. (B) HaCaT growth in response to cell aggregates after 6 days. * Significant (p<0.01) increase over XVIVO 15 negative control, # significant (p<0.01) increase in HaCaT growth over traditional conditioned medium.

3.2.3 Human Placental MSCs Expansion in Small-Scale Packed Bed Bioreactor

To demonstrate that the bioreactor had no effect on pMSCs ability to undergo mesodermal differentiation, pre-isolated passage 4 human pMSCs were expanded under static and perfusion conditions in the small packed bed bioreactor. The doubling time of bioreactor expanded cells took longer than the 2D controls (T175 flask), 63.5±1.5 hr and 43.7±1.5 hr, respectively for static (Figure 14 A) and 49.2 ± 2.6 hr and

Chapter 3: Bioreactor Application: Isolation and Expansion of Placental Derived Cells on a Packed Bed Bioreactor 63

37.9 ± 0.9 hr, respectively for perfusion (Figure 14 B). A single end point quantification protocol after seven days of expansion was adopted in the perfusion experiments, due to the disruption caused to the cells by taking a measurement every two days. This had a better impact on the MSCs expanded in the bioreactor growth, resulting in reduced differences in the doubling times between perfused bioreactor grown cells and the 2D controls. IVIS images of pMSCs expanded under perfusion conditions displayed an even fluorescent intensity from the scaffold suggesting a homogeneous distribution of cells (Figure 14 C).

A 25 BR 1 20 BR 2 BR 3 15 BR 4 10 2D 1

Fold Expansion Fold 2D 2 5 2D 3 2D 4 0 0 2 4 6 8 10 Days B 25 C

15 *

10 FoldExpansion

0 2D control Bioreactor Figure 14 (A) pMSC expansion in the small-scale 160 cm2 bioreactor (BR) in static and (B) 5 mL/day perfusion (n=4 mean + SD), with the 2D controls (2D). (C) IVIS imaging of PI stained pMSCs under perfusion conditions. IVIS images are a red (low) / yellow (high) heat map of fluorescent intensity.

The cells were characterised by mesodermal tri-lineage differentiation. No differences were observed between the bioreactor expanded cells in both static and perfusion conditions and the corresponding 2D controls when stained with Oil Red O, AlizarinRed S or Alcian Blue (Figure 15A). There was no difference in the

64 Chapter 3: Bioreactor Application: Isolation and Expansion of Placental Derived Cells on a Packed Bed Bioreactor

triglycerides amount and ALP activity (Figure 15B & C). However, there was significantly less GAG production in the cells expanded in the bioreactor under static conditions relative to the 2D control (Figure 15D).

Adipogenic Osteogenic Chondrogenic A 2D Expansion, Static Control B

10 )

Static nM ( 8 Bioreactor Expansion 6

4 2D Expansion, 2 Perfusion

Control Triglycerides per Well 0 Perfusion BR 2D BR 2D Bioreactor Static Perfusion Expansion

C 1.2 D 4 1 3.5 3 * 0.8 2.5 0.6 2

0.4 1.5

GAG (µg)/DNA GAG (µg) Absorbace/DNA(ug) 0.2 0.5 0 0 BR 2D BR 2D BR 2D BR 2D Static Perfusion Static Perfusion

3.2.4 Isolation and Expansion of Placental MSCs Directly in a Small-Scale Packed Bed Bioreactor

pMSCs were isolated from 10 g of digested placenta directly onto the bioreactor and were expanded under 5 mL/day perfusion. The growth rate of pMSCs was slower in the bioreactor compared to the 2D controls (Figure 16 A). The pMSC colonies was

Chapter 3: Bioreactor Application: Isolation and Expansion of Placental Derived Cells on a Packed Bed Bioreactor 65

observed to becoming confluent after 13 days of culture in the T175 flasks (2D controls) and at this point the cells were re-seeded back into the same flask or bioreactor, and the expansion was continued. At day 18 the 2D controls reached confluence and were harvested and characterised by tri-lineage differentiation to determine if isolation and growth had an effect on the pMSCs phenotype. The bioreactor pMSCs’ cultures were harvested at day 22.

25,000

2D Control B Adipogenic Osteogenic Chondrogenic )

2 20,000 Bioreactor

15,000 2D cells/cm

10,000 * Cell Cell Density ( 5,000 BR * # 0 Day 13 Day 18 Day 22 C D E

10 0.25 * 5

4.5 )

8 ) 0.2 4 nM ug 3.5 6 0.15 3 /DNA ( /DNA 2.5 4 0.1 2

1.5 GAG (µg)/DNA GAG (µg)

2 Absorbance 0.05 1 Triglycerides Triglycerides per Well ( 0.5 0 0 0 BR 2D BR 2D BR 2D

Figure 16 (A) Cell growth of pMSCs isolated directly from a placenta in the 160 cm2 packed bed bioreactor undergoing 5 mL/day perfusion (n=4, mean + SD). On day 13, the cells were redistributed back into the same flask or bioreactor. # indicating no 2D control result was measured as the flasks were harvested on day 18. (B) Two week tri-lineage differentiation induction of bioreactor expanded (BR) pMSCs and 2D controls down the adipogenic (Oil Red O, 10x, scale bar is 100 µm), osteogenic (Alizarin Red, 5x, scale bar 500 µm) and chondrogenic (Alcian Blue, 10x, scale bar 500 µm) lineages. Quantification of (C) triglycerides (n=4, mean + SD), (D) ALP activity (n=4, mean + SD) and (E) GAG (n=4, mean + SD) production compared to its matching 2D control. * indicates p<0.01.

The qualitative differential potential of the osteogenic, chondrogenic and adipogenic lineages were similar as shown by staining between the bioreactor expanded pMSCs and the 2D controls (Figure 16 B). However, when quantified

66 Chapter 3: Bioreactor Application: Isolation and Expansion of Placental Derived Cells on a Packed Bed Bioreactor

(Figure 16 C, D & E) there was a significant difference in ALP activity between bioreactor isolated and expanded cells compared to the 2D controls.

3.3 DISCUSSION

Although MSCs sourced from different tissue types have similar surface marker profiles, there is a difference in their epigenetic and genetic expression which affects their mesodermal differentiation capacity [26,49,50,309-311]. In our experimentation there was a significant reduction in the mesodermal differentiation capacity of pMSCs compared to bmMSCs (Figure 11). However, the conditioned medium from pMSCs and bmMSCs both promote HaCaT cell growth with no significant difference between the two donor tissues (Figure 12B). An explanation is that the secretome of MSCs from different tissue sources are similar [311,312]. However, the results were inconsistent and we could not definitively prove that conditioned medium could promote HaCaT growth and that there was a difference with varying the culture times when the experiments were repeated.

It was thought that if the medium was dialysed so that cytokines secreted could be concentrated this could improve the result by allowing the medium to be replaced with fresh medium while still having the secreted MSC cytokines. This resulted in a reduction in the HaCaT growth rate and if it did improve the growth rate it was no better than the unmodified conditioned medium (Figure 13 A). This could be the effect of continued storage at 4 degrees and also the changes in PH and ionic concentrations during the dialysis with the secreted cytokines losing function [313,314].

Another way the MSC conditioned medium can be improved is through culturing the MSCs in aggregates thereby helping to maintain their stem cell like features and this may help to improve their secretome profile [238,306,307]. It was found that the aggregate conditioned medium did improve HaCaT growth compared to the negative control (Figure 13 B). However, it was not better than the traditional culture method and the degree it improved the HaCaT growth compared to the negative control was

Chapter 3: Bioreactor Application: Isolation and Expansion of Placental Derived Cells on a Packed Bed Bioreactor 67

not consistent. Although, the aggregate conditioned medium was more likely to improve HaCaT growth consistently by some degree in repeated experiments compared to the traditional 2D culture conditioned medium that resulted in inconsistent results. This suggested that methods which better select or pre conditioned the MSCs for desired function could result in more consistent results.

The growth rate of the pMSCs in the packed bed bioreactor was less than that in the traditional tissue culture flask (Figure 14 A & B) a result that was similar to the growth rate of the GFP-mMSCs growing in the bioreactor in Chapter 2. Nevertheless, the growth rate observed in our packed bed bioreactor was similar to human MSC expansion in other bioreactor devices [226,228,231,295]. This indicates that the reduce growth rate observed in the bioreactor is not unique. A number of contributing factors could be the cause of the reduce growth rate: initial seeding distribution, seeding efficiency and nutrient concentration. These effects are explored in Chapter 4. In addition, it has been reported that cell can detach from 3D cultures during cell division [315]. However, this occur only during in high cell densities and high shear situations, neither criteria is met in our bioreactor system.

It was noticed that the pMSCs were far more sensitive to the quantification process every two days, and because of this only a single time point was measured for the perfusion expansion. It was difficult to transfer the bioreactor with the tubing circuit between the incubator and tissue culture hood for cell quantification and this disruption impacted negatively on the pMSCs. The use of AlamarBlue method was required to quantify the cell number due to the harvesting method is only 84%±11% efficient with a cell viability of 71%±15% (Figure 8 C). This will add artefact into the calculation of the growth rate in comparison to the 2D controls. In chapter 4, the model was used predict cell death because of the manipulation of the bioreactor during the media exchange for the cell quantification by AlamarBlue, and this contributed to the stunted growth in static conditions that was observed experimentally.

Despite the significant decrease in the growth rate in our bioreactor, the pMSCs maintained the capacity to differentiate down the osteogenic, adipogenic and

68 Chapter 3: Bioreactor Application: Isolation and Expansion of Placental Derived Cells on a Packed Bed Bioreactor

chondrogenic lineages. No differences in tri-lineage differentiation were observed under the perfusion expansion protocol (Figure 15 B, C & D). However, the GAG production in the static bioreactor-expanded cells diminished relative to the 2D controls. As preconditioning MSCs to a higher oxygen concentration reduces their capacity to differentiate down the chondrogenic lineage, reduced GAG production could be explained by a higher oxygen concentration present in the bioreactor compared to the traditional tissue culture flask [201,203]. The liquid layer in a traditional tissue culture flask offers greater resistance to oxygen diffusion resulting in a lower concentration of oxygen at the cell layer [316]. MSCs in the bioreactor located at the liquid-wall interface were exposed to a far higher oxygen concentration. Normoxia conditions are not ideal as for maintain MSCS which have been shown to prefer are more hypoxic 5% O2 atmosphere [199-203,223]. However, it was uncertain if oxygen diffusion alone would be enough to prevent oxygen depletion in the centre of the bioreactor without performing high level modelling in Chapter 4, so to prevent this artefact the bioreactor and flask isolation and expansion experiments were performed in normoxia.

To further characterise the bioreactor’s effect on MSCs’ stemness, and to demonstrate the flexibility of the design, pMSCs were isolated from a digested placenta directly into the bioreactor. The pMSCs produced from this protocol maintained their potential to differentiate down the adipogenic, osteogenic and chondrogenic lineages (Figure 16 B). However, the bioreactor-expanded cells exhibited greater ALP activity than the analogous traditional tissue culture flask expanded pMSCs (Figure 16 D). During isolation, large amounts of red blood cells and other debris from dead cells could not be removed and remained within the bioreactor column. As dead cells have been reported to contribute to, and possibly even induce, osteogenesis in bmMSCs [317], such debris could pre-condition bioreactor-isolated pMSCs toward osteogenic differentiation. The manual syringe driven process used in this experiment made removing the cell debris from the bioreactor difficult. A steady flow rate was difficult to achieve using this manual method, so the flow rate was kept low to avoid detaching the cells from the scaffold. Automated medium flow in future generations of the bioreactor will help prevent this build-up of cell debris, as a controlled continuous flow will allow higher flow rates

Chapter 3: Bioreactor Application: Isolation and Expansion of Placental Derived Cells on a Packed Bed Bioreactor 69

without detaching the cells from the scaffold. An increased flow rate used in the expansion process in the scaled-up bioreactor will also actively remove cell debris.

Although the growth rate of pMSCs in the bioreactor was less than the cells grown on tissue culture plastic, the few extra days required for MSCs to reach confluence in the packed bed bioreactor represents an acceptable trade-off to have the potential for an automatable method to digest, isolate and expand pMSCs in a single closed bioreactor system. An isolation/expansion process based on our design, although performed manually here, could be easily automated [231]. Thus, a single automated bioreactor imbedded with an automated isolation protocol might be developed to digest, isolate and expand pMSCs for clinical use with single procedure in a completely closed system. Failure of the rotatory plasma reactor and time limitations, needed for plasma modification in the larger columns meant that scale up of the pMSC isolation and expansion could not be done. However, in the following chapter, the scale up of the device is explored based on parameters from MSCs. The model developed was used to further aid scale-up designs and improve the culture methods used for the bioreactor. In addition, the model was used to help explain the reduced growth rate observed in the bioreactor compared to the traditional flask cultures.

70 Chapter 3: Bioreactor Application: Isolation and Expansion of Placental Derived Cells on a Packed Bed Bioreactor

Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor

Oxygen is the limiting metabolite in bioreactors due to its low solubility in cell culture medium, and thus is the most difficult to adequately supply through perfusion. As the newly developed bioreactor no longer relies solely on oxygen supplied by the medium perfusion, the flow rate can be greatly reduced to control only the glucose supply. This significantly decreases the shear stress on the cells that could induce differentiation, which is detrimental for the manufacture of MSCs for cell based therapy. Although the bioreactor achieved similar MSC growth rates to other bioreactors reported in literature [226,228,231,295], the growth rate of the MSCs in our bioreactor was significantly less than in traditional tissue culture flasks. It was hypothesised that three factors could have affected the bioreactor cell growth rate: (a) The supply of the nutrients oxygen and glucose was inadequate, leading to significant concentration gradients in the scaffold, (b) The lower flow rate ineffectually removed the waste product, lactate [318,319], (c) A homogenous cell distribution was not achieved in the scaled-up bioreactor. Initial heterogeneous cell distribution during seeding might affect the growth rate and thus and final cell number.

A mathematical model was developed to evaluate the three hypotheses and direct future experiments. Many scaffold perfusion models have been reported for cell expansion on a scaffold for tissue engineering purposes, and models designed to investigate cell growth that focus on nutrient concentration [320-322] and perfusion devices [323-326] formed the basis of our theoretical treatment. Based on previous experimental work [323], Lewis et al. modelled the interaction between oxygen profiles and cell distribution in a cartilaginous tissue construct [322]. Their cell growth

Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor 71

and uptake description were based on early tumour models [320,327] using Michaelis- Menten kinetics for cellular uptake of oxygen. However, their model did not consider cellular migration on the solid substrate or advective phenomena. Their simulation predicted a high cell proliferation around the outside of the scaffold with a region of hypoxia in the centre. Zhao et al. [326] showed the importance of perfusion on MSC growth on a polyethylene-terephthalate scaffold in a perfusion bioreactor, achieving four times the cell density compared to those achieved in static culture conditions. The 2D oxygen concentration profile model suggested the importance of perfusion, with far less depletion of oxygen within the scaffold and lower oxygen gradient compared to the static culture. However, the cell growth rate in their model was constant, the effects of oxygen concentration or cell migration on the cell density were not considered. Coletti et al. [324] developed a far more comprehensive 2D axis of symmetry model that included advection and the diffusive transport of oxygen. The culture medium flow within the bioreactor was modelled in two ways as there was of a channel around the scaffold in the centre of the bioreactor. The medium flow outside the scaffold was described by the Navier-Stokes equation for an incompressible fluid, while the flow through the scaffold was described by the Brinkman’s extension to Darcy’s Law. The cellular nutrient uptake was modelled using Michaelis-Menten kinetics. The cellular growth rate was assumed to be given by the Contois equation, which takes into consideration the nutrient concentration and the cell density. However, cell migration effects were ignored. Chung et al. [328] further developed this bioreactor model that used Michaelis-Menten kinetics for cellular uptake of nutrients, as well as a modified Contois equation for the cellular growth, to include cell migratory terms. They also explored the effect of scaffold permeability and porosity on cell growth. In their model, the cell growth causes a reduction in scaffold pore size that changed the flow through the scaffold and increases shear. This reduction in nutrient supply to the cells impacts the growth rate. Shakeel et al. [329] reduced the complexity of the fluid dynamic system by neglecting the inertial forces of the fluid flow since the Reynolds number is small and used Darcy’s Law to describe the fluid flow through the scaffold [330]. Their model suggested that both shear stresses and reduction in pore size compromised cell growth [325,329]. They also investigated the importance of the initial seeding density distribution on the final total cell number by modelling all the cells as either a central circle or along the walls of the scaffold.

72 Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor

The above models accounted for the effect of only one limiting metabolite (either oxygen or glucose) on the cell growth rate. However, the low flow rate used in the bioreactor design used here required exploring the effect of both oxygen and glucose on the cell growth rate. In addition, the lower flow rate reduced the removal rate of the waste product lactate, which is known to reduce the growth rate of cells [318,319]. The Monod model of mammalian cell growth describes the limiting effect of insufficient oxygen and glucose concentrations plus the inhibitory effect of lactate accumulation [331-333]. However, accounting for complexities of cell metabolism can be difficult. For example, the rate of consumption of glucose and the rate of production of lactate can simply be related mathematically by the stoichiometric yield of lactate- to-glucose. However, the metabolism of cells and their growth rate increases under perfusion conditions, with glucose consumption and lactate production increasing greatly compared to static conditions. The yield of lactate-to-glucose decreases under perfusion conditions, suggesting an increase aerobic respiration from the MSCs preferred anaerobic respiration, possibly associated with the increased oxygen supply due to perfusion which prevents regions of hypoxia [326]. In addition, nutrient consumption or waste production can be influence by cell morphology and density on scaffolds or cytokines levels in the medium. Example of this is normal cell culture conditions the oxygen uptake rate for MSCs is approximately 98 fmol/cell/hr, in the absence of TGF-β the levels reduce down to approximately 12 fmol/cell/hr [334].

In most bioreactors designs, the medium perfusion is the sole source of oxygen. To prevent oxygen depletion and maintain homogenous nutrient concentration in these devices, the flow rate must be increased as the cell numbers grow. To decouple the oxygen supply from the bulk medium flow, oxygen can by supplied by diffusion through interwoven gas permeable fibres in a hollow fibre bioreactor [229] or through a micro-bioreactor wall made of a gas permeable polymer [335]. Kim, et al. [336] modelled the oxygen pressure drop in a microfluidic device where the only source of oxygen was through a diffusive PDMS wall. The 2D diffusive model explored the role of the thickness of the PDMS on cellular density and the time taken for the oxygen to reach equilibrium in the medium. Depending on the cellular uptake rate, it is possible for the cell density to be affected by a wall thickness greater than 5 mm with an

Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor 73

effective diffusion time of 26 minutes, suggesting that diffusion resistance becomes high for a microfluidic scale device.

This Chapter describes a 2D axis symmetric model for MSC growth in the packed bed bioreactor. The work modelled oxygen, glucose and lactate transport in a porous media and their effects on cell growth in the scaffold. The model includes the effects of oxygen transport through the gas permeable wall of the bioreactor. Using this model, the effect of nutrient concentration and initial cell distribution on cellular growth rate, to help explain the observed reduction in growth rate in the bioreactor compared to traditional tissue culture flasks was explored. Furthermore, the effects of scale-up of our bioreactor was investigated by increasing the vessel dimensions and decreasing the particle size of the packed bed, thus increasing the scaffold surface area. The finite element solver package COMSOL was used to solve the model equations numerically.

4.1 DEVELOPMENT OF THE MATHEMATICAL MODEL

4.1.1 Model Description

It was assumed that the cells are initially seeded with a uniform density in the porous scaffold. The scaffold with a radius R1 and a length L, consists of a randomly- packed bed of polystyrene cylinders with 2.5 mm diameter by 3 mm in length. It was assumed that the pore size in the scaffold is not changed significantly by the presences of a confluent cell layer, as such the porosity and permeability of the scaffold remains constant. This is because the cell thickness is at most 1-2 µm compared to the scaffold pore size that could be greater than 1 mm, meaning that the cell layer has negligible reduction of the pore size. The fluid (cell culture medium) is assumed to be viscous, incompressible and Newtonian. It is pumped into the scaffold at the boundary z=0 and leaves at boundary z=L as shown in Figure 17. The fluid is pumped at a constant volumetric flow rate Q. The fluid contains oxygen and glucose which is consumed by the cells. A small amount of lactate is present initially in the cell culture medium

74 Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor

associated with the addition of foetal bovine serum. The bioreactor wall is made from a gas permeable polymer of a thickness R2-R1, polydimethylsiloxane (PDMS), which provides for the radial diffusion of oxygen through the bioreactor wall. The oxygen solubility in the PDMS is greater than that of the medium.

A B z z

L L

A B Oxygen Diffusive Flux A B Oxygen Diffusive Flux

r r R1 R2 R1 R2 Q C D z z L L

A Oxygen Diffusive Flux C A B Oxygen Diffusive Flux

r r R1 R2 Rc Rw R1 R2 Q Q

Figure 17 (A) Axis of symmetry diagram of perfusion bioreactor with a gas permeable wall. This allows oxygen to diffuse radially into the bioreactor. The radius of the PDMS cylinder is denoted by R2, the radius of the scaffold volume of the bioreactor is denoted by R1 and the length is denoted by L. In the equation development the radial coordinate was denoted by r and the longitudinal coordinate by z (measured from the inlet). (B) The experiments were also performed in a batch configuration where the medium was replaced every two days, however oxygen could still diffuse through the wall. (C) Traditional packed/fixed bed bioreactor where all the nutrient and oxygen are provided by medium perfusion alone. (D) Further enhancement of the design by the incorporation of a central capillary with outer wall radius Rw and lumen Rc in the centre of the reactor provides an additional oxygen source.

The base model equations stated in the next section were adjusted to explore a batch configuration (Figure 17B), and they were used to compare our bioreactor design to a traditional packed bed reactor (Figure 17C). For scaling-up the bioreactor, it may be beneficial to provide additional sources of oxygen through gas permeable capillaries. The model was further modified to simulate this situation with a central gas-permeable capillary providing oxygen (Figure 17D).

Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor 75

4.1.2 Development of Model Equations

The model leads to five partial differential equations. The first characterises the fluid flow through the porous material with velocity (U) driven by the pressure (P). The second is a mass balance for the convective and diffusive transport of the nutrients; oxygen (CO) and glucose (CG), and the waste product lactate (CL). The third equation represents the cell density (N) that includes terms describing the proliferation and migration within the scaffold.

Fluid Flow Through the Scaffold

The fluid velocity through the bioreactor scaffold is small, so that inertial forces can be neglected, thus the fluid flow can be regarded as a flow through a porous media given by Darcy’s Law. This relates velocity to the interstitial pressure as a function of the material permeability (K*) and the fluid dynamic viscosity 휇.

Given by

퐾∗ 푼 = − ∇푃, (1) 휇

where

2 3 푑푝Φ 퐾∗ = . (2) 180(1 − Φ)2

Assuming that the cells do not occupy a significant fraction of the scaffold pore space, the permeability (K*) is taken to be constant and is defined in terms of the particle diameter equivalent to a sphere (dp) and the scaffold porosity (Φ) [330]. The continuity equation the fluid which was taken to be incompressible is

∇. 푼 = 0 . (3)

76 Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor

The following boundary conditions are applied. As the fluid is viscous, there is no slip along the bioreactor wall, and axisymmetric flow was assumed, so that

푼풓 = 0 { at r=0 and R1, 0

The flow at the entrance leads to

푄 퐔풛 = 2 at 0

where Q is the constant fluid flow rate (cm3/hr).

Development of Nutrient Conservation Equation

The cells require oxygen and glucose to survive and proliferate. Lactate is produced through the metabolism of glucose and other metabolites and inhibits cell survival and proliferation. The transport equation for the nutrients is in terms of advection, diffusion and consumption of glucose and oxygen and production term for lactate. Both glucose (CG) and lactate (CL) have similar conservation of mass equations and boundary conditions. However, as lactate is a product of glucose metabolism, the lactate production rate was related to glucose consumption rate by simply multiplying by the lactate yield from glucose (YL/C) [318,319,331]. The following transport equations was obtained:

휕퐶퐺 2 푉푚푎푥,퐺푁퐶퐺 = 퐷퐺∇ 퐶퐺 − ∇. (푼퐶퐺) − , (6) 휕푡 퐾푚,퐺+퐶퐺

and

휕퐶퐿 2 푉푚푎푥,퐺푁퐶퐺 = 퐷퐿∇ 퐶퐿 − ∇. (푼퐶퐿) + 푌퐿/퐶 . (7) 휕푡 퐾푚,퐺+퐶퐺

Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor 77

The consumption of glucose can be assumed to be modelled using Michaelis-

Menten kinetics. This is a function of the maximum uptake/production rate (Vmax,I, -1 -1 mM.hr .cell ), the Michaelis-Menten constant (Km,I, mM), the cell density (N, 3 cells/cm ) and the concentration (Ci). The diffusion coefficient must be modified to account for the restriction to diffusion associated with a bioreactor that contains non- gas-permeable pellets that retard the overall diffusion. This modification is defined by the voidage (휀), tortuosity (휏) and the free diffusion coefficient (Dm,i). The tortuosity is defined in (9) as a function of the porosity and the sphericity (휑) of the particles [337]. The sphericity of non-spherical particle is given by (10) as a function of the number of the particles per unit volume of the packed bed (n), the porosity, and the surface area to volume ratio (a) of the packed bed. Defined by

휀 퐷 = 퐷 , 푖 휏 푚,푖 (8)

where

4 1.23(1−휀)3 (9) 휏 = , 휀×휑2

and

1 (36×휋×푛(1−휀)2)3 휑 = . (10) 푎

A constant concentration of glucose and lactate at the inlet (11) and that there is no longitudinal diffusive flux at the outlet (12) and the at reactor wall (13) was assumed. The symmetry condition gives a zero flux for all species at r=0 (14). In summary, these are given by

퐶푖 = 퐶푖푛,푖 at z=0, 0

휕퐶푖 = 0 at z=L, 0

78 Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor

휕퐶푖 = 0 at r=R1, 0

휕퐶 푖 = 0 at r=0, 0

Oxygen conservation leads to two equations, one to define the oxygen concentration, in the scaffold (15) and the other for the oxygen diffusion in the wall (16), these are given by

휕퐶푂,퐴 2 푉푚푎푥,푂푁퐶푂,퐴 = 퐷푂∇ 퐶푂,퐴 − ∇. (푼퐶푂,퐴) − for 0

휕퐶푂,퐵 2 = 퐷 ∇ 퐶 for R1

For expression 15 the boundary conditions are set as no flux through inlet and outlet wall with constant concentration of oxygen at the outer wall. This is defined as the solubility of oxygen in PDMS (SPDMS, mM)

휕퐶푂,퐵 = 0 at z=0, L, R1

퐶푂,퐵 = 푆푃퐷푀푆 at r=R2, 0

At the interface between the wall and the scaffold, continuity of the diffusive flux (19) and a change in concentration due to the difference in solubilities (21) was required. The solubility of oxygen in the fluid is denoted by SM (mM). The boundary conditions are

Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor 79

휕퐶푂,퐴 휕퐶푂,퐵 퐷 = 퐷 at r=R1, 0

and

퐶푂,퐴 퐶푂,퐵 = at r=R1, 0

Within the scaffold itself, the inlet fluid pumped is oxygenated, there is no diffusive flux at the outlet and at the axis of symmetry (25).

퐶푂,퐴 = 푆푀 at z=0, 0

휕퐶푂,퐴 = 0 at z=L, 0

휕퐶 푂,퐴 = 0 at r=0, 0

Development of Cell Growth Equation

The local cell density changes over time, due to proliferation and the cells are free to migrate on the pellets of the scaffold. The proliferation rate is taken to be a 3 logistic kinetic form with maximum cell density (Nmax, cell/cm ) and an intrinsic growth rate (k, hr-1) [320]. The growth rate dependence is taken as a Monod kinetic equation that is dependent on the nutrient concentrations (Ci), maximum growth (Kmax, -1 hr ) and the Michaelis-Menten constants (KG, KO and KL, mM) [321,331]. The migration is governed by the diffusion coefficient (DN) and the cell density, so that

휕푁 푁 2 = 푘푁 (1 − ) + 퐷푁∇ 푁, (24) 휕푡 푁푚푎푥 where 퐶퐺 퐶푂,퐴 퐾퐿 푘 = 퐾푚푎푥( ). 퐾퐺+퐶퐺 퐾푂+퐶푂,퐴 퐾퐿+퐶퐿 (25)

The boundary condition for the cell growth is derived as follows. The cells cannot leave the scaffold and r=0 axis of symmetry. To obtain

80 Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor

휕푁 = 0 at z=L and z=0, 0

휕푁 = 0 at r=R1, 0

휕푁 = 0 at r=0, 0

Shear Stress

The shear stress (휎, Pa) acting on the cells can be estimated from the Darcy velocity (U), porosity (ε), tortuosity (τ), viscosity (µ) of the medium and the pore diameter of the packed bed (DH) as follows

8휇휏‖푈‖ 휎 = ∗ , (29) 퐷퐻퐾

The pore diameter is a function of the voidage and the particle diameter [338] so that

2휖푑 퐷 = 푝 . 퐻 3(1−휀) (30)

4.1.3 Initial Conditions and Parameters

Initially the cells are seeded with a uniform density of 1000 cells/cm2, which must be converted to cells per unit volume by multiplying the surface area to volume ratio (a).

so that

푁(푟, 푧, 0) = 푁푖 at 0

Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor 81

It is assumed that the medium and the PDMS are initially in equilibrium with the atmospheric oxygen concentration so that

퐶푂,퐴(푟, 푧, 0) = 푆푀 at 0

The reactor is filled with cell culture medium initially so that

퐶퐺(푟, 푧, 0) = 퐶푖푛,퐺 at 0

퐶퐿(푟, 푧, 0) = 퐶푖푛,퐿 at 0

Table 5 summarises the input parameters for the model, with * representing a measure value from our experiments.

82 Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor

Table 5 Model parameters and values used in model

Parameter Description Value Units Ref Bioreactor Parameters R1 Inner wall radius 0.75 cm * R2 Outer wall radius 1.25 cm * L Scaffold length 7.5 cm * Growth area of scaffold 158.8 cm2 * Scaffold volume 13.25 cm3 휺 Voidage 0.47 - * 횽 Porosity 0.47 - * a Surface area to volume ratio of the scaffold 12.013 cm-1 * n Number of particles per volume of packed 36 cm-3 * bed dp Particle diameter (equivalent to a sphere) 0.304 cm * Q Medium flow rate 0.208 cm3/h * Wall Properties SPDMS Solubility of oxygen in PDMS 1.3 mM [339] -5 2 DPDMS Oxygen diffusion coefficient in PDMS 5x10 cm /s [326] Fluid Properties Sm Solubility of oxygen in medium 0.2 mM * Cin,G Initial glucose concentration 4.81 mM * Cin,L Initial lactate concentration 1.63 mM * -5 2 Dm,O Oxygen diffusion coefficient in medium 3.290x10 cm /s [326] -6 2 Dm,G Glucose diffusion coefficient in medium 5.4x10 cm /s [340] -5 2 Dm,L Lactate diffusion coefficient in medium 1.1x10 cm /s [341] 흆 Density of water @ 37°C 0.9933 g/cm3 [338] 흁 Viscosity of water @ 37°C 0.00692 kg/(m.s) [338] Cell Properties 2 Nmax Maximum cell density 23,000 cells/cm * 3 Ni Initial cell value 12013 cells/cm -10 2 DN Diffusion coefficient for cell movement 1.38x10 cm /s [342] Vmax,O Maximum uptake rate of oxygen 100 fmol/h/cell [334] Vmax,G Maximum uptake rate of glucose 272 fmol/h/cell [334] YL/G Yield of lactate from glucose 2.0 - [318] -9 3 Km,O Michaelis-Menten constant for oxygen 4.05x10 mol/cm [343] uptake -7 3 Km,G Michaelis-Menten constant for glucose 3.5x10 mol/cm [344] uptake -1 Kmax Max growth rate 0.02631 h * KO Oxygen Michaelis-Menten growth 0.001 mM [345] constant KG Glucose Michaelis-Menten growth 0.006 mM [345] constant KL Lactate Michaelis-Menten growth 43 mM [318] constant

Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor 83

4.1.4 Solution Method

The model consists of five partial differential equations representing fluid flow through the scaffold (1), nutrient transport of glucose (6), lactate (7), oxygen (15 & 16) and the cell density (24). Due to the complexities of the system, an analytical solution cannot be found. The commercially available finite element software COMSOL (vers. 4.4) was used to solve the governing equation.

A free triangular mesh was generated and refined successively until a convergent result was achieved (Appendix 4). The refined mesh consisted of 9,628 elements with 500 boundary elements. The dependent variables are approximated by a quadratic shape function and the system was solved for 43,493 degrees of freedom. The time step was set to 0.1 hr for a total of 8 days (192 hr).

4.2 RESULTS AND DISCUSSION

4.2.1 Base Case- Decoupling the Oxygen Supply from the Bulk Medium is Superior to a Traditional Packed Bed Bioreactor

In experimental studies of the packed bed bioreactor, a significant reduction in the MSC growth in the bioreactor compared to cells grown in a traditional T175 flask was observed in Chapters 2 and 3. The axial symmetric model of the bioreactor was developed to explore this phenomenon and to further develop and optimise this bioreactor design.

84 Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor

A B

C D

Figure 18 Model results: heat map of (A) cell density, (B) glucose, (C) lactate and (D) oxygen concentration in the wall and in the vessel on day 7. The white arrow on the oxygen concentration heat map represents the direction of total flux of oxygen while the black arrow represents the direction of the diffusive flux.

In the small scale experiments, medium was perfused at a flow rate of 0.208 mL/hr, as this flow rate was assumed to provide enough glucose to prevent depletion, and that enough oxygen was supplied by radial diffusion through the bioreactor wall. An initial uniform cell distribution resulted in a final non-uniform cell distribution by the end of the culture period (Figure 18A). For an understanding of how to read the heat map figures and explanation of an axial symmetric model refer to Appendix 5. It was concluded that the cell density gradient is driven strongly by glucose or lactate concentrations, rather than by oxygen, as there is a large difference in the cell density between the inlet and outlet with only a small difference in the cell density in the radial

Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor 85

direction. These concentration differences and cell gradients increased over time, driven by cell growth, consumption of nutrients and production of lactate (Appendix 6, Appendix 7, Appendix 8 and Appendix 9). There was a 51% reduction in the glucose concentration at the outlet compared to the inlet (Figure 18 B) and a similar increase in the lactate concentration (Figure 18 C). Nevertheless, the model supports the hypothesis that radial diffusion of oxygen is sufficient to support the cell growth in the bioreactor without depleting oxygen in the centre of the bioreactor (Figure 18 D).

When the model was modified to explore the increasing the wall thickness, it was found that the wall offered little resistance to the diffusive flux of oxygen (Appendix 10), and had little overall effect on the cell density (Figure 19 A). When the base model was adjusted to represent a traditional packed bed bioreactor (Figure 19 B) by removing the oxygen equation for the wall expression 16 and replacing the boundary condition equations (17-20) with the boundary condition of no oxygen flux through the wall, that is

휕퐶푂,퐴 = 0 at r=R1, 0

The flow rate used in the experiments (0.208 ml/hr) was not adequate to support the cell growth in a traditional packed bed reactor. A flow rate of 1.5 ml/min was required to achieve a similar cell growth as in the base model (Figure 19 B).

86 Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor

25000 Base model 0.1 cm 20000 0.25 cm 15000 0.5 cm 1 cm 10000 2 cm 5000 5 cm 0 Cell density (cell/cm2) density Cell 0 1 2 3 4 5 6 7 8 Day B 25000 Base model 20000 0.208 ml/h 0.5 ml/h 15000 1 ml/h 10000 1.5 ml/h 5000

Cell/density (cell/cm2) Cell/density 0 0 1 2 3 4 5 6 7 8 Day C 25000 Base model 20000 Experimental bioreactor 15000 expansion Experimental T175 flask 10000 expansion 5000

Cell density (cells/cm2) density Cell 0 0 1 2 3 4 5 6 7 8 D Day 25000 Base model 20000 0.5 ml/hr 1 ml/hr 15000 1.5 ml/hr 10000 3 ml/hr 5000 4 ml/hr

Cell density (cells/cm2) density Cell 0 0 1 2 3 4 5 6 7 8 Day Figure 19 (A) The average cell density change as a result of increasing the wall thickness from 0.1 to 5 cm. (B) A bioreactor model average cell density without wall diffusion requires a higher medium flow rate compared to our bioreactor with wall diffusion (Base model). (C) The average cell densities achieved experimentally from T175 flask and bioreactor expanded cells were compared to base model output. Note that only final cell numbers were measured for the bioreactor expanded cells; the curve represents exponential approximations of the previous days based on the overall growth rate from the experiment. (D) Predicted average cell density at different flow rates in a gas permeable wall packed bed bioreactor.

Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor 87

The model fits well to the traditional T175 flask expanded cell results but not to the results obtained in bioreactor expansion (Figure 19 C). The base model was modified to explore why the bioreactor growth is significantly less than that of the traditional flask expanded cells. Increasing the model flow rate to 0.5, 1, 1.5, 3 and 4 mL/hr results in only a minor increase in the average cell density (Figure 19 D). Although the cells became more evenly distributed (Figure 20A-C) as a result of the improved removal of lactate (Figure 20D-F) and the increased supply of oxygen (Figure 20G-I), this does suggest that the flow rate used in the experiments (0.208 ml/hr) was near optimal. The shear stress acting on the cells was also significantly less at all flow rates than the 0.015 Pa shown to cause osteogenic differentiation of MSCs (Appendix 11) [218-221].

A B C

D E F

G H I

Figure 20 Heat maps of (A-C) cell density distribution (D-F) lactate concentration (G-I) and oxygen concentration at 0.5, 1.5, 3 cm3/hr respectively.

88 Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor

To determine if the 0.208 ml/hr flow rate used experimental in the small-scale bioreactor was adequate, and to also determine which nutrient or if the waste product was limiting the growth rate, the intrinsic growth rate was modified in the cell balance expression (24). This allows the investigation of the dependence of the average cell density on each nutrient as well as the cell contact inhibition, as presented in equations in Table 6. Other than cell contact inhibition, lactate showed the largest effect in reducing cell growth (Figure 21 A). This suggested that glucose and oxygen were both in sufficient supply in our bioreactor, and the retarded cell growth resulted from the compromised removal of lactate due to the low flow rate. It was concluded that, lactate is the main driving force in developing the cell gradient across the length of the reactor. The KL value, which is the concentration of lactate to reduce the growth rate by half, is only 26 times larger than the initial concentration of lactate whereas the Ko is 200 times smaller than the initial oxygen concentration. As a consequence, the final cell number is much more sensitive to changes in lactate concentration than it is to the changes in the oxygen concentration. This is the case even if the oxygen concentration is reduced by a factor of two below the current inlet value.

Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor 89

Table 6 Cell growth rate equations adjusted to account for different limitations in nutrient concentrations and cell contact inhibition.

Condition Cell density equation

휕푁 퐶퐺 퐶푂,퐴 퐾퐿 푁 2 Base case = 퐾푚푎푥( )푁 (1 − ) + 퐷푁∇ 푁 휕푡 퐾퐺 + 퐶퐺 퐾푂 + 퐶푂,퐴 퐾퐿 + 퐶퐿 푁푚푎푥 Cell contact 휕푁 푁 2 = 푁 (1 − ) + 퐷푁∇ 푁 inhibition 휕푡 푁푚푎푥 휕푁 퐶푂,퐴 푁 2 Oxygen = 퐾푚푎푥( )푁 (1 − ) + 퐷푁∇ 푁 휕푡 퐾푂 + 퐶푂,퐴 푁푚푎푥 휕푁 퐶퐺 푁 2 Glucose = 퐾푚푎푥( )푁 (1 − ) + 퐷푁∇ 푁 휕푡 퐾퐺 + 퐶퐺 푁푚푎푥 휕푁 퐾퐿 푁 2 Lactate = 퐾푚푎푥( )푁 (1 − ) + 퐷푁∇ 푁 휕푡 퐾퐿 + 퐶퐿 푁푚푎푥 Oxygen and 휕푁 퐶퐺 퐶푂,퐴 푁 2 = 퐾푚푎푥( )푁 (1 − ) + 퐷푁∇ 푁 glucose 휕푡 퐾퐺 + 퐶퐺 퐾푂 + 퐶푂,퐴 푁푚푎푥 Oxygen and 휕푁 퐶푂,퐴 퐾퐿 푁 2 = 퐾푚푎푥( )푁 (1 − ) + 퐷푁∇ 푁 lactate 휕푡 퐾푂 + 퐶푂,퐴 퐾퐿 + 퐶퐿 푁푚푎푥 Glucose and 휕푁 퐶퐺 퐾퐿 푁 2 = 퐾푚푎푥( )푁 (1 − ) + 퐷푁∇ 푁 lactate 휕푡 퐾퐺 + 퐶퐺 퐾퐿 + 퐶퐿 푁푚푎푥 Oxygen, 휕푁 퐶퐺 퐶푂,퐴 퐾퐿 2 glucose and = 퐾푚푎푥( )푁 + 퐷푁∇ 푁 lactate only 휕푡 퐾퐺 + 퐶퐺 퐾푂 + 퐶푂,퐴 퐾퐿 + 퐶퐿

The oxygen supply would need to be close to depletion for it to strongly affect the MSC growth rate. This result should be interpreted carefully, as the oxygen sensitivity can vary depending upon the cell type. MSCs prefer anaerobic respiration, but other cell types that prefer aerobic respiration result in a far higher Ko [334]. Given that the lactate/oxygen yield YL/G ≈ 2 for most cell types and culture conditions, the removal of waste products is more important than the supply of nutrients [261,334,346]. Adding glucose to the medium without increasing the flow rate could lead to accumulation of lactate in the reactor, thereby diminishing the cell growth rate. Nevertheless, in our bioreactor system these simulations suggest that nutrient limitations are only a minor influence in the reduction of the experimental growth rate observed in the bioreactor. Therefore, other factors must be taken into account to explain the experimental cell growth in the packed bed bioreactor.

90 Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor

83,020 cells/cm2 25,000 Base model

20,000 Cell contact inhibition ) 2 Oxygen only

15,000 Glucose only

Lactate only 10,000 Cell density (cells/cm densityCell Oxygen and glucose

Oxygen and Lactate 5,000 Glucose and lactate

0 Oxygen, glucose and lactate only 0 1 2 3 4 5 6 7 8 Day B Base model

25,000 500 cell/cm2

20,000 600 cell/cm2

) 2 700 cell/cm2 15,000

800 cell/cm2 10,000

Cell density (cells/cm densityCell 900 cell/cm2

5,000 Experimental bioreactor expansion

0 Experimental T175 flask 0 1 2 3 4 5 6 7 8 expansion Day

Figure 21 (A) The average cell density with different cell growth rate equations defined in Table 6. (B) The average cell density with the initial homogeneous cell distribution adjusted from 500 cell/cm2 to 1000 cell/cm2 (base model).

4.2.2 Investigating the Effect of Cell Distribution on Growth

In this treatment it was assumed that, initially, all the cells were attached to the scaffold with a homogeneous distribution. However, this is almost impossible to achieve. It is more likely that the cell seeding method would naturally give rise to a heterogeneous cell distribution. Although having fewer cells attached to the scaffold reduces the final cell number (Figure 21 B). However, this was not observed to be

Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor 91

experimentally significant (see Chapter 3, Figure 9 A). The cell seeding method strongly influenced the cell distribution heterogeneity and the cell viability [347].

To model a heterogeneous cell distribution, but while still maintain same total cell number. The initial cell distribution was assumed to linearly decrease in either the axial (z) or radial (r) direction by specifying either the inlet or centre cell density (Nin) between 0 cell/cm2 to 2000 cell/cm2 (note that this is converted to a volume concentration by multiplying by the factor, a, for the simulation). This is to represent poor seeding efficiency with cells getting trapped at the inlet or settling to the bioreactor wall during the rocker roller seeding process. The linear equations represent this are as follows:

푧 푁 = 푁 + (푝 − 푁 ) at 0

for a z-direction linear cell distribution and

푟 푁푖푟 = 푁푖푛 + (푝푝 − 푁푖푛) at 0

for the r-direction linear cell distribution.

For both the cell distribution the constants variable p & pp have to be found to 2 2 ensure the total cell number, T, remain fixed. T=1000 cells/cm x 158.8 cm ≈ 158,800 2 cells. Expression (37) needed to be integrated between z=0 cm to z=L for πR1 Niz=T, therefore

푇 퐿 푧 2 = ∫ [푁푖푛 + (푝 − 푁푖푛) ] dz , (39) 휋푅1 0 퐿

2푇 푝 = 2 − 푁푖푛. (40) 휋푅1퐿

92 Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor

To solve pp, expression (38) needs to be integrated between r= 0 and r= R1 with

2πR1LNiz=T, therefore

푇 푅1 푟 = ∫ 푁푖푛 + (푝푝 − 푁푖푛) dz , (41) 2휋푅1 퐿 0 푅1

푇 푁푖푛 푝푝 = 3 ( 2 − )at 0

The initial cell distribution as given by (37) and (38) can be seen in Figure 22A & D, respectively.

Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor 93

A B

C D

E F

Figure 22 (A) Cell density heat map of linear cell distribution in the z direction at day 0 and with an 2 2 inlet concentration of 2000 cell/cm . (B) Cell distribution at day 7 with Nin=2000 cell/cm with a linear 2 cell distribution in the z direction. (C) Lactate concentration at day 7 with a Nin=2000 cell/cm with a linear cell distribution in the z direction. (D) The initial cell density of linear cell distribution in the r direction with a centre concentration of 2000 cell/cm2. (E) Cell density of linear cell distribution in the 2 r direction at day 7 with Nin=2000 cell/cm . (F) Lactate concentration for linear cell distribution in the 2 r direction with Nin=2000 cell/cm at day 7.

A linear heterogeneous cell distribution strongly impacts the final average cell density, more so in the z-direction than in the r-direction (Figure 23 A & Figure 24 A). This is shown in the cell density heat maps where a larger final cell distribution is developed in the z-direction. The cell density in the inlet half of the bioreactor almost reaches confluence, while the outlet half has a cell density of 114 cell/cm2 (Figure 22

94 Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor

B). Conversely, the cell density gradient is reduced with an initial non-uniform linear cell distribution in the r direction, with a difference of 2x104 cells/cm2 in the centre of the bioreactor compared to 1.35x104 cells/cm2 at the reactor wall (Figure 22 E). This difference is firstly due to the fact that the initial distribution in the r direction is smaller than in the z-direction (Figure 22 A & D). However, the main driver of a large cell density gradient is the lactate build-up inhibiting the cell growth, which is more significant when the initial cell distribution in the z-direction is non-uniform. Although the lactate concentration does not reach as high a level in the z-direction heterogeneous seeding, a larger volume of the bioreactor is exposed to a higher lactate concentration (Figure 22 C & F). This results in a reduced growth rate for the cells located closer to the outlet. When this effect was observed in the heterogeneous initial cell distribution in the r direction, a cell density gradient developed along the length of the reactor (Figure 22 E).

We concluded that a heterogeneous initial cell distribution alone is insufficient to cause the significant difference that was observed between the cell yield in the experimental bioreactor expansion and the traditional flask culture. However, 24 to 48 hour delay in the cell growth was consistently observed in all experimental expansions, consistent with literature reports [199,223,239]. The quantified cell density measured on day two illustrates this lag in the experimental static bioreactor expansion growth curve (Figure 25). We hypothesise that this represents a phase in which the cells were recovering from extra handling and stress during the seeding process. We investigated the effect of a time lag on cell proliferation by examining 2 cases (24 hr and 48 hr). For the first 24 hr (48 hr) the cells could migrate but not proliferate. The growth term in equation (25) was modified as follows:

퐶퐺 퐶푂,퐴 퐾퐿 푘 = 퐻 (퐾푚푎푥 ( )). (43) 퐾퐺+퐶퐺 퐾푂+퐶푂,퐴 퐾퐿+퐶퐿

H is the Heaviside step function

0, 푡 < 24 (48) ℎ푟 퐻[푡] = { 1, 푡 ≥ 24 (48) ℎ푟. (44)

Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor 95

25,000

20,000

15,000

10,000

Cell density (cells/cm2) density Cell 5,000

0 0 1 2 3 4 5 6 7 8 Day B

25,000

20,000

15,000

10,000

Cell density (cells/cm2) density Cell 5,000

0 0 1 2 3 4 5 6 7 8 C Day

25,000

20,000

15,000

10,000

Cell density (cells/cm2) density Cell 5,000

0 0 1 2 3 4 5 6 7 8 Day

Figure 23 Heterogeneous cell distribution are modelled as a linear function in the z direction, setting the cell concentration at the inlet (z= 0 cm) between 0 to 2000 cell/cm2with (A) 0 hr, (B) 24 hr and (C) 48 hr cell quiescence. Note that the total initial cell number is the same in all conditions (158,800 cells).

96 Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor

25,000

20,000

15,000

10,000

Cell density (cells/cm2) density Cell 5,000

0 0 1 2 3 4 5 6 7 8 Day B

25,000

20,000

15,000

10,000

Cell density (cells/cm2) density Cell 5,000

0 0 1 2 3 4 5 6 7 8 Day C

25,000

20,000

15,000

10,000

Cell density (cells/cm2) density Cell 5,000

0 0 1 2 3 4 5 6 7 8 Day

Figure 24 Heterogeneous cell distribution modelled as a linear function in the r direction by setting the centre cell density (r=0) between 0 to 2000 cell/cm2 with 0 hr (F), 24 hr (G) and 48 hr (H) cell quiescence. Note that the total initial cell number is the same in all conditions (158,800 cells)

Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor 97

The modification to the model resulted in a better fit with the experimental bioreactor growth data (Figure 23 B & C, Figure 24 B & C). This suggests that a combination of a non-uniform initial cell distribution in the axial direction and an initial time lag in cell proliferation may explain the reduced cell yield seen in the bioreactor compared with the yield in the traditional flask culture. Thus, distributing cells uniformly during the seeding process while minimising stress to the cells is the most important factor in achieving efficient cell expansion in a 3D culture device such as a bioreactor. This is very difficult to achieve in practice. While many seeding methods have been reported [296-298,347,348], inhomogeneity and cell stress have greatly affected the overall cell expansion in each.

4.2.3 Modelling a Batch Bioreactor

The model was modified to a batch reactor configuration, to investigate whether the cell distribution and cell growth lag was the main factor in bioreactor reduced growth rate compared to a flask culture. To do this, the terms involving the fluid flow through the scaffold were removed, and the mass balance equations for glucose (6), lactate (7) and oxygen (15) were modified as follows,

휕퐶퐺 푉푚푎푥,퐺푁퐶퐺 = ∇. (퐷퐺∇퐶퐺) − , (45) 휕푡 퐾푚,퐺+퐶퐺 and

휕퐶퐿 푉푚푎푥,퐺푁퐶퐺 = ∇. (퐷퐿∇퐶퐿) + 푌퐿/퐶 , (46) 휕푡 퐾푚,퐺+퐶퐺 and

휕퐶푂,퐴 푉푚푎푥,푂푁퐶푂,퐴 = ∇. (퐷푂∇퐶푂,퐴) − for 0

The source boundary conditions (11) and (21) were also removed, with the addition of no flux boundary condition at the inlet for glucose, lactate and oxygen, defined as

휕퐶푖 = 0 at z=0, 0

98 Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor

In the experimental protocol, a full volume medium exchange was performed every two days. To model this, the simulation was run for 48 hours. The cell density output was used as the initial cell distribution for the next simulation of 48 hours. This was repeated for the required 8 days of cell growth.

25,000 Base model

Experimental bioreactor expansion 20,000 Experimental T175 flask expansion

Experimental static bioreactor expansion Full volume medium exchange every 2 days 15,000 No media exchange

10% cell death each media exchange 24 hr quiescence- 10% cell death

10,000 Cell density density Cell (cells/cm2) 48 hr quiescence- 10% cell death in media exchange 48 hr quiescence- 10% death- 2000 cells/cm2 inlet cell density 5,000 48 hr quiesences- 15% cell death in media exchange 48 hr quiescence - 15% death - 2000 cells/cm2 inlet cell density

0 0 1 2 3 4 5 6 7 8 Day Figure 25 Experimental and model average cell densities of a batch bioreactor with a full volume exchange every two days, exploring the effect of no medium exchange, cell death with every medium exchange, quiescence due to cell shock and heterogeneous cell distribution

The cell growth in a batch reactor system was close to predicted for the perfusion bioreactor. The batch reactor had a similar growth rate even with no medium exchanges (Figure 25). However, the batch bioreactor experimental results were significantly less than the experimental perfusion results. This was likely due to the process of measuring the cells every two days, which required multiple full-volume medium exchanges as the metabolic probe in the medium was used to infer the cell

Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor 99

numbers. The resulting stress to the cells possibly killed or detached and washed away some of the cells. To explore this hypothesis, the model was modified so that this was reduced by 10% to 15% when the output cell density from the previous time period was used as the initial value. A heterogeneous initial cell distribution and lag phase was also added in order to produce a model output that fitted the static bioreactor expansion. This further confirmed that heterogeneous cell distribution and a lag phase are the main driving factors of the reduced growth in the bioreactor. The output confirms that measuring the cell numbers after each medium exchange was killing or detaching a proportion of the cells. These data support the change in protocol for the perfusion bioreactor expansion to cell quantification only at the end of the culture period.

4.2.4 Using the Model to Investigate Scaling Up

A packed bed bioreactor can be scaled-up in two ways by either increasing the dimensions of the reactor itself, or by reducing the size of individual packing material to increase the surface area. In chapter 2, scaled-up of the bioreactor to increase the reactor volume while retaining the same pellet size was achieved using GFP-mMSCs. However, equipment limitations prevented us from optimising the flow rate. A cyclic perfusion system was employed, rather than the preferred single pass circuit, and the lowest flow rate available was 30 ml/hr. In order to direct further scale up experiments and their optimisation, simulations were performed using the model that has been developed.

The geometry of the model was changed to represent the bioreactor used in the trial scale-up experiments. The radius was adjusted so that the scaffold radius, R1= 2.5 cm while the wall thickness remained at 0.5 cm, therefore R2= 3 cm. The length was increased to L=12 cm, which increased the reactor volume to 235.6 cm2. With the same pellet size as the base model this increased the reactor surface area to 2830 cm2. Assuming a cell density of 20,000 cells/cm2, this surface area is enough to support 5.66x107 cells.

100 Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor

A B

25,000 0.0005

20,000 0.0004

15,000 0.0003

0.0002

10,000 (Pa) force Shear cell density (cells/cm2) density cell

5,000 0.0001

0 0 0 2 4 6 8 30 ml/hr 20 ml/hr 15 ml/hr 10 ml/hr Day Wall diffusion 30 ml/hr Wall diffusion 20 ml/hr Wall diffusion 15 ml/hr Wall diffusion 10 ml/hr No wall diffusion 30 ml/hr No wall diffusion 20 ml/hr No wall diffusion 15 ml/hr No wall diffusion 10 ml/hr Center and wall diffusion 30 ml/hr Center and wall diffusion 20 ml/hr Center and wall diffusion 15 ml/hr Center and wall diffusion 10 ml/hr Small scale base model

The average cell density achieved in the scaled up bioreactor was similar to that obtained in the base model with a medium perfusion rate of 15 ml/hr (Figure 26 A). The shear stress was significantly below the 0.015 Pa which induces osteogenic differentiation at all flow rates (Figure 26 B). However, below 15 ml/hr it was apparent that the radial oxygen diffusion may not be enough to support cell growth, with the oxygen concentration coming close to depletion (Figure 27 A & D). This suggests that additional oxygen sources are required if the diffusion length is greater than 1 cm. Nevertheless, the design superiority of the wall diffusion bioreactor over the traditional packed bed bioreactor, with no oxygen sources other than through medium perfusion, is more apparent in the scaled up bioreactor. With medium perfusion flow rates below 20 ml/hr one starts to observe a decrease in the average cell density (Figure 26 A). At flow rates of 15ml/hr there is an increasing region of oxygen depletion at the bioreactor outlet resulting in a decreased cell density at the outlet in the traditional packed bed

Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor 101

bioreactor (Figure 27 B & E). At the higher flowrates required for the scale up version, the lactate is removed at sufficiently rapid rate to ensure that its concentration does not reach a level to cause a significant deleterious effect (Appendix 12).

To counteract this oxygen depletion and allow for lower flow rates to be used, the model was extended to include additional oxygen sources in the form of capillaries running through the bioreactor. Due to the limitations inherent in the axial symmetric model only one capillary could be added at the centre of the reactor. This is because axial symmetric model gives a 3D approximation of the 2D model by the rotation around the central axis of the rectangular geometry. The additions of a capillary a quarter of the way through the radius will actual result in an annulus, which is not an accurate representation and will require a full 3D model to do this accurately. The geometry is outlined in (Figure 17 D). The lumen radius, Rc is 0.02 cm with the capillary wall outer radius wall, Rw is 0.12 cm, the vessel radius R1 is 2.62 cm and the outer wall R2 is 3.12 cm, the reactor length L=12 cm. We assumed that the capillary wall is composed of polymer with the same properties as PDMS. The boundary conditions for the oxygen transport within the vessel needed to be modified because of oxygen diffusion through the central wall. Therefore, the concentration of oxygen within the capillary wall (CO,C) is given by

휕퐶푂,퐶 2 = 퐷 ∇ 퐶 for Rc

The same boundary conditions are applied at the outer vessel wall and for the capillary. There was zero flux of oxygen through the inlet and the outlet.

휕퐶푂,퐵 = 0 at z=0, L, Rc

It was assumed that that there is constant oxygen concentration at the capillary lumen wall

퐶푂,퐶 = 푆푃퐷푀푆 at r=Rc, 0

102 Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor

At the interface between the capillary wall and the medium, the boundary condition equation (23) is now replaced by

휕퐶푂,퐴 휕퐶푂,퐶 퐷 = 퐷 at r=Rw, 0

퐶푂,퐴 퐶푂,퐶 = at r=Rw, 0

This design change to the bioreactor model results in superior cell expansion at lower flow rates allowing the medium to be perfused at 10ml/hr resulting in a similar growth rate to the base model (Figure 26 A). There is no depletion of oxygen at the outlet of the bioreactor (Figure 27 F). However, the oxygen concentration is 10 fold less than the Michaelis-Menten oxygen growth constant KO. These results in a slower growth rate in at r= 1.25 cm in the bioreactor (Figure 27 C). Nevertheless, this concept warrants further investigation in the future. Further modelling analyses could be carried out using a three dimensional geometric models to allow for the addition of more oxygen providing capillaries the optimal since distance between capillaries needs to be determined.

Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor 103

A D

B E

C F

Figure 27 Cell density heat map of bioreactor with (A) oxygen provided by diffusion through the wall, (B) no wall diffusion and (C) oxygen provided by both diffusion through the wall and central capillary at day 7 at a flow rate 15 ml/hr. Oxygen concentration with (D) oxygen provided by diffusion through the wall, (E) no wall diffusion and (F) oxygen provided by both diffusion through the wall and central capillary at day 7.

A decrease in the particle size of the pellets further increases the available surface area. By decreasing the cylindrical pellet size to 1mm diameter by 1mm in length the reactor surface area increases to 7,493 cm2 which is enough to support 1.5x108 MSCs. By further decreasing the cylinder pellet size to 0.5 mm diameter by 0.5 mm in length the surface area increases to 14,985 cm2 which can support 3x108

104 Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor

MSCs. The probable improvement in yield can be explored by modifying the current model.

The simulations show that the average cell density is significantly less with both smaller pellet bioreactor designs with or without wall diffusion, unless the flow rate is increased to 50 ml/hr for the 1mm by 1mm pellets when the average cell density reaches 19000 cells/cm2 (Figure 28 A & B). Both designs with or without wall diffusion failed to support a pellet size of 0.5mm by 0.5mm because of oxygen depletion unless there is a significant increase in flow rate. The increased flow rate is possible as the shear stress acting is still below the threshold 0.015 Pa even with a flow rate of 90ml/hr (Figure 29 C). This high flow rate requires an increase in medium with consequent increased cost. A decrease in flow rate requires adding capillaries to provide additional oxygen sources to allow sufficient growth rate (Figure 28C). To further decrease the flow rate would require additional oxygen sources beyond the single central capillary.

Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor 105

A 20,000 18,000 16,000 14,000 12,000 10,000 8,000 6,000

Cell density (cells/cm2) density Cell 4,000 2,000 0 0 2 4 6 8 Day B 20,000 18,000 16,000 14,000 12,000 10,000 8,000 6,000

Cell density (cells/cm2) density Cell 4,000 2,000 0 0 2 4 6 8 Day C 20,000 18,000 16,000 14,000 12,000 10,000 8,000 6,000

4,000 Cell density (cells/cm2) density Cell 2,000 0 0 2 4 6 8 Day

106 Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor

25000 2.5 cm R by 12 cm L bioreactor

20000 1 cm R by 75 cm L bioreactor

0.75 cm R by 133.33 cm L bioreactor 15000 8.66 cm R by 1 cm L bioreactor

10000 6.12 cm R by 2 cm L bioreactor Cell density (cells/cm2) density Cell 5000

0 0 1 2 3 4 5 6 7 8 Day

B 25000 0.75 by 133 cm bioreactor with 0.5 by 0.5 pellets 30 ml/hr

20000 0.75 by 133 cm bioreactor with 0.5 by 0.5 pellets 20 ml/hr 15000 0.75 by 133 cm bioreactor with 0.5 by 0.5 pellets 15 ml/hr 10000

0.75 by 133 cm bioreactor with 0.5 by Cell density (cells/cm2) density Cell 5000 0.5 pellets 10 ml/hr

0 0 1 2 3 4 5 6 7 8 Day C D 2.5 cm R by 12 cm L 0.007 0.03 bioreactor 30 ml/hr 2.5 mm R by 3 mm L 0.006 pellets 30 ml/hr 8.66 cm R by 1 cm L 0.025 bioreactor 30 ml/hr 1 mm R by 1 mm L 0.005 pellets 30 ml/hr 6.12 cm R by 2 cm L Threshold shear bioreactor 30 ml/hr ) 1 mm R by 1 mm L 0.02 stress that

(Pa) 1 cm R by 75 cm L 0.004 pellets 40 ml/hr induces osteogenesis. bioreactor 30 ml/hr

1 mm R by 1 mm L 0.015 Stress (Pa Stress pellets 50 ml/hr Stress 0.75 cm R by 133.33 cm L 0.003 bioreactor 30 ml/hr

0.5 mm R by 0.5 mm L Shear Shear pellets 30 ml/hr Shear 0.01 0.75 cm R by 133.33 cm L 0.002 bioreactor 20 ml/hr 0.5 mm R by 0.5 mm L pellets 60 ml/hr 0.005 0.75 cm R by 133.33 cm L 0.001 bioreactor 15 ml/hr 0.5 mm R by 0.5 mm L pellets 90 ml/hr 0.75 cm R by 133.33 cm L 0 0 bioreactor 10 ml/hr

Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor 107

Another way to reduce the flow rate without increasing the number of capillaries providing oxygen into the reactor is by changing the reactor shape. To explore this alternative, the model reactor dimensions were changed while the reactor volume remained constant. Figure 29 A shows the average cell density bioreactor with oxygen supplied through the wall and centre maintaining a reactor volume of 235.6 cm3 filled with 0.5 mm by 0.5mm cylindrical pellets changing the geometry of the reactor by adjusting the radius, R1 and the length, L. While still maintaining a medium perfusion rate at 30 ml/hr. It is clear that the long thin bioreactor performs the best, while the short fat reactor requires a higher flow rate to achieve the same growth rate. This is because oxygen is able to diffuse through to the centre of the long, thin vessel. This suggests that the maximum distance between oxygen sources should be between 0.75 cm to 1 cm. In this way the flow rate can to be reduced without a significant effect on the cell growth rate (Figure 29 B). And since the shear stresses for the long thin reactor are above the threshold of 0.015 Pa if the flow rate is above 15 ml/hr this provides a suitable alternative reactor design (Figure 29 D).

The modelling of this bioreactor type supports that this packed bed bioreactor design is superior over a traditional packed bed device, allowing for a reduced medium flow rate to be used, thus decreasing the cost of the cell expansion. Furthermore, the model gave further insight into future development and scale-up of the bioreactor design with the addition of diffusive oxygen sources. Moreover, the model provided an insight into the experimental results by highlighting the importance of the initial cell distribution on the final cell number achievable within the bioreactor.

108 Chapter 4: Explanation of Experimental Results, and Further Design and Development: Modelling Mesenchymal Stromal Cell Growth in a Packed Bed Bioreactor

Chapter 5: Summary of Findings and Future Directions

For MSC based therapy to become routinely used in the clinic, the manufacture of the MSCs needs to become more efficient by replacing manual culture methods with an automated bioreactor system [224]. Bioreactors must have a large surface area for cell attachment and be optimised for the mass transport of nutrients and the elimination of waste [227].

To provide the large surface area for MSC expansion a packed bed bioreactor design was selected over other high surface area devices that use microcarriers such as stirred tank and wave bag bioreactors. As these devices have no connectivity between individual micro-carriers, and empty micro-carriers do not contribute to the total surface area available to the culture. As a result, some micro-carriers rapidly reach confluence, whilst others remain empty; this requires frequent passaging to overcome localized space limitations [231,264]. Serum coated glass beads were first used as a packing material as they allowed for easier sterilisation of the device. However, the cell attachment to the beads was very weak, and the cells were easily delaminated. Thus, the scaffold material was changed to 2.5 mm radius by 3 mm in length cylindrical polystyrene pellets. This was due to material availability, ideally smaller 500 µm spherical particles would have been better suited to be the bioreactor scaffold. As spheres have the largest surface area to volume ratio compared to cylinders also a smaller size gives a larger surface area in the same reactor volume. However, additional issue observed in the experiments using glass packing was the beads move during the seeding process could damage the cells. To prevent this the polystyrene cylinders were fused together to give one continuous scaffold that could easily be scaled up from 1.5 cm diameter by 7.5 cm length for the small scale bioreactor, providing 160cm2 of surface area to a 5 cm diameter by 12 cm length providing 2830 cm2 of surface. This larger bioreactor size successfully expanded 1.12x108 GFP- mMSCs and theoretically this could support up to 5.6x107 human MSCs.

Chapter 5: Summary of Findings and Future Directions 109

Due to the hydrophobic nature and exhibit low surface energy of polystyrene, the scaffold surface needed to be modified to allow cell attachment spread and proliferation [263,266-268]. Rotary air plasma treatment was selected method to modify the scaffold surface as it provides similar growth surface chemistry to commercial TCP [265], which was observed experimentally Figure 7. In addition, this surface modification is scalable and is suitable to modify scaffold and the small particle size necessary for increasing the bioreactor surface area [269,274,287,288]. Air plasma surface was selected as surface characteristics are known to influence cell growth and behaviour or exhibit toxicity [260,349-353]. So that any experimental abnormalities observed in the bioreactor would be independent of the surface chemistry. See Appendix 1for the importance of testing modified surfaces for their effect on cell culture.

Many other surface modifications are available using polymer coats that add more functionality to the bioreactor scaffold. Polymers such as poly-(N-isopropyl acrylamide) (pNIPAM), which changes the surface charge as a function of temperature allowing for non-enzymatic detachment of cells would be ideal candidate for the next evolution of the packed bed bioreactor design [354,355]. As this surface modification allows for better recovery of the cells as the process is less harsh [354,355]. A modification like this will enable better cell redistribution partway through the isolation and expansion process to redistribute the cell colonies to a more homogeneous distribution, which will greatly improve the final cell numbers harvest from the bioreactor. In addition to plasma polymerisation onto surfaces, by changing the surface topology, charge, chemical composition or hydrophilicity, stem cell fate can be directed to help maintain MSC phenotype and multipotency through physical actions through the interaction with Nano patterns or functional moieties can be subsequently used to immobilise proteins or other biomolecules to the surface once the plasma treated [288,290,353,354,356-361].

The other problem that a bioreactor design needs to take into account is the limiting metabolite oxygen due to its low solubility in the culture medium. Traditional bioreactors require high flow and mixing rates to provide an adequate oxygen supply. However, for MSC expansion, one of the main known limitations of perfusion

110 Chapter 5: Summary of Findings and Future Directions

bioreactors is shear stress, as a force of 0.015 Pa has been shown to up-regulate the osteogenic pathways in human bmMSCs [218-221]. This shear stress threshold greatly limits the scalability of bioreactor devices that provide oxygen through the bulk media flow. To counteract this problem, the packed bed bioreactor design incorporated gas permeable wall made from gas permeable polymer PDMS that allows gas to diffuse radially into the bioreactor that enables the decoupling of oxygen from the bulk nutrient supply. This allowed a lower medium flow rate to be used in the experiments than has been previously been reported [226,228,231,295]. Through the development of 2D axial symmetric model in COMSOL in Chapter 4, showed that the flow rates used in the bioreactor experiments were below the shear stress threshold of 0.015 Pa, and that oxygen diffusion through the wall was sufficient to support MSC growth even when the bioreactor was scaled up. In addition, the model supported the hypothesis that decoupling the oxygen supply through the bulk nutrient flow by allowing oxygen to diffuse through the wall allows a lower perfusion rate. When a traditional packed bed bioreactor (without a gas permeable wall) was modelled, a 5 times higher flow rate of the medium was required to achieve a similar growth rate to our packed bed bioreactor. This difference represents a significant potential cost saving in the manufacturing process, especially if the medium is changed to a more expensive fully- defined xeno-free medium.

To further develop the bioreactor design, the scaled-up 2830cm2 bioreactor was modelled. The model did show a depletion of oxygen in the centre of the large scale bioreactor at lower flow rates, suggesting that oxygen diffusion through the wall alone could not keep up with the cell consumption at larger scale. To mitigate this, an additional diffusive source of oxygen was added to the bioreactor model in the form of a gas permeable capillary running through the centre of the bioreactor. The model predicts that this addition prevents oxygen depletion in the bioreactor. Furthermore, by scaling-up the bioreactor by changing the geometry of the bioreactor and decreasing the particle size, sufficient oxygen was provided by diffusion through the wall and the central capillary. A bioreactor modelled with a volume of 235.6 cm3 was able to support a cell number of 3x108 cells. Although a 2D axisymmetric model provided insight to the observed experimental results and helped to direct the design by demonstrating that additional diffusive oxygen sources can provide adequate oxygen

Chapter 5: Summary of Findings and Future Directions 111

to the cells in the scaffold. A full 3D model is required to find the optimal geometry and placement of the oxygen delivery capillaries that will eliminate oxygen gradients in the bioreactor at high cell densities. This will potentially help to design a higher surface area bioreactor device that will allow 1010 MSCs to be manufactured, which is enough for approximately 100 doses. Limitations in the current model included the assumption that the dissolution of oxygen was instantaneous and that the concentration of oxygen at the wall-air interface was constant. The rate of oxygen diffusion from the air to the wall and from the wall to the medium will greatly affect the penetration of the oxygen into the bioreactor and the amount of oxygen available to the cells. Any new model must include the rates of dissolution of oxygen in the PDMS wall and into the medium, which is described by the Noyes-Whitney equation [362]. Ultimately, this will likely require additional oxygen capillaries due to an increase in time for the oxygen to diffuse through the membrane and will not be able to keep up with the rate of consumption especially at high cell densities. However, the experiments and modelling done in the thesis have been performed in normoxia. Hypoxic conditions have been reported to support MSC expansion and to maintain their stemness [200- 203,223]. Under hypoxic conditions, the highest oxygen concentration available would be 0.05 mM. Hypoxia would be very difficult to achieve while maintaining a shear stress below the 0.015 Pa threshold in a traditional packed bed bioreactor without a decoupled oxygen supply. It is essential that future scale-up experiments of the isolation and expansion of pMSCs will have to be performed in hypoxia in a packed bed bioreactor designed directed by 3D model incorporating many capillaries to provide additional diffusive sources of oxygen in addition to wall source, in order to expand MSCs in their preferred environment.

Although adequate nutrients were supplied to the cells in the bioreactor, the model indicated that lactate impacted the cell growth in the small scale bioreactor. This highlighted the importance of waste product removal in low flow rate devices, especially in micro devices were very low flow rates are used at very high cell densities. An additional problem with the lower flow rate used in the bioreactor with the removal of the non-attached cells and debris that may have induced osteogenic differentiation [317] that was observed in the small scale experiments in the isolation and expansion of pMSCs in Chapter 3. Through the addition of process control will

112 Chapter 5: Summary of Findings and Future Directions

allow for a better process to remove the cell debris from the bioreactor during the isolation process. As this process was done manually by a syringe which resulted in discontinuous flow and low flow rate, due to no controls in place to prevent detaching the MSCs. Automation of this process will allow the flow rate to be controlled so that it will be under the 0.015 Pa shear stress threshold and the flow will be continuous to better wash away the cell debris to prevent the osteogenic differentiation. The addition of temperature, pH, glucose, oxygen and flow rate probes would allow further control to of the environment to help maintain the required MSC phenotype. Furthermore, automation of the isolation of the pMSCs from placenta is possible which allows for the construction of a fully closed system bioreactor device for the isolation and expansion of pMSCs [210]. Due to limitations the regulation of medical devices, which stipulates that all components of such a device requires approval by governmental regulatory authority, it would be prudent to adapt the bioreactor to be incorporated into approved bioreactor control system and circuitry that is currently on the market, such as Quatum® Cell Expansion System (TerumoBCT) or FlexAct® (Sartorius stedim biotech). This will greatly reduce the registration cost and bring this packed bed bioreactor quicker to the market.

The packed bed bioreactor system developed was able to achieve a growth rate for human pMSCs growing similar to that reported in the literature for other perfusion bioreactor devices [226,228,231,242,295]. In addition, we observed no dramatic difference in MSCs’ ability to differentiate after being expanded in the bioreactor compared to traditional flasks cultures. However, based on observations in both scales of bioreactor and with two cells types (GFP-mMSCs (Chapter 2) and human pMSCs (Chapter 3)), the growth rate of the cells expanding in the bioreactor was observed to be significantly less than in traditional flask cultures. This suggests that the reduced growth observed in our bioreactor is a common phenomenon among perfusion bioreactors and are unlikely to be due to the surface chemistry or culture conditions that are unique to our packed bed bioreactor. The finite element modelling in COMSOL showed that the reduced growth rate might be caused by a non-uniform cell distribution, most likely in the axial direction, with a cell lag phase of between 24 to 48 hours due to the handling of the cells during the seeding method. This model output highlights the importance of achieving homogeneous cell distribution, while limiting

Chapter 5: Summary of Findings and Future Directions 113

the stress on the cells during the seeding process. This was further supported when the bioreactor was modelled under static conditions, as it suggested that the MSCs were sensitive to manual manipulation. The reduced growth rate observed in the MSCs grown under static compared to perfusion conditions were predicted by a model that included cell death during the medium exchanges. With the higher flow rates used in the scaled-up bioreactor, the lactate concentration did not cause the decrease in the growth rate that had previously been observed in the smaller reactor. Modelling could help in directing the development of a better method of cell seeding through use of particle settling and cell adhesion kinetics [363,364]. However, given the number of seeding methods currently in the literature, a homogeneous cell distribution is probably not achievable, and this should be considered the trade off when expanding cells in a bioreactor [239,296,297,300,365,366].

pMSCs were selected over the more traditional source, bmMSCs, as a source of allogeneic MSCs to expand as a single placenta (500-700 g tissue) is sufficient for manufacturing several thousand units of allogeneic MSCs once expanded [231]. Furthermore, placental tissue can be harvested aseptically during a Caesarean section with no added risk to the mother or child. Although the ability for pMSCs to undergo trilineage mesoderm differentiation was significantly less than bmMSCs, the ability for both cell types conditioned medium to promote HaCaT growth was the equivalent. No correlation was observed between passage number and the MSCs’ ability to promote HaCaT growth, but this was largely due to data inconstancy. The conditioned medium derived from MSC aggregates was more consistent in promoting HaCaT growth, suggesting that modifying the cell culture to promote a specific phenotype may result in better and more consistent outcomes [7,203,367].

The difficulty in promoting HaCaT growth with MSC conditioned medium was the reproducibility, as no consistent result could be achieved. This was frustrating since the literature was showing that MSCs could support cell growth and promote tissue regeneration in animals and in lower level clinical trials. However, higher level clinical results have been published recently showing mixed results overall; there is no significant improvement when treating a wide range of conditions with MSC therapy [22,56,74,140,368]. In the case of ischemic heart repair there was no difference in

114 Chapter 5: Summary of Findings and Future Directions

recovery between MSC treatment and placebo in clinical outcomes [129,369]. By reviewing our understanding of what defines a MSC and determining the function and mechanism of action, we can see that there are very large knowledge gaps. By far the largest deficiency is that there is no set of definitive marker. Where even selecting a more robust set of markers such as STRO-1, CD49a, CD105, CD133, CD146, CD271, SSEA-4 it is still no better than just relying on tissue culture plastic adherence to isolating a multipotent self-renewal MSCs [44,48,56,74,370]. This means that we are still dealing with a heterogeneous cell population that can change during culture [44,158,370]. This may explain the inconsistent results observed in the work presented here and the failure to achieve significant clinical outcomes. Also it has been assumed that MSCs are immune privileged, however they do activate the innate immune system meaning that intravenous injections of MSCs will be killed within 24 hours especially at high dosages [75,94]. In addition, when MSCs start to differentiate, they begin to express MHC receptors meaning that the use of allogenic MSCs for bone or cartilage repair is not ideal [371,372]. Most of the characterisation of MSCs have been done in vitro, which gives a false impression of their biological function, given that even fibroblasts can undergo trilineage mesodermal differentiation under the same protocol [56,372,373]. Many clinical applications are reliant on the “injury drug store” function of MSCs[8]. However, one cannot hope that the injected MSCs would respond to an injured or inflamed environment by secreting the right kind of cytokines as well as modulating the immune system; this suggests that this approach is neither evidence based nor scientific. Intravenous injected MSCs are embolised in the lungs which results in endothelial damage due to the activation of the innate immune system. As previously stated the paracrine effect that is said to happen, must occur in a matter of hours, and it could even be related to cell death processes [56,105,197]. It is also noted that neural repair cannot happen without the presence of the peripheral immune system [374]. This all indicates that the tissue repair mechanism has more to do with activating the immune system than any direct regeneration by MSCs. Knowing the true mechanism of action of how MSCs respond in such an environment and discovering the true function of MSCs is paramount.

Although this thesis has focused on expanding MSCs in the packed bed bioreactor, this bioreactor can be used for expansion or production of other cell types

Chapter 5: Summary of Findings and Future Directions 115

such as platelets, or the production of protein, anti-bodies or other biological product production that requires adherent cells [375]. As MSCs conditioned medium can potentially promote tissue repair, the bioreactor can be used as a “conditioned medium factory”, as the large surface area and low rates our bioreactor can achieve allows for a more concentrated product requiring less downstream manipulation. The packed bed bioreactor could be used as a platform for an artificial liver [376-378]. With the advantage of an independent oxygen supply, the culture environment is better controlled with our device, meaning that this device is superior to the traditional packed bed bioreactor.

In summation this thesis demonstrates that a packed bed bioreactor that has a gas permeable wall which allows for the decoupling of oxygen from the bulk nutrient flow can be used to isolate and expand pMSCs. This bioreactor is scalable, and finite element modelling of the bioreactor provided further direction to scale-up and improve future bioreactor designs. The model highlighted the most important factor in the cell growth in the bioreactor is the seeding distribution. Although the growth rate of pMSCs in the bioreactor was less than the cells grown in traditional on tissue culture flasks, the few extra days required for MSCs to reach confluence in the packed bed bioreactor represents an acceptable trade-off to have the potential for an automatable method to digest, isolate and expand pMSCs in a single closed bioreactor system.

116 Chapter 5: Summary of Findings and Future Directions

References

1. Mundra V, Gerling IC, Mahato RI (2013) Mesenchymal stem cell-based therapy. Mol Pharm 10: 77-89. 2. Sensebe L, Krampera M, Schrezenmeier H, Bourin P, Giordano R (2010) Mesenchymal stem cells for clinical application. Vox Sang 98: 93-107. 3. Wang J, Liao L, Tan J (2011) Mesenchymal-stem-cell-based experimental and clinical trials: current status and open questions. Expert Opinion on Biological Therapy 11: 893-909. 4. Kidd S, Spaeth E, Dembinski JL, Dietrich M, Watson K, et al. (2009) Direct Evidence of Mesenchymal Stem Cell Tropism for Tumor and Wounding Microenvironments Using In Vivo Bioluminescent Imaging. STEM CELLS 27: 2614-2623. 5. Sasaki M, Abe R, Fujita Y, Ando S, Inokuma D, et al. (2008) Mesenchymal Stem Cells Are Recruited into Wounded Skin and Contribute to Wound Repair by Transdifferentiation into Multiple Skin Cell Type. The Journal of Immunology 180: 2581-2587. 6. Wise AF, Williams TM, Kiewiet MBG, Payne NL, Siatskas C, et al. (2014) Human mesenchymal stem cells alter macrophage phenotype and promote regeneration via homing to the kidney following ischemia-reperfusion injury; Wise AF, Williams TM, Kiewiet MBG, Payne NL, Siatskas C et al., editors: American Journal of Physiology - Renal Physiology. F1222-F1235 p. 7. Doorn J, Moll G, Le Blanc K, van Blitterswijk C, de Boer J (2012) Therapeutic applications of mesenchymal stromal cells: paracrine effects and potential improvements. Tissue Eng Part B Rev 18: 101-115. 8. Caplan Arnold I, Correa D (2011) The MSC: An Injury Drugstore. Cell Stem Cell 9: 11-15. 9. Kode JA, Mukherjee S, Joglekar MV, Hardikar AA (2009) Mesenchymal stem cells: immunobiology and role in immunomodulation and tissue regeneration. Cytotherapy 11: 377-391. 10. Amati E, Di Trapani M, Liotta F, Annunziato F, Perbellini O, et al. (2013) In Vitro Study Of The Mechanisms Involved In The Bone Marrow Mesenchymal Stromal Cell Modulatory Effect On B Cell Function. Blood 122: 1053. 11. English K (2013) Mechanisms of mesenchymal stromal cell immunomodulation. Immunol Cell Biol 91: 19-26. 12. Lamirault G, De Bock E, Roncalli J, Van Belle E, Piot C, et al. (2013) Sustained quality of life improvement after intracoronary injection of autologous bone marrow cells in the setting of acute myocardial infarction. European Heart Journal 34. 13. Rodrigo S, van Ramshorst J, Beeres S, Younis IA, Dibbets-Schneider P, et al. (2012) Intramyocardial Injection of Bone Marrow Derived, Ex Vivo Expanded, Mesenchymal Stem Cells in Acute Myocardial Infartion Patients is Safe and Improves Myocardial Perfusion at 3 Years Follow up. Journal of the American College of Cardiology 59: E478-E478. 14. Flynn A, O'Brien T (2011) Stem cell therapy for cardiac disease. Expert Opin Biol Ther 11: 177-187.

References 117

15. Blumberg SN, Berger A, Hwang L, Pastar I, Warren SM, et al. (2012) The role of stem cells in the treatment of diabetic foot ulcers. Diabetes Research and Clinical Practice 96: 1-9. 16. Vojtassak J, Danisovic L, Kubes M, Bakos D, Jarabek L, et al. (2006) Autologous biograft and mesenchymal stem cells in treatment of the diabetic foot. Neuro Endocrinol Lett 27 Suppl 2: 134-137. 17. Panetta NJ, Gupta DM, Quarto N, Longaker MT (2009) Mesenchymal cells for skeletal tissue engineering. Panminerva Med 51: 25-41. 18. Orozco L, Soler R, Morera C, Alberca M, Sánchez A, et al. (2011) Intervertebral Disc Repair by Autologous Mesenchymal Bone Marrow Cells: A Pilot Study. Transplantation 92: 822-828 810.1097/TP.1090b1013e3182298a3182215. 19. Dawson JI, Kanczler J, Tare R, Kassem M, Oreffo ROC (2014) Concise Review: Bridging the Gap: Bone Regeneration Using Skeletal Stem Cell-Based Strategies—Where Are We Now? STEM CELLS 32: 35-44. 20. Griffin MD, Elliman SJ, Cahill E, English K, Ceredig R, et al. (2013) Concise review: Adult mesenchymal stromal cell therapy for inflammatory diseases: How well are we joining the dots? STEM CELLS 31: 2033-2041. 21. Kaplan JM, Youd ME, Lodie TA (2011) Immunomodulatory activity of mesenchymal stem cells. Curr Stem Cell Res Ther 6: 297-316. 22. Kramer M, Wermke M, Röllig C, Alakel N, Stölzel F, et al. (2013) Response and Long-Term Outcome After Treatment With Third-Party Mesenchymal Stromal Cells - Updated Results In 58 Patients With Steroid-Refractory Acute Graft- Versus Host Disease. Blood 122: 4612. 23. Charbord P (2010) Bone marrow mesenchymal stem cells: historical overview and concepts. Hum Gene Ther 21: 1045-1056. 24. Mosna F, Sensebe L, Krampera M (2010) Human bone marrow and adipose tissue mesenchymal stem cells: a user's guide. Stem Cells Dev 19: 1449-1470. 25. Jager M, Zilkens C, Bittersohl B, Krauspe R (2009) Cord blood--an alternative source for bone regeneration. Stem Cell Rev 5: 266-277. 26. Barlow S, Brooke G, Chatterjee K, Price G, Pelekanos R, et al. (2008) Comparison of human placenta- and bone marrow-derived multipotent mesenchymal stem cells. Stem Cells Dev 17: 1095-1107. 27. Neofytou E, Deuse T, Beygui RE, Schrepfer S (2015) Mesenchymal Stromal Cell Therapy: Different Sources Exhibit Different Immunobiological Properties. Transplantation 99: 1113-1118. 28. Williams AR, Hare JM (2011) Mesenchymal Stem Cells: Biology, Pathophysiology, Translational Findings, and Therapeutic Implications for Cardiac Disease. Circulation Research 109: 923-940. 29. Lataillade JJ, Doucet C, Bey E, Carsin H, Huet C, et al. (2007) New approach to radiation burn treatment by dosimetry-guided surgery combined with autologous mesenchymal stem cell therapy. Regen Med 2: 785-794. 30. Hanley PJ, Mei Z, da Graca Cabreira-Hansen M, Klis M, Li W, et al. (2013) Manufacturing mesenchymal stromal cells for phase I clinical trials. Cytotherapy 15: 416-422. 31. Cierpka K, Elseberg CL, Niss K, Kassem M, Salzig D, et al. (2013) hMSC Production in Disposable Bioreactors with Regards to GMP and PAT. Chemie Ingenieur Technik 85: 67-75. 32. Grassel S, Ahmed N (2007) Influence of cellular microenvironment and paracrine signals on chondrogenic differentiation. Front Biosci 12: 4946-4956.

118 References

33. Wang J, Liao L, Tan J (2011) Mesenchymal-stem-cell-based experimental and clinical trials: current status and open questions. Expert Opin Biol Ther 11: 893-909. 34. Beyer Nardi N, Silva Meirelles L (2006) Mesenchymal Stem Cells: Isolation, In Vitro Expansion and Characterization Stem Cells. In: Wobus A, Boheler K, editors: Springer Berlin Heidelberg. pp. 249-282. 35. Prockop DJ (1997) Marrow Stromal Cells as Stem Cells for Nonhematopoietic Tissues. Science 276: 71-74. 36. Sacchetti B, Funari A, Michienzi S, Di Cesare S, Piersanti S, et al. (2007) Self- renewing osteoprogenitors in bone marrow sinusoids can organize a hematopoietic microenvironment. Cell 131: 324-336. 37. Mendez-Ferrer S, Michurina TV, Ferraro F, Mazloom AR, MacArthur BD, et al. (2010) Mesenchymal and haematopoietic stem cells form a unique bone marrow niche. Nature 466: 829-834. 38. Crisan M, Yap S, Casteilla L, Chen CW, Corselli M, et al. (2008) A perivascular origin for mesenchymal stem cells in multiple human organs. Cell Stem Cell 3: 301-313. 39. Caplan AI (2008) All MSCs Are Pericytes? Cell Stem Cell 3: 229-230. 40. Dominici M, Le Blanc K, Mueller I, Slaper-Cortenbach I, Marini F, et al. (2006) Minimal criteria for defining multipotent mesenchymal stromal cells. The International Society for Cellular Therapy position statement. Cytotherapy 8: 315-317. 41. Russell KC, Phinney DG, Lacey MR, Barrilleaux BL, Meyertholen KE, et al. (2010) In vitro high-capacity assay to quantify the clonal heterogeneity in trilineage potential of mesenchymal stem cells reveals a complex hierarchy of lineage commitment. Stem Cells 28: 788-798. 42. Mendez-Ferrer S, Michurina TV, Ferraro F, Mazloom AR, Macarthur BD, et al. (2010) Mesenchymal and haematopoietic stem cells form a unique bone marrow niche. Nature 466: 829-834. 43. Lv FJ, Tuan RS, Cheung KM, Leung VY (2014) Concise review: the surface markers and identity of human mesenchymal stem cells. Stem Cells 32: 1408- 1419. 44. James S, Fox J, Afsari F, Lee J, Clough S, et al. (2015) Multiparameter Analysis of Human Bone Marrow Stromal Cells Identifies Distinct Immunomodulatory and Differentiation-Competent Subtypes. Stem Cell Reports 4: 1004-1015. 45. Pittenger MF, Mackay AM, Beck SC, Jaiswal RK, Douglas R, et al. (1999) Multilineage Potential of Adult Human Mesenchymal Stem Cells. Science 284: 143-147. 46. Buhring HJ, Battula VL, Treml S, Schewe B, Kanz L, et al. (2007) Novel markers for the prospective isolation of human MSC. Ann N Y Acad Sci 1106: 262- 271. 47. Meirelles LdS, Chagastelles PC, Nardi NB (2006) Mesenchymal stem cells reside in virtually all post-natal organs and tissues. Journal of Cell Science 119: 2204- 2213. 48. Bianco P (2011) Back to the future: moving beyond "mesenchymal stem cells". J Cell Biochem 112: 1713-1721. 49. Fazekasova H, Lechler R, Langford K, Lombardi G (2011) Placenta-derived MSCs are partially immunogenic and less immunomodulatory than bone marrow- derived MSCs. J Tissue Eng Regen Med 5: 684-694.

References 119

50. Heazlewood CF, Sherrell H, Ryan J, Atkinson K, Wells CA, et al. (2014) High Incidence of Contaminating Maternal Cell Overgrowth in Human Placental Mesenchymal Stem/Stromal Cell Cultures: A Systematic Review. Stem Cells Translational Medicine. 51. Dellavalle A, Sampaolesi M, Tonlorenzi R, Tagliafico E, Sacchetti B, et al. (2007) Pericytes of human skeletal muscle are myogenic precursors distinct from satellite cells. Nat Cell Biol 9: 255-267. 52. Gronthos S, Brahim J, Li W, Fisher LW, Cherman N, et al. (2002) Stem cell properties of human dental pulp stem cells. J Dent Res 81: 531-535. 53. Bianco P, Kuznetsov SA, Riminucci M, Gehron Robey P (2006) Postnatal skeletal stem cells. Methods Enzymol 419: 117-148. 54. Bianco P, Robey PG, Simmons PJ (2008) Mesenchymal Stem Cells: Revisiting History, Concepts, and Assays. Cell stem cell 2: 313-319. 55. Trounson A, Thakar RG, Lomax G, Gibbons D (2011) Clinical trials for stem cell therapies. BMC Med 9: 52. 56. Bianco P, Cao X, Frenette PS, Mao JJ, Robey PG, et al. (2013) The meaning, the sense and the significance: translating the science of mesenchymal stem cells into medicine. Nat Med 19: 35-42. 57. El Haddad N, Heathcote D, Moore R, Yang S, Azzi J, et al. (2011) Mesenchymal stem cells express serine protease inhibitor to evade the host immune response. Blood 117: 1176-1183. 58. Griffin MD, Ritter T, Mahon BP (2010) Immunological aspects of allogeneic mesenchymal stem cell therapies. Hum Gene Ther 21: 1641-1655. 59. Cho YM, Kim JH, Kim M, Park SJ, Koh SH, et al. (2012) Mesenchymal Stem Cells Transfer Mitochondria to the Cells with Virtually No Mitochondrial Function but Not with Pathogenic mtDNA Mutations. PLoS One 7: e32778. 60. Islam MN, Das SR, Emin MT, Wei M, Sun L, et al. (2012) Mitochondrial transfer from bone-marrow-derived stromal cells to pulmonary alveoli protects against acute lung injury. Nat Med 18: 759-765. 61. Plotnikov EY, Khryapenkova TG, Vasileva AK, Marey MV, Galkina SI, et al. (2008) Cell-to-cell cross-talk between mesenchymal stem cells and cardiomyocytes in co-culture. Journal of Cellular and Molecular Medicine 12: 1622-1631. 62. Salem HK, Thiemermann C (2010) Mesenchymal stromal cells: current understanding and clinical status. Stem Cells 28: 585-596. 63. Marie P (2011) Cell and gene therapy for bone repair. Osteoporosis International 22: 2023-2026. 64. Keeney M, Lai JH, Yang F (2011) Recent progress in cartilage tissue engineering. Current Opinion in Biotechnology 22: 734-740. 65. Li J, Tao R, Wu W, Cao H, Xin J, et al. (2010) 3D PLGA scaffolds improve differentiation and function of bone marrow mesenchymal stem cell-derived hepatocytes. Stem Cells Dev 19: 1427-1436. 66. Wright KT, Masri WE, Osman A, Chowdhury J, Johnson WEB (2011) Concise Review: Bone Marrow for the Treatment of Spinal Cord Injury: Mechanisms and Clinical Applications. Stem Cells 29: 169-178. 67. Chao KC, Chao KF, Fu YS, Liu SH (2008) Islet-like clusters derived from mesenchymal stem cells in Wharton's Jelly of the human umbilical cord for transplantation to control type 1 diabetes. PLoS One 3: e1451. 68. De Napoli IE, Scaglione S, Giannoni P, Quarto R, Catapano G (2011) Mesenchymal stem cell culture in convection-enhanced hollow fibre

120 References

membrane bioreactors for bone tissue engineering. Journal of Membrane Science 379: 341-352. 69. Neuhuber B, Gallo G, Howard L, Kostura L, Mackay A, et al. (2004) Reevaluation of in vitro differentiation protocols for bone marrow stromal cells: Disruption of actin cytoskeleton induces rapid morphological changes and mimics neuronal phenotype. Journal of Neuroscience Research 77: 192-204. 70. Tsuji H, Miyoshi S, Ikegami Y, Hida N, Asada H, et al. (2010) Xenografted Human Amniotic Membrane–Derived Mesenchymal Stem Cells Are Immunologically Tolerated and Transdifferentiated Into Cardiomyocytes. Circulation Research 106: 1613-1623. 71. Nie C, Yang D, Xu J, Si Z, Jin X, et al. (2011) Locally administered adipose- derived stem cells accelerate wound healing through differentiation and vasculogenesis. Cell Transplant 20: 205-216. 72. Wu Y, Chen L, Scott PG, Tredget EE (2007) Mesenchymal stem cells enhance wound healing through differentiation and angiogenesis. Stem Cells 25: 2648- 2659. 73. Coyne TM, Marcus AJ, Woodbury D, Black IB (2006) Marrow Stromal Cells Transplanted to the Adult Brain Are Rejected by an Inflammatory Response and Transfer Donor Labels to Host Neurons and Glia. Stem Cells 24: 2483- 2492. 74. Bianco P (2014) "Mesenchymal" stem cells. Annu Rev Cell Dev Biol 30: 677-704. 75. Moll G, Jitschin R, von Bahr L, Rasmusson-Duprez I, Sundberg B, et al. (2011) Mesenchymal Stromal Cells Engage Complement and Complement Receptor Bearing Innate Effector Cells to Modulate Immune Responses. PLoS ONE 6: e21703. 76. Liu H, Lu K, MacAry PA, Wong KL, Heng A, et al. (2012) Soluble molecules are key in maintaining the immunomodulatory activity of murine mesenchymal stromal cells. Journal of Cell Science 125: 200-208. 77. Badillo AT, Beggs KJ, Javazon EH, Tebbets JC, Flake AW (2007) Murine Bone Marrow Stromal Progenitor Cells Elicit an In Vivo Cellular and Humoral Alloimmune Response. Biology of blood and marrow transplantation : journal of the American Society for Blood and Marrow Transplantation 13: 412-422. 78. Hoogduijn MJ, Roemeling-van Rhijn M, Korevaar SS, Engela AU, Weimar W, et al. (2011) Immunological aspects of allogeneic and autologous mesenchymal stem cell therapies. Hum Gene Ther 22: 1587-1591. 79. Yoshimura A, Muto G (2011) TGF-beta function in immune suppression. Curr Top Microbiol Immunol 350: 127-147. 80. Yang SH, Park MJ, Yoon IH, Kim SY, Hong SH, et al. (2009) Soluble mediators from mesenchymal stem cells suppress T cell proliferation by inducing IL-10. Exp Mol Med 41: 315-324. 81. Krampera M, Cosmi L, Angeli R, Pasini A, Liotta F, et al. (2006) Role for Interferon-γ in the Immunomodulatory Activity of Human Bone Marrow Mesenchymal Stem Cells. Stem Cells 24: 386-398. 82. Akiyama K, Chen C, Wang D, Xu X, Qu C, et al. (2012) Mesenchymal-Stem-Cell- Induced Immunoregulation Involves FAS-Ligand-/FAS-Mediated T Cell Apoptosis. Cell stem cell 10: 544-555. 83. Li W, Ren G, Huang Y, Su J, Han Y, et al. (2012) Mesenchymal stem cells: a double-edged sword in regulating immune responses. Cell Death Differ 19: 1505-1513.

References 121

84. Casiraghi F, Remuzzi G, Abbate M, Perico N (2013) Multipotent Mesenchymal Stromal Cell Therapy and Risk of Malignancies. Stem Cell Reviews and Reports 9: 65-79. 85. Mao Y, Keller E, Garfield D, Shen K, Wang J (2013) Stromal cells in tumor microenvironment and breast cancer. Cancer and Metastasis Reviews 32: 303- 315. 86. Klopp AH, Gupta A, Spaeth E, Andreeff M, Marini F (2011) Concise Review: Dissecting a Discrepancy in the Literature: Do Mesenchymal Stem Cells Support or Suppress Tumor Growth? STEM CELLS 29: 11-19. 87. Kucerova L, Zmajkovic J, Toro L, Skolekova S, Demkova L, et al. (2015) Tumor- driven Molecular Changes in Human Mesenchymal Stromal Cells. Cancer Microenvironment 8: 1-14. 88. Bergfeld SA, Blavier L, DeClerck YA (2014) Bone Marrow–Derived Mesenchymal Stromal Cells Promote Survival and Drug Resistance in Tumor Cells. Molecular Cancer Therapeutics 13: 962-975. 89. Sharma RR, Pollock K, Hubel A, McKenna D (2014) Mesenchymal stem or stromal cells: a review of clinical applications and manufacturing practices. Transfusion 54: 1418-1437. 90. Hoogduijn MJ, Popp F, Verbeek R, Masoodi M, Nicolaou A, et al. (2010) The immunomodulatory properties of mesenchymal stem cells and their use for immunotherapy. International Immunopharmacology 10: 1496-1500. 91. Eliopoulos N, Stagg J, Lejeune L, Pommey S, Galipeau J (2005) Allogeneic marrow stromal cells are immune rejected by MHC class I– and class II– mismatched recipient mice. Blood 106: 4057-4065. 92. Le Blanc K, Tammik C, Rosendahl K, Zetterberg E, Ringden O (2003) HLA expression and immunologic properties of differentiated and undifferentiated mesenchymal stem cells. Exp Hematol 31: 890-896. 93. Zangi L, Margalit R, Reich-Zeliger S, Bachar-Lustig E, Beilhack A, et al. (2009) Direct imaging of immune rejection and memory induction by allogeneic mesenchymal stromal cells. Stem Cells 27: 2865-2874. 94. Moll G, Rasmusson-Duprez I, von Bahr L, Connolly-Andersen A-M, Elgue G, et al. (2012) Are Therapeutic Human Mesenchymal Stromal Cells Compatible with Human Blood? STEM CELLS 30: 1565-1574. 95. Chan JL, Tang KC, Patel AP, Bonilla LM, Pierobon N, et al. (2006) Antigen- presenting property of mesenchymal stem cells occurs during a narrow window at low levels of interferon-gamma. Blood 107: 4817-4824. 96. Stagg J, Pommey S, Eliopoulos N, Galipeau J (2006) Interferon-gamma-stimulated marrow stromal cells: a new type of nonhematopoietic antigen-presenting cell. Blood 107: 2570-2577. 97. Francois M, Romieu-Mourez R, Stock-Martineau S, Boivin MN, Bramson JL, et al. (2009) Mesenchymal stromal cells cross-present soluble exogenous antigens as part of their antigen-presenting cell properties. Blood 114: 2632- 2638. 98. Ren G, Su J, Zhang L, Zhao X, Ling W, et al. (2009) Species variation in the mechanisms of mesenchymal stem cell-mediated immunosuppression. Stem Cells 27: 1954-1962. 99. Spaggiari GM, Capobianco A, Becchetti S, Mingari MC, Moretta L (2006) Mesenchymal stem cell-natural killer cell interactions: evidence that activated NK cells are capable of killing MSCs, whereas MSCs can inhibit IL-2-induced NK-cell proliferation. Blood 107: 1484-1490.

122 References

100. Sundin M, Ringden O, Sundberg B, Nava S, Gotherstrom C, et al. (2007) No alloantibodies against mesenchymal stromal cells, but presence of anti-fetal calf serum antibodies, after transplantation in allogeneic hematopoietic stem cell recipients. Haematologica 92: 1208-1215. 101. Chen L, Tredget EE, Wu PY, Wu Y (2008) Paracrine factors of mesenchymal stem cells recruit macrophages and endothelial lineage cells and enhance wound healing. PLoS One 3: e1886. 102. Kinnaird T, Stabile E, Burnett MS, Lee CW, Barr S, et al. (2004) Marrow-Derived Stromal Cells Express Genes Encoding a Broad Spectrum of Arteriogenic Cytokines and Promote In Vitro and In Vivo Arteriogenesis Through Paracrine Mechanisms. Circulation Research 94: 678-685. 103. Rehman J, Traktuev D, Li J, Merfeld-Clauss S, Temm-Grove CJ, et al. (2004) Secretion of Angiogenic and Antiapoptotic Factors by Human Adipose Stromal Cells. Circulation 109: 1292-1298. 104. Choi H, Lee RH, Bazhanov N, Oh JY, Prockop DJ (2011) Anti-inflammatory protein TSG-6 secreted by activated MSCs attenuates zymosan-induced mouse peritonitis by decreasing TLR2/NF-κB signaling in resident macrophages. Blood 118: 330-338. 105. Lee RH, Pulin AA, Seo MJ, Kota DJ, Ylostalo J, et al. (2009) Intravenous hMSCs improve myocardial infarction in mice because cells embolized in lung are activated to secrete the anti-inflammatory protein TSG-6. Cell Stem Cell 5: 54- 63. 106. Qi Y, Jiang D, Sindrilaru A, Stegemann A, Schatz S, et al. (2014) TSG-6 Released from Intradermally Injected Mesenchymal Stem Cells Accelerates Wound Healing and Reduces Tissue Fibrosis in Murine Full-Thickness Skin Wounds. J Invest Dermatol 134: 526-537. 107. Nowakowski A, Walczak P, Janowski M, Lukomska B (2015) Genetic engineering of mesenchymal stem cells for regenerative medicine. Stem Cells and Development. 108. Jackson WM, Nesti LJ, Tuan RS (2011) Concise Review: Clinical Translation of Wound Healing Therapies Based on Mesenchymal Stem Cells. Stem Cells Translational Medicine 1: 44-50. 109. Tejedor LS, Berner G, Jacobsen K, Gudi V, Jungwirth N, et al. (2015) Mesenchymal stem cells do not exert direct beneficial effects on CNS remyelination in the absence of the peripheral immune system. Brain, Behavior, and Immunity. 110. Walter MN, Wright KT, Fuller HR, MacNeil S, Johnson WE (2010) Mesenchymal stem cell-conditioned medium accelerates skin wound healing: an in vitro study of fibroblast and keratinocyte scratch assays. Exp Cell Res 316: 1271-1281. 111. Takahashi M, Li T-S, Suzuki R, Kobayashi T, Ito H, et al. (2006) Cytokines produced by bone marrow cells can contribute to functional improvement of the infarcted heart by protecting cardiomyocytes from ischemic injury. American Journal of Physiology - Heart and Circulatory Physiology 291: H886-H893. 112. Selmani Z, Naji A, Zidi I, Favier B, Gaiffe E, et al. (2008) Human leukocyte antigen-G5 secretion by human mesenchymal stem cells is required to suppress T lymphocyte and natural killer function and to induce CD4+CD25highFOXP3+ regulatory T cells. Stem Cells 26: 212-222.

References 123

113. Di Nicola M, Carlo-Stella C, Magni M, Milanesi M, Longoni PD, et al. (2002) Human bone marrow stromal cells suppress T-lymphocyte proliferation induced by cellular or nonspecific mitogenic stimuli. Blood 99: 3838-3843. 114. da Silva Meirelles L, Caplan AI, Nardi NB (2008) In search of the in vivo identity of mesenchymal stem cells. Stem Cells 26: 2287-2299. 115. Ren G, Zhang L, Zhao X, Xu G, Zhang Y, et al. (2008) Mesenchymal Stem Cell- Mediated Immunosuppression Occurs via Concerted Action of Chemokines and Nitric Oxide. Cell stem cell 2: 141-150. 116. Chapel A, Bertho JM, Bensidhoum M, Fouillard L, Young RG, et al. (2003) Mesenchymal stem cells home to injured tissues when co-infused with hematopoietic cells to treat a radiation-induced multi-organ failure syndrome. J Gene Med 5: 1028-1038. 117. Spaeth E, Klopp A, Dembinski J, Andreeff M, Marini F (2008) Inflammation and tumor microenvironments: defining the migratory itinerary of mesenchymal stem cells. Gene Ther 15: 730-738. 118. Ponte AL, Marais E, Gallay N, Langonne A, Delorme B, et al. (2007) The in vitro migration capacity of human bone marrow mesenchymal stem cells: comparison of chemokine and growth factor chemotactic activities. Stem Cells 25: 1737-1745. 119. Schmidt A, Ladage D, Steingen C, Brixius K, Schinkothe T, et al. (2006) Mesenchymal stem cells transmigrate over the endothelial barrier. Eur J Cell Biol 85: 1179-1188. 120. Segers VF, Van Riet I, Andries LJ, Lemmens K, Demolder MJ, et al. (2006) Mesenchymal stem cell adhesion to cardiac microvascular endothelium: activators and mechanisms. Am J Physiol Heart Circ Physiol 290: H1370- 1377. 121. Ruster B, Gottig S, Ludwig RJ, Bistrian R, Muller S, et al. (2006) Mesenchymal stem cells display coordinated rolling and adhesion behavior on endothelial cells. Blood 108: 3938-3944. 122. Rombouts WJC, Ploemacher RE (2003) Primary murine MSC show highly efficient homing to the bone marrow but lose homing ability following culture. Leukemia 17: 160-170. 123. Barbash IM, Chouraqui P, Baron J, Feinberg MS, Etzion S, et al. (2003) Systemic Delivery of Bone Marrow–Derived Mesenchymal Stem Cells to the Infarcted Myocardium. Circulation 108: 863-868. 124. Zhang M, Mal N, Kiedrowski M, Chacko M, Askari AT, et al. (2007) SDF-1 expression by mesenchymal stem cells results in trophic support of cardiac myocytes after myocardial infarction. The FASEB Journal 21: 3197-3207. 125. Sackstein R, Merzaban JS, Cain DW, Dagia NM, Spencer JA, et al. (2008) Ex vivo glycan engineering of CD44 programs human multipotent mesenchymal stromal cell trafficking to bone. Nat Med 14: 181-187. 126. Sarkar D, Spencer JA, Phillips JA, Zhao W, Schafer S, et al. (2011) Engineered cell homing. Blood 118: 184-191. 127. Nejadnik H, Hui JH, Feng Choong EP, Tai B-C, Lee EH (2010) Autologous Bone Marrow–Derived Mesenchymal Stem Cells Versus Autologous Chondrocyte Implantation: An Observational Cohort Study. The American Journal of Sports Medicine 38: 1110-1116. 128. Wong KL, Lee KB, Tai BC, Law P, Lee EH, et al. (2013) Injectable cultured bone marrow-derived mesenchymal stem cells in varus knees with cartilage defects

124 References

undergoing high tibial osteotomy: a prospective, randomized controlled clinical trial with 2 years' follow-up. Arthroscopy 29: 2020-2028. 129. Mathiasen AB, Qayyum AA, Jørgensen E, Helqvist S, Fischer-Nielsen A, et al. (2015) Bone marrow-derived mesenchymal stromal cell treatment in patients with severe ischaemic heart failure: a randomized placebo-controlled trial (MSC-HF trial). European Heart Journal. 130. Chen S-l, Fang W-w, Ye F, Liu Y-H, Qian J, et al. (2004) Effect on left ventricular function of intracoronary transplantation of autologous bone marrow mesenchymal stem cell in patients with acute myocardial infarction. The American Journal of Cardiology 94: 92-95. 131. Heldman AW, DiFede DL, Fishman JE, et al. (2014) Transendocardial mesenchymal stem cells and mononuclear bone marrow cells for ischemic cardiomyopathy: The tac-hft randomized trial. JAMA 311: 62-73. 132. Perin EC, Borow KM, Silva GV, DeMaria AN, Marroquin OC, et al. (2015) A Phase II Dose-Escalation Study of Allogeneic Mesenchymal Precursor Cells in Patients With Ischemic or Non-Ischemic Heart Failure. Circulation Research. 133. Hare JM, Traverse JH, Henry TD, Dib N, Strumpf RK, et al. (2009) A Randomized, Double-Blind, Placebo-Controlled, Dose-Escalation Study of Intravenous Adult Human Mesenchymal Stem Cells (Prochymal) After Acute Myocardial Infarction. Journal of the American College of Cardiology 54: 2277-2286. 134. Chullikana A, Majumdar AS, Gottipamula S, Krishnamurthy S, Kumar AS, et al. (2015) Randomized, double-blind, phase I/II study of intravenous allogeneic mesenchymal stromal cells in acute myocardial infarction. Cytotherapy 17: 250-261. 135. Lee JS, Hong JM, Moon GJ, Lee PH, Ahn YH, et al. (2010) A Long-Term Follow- Up Study of Intravenous Autologous Mesenchymal Stem Cell Transplantation in Patients With Ischemic Stroke. Stem Cells 28: 1099-1106. 136. Bhasin A, Padma Srivastava MV, Mohanty S, Bhatia R, Kumaran SS, et al. (2013) Stem cell therapy: A clinical trial of stroke. Clinical Neurology and Neurosurgery 115: 1003-1008. 137. Liao GP, Harting MT, Hetz RA, Walker PA, Shah SK, et al. (2015) Autologous Bone Marrow Mononuclear Cells Reduce Therapeutic Intensity for Severe Traumatic Brain Injury in Children. Pediatric Critical Care Medicine 16: 245- 255. 138. Kharaziha P, Hellström PM, Noorinayer B, Farzaneh F, Aghajani K, et al. (2009) Improvement of liver function in liver cirrhosis patients after autologous mesenchymal stem cell injection: a phase I-II clinical trial. European Journal of Gastroenterology & Hepatology 21: 1199-1205 1110.1097/MEG.1190b1013e32832a32831f32836c. 139. Duijvestein M, Vos ACW, Roelofs H, Wildenberg ME, Wendrich BB, et al. (2010) Autologous bone marrow-derived mesenchymal stromal cell treatment for refractory luminal Crohn's disease: results of a phase I study. Gut 59: 1662- 1669. 140. Forbes GM, Sturm MJ, Leong RW, Sparrow MP, Segarajasingam D, et al. (2014) A Phase 2 Study of Allogeneic Mesenchymal Stromal Cells for Luminal Crohn's Disease Refractory to Biologic Therapy. Clinical Gastroenterology and Hepatology 12: 64-71.

References 125

141. Le Blanc K, Frassoni F, Ball L, Locatelli F, Roelofs H, et al. (2008) Mesenchymal stem cells for treatment of steroid-resistant, severe, acute graft-versus-host disease: a phase II study. The Lancet 371: 1579-1586. 142. Maziarz RT, Devos T, Bachier CR, Goldstein SC, Leis JF, et al. (2014) Single and Multiple Dose MultiStem® (Multi-Potent Adult Progenitor Cell) Therapy Prophylaxis of Acute GVHD in Myeloablative Allogeneic Hematopoietic Cell Transplantation: A Phase I Trial. Biology of Blood and Marrow Transplant. 143. Llufriu S, Sepulveda M, Blanco Y, Marin P, Moreno B, et al. (2014) Randomized placebo-controlled phase II trial of autologous mesenchymal stem cells in multiple sclerosis. PLoS One 9: e113936. 144. Falanga V, Iwamoto S, Chartier M, Yufit T, Butmarc J, et al. (2007) Autologous bone marrow-derived cultured mesenchymal stem cells delivered in a fibrin spray accelerate healing in murine and human cutaneous wounds. Tissue Eng 13: 1299-1312. 145. Dash NR, Dash SN, Routray P, Mohapatra S, Mohapatra PC (2009) Targeting nonhealing ulcers of lower extremity in human through autologous bone marrow-derived mesenchymal stem cells. Rejuvenation Res 12: 359-366. 146. Hatzistergos KE, Quevedo H, Oskouei BN, Hu Q, Feigenbaum GS, et al. (2010) Bone marrow mesenchymal stem cells stimulate cardiac stem cell proliferation and differentiation. Circ Res 107: 913-922. 147. Sassoli C, Pini A, Mazzanti B, Quercioli F, Nistri S, et al. (2011) Mesenchymal stromal cells affect cardiomyocyte growth through juxtacrine Notch-1/Jagged- 1 signaling and paracrine mechanisms: clues for cardiac regeneration. J Mol Cell Cardiol 51: 399-408. 148. Timmers L, Lim SK, Hoefer IE, Arslan F, Lai RC, et al. (2011) Human mesenchymal stem cell-conditioned medium improves cardiac function following myocardial infarction. Stem Cell Res 6: 206-214. 149. Gnecchi M, Zhang Z, Ni A, Dzau VJ (2008) Paracrine mechanisms in adult stem cell signaling and therapy. Circ Res 103: 1204-1219. 150. Gao LR, Chen Y, Zhang NK, Yang XL, Liu HL, et al. (2015) Intracoronary infusion of Wharton's jelly-derived mesenchymal stem cells in acute myocardial infarction: double-blind, randomized controlled trial. BMC Med 13: 162. 151. Katritsis DG, Sotiropoulou PA, Karvouni E, Karabinos I, Korovesis S, et al. (2005) Transcoronary transplantation of autologous mesenchymal stem cells and endothelial progenitors into infarcted human myocardium. Catheter Cardiovasc Interv 65: 321-329. 152. Chen S, Liu Z, Tian N, Zhang J, Yei F, et al. (2006) Intracoronary transplantation of autologous bone marrow mesenchymal stem cells for ischemic cardiomyopathy due to isolated chronic occluded left anterior descending artery. J Invasive Cardiol 18: 552-556. 153. Ripa RS, Haack-Sorensen M, Wang Y, Jorgensen E, Mortensen S, et al. (2007) Bone marrow derived mesenchymal cell mobilization by granulocyte-colony stimulating factor after acute myocardial infarction: results from the Stem Cells in Myocardial Infarction (STEMMI) trial. Circulation 116: I24-30. 154. Mohyeddin-Bonab M, Mohamad-Hassani MR, Alimoghaddam K, Sanatkar M, Gasemi M, et al. (2007) Autologous in vitro expanded mesenchymal stem cell therapy for human old myocardial infarction. Arch Iran Med 10: 467-473.

126 References

155. Katritsis DG, Sotiropoulou P, Giazitzoglou E, Karvouni E, Papamichail M (2007) Electrophysiological effects of intracoronary transplantation of autologous mesenchymal and endothelial progenitor cells. Europace 9: 167-171. 156. Hare JM, Traverse JH, Henry TD, Dib N, Strumpf RK, et al. (2009) A randomized, double-blind, placebo-controlled, dose-escalation study of intravenous adult human mesenchymal stem cells (prochymal) after acute myocardial infarction. J Am Coll Cardiol 54: 2277-2286. 157. Burns TC, Verfaillie CM, Low WC (2009) Stem cells for ischemic brain injury: A critical review. The Journal of Comparative Neurology 515: 125-144. 158. Bara JJ, Richards RG, Alini M, Stoddart MJ (2014) Concise Review: Bone Marrow-Derived Mesenchymal Stem Cells Change Phenotype Following In Vitro Culture: Implications for Basic Research and the Clinic. STEM CELLS 32: 1713-1723. 159. Noble LJ, Donovan F, Igarashi T, Goussev S, Werb Z (2002) Matrix metalloproteinases limit functional recovery after spinal cord injury by modulation of early vascular events. J Neurosci 22: 7526-7535. 160. Carlson SL, Parrish ME, Springer JE, Doty K, Dossett L (1998) Acute Inflammatory Response in Spinal Cord Following Impact Injury. Experimental Neurology 151: 77-88. 161. Oliveri RS, Bello S, Biering-Sørensen F (2014) Mesenchymal stem cells improve locomotor recovery in traumatic spinal cord injury: Systematic review with meta-analyses of rat models. Neurobiology of Disease 62: 338-353. 162. Karamouzian S, Nematollahi-Mahani SN, Nakhaee N, Eskandary H (2012) Clinical safety and primary efficacy of bone marrow mesenchymal cell transplantation in subacute spinal cord injured patients. Clinical Neurology and Neurosurgery 114: 935-939. 163. Yoon SH, Shim YS, Park YH, Chung JK, Nam JH, et al. (2007) Complete Spinal Cord Injury Treatment Using Autologous Bone Marrow Cell Transplantation and Bone Marrow Stimulation with Granulocyte Macrophage-Colony Stimulating Factor: Phase I/II Clinical Trial. STEM CELLS 25: 2066-2073. 164. Wang S, Cheng H, Dai G, Wang X, Hua R, et al. (2013) Umbilical cord mesenchymal stem cell transplantation significantly improves neurological function in patients with sequelae of traumatic brain injury. Brain Research 1532: 76-84. 165. Ciccocioppo R, Bernardo ME, Sgarella A, Maccario R, Avanzini MA, et al. (2011) Autologous bone marrow-derived mesenchymal stromal cells in the treatment of fistulising Crohn's disease. Gut 60: 788-798. 166. Mohyeddin Bonab M, Yazdanbakhsh S, Lotfi J, Alimoghaddom K, Talebian F, et al. (2007) Does mesenchymal stem cell therapy help multiple sclerosis patients? Report of a pilot study. Iran J Immunol 4: 50-57. 167. Uccelli A, Laroni A, Freedman MS (2011) Mesenchymal stem cells for the treatment of multiple sclerosis and other neurological diseases. Lancet Neurol 10: 649-656. 168. Rice CM, Kemp K, Wilkins A, Scolding NJ Cell therapy for multiple sclerosis: an evolving concept with implications for other neurodegenerative diseases. The Lancet 382: 1204-1213. 169. Ringdén O, Uzunel M, Rasmusson I, Remberger M, Sundberg B, et al. (2006) Mesenchymal Stem Cells for Treatment of Therapy-Resistant Graft-versus- Host Disease. Transplantation 81: 1390-1397 1310.1097/1301.tp.0000214462.0000263943.0000214414.

References 127

170. Fang B, Song Y, Liao L, Zhang Y, Zhao RC (2007) Favorable Response to Human Adipose Tissue-Derived Mesenchymal Stem Cells in Steroid- Refractory Acute Graft-Versus-Host Disease. Transplantation Proceedings 39: 3358-3362. 171. Fang B, Song YP, Liao LM, Han Q, Zhao RC (2006) Treatment of severe therapy- resistant acute graft-versus-host disease with human adipose tissue-derived mesenchymal stem cells. Bone marrow transplantation 38: 389-390. 172. Lazarus HM, Koc ON, Devine SM, Curtin P, Maziarz RT, et al. (2005) Cotransplantation of HLA-Identical Sibling Culture-Expanded Mesenchymal Stem Cells and Hematopoietic Stem Cells in Hematologic Malignancy Patients. Biology of Blood and Marrow Transplantation 11: 389-398. 173. Pérez-Simon JA, López-Villar O, Andreu EJ, Rifón J, Muntion S, et al. (2011) Mesenchymal stem cells expanded in vitro with human serum for the treatment of acute and chronic graft-versus-host disease: results of a phase I/II clinical trial. 1072-1076 p. 174. Yin F, Battiwalla M, Ito S, Feng X, Chinian F, et al. (2014) Bone Marrow Mesenchymal Stromal Cells to Treat Tissue Damage in Allogeneic Stem Cell Transplant Recipients: Correlation of Biological Markers with Clinical Responses. STEM CELLS 32: 1278-1288. 175. Yew TL, Hung YT, Li HY, Chen HW, Chen LL, et al. (2011) Enhancement of wound healing by human multipotent stromal cell conditioned medium: the paracrine factors and p38 MAPK activation. Cell Transplant 20: 693-706. 176. Peura M, Bizik J, Salmenperä P, Noro A, Korhonen M, et al. (2009) Bone marrow mesenchymal stem cells undergo nemosis and induce keratinocyte wound healing utilizing the HGF/c-Met/PI3K pathway. Wound Repair and Regeneration 17: 569-577. 177. Nemeth K, Leelahavanichkul A, Yuen PST, Mayer B, Parmelee A, et al. (2009) Bone marrow stromal cells attenuate sepsis via prostaglandin E2-dependent reprogramming of host macrophages to increase their interleukin-10 production. Nat Med 15: 42-49. 178. Reitamo S, Remitz A, Tamai K, Uitto J (1994) Interleukin-10 modulates type I collagen and matrix metalloprotease gene expression in cultured human skin fibroblasts. J Clin Invest 94: 2489-2492. 179. Jeon YK, Jang YH, Yoo DR, Kim SN, Lee SK, et al. (2010) Mesenchymal stem cells' interaction with skin: Wound-healing effect on fibroblast cells and skin tissue. Wound Repair and Regeneration 18: 655-661. 180. Gruber R, Kandler B, Holzmann P, Vogele-Kadletz M, Losert U, et al. (2005) Bone marrow stromal cells can provide a local environment that favors migration and formation of tubular structures of endothelial cells. Tissue Eng 11: 896-903. 181. Kaigler D, Krebsbach PH, Polverini PJ, Mooney DJ (2003) Role of vascular endothelial growth factor in bone marrow stromal cell modulation of endothelial cells. Tissue Eng 9: 95-103. 182. Lozito TP, Taboas JM, Kuo CK, Tuan RS (2009) Mesenchymal stem cell modification of endothelial matrix regulates their vascular differentiation. J Cell Biochem 107: 706-713. 183. Au P, Tam J, Fukumura D, Jain RK (2008) Bone marrow-derived mesenchymal stem cells facilitate engineering of long-lasting functional vasculature. Blood 111: 4551-4558.

128 References

184. Renault MA, Roncalli J, Tongers J, Misener S, Thorne T, et al. (2009) The Hedgehog transcription factor Gli3 modulates angiogenesis. Circ Res 105: 818-826. 185. Li L, Zhang Y, Li Y, Yu B, Xu Y, et al. (2008) Mesenchymal stem cell transplantation attenuates cardiac fibrosis associated with isoproterenol- induced global heart failure. Transpl Int 21: 1181-1189. 186. Kim CH, Lee JH, Won JH, Cho MK (2011) Mesenchymal stem cells improve wound healing in vivo via early activation of matrix metalloproteinase-9 and vascular endothelial growth factor. J Korean Med Sci 26: 726-733. 187. Schievenbusch S, Strack I, Scheffler M, Wennhold K, Maurer J, et al. (2009) Profiling of anti-fibrotic signaling by hepatocyte growth factor in renal fibroblasts. Biochem Biophys Res Commun 385: 55-61. 188. Mou S, Wang Q, Shi B, Gu L, Ni Z (2009) Hepatocyte growth factor suppresses transforming growth factor-beta-1 and type III collagen in human primary renal fibroblasts. Kaohsiung J Med Sci 25: 577-587. 189. Bevan D, Gherardi E, Fan TP, Edwards D, Warn R (2004) Diverse and potent activities of HGF/SF in skin wound repair. J Pathol 203: 831-838. 190. Shukla MN, Rose JL, Ray R, Lathrop KL, Ray A, et al. (2009) Hepatocyte growth factor inhibits epithelial to myofibroblast transition in lung cells via Smad7. Am J Respir Cell Mol Biol 40: 643-653. 191. Gabr H, Hedayet A, Imam U, Nasser M (2011) Limb salvage using intramuscular injection of unfractionated autologous bone marrow mononuclear cells in critical limb ischemia: A prospective pilot clinical trial. Experimental and Clinical Transplantation 9: 197-202. 192. Bartsch T, Brehm M, Falke T, Kögler G, Wernet P, et al. (2005) Rapid healing of a therapy-refractory diabetic foot after transplantation of autologous bone marrow stem cells. Schnelle abheilung eines therapierefraktären diabetischen fußes nach autologer knochenmarkstammzelltransplantation 100: 676-680. 193. Kawamura A, Horie T, Tsuda I, Ikeda A, Egawa H, et al. (2005) Prevention of limb amputation in patients with limbs ulcers by autologous peripheral blood mononuclear cell implantation. Therapeutic Apheresis and Dialysis 9: 59-63. 194. Huang P, Li S, Han M, Xiao Z, Yang R, et al. (2005) Autologous transplantation of granulocyte colony-stimulating factor-mobilized peripheral blood mononuclear cells improves critical limb ischemia in diabetes. Diabetes Care 28: 2155-2160. 195. Prockop DJ, Prockop SE, Bertoncello I (2014) Are Clinical Trials With Mesenchymal Stem/Progenitor Cells too Far Ahead of the Science? Lessons From Experimental Hematology. STEM CELLS 32: 3055-3061. 196. Steinert AF, Rackwitz L, Gilbert F, Nöth U, Tuan RS (2012) Concise Review: The Clinical Application of Mesenchymal Stem Cells for Musculoskeletal Regeneration: Current Status and Perspectives. Stem Cells Translational Medicine 1: 237-247. 197. Schrepfer S, Deuse T, Reichenspurner H, Fischbein MP, Robbins RC, et al. (2007) Stem cell transplantation: the lung barrier. Transplant Proc 39: 573-576. 198. Jaramillo-Ferrada PA, Wolvetang EJ, Cooper-White JJ (2012) Differential mesengenic potential and expression of stem cell-fate modulators in mesenchymal stromal cells from human-term placenta and bone marrow. Journal of Cellular Physiology 227: 3234-3242. 199. dos Santos F, Andrade PZ, Boura JS, Abecasis MM, da Silva CL, et al. (2010) Ex vivo expansion of human mesenchymal stem cells: A more effective cell

References 129

proliferation kinetics and metabolism under hypoxia. Journal of Cellular Physiology 223: 27-35. 200. Du L, Yu Y, Ma H, Lu X, Ma L, et al. (2014) Hypoxia Enhances Protective Effect of Placental-Derived Mesenchymal Stem Cells on Damaged Intestinal Epithelial Cells by Promoting Secretion of Insulin-Like Growth Factor-1. International Journal of Molecular Sciences 15: 1983-2002. 201. Leijten J, Georgi N, Moreira Teixeira L, van Blitterswijk CA, Post JN, et al. (2014) Metabolic programming of mesenchymal stromal cells by oxygen tension directs chondrogenic cell fate. Proceedings of the National Academy of Sciences 111: 13954-13959. 202. Markway BD, Tan GK, Brooke G, Hudson JE, Cooper-White JJ, et al. (2010) Enhanced chondrogenic differentiation of human bone marrow-derived mesenchymal stem cells in low oxygen environment micropellet cultures. Cell Transplant 19: 29-42. 203. Muscari C, Giordano E, Bonafe F, Govoni M, Pasini A, et al. (2013) Priming adult stem cells by hypoxic pretreatments for applications in regenerative medicine. Journal of Biomedical Science 20: 63. 204. Dissmann P, Kansy B, Bruderek K, Dumitru C, Lang S, et al. (2015) 24 Cross- talk of tumor cells with mesenchymal stromal cells enhances tumor progression in head and neck cancer. Oral Oncology 51: e35. 205. Hernanda PY, Pedroza-Gonzalez A, Sprengers D, Peppelenbosch MP, Pan Q (2014) Multipotent mesenchymal stromal cells in liver cancer: Implications for tumor biology and therapy. Biochimica et Biophysica Acta (BBA) - Reviews on Cancer 1846: 439-445. 206. Zeitouni S, Krause U, Clough BH, Halderman H, Falster A, et al. (2012) Human mesenchymal stem cell-derived matrices for enhanced osteoregeneration. Sci Transl Med 4: 132ra155. 207. Ala-Kokko L, Günzler V, Hoek JB, Rubin E, Prockop DJ (1992) Hepatic fibrosis in rats produced by carbon tetrachloride and dimethylnitrosamine: Observations suggesting immunoassays of serum for the 75 fragment of type IV collagen are a more sensitive index of liver damage than immunoassays for the NH2-terminal propeptide of type III procollagen. Hepatology 16: 167-172. 208. Davidson JM, Yu F, Opalenik SR (2013) Splinting Strategies to Overcome Confounding Wound Contraction in Experimental Animal Models. Adv Wound Care (New Rochelle) 2: 142-148. 209. Serra M, Brito C, Correia C, Alves PM (2012) Process engineering of human pluripotent stem cells for clinical application. Trends in Biotechnology 30: 350-359. 210. Timmins NE, Kiel M, Günther M, Heazlewood C, Doran MR, et al. (2012) Closed system isolation and scalable expansion of human placental mesenchymal stem cells. Biotechnology and Bioengineering 109: 1817-1826. 211. Godara P, McFarland CD, Nordon RE (2008) Design of bioreactors for mesenchymal stem cell tissue engineering. Journal of Chemical Technology & Biotechnology 83: 408-420. 212. Rodrigues CA, Fernandes TG, Diogo MM, da Silva CL, Cabral JM (2011) Stem cell cultivation in bioreactors. Biotechnol Adv 29: 815-829. 213. Cao C, Zhang H, Gong L, He Y, Zhang N (2012) High glucose conditions suppress the function of bone marrow-derived endothelial progenitor cells via inhibition of the eNOS-caveolin-1 complex. Mol Med Rep 5: 341-346.

130 References

214. Ferrari C, Olmos E, Balandras F, Tran N, Chevalot I, et al. (2014) Investigation of growth conditions for the expansion of porcine mesenchymal stem cells on microcarriers in stirred cultures. Appl Biochem Biotechnol 172: 1004-1017. 215. Li YM, Schilling T, Benisch P, Zeck S, Meissner-Weigl J, et al. (2007) Effects of high glucose on mesenchymal stem cell proliferation and differentiation. Biochem Biophys Res Commun 363: 209-215. 216. Ryu JM, Lee MY, Yun SP, Han HJ (2010) High glucose regulates cyclin D1/E of human mesenchymal stem cells through TGF-beta1 expression via Ca2+/PKC/MAPKs and PI3K/Akt/mTOR signal pathways. J Cell Physiol 224: 59-70. 217. Lo T, Ho JH, Yang MH, Lee OK (2011) Glucose reduction prevents replicative senescence and increases mitochondrial respiration in human mesenchymal stem cells. Cell Transplant 20: 813-825. 218. Weber C, Pohl S, Poertner R, Pino-Grace P, Freimark D, et al. (2010) Production process for stem cell based therapeutic implants: expansion of the production cell line and cultivation of encapsulated cells. Adv Biochem Eng Biotechnol 123: 143-162. 219. Yeatts AB, Choquette DT, Fisher JP (2013) Bioreactors to influence stem cell fate: augmentation of mesenchymal stem cell signaling pathways via dynamic culture systems. Biochim Biophys Acta 1830: 2470-2480. 220. Zhao F, Chella R, Ma T (2007) Effects of shear stress on 3-D human mesenchymal stem cell construct development in a perfusion bioreactor system: Experiments and hydrodynamic modeling. Biotechnol Bioeng 96: 584- 595. 221. Li D, Tang T, Lu J, Dai K (2009) Effects of flow shear stress and mass transport on the construction of a large-scale tissue-engineered bone in a perfusion bioreactor. Tissue Eng Part A 15: 2773-2783. 222. Kinney MA, Sargent CY, McDevitt TC (2011) The Multiparametric Effects of Hydrodynamic Environments on Stem Cell Culture. Tissue Engineering Part B: Reviews 17: 249-262. 223. Grayson WL, Zhao F, Izadpanah R, Bunnell B, Ma T (2006) Effects of hypoxia on human mesenchymal stem cell expansion and plasticity in 3D constructs. Journal of Cellular Physiology 207: 331-339. 224. Liu N, Zang R, Yang S-T, Li Y (2014) Stem cell engineering in bioreactors for large-scale bioprocessing. Engineering in Life Sciences 14: 4-15. 225. King JA, Miller WM (2007) Bioreactor development for stem cell expansion and controlled differentiation. Current Opinion in Chemical Biology 11: 394-398. 226. Weber C, Freimark D, Portner R, Pino-Grace P, Pohl S, et al. (2010) 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: 512-525. 227. Dubois J, Martin Y, Chouinard JA, Lecomte R, Vermette P (2011) 2.28 - Bioreactors for Tissue Engineering: Design, Applications, and Monitoring. In: Editor-in-Chief: Murray M-Y, editor. Comprehensive Biotechnology (Second Edition). Burlington: Academic Press. pp. 383-399. 228. Santos F, Andrade PZ, Abecasis MM, Gimble JM, Chase LG, et al. (2011) 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: 1201-1210.

References 131

229. Housler GJ, Miki T, Schmelzer E, Pekor C, Zhang X, et al. (2012) Compartmental Hollow Fiber Capillary Membrane–Based Bioreactor Technology for In Vitro Studies on Red Blood Cell Lineage Direction of Hematopoietic Stem Cells. Tissue Engineering Part C, Methods 18: 133-142. 230. Li M, Tilles AW, Milwid JM, Hammad M, Lee J, et al. (2012) Phenotypic and functional characterization of human bone marrow stromal cells in hollow- fibre bioreactors. J Tissue Eng Regen Med 6: 369-377. 231. Timmins NE, Kiel M, Gunther M, Heazlewood C, Doran MR, et al. (2012) Closed system isolation and scalable expansion of human placental mesenchymal stem cells. Biotechnol Bioeng 109: 1817-1826. 232. Yeatts AB, Both SK, Yang W, Alghamdi HS, Yang F, et al. (2014) In vivo bone regeneration using tubular perfusion system bioreactor cultured nanofibrous scaffolds. Tissue Eng Part A 20: 139-146. 233. Kirouac DC, Zandstra PW (2008) The Systematic Production of Cells for Cell Therapies. Cell Stem Cell 3: 369-381. 234. Eibl R, Werner S, Eibl D (2010) Bag Bioreactor Based on Wave-Induced Motion: Characteristics and Applications Disposable Bioreactors. In: Eibl R, Eibl D, editors: Springer Berlin / Heidelberg. pp. 55-87. 235. Singh V (1999) Disposable bioreactor for cell culture using wave-induced agitation. Cytotechnology 30: 149-158. 236. Kehoe DE, Jing D, Lock LT, Tzanakakis ES (2010) Scalable stirred-suspension bioreactor culture of human pluripotent stem cells. Tissue Eng Part A 16: 405- 421. 237. Storm MP, Orchard CB, Bone HK, Chaudhuri JB, Welham MJ (2010) Three- dimensional culture systems for the expansion of pluripotent embryonic stem cells. Biotechnology and Bioengineering 107: 683-695. 238. Frith JE, Thomson B, Genever PG (2010) Dynamic three-dimensional culture methods enhance mesenchymal stem cell properties and increase therapeutic potential. Tissue Eng Part C Methods 16: 735-749. 239. Eibes G, dos Santos F, Andrade PZ, Boura JS, Abecasis MMA, et al. (2010) Maximizing the ex vivo expansion of human mesenchymal stem cells using a microcarrier-based stirred culture system. Journal of Biotechnology 146: 194- 197. 240. Yu Y, Li K, Bao C, Liu T, Jin Y, et al. (2009) Ex Vitro Expansion of Human Placenta-Derived Mesenchymal Stem Cells in Stirred Bioreactor. Applied Biochemistry and Biotechnology 159: 110-118. 241. Schop D, Janssen FW, Borgart E, de Bruijn JD, van Dijkhuizen-Radersma R (2008) Expansion of mesenchymal stem cells using a microcarrier-based cultivation system: growth and metabolism. Journal of Tissue Engineering and Regenerative Medicine 2: 126-135. 242. Carmelo JG, Fernandes-Platzgummer AM, Diogo MM, Lobato da Silva C, Cabral JM (2015) A xeno-free microcarrier-based stirred culture system for the scalable expansion of human mesenchymal stem/stromal cells isolated from bone marrow and adipose tissue. Biotechnol J. 243. Kaiser SC, Eibl R, Eibl D (2011) Engineering characteristics of a single-use stirred bioreactor at bench-scale: The Mobius CellReady 3L bioreactor as a case study. Engineering in Life Sciences 11: 359-368. 244. D J, N S, S P, M A (2013) Growth Kinetics of Human Mesenchymal Stem Cells in a 3-L Single-Use, Stirred-Tank Bioreactor. BioPharm Int: 28-38.

132 References

245. Jossen V, Kaiser SC, Schirmaier C, Herrmann J, Tappe A, et al. (2014) Modification and qualification of a stirred single-use bioreactor for the improved expansion of human mesenchymal stem cells at benchtop scale. Pharmaceutical Bioprocessing 2: 311-322. 246. Shanglong X, Pingan D, Youzhuan X, Yang Y (2008) Cell distribution in a scaffold with random architectures under the influence of fluid dynamics. J Biomater Appl 23: 229-245. 247. Chen X, Xu H, Wan C, McCaigue M, Li G (2006) Bioreactor Expansion of Human Adult Bone Marrow-Derived Mesenchymal Stem Cells. Stem Cells 24: 2052-2059. 248. Song K, Yang Z, Liu T, Zhi W, Li X, et al. (2006) Fabrication and detection of tissue-engineered bones with bio-derived scaffolds in a rotating bioreactor. Biotechnology and Applied Biochemistry 45: 65-74. 249. Granet C, Laroche N, Vico L, Alexandre C, Lafage-Proust M (1998) Rotating- wall vessels, promising bioreactors for osteoblastic cell culture: comparison with other 3D conditions. Medical and Biological Engineering and Computing 36: 513-519. 250. Cherry RS, Papoutsakis ET (1986) Hydrodynamic effects on cells in agitated tissue culture reactors. Bioprocess and Biosystems Engineering 1: 29-41. 251. Nordon RE, Schindhelm K (1997) Design of Hollow Fiber Modules for Uniform Shear Elution Affinity Cell Separation. Artificial Organs 21: 107-115. 252. Martin Y, Vermette P (2005) Bioreactors for tissue mass culture: Design, characterization, and recent advances. Biomaterials 26: 7481-7503. 253. Piret JM, Cooney CL (1991) Model of oxygen transport limitations in hollow fiber bioreactors. Biotechnology and Bioengineering 37: 80-92. 254. Brotherton JD, Chau PC (1995) Protein-free human–human hybridoma cultures in an intercalated-spiral alternate-dead-ended hollow fiber bioreactor. Biotechnology and Bioengineering 47: 384-400. 255. Sladkova M, de Peppo G (2014) Bioreactor Systems for Human Bone Tissue Engineering. Processes 2: 494. 256. Meissner P, Schröder B, Herfurth C, Biselli M (1999) Development of a fixed bed bioreactor for the expansion of human hematopoietic progenitor cells. Cytotechnology 30: 227-234. 257. Jelinek N, Schmidt S, Hilbert U, Thoma S, Biselli M, et al. (2002) Novel Bioreactors for the ex vivo Cultivation of Hematopoietic Cells. Engineering in Life Sciences 2: 15-18. 258. Mizukami A, Orellana MD, Caruso SR, de Lima Prata K, Covas DT, et al. (2013) Efficient expansion of mesenchymal stromal cells in a disposable fixed bed culture system. Biotechnol Prog 29: 568-572. 259. Ang LP, Cheng ZY, Beuerman RW, Teoh SH, Zhu X, et al. (2006) The development of a serum-free derived bioengineered conjunctival epithelial equivalent using an ultrathin poly(epsilon-caprolactone) membrane substrate. Invest Ophthalmol Vis Sci 47: 105-112. 260. Fiedler J, Özdemir B, Bartholomä J, Plettl A, Brenner RE, et al. (2013) The effect of substrate surface nanotopography on the behavior of multipotnent mesenchymal stromal cells and osteoblasts. Biomaterials 34: 8851-8859. 261. Li Y, Kniss DA, Lasky LC, Yang ST (2003) Culturing and differentiation of murine embryonic stem cells in a three-dimensional fibrous matrix. Cytotechnology 41: 23-35.

References 133

262. Tan G-K, Dinnes DLM, Cooper-White JJ (2011) Modulation of collagen II fiber formation in 3-D porous scaffold environments. Acta Biomaterialia 7: 2804- 2816. 263. Webb K, Hlady V, Tresco PA (1998) Relative importance of surface wettability and charged functional groups on NIH 3T3 fibroblast attachment, spreading, and cytoskeletal organization. Journal of Biomedical Materials Research 41: 422-430. 264. Bock A, Sann H, Schulze-Horsel J, Genzel Y, Reichl U, et al. (2009) Growth behavior of number distributed adherent MDCK cells for optimization in microcarrier cultures. Biotechnol Prog 25: 1717-1731. 265. van Kooten TG, Spijker HT, Busscher HJ (2004) Plasma-treated polystyrene surfaces: model surfaces for studying cell-biomaterial interactions. Biomaterials 25: 1735-1747. 266. Morent R, De Geyter N, Trentesaux M, Gengembre L, Dubruel P, et al. (2010) Influence of Discharge Atmosphere on the Ageing Behaviour of Plasma- Treated Polylactic Acid. Plasma Chemistry and Plasma Processing 30: 525- 536. 267. Pu FR, Williams RL, Markkula TK, Hunt JA (2002) Effects of plasma treated PET and PTFE on expression of adhesion molecules by human endothelial cells in vitro. Biomaterials 23: 2411-2428. 268. Ramires PA, Mirenghi L, Romano AR, Palumbo F, Nicolardi G (2000) Plasma- treated PET surfaces improve the biocompatibility of human endothelial cells. Journal of Biomedical Materials Research 51: 535-539. 269. Desmet T, Morent R, Geyter ND, Leys C, Schacht E, et al. (2009) Nonthermal Plasma Technology as a Versatile Strategy for Polymeric Biomaterials Surface Modification: A Review. Biomacromolecules 10: 2351-2378. 270. Hambardzumyan A, Biltresse S, Dufrêne Y, Marchand-Brynaert J (2002) An Unprecedented Surface Oxidation of Polystyrene Substrates by Wet Chemistry under Basic Conditions. Journal of Colloid and Interface Science 252: 443- 449. 271. Yoon Sung N, Joon Jin Y, Jae Gwan L, Tae Gwan P (1999) Adhesion behaviours of hepatocytes cultured onto biodegradable polymer surface modified by alkali hydrolysis process. Journal of Biomaterials Science, Polymer Edition 10: 1145-1158. 272. Dalton BA, McFarland CD, Gengenbach TR, Griesser HJ, Steele JG (1998) Polymer surface chemistry and bone cell migration. Journal of Biomaterials Science, Polymer Edition 9: 781-799. 273. Curtis AS, Forrester JV, McInnes C, Lawrie F (1983) Adhesion of cells to polystyrene surfaces. The Journal of Cell Biology 97: 1500-1506. 274. Goddard JM, Hotchkiss JH (2007) Polymer surface modification for the attachment of bioactive compounds. Progress in Polymer Science 32: 698-725. 275. Desai S, Singh RP (2004) Surface Modification of Polyethylene. In: Albertsson A-C, editor. Long Term Properties of Polyolefins: Springer Berlin Heidelberg. pp. 231-294. 276. Chong MSK, Lee CN, Teoh SH (2007) Characterization of smooth muscle cells on poly(ε-caprolactone) films. Materials Science and Engineering: C 27: 309- 312. 277. Ho MH, Lee JJ, Fan SC, Wang DM, Hou LT, et al. (2007) Efficient modification on PLLA by ozone treatment for biomedical applications. Macromol Biosci 7: 467-474.

134 References

278. Koo G-H, Jang J (2008) Surface modification of poly(lactic acid) by UV/Ozone irradiation. Fibers and Polymers 9: 674-678. 279. Place ES, George JH, Williams CK, Stevens MM (2009) Synthetic polymer scaffolds for tissue engineering. Chem Soc Rev 38: 1139-1151. 280. Clough RL (2001) High-energy radiation and polymers: A review of commercial processes and emerging applications. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 185: 8-33. 281. Ryan J (2008) Evolution of cell culture surfaces. BioFiles 3: 8,21. 282. Ho MH, Hou LT, Tu CY, Hsieh HJ, Lai JY, et al. (2006) Promotion of cell affinity of porous PLLA scaffolds by immobilization of RGD peptides via plasma treatment. Macromol Biosci 6: 90-98. 283. Morent R, De Geyter N, Desmet T, Dubruel P, Leys C (2011) Plasma Surface Modification of Biodegradable Polymers: A Review. Plasma Processes and Polymers 8: 171-190. 284. Morent R, De Geyter N, Gengembre L, Leys C, Payen E, et al. (2008) Surface treatment of a polypropylene film with a nitrogen DBD at medium pressure. The European Physical Journal - Applied Physics 43: 289-294. 285. Lane JM, Hourston DJ (1993) Surface treatments of polyolefins. Progress in Organic Coatings 21: 269-284. 286. Wang MJ, Chang YI, Poncin-Epaillard F (2005) Acid and basic functionalities of nitrogen and carbon dioxide plasma-treated polystyrene. Surface and Interface Analysis 37: 348-355. 287. Michl TD, Coad BR, Hüsler A, Vasilev K, Griesser HJ (2015) Laboratory Scale Systems for the Plasma Treatment and Coating of Particles. Plasma Processes and Polymers 12: 305-313. 288. Lee S-D, Hsiue G-H, Kao C-Y, Chang PC-T (1996) Artificial cornea: surface modification of silicone rubber membrane by graft polymerization of pHEMA via glow discharge. Biomaterials 17: 587-595. 289. Chan CM, Ko TM, Hiraoka H (1996) Polymer surface modification by plasmas and photons. Surface Science Reports 24: 3-54. 290. Akhavan B, Jarvis K, Majewski P (2013) Tuning the hydrophobicity of plasma polymer coated silica particles. Powder Technology 249: 403-411. 291. Toepke MW, Beebe DJ (2006) PDMS absorption of small molecules and consequences in microfluidic applications. Lab on a Chip 6: 1484-1486. 292. Christensen ME, Turner BE, Sinfield LJ, Kollar K, Cullup H, et al. (2010) Mesenchymal stromal cells transiently alter the inflammatory milieu post- transplant to delay graft-versus-host disease. Haematologica 95: 2102-2110. 293. Schaefer BC, Schaefer ML, Kappler JW, Marrack P, Kedl RM (2001) Observation of antigen-dependent CD8+ T-cell/ dendritic cell interactions in vivo. Cell Immunol 214: 110-122. 294. Regehr KJ, Domenech M, Koepsel JT, Carver KC, Ellison-Zelski SJ, et al. (2009) Biological implications of polydimethylsiloxane-based microfluidic cell culture. Lab Chip 9: 2132-2139. 295. Nold P, Brendel C, Neubauer A, Bein G, Hackstein H (2013) 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: 325-330.

References 135

296. Chen J-P, Lin C-T (2006) Dynamic seeding and perfusion culture of hepatocytes with galactosylated vegetable sponge in packed-bed bioreactor. Journal of Bioscience and Bioengineering 102: 41-45. 297. Kim SS, Sundback CA, Kaihara S, Benvenuto MS, Kim BS, et al. (2000) Dynamic seeding and in vitro culture of hepatocytes in a flow perfusion system. Tissue Eng 6: 39-44. 298. Zhu XH, Arifin DY, Khoo BH, Hua J, Wang C-H (2010) Study of cell seeding on porous poly(d,l-lactic-co-glycolic acid) sponge and growth in a Couette– Taylor bioreactor. Chemical Engineering Science 65: 2108-2117. 299. Timmins NE, Scherberich A, Fruh JA, Heberer M, Martin I, et al. (2007) Three- dimensional cell culture and tissue engineering in a T-CUP (tissue culture under perfusion). Tissue Eng 13: 2021-2028. 300. Wendt D, Marsano A, Jakob M, Heberer M, Martin I (2003) Oscillating perfusion of cell suspensions through three-dimensional scaffolds enhances cell seeding efficiency and uniformity. Biotechnology and Bioengineering 84: 205-214. 301. Li X, Ling W, Pennisi A, Wang Y, Khan S, et al. (2011) Human placenta-derived adherent cells prevent bone loss, stimulate bone formation, and suppress growth of multiple myeloma in bone. Stem Cells 29: 263-273. 302. Pountos I, Panteli M, Georgouli T, Giannoudis PV (2014) Do Mesenchymal Stem Cells Have a Role to Play in Cutaneous Wound Healing? Cell & Tissue Transplantation & Therapy 6: 11-17. 303. Cronk SM, Kelly-Goss MR, Ray HC, Mendel TA, Hoehn KL, et al. (2015) Adipose-Derived Stem Cells From Diabetic Mice Show Impaired Vascular Stabilization in a Murine Model of Diabetic Retinopathy. Stem Cells Translational Medicine 4: 459-467. 304. Elman JS, Murray RM, Wang F, Shen K, Gao S, et al. (2014) Pharmacokinetics of Natural and Engineered Secreted Factors Delivered by Mesenchymal Stromal Cells. PLoS ONE 9: e89882. 305. Brooke G, Rossetti T, Pelekanos R, Ilic N, Murray P, et al. (2009) Manufacturing of human placenta-derived mesenchymal stem cells for clinical trials. British Journal of Haematology 144: 571-579. 306. Babur B, Futrega K, Lott W, Klein T, Cooper-White J, et al. (2015) High- throughput bone and cartilage micropellet manufacture, followed by assembly of micropellets into biphasic osteochondral tissue. Cell and Tissue Research: 1-14. 307. Babur BK, Ghanavi P, Levett P, Lott WB, Klein T, et al. (2013) The Interplay between Chondrocyte Redifferentiation Pellet Size and Oxygen Concentration. PLoS ONE 8: e58865. 308. An Y, Wei W, Jing H, Ming L, Liu S, et al. (2015) Bone marrow mesenchymal stem cell aggregate: an optimal cell therapy for full-layer cutaneous wound vascularization and regeneration. Scientific reports. pp. 17036. 309. Gotherstrom C, West A, Liden J, Uzunel M, Lahesmaa R, et al. (2005) Difference in gene expression between human fetal liver and adult bone marrow mesenchymal stem cells. 1017-1026 p. 310. Al-Nbaheen M, vishnubalaji R, Ali D, Bouslimi A, Al-Jassir F, et al. (2013) Human Stromal (Mesenchymal) Stem Cells from Bone Marrow, Adipose Tissue and Skin Exhibit Differences in Molecular Phenotype and Differentiation Potential. Stem Cell Reviews and Reports 9: 32-43. 311. Strioga M, Viswanathan S, Darinskas A, Slaby O, Michalek J (2012) Same or Not the Same? Comparison of Adipose Tissue-Derived Versus Bone Marrow-

136 References

Derived Mesenchymal Stem and Stromal Cells. Stem Cells and Development 21: 2724-2752. 312. Kupcova Skalnikova H (2013) Proteomic techniques for characterisation of mesenchymal stem cell secretome. Biochimie 95: 2196-2211. 313. Wang J, Zhu HH, Xue JH, Wu SS, Chen Z (2015) Effects of storage conditions on the stability of serum CD163, NGAL, HMGB1 and MIP2. International Journal of Clinical and Experimental Pathology 8: 4099-4105. 314. Panicker G, Meadows KS, Lee DR, Nisenbaum R, Unger ER (2007) Effect of storage temperatures on the stability of cytokines in cervical mucous. Cytokine 37: 176-179. 315. Quesney S, Marvel J, Marc A, Gerdil C, Meignier B (2001) Characterization of Vero cell growth and death in bioreactor with serum-containing and serum-free media. Cytotechnology 35: 115-125. 316. Allen CB, Schneider BK, White CW (2001) Limitations to oxygen diffusion and equilibration in in vitro cell exposure systems in hyperoxia and hypoxia. L1021-L1027 p. 317. Fujita H, Yamamoto M, Ogino T, Kobuchi H, Ohmoto N, et al. (2014) Necrotic and apoptotic cells serve as nuclei for calcification on osteoblastic differentiation of human mesenchymal stem cells in vitro. Cell Biochemistry and Function 32: 77-86. 318. Schop D, Janssen FW, van Rijn LD, Fernandes H, Bloem RM, et al. (2009) Growth, metabolism, and growth inhibitors of mesenchymal stem cells. Tissue Eng Part A 15: 1877-1886. 319. Higuera G, Schop D, Janssen F, van Dijkhuizen-Radersma R, van Boxtel T, et al. (2009) Quantifying in vitro growth and metabolism kinetics of human mesenchymal stem cells using a mathematical model. Tissue Eng Part A 15: 2653-2663. 320. McElwain DLS, Ponzo PJ (1977) A model for the growth of a solid tumor with non-uniform oxygen consumption. Mathematical Biosciences 35: 267-279. 321. Miller WM, Blanch HW, Wilke CR (1988) A kinetic analysis of hybridoma growth and metabolism in batch and continuous suspension culture: effect of nutrient concentration, dilution rate, and pH. Biotechnol Bioeng 32: 947-965. 322. Lewis MC, Macarthur BD, Malda J, Pettet G, Please CP (2005) Heterogeneous proliferation within engineered cartilaginous tissue: the role of oxygen tension. Biotechnol Bioeng 91: 607-615. 323. Malda J, Rouwkema J, Martens DE, Le Comte EP, Kooy FK, et al. (2004) Oxygen gradients in tissue-engineered PEGT/PBT cartilaginous constructs: measurement and modeling. Biotechnol Bioeng 86: 9-18. 324. Coletti F, Macchietto S, Elvassore N (2006) Mathematical Modeling of Three- Dimensional Cell Cultures in Perfusion Bioreactors. Industrial & Engineering Chemistry Research 45: 8158-8169. 325. Shakeel M (2013) 2-D coupled computational model of biological cell proliferation and nutrient delivery in a perfusion bioreactor. Math Biosci 242: 86-94. 326. Zhao F, Pathi P, Grayson W, Xing Q, Locke BR, et al. (2005) Effects of oxygen transport on 3-d human mesenchymal stem cell metabolic activity in perfusion and static cultures: experiments and mathematical model. Biotechnol Prog 21: 1269-1280. 327. Greenspan H (1972) Models for the growth of a solid tumor by diffusion. Stud Appl Math 51: 317-340.

References 137

328. Chung CA, Chen CW, Chen CP, Tseng CS (2007) Enhancement of cell growth in tissue-engineering constructs under direct perfusion: Modeling and simulation. Biotechnology and Bioengineering 97: 1603-1616. 329. Shakeel M, Matthews PC, Graham RS, Waters SL (2011) A continuum model of cell proliferation and nutrient transport in a perfusion bioreactor. Mathematical Medicine and Biology. 330. Bear J (1988) Dynamics of fluids in porous media: Courier Dover Publications. 331. Craven S, Shirsat N, Whelan J, Glennon B (2013) Process model comparison and transferability across bioreactor scales and modes of operation for a mammalian cell bioprocess. Biotechnology Progress 29: 186-196. 332. Miller WM, Blanch HW, Wilke CR (1988) A kinetic analysis of hybridoma growth and metabolism in batch and continuous suspension culture: Effect of nutrient concentration, dilution rate, and pH. Biotechnology and Bioengineering 32: 947-965. 333. Zeng AP, Hu WS, Deckwer WD (1998) Variation of stoichiometric ratios and their correlation for monitoring and control of animal cell cultures. Biotechnol Prog 14: 434-441. 334. Pattappa G, Heywood HK, de Bruijn JD, Lee DA (2011) The metabolism of human mesenchymal stem cells during proliferation and differentiation. J Cell Physiol 226: 2562-2570. 335. Polinkovsky M, Gutierrez E, Levchenko A, Groisman A (2009) Fine temporal control of the medium gas content and acidity and on-chip generation of series of oxygen concentrations for cell cultures. Lab Chip 9: 1073-1084. 336. Kim M-C, Lam RW, Thorsen T, Asada HH (2013) Mathematical analysis of oxygen transfer through polydimethylsiloxane membrane between double layers of cell culture channel and gas chamber in microfluidic oxygenator. Microfluidics and Nanofluidics 15: 285-296. 337. Lanfrey PY, Kuzeljevic ZV, Dudukovic MP (2010) Tortuosity model for fixed beds randomly packed with identical particles. Chemical Engineering Science 65: 1891-1896. 338. Green D, Perry R (2007) Perry's Chemical Engineers' Handbook, Eighth Edition: McGraw-Hill Education. 339. Baltz JM, Biggers JD (1991) Oxygen transport to embryos in microdrop cultures. Molecular Reproduction and Development 28: 351-355. 340. Kleinstreuer C, Agarwal SS (1986) Analysis and simulation of hollow-fiber bioreactor dynamics. Biotechnology and Bioengineering 28: 1233-1240. 341. Øyaas J, Storrø I, Svendsen H, Levine DW (1995) The effective diffusion coefficient and the distribution constant for small molecules in calcium- alginate gel beads. Biotechnology and Bioengineering 47: 492-500. 342. Sarkar S, Bustard BL, Welter JF, Baskaran H (2011) Combined experimental and mathematical approach for development of microfabrication-based cancer migration assay. Ann Biomed Eng 39: 2346-2359. 343. Donato D, Napoli IED, Catapano G (2014) Model-Based Optimization of Scaffold Geometry and Operating Conditions of Radial Flow Packed-Bed Bioreactors for Therapeutic Applications. Processes 2: 34-57. 344. Mohebbi-Kalhori D, Behzadmehr A, Doillon C, Hadjizadeh A (2012) Computational modeling of adherent cell growth in a hollow-fiber membrane bioreactor for large-scale 3-D bone tissue engineering. Journal of Artificial Organs 15: 250-265.

138 References

345. Campolo M, Curcio F, Soldati A (2013) Minimal perfusion flow for osteogenic growth of mesenchymal stem cells on lattice scaffolds. AIChE Journal 59: 3131-3144. 346. Andrade PZ, de Soure AM, dos Santos F, Paiva A, Cabral JMS, et al. (2013) Ex vivo expansion of cord blood haematopoietic stem/progenitor cells under physiological oxygen tensions: clear-cut effects on cell proliferation, differentiation and metabolism. Journal of Tissue Engineering and Regenerative Medicine: n/a-n/a. 347. Wendt D, Marsano A, Jakob M, Heberer M, Martin I (2003) Oscillating perfusion of cell suspensions through three-dimensional scaffolds enhances cell seeding efficiency and uniformity. Biotechnol Bioeng 84: 205-214. 348. Kitagawa T, Yamaoka T, Iwase R, Murakami A (2006) Three-dimensional cell seeding and growth in radial-flow perfusion bioreactor for in vitro tissue reconstruction. Biotechnology and Bioengineering 93: 947-954. 349. Mei Y, Goldberg M, Anderson D (2007) The development of high-throughput screening approaches for stem cell engineering. Current opinion in chemical biology 11: 388-393. 350. Anselme K, Linez P, Bigerelle M, Le Maguer D, Le Maguer A, et al. (2000) The relative influence of the topography and chemistry of TiAl6V4 surfaces on osteoblastic cell behaviour. Biomaterials 21: 1567-1577. 351. Ponsonnet L, Reybier K, Jaffrezic N, Comte V, Lagneau C, et al. (2003) Relationship between surface properties (roughness, wettability) of titanium and titanium alloys and cell behaviour. Materials Science and Engineering: C 23: 551-560. 352. Lavenus S, Pilet P, Guicheux J, Weiss P, Louarn G, et al. (2011) Behaviour of mesenchymal stem cells, fibroblasts and osteoblasts on smooth surfaces. Acta Biomaterialia 7: 1525-1534. 353. Jacobs T, Morent R, De Geyter N, Dubruel P, Leys C (2012) Plasma Surface Modification of Biomedical Polymers: Influence on Cell-Material Interaction. Plasma Chemistry and Plasma Processing 32: 1039-1073. 354. Canavan HE, Cheng X, Graham DJ, Ratner BD, Castner DG (2006) A Plasma- Deposited Surface for Cell Sheet Engineering: Advantages over Mechanical Dissociation of Cells. Plasma Processes and Polymers 3: 516-523. 355. Akiyama Y, Kikuchi A, Yamato M, Okano T (2004) Ultrathin Poly(N- isopropylacrylamide) Grafted Layer on Polystyrene Surfaces for Cell Adhesion/Detachment Control. Langmuir 20: 5506-5511. 356. Kilian KA, Bugarija B, Lahn BT, Mrksich M (2010) Geometric cues for directing the differentiation of mesenchymal stem cells. Proceedings of the National Academy of Sciences 107: 4872-4877. 357. Guilak F, Cohen DM, Estes BT, Gimble JM, Liedtke W, et al. (2009) Control of Stem Cell Fate by Physical Interactions with the Extracellular Matrix. Cell Stem Cell 5: 17-26. 358. McMurray RJ, Gadegaard N, Tsimbouri PM, Burgess KV, McNamara LE, et al. (2011) Nanoscale surfaces for the long-term maintenance of mesenchymal stem cell phenotype and multipotency. Nat Mater 10: 637-644. 359. Lutolf MP, Gilbert PM, Blau HM (2009) Designing materials to direct stem-cell fate. Nature 462: 433-441. 360. Kolind K, Leong KW, Besenbacher F, Foss M (2012) Guidance of stem cell fate on 2D patterned surfaces. Biomaterials 33: 6626-6633.

References 139

361. Oh S, Brammer KS, Li YSJ, Teng D, Engler AJ, et al. (2009) Stem cell fate dictated solely by altered nanotube dimension. Proceedings of the National Academy of Sciences 106: 2130-2135. 362. Dokoumetzidis A, Papadopoulou V, Macheras P (2006) Analysis of Dissolution Data Using Modified Versions of Noyes–Whitney Equation and the Weibull Function. Pharmaceutical Research 23: 256-261. 363. DiMilla PA, Barbee K, Lauffenburger DA (1991) Mathematical model for the effects of adhesion and mechanics on cell migration speed. Biophysical Journal 60: 15-37. 364. Webb K, Hlady V, Tresco PA (2000) Relationships among cell attachment, spreading, cytoskeletal organization, and migration rate for anchorage- dependent cells on model surfaces. Journal of Biomedical Materials Research 49: 362-368. 365. Vunjak-Novakovic G, Obradovic B, Martin I, Bursac PM, Langer R, et al. (1998) Dynamic Cell Seeding of Polymer Scaffolds for Cartilage Tissue Engineering. Biotechnology Progress 14: 193-202. 366. Zhao F, Chella R, Ma T (2007) Effects of shear stress on 3-D human mesenchymal stem cell construct development in a perfusion bioreactor system: Experiments and hydrodynamic modeling. Biotechnology and Bioengineering 96: 584-595. 367. Sheng H, Wang Y, Jin Y, Zhang Q, Zhang Y, et al. (2008) A critical role of IFN[gamma] in priming MSC-mediated suppression of T cell proliferation through up-regulation of B7-H1. Cell Res 18: 846-857. 368. Yamasaki S, Mera H, Itokazu M, Hashimoto Y, Wakitani S (2014) Cartilage Repair With Autologous Bone Marrow Mesenchymal Stem Cell Transplantation: Review of Preclinical and Clinical Studies. Cartilage 5: 196- 202. 369. Smits PC, Nienaber C, Colombo A, Ince H, Carlino M, et al. (2006) Myocardial repair by percutaneous cell transplantation of autologous skeletal myoblast as a stand alone procedure in post myocardial infarction chronic heart failure patients. EuroIntervention 1: 417-424. 370. Busser H, Najar M, Raicevic G, Pieters K, Velez Pombo R, et al. (2015) Isolation and characterization of human mesenchymal stromal cells subpopulations: comparison of bone marrow and adipose tissue. Stem Cells Dev. 371. Liu H, Kemeny DM, Heng BC, Ouyang HW, Melendez AJ, et al. (2006) The Immunogenicity and Immunomodulatory Function of Osteogenic Cells Differentiated from Mesenchymal Stem Cells. The Journal of Immunology 176: 2864-2871. 372. Ankrum JA, Ong JF, Karp JM (2014) Mesenchymal stem cells: immune evasive, not immune privileged. Nat Biotech 32: 252-260. 373. Lorenz K, Sicker M, Schmelzer E, Rupf T, Salvetter J, et al. (2008) Multilineage differentiation potential of human dermal skin-derived fibroblasts. Experimental Dermatology 17: 925-932. 374. Salinas Tejedor L, Berner G, Jacobsen K, Gudi V, Jungwirth N, et al. (2015) Mesenchymal stem cells do not exert direct beneficial effects on CNS remyelination in the absence of the peripheral immune system. Brain Behav Immun. 375. De Jesus M, Wurm FM (2011) Manufacturing recombinant proteins in kg-ton quantities using animal cells in bioreactors. European Journal of Pharmaceutics and Biopharmaceutics 78: 184-188.

140 References

376. Takahashi M, Matsue H, Matsushita M, Nakajima Y, Uchino J (1993) Isolation and Culture of Human Hepatocytes from Resected Liver Tissue as a Bioreactor for a Hybrid Artificial Liver. Artificial Organs 17: 653-659. 377. Ohshima N, Yanagi K, Miyoshi H (1997) Packed-Bed Type Reactor to Attain High Density Culture of Hepatocytes for Use as a Bioartificial Liver. Artificial Organs 21: 1169-1176. 378. Sarika PR, Sidhy Viha CV, Sajin Raj RG, Nirmala RJ, Anil Kumar PR (2015) A non-adhesive hybrid scaffold from gelatin and gum Arabic as packed bed matrix for hepatocyte perfusion culture. Materials Science and Engineering: C 46: 341-347. 379. Schairer DO, Chouake JS, Nosanchuk JD, Friedman AJ (2012) The potential of nitric oxide releasing therapies as antimicrobial agents. Virulence 3: 271-279. 380. McDonnell G, Russell AD (1999) Antiseptics and Disinfectants: Activity, Action, and Resistance. Clinical Microbiology Reviews 12: 147-179. 381. Michl TD, Coad BR, Doran M, Husler A, Valentin JDP, et al. (2014) Plasma polymerization of 1,1,1-trichloroethane yields a coating with robust antibacterial surface properties. RSC Advances 4: 27604-27606. 382. Michl TD, Coad BR, Doran M, Osiecki M, Kafshgari MH, et al. (2015) Nitric oxide releasing plasma polymer coating with bacteriostatic properties and no cytotoxic side effects. Chem Commun (Camb) 51: 7058-7060. 383. Ferreira MV, Jahnen-Dechent W, Neuss S (2011) Standardization of Automated Cell-Based Protocols for Toxicity Testing of Biomaterials. Journal of Biomolecular Screening 16: 647-654. 384. Hanks CT, Wataha JC, Sun Z (1996) In vitro models of biocompatibility: A review. Dental Materials 12: 186-193. 385. Kurata S, Morishita K, Kawase T, Umemoto K (2011) Cytotoxic effects of polybasic acids, poly(alkenoic acid)s, and the monomers with various functional groups on human pulp fibroblasts. Dental Materials Journal 30: 904- 912. 386. Rabe M, Verdes D, Seeger S (2011) Understanding protein adsorption phenomena at solid surfaces. Advances in Colloid and Interface Science 162: 87-106. 387. Jiang X, Jeffries R, Acosta M, Tikunov A, Macdonald J, et al. (2015) Biocompatibility of Tygon® tubing in microfluidic cell culture. Biomedical Microdevices 17: 1-7. 388. Kinsella AR, Haran MS (1991) Decreasing Sensitivity to Cytotoxic Agents Parallels Increasing Tumorigenicity in Human Fibroblasts. Cancer Research 51: 1855-1859. 389. Zhu Y, Li W, Li Q, Li Y, Li Y, et al. (2009) Effects of serum proteins on intracellular uptake and cytotoxicity of carbon nanoparticles. Carbon 47: 1351- 1358. 390. Brunelli L, Crow JP, Beckman JS (1995) The comparative toxicity of nitric oxide and peroxynitrite to Escherichia coli. Arch Biochem Biophys 316: 327-334. 391. Yang J, Rose FRAJ, Gadegaard N, Alexander MR (2009) A High-Throughput Assay of Cell-Surface Interactions using Topographical and Chemical Gradients. Advanced Materials 21: 300-304. 392. Voth DE, Hamm EE, Nguyen LG, Tucker AE, Salles II, et al. (2005) Bacillus anthracis oedema toxin as a cause of tissue necrosis and cell type-specific cytotoxicity. Cellular Microbiology 7: 1139-1149.

References 141

393. Sohaebuddin S, Thevenot P, Baker D, Eaton J, Tang L (2010) Nanomaterial cytotoxicity is composition, size, and cell type dependent. Particle and Fibre Toxicology 7: 22. 394. Coulter JA, Jain S, Butterworth KT, Taggart LE, Dickson GR, et al. (2012) Cell type-dependent uptake, localization, and cytotoxicity of 1.9 nm gold nanoparticles. Int J Nanomedicine 7: 2673-2685. 395. Furley A, Reeves B, Mizutani S, Altass L, Watt S, et al. (1986) Divergent molecular phenotypes of KG1 and KG1a myeloid cell lines. 1101-1107 p. 396. Liu L-L, Long Z-J, Wang L-X, Zheng F-M, Fang Z-G, et al. (2013) Inhibition of mTOR Pathway Sensitizes Acute Myeloid Leukemia Cells to Aurora Inhibitors by Suppression of Glycolytic Metabolism. Molecular Cancer Research 11: 1326-1336. 397. Yao X, Liu Y, Gao J, Yang L, Mao D, et al. (2015) Nitric oxide releasing hydrogel enhances the therapeutic efficacy of mesenchymal stem cells for myocardial infarction. Biomaterials 60: 130-140. 398. Li N, Lu X, Zhao X, Xiang F-L, Xenocostas A, et al. (2009) Endothelial Nitric Oxide Synthase Promotes Bone Marrow Stromal Cell Migration to the Ischemic Myocardium via Upregulation of Stromal Cell-Derived Factor-1α. STEM CELLS 27: 961-970. 399. Mujoo K, Krumenacker JS, Murad F (2011) Nitric oxide-cyclic GMP signaling in stem cell differentiation. Free Radic Biol Med 51: 2150-2157.

142 References

Appendices

Appendix 1

Biological testing of plasma modified surfaces

As part of our collaboration with Prof. Hans Griesser and Thomas Michl, from the University of South Australia and to further understand plasma surface modification. A method was developed to characterise two types of antibacterial surfaces, which were obtained from our collaborators, who performed the plasma surface modification of the bioreactor scaffolds. These surfaces released either nitric oxide (NOX) or chlorine (TCE) to inhibit bacterial growth. The main mechanisms of the antibacterial properties of nitric oxide is through the damage of DNA by reactive nitric oxide species that were generated by the autoxidation of nitric oxide, inhibition of DNA repair mechanisms, and increase generation of alkylating agents and hydrogen peroxide [379]. Chlorine has long been used as an antiseptic and disinfectant, through its action as an oxidising agent [380]. For potential medical application these surfaces must be assessed for human cell toxicity. Some of the data presented here, have been published in two of their papers [381,382].

As each cell type and application design is different, all modified surfaces must be tested for their biological properties and toxicities to ensure that they are suitable for the specific purpose. Cell adherence, cell morphology and cell growth in adherent cell lines are standard testing parameters [383-385]. Of course, these are not suitable for an application that demands non-adherent surfaces. In addition, it is difficult to correlate an observed reduced growth rate on a surface with surface characteristics that are known to either influence cell growth and behaviour or exhibit toxicity [260,349- 353]. Standards used to evaluate these surfaces must also address the potential depletion of essential media components, such as growth factors, proteins or lipids that are required for cell growth, which could be absorbed by the scaffold [386,387]. Finally, the use of serum supplemented media can complicate the analysis by either

Appendices 143

contributing to or inhibiting the cytotoxic effects of the material being tested [388,389].

In our system, a non-adherent myeloid leukemic cell line KG1a grown in a fully defined serum-free medium (X-VIVO 15) was used to evaluate the toxicity of a test surface independently of cell adhesion and complications due to serum. To test the biocompatibility of the surfaces, the cells were cultured in either direct contact with the test surface, cultured with the test surface but without direct cell contact, or in a 24-hour conditioned medium away from the test surface. In addition to ethanol as a toxicity standard, a standard dilution of PBS was used to determine if any medium components were depleted. This pre-screening process determined if an observed reduced growth rate or cell depth on a test surface resulted from a toxic effect or a surface chemistry difference when compared to traditional tissue culture plastic. For surfaces that passed this screening, primary human bmMSCs were subsequently used to study the growth and morphology of adherent cells growing on such surfaces.

Methods

Cell Culture

144 Appendices

then incubated in 20% oxygen, 5% CO2 at 37 ⁰C for 24 hours to allow adherent cells to attach. The medium was then replaced with fresh medium and the flask was placed into a hypoxic incubator (2% oxygen and 5% CO2 at 37 ⁰C). The medium was replaced twice a week until the monolayer was 80-90% confluent before passaging.

Kg1a myeloid leukaemia cell line was sourced from ATCC and was maintained in RPMI (Invitrogen) supplemented with 10% fetal bovine serum and 100 units/mL penicillin, and 100 µg/mL streptomycin (referred to RPMI 10% FBS from here on) in a T25 flask and incubated at standard cell culture conditions (5% CO2 at 37 ⁰C). The medium was changed twice weekly and the cells were passaged when they reached 80-90% confluent.

Surface Preparation

Isopentyl nitrate that produces nitric oxide upon contact with water and is sufficiently volatile to be plasma polymerised onto a polyethylene terephthalate surface to give rise to the NOX surface. 1,1,1-Trichlorethane can be plasma - - polymerised due to volatility and release Cl and Cl2 ion in contact with water. The surfaces were either unwashed (dry), rinsed with 1 ml of distilled water for 15 minutes (wet) or repeating the wet process 20 times (washed). The method for plasma modification of the surfaces has been previously described by others [381,382]. The surface composition was unknown when the biological testing was performed.

MSC Cytotoxicity Assay

Passage 4 MSCs were seeded at 2000 cells/cm2 suspended in DMEM 10% FBS in a 24 well plate that had the surface modified (n=4). For the controls the 2000 cell/cm2 were seeded into a normal tissue culture treated 24 well plate suspended in DMEM 10% FBS for the positive growth control (n=4) and serum free DMEM for the negative control (n=4). For the non-surface contact controls, 2000 cells/cm2 were seeded in DMEM 10% FBS into a normal tissue culture plate but with the surface

Appendices 145

modified insert placed vertically into the well (n=4) and 2000 cell/cm2 were seeded in conditioned DMEM 10% FBS that has been in contact with the modified surface for

24 hours in an incubator (5% CO2 at 37 ⁰C) (n=4). The cells were allowed to grow for

4 days in standard tissue culture conditions (37⁰C and 5% CO2). The cells were quantified by the metabolic probe AlamarBlue (Invitrogen). The medium was replaced with a 1:50 dilution of AlamarBlue in DMEM 10% FBS and incubated for 3 hours. The fluorescence signal of the AlamarBlue in media was measured in a plate reader (FLUOstar Omega, BMG Labtech, Germany) along with a cell titration to infer cell numbers using excitation and emission filters of 544 and 590.

KG1a Cytotoxicity Assay

Kg1a cells were seeded at 20,000 cells/ml suspended in serum free media X- VIVO 15 (Lonza) in a 24 well plate that had the surface modified (n=4). For the controls the 20,000 cells/ml were seeded into a normal tissue culture treated 24 well plate suspended in X-VIVO 15 for the positive growth control (n=4) and serum free RPMI for the negative control (n=4). For the non-contact controls, 20,000 cells/ml were seeded in X-VIVO 15 into a normal tissue culture plate but with the surface modified insert placed vertically into the well (n=4) and were seeded in conditioned X-VIVO 15 that had been in contact with the modified surface for 24 hours (n=4). The cells were grown for 4 days under standard tissue culture conditions (37⁰C and 5%

CO2). The cells were counted by flow cytometry (FC500, Beckman Coulter, USA) using Flow-CountTM Fluorospheres (Beckman Coulter) with 3 µM propidium iodide to gate out the dead cells.

Cytotoxicity Standards

The MSCs 2000 cells/cm2 suspended in DMEM 10% FBS in a 24 well plate were titrated with PBS and ethanol (200 proof, >99.5%) to give a dilution and toxicity curve. For the KG1a cells, the cells were seeded at 20,000 cells/ml suspended in X- VIVO 15 or RPMI 10% FBS and were titrated with PBS and ethanol. The cells were

146 Appendices

allowed to grow for 4 days and the cell numbers were quantified by the AlamarBlue method for the MSCs and by flow cytometry for the KG1a cells.

MSC Morphology on Modified Surfaces

The medium was removed and the cells washed before being fixed with 4% paraformide in PBS for 10 minutes and then washed with PBS. The MSCs were permeabilised by 0.1% TRITON X-100 (Invitrogen) in PBS and washed again in PBS. The cells were stained with 50 µg/ml of phalloidin (Invitrogen) solution in 1% bovine serum albumin (BSA, Sigma) and PBS for 20 minutes at room temperature. The cells were washed again and stained with 5 µg/ml DAPI for 5 minutes. The cells were then washed with PBS and imaged with an ECLIPSE Ti epifluorescent microscope (Nikon, Japan).

Statistics

Refer to section 2.1.15.

Appendices 147

Results

Cytotoxicity Standards

KG1a cell growth was first compared in different media types to demonstrate that XVIVO-15 medium did not adversely affect growth. No difference was observed between XVIVO-15 and the standard RPMI 10% FBS, even when XVIVO-15 was supplemented with FBS (Figure 30). However, non-supplemented RPMI did not support KG1a cell growth.

600,000

500,000

400,000

300,000

(cells/ml) 200,000

100,000

Viable Viable KG1a concentration * 0 RPMI 10% RPMI XVIVO 15 XVIVO 15 FBS +10% FBS

Figure 30 KG1a cell growth after four days with different medium types. * indicates P<0.01.

Increasing concentrations of ethanol were used as toxicity standards. For the KG1a cell standard, a difference in the level of toxicity between the two media types used was observed, with the ethanol having a greater toxic effect at lower concentration in the XVIVO 15 with a AC50 of 0.275 %vol/vol compared to the serum- containing medium with a AC50 of 3x10-10 %vol/vol (Figure 31A). The cell type used is also an important consideration, as bmMSCs are far more resistant to the toxic effects of ethanol than KG1a cells with an AC50 of 1.55 % vol/vol (Figure 31 C). PBS dilution reduced the cell growth rate, but did not induce cell death, thus demonstrating that it is a suitable model for nutrient depletion (Figure 31 B&D). However, there was no significant reduction observed in final KG1a cell number until 80% vol/vol dilution of the medium with PBS.

148 Appendices

A B

350,000 350,000 RPMI 10% FBS RPMI 10% FBS 300,000 300,000 XVIVO 15 XVIVO 15 250,000 250,000 200,000 200,000 150,000 150,000

concentration(cells/ml) 100,000 100,000

50,000 50,000 KG1a KG1a

0 0 Viable Viable KG1a concentration(cells/ml) Viable Viable 0.0% 0.2% 0.4% 0.6% 0.8% 1.0% 0% 20% 40% 60% 80% 100% Ethanol (% vol/vol) PBS (% vol/vol) C D

14,000 14000

)

2 2 12,000 12000 10,000 10000 8,000 8000 6,000 6000 4,000 4000

2,000 2000 Viable MSCdensity (cells/cm Viable MSCdensity (cells/cm 0 0 0% 2% 4% 6% 8% 10% 0% 20% 40% 60% 80% 100% Ethanol in DMEM 10% FBS (% vol/vol) PBS in DMEM 10% FBS (% vol/vol)

Figure 31 (A) KG1a day 4 cell concentration exposed to ethanol in RPMI 10% FBS or XVIVO 15. (B) KG1a day 4 cell concentration after being diluted with PBS in RPMI 10% FBS or XVIVO 15. (C) bmMSC cell density after 4 days growth in DMEM 10% FBS exposed to ethanol. (D) bmMSC cell density after 4 days growth in DMEM 10% FBS diluted with PBS.

Appendices 149

Toxicity of Chlorine Surfaces

A B 350000 350000

300000 300000

250000 250000

200000 200000

150000 150000 * 100000 100000

50000 * * 50000

0 0

Viable Viable ConcentrationKG1a (cells/ml) Viable Viable KG1a Concentration(cells/ml)

) 18000 2 16000 14000 12000 10000 8000 * 6000 4000

2000 Viable (Cells/cm Viable density MSC 0

Figure 32 (A) KG1a growth after 4 days on two different TCE surfaces n=4. (B) KG1a growth after 4 days on a TCE surfaces that have been unwashed (Dry), washed and dry (washed), and washed and remained wet (wet), the cells were either in contact or grown in conditioned media n=4. (C) Human bmMSC grown in contact or in conditioned media on surfaces that have been unwashed (dry), rinsed briefly (wet) or washed by rigorous method (washed) n=4. * indicates P<0.01.

The surfaces being tested were plasma polymerised 1,1,1-trichloroethane (TCE) onto polystyrene 48 well plates. For the non-contact surface, the TCE was polymerised onto a surface of polyethylene terephthalate (PET) [381]. Surfaces were tested with the KG1a cells that underwent a variety of different vapour pressures, input powers and treatment times. All surfaces showed toxicity to KG1a cells in all surface types compared to KG1a cells grown in XVIVO 15 on commercial TCP. Figure 32A is an example of such results of direct surface contact. The surfaces were washed to

150 Appendices

eliminate the possibility of unreacted polymer leaching from the surface into the medium. However, this had no effect on the toxicity for either KG1a cells or human bmMSCs, as both in contact and the conditioned media killed the cells (Figure 32 B&C).

Biocompatibility of Nitric Oxide Releasing Surfaces

Isopentyl nitrite was plasma polymerised to untreated 48 well polystyrene tissue culture plates and PET coverslips. This surface releases NO upon contact with water, which has a toxic effect on bacteria [382,390]. This surface is non-cytotoxic to human cells, and KG1a cells shows no difference in cells grown on the NOX surfaces compared to cells grown on tissue culture plastic (Figure 33 A). A decrease in cell density for bmMSCs grown on the NOX surface was observed (Figure 33 B). However, no difference in cell density was observed when the cells were grown on commercial TCP and exposed to, but do not come in contact with, the NOX surfaces or with the conditioned medium. This suggests that the observed reduced growth rate of the bmMSCs was due to surface effects, but not to cytotoxic effects. No difference in the morphology of the bmMSC growing on either the NOX surfaces (Figure 33 D&E) or commercial TCP (Figure 33 F&G) was observed. However, when the NOX surfaces were modified to have increased NO released, a significant reduction in the growth rate of the bmMSCs was noted for cells grown on the surface, no contact with surface or the conditioned media (Figure 33 C). This suggests that NO has an effect on the bmMSCs at higher concentrations. Although these modified NOX surfaces exhibited enhanced antibacterial properties, the surface type used in Figure 33 A&B was considered to be the most suitable for medical applications and still had adequate antibacterial properties [382].

Appendices 151

A B

500,000 16000 ) 2 14000 400,000 12000 10000 300,000 * 8000 200,000 6000 100,000 4000 2000

0 0

Viable Viable MSC density(cells/cm Viable Viable KG1a concentration(cells/ml) C

) * 2 16000 * 14000 12000 10000 8000 6000 4000 2000

0 Viable Viable MSC density(Cells/cm

D E

100 µm 100 µm F G

100 µm 100 µm Figure 33 KG1a growth after 4 days exposed to NOX surface n=4 (A). Human bmMSC growth after 4 days exposed to NOX surface n=4 (B). Human bmMSC growth after 4 days on different types NOX surfaces that have stronger antibacterial properties n=4 (C). Human bmMSC growing on NOX surface (D & E) and tissue culture plastic (F & G) stained with phalloidin to show actin filaments (red) and DAPI to stain the nucleus (blue). * indicates P<0.01.

152 Appendices

Discussion

The correct controls and method to test the cytotoxicity of biomaterials and modified surfaces is important for the development of functionalised surfaces in bioreactors to satisfactorily control culture conditions [275,283,288]. A method that tests for cytotoxicity independently of cell adhesion is important to establish if an observed reduced growth rate resulted from either surface effects or from toxicity of the biomaterial or surface [260,349-353]. Non-adherent cells allow for greater flexibility for high throughput methods for cytotoxicity testing using flow cytometry [349,383,391]. Because serum can mask cytotoxicity, an appropriate serum-free medium is necessary [388,389]. In our experiments (in Figure 31A), the serum- containing RPMI medium masked the toxic effect of ethanol up to 0.25 %vol/vol, while a toxic effect could be observed at 0.1%vol/vol in the serum-free media XVIVO 15. The cell type was also important for detecting cytotoxicity; the KG1a cells being far more sensitive to ethanol compared to bmMSCs (Figure 31 A&C). This has been observed with other cytotoxic materials and chemicals, and demonstrates the importance of pre-screening. It also highlights the importance of testing and optimising the bioreactor surface for the cell type that will be actually growing in the bioreactor [392-394].

In addition to the cytotoxicity standard, an absorption or deficit of important media components standard was assayed by diluting the medium with PBS. Although complete cell death was not observed, the growth rate was reduced in both the KG1a cells and bmMSC as a result of the reduced available nutrients and growth factors (Figure 31 B&D). The bmMSC showed a steady decrease in cell density, resulting in half the final cell numbers at 80% vol/vol dilution compared with the positive control. However, no significant reduction in the KG1a cell number was evident until 80% vol/vol dilution with PBS (Figure 31B). It should be noted that KG1a is an immortalised cell line that can grow in the absence of growth factors, therefore its growth is limited only by nutrient concentration [395]. The uptake rate of nutrients is higher in KG1a cells compared to bmMSCs, for example glucose uptake rate is 346 fmol/cell/hr for KG1a cells and 200 fmol/cell/hr for bmMSCs [334,396]. In addition, more KG1a cells per well were observed compared to bmMSCs, resulting in a higher

Appendices 153

consumption of nutrients and other medium components that could result in a reduced growth rate.

These methods were also used to test TCE plasma polymerised surfaces, which release chlorine and have an anti-bacterial effect [287]. The TCE surfaces were cytotoxic under all conditions and cell types (Figure 32). Although the TCE surface cannot be used clinically, this surface modification might still prove useful for non- fouling piping or as an antibacterial coating on surgical gowns, masks and other personal protection equipment.

The KG1a cells grown on the NOX surface displayed no reduction in the final cell number (Figure 33A), indicating that the surface was non-cytotoxic. However, when the bmMSCs were grown on the NOX surface there was a significant reduction in the final cell number, compared to bmMSCs grown on commercial TCP that were exposed to either the conditioned medium or grown in the presence of, but not in contact with, the NOX surface (Figure 33B). However, there was no difference in the morphology of the cells grown on the NOX surface (Figure 33 D, E, F &G) relative to TCP, suggesting that the reduction in cell number of the bmMSCs grown on the NOX surface was not a result of toxicity. Surface chemistry is known to affect cell growth and could account for the reduction of bmMSC growth rate [260,349-353]. In addition, nitric oxide can affect stem cell fate by inducing mobilisation and differentiation of MSCs [115,397-399]. The stronger NOX surface in Figure 33C might have reduced the bmMSC cell growth by signalling the bmMSC to alter the cell fate. In summary, the NOX surface is non-cytotoxic and is suitable for clinical applications. However, cell function alteration through nitric oxide signalling must be considered when using this surface.

The method described here for evaluating material surfaces is effective for distinguishing between true toxic effects and other phenomena, such as changing cell fate through nitric oxide. This will be important for further development of bioreactor scaffolds to either add functionality to influence cell fate or to subsequently attach selection molecules to the surface to help isolate a preferred cell type [265,268,269].

154 Appendices

In addition, the reduce growth of bmMSCs on the NOX surface in comparison to traditional tissue culture plastic, highlights the importance of the surface characteristics on cell growth that has to be optimised for each cell type grown in a bioreactor.

Appendices 155

Appendix 2

Difference in media type and conditioning time on HaCaT growth

To determine the most suitable media type and to optimise the method for producing conditioned media, HaCaT cell growth was measured with different media types and different lengths of media conditioning. After 4 days, pMSC and bmMSC XVIVO 15 conditioned medium yielded greater growth characteristics comparing the XVIVO 15 negative control. No effect of the conditioning time was observed on overall cell growth (Figure 34A). However, the pMSC and bmMSC DMEM conditioned medium only increased cell growth over the negative control after 30 hours of conditioning. Also significantly greater growth was achieved in the XVIVO 15 control medium compared to the DMEM conditioned medium.

After 6 days of HaCaT growth in the condition medium, only after 24 hours conditioning the XVIVO 15 medium with pMSCs did the HaCaT number increase was significantly higher than the XVIVO 15 negative control. While it took 30 hours for the pMSC DMEM conditioned medium to be greater than the DMEM negative control. For the bmMSC condition medium, all time points for the XVIVO 15 conditioned medium resulted in an increase in HaCaT growth after 6 days compared to the XVIVO 15 negative control. However, the DMEM bmMSC conditioned medium had no effect on HaCaT growth regardless of how long the medium was conditioned (Figure 34B). From this result it was concluded that XVIVO 15 was the best medium to use, firstly because it is a fully defined medium that not only contains glucose, vitamins and minerals but also includes fats as compared to DMEM that is a very basic medium requiring the addition of serum to support MSC growth. Secondly, it appears MSCs are in a better state to secrete any growth factors when grown in the presence of XVIVO with a far better impact on the HaCaT growth. It does appear that 30 hours of conditioning results in the best growth in all cases, we opted to use 48 hours as it is an easier time point to manage and it was the conditioning time of other publications [101,110,175,238,312].

156 Appendices

A ** **

45,000

) 40,000 # 2 35,000 # 30,000 # 25,000 # 20,000 #

density(cells/cm 15,000 * 10,000 *

HaCaT * 5,000 * 0

** ** B

200,000 #

) 180,000 2 160,000 # # 140,000 120,000 100,000 80,000 * density(cells/cm 60,000 * 40,000

HaCaT 20,000 0

Figure 34 HaCaT cell density in response to pMSC and bmMSC for different conditioning times and medium types at (A) day 4 and (B) day 6. * indicates cell density significantly greater (p<0.01) over the DMEM negative control, # indicates significantly greater (p<0.01) over XVIVO 15 control and ** indicates a significate difference (p<0.01) between XVIVO 15 and DMEM media types.

Appendices 157

Appendix 3

Repeated conditioned medium experiments

Repeat experiments where performed using different donors and different batches of conditioned medium (Figure 35 A & B). However, the results were inconsistent with the degree of HaCaT growth improvement was highly variable between the experiments.

160,000 ) 2 140,000 120,000 100,000 * * 80,000 * * * * * 60,000 40,000 *

20,000 HaCaT density(cells/cm 0

140,000 ) 2 120,000 100,000 80,000 60,000 40,000 20,000 * * *

HaCaT HaCaT density(cells/cm * * * * * * * * * * * * 0

Figure 35 (A & B) HaCaT growth in response to pMSC and bmMSC conditioned medium from different donors and cell aggregates two different experimental replicates after 6 days. * Significant (p<0.01) increase over XVIVO 15 negative control.

158 Appendices

Appendix 4

Mesh refinement, difference in final average cell density (cell/cm2) at different mesh sizes (cm)

A mesh refinement study was performed to find convergence of the result while not using excessive computing time. The maximum mesh size was reduced from 0.1cm to 0.02 cm in order to reduce the error of the approximate solution. The difference between the average cell different from 0.02 cm mesh size and the large mesh size was graphed to find where the result was convergent below the acceptable tolerance of 0.5 cell/cm2. A mesh size of 0.05 cm was the point where the solution converged within the acceptable tolerance and computing time.

-1 ) ) density

2 -2 -3 -4 -5 -6 -7 -8 -9 -10

Difference in final densitycell (cells/cm Difference -11 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 Max mesh size (cm)

Appendices 159

Appendix 5

Understanding Axis symmetric models and heat map figures. Flux direction arrows: the black arrow shows the direction of diffusive flux which is towards the centre of the reactor, the white arrow represents the combined flux of the diffusion and convection as the medium is pumped through the reactor. Maximum concentration present in reactor

Colour Scale for nutrient concentration: blue being low concentration and dark red being high concentration

Minimum concentration present in As the wall has a higher solubility of oxygen than the reactor medium in the reactor there needs to be two scales for the oxygen concentration: one for the wall and the other for the medium in the reactor.

A 2D axial symmetric model give an approximation of 3D object by rotation around the axis. This requires less computing power to solve the equations compared to a full 3D model.

160 Appendices

Appendix 6

Average cell density of the base model

Heat map of the cell density within the bioreactor calculated from the base model for every two days. It shows the development of the non-uniform cell gradient from the inlet to the outlet which increases over time.

Cell density day (cells/cm2) day 2

Appendices 161

Cell density day (cells/cm2) day 4

Cell density (cells/cm2) day 6

162 Appendices

Cell density (cells/cm2) day 8

Appendices 163

Appendix 7

Oxygen concentration (mM) in reactor vessel and PDMS wall of the base model. White arrows represent direction of combined diffusive and convective flux and the black arrow represents the direction of the diffusive flux.

Oxygen concentration (mM) day 2

Oxygen concentration (mM) day 4

164 Appendices

Oxygen concentration (mM) day 6

Oxygen concentration (mM) day 8

Appendices 165

Appendix 8

Glucose concentrate (mM) for the base model.

Glucose concentration (mM) day 2

Glucose concentration (mM) day 4

166 Appendices

Glucose concentration (mM) day 6

Glucose concentration (mM) day 8

Appendices 167

Appendix 9

Lactate concentration (mM) of the base model.

Lactate concentration (mM) day 2

Lactate concentration (mM) day 4

168 Appendices

Lactate concentration (mM) day 6

Lactate concentration (mM) day 8

Appendices 169

Appendix 10

Changing wall thickness on oxygen concentration

Oxygen concentration on day 7 in the centre of the bioreactor.

170 Appendices

Oxygen concentration at the wall interface

Appendices 171

Appendix 11

Shear stress with varying flow rates.

From the shear force was calculated from the Darcy flow as described in 4.1.2 Shear Stress. The base model of 0.208 ml/hr was the flow rate used in the small scale experiments.

0.0007

0.0006

0.0005

0.0004

0.0003

Shearforce (Pa) 0.0002

0.0001

0 Base 0.5 ml/hr 1 ml/hr 1.5 ml/hr 3 ml/hr 4 ml/hr model- 0.208 ml/hr

172 Appendices

Appendix 12

Glucose and oxygen concentration based on different flow rates in the wall diffusion bioreactor.

30 ml/hr Glucose concentration day 7

30 ml/hr Oxygen concentration day 7

Appendices 173

20ml/hr Glucose concentration day 7

20 ml/hr Oxygen concentration day 7

174 Appendices

15 ml/hr Glucose concentration day 7

15 ml/hr Oxygen concentration day 7

Appendices 175

10 ml/hr Glucose concentration day 7

10 ml/hr Oxygen concentration day 7

176 Appendices

Appendix 13

Figure 1 Approval

Our Ref: LA/IEBT/P6456

19 February 2016

Dear Michael Osiecki,

Material requested: Figure 1 from Jin Wang, Lianming Liao & Jianming Tan (2011) Mesenchymal-stem-cellbased experimental and clinical trials: current status and open questions, Expert Opinion on Biological Therapy, 11:7, 893-909

Thank you for your correspondence requesting permission to reproduce the above mentioned material from our Journal in your printed thesis entitled ‘Isolation And Expansion of Placental Derived Mesenchymal Stromal Cells in A Packed Bed Bioreactor’ and to be posted in your university’s repository – Queensland University of Technology.

We will be pleased to grant entirely free permission on the condition that you acknowledge the original source of publication and insert a reference to the Journal’s web site: http://www.tandfonline.com

Please note that this licence does not allow you to post our content on any third party websites or repositories.

Thank you for your interest in our Journal.

Yours sincerely

Lee-Ann

Lee-Ann Anderson – Permissions & Licensing Administrator, Journals Routledge, Taylor & Francis Group 3 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN, UK. Tel: +44 (0)20 7017 7932 Fax:+44 (0)20 7017 6336 Web: www.tandfonline.com e-mail: [email protected]

Taylor & Francis is a trading name of Informa UK Limited, registered in England under no. 1072954

Appendices 177

Figure 2 Approval

JOHN WILEY AND SONS LICENSE TERMS AND CONDITIONS Jan 20, 2016 This Agreement between Michael Osiecki ("You") and John Wiley and Sons ("John Wiley and Sons") consists of your license details and the terms and conditions provided by John Wiley and Sons and Copyright Clearance Center. License Number 3792880057369 License date Jan 20, 2016 Licensed Content Publisher John Wiley and Sons Licensed Content Publication Transfusion Licensed Content Title Mesenchymal stem or stromal cells: a review of clinical applications and manufacturing practices Licensed Content Author Ratti Ram Sharma,Kathryn Pollock,Allison Hubel,David McKenna Licensed Content Date Oct 16, 2013 Pages 20 Type of use Dissertation/Thesis Requestor type University/Academic Format Print and electronic Portion Figure/table Number of figures/tables 1 Original Wiley figure/table number(s) figure 2 Will you be translating? No Title of your thesis / dissertation ISOLATION AND EXPANSION OF PLACENTAL DERIVED MESENCHYMAL STROMAL CELLS IN A PACKED BED BIOREACTOR Expected completion date Mar 2016 Expected size (number of pages) 200 Requestor Location Michael Osiecki 2 George St Brisbane, Australia 4000 Attn: Michael Osiecki Billing Type Invoice Billing Address Michael Osiecki 2 George St Brisbane, Australia 4000 Attn: Michael Osiecki Total 0.00 AUD Terms and Conditions TERMS AND CONDITIONS This copyrighted material is owned by or exclusively licensed to John Wiley & Sons, Inc. orone of its group companies (each a"Wiley Company") or handled on behalf of a society with which a Wiley Company has exclusive publishing rights in relation to a particular work

178 Appendices

(collectively "WILEY"). By clicking "accept" in connection with completing this licensing transaction, you agree that the following terms and conditions apply to this transaction (along with the billing and payment terms and conditions established by the Copyright Clearance Center Inc., ("CCC's Billing and Payment terms and conditions"), at the time that you opened your RightsLink account (these are available at any time at http://myaccount.copyright.com). Terms and Conditions The materials you have requested permission to reproduce or reuse (the "Wiley Materials") are protected by copyright. You are hereby granted a personal, nonexclusive, nonsub licensable (on a standalone basis), nontransferable, worldwide, limited license to reproduce the Wiley Materials for the purpose specified in the licensing process. This license, and any CONTENT (PDF or image file) purchased as part of your order, is for a onetime use only and limited to any maximum distribution number specified in the license. The first instance of republication or reuse granted by this license must be completed within two years of the date of the grant of this license (although copies prepared before the end date may be distributed thereafter). The Wiley Materials shall not be used in any other manner or for any other purpose, beyond what is granted in the license. Permission is granted subject to an appropriate acknowledgement given to the author, title of the material/book/journal and the publisher. You shall also duplicate the copyright notice that appears in the Wiley publication in your use of the Wiley Material. Permission is also granted on the understanding that nowhere in the text is a previously published source acknowledged for all or part of this Wiley Material. Any third party content is expressly excluded from this permission. With respect to the Wiley Materials, all rights are reserved. Except as expressly

Appendices 179

granted by the terms of the license, no part of the Wiley Materials may be copied, modified, adapted (except for minor reformatting required by the new Publication), translated, reproduced, transferred or distributed, in any form or by any means, and no derivative works may be made based on the Wiley Materials without the prior permission of the respective copyright owner.For STM Signatory Publishers clearing permission under the terms of the STM Permissions Guidelines only, the terms of the license are extended to include subsequent editions and for editions in other languages, provided such editions are for the work as a whole in situ and does not involve the separate exploitation of the permitted figures or extracts, You may not alter, remove or suppress in any manner any copyright, trademark or other notices displayed by the Wiley Materials. You may not license, rent, sell, loan, lease, pledge, offer as security, transfer or assign the Wiley Materials on a standalone basis, or any of the rights granted to you hereunder to any other person. The Wiley Materials and all of the intellectual property rights therein shall at all times remain the exclusive property of John Wiley & Sons Inc, the Wiley Companies, or their respective licensors, and your interest therein is only that of having possession of and the right to reproduce the Wiley Materials pursuant to Section 2 herein during the continuance of this Agreement. You agree that you own no right, title or interest in or to the Wiley Materials or any of the intellectual property rights therein. You shall have no rights hereunder other than the license as provided for above in Section 2. No right, license or interest to any trademark, trade name, service mark or other branding ("Marks") of WILEY or its licensors is granted hereunder, and you agree that you 1/20/2016 RightsLink Printable License https://s100.copyright.com/AppDispatchServlet 3/5 shall not assert any such right, license or interest with respect thereto NEITHER WILEY NOR ITS LICENSORS MAKES ANY WARRANTY OR REPRESENTATION OF ANY KIND TO YOU OR ANY THIRD PARTY, EXPRESS, IMPLIED OR STATUTORY, WITH RESPECT TO THE MATERIALS

180 Appendices

OR THE ACCURACY OF ANY INFORMATION CONTAINED IN THE MATERIALS, INCLUDING, WITHOUT LIMITATION, ANY IMPLIED WARRANTY OF MERCHANTABILITY, ACCURACY, SATISFACTORY QUALITY, FITNESS FOR A PARTICULAR PURPOSE, USABILITY, INTEGRATION OR NONINFRINGEMENT AND ALL SUCH WARRANTIES ARE HEREBY EXCLUDED BY WILEY AND ITS LICENSORS AND WAIVED BY YOU. WILEY shall have the right to terminate this Agreement immediately upon breach of this Agreement by you. You shall indemnify, defend and hold harmless WILEY, its Licensors and their respective directors, officers, agents and employees, from and against any actual or threatened claims, demands, causes of action or proceedings arising from any breach of this Agreement by you. IN NO EVENT SHALL WILEY OR ITS LICENSORS BE LIABLE TO YOU OR ANY OTHER PARTY OR ANY OTHER PERSON OR ENTITY FOR ANY SPECIAL, CONSEQUENTIAL, INCIDENTAL, INDIRECT, EXEMPLARY OR PUNITIVE DAMAGES, HOWEVER CAUSED, ARISING OUT OF OR IN CONNECTION WITH THE DOWNLOADING, PROVISIONING, VIEWING OR USE OF THE MATERIALS REGARDLESS OF THE FORM OF ACTION, WHETHER FOR BREACH OF CONTRACT, BREACH OF WARRANTY, TORT, NEGLIGENCE, INFRINGEMENT OR OTHERWISE (INCLUDING, WITHOUT LIMITATION, DAMAGES BASED ON LOSS OF PROFITS, DATA, FILES, USE, BUSINESS OPPORTUNITY OR CLAIMS OF THIRD PARTIES), AND WHETHER OR NOT THE PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. THIS LIMITATION SHALL APPLY NOTWITHSTANDING ANY FAILURE OF ESSENTIAL PURPOSE OF ANY LIMITED REMEDY PROVIDED HEREIN. Should any provision of this Agreement be held by a court of competent jurisdiction to be illegal, invalid, or unenforceable, that provision shall be deemed amended to

Appendices 181

achieve as nearly as possible the same economic effect as the original provision, and the legality, validity and enforceability of the remaining provisions of this Agreement shall not be affected or impaired thereby. The failure of either party to enforce any term or condition of this Agreement shall not constitute a waiver of either party's right to enforce each and every term and condition of this Agreement. No breach under this agreement shall be deemed waived or excused by either party unless such waiver or consent is in writing signed by the party granting such waiver or consent. The waiver by or consent of a party to a breach of any provision of this Agreement shall not operate or be construed as a waiver of or consent to any other or subsequent breach by such other party. This Agreement may not be assigned (including by operation of law or otherwise) by you without WILEY's prior written consent. Any fee required for this permission shall be nonrefundable after thirty (30) days 1/20/2016 RightsLink Printable License https://s100.copyright.com/AppDispatchServlet 4/5 from receipt by the CCC. These terms and conditions together with CCC's Billing and Payment terms and conditions (which are incorporated herein) form the entire agreement between you and WILEY concerning this licensing transaction and (in the absence of fraud) supersedes all prior agreements and representations of the parties, oral or written. This Agreement may not be amended except in writing signed by both parties. This Agreement shall be binding upon and inure to the benefit of the parties' successors, legal representatives, and authorized assigns. In the event of any conflict between your obligations established by these terms and conditions and those established by CCC's Billing and Payment terms and conditions, these terms and conditions shall prevail. WILEY expressly reserves all rights not specifically granted in the combination of (i) the license details provided by you and accepted in the course of this licensing

182 Appendices

transaction, (ii) these terms and conditions and (iii) CCC's Billing and Payment terms and conditions. This Agreement will be void if the Type of Use, Format, Circulation, or Requestor Type was misrepresented during the licensing process. This Agreement shall be governed by and construed in accordance with the laws of the State of New York, USA, without regards to such state's conflict of law rules. Any legal action, suit or proceeding arising out of or relating to these Terms and Conditions or the breach thereof shall be instituted in a court of competent jurisdiction in New York County in the State of New York in the United States of America and each party hereby consents and submits to the personal jurisdiction of such court, waives any objection to venue in such court and consents to service of process by registered or certified mail, return receipt requested, at the last known address of such party. WILEY OPEN ACCESS TERMS AND CONDITIONS Wiley Publishes Open Access Articles in fully Open Access Journals and in Subscription journals offering Online Open. Although most of the fully Open Access journals publish open access articles under the terms of the Creative Commons Attribution (CC BY) License only, the subscription journals and a few of the Open Access Journals offer a choice of Creative Commons Licenses. The license type is clearly identified on the article. The Creative Commons Attribution License

The Creative Commons Attribution License (CCBY) allows users to copy, distribute and transmit an article, adapt the article and make commercial use of the article. The CCBY license permits commercial and nonCreative Commons Attribution NonCommercial License The Creative Commons Attribution NonCommercial (CCBYNC) License permits use,

Appendices 183

distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.(see below) Creative Commons AttributionNonCommercialNoDerivs License The Creative Commons Attribution NonCommercialNoDerivs License (CCBYNCND) permits use, distribution and reproduction in any medium, provided the original work is properly cited, is not used for commercial purposes and no modifications or adaptations are made. (see below) Use by commercial "forprofit" organizations 1/20/2016 RightsLink Printable License https:/

184 Appendices

Figure 3 Approval

JOHN WILEY AND SONS LICENSE TERMS AND CONDITIONS Jan 20, 2016 This Agreement between Michael Osiecki ("You") and John Wiley and Sons ("John Wiley and Sons") consists of your license details and the terms and conditions provided by John Wiley and Sons and Copyright Clearance Center. License Number 3792880565341 License date Jan 20, 2016 Licensed Content Publisher John Wiley and Sons Licensed Content Publication Journal of Chemical Technology & Biotechnology Licensed Content Title Design of bioreactors for mesenchymal stem cell tissue engineering Licensed Content Author Pankaj Godara,Clive D McFarland,Robert E Nordon Licensed Content Date Feb 18, 2008 Pages 13 Type of use Dissertation/Thesis Requestor type University/Academic Format Print and electronic Portion Figure/table Number of figures/tables 1 Original Wiley figure/table number(s) figure 3 Will you be translating? No Title of your thesis / dissertation ISOLATION AND EXPANSION OF PLACENTAL DERIVED MESENCHYMAL STROMAL CELLS IN A PACKED BED BIOREACTOR Expected completion date Mar 2016 Expected size (number of pages) 200 Requestor Location Michael Osiecki 2 George St Brisbane, Australia 4000 Attn: Michael Osiecki Billing Type Invoice Billing Address Michael Osiecki 2 George St Brisbane, Australia 4000 Attn: Michael Osiecki TERMS AND CONDITIONS This copyrighted material is owned by or exclusively licensed to John Wiley & Sons, Inc. or 1/20/2016 RightsLink Printable License https://s100.copyright.com/AppDispatchServlet 2/5 one of its group companies (each a"Wiley Company") or handled on behalf of a society with which a Wiley Company has exclusive publishing rights in relation to a particular work

Appendices 185

186 Appendices

Appendices 187

188 Appendices

Appendices 189

190 Appendices

and is not used for commercial purposes.(see below) Creative Commons AttributionNonCommercialNoDerivs License The Creative Commons Attribution NonCommercialNoDerivs License (CCBYNCND) permits use, distribution and reproduction in any medium, provided the original work is properly cited, is not used for commercial purposes and no modifications or adaptations are made. (see below) Use by commercial "forprofit" organizations 1/20/2016 RightsLink Printable License https://s100.copyright.com/AppDispatchServlet 5/5 Use of Wiley Open Access articles for commercial, promotional, or marketing purposes requires further explicit permission from Wiley and will be subject to a fee. Further details can be found on Wiley Online Library http://olabout.wiley.com/WileyCDA/Section/id410895. html Other Terms and Conditions: v1.10 Last updated September 2015 Questions? [email protected] or +18552393415 (toll free in the US)

Appendices 191

Appendices 193