UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Process development for the production of pancreatic islet equivalents Luu, Minh 2004

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2004-0128.pdf [ 6.49MB ]
Metadata
JSON: 831-1.0058660.json
JSON-LD: 831-1.0058660-ld.json
RDF/XML (Pretty): 831-1.0058660-rdf.xml
RDF/JSON: 831-1.0058660-rdf.json
Turtle: 831-1.0058660-turtle.txt
N-Triples: 831-1.0058660-rdf-ntriples.txt
Original Record: 831-1.0058660-source.json
Full Text
831-1.0058660-fulltext.txt
Citation
831-1.0058660.ris

Full Text

PROCESS DEVELOPMENT FOR T H E PRODUCTION OF PANCREATIC ISLET EQUIVALENTS  by MINH L U U  B.A.Sc, University of Waterloo, 2001  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTERS OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES Department of Chemical and Biological Engineering and The Biotechnology Laboratory  We accept this thesis as conforming to the required standard  The University of British Columbia February 2004 © Minh Luu, 2004  Library Authorization  In presenting this thesis in partial fulfillment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission.  Name of Author  Title of Thesis:  Degree:  Date (dd/mm/yyyy)  (please print)  flQ-fl  pfO£&£D  I>MJO^MJUA^  Year:  . 3D,  Department of L%MW,idjjL cUil The University of British Columbia Vancouver, BC Canada  Jh^^jjUL,  ifiteWi&dL  ~U  d-<OD f L  Q^ueJUd** "  U  fitjj  f^hfrJj^/^if /r  {j^^roj^s^^  ABSTRACT The Edmonton Protocol has shown that islet transplantation can reverse type-1 diabetes, but there are not enough donors available to make it a widespread application. The purpose of this work was to apply techniques, such as design of experiments and bioreactor culture methods, to facilitate islet derivation from pancreatic stem cells. Neonatal porcine pancreatic cell clusters (NPCCs) were used as a model system and a two-step approach of expansion followed by differentiation was taken. When cultured in suspension, total cell numbers decreased over time and a factorial experiment involving the addition of EGF, VEGF, KGF and HGF did not yield any significant improvements. A dose response experiment indicated that the tested concentrations of 50 ng/mL was higher than necessary.  The heterogeneity of the cultures required that specific  populations be examined. Based upon studies suggesting that ductal cells contained islet progenitors, ductal cells were selected as a target population to expand. When NPCCs were dissociated into a single cell suspension and cultured in monolayers, the proportion of cells expressing CK7 (a ductal cell marker) increased to 95% 9 days after isolation but decreased to 39% by day 12. To elucidate which growth factors might sustain CK7+ cell proliferation, a factorial analysis of EGF, VEGF, KGF, HGF and bFGF was performed. HGF had the most promising effects in that it promoted both cell attachment and CK7+ cell proliferation.  The maturation of NPCCs was also examined.  Insulin/DNA of  NPCCs cultured in suspension increased 2-fold whereas culturing NPCCs in alginate and 5% NPS increased insulin/DNA even further (4-fold compared to controls). The existing forms of alginate immobilization are either not readily scalable (slabs) or require a generation unit (beads). A hollow-fiber bioreactor (HFBR) was investigated as a novel  ii  system for generating -100 mL alginate cultures, which would meet the 10 islets 6  required per patient. The alginate was loaded into the extracapillary space and gelled by passing a Ca  2+  solution through the intracapillary space (ICS). High ICS flow rates (300  mL/min) reduced alginate plugging of the fibers as assessed by residence time distribution analysis.  Cell recoveries after de-gelling were >90%.  A cell growth  experiment with CHO cells demonstrated that the HFBR provides a rapid, closed means of culturing alginate-immobilized cells at a patient scale.  iii  TABLE OF CONTENTS  Abstract  ii  Table of Contents  iv  List of Figures  vi  List of Tables  vii  Nomenclature  .  Acknowledgements 1  ix xi  Introduction  1  2  Background 3 2.1 The Pancreas 3 2.2 Pancreatic Stem Cell Candidates 5 2.2.1 Evidence of a Pancreatic Progenitor in Post-natal Animals 5 2.2.2 Nestin Positive Cells 6 2.2.3 Pdx-1 Positive Cells 7 2.2.4 Ductal Cells 8 2.3 Strategies in Developing Pancreatic Stem Cell Cultures 9 2.4 Neonatal Pancreatic Porcine Cell Clusters (NPCCs) As A Model System 13 2.4.1 NPCC Suspension Culture 13 2.4.2 NPCC Monolayer Culture 14 2.4.3 Maturation of NPCCs in Alginate and Autologous Serum 15 2.5 Cell Immobilization Methods 16 2.5.1 Immobilization of Mammalian Cells 16 2.5.2 Alginate 16 2.5.3 Systems for Alginate Immobilization 17 2.5.4 Hollow-Fibre Bioreactors 17  3  Materials and Methods 3.1 NPCC Cultures and Analysis 3.1.1 Isolation and Shipment 3.1.2 Seeding and Sampling Technique of Suspension Cultures 3.1.3 Dissociation and Technique for Monolayer Cultures 3.1.4 Cell Enumeration : 3.1.5 Insulin Analysis 3.1.6 Cell Dissociation and Fixation for Immunocytochemistry Slides 3.1.7 Immunocytochemistry Staining Procedure 3.1.8 Cell Counting and Image Processing for Immunocytochemistry 3.1.9 Suspension Culture Experiments 3.1.10 Monolayer Culture Experiments 3.1.11 Alginate Cultures iv  20 20 20 20 21 22 22 23 24 25 26 28 28  3.2 HFBR Cultures 3.2.1 CHO KI Cells 3.2.2 Apparatus and Setup 3.2.3 HFBR Alginate Loading, Gelling and Degelling 3.2.4 HFBR Experiments 3.3 Statistical Analysis  4  Results and Discussion  29 29 29 30 32 34  36  4.1 NPCC Growth In Suspension Culture 36 4.1.1 Assessment of NPCC Survival in Suspension Culture 36 4.1.2 Growth Factor Effects on NPCC Survival in Suspension Cultures 38 4.1.3 Serum vs. Serum Substitutes and Insulin vs. IGF-1 41 4.2 NPCC Growth In Monolayer Culture 43 4.2.1 Assessment of NPCC Survival in Monolayer Culture 43 4.2.2 Analysis of CK7+ Cells in NPCC Monolayer Culture 44 4.2.3 Targeted Expansion of CK7+ Cells Using Growth Factors 45 4.3 (3-CellMaturation in NPCCs 57 4.3.1 Assessment of Insulin Content of NPCCs in Suspension Culture 58 4.3.2 Growth Factor Effects on Insulin/DNA of NPCCs in Suspension Culture.. 58 4.3.3 Alginate and Serum Effects on NPCC Maturation 60 4.4 Alginate Immobilization in HFBR 61 4.4.1 Use of CHO KI Cells to Model Alginate Immobilization of Cells in HFBRs 62 4.4.2 Determination of Alginate Plugging 62 4.4.3 Degree of Gelling 66 4.4.4 Cell Recovery 66 4.4.5 Cell Growth in Alginate in the HFBR 68  5  Conclusions and Recommendations  71  6  References  75  7  Appendix  80  v  LIST O F F I G U R E S  Figure 2.1: Organization of Tissues in the Pancreas Figure 2.2: Pancreatic Organogenesis Figure 2.3: Process for Immobilizing Cells in Alginate Beads or Slabs Figure 2.4: Schematic of HFBR Figure 3.1: Representative Images Used in Determining CK7 and Ki67 Staining Figure 3.2: Setup of HFBR System for Cell Growth Experiment Figure 4.1: Cell Concentration Profile of NPCC Suspension Culture. Figure 4.2: Cell Losses in NPCC Culture, Shipment Effects Figure 4.3: Effects of Additional Medium Change on NPCC Survival Figure 4.4: Effects of Using Supernatants From Previous Cultures on NPCC Survival. Figure 4.5: NPCC Dose Response Curves of EGF and HGF Figure 4.6: Serum, Serum-Free and BIT Supplemented Effects on NPCC Survival Figure 4.7: Comparison of IGF-1 and Insulin Effects on NPCC Survival Figure 4.8: Cell Concentration Profile of NPCCs in Monolayer Culture Figure 4.9: CK7 Staining in Monolayer Cultures Figure 4.10: Interaction Plots for Day 4 Supernatant Cell Concentration Figure 4.11: Residual Plot for Day 7 Monolayer Cell Concentration Figure 4.12: Typical Insulin Content Profile in NPCC Cultures Figure 4.13: Effect of Alginate and NPS on Insulin/DNA of NPCCs Figure 4.14: Comparison of CHO Cells Grown In Alginate Figure 4.15: Effect of Rinsing Time on Azoalbumin Elution Curves Figure 4.16: Azoalbumin Elution Curves Before and After Loading Alginate Figure 4.17: Cumulative Cell Recovery After Each Degelling Pass Figure 4.18: Total Cell Recovery at Different Alginate Concentrations Figure 4.19: Glucose and Lactate Profiles in Alginate Immobilized HFBR and Slab Cultures Figure 4.20: pH Profile in Alginate Immobilized HFBR Culture Figure 4.21: Cell Concentration in HFBR Cell Growth Experiment  vi  3 4 18 19 27 30 36 37 38 38 40 42 42 43 44 47 49 58 61 62 64 65 67 68 69 70 70  LIST OF TABLES Table 2.1: Reported Mitogens for Pancreatic Cell Culture 10 Table 2.2: Reported Differentiating Factors for Pancreatic Cell Culture 12 Table 3.1: Antibodies Used in Immunocytochemistry 24 Table 3.2: Operating Conditions of Different HFBR Runs 31 Table 4.1: Factorial Analysis of Day 4 NPCC Survival in Suspension 39 Table 4.2: Factorial Analysis of Day 8 NPCC Survival in Suspension 39 Table 4.3: CK7+ Cells in Monolayer Cultures 44 Table 4.4: Factorial Analysis of Day 4 Supernatant Cell Concentration 46 Table 4.5: Factorial Analysis of Day 4 to 7 Supernatant Cell Expansion 47 Table 4.6: Factorial Analysis of Monolayer Total Cell Concentration 48 Table 4.7: Factorial Analysis of Monolayer Total Cell Concentration, Reduced Model 49 Table 4.8: Factorial Analysis of Proportion of CK7+ Cells in Day 7 Monolayers 50 Table 4.9: Factorial Analysis of Proportion of Ki67+ Cells in Day 7 Monolayers 51 Table 4.10: Factorial Analysis of Proportion of Ki67+ Cells in Day 7 Monolayers, Reduced Model 51 Table 4.11: Factorial Analysis of Proportion of CK7+ Cells Also Expressing Ki67 in Day 7 Monolayers 52 Table 4.12: Factorial Analysis of Day 7 Concentration of CK7+ Cells in Monolayers.. 53 Table 4.13: Factorial Analysis of Day 7 Concentration of CK7+ Cells in Monolayers, Reduced Model 54 Table 4.14: Factorial Analysis of Day 4 Insulin/DNA of NPCC Suspension Cultures... 59 Table 4.15: Factorial Analysis of Day 8 Insulin/DNA of NPCC Suspension Cultures... 59 Table 4.16: Effect of Alginate on Average Residence Times of Azoalbumin in HFBR. 66 Table 7.1: Error in Seeding Technique 80 Table 7.2: Comparison of Nuclei Counting to Trypan Blue Method 80 Table 7.3: Design and Data for Growth Factor Experiment in Suspension Culture 80 Table 7.4: ANOVA for Model in Table 4.1 81 Table 7.5: ANOVA for Model in Table 4.2 81 Table 7.6: Immunocytochemistry Images Counts - Total Cell Number 81 Table 7.7: Immunocytochemistry Images Counts - CK7+ Cells 82 Table 7.8: Immunocytochemistry Images Counts - Ki67 + Cells 82 Table 7.9: Immunocytochemistry Images Counts - Double Stained Cells 83 Table 7.10: Selection of Intensity Threshold for Ki67 Threshold 83 Table 7.11: Summary of Counts 84 Table 7.12: Design and Data for Growth Factor Experiment in Monolayer Culture 85 Table 7.13: ANOVA for Model in Table 4.4 86 Table 7.14: ANOVA for Model in Table 4.5 86 Table 7.15: ANOVA for Model in Table 4.6 87 Table 7.16: ANOVA for Model in Table 4.7 87 Table 7.17: ANOVA for Model in Table 4.8 88 Table 7.18: ANOVA for Model in Table 4.9 88 Table 7.19: ANOVA for Model in Table 4.10 89 Table 7.20: ANOVA for Model in Table 4.11 89 Table 7.21: ANOVA for Model in Table 4.12 90  vii  Table 7.22: Table 7.23: Table 7.24: Table 7.25:  ANOVA for Model in Table 4.13 90 Prediction for CK7+ Cell Concentration and % Proliferating CK7+ Cells.. 91 ANOVA for Model in Table 4.14 91 ANOVA for Model in Table 4.15 92  viii  NOMENCLATURE  A  total cross sectional area of the fibres  ABC  avidin-biotinylated enzyme complex  ANOVA  analysis of variance  BCEM  bovine corneal endothelial cell matrix  bFGF  basis fibroblast growth factor  BIT  bovine serum albumin, insulin, transferrin  BrDU  5-bromo-2-deoxyuridine  BSA  bovine serum albumin  CHO  Chinese hamster ovary  CK  cytokeratin  DAB  diaminobenzidine  ECM  extracellular matrix  ECS  extracapillary space  EGF  epidermal growth factor  FBS  fetal bovine serum, same as FCS  FCS  fetal calf serum, same as FBS  FGF10  fibroblast growth factor 10  Flk-1  PTK fetal liver kinase-1  GLP-1  glucagon-like peptide 1  HFBR  hollowfibrebioreactor  HBSS  Hank's balanced salt solution  HGF  hepatocyte growth factor  ix  HTB-9  human tumour bladder-9 cells  ICS  intracapillary space  IGF-1  insulin-like growth factor 1  KGF  keratinocyte growth factor  L  length of the fibres  Ngn-3  neurogenin-3  NGS  normal goat serum  NPCC  neonatal pancreatic cell clusters  NPS  neonatal pig serum  PBS  phosphate buffered saline  Pdx-1  pancreatic and duodenal homeobox gene-1  Q  flow rate  RT-PCR  reverse transcriptase polymerase chain reaction  TGF  transforming growth factor  tR,ave  average residence time  v  ave  VEGF  average velocity vascular endothelial growth factor  X  ACKNOWLEDGEMENTS  First, I would like to thank my supervisor, Dr. Jamie Piret, for allowing me to work on such a challenging and interesting project and for giving much guidance and support. I thank Dr. Greg Korbutt for providing us with the cells and information that was necessary to start. His feedback on this manuscript was also very useful. I also thank Dr. Bruce Bowen for his feedback on this manuscript. I am also very appreciative of the assistance that I received from Cale Street and Corinne Hoesli with some of the analytical work. I would like to thank the Natural Sciences and Engineering Research Council of Canada for their financial support. It was not always an easy road, but I am glad that I finished. I hope that I am able to lay at least one brick in the long road to curing diabetes. This work is dedicated to my long list of supporters which include Jamin, Luong, Thanh, Suly, my parents, Clive, Hans, Volker, Jay, Leah, Christine, and Sumi.  xi  1  INTRODUCTION Diabetes Mellitus is a disease characterized by elevated levels of glucose in the  blood and insulin is a hormone that regulates glucose absorption in cells. Type 1 diabetes occurs when the insulin producing cells in the pancreatic islets are destroyed, either through an autoimmune attack or infection, whereas type 2 diabetes occurs mainly when cells become insulin resistant.  According to Health Canada and the World Health  Organization, approximately 2 million Canadians and 150 million people worldwide have diabetes. Type 1 diabetes comprises only 10% of cases, but it tends to strike at an earlier age and requires treatment with frequent insulin injections. Most type 2 diabetes can be treated with diet modification and drugs, although 20 to 30% of type 2 diabetics also require insulin injections (Lechner et al., 2003). These treatments are not a cure and tight glucose control is difficult to achieve. Long term complications associated with diabetes include high blood pressure, stroke, blindness, loss of limbs, neuropathy and kidney failure. A  potential  cure  for  diabetes  is  islet  transplantation  immunosuppression to prevent rejection of the donor tissue.  coupled  with  At the University of  Alberta, researchers were able to reverse type 1 diabetes in seven out of seven patients by injecting approximately 10 islet equivalents per kg into their liver portal veins and using 4  a novel regime of glucocorticoid-free immunosuppressants (Shapiro et al., 2000). Follow-up studies on what is now called the Edmonton Protocol report that 87% of the 35 patients who have since undergone this procedure remain insulin-independent for at least 1 year and for as long as 3 years (Burridge et al., 2002).  1  Despite this recent success, there are major barriers to making islet transplantation a routine treatment for diabetes. Immunosuppressive drugs have many side effects and other methods of preventing rejection are not yet available. On average 2 cadaveric pancreata are required per patient to reverse hyperglycemia.  Even if islet isolation  methods are optimized, the current donor supply only meets approximately 0.5% of the demand (Burridge et al., 2002). Therefore, multiple avenues of generating more tissue are being explored including xenotransplantation - the use of organs from other species (Rayat et al., 1999), genetic engineering (Cheung et al., 2000), embryonic stem cells (Lumelsky et al., 2001) and adult stem cells (Lechner and Habener, 2003). A stem cell has the ability to both self-renew and differentiate into other cell types.  Embryonic stem cells are derived from the inner cell mass of a blastocyst,  whereas adult stem cells persist in the body. Although the therapeutic potential of stem cells is enormous, their identity and mechanisms of self-renewal and differentiation are not fully understood. There also are many ethical issues surrounding the use of stem cells, especially embryonic stem cells. The purpose of this project was to apply bioprocessing techniques, such as design of experiments and bioreactor culture methods, to facilitate pancreatic stem cell research and development.  Neonatal pancreatic porcine cell clusters (NPCCs) were used as a  model system and alginate immobilization in a hollow-fibre bioreactor (HFBR) was investigated as a system for culturing these cells.  2  BACKGROUND  2 2.1  T h e Pancreas  The pancreas is composed of acinar, ductal and endocrine cells (Figure 2.1). The acinar, or exocrine, cells comprise 85 to 90% of the pancreas and produce digestive enzymes, which are secreted into the gut. The ductal cells line the main and intercalated ducts of the pancreas and make up about 5% of the pancreas. Endocrine cells comprise the remainder of the pancreas and are organized into clusters called Islets of Langerhans. There are approximately 10 islets per adult pancreas, and each islet consists of a, (3, 5 6  and PP cells which produce glucagon, insulin, somatostatin and pancreatic-polypeptide, respectively. 60 to 80% of the cells in an islet are (3-cells (Edlund, 2002), but as much as 15% of (3-cells are single cells scattered throughout the exocrine tissue (Bouwens, 1998b).  Figure 2.1: Organization of Tissues in the Pancreas  Adapted from arbl.cvmbs.colostate.edu/.../' pancreas/histo_endo.html and Edlund ( 2002). During embryogenesis, the pancreas actually develops from two regions, dorsal and ventral, in the primitive foregut that bud and fuse into one organ. A sheet of endoderm cells in each region committed to a pancreatic fate (Figure 2.2A) invaginates into the embryonic connective tissue (Figure 2.2A), or mesenchyme, and branches off  creating ducts in the process (Figure 2.2C). Acinar and endocrine cells form around these ducts, and endocrine cells migrate further into the exocrine mass (Figure 2.2D). All functional pancreatic tissue is of endoderm origin; however, it has been shown that interaction with mesenchyme is required for development.  When cultured alone,  pancreatic extracts failed to grow, whereas co-culture with mesenchyme resulted in growth and morphogenesis.  When these two extracts were cultured together but  separated by a membrane, the extracts still grew to a lesser extent, indicating that soluble factors may play a role in directing pancreatic differentiation (Edlund, 2002).  Figure 2.2: Pancreatic Organogenesis Source: Wells (2003).  4  The transcription factor, pancreatic and duodenal homeobox gene-1 (Pdx-1), is expressed in pancreatic precursors (Figure 2.2B) and is required for pancreatic development (Jonsson et al., 1994).  Lineage tracing studies have shown that cells  expressing neurogenin-3 (Ngn-3), another transcription factor, become islet cells; however, their progeny do not express ngn-3, indicating that these progenitors do not self-renew and they represent a transient state of differentiation (Gu et al., 2001). 2.2  Pancreatic Stem Cell Candidates  2.2.1 Evidence of a Pancreatic Progenitor in Post-natal Animals The pancreas exhibits a limited self-renewal capacity. P-cell regeneration was observed after induced pancreatic injury from cellophane wrapping (Rosenberg et al., 1983), partial pancreatectomy (Bonner-Weir et al., 1993) and duct ligation (Wang et al., 1995). Rosenberg et al. (1983), in their "cellophane wrapping" model, induced ductal enlargement in hamster pancreas by tying cellophane loosely around the head of the organ. Islet cells appeared to "bud" from the enlarged ducts, and 6 weeks after wrapping, a 2.6-fold increase in islets cells was observed compared to control animals. In another rodent study, injection of the toxin streptozotocin into mice initially reduced the P-cell numbers to undetectable levels, but within five days, they were restored to 32% of the amount before injection (Fernandes et al., 1997). Also, short-term glucose infusion was shown to increase P-cell mass (Bernard et al., 1999). A kinetic analysis in post-natal rats indicated that P-cell mass increased with age until 3 months and that the pancreas underwent a massive remodeling with a large degree of apoptosis (Finegood et al., 1995). The rates of P-cell replication and enlargement were  5  insufficient to maintain or increase cell mass, implying that neogenesis, which is the formation of differentiated cellsfromprecursors, had to occur. Although these results show the possible existence of a pancreatic stem cell, it is unclear which cells are the precursor cells, where they persist, and if they have the unlimited proliferative capacity of stem cells. Both in vitro and in vivo studies have identified some candidates, some of which will be discussed in further detail in the following sections. However, it is important to first note some difficulties in identifying the pancreatic stem cell.  First, the cultures in most studies are heterogeneous  populations. So although certain fractions may be assessed, i.e. an islet rich digest, they still contain contaminating tissue such as exocrine, ductal, and connective (fibroblasts) cells.  Second, the insulin expression and secretion levels of cultured "differentiated"  cells are generally lower than in freshly isolated islets. Last, different cell processes can occur such as proliferation, differentiation, dedifferentiation, and transdifferentiation although the contribution of the latter two is debatable - and a population of cells may undergo several processes at the same time. Unless these processes are individually tracked, it is difficult to attribute increases or decreases to a particular mechanism. 2.2.2 Nestin Positive Cells Zulewski et al. (2001) found that cells isolated from rat and human islets could be expanded in vitro with a doubling time of 12 to 15 hours. The cells grew in monolayers and stained positive for nestin, a neural stem cell marker. It was shown that these nestin positive cells were a distinct population because they did not co-stain for any islet hormones or cytokeratin 19 (CK19), a marker for ductal cells in rats, pigs, and humans (Bouwens, 1998a). When these cultures reached confluence, three-dimensional clusters  6  grew out of the monolayers and expressed Pdx-1, proglucagon, and insulin, suggesting that these nestin positive cells could participate in islet neogenesis. In a second report by the same group, the addition of glucagon-like peptide 1 (GLP-1) or exendin-4, a long-acting analog of GLP-1, further stimulated differentiation of nestin positive cells into insulin producing cells as shown by double-staining for nestin and insulin (Abraham et al., 2002). However, insulin expression was not confirmed with RT-PCR so it was possible that cells instead took up insulin that was present in the media. Furthermore, nestin was found only in the surrounding mesenchyme, not the endodermal tissue in the developing mouse (Selander and Edlund, 2002). Gao et al. (2003) found proliferating nestin positive cells in their cultures of human ductal-rich tissue, but these cells co-expressed vimentin, a fibroblast marker, and had a spindle shape similar to that of fibroblasts. Also, the endocrine clusters that formed contained few nestin positive cells. 2.2.3 Pdx-1 Positive Cells In addition to being critical in pancreatic development, Pdx-1 is expressed in mature /3-cells (Soria, 2001). In an attempt to grow (S-cells, Beattie et al. (1999) isolated cells from the inner fraction of human islets and expanded them in monolayer cultures on dishes coated with human tumor bladder-9 (HTB-9) matrix. The cells, which initially expressed both Pdx-1 and insulin, were passaged 15 times, maintained Pdx-1 expression but lost insulin expression. To demonstrate that proliferating cells were originally islet cells, they cultured ductal fractions in a similar manner and found that the proliferating cells contained CK19, and not Pdx-1. It is possible that these proliferating Pdx-1+ cells  7  represent a dedifferentiated /3-cell (Bonner-Weir, 2000), but there has been little success in restoring the /3-cell function after expansion. 2.2.4 Ductal Cells Because endocrine and exocrine cells originate from the ductal epithelium in the developing pancreas, numerous studies have examined ductal tissue as a source of islet progenitors. Bonner-Weir et al. (2000) isolated and cultured ductal-rich human tissue in monolayers for 1 to 2 weeks, with a maximum expansion of 7-fold. Most cells stained positive for CK19, and a few for insulin. When confluent cultures were overlaid with Matrigel, a commercial extracellular matrix (ECM) preparation, clusters formed which stained positive for dithizone (a /3-cell marker) and CK19, and RT-PCR was used to confirmed the expression of insulin. Gao et al. (2003) also expanded ductal-richfractionsabout 3-fold in monolayers and found two peaks of 5-bromo-2-deoxyuridine (BrDU) labeling, which was used to detect proliferation. On days 3 to 4, the majority of proliferating cells were CK19 positive whereas on days 7 to 8, nestin-positive cell proliferation dominated.  This  suggested that there was a short window of ductal cell proliferation before the overgrowth of nestin-positive cells. At about 80% confluence, serum was omitted from the medium and the monolayers were overlaid with Matrigel. Similarly to the cultures by Bonner-Weir et al. (2000), clusters containing insulin-positive cells formed and the insulin/DNA ratios increased 4-fold. Overall, these studies provide a promising starting point for expanding pancreatic stem cells.  8  2.3  Strategies in Developing Pancreatic Stem Cell Cultures With the exception of the Pdx-1 positive cells that Beattie et al. (1999) cultured,  the amount of growth in primary pancreatic cultures has been limited and when they did grow, they lost their differentiated phenotype (Bouwens et al., 1998). Therefore, a twostep protocol of progenitor expansion followed by differentiation is a logical strategy for generating more (3-cells. Factors that have been reported to stimulate proliferation (Table 2.1) and differentiation (Table 2.2) include soluble media components and extracellular matrix proteins. Interestingly, the processes of expansion and differentiation are often not described separately, as an increase in /3-cell mass is usually the goal. However, the factors that stimulate progenitor growth may actually reduce /3-cell mass. Of the factors reported in Table 2.1, several are potential mitogens for ductal or progenitor cells specifically.  Epidermal growth factor (EGF) stimulated overall cell  growth but decreased the proportion of endocrine cells. Upon its removal, the proportion of insulin+ cells increased by 3-fold compared to controls (Cras-Meneur et al., 2001). Keratinocyte growth factor (KGF) was used by Bonner-Weir et al. (2000) and Gao et al. (2003) to expand their ductal extracts, and its injection into rats stimulated ductal cell growth in vivo (Yi et al., 1994).  Hepatocyte growth factor (HGF) was shown to  specifically stimulate CK19+ cells (Lefebvre et al., 1998). The receptor for vascular endothelial growth factor (VEGF), PTK fetal liver kinase-1 (flk-1), is localized in pancreatic ducts (Oberg et al., 1994). Most studies examined one or two factors at a time, but it is possible that a combination of factors may have synergistic effects that result in greater levels of expansion. Vila et al. (1995) reported "synergistic" effects between EGF, HGF and IGF-  9  -3  1 on [ H]-thymidine (a proliferation marker) labeling. However, the values reported for the combinations were not higher than the sum of the individual effects, as the term "synergy" implies.  Statistical methods, such as factorial experiments, allow the  assessment of multiple factors and their interactions in a minimal number of runs. They differ from one-factor-at-a-time approaches in that factors are varied simultaneously. They have been used to optimize recombinant protein production processes (Castro et al., 1992 and 1995; Moran et al., 2000; Liu et al., 2000; Chun et al., 2003) but have not been as widely applied in stem cell research (Zandstra et al., 1998; Audet et al., 2002). Table 2.1: Reported Mitogens for Pancreatic Cell Culture Factor Concentration  Culture System and Media Components  Animal Tissue Source  Effect  Reference  betacellulin  suspension 10% human serum  human fetal pancreas  2.6X t D N A synthesis  4nM bFGF 20 ng/mL  monolayers 20 ng/mL E G F  rat and human islets  enough growth to p a s s a g e cells 8 to 10 times over 8 months  EGF 10 nM  monolayers on collagen  guinea pig duct epithelium  1.5X t monolayer area  2.5% F B S 5 ug/mL insulin  Demeterco et al., 2000 Zulewski et al., 2001 V e r m e and Hootman, 1990  5 ug/mL transferrin 5 ng/mL selenite EGF  monolayer  Bhattacharyya  2.5% F B S  guinea pig duct epithelium  1 . 6 X t B r D U labeling and  30 nM  monolayer area  e t a l . , 1995  EGF 10 ng/mL  monolayer  human  Upto2Xt  Vila et al.,  1% F B S  duct epithelium  labeling  [ H]thymidine 3  1995  50 ug/ml fibroblast inhibitor EGF  3D type 1 collagen matrix  rat  2 X t bud size  Cras-Meneur  50 ng/mL  10% F C S  embryonic pancreas  3X t insulin expression  et al., 2001  B r D U labeled p a n - C K containing cells  NA  FGF10 50 ng/mL  serum-free 30 ug/ml transferrin monolayer on fibronectin  glucose 11.1 -26.1  mM  20 m M glutamine 5% F C S  HGF 10 ng/mL  monolayer 1%FBS  mouse embryonic pancreatic buds  t ductal proliferation 1 day cultures  rat neonatal islets "fibroblast free"  4X1 H-thymidine labeling at 21 m M no effect on insulin content  human duct epithelium  50 ug/ml fibroblast inhibitor  10  Pulkkinen et al., 2003  no effect on 5 day cultures 3  1.5X-  12Xt  [ H]thymidine labeling 3  Hulinsky et al., 1995  Vila e t a l . , 1995  Table 2.1 Continued Factor Concentration  Culture System and Media Components  HGF 10 ng/mL  monolayer on 8 0 4 G matrix  Animal Tissue Source  Effect  Reference  human  3X t in D N A  Beattie et al.,  fetal islets  I insulin Aggregation & H G F removal restored insulin content.  1996  monolayer on H T B - 9  human  200X T with H G F  Beattie et al.,  matrix 10% human serum  adult islets  1 0 0 X T without H G F H G F removal restored insulin content  1997  HGF 25 ng/mL  monolayer on H T B - 9 matrix  human inner core of adult islets  30000X t in cell number. Continued expression of PDX-1 but loss of insulin  Beattie et al., 1999  HGF  human  3X t BrdU labeling in  Lefebvre et  10-15 ng/mL  monolayer on 8 0 4 G matrix  adult islets  CK19+ cells  al., 1998  IGF-1  monolayer  Vila et al.,  1% F B S  human duct epithelium  3.7X t [ H]thymidine  10 ng/mL  10% human serum  HGF 10-25  ng/mL  10% F B S  10% human serum 3  1995  50 ug/ml fibroblast inhibitor Insulin 1 uM  monolayer on treated dishes or culture inserts 2.5% F B S  guinea pig duct epithelium  tmonolayer area.  KGF 50 ng/mL  3D collagen gel matrix 1%FBS  rat embryonic pancreatic epithelium  2.8X expansion I endocrine cells 3X expansion in acinar cells upon K G F removal  KGF  serum-free  Pulkkinen et  30 ug/ml transferrin  mouse embryonic dorsal and ventral buds, duodenal loop  t ductal proliferation 1 day  50 ng/mL  cultures  al., 2003  TGF-a 10 nM  monolayer 2.5% F B S  1.8X t B r D U labeling  •Bhattacharyya -.et al., 1995 Elghazi et al., 2002  no effect on 5 day cultures 4- insulin and glucagons  guinea pig duct epithelium  1 . 5 X T B r D U labeling  Bhattacharyya  t monolayer area  e t a l . , 1995  guinea pig duct epithelium  50% I in B r D U labeling and monolayer area  Bhattacharyya et al., 1995  TGF-p  monolayer  1 nM  2.5% F B S  VEGF  attachment dishes  rat  2 X t insulin/DNA ratios  Oberg et al.,  20 ng/mL  1% F C S  fetal pancreas  f H-thymidine labeling  1994  3  insulin -ve cells no effect on insulin +ve cells 2-3X t BrdU labeling  R o o m a n et  no effects on insulin  al., 1997  VEGF 10 ng/mL  5% F C S preculture, serum-free for experiments  rat duct epithelium  BCEM  monolayer  rat islets 10% F B S  2-5X overall expansion % endocrine cells remained constant  Thivolet et al.,  rat neonatal islets  no effect  10% F B S  Hayek et al., 1988  monolayer  human  increased attachment,  Lucas-Clerc  7% F B S  human islets  increased insulin secretion  et al., 1993  monolayer  guinea pig  2.5% F B S  duct epithelia  monolayer  human fetal islets  collagen collagen 1  monolayer  staining 1985  vs. control collagen 1 HTB-9  11  2X t morphological area  Bhattacharyya etal.,  100X t D N A content increase, but loss of insulin content  1995  Beattie et al., 1997  Table 2.1 Continued Factor Concentration fibronectin fibronectin  Culture System and Media Components  Animal Tissue Source  Effect  Reference  No effects  Bhattacharyya et al., 1995  monolayer  guinea pig  2.5% F B S  duct epithelia  attachment  rat  20 m M glutamine  neonatal islets "fibroblast free"  15% F C S fibrin  laminin  2X t H-thymidine labeling 5X t B r D U staining no difference in insulin content J  3D gels 10% F B S 20 ng/mL H G F 1 ng/mL F G F 4 10 m M nicotinamide  human adult islets  monolayer on culture  guinea pig duct epithelia  50% T morphological area  Bhattacharyya et al., 1995  inserts 2.5% F B S  3X t D N A 3X t beta cells per islet, improved islet function in vivo, no differences  Hulinsky et al., 1995  Beattie et al., 2002  observed when growth factors were used in suspension cultures  Matrigel  suspension 10% F B S 22 m M glucose  rat neonatal islets  no effect  Hayek et al., 1988  Matrigel  monolayer on culture inserts  guinea pig duct epithelia 2.5% F B S  50% I morphological area  Bhattacharyya et al., 1995  Table 2.2: Reported Differentiating Factors for Pancreatic Cell Culture Animal Tissue Source  Effect  human  1.5-2X t in insulin content  Reference  Factor Concentration  Culture System and Media Components  activin-A 4 nM  suspension, clusters 10% human serum  fetal pancreas  GLP-1 10 n M  monolayer  human  10% F B S  P A N C - 1 cells transfected with PDX-1  GLP-1  monolayer  rat  aggregation of cells  Hui e t a l . ,  10 n M  10% F B S  A R I P cell line  insulin production and  2001  GLP-1,  monolayer  human  t insulin staining, had  Abraham et  exendin-4 10 n M  10% F B S  islets (nestin positive)  synergistic effects with PDX-1 transfection  al., 2002  Demeterco et al., 2000  aggregation of cells insulin production and secretion  Hui et al., 2001  secretion  71 u M p-mercaptoethanol  only a subset of cells produced insulin  20 ng E G F 20 ng b F G F B-27 supplement exendin-4 10 n M  monolayer on B C E M  human  t PDX-1 expression but  Movassat et  10% human serum  fetal islet  no differences in insulin  al., 2002  nicotinamide  Matrigel  chick  t insulin stained cells.  Mngomezulu  5-10 m M  5ug/mL insulin  embryonic  BrdU staining of insulin+  et al., 2000  5ug/mL transferrin  pancreatic  cells was low  1 n M selenium  dorsal buds  content  12  Table 2.2 Continued Factor Concentration  Culture System and Media Components  Animal Tissue Source  Effect  Reference  collagen  3D matrix  rat neonatal pancreas  reorganization into smooth, 3D islet -like aggregates  al., 1983  10% F C S collagen 1  Montesano et  3D matrix  human  reorganization into  Lucas-Clerc  7% F C S  adult islets  smooth, 3D islet -like  et al., 1993  aggregates t insulin secretion v s . monolayer collagen culture Matrigel  overlay of monolayer cultures 5 (xg/mL insulin  Matrigel  5 ug/mL transferrin 5 ng/mL selenite 10 ng/mL K G F 10 m M nicotinamide overlay of monolayer cultures 10 ng/mL K G F 10 m M nicotinamide  2.4  human adult ductal and islets rich tissue that adhered to non-treated dishes  formation of 3D cystic structures  Bonner-Weir et al., 2000  human ductal rich fragments  10X t insulin/DNA ratios of monolayers  G a o et al., 2003  Neonatal Pancreatic Porcine Cell Clusters (NPCCs) as a Model System Due to their fast breeding, large litter size and physiological similarities to  humans, pigs provide a good model system for studying islet neogenesis and are a potential tissue source for xenotransplantation. Neonatal pancreas is favoured over adult tissue because it is less fragile and variable, and it may be rich in progenitor cells since part of the maturation process of P-cells occurs post-natally in the pig (Korbutt et al., 1996; Rayatetal., 1999). 2.4.1 NPCC Suspension Culture Korbutt et al. (1996) characterized NPCCs' survival and their ability to correct hypoglycemia. Cell clusters were formed by digesting the pancreas with collagenase and culturing cells in serum-free suspension. After 9 days, cell numbers decreased to 10% of the initial values and about 80% of the decrease was due to the loss of acinar cells (Korbutt et al., 1996). In contrast, the proportion of insulin positive cells and insulin  13  content increased by 4.8-fold and 6.7-fold, respectively.  Non-granulated cells, which  would include ductal cells, increased from 10 to 57%. Transplantation of 2 x 10 clusters 3  into the kidney capsule of alloxan-induced diabetic mice corrected their hypoglycemia within 8 weeks. Yoon et al. (1999) reported similar insulin profiles in their NPCCs and found that approximately 18% of insulin-positive cell co-stained with cytokeratin-7 (CK7), a pig ductal cell marker (Bouwens, 1998a). Trivedi et al. (2001) assessed the grafts of NPCCs and found that 10 days after transplantation, 9% of fi-cells expressed Ki67, a proliferation marker, but this decreased to undetectable amounts at 20 weeks. Approximately 6% were ductal cells, which were identified by their expression of CK7, at day 10. Further proliferation was not assessed at later time points because of the paucity of ductal cells. At day 10, approximately 20% of CK7+ cells co stained for insulin. Together, these results suggest that both in vivo and in vitro, NPCCs undergo a maturation process which may include the preferential survival of P-cells, P-cell neogenesis from ductal cells, and (3-cell replication. 2.4.2 NPCC Monolayer Culture NPCCs have been cultured in monolayers (Tatarkiewicz et al., 2003) to favour the growth of ductal cells. For this purpose, collagenase digested pancreatic clusters were further digested with trypsin and DNase into a single cell suspension and allowed to adhere to tissue-culture treated dishes. Fibroblasts generally attach faster so cells that had adhered within the first 3 hours of incubation were discarded to reduce the proportion of fibroblasts. The total DNA content fell by 50% after 3 days - attributed to acinar cell death - but almost fully recovered by day 12. At day 3, 35% of CK7+ cells co-expressed  14  Ki67. At day 9, the number of proliferating ductal cells fell to 10%. Since the authors did not report the number of CK7+ cells for day 4, it is difficult to assess whether the 70% ductal proportion on day 9 was an increase or decrease. However, the decrease in the proportion of proliferating ductal cells was attributed to the overgrowth of fibroblasts on day 12. Cells from day 9 were reaggregated and transplanted into nude, non-diabetic mice. When examined 4 weeks after transplantation, the grafts were found to consist mainly of endocrine cells. 2.4.3 Maturation of NPCC's in Alginate and Autologous Serum Although NPCCs have corrected hyperglycemia in diabetic mice, recovery rates were as long as 8 weeks (Vizzardelli et al., 2002; Korbutt et al., 1996), possibly due to the required maturation of cells.  Insulin content was increased when NPCCs were  cultured in alginate and media supplemented with 10% fetal bovine serum (FBS) (Tatarkiewicz et al., 2001) or 5% neonatal pig serum (NPS) (Korbutt et al., 1997). Both serum and alginate were necessary because no change in insulin content was observed when cells were cultured in alginate alone, and cultures in NPS alone resulted in a large degree of cell clumping that lead to necrosis.  Tatarkiewicz et al. reported that FBS  yielded higher insulin/DNA ratios than NPS, but their protocol failed to reduce the time to achieve normoglycemia in diabetic mice after transplantation (Lopez-Avalos et al., 2001), whereas Korbutt et al. reported recoveries between 3 and 14 days when encapsulated cells were transplanted. The difference could have been due to the serum or the fact that Korbutt's group transplanted cells that were still encapsulated in alginate.  15  2.5  Cell Immobilization Methods  2.5.1 Immobilization of Mammalian Cells Examples of ECM components include collagen, fibrin andfibronectin.Matrigel is a commercial ECM extract prepared from mouse sarcomas. ECM substitutes that are used in tissue engineering include natural gels derived from non-mammalian cells such as alginate, carrageenan, agar, agarose and synthetic gels such as polyurethane, polyethylene glycol, polylactic acid, polyglycolic acid and polylactic-co-glycolic acid. In addition to providing support and structure to tissue, ECM may influence stem cell growth and differentiation.  Proteins in the ECM provide signals that regulate cell processes and  influence cell attachment, which is critical for growth in some cell types. ECM also directs the positioning of cells, which may induce cell-cell interactions that are also believed to play a role regulating stem cell behaviour (Levenberg et al., 2003; Beattie et al., 1996). 2.5.2  Alginate Alginate is a biopolymer derived from brown sea algae.  It is composed of  polysaccharide chains of 1,4-linked p-D-mannuronic and a-L-guluronic acid. These polymers gel in the presence of divalent cations, such as Ca  2+  and Ba . Alginate can be 2+  de-gelled by displacing the divalent cations with Na and chelating them with agents such +  as EDTA, citrate or phosphate. Its advantages are that it is biocompatible, can be gelled without harsh chemicals or extreme temperatures, and the gelling is reversible. Besides NPCCs (Section 2.4.3), alginate has been used as a scaffold for culturing cells such as cartilage (Kavalkovich et al., 2002; Steinert et al., 2003), Schwann (Mosahebi et al., 2003), bone marrow (Wang et al., 2003) and embryonic stem cells (Magyar et al., 2001).  16  2.5.3 Systems for Alginate Immobilization Alginate immobilization is typically in the form of beads or slabs (Figure 2.3). To generate the ~10 islets, or 1.5 x 10 cells, per patient required for transplantation 6  9  (Shapiro et al., 2000), culture volumes would have to be around 1000 mL, which limits the practicality of using slabs in a scaled-up system. Koch et al. (2003) compared 3 methods of bead generation: 1) vibrating nozzle, 2) coaxial gas flow extrusion, and 2) JetCutter technology which had processing rates of 10 mL/h, 130 to 240 mL/h, and 330 mL/h, respectively. The 10 hours that would be required for bead generation in coaxial gas flow extrusion limits its applicability at the clinical scale. Another criterion is that beads must be mechanically stable enough to withstand agitation and hence, a higher viscosity alginate is usually used. The vibrating nozzle device required viscosities less than 0.2 Pa-s in order to produce uniform beads smaller than 300 (am. Even if a sufficient number of beads can be generated, the additional unit and operating constraints make bead immobilization a less desirable method. 2.5.4 Hollow-Fibre Bioreactors A hollow-fibre bioreactor (HFBR) is a continuous perfusion system that is widely used in the production of monoclonal antibodies at the laboratory scale (Figure 2.4). Although no reports of culturing cells in a gel in a HFBR have been found, the extracapillary space (ECS) has been filled with a 3.5% gelatin and 1.25% agarose mixture (Wei and Russ, 1977) and 2% agarose (Koska et al., 1997) to simulate packed cell densities for mass transport studies. Gelatin and agarose are thermally gelled, so the process involved slowly filling the ECS with agarose at a high temperature (60 to 70°C) and allowing it to cool overnight. A similar protocol could be used to fill the ECS with  17  alginate, but instead of cooling the cartridge, a gelling solution of Ca could be passed through the intracapillary space (ICS). The advantage of culturing cells in a HFBR are that it is relatively inexpensive, is disposable, can support very high cell densities ( > 1 0 cells/mL) and more importantly, is a closed system, thereby reducing the risk of contamination. Another advantage of using a HFBR is that the alginate would not be exposed to high shear rates, so the constraints on alginate viscosity or concentration discussed in Section 2.5.3 would not necessarily apply. cell clusters s u s p e n d e d in liquid alginate  bead generator  Overlay with C a to form alginate slab  Ca solution 2 +  2 +  Replace Ca with medium 2 +  W a s h and transfer beads to culture dishes or spinner flasks  Cells in alginate slabs  Slabs  t.°o°o°o°e°o°o°o°o°o°o°o°o°o°o'°l  Beads Figure 2.3: Process for Immobilizing Cells in Alginate Beads or Slabs  18  ICS Manifold Cell Loading  Figure 2.4: Schematic of HFBR  19  3  MATERIALS AND METHODS  3.1  NPCC Cultures and Analysis  3.1.1 Isolation and Shipment NPCCs from 1-3 day old pigs (Day 0) were obtained from the University of Alberta's Department of Surgery (Korbutt et al., 1996). Briefly, each pancreas was cut into pieces, digested with collagenase, washed four times with Hank's balanced salt solution (HBSS), and placed into 4-15 cm diameter non-treated dishes (Fisher Scientific, Pittsburg, PA), each containing 35 mL Ham's F10 medium supplemented with 10 mM glucose, 50 mM isobutylmethylxanthine (IBMX), 0.5% (w/v) bovine serum albumin (BSA) (Sigma, St. Louis, MO), 2 mM L-glutamine, 10 mM nicotinamide, 100 U/mL penicillin, and 100 mg/mL streptomycin. Dishes were placed in a 37°C incubator with humidified air containing 5% CO2. Media were changed on days 1 and 3.  Unless  otherwise indicated, cells were shipped on day 4 and arrived at the Biotechnology Laboratory at the University of British Columbia on day 5. Cells were shipped without temperature control in 50 mL conical tubes with 0.5 to 1 cm headspaces. Each pancreas yielded about 4 x 10 cells on day 5. 7  It was observed that when cells were received on  day 5 and cultured in serum-free medium, adhesion to tissue culture treated dishes was minimal, whereas cells received immediately after isolation would adhere even when non-treated dishes were used. 3.1.2 Seeding and Sampling Technique of Suspension Cultures On arrival, cells were centrifuged for 1 to 2 min at 200 g, resuspended in fresh media, and combined into one 50 mL conical tube.  Because NPCC clusters settled  rapidly in the medium, a special technique was used to ensure a consistent seeding  20  density among cultures. A concentrated stock was made by suspending cells in just enough medium to aliquot 200 \xL per dish, t-flask, or well. Using a 1000 u.L pipettor, the cell stock was quickly aspirated 2 to 3 times before drawing up 200 uL for seeding. The standard deviation of this method was about 7.5% (Table 7.1). The same technique was used when sampling, except that the draw-up volume for nuclei counting varied and cultures were sometimes diluted or concentrated depending on the cell concentration and amount of tissue available.  Seeding densities were between 7.5 x 10 and 4 x 10 5  6  cells/mL, depending on the cells needed for analysis. 3.1.3 Dissociation and Technique for Monolayer Cultures For monolayer cultures, cells were shipped on the day of isolation. The protocol was adapted from Tatarkiewicz et al. (2003). On arrival, tissue was combined and washed twice with and resuspended in 40 mL per pancreas of dissociation medium 1 (Ca+ and Mg+ free HBSS supplemented with 1 mM EDTA, 10 mM HEPES, 0.5% BSA) in a 75 cm T-flask (Sarstedt, Germany). Cells were shaken at 35 rpm, 37°C for 10 2  minutes. 25 u.g/mL trypsin (Roche, Laval, QC) and 4 u.g/mL DNase (Roche) were added and cells were shaken for another 10 min, resulting in a single cell suspension with some large clumps of cells.  The suspension was carefully decanted and saved, and the  dissociation process was repeated with the remaining clumps. dissociation, the two suspensions  After the second  were combined, resuspended in RPMI-1640  supplemented with 10% FBS (Cansera, Toronto, ON), 100 U/mL penicillin, 100 u.g/mL streptomycin, 10 mM nicotinamide, 10 mM HEPES at about 1 x 10 cells/mL, and 6  transferred into tissue culture treated dishes. Cells were incubated for 3 hr at 37°C in a humidified atmosphere with 5% CO2, allowing mainly contaminating fibroblasts to  21  adhere. The supernatant was drawn off and aliquoted into culture dishes at 4 to 7 x 10 cells/mL. 3.1.4 Cell Enumeration If a single cell suspension could be obtained, cells were counted in a 0.2 mm deep hemocytometer, using trypan blue exclusion to assess cell viability. In most cases, at least 200 cells were counted to obtain an accurate estimate of cell density.  In many  NPCCs, cells were too aggregated to be accurately counted this way, so DNA measurements or nuclei counts were used to enumerate cells. For the DNA analysis, 200 u.L samples were taken in duplicate or triplicate and washed twice in citrate buffer (150 mM NaCl, 15 mM citric acid, 3 mM EDTA, pH with NaOH to 7.4). Cell pellets were stored at -20°C until analysis.  Samples were shipped on dry ice overnight to the  Department of Surgery at the University of Alberta where the DNA content was analyzed with the fluorescent reagent, Pico Green (Korbutt et al., 1996). To calculate cell numbers from DNA, a conversion of 7.1 pg DNA per cell was assumed (Korbutt et al., 1996). For each nuclei count, a 0.1 to 3 mL, depending on the cell concentration, sample was transferred into a microcentrifuge, 15 mL conical or 50 mL conical tube. At least 7 times the sample volume of 0.1 M citric acid was added, and the sample was incubated at 37°C for 1 to 3 hr. Cell membranes were disrupted with periodic vortexing. Nuclei were stained with 0.1% (w/v) crystal violet in 0.1 M citric acid and counted in a 0.1 mm deep hemocytometer, which required less refocusing than a 0.2 mm deep hemocytometer. 3.1.5 Insulin Analysis 1 mL of Azol (11.4% (v/v) glacial acetic acid, 0.25% BSA) was added to each 200 u.L sample, taken in duplicate or triplicate and stored at -20°C until analysis.  22  Samples were shipped overnight on dry ice to the Department of Surgery at the University of Alberta where cells were disrupted and the insulin content analyzed by radioimmunoassay (Korbutt et al., 1996). 3.1.6 Cell Dissociation and Fixation for Immunocytochemistry Slides When cultured in suspension, NPCCs needed to be dissociated into single cells before fixation on slides.  Islets were transferred into a 15 mL conical tube and  centrifuged at 200 g for 1 min. Cells were resuspended in 10 mL of dissociation medium 2 (135.87 mM NaCl, 0.81 mM MgS0 *7H 0, 5.36 mM KC1, 1 mM Na HP0 , 0.44 mM 4  2  2  4  KH P0 , 10 mM HEPES, 2.78 mM glucose, 1 mM EGTA, 4.17 NaHC0 prepared in 2  4  3  Nanopure water, pH adjusted to 7.4) and placed into a 37°C water bath and aspirated with a siliconized Pasteur pipette for 7 min. 25 ug/mL trypsin and 4 ng/mL DNase were added and cells were aspirated for 4 min. At this point, cells were checked under a microscope to see if they were dissociated adequately. If there still were large clusters, cells were further aspirated until most clusters were smaller than 3 to 5 cells. Two 0.8 cm circles were etched onto each silane-coated glass slide (Marienfeld, Germany or Electron Microscopy Sciences, Ft. Washington, PA) using a diamond-tipped pen. Cells were washed once with phosphate buffered saline (PBS) and resuspended in PBS. 50 fiL of the cell stock was applied to each circle, and slides were left for at least 15 minutes to allow cells to adhere. Slides were placed into a Coplin jar containing Bouin's fixative (8.8% formaldehyde, 4.8% glacial acetic acid) and incubated at room temperature for 15 min, followed by rinsing 3 times and storage in 70% ethanol at 4°C until staining.  23  3.1.7 Immunocytochemistry Staining Procedure Single staining was performed with the avidin-biotinylated enzyme complex with diaminobenzidine as the chromagen (ABC-DAB) whereas double staining was detected with fluorescent tags, which were generously donated by Dr. B. Finlay (Biotechnology Laboratory, University of British Columbia). Table 3.1 lists the antibodies, dilutions, and incubation times used. Slides were immersed in 10 mM citrate (pH 6.0) and microwaved on high 6 times at 5 second intervals for antigen retrieval. When ABC-DAB was used, slides were incubated in a 10% H2O2 solution prepared in methanol for 6 minutes to block endogenous peroxidase activity. Slides were rinsed under running tap water for 3 minutes, removed one at a time, and shaken to remove the water. A kimwipe was used to dry the area around each dot, and a dime-sized circle around the dot was drawn with a Dako pen (DakoCytomation, Denmark) to create a water barrier. A drop of PBS was applied to each dot and slides were placed in a moist chamber to prevent dehydration during incubation. Table 3.1: Antibodies Used in Immunocytochemistry Antibody  Dilution  Source  Incubation Time (min) Single Staining  Double Staining  Anti-human C K 7 (raised in mouse)  DakoCytomation  1:200  30  60  Anti-human Ki67 (raised in guinea pig)  Santa C r u z Biotech  1:50  30  60  Biotinylated anti-mouse IgG  Vector Laboratories  1:200  20  NA  Biotinylated anti-guinea pig IgG  Vector Laboratories  1:200  20  NA  Alexa Fluor 568 (red) goat anti-mouse  Molecular Probes,  1:200  NA  60  IgG  Eugene, O R  Alexa Fluor 488 (green) goat anti-  Molecular Probes  1:100  NA  60  guinea pig IgG  To block non-specific binding of the antibodies, the PBS was shaken off and 50 uL of 20% normal goat serum (NGS, Sigma) prepared in PBS was applied to each dot.  24  After 15 min, the serum was shaken off and 50 JJ.L of the 1° antibody was applied to the first dot on each slide and 50 uL of 20% NGS diluted by the same amount as the 1° antibody was applied to the second dot on each slide for the negative control. 1° antibodies were applied at the same time for double staining. After incubation, slides were rinsed 3 times with PBS, avoiding mingling of fluids between dots. The PBS was shaken off and 50 uL of the 2° antibody was applied to each dot and incubated. 2° antibodies were applied at the same time for double staining. After incubation, each slide was rinsed 3 times with PBS. To develop the colour in single staining, a few drops of avidin-biotinylated enzyme complex (ABC Standard Kit, Vector Lab, Burlingame, CA) were applied to each dot to amplify the signal. After 40 min, slides were rinsed with PBS, and a few drops of DAB solution (Ready-to-Use Liquid DAB Substrate Pack, Biogenex, San Ramon, CA) were applied. After 5 min, slides were rinsed 3 times with distilled water. Slides were preserved with SuperMount (Biogenex) followed by Permount for colour slides or antifade solution (2.5% DABCO, 90% glycerol in PBS, pH 8.0) for fluorescent slides and covered with a coverslip. 3.7.5  Cell Counting and Image Processing for Immunocytochemistry Cells from single staining slides were counted under a light microscope.  Fluorescent images were captured using a Zeiss Axioplan 2 fluorescent microscope with a Qlmaging Retiga 1300 camera and Northern Eclipse 6.0 software (Mississauga, ON). Figure 3.1 shows some representative images that were processed. Monochrome pictures were taken to obtain a higher resolution and were processed using a trial version of SigmaScan Pro 5.0 (Chicago, IL). The negative controls for C K 7 showed little red  25  fluorescence, so a low intensity threshold, SigmaScan intensity 14, was used for counting a cell as positively stained. Negative controls for Ki67 showed a low degree of green fluorescence so a higher threshold for a positive result was used. The green threshold was selected by examining each negative control image and adjusting the intensity threshold to a count of zero spots. Extreme thresholds were ignored and a SigmaScan intensity threshold of 68 was used, which gave a rate of counting false positives out of total positives of 6% (Table 7.10).  After applying the appropriate thresholds, the  resulting red and green images were combined and double staining appeared yellow. Images were processed with a macro script, but counting was done manually. 3.1.9 Suspension Culture Experiments Unless otherwise specified, cells were cultured in serum-free Ham's F10 medium described in Section 3.1.1 and media were changed every 2 days. A 2 " factorial design 4  1  with 4 centre points (Table 7.3) was used to assess the main and interactive effects of EGF (Stemcell Technologies, Vancouver, BC), HGF, VEGF and KGF (all Sigma) on cell survival and insulin/DNA. Growth factors were prepared according to manufacturers' instructions and added to 3 mL cultures in 6-well tissue culture treated plates (Nunclon, Denmark) with final concentrations of 0 (low value) or 50 ng/mL. Nuclei, insulin, and DNA were quantified after 4 and 8 day exposure to growth factors. Dose response experiments were carried out with EGF and HGF at 0, 1,5 and 25 ng/mL with 5% FBS. Comparisons were also made between 5% FBS and a serum replacement consisting of 1.5% BSA, 10 u.g/mL insulin (Recombulin Zn, Invitrogen, San Diego, CA), and 200 u.g/mL transferrin (Invitrogen). A long-acting form of IGF-1 (JRH Biosciences, Lenexa, KS) at 50 ng/mL was also tested as an insulin replacement.  26  7 Negative  i67 N e g a t i v e C o n t r o l  C K 7 a n d £ i 6 7 Combined, Threshold Applied  Figure 3.1: Representative Images Used in Determining CK7 and Ki67 Staining Red and greenfluorescenttags were used to stain cells for CK7 and Ki67, respectively. Primary antibodies were not applied in negative controls so anyfluorescencewould be due to non-specific binding of the secondary antibody (fluorescent tag) or autoflourescence of the cells.  27  3.1.10 Monolayer Culture Experiments Cells were cultured in serum-containing RPMI-1640 medium described in Section 3.1.3; media were changed 4, 7 and 9 days after isolation. A 2 " design with 4 centre 5  1  points (Table 7.12) was used to assess the main effects and all two factor interactions of EGF, VEGF, KGF, HGF, and bFGF on cell numbers and the expression of Ki67 and CK7. All growth factors were added to cultures at either 0 or 20 ng/mL, except HGF which had a high concentration of 16.7 ng/mL. The growth factor stocks from Section 3.1.9 were used for EGF, VEGF, and HGF. KGF was purchased from R&D Systems (Minneapolis, MN) and bFGF was purchased from Sigma. 3.1.11 Alginate Cultures Cells were shipped on the day of isolation, along with a stock of heat inactivated NPS from other isolations. These were cultured for 8 days in serum-free Ham's F10 media in non-treated dishes, allowing clusters to form.  The tissue was then evenly  divided into four 10 cm dishes: 2 for controls and 2 for alginate immobilization. For the alginate cultures, cells were suspended in 2.5 mL per dish of 0.75% MVG Pronova alginate (FMC Biopolymer, Philadelphia, PA), prepared in saline (0.9% NaCl, 10 mM HEPES). The surface of the treated dishes (Sarstedt) had been wetted with saline to allow an even spreading of the alginate. Dishes were tilted and rotated, and were left for several minutes. To gel the alginate, 10 to 12 mL of a 75 mM CaCb, 75 mM NaCl, 10 mM HEPES solution was added dropwise to each plate and left for approximately 5 minutes. The resulting alginate slabs, which had nominal thicknesses of about 0.4 mm, were washed twice with saline and twice with PBS, and covered with 10 mL Ham's F10 media (Section 3.1.1) supplemented with 5% NPS. Controls were cultured in 10 mL  28  medium without serum. On day 18, cells were recovered from the alginate by replacing the medium with 55 mM sodium citrate buffered with 10 mM HEPES and shaking (35 rpm) for about 20 minutes. 3.2  HFBR Cultures  3.2.1 CHOK1 Cells Chinese hamster ovary CHO-K1 cells (#CCL-61, ATCC, Manassas, VA) were used as a model system for testing alginate cultures in the HFBR, and were generously donated by Dr. C. Brown (Department of Medical Genetics, University of British Columbia). Cells were thawed and cultured in a 25 cm tissue culture flask in DMEM 2  supplemented with 5% FBS. For the HFBR growth experiment, FBS levels were increased to 20% because of difficulties with reduced growth at lower levels. To promote the formation of clusters simulating NPCCs, the KI cells, which grow in monolayers, were transferred to non-treated bacteriological dishes and passaged at least once before being placed in alginate. Cells were immobilized in 1% plant cell culture tested alginic acid (Sigma) and gelled in slabs as described in Section 3.1.11 or in the HFBR as described below. 3.2.2 Apparatus and Setup All HFBR experiments were carried out at room temperature under non-sterile conditions and without aeration of the fluids, except for the run where growth in the HFBR was tested.  The setup for the cell growth experiment, which was the most  complex is shown in Figure 3.2. A Gambro 15 dialyzer (Stockholm, Sweden) was used as the HFBR. Fluid flow was controlled with Masterflex peristaltic pumps (Cole-Parmer, Vernon Hills, IL) with standard pump heads.  29  Masterflex Pharmed tubing was used  throughout, except in the aeration unit, which consisted of 2 metres of 1/8" OD silicon tubing (Cole-Parmer) coiled and encased in a plastic box with a small vent at the top. Air mixed with 10% CO2 was fed into the box at a flow rate of lL/hr. The heating jacket of a Cell-Pharm System 1500 (Unisyn Technologies, Hopkinton, MA) biocontroller unit with UniNet 1.0 software (Unisyn Technologies) was used to maintain the temperature at 37°C. To assemble the system aseptically, sections of tubing and bottles were autoclaved separately and either assembled in a laminar flow hood or connected with a SCD ITB sterile tubing welder (Terumo, Elkton, MD).  Figure 3.2: Setup of HFBR System for Cell Growth Experiment 3.2.3 HFBR Alginate Loading, Gelling and Degelling For each run, a new HFBR cartridge wasrinsedwith 2 L of deionized water, and then 2 L of deionized water was recirculated through the ICS for at least 24 hours to flush  30  out bubbles in the fibers. Table 3.2 lists the flow rates and solutions used for the different experiments. At least 150 mL of alginate was prepared to fill the 120 mL ECS to allow for losses in the tubing.  The ECS was emptied by flushing it with air and was  subsequently loaded with alginate through the bottom port at a rate of 160 mL/min without cells and 50 mL/min with cells.  The alginate was recirculated for 5 to 10  minutes and the cartridge gently tapped to remove bubbles. The ECS pump was stopped and the ECS outlet was clamped to prevent fluid losses while recirculating the gelling solution through the ICS. Table 3.2: Operating Conditions of Different HFBR Runs Experiment  Alginate Concentration  Gelling Solution  Gelling Time (hr)  ICS Flow rate (mL/min)  Alginate Plugging  0.75% M V G  100 mM C a C I , 100 m M N a C l  1  115,300  Degree of Gelling  1%, 1.25%, 3% Sigma  75 m M C a C I , 75 m M N a C l  0.5, 1  300  Cell Recovery  1%, 3% Sigma  75 m M C a C I , 75 m M N a C l , 10 mM H E P E S  0.5  300  Cell Growth  1% Sigma  75 m M C a C I , 75 m M N a C l , 10 mM H E P E S  0.5  250  2  2  2  2  The alginate was degelled by recirculating 2 L of a 55 mM sodium citrate solution buffered with 10 mM HEPES through the ICS for 1 hr. This mainly degelled the alginate between the fibers, as the alginate between the fiber bundle and HFBR shell was still visibly gelled. Air was pumped into the ECS to push out 50 to 70 mL of liquefied alginate solution. This left a little space in the ECS, which was subsequently half filled with citrate solution. The cartridge was then manually rotated to distribute the citrate solution throughout the ECS. The process of removing liquid alginate, partially filling the ECS with citrate solution, and manually rotating the cartridge was repeated 5 to 6  31  times until no gelled alginate was observed. The liquid alginate containing the cells was collected and counted. 3.2.4 HFBR Experiments The degree of alginate plugging was determined by analyzing the residence time distribution of a protein that was pulsed through the HFBR before and after loading alginate.  Cells were not used in these experiments.  Azoalbumin, a reddish-orange  protein, was used as a tracer and prepared at 5 g/L in saline. It was pulsed for 5 sec, passed through the ICS, and samples were collected at the ICS outlet at 2 second intervals using an Advantec SF2120 fraction collector (Dublin, CA). The concentrations were determined with spectrophotometer readings at 450 nm. Given: Q =v A  (3.1)  ave  = l  R,ave  _L_  (3.2)  V ave  where Q is the flow rate v  ave  is the average velocity  A is the total cross sectional area of the fibres tR,ave  is the average residence time  L is the length of the fibres at constant Q and L, a decrease in A would increase v  ave  and decrease  tR, eaV  Hence  alginate plugging would result in faster elution of the protein. It was assumed that the protein did not interact with fibres or alginate and hence would move at the same velocity as the fluid. However, the cartridge did retain some red colour, indicating that there was some binding of the azoalbumin. The average residence time ( t R the measurements by: 32  )ave  )  was calculated from  (3.3) t=0  t  {C(t)dt t=0  where C = concentration and t = time after start of pulse. The integrals were calculated from the area under the elution curves using the trapezoidal rule. The degree of gelling was assessed at 3 concentrations and two gelling times (Table 3.2). An attempt was made to determine the degree of gelling quantitatively by performing a mass balance of alginate around the HFBR. Alginate concentrations could be measured with UV absorbance at 210 nm; however, many other substances such as HEPES, citrate, EDTA, and possibly other residues from the HFBR, interfered with the UV reading so eventually the degree of gelling was not assessed quantitatively. To assess the degree of gelling qualitatively, cartridges were sawed open and examined. In the cell recovery experiments, cells were suspended in the alginate at a concentration around 1.2 to 1.5 x 10 cells/mL. At this density and room temperature, the 5  oxygen consumption rate of the cells did not require active aeration of the solutions. The alginate was loaded, gelled and degelled within 4 hr. Cell recovery was calculated by: cell recovery =  cells recovered cells loaded  (3.4)  xl00%  In the cell growth experiment, the alginate was agitated with a magnetic stirrer at around 100 rpm while loading cells at around 4.4 x 10 cells/mL. After gelation, 1 L of 5  saline solution was flushed through the system to remove any excess Ca  2+  that would  sometimes form a precipitate with the phosphate in the medium. 0.6 L of medium was recirculated through the ICS for 5 days. Samples were taken daily and pH, glucose, lactate, and dissolved gases were measured on the Stat Profile 10 (NOVA Biomedical,  33  Waltham, MA), YSI 7100 MBS (YSI Life Sciences, Yellow Springs, OH), and RapidLab 348 (Bayer, Tarrytown, NY). On the fifth day, the silicon tubing in the aeration unit tore, ending the run. Cells were recovered and counted.  3.3  Statistical Analysis Factorial experiments were analyzed using a trial version of Design Expert 6.0  (Statease Inc, Minneapolis, MN) and JMPLN 4.0 (SAS Institute, Cary, NC). Factor values were first coded so that -1, 0, 1 represented the low, centre point, and high concentrations, respectively. Linear regression was used to fit the following equations to the data: y = /3 + /3 EGF + 0 0  EGF  VEGF  V E G F + |S KGF + |3 HGF KGF  @ EGFx VEGF aliased with KGFxHGF  (  E G F  X V  E  G  (3.5)  HGF  aliped with KGF X HGF) +  F  /WKGF aliased wi.h VEGFXHGF (EGF x KGF aliased with VEGF x HGF) + @ EGFx HGF aliased with VEGFxKGF  y = 8 + j8 EGF + j3 0  EGF  /W GFE  G F  V E  /3  E G F  xbFGFE  @ vEGFxbFGF  G F  VEGF  (  E G F  X H  G  F  aliased W i t h VEGF X KGF)  VEGF + /3 KGF + /3 HGF + /3  x VEGF + /3  KGF  EGFxKGF  HGF  EGF x KGF + /3  KGF)<HGF  bFGF +  (3.6)  HGFE x HGF + gf  EGFx  x bFGF + $ VEGFXKGF VEGF x KGF + /3  VEGF x bFGF + /3  bFGF  KGF x HGF + / ?  VEGFxHGF  KGFxbFGF  VEGF x HGF +  K G F x bFGF +  /^HGFxbFGFHOFxbFGF  where Equation 3.5 and Equation 3.6 applies to Section 3.1.9 and 3.1.10, respectively, y is the response, EGF represents the coded value of the EGF concentration, EGF x VEGF represents the interaction between EGF and VEGF,  (3EGF represents  the effect of the EGF,  etc. The significance of each P term was tested using analysis of variance (ANOVA). The fullest model was first tested with all non-aliased terms included. For each response, the models were then refined by the following process. The non-significant terms (p > 0.05) were removed and the new model was examined individually for overall  34  significance o f the model, an insignificant lack-of-fit, and residual plots without any outliers or unusual patterns. The fullest model that met those criteria was selected as a refined model. M o d e l hierarchy was obeyed so i f a factor was in a significant interaction, but its main effect was insignificant, the main effect would still be included in the model. B o x - C o x plots were also used to determine i f transformation of the response necessary.  was  In some cases, no factors were significant. For brevity, only the full model  and refined model, i f applicable, are presented in the Results section.  The A N O V A  tables can be found in Section 7. Unless otherwise stated, reported values are averages ± standard errors of the mean.  35  4  RESULTS AND DISCUSSION  4.1  NPCC Growth In Suspension Culture  4.1.1 Assessment of NPCC Survival in Suspension Culture Cells were isolated (day 0) and cultured for several days before overnight shipment on day 4. After arrival, when NPCCs were cultured in suspension in serum-free standard Ham's F10 medium, the cell concentration decreased by 50 to 80% (Figure 4.1). Nuclei counting resulted in lower estimates than DNA analysis, as expected since DNA analysis can detect non-viable cells whereas nuclei counting should not (Sanford et al., 1956). The losses were comparable to 90% losses reported by Korbutt et al. (1996) over 9 days of culture and 84% reported by Yoon et al., (1999) over 8 days of culture.  1x10 n •S 8x10  5  IT 6x10 o  5  co 1 4x10 H 5  o  c  o = 2x10  5  CD  O  6  8  10  12  —i  14  Day After Isolation  Figure 4.1: Cell Concentration Profile of NPCC Suspension Culture. A = counts based on DNA, • = counts based on nuclei counts. Cells were cultured in suspension in standard Ham's F10 medium. Error bars = SEM, n = 3. It was initially believed that overnight shipment without temperature control did not affect cells and that the rate of cell death was correlated to the time after isolation. However, when cells were shipped at times other than 4 days after isolation, an initial  36  sharp decrease in viable cell concentration was often observed in the day after shipment (Figure 4.2), indicating that shipment did contribute to cell death.  Day After Isolation Figure 4.2: Cell Losses in NPCC Culture, Shipment Effects.  Each line represents a different experiment and start of each line indicates day of arrival. To determine if cell lysis products would affect the growth of surviving cells, an additional medium change on the day after arrival was performed to see if it would reduce cell losses; however, no difference was observed (Figure 4.3). Also, supernatants from thefirstmedium change of a previous experiment were saved, frozen and reused in a subsequent culture. Again, no differences were observed (Figure 4.4). It is possible that freezing and thawing the supernatants would deactivate apoptotic factors and the difference between the controls from the two experiments indicated that the freezethawed medium was inferior (Figure 4.3 and Figure 4.4). However, this was not further investigated because the additional medium change did not have an effect.  37  — 2x10° E  "53 ~ 1x10  6  o ro  8 5x1 (f c o  O  "55 O  0  Control  Additional Medium Change  Figure 4.3: Effects of Additional Medium Change on NPCC Survival.  Medium was changed on arrival (day 1) and day 3 for both conditions. The additional medium change was performed on day 2. Cells were counted on day 4. Error bars SEM, n = 9. ^ 2x10  6  E "55  ~ 1x10 o  6  -4—»  CD  § 5x10 c o  5  O  "55 O  0  Control  Supernatant Medium  Figure 4.4: Effects of Using Supernatants From Previous Cultures on NPCC Survival.  Cells were counted 4 days after arrival. Unused medium that was frozen at the same time as supernatants was used in control cultures. Error bars = SEM, n = 3. 4.1.2  Growth Factor Effects on NPCC Survival in Suspension Cultures A preliminary factorial experiment was executed to determine if cell survival in  suspension cultures could be improved by adding growth factors. Since ductal cells were thought to be p-cell progenitors, EGF, HGF, VEGF and KGF were selected based on  38  literature reports that they stimulated ductal cell growth (Table 2.1). A 2 " design with 4 4  1  centre points allowed the assessment of the main effects, several confounded interactions, and curvature (Table 7.3).  The addition of the curvature term reduced the degrees of  freedom in the error, which increased the p-values of all other factors. The cell survival, which is the final concentration divided by the initial concentration, was assessed as a response variable. The losses ranged between 75 and 94% after 8 days. Using p < 0.05 as the criterion, none of the growth factors had a significant effect after 4 days (Table 4.1) or 8 days (Table 4.2) of culture. Table 4.1: Factorial Analysis of Day 4 NPCC Survival in Suspension  Coefficients given are for coded growth factor concentrations. ANOVA is given in Table 7.4. Coefficient  Source Model Intercept EGF VEGF KGF HGF E G F * V E G F aliased with K G F * H G F E G F * K G F aliased with V E G F * H G F E G F * H G F aliased with V E G F * K G F Curvature  P-value 0.3053  0.741 0.069 0.051 -0.039 -0.031 0.019  0.1001 0.1778 0.2768 0.3633 0.5668  -0.031  0.3633  -0.029 -0.056  0.3978 0.3475  Table 4.2: Factorial Analysis of Day 8 NPCC Survival in Suspension  Coefficients given are for coded growth factor concentrations. ANOVA is given in Table 7.5. Coefficient  Source  P-value 0.8641  Model Intercept  0.130  EGF  0.029  0.4695  VEGF KGF HGF E G F * V E G F aliased with K G F * H G F  0.8690 0.5057  E G F * K G F aliased with V E G F * H G F E G F * H G F aliased with V E G F * K G F  0.006 0.026 0.029 -0.016 0.014 0.021  Curvature  0.024  0.7201  39  0.4695 0.6725 0.7193 0.5848  Because the media in the previously reported growth factor studies contained serum, it was possible that the lack of significant effects in the factorial experiments was due to the use of serum-free media.  It was also unclear if the growth factor  concentrations were in the appropriate range. Therefore, dose-response experiments with EGF and HGF, in cultures containing 5% FBS were carried out. There was a trend of decreased cell survival when EGF and HGF were added with saturation of the effect around 5 ng/mL (Figure 4.5).  The data were fit to an exponential decay model and  parameters for EGF, but not HGF, were significant (p<0.05), but the trend of decreasing cell density with the addition of growth factor was the opposite of what was expected. The curvature effect observed in these results was not captured in the factorial experiment, most likely due to the fact that the tested concentrations were much higher. Thus, in the factorial experiments, it was likely that serum was not a missing factor and that the growth factor concentrations were excessive at 50 ng/mL.  •J 2.5x10 i g 2.0x10 H 6  1 2  1.5x10  § c o ^  1.0x10 H  6  6  5.0x10  5  CD  2 c  CO  0.0  0  5  10  15  20  25  Growth Factor Concentration (ng/mL)  Figure 4.5: NPCC Dose Response Curves of EGF and HGF A = EGF, • = HGF. Cells were cultured in standard Ham's F10 supplemented with 5% FBS. Cells were counted after 10 day exposure to growth factors. An exponential decay function was fit to the data: y = 1.23xl0 + 6.62x1 (fexp(-x) for EGF and y = 1.49xl0 +4.0 x Hfexp(-x) for HGF. 6  6  40  In hindsight, the insignificance of terms in the factorial experiment was probably partly due to experimental error as it was performed at a time when there was little experience in handling the tissue. The seeding technique discussed in Section 3.1.2 had not been established at this time, which means that cells may not have been evenly distributed between runs. Nuclei counting was also a method that required some practice. Finally, the addition of serum and some growth factors promoted adhesion of some cells that would have been lost in media changes. In any case, if added factors stimulated growth in some cell populations but not others, total cell numbers would not be an appropriate response variable to assess. Although there are no clear pancreatic stem cell markers, it would have been more useful to study known populations using markers such as vimentin (fibroblasts) and CK7 (ductal cells). The /3-cells were assessed indirectly using insulin/DNA measurements and those results are discussed in Section 2.4.3. 4.1.3 Serum vs. Serum Substitutes and Insulin vs. IGF-1 Serum is widely used in pancreatic studies, even though it is not fully defined and varies from batch to batch. Its richness in proteins, including insulin-like growth factors (Thomas and Fung, 1993), adds variability and confounding effects to experiments. Thus, BIT (BSA, insulin, transferrin), a serum-free medium supplement, was compared to FBS to determine if its effect on cell survival would be different. On average, all cultures had a 60% loss in the first 2 days, but BIT cultures underwent a subsequent growth period whereby 30% of those losses were recovered (Figure 4.6). Although serum contains other components such as adhesion factors and antitrypsin activity (Freshney, 2000), it was demonstrated that replacing serum with a supplement like BIT, may be feasible.  41  — E 1  2.5x10° 2.0x10 H 6  •3,. | 1.5x10 ro c 1.0x10  6  6  CD O  c3 CD  O  5.0x10  5  0.0  8  9 10 11 12 13 14 Day After Isolation  Figure 4.6: Serum, Serum-Free and BIT Supplemented Effects on NPCC Survival • = Ham's F10 alone, A = Ham's FI0 with BIT, O = Ham's F10 with 5% FBS. Error bars = SEM, n = 3. On day 14,p<0.05 for BIT vs. control and BIT vs. 5% FBS. Insulin in the medium can interfere with the identification offi-cells(Rajagopal et al., 2003), so insulin-like growth factor (IGF-1), which shares extensive homology with insulin (Thomas and Fung, 1993), was examined as an insulin substitute. The same cell recovery was observed when a long acting form of IGF-1 was used in place of insulin (Figure 4.7).  However, the cell types were not assessed in these experiments so the  utility of adding IGF-1 or insulin remains unclear.  ~ E  1.0x10 6  g 7.5x10^ c o "co 5.0x10 c 0  5  § 2.5x10 O CD  o  5  0.0  —i—•—i—•—i—•—i—•—i—i—i—i—i—  5  6 7 8 9 10 11 Day After Isolation  Figure 4.7: Comparison of IGF-1 and Insulin Effects on NPCC Survival = Ham's F10 with BIT, A = Ham's FI0 with 50 ng/mL IGF-1. Error bars = SEM, n =  42  4.2  NPCC Growth In Monolayer Culture  4.2.1 Assessment of NPCC Survival in Monolayer Culture NPCCs were cultured in monolayers following the protocol of Tatarkiewicz et al. (2003), which involved dispersion of clusters into a single cell suspension and a preincubation step to selectively remove contaminating fibroblasts. Cells were shipped on the day of isolation rather than 4 days after, as was the case in the suspension culture experiments (Section 4.1) and a RPMI-based medium with 10% FBS was used instead of the serum-free Ham's F10. Serum was used to promote the adhesion of cells to the tissue culture dishes. A sharp decrease in the cell density was observed in the first few days of culture (Figure 4.8), as seen in the suspension cultures and reported by the Tatarkiewicz group. However, the viability remained high because non-viable cells would detach from the surface and be removed by medium changes. Unlike in Figure 4.8, Tatarkiewicz et al. observed a recovery in cell numbers to initial values. One possible reason for the difference is that their cells did not undergo shipment. Also, they used DNA content to evaluate cell numbers whereas manual counts were used in this experiment. 100  8x10  a  E  ^ 6x10  75 _  5  CD  50 =  I 4x10 1 5  CD  CO  C  CD O  >  25  c 2x10 o O  s  "55  O  0  2  4  6  n  8  1  1  10  •  r  12  Day After Isolation Figure 4.8: Cell Concentration Profile of NPCCs in Monolayer Culture  • = cell concentration, A = viability. Cells were cultured in suspension in RPMI based media supplemented with 10% FBS. Error bars = SEM, n = 2.  43  4.2.2 Analysis of CK7+ Cells in NPCC Monolayer Culture After dissociation and fibroblast removal, the ductal cell population (putative islet progenitors) was evaluated in monolayer cultures using immunocytochemistry with CK7 as a marker (Table 4.3). Although total cell numbers decreased, an increase in CK7+ cell numbers was observed between days 1 and 4. Given the increase and assuming that all CK7+ cells were proliferating, they would have had a doubling time of 17 hr, a relatively short time for mammalian cells (Freshney, 2000). An alternative to the increases being due to proliferation is that cells that did not initially express CK7 began to express it in cultures. This is supported by the fact that some cells on days 4 and 9 were relatively weakly stained (Figure 4.9). Whether or not these cells were ductal progenitors was not clear, but the culture conditions did not seem to favour long term expression of CK7, as it was reduced by day 12. Tatarkiewicz et al. (2003) also observed a similar profile in CK7 expression and found that about half of the total cells stained for vimentin by day 12, indicating fibroblast overgrowth.  Table 4.3: CK7+ Cells in Monolayer Cultures Day After  Cell Concentration (cells/mL)  Proportion of  Total C K 7 +  Isolation  C K 7 + Cells  Cells (cells/mL)  1 4 9 12  6.43x10 3.16x10 2.92x10 1.77x10  3% >95% >95% 39%  1.72x10 3.00x10 2.77x10 6.95x10  s  5  5  5  4  5  5  4  B  Figure 4.9: CK7Staining in Monolayer Cultures A: weak staining observed on days 4 and 9. B: dark staining observed on all days.  44  4.2.3  Targeted Expansion of CK7+ Cells Using Growth Factors A factorial experiment was carried out to determine if CK7+ cells could be further  expanded in monolayer cultures. EGF, VEGF, KGF and HGF, the same growth factors as in the previous factorial experiment (Section 4.1.2), were tested. Basic fibroblast growth factor (bFGF) was added because it was used to expand nestin+ cells (Habener, 2001) and NPCC monolayers, although ductal cell expansion was not assessed (Korbutt, personal communication).  A 2 " design with A centre points was used to provide 5  1  estimates for all main and 2-factor interaction effects.  All high concentrations were  reduced from the last factorial experiment value of 50 ng/mL to 20 ng/mL given the dose response experiment results that showed 50 ng/mL was too high (Section 4.1.2), except for HGF, which was reduced to 16.7 ng/mL due to the quantity that was readily available. Low concentrations were tested at 0 ng/mL. Proliferation was assessed by expression of the proliferation marker, Ki67. The initial density was 4.1xl0 cells/mL, and only 3% 6  and 2% of cells were CK7+ and Ki67+, respectively.  Rather than discarding the  supernatant from the media change on day 4, it was counted and transferred to new tissue cultures dishes. Cell concentrations from day 4 supernatants, day 7 supernatant, and day 7 monolayer cultures were analyzed. The proportions of CK7+, Ki67, CK7+ without Ki67 (proliferation in CK7+ population), and total concentrations of CK7+ were also assessed. 4.2.3.1 Supernatant Cell Concentration The day 4 supernatant cell concentration represents the cells that did not adhere to the tissue culture treated dishes, and would normally be discarded during medium changes. Therefore a negative effect meant that more cells remained attached to the  45  dishes. It is evident that all of the main and interactive effects were significant and that EGF, HGF and bFGF promoted cell adhesion whereas VEGF and KGF promoted suspension cultures (Table 4.4).  EGF and HGF acted synergistically (more than  additively) as the coefficient for their interaction had the same sign as their main effects. Similarly, EGF*bFGF and HGF*bFGF had synergistic relationships. VEGF and KGF reduced the effects of EGF, HGF, and bFGF because their interactions were positive. The VEGF and KGF interaction was negative such that their positive main effects were overlapping and less than additive. Figure 4.10 illustrates the interactions between some of the growth factors. Table 4.4: Factorial Analysis of Day 4 Supernatant Cell Concentration Coefficients given are for coded growth factor concentrations. Cell concentration is in cells/mL. One of the runs was missing a value, which aliased HGF*bFGF with all of the other terms. ANOVA is given in Table 7.5. Source Model  Coefficient  Intercept  EGF VEGF KGF HGF bFGF EGF*VEGF EGF*KGF EGF*HGF EGF*bFGF VEGF*KGF VEGF*HGF VEGF*bFGF KGF*HGF KGF*bFGF HGF*bFGF  1.25x10  P-Value 0.0008  s  -2.76x10" 5.23x10 4  4.50x10 -5.32x10  4  -4.13x10  4  4  3.26x10" 5.01x10" -2.84x10" -2.78x10" -3.26x10" 3.07x10" 4.19x10" 4.47x10" 3.91x10" -2.36x10"  0.0009 0.0001 0.0002 0.0001 0.0003 0.0006 0.0002 0.0008 0.0009 0.0006 0.0007 0.0003 0.0002 0.0003 0.0014  The supernatant cells were cultured for an additional 3 days in the spent media and their expansions also analyzed (Table 4.5). No factors were significant, supporting the interpretation that the results observed on day 4 were due to influences on cell adhesion rather than on cell proliferation. This could also explain the results in the dose  46  response experiments where EGF and HGF were found to decrease the final cell concentration in suspension cultures (Section 4.1.2). The promotion of cell adhesion likely increased the amount of cells lost from medium changes. c  •S  2.50x10 " s  0  5  10  15  20  0  5  EGF (ng/mL)  10  15  20  E G F (ng/mL)  0  5  10  15  20  KGF (ng/mL)  Figure 4.10: Interaction Plots for Day 4 Supernatant Cell Concentration Interaction plots for EGF*HGF (A), VEGF*EGF (B), VEGF*KGF (C). For each plot, other factors are at centre point values. Synergistic - EGF has an effect only when HGF is added (A). Antagonistic - VEGF addition reduces the effect of EGF (B) and KGF (C). Table 4.5: Factorial Analysis of Day 4 to 7 Supernatant Cell Expansion Coefficients given are for coded growth factor concentrations. Cell concentration is in cells/mL. ANOVA is given in Table 7.14. Source Model Intercept EGF VEGF KGF HGF bFGF EGF*VEGF EGF*KGF EGF*HGF EGF*bFGF VEGF*KGF VEGF*HGF VEGF*bFGF KGF*HGF KGF*bFGF HGF*bFGF  Coefficient 0.300 0.050 -0.045 -0.096 0.075 0.074 -0.035 -0.100 0.082 0.031 0.043 -0.019 -0.085 -0.068 -0.100 0.040  47  P-Value 0.4036 0.4201 0.4635 0.1730 0.2572 0.2655 0.5639 0.1493 0.2258 0.6082 0.4871 0.7435 0.2148 0.2979 0.1591 0.5083  4.2.3.2 Day 7 Monolayer Total Cell Concentration When the total cell concentrations of the monolayers at day 7 were analyzed, HGF, VEGF*HGF, VEGF*bFGF, and curvature were found to be significant in the full model (Table 4.6). The overall modelfitwas not significant; therefore, a second model was made which included VEGF, HGF, bFGF, VEGF*HGF, VEGF*bFGF (data not shown). Although VEGF and bFGF were not significant on their own, they were still included because they were involved in an apparently significant interaction. The residual plots of the second model showed run 19 to be an outlier (Figure 4.11), and when this point was excluded, VEGF*bFGF became insignificant and VEGF became significant. The reduced model is shown in Table 4.7. HGF increased monolayer cell concentrations, whereas VEGF decreased them, which may be attributable to the VEGF reduction of cell adhesion discussed in the supernatant analysis. The positive curvature term indicates that a local maximum may exist within the range of the concentrations tested. Table 4.6: Factorial Analysis of Monolayer Total Cell Concentration Coefficients given are for coded growth factor concentrations. Cell concentration is in cells/mL. ANOVA is given in Table 7.15. Source  Coefficient  0.0766  Model Intercept EGF  P-Value  1.97x10  s  -8.75x10 -1.69x10" 1.50x10 3.16x10" 1.41x10"  0.2863  EGF*VEGF EGF'KGF  1.00x10 1.74x10"  EGF*HGF EGF*bFGF  -1.00x10 -2.03x10"  0.8918 0.0825 0.8918  VEGF*KGF VEGF*HGF  -6.25x10 -2.81x10"  VEGF*bFGF  2.79x10"  0.0253 0.0259  1.40x10"  0.1302  6.75x10 -1.54x10" 7.31x10"  0.3918  VEGF KGF HGF bFGF  KGF*HGF KGF*bFGF HGF*bFGF Curvature  3  3  3  3  3  3  48  0.0881 0.8387 0.0185 0.1279  0.0579 0.4235  0.1076 0.0169  3.001"  S  1.501  0)  •  ce  | o.oot  °  •  •  •  •  •  •  o  •  -1.501  -3.001" I  1  1  I  1  1  4  ' I  1  1  I  7  1  ' I  10  1  1  I  13  ' I  1  1  16 19  Run Number  Figure 4.11: Residual Plot for Day 7 Monolayer Cell Concentration Model includes VEGF, HGF, bFGF, VEGF*HGF,  VEGF*bFGF  and curvature terms.  Table 4.7: Factorial Analysis of Monolayer Total Cell Concentration, Reduced Model Coefficients given are for coded growth factor concentrations. cells/mL. ANOVA is given in Table 7.16. Source Model  Coefficient  Intercept  2.06x10  VEGF HGF VEGF*HGF Curvature  -2.57x10 4.05x10 -3.70x10  Cell concentration  is in  P-Value 0.0009  5  4  4  4  6.43x10"  Lack of Fit  0.0311 0.0021 0.0039 0.0151 0.2255  4.2.3.3 Day 7 Proportion of CK7+ Cells The proportion of cells that were CK7+ was assessed. On average, only 50% of cells were CK7+, which was reduced from the case where no growth factors were added (Section 4.2.2). In the full model, VEGF, all EGF interactions, VEGF*HGF, KGF*HGF, KGF*bFGF, HGF*bFGF and curvature were significant (Table 4.8). Removal of the non-significant interactions from the model did not change the significance of the other terms (data not shown). The positive curvature term suggests that a maximum may have been somewhere between the tested concentration levels.  49  Table 4.8: Factorial Analysis of Proportion of CK7+ Cells in Day 7 Monolayers Coefficients given are for coded growth factor concentrations. Proportion is in %. ANOVA is given in Table 7.17. Source Model Intercept EGF VEGF KGF : HGF bFGF EGF*VEGF EGF*KGF EGF*HGF EGF*bFGF VEGF*KGF VEGF*HGF VEGF*bFGF KGF*HGF KGF*bFGF HGF*bFGF Curvature  Coefficient 48.50 -1.61 6.50 -2.25 5.09 1.42 -22.17 9.19 -8.61 10.51 -2.42 -17.05 3.23 -8.32 -9.94 -11.97 41.59  P-Value 0.0090 0.4432 0.0382 0.3067 0.0692 0.4963 0.0012 0.0153 0.0182 0.0105 0.2786 0.0026 0.1759 0.0200 0.0123 0.0073 0.0020  4.2.3.4 Day 7 Proportion of Ki67+ Cells The fraction of proliferating cells in the total population was measured by the analysis of Ki67 expression.  The average proportion of proliferating cells was only  around 5%, which was only a little higher than the 2% observed on day 1. The full model showed that EGF, EGF*VEGF, EGF*KGF, VEGF*HGF, and KGF*bFGF were significant (Table 4.9). The model was reduced by removing all other interaction terms, except EGF*bFGF, which was found to be significant once the model was reduced (Table 4.10). It is interesting EGF was the only significant main effect, despite the reported mitogenic activity of the other factors.  KGF, HGF, and bFGF appeared to  enhance EGF's effect, whereas VEGF decreased it. The strongest interaction did not actually involve EGF, but rather VEGF and HGF. VEGF*HGF was a recurring strong antagonistic interaction. Curvature was not significant in this analysis, indicating that a linear relationship was appropriate.  50  Table 4.9: Factorial Analysis of Proportion of Ki6 7+ Cells in Day 7 Monolayers  Coefficients given are for coded growth factor concentrations. Proportion ANOVA is given in Table 7.18. Source Model  Coefficient  Intercept  4.53 1.30 0.13 -0.65 -0.30  EGF VEGF KGF HGF bFGF  -0.69 -1.65 1.53  EGF*VEGF EGF*KGF EGF*HGF EGF*bFGF VEGF*KGF  VEGF*HGF VEGF*bFGF KGF'HGF  KGF*bFGF HGF'bFGF Curvature  0.13 1.17 -0.72 -2.84 -0.86 0.029 -1.28 0.16 1.01  is in %.  P-Value 0.0489 0.0443 0.7561 0.1943 0.4922 0.1768  0.0241 0.0294 0.7539 0.0577 0.1615  0.0053 0.1132 0.9458  0.0459 0.7045 0.3304  Table 4.10: Factorial Analysis of Proportion of Ki67+ Cells in Day 7 Monolayers, Reduced Model  Coefficients given are for coded growth factor concentrations. Proportion is in %. ANOVA is given in Table 7.19. Source Model  Coefficient  P-Value 0.0017  Intercept  4.74  EGF VEGF  1.30  0.0213  0.13  0.7833  KGF  -0.65  0.1987  HGF bFGF  -0.30 -0.69 -1.65  0.5322  EGF*VEGF EGF*KGF EGF*bFGF VEGF*HGF KGF*bFGF  1.53 1.17 -2.84 -1.28  0.1770  0.0064 0.0097 0.0340 0.0002 0.0227 0.3624  Lack of Fit  4.2.3.5 Day 7 Proportion of CK7+ Cells with Ki67 The proportion of CK7+ cells with Ki67 was a measure of the desired proliferation in the CK7+ population alone. Factors that stimulated proliferation in the  51  CK7+ population would be candidates for subsequent studies. In the full model (Table 4.11), V E G F was found to reduce the proliferation of CK7+ cells on day 6. This would normally be surprising given that V E G F increased the proportion of CK7+ cells (Table 4.8). However, V E G F also decreased cell adhesion (Table 4.4) and may have reduced the amount of fibroblasts in the cultures to yield a higher proportion of CK7+ cells. K G F and HGF, on their own, increased proliferation, and they interacted synergistically (Table 4.11). E G F and V E G F reduced the effects of K G F and HGF, whereas bFGF enhanced the effect of HGF (Table 4.11). The negative curvature term indicates that local minima may exist in the concentration ranges tested. A refined model is not presented because the removal of non-significant terms did not change the significance of other terms. Table 4.11: Factorial Analysis of Proportion of CK7+ Cells Also Expressing Ki67 in Day 7 Monolayers Coefficients given are for coded growth factor concentrations. Proportion is in %. ANOVA is given in Table 7.20. Source  Coefficient  Model Intercept  P-Value 0.0016  5.70  EGF VEGF KGF HGF bFGF EGF*VEGF EGF*KGF EGF*HGF EGF*bFGF VEGF*KGF VEGF*HGF VEGF*bFGF KGF*HGF KGF*bFGF HGF*bFGF Curvature  -0.26  0.2530  -1.09 1.27 1.01 0.52  0.0091 0.0059 0.0114  0.31  0.0631 0.1799  0.31  0.1861  -1.63  0.0029 0.0303 0.0020 0.0002 0.0005 0.0110  -0.70 -1.85 -4.38 -2.89 1.02 0.34 2.32 -1.99  52  0.1558  0.0010 0.0160  4.2.3.6 Day 7 Concentration of CK7+ Cells The proportion of CK7+ cells was multiplied by the monolayer cell concentration to obtain the total concentration of CK7+ cells. In the full model (Table 4.12), HGF significantly increased the total number of CK7+ cells but in the reduced model (Table 4.13), it lost its significance as a main effect. HGF still interacted antagonistically with VEGF. bFGF also interacted antagonistically with HGF. All main effects, except HGF, were negative, meaning that overall they reduced the number of CK7+ cells in the monolayer cultures. The curvature appeared to be highly significant in that the average of the centre points was about 50% higher than the average of the factorial points. Thus, it was likely that the tested growth factor concentrations were higher than necessary. Table 4.12: Factorial Analysis of Day 7 Concentration of CK7+ Cells in Monolayers  Coefficients given are for coded growth factor concentrations. cells/mL. ANOVA is given in Table 7.21. Source Model Intercept  EGF VEGF KGF HGF bFGF EGF*VEGF EGF*KGF EGF*HGF EGF*bFGF VEGF*KGF VEGF*HGF VEGF*bFGF KGF*HGF KGF*bFGF HGF*bFGF Curvature  Coefficient 1.02x10 -4.56x10  P-Value 0.0257  s  -1.24x10 -6.68x10 2.62x10 -2.72x10  3  2  3  4  -4.33x10 3.04x10  3  4  4  0.5441 0.9863 0.3911  0.0294 0.7107  0.0074 0.0199  -1.06x10"  0.2100  6.37x10 -1.13x10"  0.4106 0.1883  -4.67x10" 3.58x10" -1.63x10"  0.0060 0.0127  -2.16x10"  0.0481  -3.64x10" 1.42x10  0.0122 0.0025  3  5  53  0.0921  Concentration is in  Table 4.13: Factorial Analysis of Day 7 Concentration of CK7+ Cells in Monolayers, Reduced Model  Coefficients given are for coded growth factor concentrations. cells/mL. ANOVA is given in Table 7.22. Source  Coefficient  Model Intercept EGF VEGF KGF HGF bFGF EGF*VEGF EGF*KGF VEGF*HGF VEGF*bFGF HGF*bFGF  Curvature  Concentration is in  P-Value 0.0100  1.02x10 -4.56x10 -1.24x10 -6.68x10 2.62x10 -2.72x10 -4.33x10" 3.04x10 5  3  2  3  4  3  4  -4.67x10" 3.58x10" -3.64x10" 1.42x10 5  Lack of Fit  0.7136 0.9920 0.5926 0.0600 0.8258 0.0068 0.0350 0.0046 0.0174 0.0162 0.0007 0.1212  4.2.3.7 Summary of Targeted Expansion of CK7+ Cells Using Growth Factors In this experiment, EGF promoted cell attachment (Table 4.4) and proliferation (Table 4.9), but not CK7+ cell proliferation (Table 4.11). Cras-Meneur et al. (2001) reported that EGF enhanced overall cell growth but reduced endocrine proportions in rat embryonic pancreatic epithelia.  Upon its removal, endocrine proportions increased,  suggesting that EGF stimulated growth in islet progenitors. The cells that were expanded stained positive for CK (type not specified). The discrepancy between these results and Cras-Meneur et al.'s could be due to the different systems used (embryonic rat vs. neonatal pig; collagen matrix vs. monolayer). VEGF reduced cell attachment (Table 4.4) and decreased ductal cell proliferation (Table 4.11). Its positive effect on the proportion of CK7+ cells (Table 4.8) was likely due to a reduction of contaminating fibroblasts, by reducing cell attachment. Oberg et al. (1994) observed that 20 ng/mL of VEGF increased proliferation of ductal extracts, determined by [ H]thymidine (proliferation marker) labeling, but they did not assess 3  54  proliferation in the ductal cell population specifically.  Rooman et al. (1997) assessed  cultures with double staining of BrDU (proliferation marker) and C O O (rat ductal cell marker), and found that VEGF increased proliferation of ductal cells by about 2-fold. However, CK expression in ductal cells can vary (Bouwens, 1998a). KGF was shown to stimulate proliferation of CK7+ cells (Table 4.11), but it did not have an effect on the overall number of CK7+ cells (Table 4.8 and Table 4.13). It is possible that this was due to a reduction in ductal cell attachment (Table 4.4). BonnerWeir et al. (2000) used 10 ng/mL of KGF to expand their ductal cultures, but they did not actually assess its effect on cells by comparison with cultures without KGF. Elghazi et al. (2002) observed a 3-fold expansion in rat embryonic epithelia when KGF was added, but they cultured their cells in a 3D collagen matrix where cells would not be divided into two phases (supernatant vs. monolayer). Of the five factors tested, HGF yielded the most promising results in that it promoted both cell attachment (Table 4.4) and ductal cell proliferation (Table 4.11). There seemed to be a strong antagonistic interaction between VEGF and HGF, and it is likely that HGF's effects were partly masked in this experiment. Lefebvre et al. (1998) found that HGF specifically stimulated proliferation in CK19+ (human ductal marker) cells and Beattie et al. (1999) used it to expand their Pdx-1+ cells by 30000-fold. Habener et al. (2001) added bFGF to expand their nestin positive cells, but they did not report the effect of not adding bFGF. bFGF promoted cell attachment (Table 4.4), but not proliferation of CK7+ cells specifically (Table 4.11). In terms of overall CK7+ cell number, bFGF did not have a significant effect on its own, but it reduced the  55  effects of HGF (Table 4.13). Overall, it was not clear if bFGF was beneficial to the ductal cultures. The interactions were often more significant than the main effects. Although examining effects one-factor-at-a-time would miss these interactions, it is still the most popular experimental approach in biological studies. A pitfall of including interactions is that one main effect can be masked by another, as HGF was by VEGF. To address this, a one-factor-at-a-time approach could have been used to compare all of the main effects and two-factor interactions to a control. But for 5 factors and 10 two-factor interactions, this would have required 32 runs (16 runs multiplied by 2 replicates), not including a curvature estimate. The relative efficiency of the 2 " design with 4 centre points (20 5  1  runs) is 1.6(32/20). However, the above comparison is relatively modest because factorial designs have an inherent high degree of replication. For each factor, half the runs in a two-level factorial are at a high level, and in the other half are at a low level. Thus, the degree of replication is equal to the number of runs divided by 2. For a 2 " design, not including 5  1  centre points, the degree of replication is 8 (2 "'/2). To achieve the same degree of 5  replication in the above one-factor-at-a time example requires 128 runs (16x8). Even when interactions are not estimated, the factorial design is more efficient than the onefactor-at-a-time approach, and the relative efficiency increases as the number of factors increase (Montgomery 1997). The factorial design performed was a screening tool to help select the most significant factors for further examination. Many of the growth factors actually reduced CK7+ proliferation and attachment, which was probably why the proportions of CK7+  56  cells were lower than in cultures where no growth factors were added (Section 4.2.2). In a subsequent experiment, the promising factors, such as KGF and HGF, could be examined in further detail without the antagonistic interactions of the other variables. However, KGF's negative effect on cell attachment would have to be considered. An alternative approach would be to decouple cell attachment and proliferation. Using the reduced models to predict responses, it was found that adding VEGF and HGF would maximize the concentration of CK7+ cells and adding KGF, HGF, and bFGF would maximize their proliferation (Table 7.23). A two step approach could be taken whereby VEGF and HGF are added to increase CK7+ cell selection and attachment, followed by VEGF removal and KGF and bFGF addition to stimulate proliferation. Curvature was also significant in many of the responses, but it was not clear which factors had non-linear effects. An advantage of factorial design is that runs can be subsequently added to estimate curvature effects for individual factors (Montgomery, 1997). The significance of the curvature also highlighted a difficulty in factorial design. Selection of the factor ranges is important in determining the outcome, as demonstrated in the first factorial experiment (Section 4.1.2) where the growth factor concentrations were excessive. A dose response experiment, can be used first to elucidate the proper concentration ranges, but the number of runs required can become too high as more factors are added.  4.3  P-Cell Maturation in NPCCs Sections 4.1 and 4.2 discussed ways in which NPCCs, in particular the ductal  (putative progenitor) population, could be expanded. Once the progenitors are expanded, they must be differentiated towards a /3-cell phenotype.  57  4.3.1  Assessment of Insulin Content of NPCCs in Suspension  Culture  When NPCCs were cultured in suspension for 9 days, the total insulin values remained relatively constant while the insulin/DNA increased over time (Figure 4.12). Yoon et al. (1999) reported a similar 2-fold increase in insulin/DNA over 7 days whereas Korbutt et al. (1996) found a 6-fold increase over 9 days. The differences could be attributed to the variability among pigs and the fact that data collection started 5 days after isolation (Figure 4.12) rather than immediately. Nevertheless, a trend of increasing insulin content was observed.  0.5  0.10 0.08  % 0.4  1  A  ' \\  0.3  •  w 0.2  1 I-  \  0.1 0.0  0.06  8  10  12  <  0.04 Q I  0.02 6  "a. 5)  "5 co c  0.00  Day After Isolation  Figure 4.12: Typical Insulin Content Profile in NPCC Cultures. A = insulin/DNA, • total insulin. I cell = 7.1 pg DNA. suspension in standard Ham's F10 medium. 4.3.2  Growth Factor Effects on Insulin/DNA  Cells were cultured in  of NPCCs in Suspension  Culture  To examine the effects of different growth factors on insulin content, insulin/DNA ratios from the factorial experiment (Section 4.1.2) were analyzed after 4 and 8 day exposures to EGF, VEGF, HGF and bFGF, all at 50 ng/mL.. When the full model was assessed for the day 4 response, EGF*VEGF aliased with KGF*HGF was the only significant term, but the model itself was not significant (Table 4.14). Reducing the  58  model did not change the significance of terms or the model (data not shown). When the values were examined on day 8, EGF and EGF*HGF aliased with VEGF*HGF were significant but again, the model itself was not (Table 4.15). Refining the model increased the model significance, but the EGF*HGF interaction was no longer significant (data not shown). Overall there was only strong evidence to say that EGF had a negative effect on insulin/DNA and it was only observed after 8 days of exposure.  Cras-Meneur et al.  (2002) similarly found EGF reduced the insulin content in rat embryonic pancreatic epithelia, and a recovery was only observed after its removal. Table 4.14: Factorial Analysis of Day 4 Insulin/DNA of NPCC Suspension Cultures  Coefficients are for coded terms. Insulin/DNA is in [ig/ptg. ANOVA is given in Table 7.24 Coefficient  Source Model Intercept EGF VEGF KGF HGF EGF*VEGF aliased with KGF*HGF EGF*KGF aliased with VEGF*HGF EGF*HGF aliased with VEGF*HGF  P-Value 0.3060  6.85x10" 0 2.00x10" 5.25x10" 2.50x10  2  1.0000 0.6024 0.2250 0.9467  3  3  -4  1.10x10' 1.75x10" 3.25x10"  0.0496 0.6466 0.4152  2  3  3  1.0000  0  Curvature  Table 4.15: Factorial Analysis of Day 8 Insulin/DNA of NPCC Suspension Cultures Coefficients are for coded terms. Insulin/DNA is in fig/fig. ANOVA is given in  Table 7.25 Coefficient  Source  P-Value 0.0537  Model 5.58x10" -1.75x10"  Intercept EGF  2  2  0.0092 1.0000  0  VEGF KGF  -7.00x10-  3  0.0952  HGF EGF*VEGF aliased with KGF*HGF  -3.50x10" 1.75x10"  3  0.3149 0.5896  EGF*KGF aliased with VEGF*HGF EGF*HGF aliased with VEGF*HGF  5.75x10" 9.75x10" -1.10x10'  Curvature  59  3  3  3  2  0.1423 0.0439 0.1168  4.3.3 Alginate and Serum Effects on NPCC Maturation When NPCCs were immobilized in 0.75% alginate and cultured with 5% neonatal pig serum (NPS), the increases in insulin/DNA were enhanced 3.9-fold (p<0.05) in comparison to controls (Figure 4.13) and approached values found in adult pancreata of 2 fig/fig  (Yoon et al., 1999). No difference was observed in DNA content (640 ± 112 ng in  controls vs. 645 ± 168 ng in alginate/NPS cultures), indicating that cell survival was not improved. The increase in insulin/DNA could have corresponded to increased insulin production in existing /3-cells, increased numbers of /3-cells, preferential survival of /3cells or a combination of the three. Unfortunately, when the immunocytochemistry slides were examined, there was no staining of CK7, vimentin, insulin or Ki67, so the numbers of insulin positive cells could not be obtained. It is possible that alginate interfered with the analysis. Tatarkiewicz et al. (2001) stained their NPCC sections for insulin and found no noticeable difference between cells cultured in serum-free media or 10% FBS and alginate. This would suggest that FBS and alginate did not increase /3-numbers, and it is possible that insulin was taken up from the serum. When NPCCs were immobilized in alginate and cultured with 5% FBS instead of 5% NPS, no difference was observed between control and alginate cultures (0.18 ± 0.05 U-g/u-g and 0.22 u.g/ug, respectively). This indicated that the efficacy of the serum may be species dependant. The mechanism for the alginate maturation effect on NPCCs was not clear. However, the formation of aggregates in a 3-dimension matrix was reported to have a differentiating effect on pancreatic cultures (Table 2.2). Also, differentiation was shown to occur when monolayers reach confluency (Zulewski et al., 2001). It is possible that alginate has a constricting effect on cells that promotes differentiation.  60  2.0  .a.  < z  1.0-  a  I- .' ' • I  Day 1  lr • •  .'VI  Day 8  ]'  Day 18 Control  I  I  I  Day 18 Alginate +5% NPS  Figure 4.13: Effect of Alginate and NPS on Insulin/DNA of NPCCs Mean insulin/DNA ratios of cultures on arrival (Day 1 after isolation), after one week of culture in standard medium (Day 8), and end of experiment (Day 18). Cultures in alginate slabs overlaid with medium supplemented with 5% NPS were started on day 8 with a parallel control culture in standard medium. Error bars = standard deviation. 4.4  Alginate Immobilization in HFBR Approximately 10 islet equivalents, or 1.5 x 10 cells, are required per patient 6  9  (assuming a 100 kg person) to correct hyperglycemia (Shapiro et al., 2000). Alginate cultures are currently in the form of beads or slabs. Bead cultures are readily scaleable, but require a bead generator with sufficient capacity. Alginate slabs in stationary cultures are limited to densities around 10 cells/mL. The required volume of 1500 mL would not 6  be practical to implement in a stationary culture. In contrast, HFBRs can support very high cell densities (>10 cells/mL). A 100 mL HFBR would have the capacity to culture 8  enough cells for 6 patients, but a single HFBR could also be used on a per patient basis. To determine the feasibility of alginate immobilization of cells in a HFBR, 3 issues were examined: potential plugging of ICS fibres, degree of gelling and cell recovery. Once a working protocol was determined, cell growth was tested in a full bioreactor run.  61  4.4.1  Use of CHO KI Cells to Model Alginate Immobilization of Cells in HFBRs It was not practical to use NPCCs in the H F B R studies due to the large quantity of  cells that was required. Therefore, CHO K I cells were used to test the scale-up in the HFBR.  A lower grade and less costly alginate (Sigma) was compared to Pronova  alginate. The normally adherent CHO cells were adapted to form cell clusters resembling NPCCs by culturing them on non-treated bacteriological dishes. Regardless of alginate brand, clusters cultured in alginate did not grow as fast as in suspension (Figure 4.14), but they remained viable. Hence, it was concluded that C H O cells immobilized in Sigma alginate were suitable for scale-up studies. 10  c o c  'tn  CO Q.  X  UJ  "55  O >. co Q  mm  Control Alginate Pronova  Control Alginate Sigma  Figure 4.14: Comparison of CHO Cells Grown In Alginate  4 day cell expansion of CHO cells in Pronova and Sigma alginate. Cells were either cultured in medium alone (Control) or in alginate slabs overlaid with medium (Alginate). Error bars = SEM. 4.4.2  Determination of Alginate Plugging Alginate has a molecular weight range of 12 to 80 kDa (Sigma technical support)  and is linear in its configuration, whereas the molecular weight cut-off of the fibre membranes for globular proteins is approximately 18 kDa. It was anticipated when the alginate was added to the ECS, that alginate would cross into the ICS and plug the fibers when gelled. The degree of plugging was determined by residence time distribution  62  analysis. Azoalbumin has a molecular weight of 70 kDa and was not expected to cross the fibre membrane. Therefore, when it was pulsed through the ICS, the protein elution curve was indicative of the fluid velocity through the fibres. The general assumption of the analysis was that if fibers were plugged, the cross-sectional area for flow would be reduced, the protein would elute faster, and residence times would decrease. In the first experiment, the opposite result was observed when the protein residence times were longer after the cartridge was filled with alginate (data not shown). However, the alginate was loaded a day after the alginate-free runs were executed and distilled water was recirculated in that time period.  It was shown in a subsequent  experiment that the residence times would increase as rinsing time increased (Figure 4.15). It was also observed that air bubbles were often trapped in the fibers and it took several days flush them out. The bubbles would have reduced the cross-sectional area of the passable fibers, increasing the fluid velocity. This indicated that the fibers would have to be adequately flushed before starting the pulse trials. The experiment was repeated with the incorporation of longer rinsing times (> 48 h) and alginate was only loaded after sequential runs yielded similar elution curves. A comparison of azoalbumin (5 g/L in saline) elution curves before and after loading alginate was first examined at an ICS flow rate of 115 mL/min (Figure 4.16A). At this low flow rate, the shorter residence times after loading the alginate implied that alginate plugged the ICS fibres. In an attempt to reduce the plugging, the ICS flow rate was increased to 300 mL/min and the Ca while loading the alginate in the ECS. alginate loading (Figure 4.16B).  2+  solution was recirculated through the ICS  However, the protein still eluted faster after  When the flow rate was simply increased to 300  63  mL/min, without Ca  z+  recirculation during alginate loading, the before and after curves  became nearly indistinguishable (Figure 4.16C).  1.5  n  Time After Start of Pulse (s)  Figure 4.15: Effect of Rinsing Time on Azoalbumin Elution Curves  Arrow indicates direction of increasing rinse time: 0, 24, 47, 91 hr. As the HFBR was rinsed for longer periods of time, the elution velocities decreased, and residence times increased. Experimental flow rate = 100 mL/min. Rinsing flow rate = 150 mL/min. The elution curves had a Poisson-like distribution (Figure 4.15). The flow in the tubing was laminar, yielding a parabolic velocity profile. The fluid along the channel walls would be stagnant resulting in longer residence times and a tailing effect in the elution curve. If the length traveled in the tubing were longer, Taylor dispersion in the tubing and ICS likely would give a Gaussian shape to the elution curve. Tailing of the curves could be also due to protein interactions with the alginate and fibres that would slow the elution of the protein. Another explanation for the tailing effect is that the entry point of the ICS manifold (Figure 2.4) is central to the fibre bundle. Some of the protein pulse would immediately enter the fibres in the centre, whereas there would be a slight delay in reaching the fibres radially farther from the entry point. This effect would be enhanced by the manifold of the ICS exit.  64  2.500n  2.0001  1.500-  1.000-  0.500 i  0.500  0.000 100  120  r  140  T i m e After Start of P u l s e (s)  Figure 4.16: Azoalbumin Elution Curves Before and After Loading Alginate Black lines = before loading alginate. Red lines = after loading alginate. Plot A: ICS flow rate = 115 mL/min. Plot B: ICS flow rate = 300 mL/min, recirculated gelling solution through ICS while loading alginate in ECS. Plot C: ICS flow rate = 300 mL/min.  65  From Equation 3.3, the calculated average residence times (t  res  ,ave)  confirmed that  increasing flow rates reduced fibre plugging whereas recirculating C a solution through 2+  the ICS during alginate loading increased it (Table 4.16). The decrease in t  res  ,ave  indicated  that overall, the degree of plugging was not very high. Table 4.16: Effect of Alginate on Average Residence Times of Azoalbumin in HFBR tres.ave  was calculated from Equation 3.3. ICS Flow rate (mL/min)  ICS Flow During Loading  tres.ave Before Loading (s)  tres.ave A f t G T  Decrease in  Loading (s)  tres.ave  115 300 300  No Yes , No  64.4 27.2 26.0  60.7 25.0 25.5  6% 9% 2%  4.4.3 Degree of Gelling To observe the degree of gel formation after recirculating a C a solution through 2+  the ICS, HFBRs were sawed open and examined.  Alginate concentrations and  recirculation times were varied from 1 to 3% and 0.5 to 1 hr, respectively. Regardless of concentration or recirculation time, gel was found between the fibres, around the bundle of fibres, and in the ECS manifold region (Figure 2.4). Because the ECS manifold is radially furthest from the ICS and was found to be filled with gelled alginate, it was concluded that 0.5 hr was a sufficient gelling time. 4.4.4 Cell Recovery For protein production, the cells do not need to be recoveredfromthe ECS at the end of a run. In a tissue engineering application where cells are the product, it was important to determine whether or not cells could easily be recovered from the reactor. To test this, HFBRs were loaded with 120 mL of alginate containing a known amount of cells. A Ca  2+  solution was recirculated through the ICS for 0.5 hr, and then a sodium  citrate solution was recirculated through the ICS for 1 hr. Recovery was tested for 1 and  66  3% alginate. Because the experiments were carried out at room temperature within 4 hours, it was assumed that cell growth was minimal. At the end of the citrate recirculation, 40 to 75 mL of liquefied alginate was collected from the ECS, but gelled alginate could still be seen in the cartridge. To recover the remaining alginate, about half of the ECS space was filled with citrate solution, leaving an air gap to facilitate mixing. The cartridge was manually rotated several times to degel the alginate, and the liquid was collected.  These steps were  repeated, and it took approximately 4 to 5 passes to degel most of the gel (Figure 4.17). The greater than 100% recovery was likely due to errors in cell counting. Since the seeding concentrations were not very high (1.2 to 1.5 x 10 cells/mL), by the time the 5  liquefied alginate was diluted with citrate solution, concentrations were as low as 5 x 10  4  cells/mL.  Number of Passes  Figure 4.17: Cumulative Cell Recovery After Each Degelling Pass A = 3% alginate, • 1% alginate. 1 pass = recovery of liquefied alginate immediately after recirculating citrate solution through ICSfor 1 hr. Error bar = 1 SEM, n = 4. st  Another run using 1% alginate was performed, although the recovery after each pass was not recorded. Figure 4.18 shows the total recovery in each run. Regardless of  67  concentration, the recoveries were high. This was surprising because it was expected that cells between the fibres would be difficult to recover. However, in HFBRs, cells usually grow up to packed cell densities (Piret and Clooney, 1990) and can attach to the fibres. It is possible that alginate immobilization actually reduced adhesion to the fibres, making cell recovery relatively easy. There was a large variability in the recovery values, which could be attributed to insufficient mixing of cells in the alginate. Nevertheless, the results were promising enough to carry out a cell growth experiment in the HFBR.  150  N  Alginate Concentration Figure 4.18: Total Cell Recovery at Different Alginate Concentrations 4.4.5  Cell Growth in Alginate in the HFBR  Once a working protocol was established for gelling the alginate and recovering the cells, a run was carried out to demonstrate that cells would grow in an alginate immobilized HFBR. CHO cells were seeded at 4.4 x 10 cells/mL alginate in a HFBR 5  and 2 slab cultures in tissue culture dishes. Because the ECS could not be sampled for cell growth directly, the glucose and lactate concentrations were used to monitor growth indirectly.  The starting glucose concentration of the medium was 25 m M , but was  diluted to 20 m M by the alginate in the slab cultures (2.5 mL alginate per 10 mL  68  medium). The dilution of the medium in the HFBR was greater because some of the medium was used to flush out the other solutions in the tubing and ICS. The final volume of diluted media that was recirculated through HFBR was approximately 0.6 L. In the slab cultures, glucose levels decreased and lactate levels increased (Figure 4.19), concomitant with cell growth that was determined visually. The changes in glucose and lactate were more gradual in the HFBR, and it appeared that the HFBR cultures were undergoing a lag period of approximately 4 days (Figure 4.19). Between day 4 and 5, the glucose dropped more dramatically, signaling the start of exponential growth.  20  5 c Q  o " 'co 10 cp -t w c O 05  » g  5 0 0  1  2  3  4  5  Day Figure 4.19: Glucose and Lactate Profiles in Alginate Immobilized HFBR and Slab Cultures Concentration of • = glucose in slab, lactate in HFBR.  A  = lactate in slab, •  = glucose in HFBR,  A =  The pH was initially around 7.6 but decreased to 7.4 within 18 hours as the medium equilibrated with the 10% C 0 in the aeration unit (Figure 4.20). Over the next 2  4 days, the pH gradually fell as a result of the increasing lactate concentration (Figure 4.19). pH was not recorded for the slab cultures. Unfortunately on day 5, the silicon tubing in the aeration box tore, leaking media, and 5 hr elapsed before this was discovered. During that period, air was recirculating  69  through the ICS and the HFBR temperature dropped to ambient temperature. Despite this accident, cells were still recovered and found to be 77% viable when counted. The HFBR cell growth was lower than in slabs (Figure 4.21), but this was expected due to the apparent lag phase. Despite the early ending, there was a 2-fold recovery of cells from the HFBR, suggesting that cells could be alginate immobilized and grown in the HFBR. These results should be confirmed in another run with an extended culture time.  7.8 7.6  Q  O-  -o-  -O  7.2 7.0  —i 0  1  1  1  1  2  1 Day  1  3  1  1  4  •  1  5  Figure 4.20: pH Profile in Alginate Immobilized HFBR  Culture  2.5x10° E j 2.0x10 3  6  CO  §  1.5x10  | o  1.0x10  S  5.0x10  6  6  5  "55  o  o.o  Day 0  Day 5 Slab  Day 4.5 HFBR  Figure 4.21: Cell Concentration in HFBR Cell Growth Experiment  White = HFBR. SEM, n=2.  Grey = alginate slab cultures run in parallel with HFBR. Error bar  70  5  CONCLUSIONS AND RECOMMENDATIONS  In this work, a two-step approach was taken to generating islet tissue. The first step involved examining ways in which potential pancreatic progenitors in NPCC cultures could be expanded. The second step entailed the maturation of NPCC cells to a P-cell like phenotype. Statistical design of experiments was used as a tool to elucidate growth factor effects, and a HFBR was adapted to culture alginate immobilized cells. In the first part, the survival of NPCC suspension cultures was assessed. 80% losses were typical, and a steep decrease in the first day after arrival was attributed to a shipment effect. The addition of EGF, VEGF, KGF and HGF was tested in a fractionalfactorial experiment, but none of the factors yielded a significant improvement. A dose response experiment with EGF and HGF showed that the concentration of 50 ng/mL was higher than necessary. FBS was compared to Ham's F10 alone and Ham's F10 supplemented with bovine serum albumin, insulin and transferrin (BIT). There was no difference in cell survival between cultures with Ham's F10 alone and cultures with 5% FBS. However, cultures with BIT showed a slight improvement. The substitution of insulin with IGF-1 in the BIT mixture resulted in no significant change in this response. The heterogeneity of the cultures and the results from the cell survival analysis in suspension cultures suggested that specific populations, rather than the total population, in pancreatic cell cultures should be assessed. The next set of experiments focused on the expansion of the CK7+ ductal population of NPCCs grown in monolayers. By  dissociating cells and preincubating the suspension to remove fibroblasts  (Tatarkiewicz et al., 2003), a monolayer culture was obtained wherein approximately  71  95% of cells were CK7+ 9 days after isolation. However, these percentages fell to 39% by day 12. To determine if the addition of reported ductal cell mitogens could sustain the CK7+ population, another factorial experiment was performed to assess the main effects and 2-factor interactions of EGF, VEGF, KGF, HGF and bFGF on monolayer cultures. Based upon the previous dose response experiment results, the factor concentrations were reduced. EGF promoted cell attachment and stimulated proliferation, but this was not specifically limited to the CK7+ population. VEGF increased the proportion of CK7+ cells in the monolayers by reducing the attachment of fibroblasts, but it did not stimulate proliferation in CK7+ cells. KGF increased CK7+ cell proliferation but not the number of CK7+ cells in the monolayers because it also decreased cell attachment. HGF was the most promising growth factor in that it promoted cell attachment and proliferation of the CK7+ population, resulting in an overall increase in CK7+ cell numbers. bFGF increased cell attachment but did not stimulate growth in CK7+ cells. The factorial design also revealed many interactions between the growth factors, especially an antagonistic relation between VEGF and HGF. These interactions could not have been elucidated in a one-factor-at-a-time experimental approach without significantly increasing the number of runs. However, a comparison between the two factorial experiments demonstrates that the detection of effects requires the proper selection of factor levels and the identification of potential confounding factors. Curvature was found to be significant in the factorial experiment, indicating that the effects are non-linear. A follow-up experiment should be performed to examine the effects of KGF and HGF in detail because they stimulated CK7+ cell proliferation.  72  However, growth factors should be added after a monolayer culture is established because KGF reduces cell attachment. A two-step approach could also be taken whereby factors that promoted adhesion be added first, followed by the factors that stimulated CK7+ proliferation. Additional time points could also be added to extend the duration of the cultures to determine if the high proportions of CK7+ cells is sustained. In the second section of the work, the maturation of NPCCs was examined. In serum-free suspension cultures, insulin/DNA of NPCC cultures was increased by approximately 2-fold between days 5 and 13 after isolation. EGF was found to reduce insulin/DNA. A 3-fold increase in comparison to controls was observed in insulin/DNA from day 8 to 18 when a previously reported method of culturing cells in 0.75% MVG alginate and 5% NPS was tested (Korbutt et al., 1997). It was not determined whether the insulin/DNA actually corresponded to an increase of P-cell numbers, increased insulin production by existing P-cells, or both. It is suggested that this experiment be repeated with quantification of P-cells using immunocytochemistry. Also, quantitative RT-PCR and glucose challenge tests (Kaczorowski et al., 2002) could be used to confirm that insulin/DNA increases were due to insulin synthesis and not insulin uptake from the medium. Based upon these results, a HFBR was adapted to culture alginate immobilized cells at a clinical scale (-100 mL). The proof-of-concept involved assessment of alginate plugging in the fibres, the degree of gelling, cell recoveries, and cell growth. CHO cells were used to carry out these experiments. Residence time distribution analysis was used to determine if alginate plugged the ICS fibres. It was found that using a high ICS flow rate kept the plugging to a minimum. The degree of alginate gelling was determined by  73  sawing open and examining the HFBR cartridge after gelling. Regardless of alginate concentration (1 or 3%) and gelling time (0.5 and 1 hr), gelling appeared to be complete. Cell recoveries were determined by filling the HFBR with cells, gelling and degelling the alginate, and counting the cells that could be retrieved from the ECS. By passing a degelling solution through the ECS in addition to the ICS, average recoveries greater than >90% could be achieved. The recovery was independent of the alginate concentration. Once a working protocol was established for loading, gelling and degelling the alginate, cell growth was tested in the HFBR. Glucose and lactate were used to indirectly measure cell growth, and it was observed that cells underwent a lag phase in the HFBR, which was absent in slab cultures. Although the run was ended after 5 days because of a burst tube, final cell concentrations indicated that the cell concentration doubled in the HFBR, and likely had just entered the exponential growth phase.  Another growth  experiment in the HFBR with CHO cells should be performed to confirm these results. Once that is successful, the maturation of NPCCs should be tested in the HFBR. Overall however, these results showed that the HFBR is a promising system for culturing cells in alginate.  74  6  REFERENCES  Abraham EJ, Leech CA, Lin JC, Zulewski H, Habener JF. 2002. Insulinotropic hormone glucagon-like peptide-1 differentiation of human pancreatic islet-derived progenitor cells into insulin-producing cells. Endocrinology 143(8):3152-61. Audet J, Miller CL, Eaves CJ, Piret JM. 2002. Common and Distinct Features of Cytokine Effects on Hematopoietic Stem and Progenitor Cells Revealed by DoseResponse Surface Analysis. Biotechnol Bioeng 80(4):393-404. Beattie GM, Rubin JS, Mally MI, Otonkoski T, Hayek A. 1996. Regulation of proliferation and differentiation of human fetal pancreatic islet cells by extracellular matrix, hepatocyte growth factor, and cell-cell contact. Diabetes 45(9): 1223-8. Beattie GM, Cirulli V, Lopez AD, Hayek A. 1997. Ex vivo expansion of human pancreatic endocrine cells. J Clin Endocrinol Metab 82(6): 1852-6. Beattie GM, Itkin-Ansari P, Cirulli V, Leibowitz G, Lopez AD, Bossie S, Mally MI, Levine F, Hayek A. 1999. Sustained proliferation of PDX-1+ cells derived from human islets. Diabetes. 48(5): 1013-9. Bernard C, Berthault MF, Saulnier C, Ktorza A. 1999. Neogenesis vs. apoptosis as main components of pancreatic beta cell mass changes in glucose-infused normal and mildly diabetic adult rats. FASEB J 13(10): 1195-1205. Bhattacharyya E, Panchal A, Wilkins TJ, de Ondarza J, Hootman SR. 1995. Insulin, transforming growth factors, and substrates modulate growth of guinea pig pancreatic duct cells in vitro. Gastroenterology 109(3):944-52. Bonner-Weir S. 2000. Perspective: Postnatal pancreatic beta cell growth. Endocrinology 141(6): 1926-9. Bonner-Weir S, Baxter LA, Schuppin GT, Smith FE. 1993. A second pathway for regeneration of adult exocrine and endocrine pancreas. A possible recapitulation of embryonic development. Diabetes 42(12): 1715-1720. Bonner-Weir S, Taneja M, Weir GC, Tatarkiewicz K, Song KH, Sharma A, O'Neil JJ. 2000. In vitro cultivation of human islets from expanded ductal tissue. Proc Natl Acad Sci USA 97(14):7999-8004. Bouwens L. 1998a. Cytokeratins and cell differentiation in the pancreas. J Pathol. 184(3):234-9. Bouwens L. 1998b. Transdifferentiation Versus Stem Cell Hypothesis for the Regeneration of Islet-Beta-Cell in the Pancreas. Micro Res Tech 43(4):332-336. Bouwens L, Pipeleers DG. 1998. Extra-insular beta cells associated with ductules are frequent in adult human pancreas. Diabetologia 41(6):629-33. Burridge PW, Shapiro JM, Ryan EA, Lakey JRT. 2002. Future Trends in Clinical Islet Transplantation. Transplant Proc 34(8):3347-3348. Castro PM, Hayter PM, Ison AP and Bull AT. 1992. Application of a statistical design to the optimization of culture medium for recombinant interferon-gamma production by Chinese hamster ovary cells. Appl Microbiol Biotechnol 38(1):8490. Castro PM, Hayter PM, Ison AP and Bull AT. 1996. CHO cell growth and recombinant interferon-gamma production: effects of BSA, Pluronic and lipids. Cytotechnology. 19(l):27-36.  75  Cheung AT, Dayanandan B, Lewis JT, Korbutt GS, Rajotte RV, Bryer-Ash M, Boylan MO, Wolfe MM, Kieffer TJ. Glucose-dependent insulin release from genetically engineered K cells. Science. 290(5498): 1959-62. Chun C, Heineken K, Szeto D, Ryll T, Chamow S, Chung JD. 2003. Application of factorial design to accelerate identification of CHO growth factor requirements. Biotechnol Prog 19(l):52-7. Cras-Meneur C, Elghazi L, Czernichow P, Scharfmann R. 2001. Epidermal growth factor increases undifferentiated pancreatic embryonic cells in vitro: a balance between proliferation and differentiation. Diabetes 50(7): 1571-9. Demeterco C, Beattie GM, Dib SA, Lopez AD, Hayek A. 2000. A role for activin A and betacellulin in human fetal pancreatic cell differentiation and growth. J Clin Endocrinol Metab 85(10):3892-7. Edlund, H. 2002. Pancreatic Organogenesis - Developmental Mechanisms and Implications for Therapy. Nat Rev Gen 3(7):524-532. Elghazi L, Cras-Meneur C, Czernichow P, Scharfmann R. 2002. Role for FGFR2IIIbmediated signals in controlling pancreatic endocrine progenitor cell proliferation. Proc Natl Acad Sci USA 99(6):3884-9. Fernandes A, King LC, Guz Y, Stein, R, Wright CVE, Teitelmand G. 1997. Differentiation of New Insulin-Producing Cells Is Induced by Injury in Adult Pancreatic Islets. Endocrinology 138(4): 1750-1762. Finegood DT, Scaglia L, Bonner-Weir S. 1995. Dynamics of beta-cell mass in the growing rat pancreas. Estimation with a simple mathematical model. Diabetes 44(3):249-56. Freshney, IR. 2000. Chapter 20: Quantitation in Culture of Animal Cells: A Manual of Basic Technique (4th Edition), pp. 309-328. Wiley-Liss, Inc. Gao R, Ustinov J, Pulkkinen MA, Lundin K, Korsgren O, Otonkoski T. 2003. Characterization of endocrine progenitor cells and critical factors for their differentiation in human adult pancreatic cell culture. Diabetes 52(8):2007-15. Gu G, Brown JR, Melton DA. 2003. Direct lineage tracing reveals the ontogeny of pancreatic cell fates during mouse embryogenesis. Mech Dev 120(l):35-43. Health Canada Website. Accessed November 12, 2003. "Facts and Figures". Revised January 1, 2003. http://www.hc-sc.gc.ca/pphb-dgspsp/ccdpc-cpcmc/diabetesdiabete/english/facts/index.html Hulinsky I, Cooney S, Harrington J, Silink M. 1995. In vitro growth of neonatal rat islet cells is stimulated by adhesion to matrix. Horm Metab Res 27(5):209-15. Hui H, Wright C, Perfetti R. 2001. Glucagon-like peptide 1 induces differentiation of islet duodenal homeobox-1-positive pancreatic ductal cells into insulin-secreting cells. Diabetes 50(4):785-96. Jonsson J, Carlsson L, Edlund T, Edlund H. 1994. Insulin-promoter-factor 1 is required for pancreas development in mice. Nature 371(6498):606-9. Kaczorowski DJ, Patterson ES, Jastromb WE, Shamblott MJ. 2002. Glucose-responsive insulin-producing cells from stem cells. Diabetes Metab Res Rev 18(6):442-50. Kavalkovich KW, Boynton RE, Murphy JM, Barry F. 2002. Chondrogenic differentiation of human mesenchymal stem cells within an alginate layer culture system. In Vitro Cell Dev Biol Anim 38(8):457-66.  76  Koch S, Schwinger C, Kressler J, Heinzen Ch, Rainov NG. 2003. Alginate encapsulation of genetically engineered mammalian cells: comparison of production devices, methods and microcapsule characteristics. J Microencapsul 20(3):303-16. Korbutt GS, Ao Z, Rayat GR, Flashner M, Rajotte RV. 1997. Effect of microencapsulation and autologous serum on the growth and maturation of neonatal porcine islets in vitro. Acta Diabetol 34:136. Korbutt GS, Elliott JF, Ao Z, Smith DK, Warnock GL, Rajotte RV. 1996. Large scale isolation, growth, and function of porcine neonatal islet cells. J Clin Invest 97(9):2119-2129. Koska J, Bowen BD, Piret JM. 1997. Protein transport in packed-bed ultrafiltration hollow-fibre bioreactors. Chem Eng Sci 52(14):2251-2263. Lechner A, Habener JF. 2003. Stem/progenitor cells derived from adult tissues: potential for the treatment of diabetes mellitus. Am J Physiol Endocrinol Metab 284(2):E259-66. Lefebvre VH, Otonkoski T, Ustinov J, Huotari MA, Pipeleers DG, Bouwens L. 1998. Culture of adult human islet preparations with hepatocyte growth factor and 804G matrix is mitogenic for duct cells but not for beta-cells. Diabetes 47(1): 134-7. Levenberg S, Huang NF, Lavik E, Rogers AB, Itskovitz-Eldor J, Langer R. 2003. Differentiation of human embryonic stem cells on three-dimensional polymer scaffolds. Proc Natl Acad Sci U S A 100(22): 12741-6. Liu C, Chu I, Hwang S. 2001. Factorial designs combined with the steepest ascent method to optimize serum-free media for CHO cells. Enzyme Microb Technol 28(4-5): 314-321. Lopez-Avalos MD, Tatarkiewicz K, Sharma A, Bonner-Weir S, Weir GC. 2001. Enhanced maturation of porcine neonatal pancreatic cell clusters with growth factors fails to improve transplantation. Transplantation 71 (8): 1154-1162. Lucas-Clerc C, Massart C, Campion JP, Launois B, Nicol M. 1993. Long-term culture of human pancreatic islets in an extracellular matrix: morphological and metabolic effects. Mol Cell Endocrinol 94(l):9-20. Lumelsky N, Blondel O, Laeng P, Velasco I, Ravin R, McKay R. 2001. Differentiation of embryonic stem cells to insulin-secreting structures similar to pancreatic islets. Science 292(5520): 1389-94. Magyar JP, Nemir M , Ehler E, Suter N, Perriard JC, Eppenberger HM. 2001. Mass production of embryoid bodies in microbeads. Ann N Y Acad Sci 944:135-43. Mngomezulu WT, Kramer B. 2000. Beneficial effect of nicotinamide on the proportion of insulin cells in developing chick pancreas. Dev Growth Differ 42(2): 187-93. Montesano R, Mouron P, Amherdt M, Orci L. 1983. Collagen matrix promotes reorganization of pancreatic endocrine cell monolayers into islet-like organoids. J Cell Biol 97(3):935-9. Montgomery, DC. 1997. Chapter 15: Response Surface Methods and Other Approaches to Process Optimization in Design and Analysis of Experiments (4 Edition), pp. 575-642. John Wiley & Sons, Inc. Mosahebi A, Wiberg M, Terenghi G. 2003. Addition offibronectinto alginate matrix improves peripheral nerve regeneration in tissue-engineered conduits. Tissue Eng 9(2):209-18. th  77  Movassat J, Beattie GM, Lopez AD, Hayek A. 2002. Exendin 4 up-regulates expression of PDX 1 and hastens differentiation and maturation of human fetal pancreatic cells. J Clin Endocrinol Metab 87(10):4775-81. Oberg C, Waltenberger J, Claesson-Welsh L, Welsh M. 1994. Expression of protein tyrosine kinases in islet cells: possible role of the Flk-1 receptor for beta-cell maturation from duct cells. Growth Factors 10(2): 115-26. Piret, JM and Cooney, CL. 1990. Mammalian Cell and Protein Distributions in Ultrafiltration Hollow Fiber Bioreactors. Biotechnol Bioeng 36(9):902-910. Pulkkinen MA, Spencer-Dene B, Dickson C, Otonkoski T. 2003. The Illb isoform of fibroblast growth factor receptor 2 is required for proper growth and branching of pancreatic ductal epithelium but not for differentiation of exocrine or endocrine cells. Mech Dev 120(2): 167-75. Rajagopal J, Anderson WJ, Kume S, Martinez Ol, Melton DA. 2003. Insulin Staining of ES Cell Progeny from Insulin Uptake. Science 299(5605):363. Rayat GR, Rajotte RV, Korbutt GS. 1999. Potential application of neonatal porcine islets as treatment for type 1 diabetes: a review. Ann NY Acad Sci 18(875): 175188. Rosenberg L, Brown RA, Duguid WP. 1983. A new model for the development of duct hyperplasia and the initiation of of nesidioblastosis. J Surg Res 35(l):63-72. Rooman I, Schuit F, Bouwens L. 1997. Effect of vascular endothelial growth factor on growth and differentiation of pancreatic ductal epithelium. Lab Invest 76(2):22532. Rooman I, Heremans Y, Heimberg H, Bouwens L. 2000. Modulation of rat pancreatic acinoductal transdifferentiation and expression of PDX-1 in vitro. Diabetologia 43(7):907-14. Sanford KK, Earle WR, Evans VJ, Waltz HK, Shannon JE. 1951. The measurement of proliferation in tissue cultures by enumeration of cell nuclei. J Natl Cane Inst 11:773-795. Shapiro AM, Lakey JR, Ryan EA, Korbutt GS, Toth E, Warnock GL, Kneteman NM, Rajotte RV. 2000. Islet transplantation in seven patients with type 1 diabetes mellitus using a glucocorticoid-free immunosuppressive regimen. N Engl J Med 343(4):230-8. SoriaB. 2001. In-vitro differentiation of pancreatic /3-cells. Differentiation 68(4-5):20519. Steinert A, Weber M, Dimmler A, Julius C, Schutze N, Noth U, Cramer H, Eulert J, Zimmermann U, Hendrich C. 2003. Chondrogenic differentiation of mesenchymal progenitor cells encapsulated in ultrahigh-viscosity alginate. J Orthop Res 21(6): 1090-7. Tatarkiewicz K, Garcia M, Lopez-Avalos M, Bonner-Weir S, Weir GC. 2001. Porcine neonatal pancreatic cell clusters in tissue culture: benefits of serum and immobilization in alginate hydrogel. Transplantation 71(11): 1518-26. Tatarkiewicz K, Lopez-Avalos MD, Yoon KH, Trivedi N, Quickel RR, Bonner-Weir S, Weir GC. 2003. Development and retroviral transduction of porcine neonatal pancreatic islet cells in monolayer culture. Dev Growth Differ 45(l):39-50.  78  Thivolet CH, Chatelain P, Nicoloso H, Durand A, Bertrand J. 1985. Morphological and functional effects of extracellular matrix on pancreatic islet cell cultures. Exp Cell Res 159(2):313-22. Thomas JN, Fung V. 1993. Comparison of long R3 IGF-1 with insulin in the support of cell growth and recombinant protein expression in CHO cells in Animal Cell Technology: Products of Today, Prospects for Tomorrow, pp. 91-95. Butterworth-Heinemann Ltd. Trivedi N, Hollister-Lock J, Lopez-Avalos MD, O'Neil JJ, Keegan M, Bonner-Weir S, Weir GC. 2001. Increase in beta-cell mass in transplanted porcine neonatal pancreatic cell clusters is due to proliferation of beta-cells and differentiation of duct cells. Endocrinology 142(5):2115-22. Verme TB, Hootman SR. 1990. Regulation of pancreatic duct epithelial growth in vitro. Am J Physiol. 258(6 Pt l):G833-40. Vila MR, Nakamura T, Real FX. 1995. Hepatocyte growth factor is a potent mitogen for normal human pancreas cells in vitro. Lab Invest 73(3):409-18. Wang RN, Kloppel G, Bouwens L. 1995. Duct to islet cell differentiation and islet growth in the pancreas of duct ligated adult rats. Diabetologia 38(12): 1405-1411. Wang L, Shelton RM, Cooper PR, Lawson M, Triffitt JT, Barralet JE. 2003. Evaluation of sodium alginate for bone marrow cell tissue engineering. Biomaterials 24(20):3475-81. World Health Organization Website. Accessed November 12, 2003. "Fact Sheet No. 138." Updated April 2002. http://www.who.int/inf-fs/en/fact 138.html Wei J, Russ MB. 1977. Convection and diffusion in tissues and tissue cultures. J Theor Biol 66:775-787 Wells JM. 2003. Genes expressed in the developing endocrine pancreas and their importance for stem cell and diabetes research. Diabetes Metab Res Rev 19(3):191-201. Yi ES, Yin S, Harclerode DL, Bedoya A, Bikhazi NB, Housley RM, Aukerman SL, Morris CF, Pierce GF, Ulich TR. 1994. Keratinocyte growth factor induces pancreatic ductal epithelial proliferation. Am J Pathol. 145(l):80-5. Yoon KH, Quickel RR, Tatarkiewicz K, Ulrich TR, Hollister-Lock J, Trivedi N, BonnerWeir S, Weir GC. 1999. Differentiation and expansion of beta cell mass in porcine neonatal pancreatic cell clusters transplanted into nude mice. Cell Transplant 8(6):673-89. Zandstra PW, Conneally E, Piret JM, Eaves CJ. 1998. Ontogeny-associated changes in the cytokine responses of primitive human haemopoietic cells. Br J Haematol 101(4):770-8. Zulewski H, Abraham EJ, Gerlach MJ, Daniel PB, Moritz W, Muller B, Vallejo M , Thomas MK, Habener JF. 2001. Multipotential nestin-positive stem cells isolated from adult pancreatic islets differentiate ex vivo into pancreatic endocrine, exocrine, and hepatic phenotypes. Diabetes 50(3):521-33.  79  7  APPENDIX  Table 7.1: Error in Seeding Technique NPCCs were seeded into 4 tubes and 4 counts were performed for each tube. Note that the standard deviation in counting is higher than in seeding. Seeding Sample Count  1  2  4  3  1  7.00x10  5  7.80x10  s  8.30x10  s  '6.30x10  s  2  5.40x10  5  8.40x10  s  6.40x10  s  6.40x10  s  3  7.98x10  s  7.96x10  s  8.77x10  s  6.00x10  5  4  7.26x10  s  7.54x10  5  6.43x10  s  8.27x10  s  Counting Average  6.91x10  s  7.93x10  s  7.48x10  s  6.74x10  s  Counting Standard Deviation  1.09x10  s  3.61 x 1 0  4  1.24x10  s  1.03x10  s  Seeding Sample Average  7.26x10  s  Seeding Standard Deviation  5.41 x 1 0  4  Table 7.2: Comparison of Nuclei Counting to Trypan Blue Method CHO cells were counted using both methods. No statistically significant difference was observed between the two methods. Trypan Blue (viable cell/mL)  Nuclei Count (cell/mL)  1  4.46x10  s  3.88x10  2  4.83x10  s  3.16x10  s  Sample  s  3  3.10x10  s  2.79x10  s  4  2.73x10  5  2.86x10  s  5  3.60x10  s  3.75x10  s  6  4.15x10  s  3.38x10  s  Average  3.81x10  5  3.30x10  s  Standard Deviation  8.12x10  4  4.51 x 1 0  4  4  4.74x10  Degrees of Freedom  5  5 8.52x10  9 5 % Confidence Interval ( ± ) of Mean  4  Table 7.3: Design and Data for Growth Factor Experiment in Suspension Culture EGF Run (ng/mL) 1 0 3 0 7 0 8 0 2 25 25 6 10 25 12 25 4 50 5 50 9 50 11 50  Day 4 (ng/mL) (ng/mL) (ng/mL) Expansion 0 0.65 0 0 0 0.7 50 50 50 0.63 0 50 0.71 50 0 50 25 0.77 25 25 25 0.6 25 25 25 25 0.63 25 25 0.74 25 25 50 0.75 50 50 1.01 50 0 0 0.73 0 50 0 0 50 0.75 0  VEGF  KGF  HGF  80  Day 8 Expansion 0.06 0.13 0.1 0.12 0.07 0.07 0.23 0.25 0.24 0.06 0.16 0.18  Day 4 Day 8 Insulin/DNA Insulin/DNA (ug/ug) 0.077 0.066 0.078 0.053 0.054 0.072 0.075 0.073 0.092 0.071 0.059 0.052  (ug/ug) 0.101 0.072 0.049 0.071 0.054 0.036 0.049 0.04 0.045 0.035 0.029 0.044  Table 7.4: ANOVA for Model in Table 4.1 Sum of  Degrees  Mean  F  Source  Squares  Of Freedom  Square  Value  Model  0.095888  7  0.013698  2.00  EGF  0.037813  1  0.037813  5.53  VEGF  0.021013  1  0.021013  3.07  KGF  0.012013  1  0.012013  1.75  HGF  0.007813  1  0.007813  1.14  E G F * V E G F (aliased with K G F * H G F )  0.002813  1  0.002813  0.41  E G F * K G F (aliased with V E G F * H G F )  0.007813  1  0.007813  1.14  E G F * H G F (aliased with V E G F * K G F )  0.006613  1  0.006613  0.96  Curvature  0.008438  1  0.008438  1.23  Pure Error  0.0205  3  0.006833  0.124825  11  C o r Total  Table 7.5: ANOVA for Model in Table 4.2  Source Model  Sum of  Degrees  Mean  F  Squares  Of Freedom  Square  Value  0.026  7  3.75x10"  J  0.39  1  6.61x10"  3  0.68  1  3.12x10"*  0.03  1  5.51 x10"  3  0.57  1  6.61 x 1 0  -3  0.68  3  EGF  6.61x10"  VEGF  3.12X10"  4  KGF  5.51 x10"  3  HGF  6.61 x10'  3  E G F * V E G F (aliased with K G F * H G F )  2.11x10"  1  2.11x10"  E G F * K G F (aliased with V E G F * H G F )  1.51x10"  3  1  1.51x10"  E G F * H G F (aliased with V E G F * K G F )  3.61 x10"  Curvature  1.50x10"  Pure Error C o r Total  3  3  0.22  3  0.16  1  3.61 x10"  0.37  1  1.50x10'  0.029  3  9.70x10"  0.057  11  3  3  3  0.16  3  3  Table 7.6: Immunocytochemistry Images Counts - Total Cell Number  Day 1 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7  Run NA 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20  Slide 1 Image 1 83 54 88 123 91 24 167 80 36 41 64 147 82 33 69 37 35 70 57 48 41  2 134 45 399 311 36 20 195 85 25 47 94 120 53 20 51 36 33 80 61 88 43  3 65 50 174 137 49 18 162 35 40 28 160 101 43 35 50 24 68 66 144 101 70  4 76 62 . 56 74 96 51 354 31 34 43 190 193 234 49 40 19 74 25 60 212 212  Slide 2 Image 1 61 110 95 74 92 160 137 57 23 46 48 160 89 66 68 209 54 50 61 128 115  2 36 . 245 102 NA 49 115 158 388 11 80 46 52 94 48 184 101 59 43 61 270 53  3 33 231 87 126 51 113 159 8 22 58 66 82 144 58 122 56 61 49 156 177 66  4 46 134 284 58 76 97 89 24 23 51 111 122 292 118 146 207 50 72 216 280 59  81  Slide 3 Image 1 64 186 146 132 89 58 161 100 178 43 70 85 46 392 NA 55 111 60 175 129 33  2 48 103 268 98 99 40 89 39 126 30 71 86 74 108 NA 111 63 58 80 114 59  3 49 74 103 125 104 57 53 25 86 12 59 146 77 41 NA 92 104 68 53 72 57  4 75 62 89 225 218 57 112 139 85 31 55 155 131 284 NA 198 140 64 63 88 44  Slide 4 Image 1 146 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA  2 165 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA  3 148 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA  4 74 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA  Table 7.7: Immunocytochemistry Images Counts - CK7+ Cells  Day 1 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7  Run NA 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20  Slide 1 Image 1 9 49 1 0 71 22 0 2 25 33 49 2 71 27 70 22 34 63 46 41 41  2 8 38 2 4 32 14 0 0 19 37 73 3 55 18 45 22 24 60 50 62 35  3 0 38 2 0 42 18 0 0 34 12 100 0 40 28 50 24 58 54 160 83 54  4 0 34 0 0 72 40 0 0 28 26 121 2 197 48 40 14 58 51 151 175 177  Slide 2 Image 1 0 0 0 63 81 141 0 0 0 32 42 3 75 53 64 142 48 62 64 100 114  2 0 4 1 NA 42 105 0 0 0 50 34 0 77 46 48 71 56 37 51 205 38  3 1 0 0 83 45 109 0 0 0 46 51 0 117 62 109 43 48 41 134 141 55  4 1 0 4 54 62 94 0 0 0 39 85 2 213 84 144 131 46 63 174 237 52  Slide 3 Image 1 0 0 2 0 61 52 0 0 62 35 48 0 43 305 NA 49 0 40 145 116 36  2 0 0 2 0 71 44 0 0 111 21 58 2 57 107 NA 97 1 43 78 94 52  3 0 6 0 0 81 55 0 0 56 10 51 0 56 41 NA 80 1 61 47 58 60  4 1 0 0 0 127 51 1 0 74 22 53 5 97 219 NA 139 1 52 64 74 39  Slide 4 Image 1 12 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA. NA NA NA NA NA  2 13 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA  3 1 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA  4 1 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA  Table 7.6V Immunocytochemistry Images Counts - Ki67 + Cells Total number of Ki67+ cell = 899. Total number of images counted = 252. Positives per image = 899/252 = 3.57.  Day 1 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7  Run NA 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20  Slide 1 Image 1 7 3 3 2 2 3 0 0 5 9 4 0 3 1 6 0 7 5 2 6 0  2 9 4 5 1 3 3 0 0 3 6 6 0 8 0 3 0 1 5 7 11 0  3 0 4 4 1 5 7 0 0 3 4 1 1 0 0 6 0 3 2 8 13 0  4 2 10 2 0 10 12 0 0 1 2 9 0 10 2 2 0 4 3 0 27 0  Slide 2 Image 1 0 6 0 1 7 5 0 0 0 7 9 0 0 5 2 3 4 3 1 15 0  2 0 0 0 0 3 5 0 0 0 13 2 0 6 2 1 0 2 1 0 22 0  4 0 2 0 0 4 4 0 0 0 9 12 0 9 4 5 0 2 1 11 21 2  3 0 0 0 7 9 6 0 0 0 10 17 0 0 5 0 0 2 2 5 10 0  82  Slide 3 Image 1 0 0 0 0 9 7 0 0 12 5 3 0 0 4 NA 8 0 9 10 18 6  2 0 0 0 0 3 1 0 0 44 2 2 0 0 3 NA 2 0 10 11 17 7  3 0 1 0 0 0 3 0 0 14 0 2 0 1 1 NA 3 0 8 2 2 4  4 1 0 0 0 1 6 0 0 20 3 6 0 1 13 NA 1 0 3 9 15 1  Slide 4 Image 1 1 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA  2 3 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA  3 4 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA  4 1 NA NA NA NA NA NA NA NA NA NA . NA NA NA NA NA NA NA NA NA NA  Table 7.9: Immunocytochemistry Images Counts - Double Stained Cells Slide 1 Image Day  Run  1 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7  NA 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20  1 0 1 0 1 2 2 0 0 2 4 0 0 1 1 5 0 6 1 1 4 0  Slide 2 Image 2 0 4 0 0 2 2 0 0 0 6 4 0 8 0 3 0 0 0 4 9 0  3 0 3 0 0 5 6 0 0 2 4 1 0 0 0 6 0 2 1 8 7 0  4 0 7 0 0 7 9 0 0 0 1 5 0 4 3 2 0 4 0 1 20 0  1 0 0 0 0 7 4 0 0 0 5 3 0 0 4 2 0 3 1 1 13 0  Slide 3 Image 2 0 0 0 0 2 3 0 0 0 11 1 0 5 1 0 0 1 1 0 18 0  3 0 0 0 3 7 1 0 0 0 8 9 0 0 0 0 0 0 0 3 5 0  4 0 0 0 0 1 1 0 0 0 8 9 0 2 3 2 0 0 0 5 16 1  1 0 0 0 0 4 4 0 0 8 0 3 0 0 1 NA 4 0 6 6 15 4  Slide 4 Image 2 0 0 0 0 2 1 0 0 34 1 0 0 0 3 NA 2 0 6 7 14 7  3 0 0 0 0 0 1 0 0 12 0 1 0 1 1 NA 1 0 5 1 2 4  4 1 0 0 0 1 4 0 0 16 3 1 0 1 6 NA 1 0 1 7 10 1  1 0 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA  2 0 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA  3 0 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA  4 0 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA • NA NA NA NA  Table 7.10: Selection of Intensity Threshold forKi67 Threshold  Negative control images were analysed for the minimum intensity threshold required to count zero cells. The number ofpositives counted per slide for regular images was 3.57 (Table 7.8). Assuming the average number of cells counted per slide are equal for negative controls and regular images, the rate offalse positives is 0.219/(3.57+0.219) = 6%.  Run  Day 1 Day 7 Run 1 Day 7 Run 2 Day 7 Run 3 Day 7 Run 4 Day 7 Run 5 Day 7 Run 6 Day 7 Run 7 Day 7 Run 8 Day 7 Run 9 Day 7 Run 10 Day 7 Run 11 Day 7 Run 12 Day 7 Run 13 Day 7 Run 14 Day 7 Run 15 Day 7 Run 16 Day 7 Run 17 Day 7 Run 18 Day 7 Run 19 Day 7 Run 20  Minimum Threshold for No False Positives  # of False Positives at Intensity Threshold of 68  Slide 1 50 61 113 92 63 50 72 61 50 50 50 256 50 50 50 50 50 50 50 50 50  Slide 1 Slide 2 Slide 3 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 8 0 1 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Total # False Positives False Positives per slide  Slide 2 50 50 50 50 50 50 109 50 50 50 50 100 50 50 50 52 154 50 50 50 50  Slide 3 Slide 4 50 50 50 NA 50 NA 256 NA 50 NA 50 NA 50 NA 50 NA 69 NA 50 NA 50 NA 177 NA 63 NA 50 NA NA NA 154 NA 64 NA 50 NA 50 NA 50 NA 162 NA  83  Slide 0 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA 14 0.219  <<<<<<<<<<<<<<<<<< z z z z z z z z z z z 2 z 2 S z z §  vo  a o  o o o r - T r d p ^ f O T ^ d o T ^  8 8  T - ' d r ^ ' i r ) ° o 6  Z  38888S>5!g!2Sr:g£88!gS  © d ^ d ^ d d d ^ c d d W c M ' d d W d ^ i b d  N d o d ~ d d ( N i d r d n N s d i r i d v ' c d 6  vo  d 6  +  o  a  R  <a  u  +  r  n  d  d  o  n  a  +  + ro  d  «j a) oi  T-  o j cvi d d O  d  CN) CO  m h- K. T - Q © cn Tt ^ co o o T - : d T-^ i o d d  r  '  r  d  o  N  '  j  C  o  N  O  O  C  M  0 - ^ 9 0 0  r  t  N  d  C  N  W  T  -  T  -  O  O CO o  8 8 8 5 8 8 8  OO  d d xj uS ci d d  <<<<<<<<<<<< z z z 2 z 2 z z z 2 z z  8 8 < 8S 8< ° 2  i  § S !N (N 8 8 8 S S 8 8 g f c S 8 S 3 & S  r ^  O O  -R  r  ' - h - ^ o O ' - ^ . c N c o N - - T - o o ( N ^ n c c o  +  ro  ON  3  VO  +  oo  d  z I I zS IS VO  O  d  s  O O  8n  ^  ^ CM O O  z  ?! °  < 2  8  255 ? S  O) CO ^  00  3 "CT  60 ccj  + ro cu  *•>» e P 00 •C5 ccJ  S  s  +  CN U 60 60 c0 <D CO  58  fc|g8SsS88S85°8  8 Si8 ° 8 S 2  • _• in  N CO  0  s °  1 0 T- CD SJ Q C O CO ~ O  0 ^ m  n S j t N o n r i s d d d n ^ d  Q  3 0 d cd! ^ d d  r--  o  OJ  co  Tf  dddc\icddd®odiridd  CM  n0  z  06 ^ oi  . ^ S . - i J S T - q o q ^ ^ S ^ i n u l i i n ' c t t N ^ o i O co CO o o o 0 ( N i n < - o v n f i N O  d^di^'^ddd^rodu^cvddr^'diri^d  -cs cu R  8  CD 60  CN  bo +'  ~t3  0 0 0  £ 6 +  v.  ^£  <L> 00 ea  ca  60 CCJ  S a  R <a cu ^ 5 ca P \  CU ~<3 ca ^3  fI R  ^  ca  -s;  Si  s z z z II  U  ; 8 8 » " S 8 " 8 S ? ^ < S K ? 8 K - d d g $ d d S i e g ' - f : 5 8 ° S S S S z  SS g 8 ° ° ° g fe  T-  CO CM  m _ CSJ CO CN I - CO o o J m H O W CN i n CD •H- O O CM CD CN '  m  s  in co CM  o  CO  CM  O O I - N T - ^ B S  N ^ - C 0 O ( N l O T - T - t - 0 ) 0 ) T - ( D  ? E W _ C M — O -rCM i t CO O N O O N N (N m S N » £ * S I O 11) CN CN T T- O ) O 3 CO O CO CO CN „ ~ ~ C M 2 ° 2 2 ' * S > C N ' ,^ 3_ 00 CM C D C O w r r O N r- m T- ^ -  31 £1  c  — + It N S O r .9 ^  CD TJ- o  CO 05 m TO i -  Jjcoocncno)  •>£ CN 00 CN o o o CD o> 1 - T - 00  5  .  0  f»  in co co co  it: CD CD CN CO  CD  dddr o  U  C O N N I D C O C O M T - m c O ^ CD n O ) 0 ) 0 0 0 * ( O O r O O I 0 4 o o r i r i r i d ^ - d d ( \ i * - N d  j f ^ i n c o c D o o i n c N o o °i " : ^ •* o '-CMTf'd^coeOT-cN h  s ^ o  N  o CN i n o o o  ^] t -  1 0  ^  <£ C§-co co co !?- ,1 5 £ S _ O O O C O T - 0 0  0  1  0  5  CO 00 CO C O C N CN N CO T- # • O O S V C O C N C O - S ^ C C I N O « CO * m N in T- co  a> in T - ^ " m co co to r^- CM 00 c N c o T O c o o i n c o t n r ^ c n c D i n • : co UOOOSCSOOTtSNT-cOCOCOMOCOOICoio +  00  St  =5  53  s  <3 S .&>  3  ra co E 2 2 S S S ° ° ° ° O O Q O O O O O O O O o o g o o o o o o o o S o o o o o o S o >, cQ I o o o o o g o o o o o o o o o o o o o o o raoo a S£K^3;r^<M^OTi-co>a-cncoi'^oiocD'>* Q S CMr-CMT-rtCMT-r-^CM'--r-CMCNCMCNCOCM't-CM  i _ 5 ^ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 o o> in c t - = N S S C \ I C \ I O N M N s S s i n ? . N CM 55 ra§"m«£S°^ ^^^^«^P^coincoco CO O i CM -S p™ W w w r r r r r ( \ | r P J O I r r r r r r r r 1  •* v t  jQ^oininpiniiilBoinininu)5BoSB O  N  ra  o in CM N CO  "r E r- O S ! N  O  O N  O O N O N O  O  O  O N O  O r O  O  «  O O O O O O ° ? ; N N  E D)  5  t ^ N - m i « 1 0i nm s s C „ C O CD CO M CD * 1^ * C< '* C Ccd O C cd O C DC D '' ^:CIDD °? > ~ - O O * - O T - T - O O O O T - T - 0 0 ' - O C 0 C 0 C 0 ' - ? "  u. O  1  T~ T- CO  cogoeomoitcoco cprt^inscftcooco ^ C M T t O C O C N C O C O C N  0 00 CO C N o m e o cnooo O M t o o 6 ^ s o o « ) « i « o 3  2 5  n rn T- c o  O) *m cso o M  "D  'm —  +o  o s t o O ) O ) m iT» C N  r  fl)  I  c o s s CN  c » c \ i O T - s s i n o i n o o p ^ S n i D S i n r T - C N ^ C O ^ C O i n T - ^ C N ^ C N C M ^ C M C M C M ^ C M  •9 = 3  M  f-  o*  S  oo O O in in T < i n cn t CO CO  CM Tt^- co co g e o c o s ( M i n T - o c o O 00  5 -t •"HA  >  c O O O O O O O O O O O O i O O N I N I M C N C M N O O O I N - O N r - J - r - O O E 01 e _  _  _  _  _  o  o  o  o  o  o  o  O O O O  O  J E ~D) C _ P_ O O O O O O O O O O O -=•0 t \ l 0 1 N C M N 0 ( M 0 N ( M 0 r 0 0 r - i - r ( N 0  c  D3i r - C M W M - i n C O S C O O )  O r - c N O ^ l O C O M l O f f l O  10 00  Table 7.13: ANOVA for Model in Table 4.4 Source  Sum of Squares  Model EGF  2.576x10  1U  2.026x10 7.298x10 5.400x10  VEGF KGF HGF bFGF EGF'VEGF EGF*KGF EGF*HGF EGF*bFGF VEGF'KGF VEGF*HGF VEGF*bFGF KGF*HGF KGF*bFGF HGF*bFGF Error Total  9  9  9  Degrees of Freedom  Mean Squares  15 1 1 1  1.718x10  7.535x10  9  1  4.558x10  9  1  2.838x10 6.683x10 2.152x10 2.067x10  1 1 1 1  9  s  s  s  2.838x10 2.506x10 4.683x10 5.333x10 4.076x10 1.488x10 3.417x10 2.580x10  s  s  9  9  9  s  7  10  1 1 1 1 1 1 3 18  2.026x10 7.298x10 5.400x10 7.535x10 4.558x10 2.838x10 6.683x10 2.152x10  F-Value a  150.79  s  177.85 640.67 474.07  s  s  661.50 400.17  s  s  249.19 586.74 188.91 181.50  s  s  s  2.067x10 2.838x10 2.506x10 4.683x10 5.333x10 4.076x10 1.488x10 1.139x10  s  249.19 220.02 411.13 468.17  s  9  9  9  357.80 130.67  9  s  7  Table 7.14: ANOVA for Model in Table 4.5 Source  Sum of Squares  Model EGF  0.18 6.726x10"  VEGF KGF  5.439x10" 0.025 0.015 0.014  3  3.238x10" 0.029  3  HGF bFGF EGF*VEGF EGF*KGF  3  Degrees of Freedom  Mean Squares  F-Value  15 1  0.012 6.726x10"  1 1 1 1  5.439x10" 0.025 0.015 0.014  3  1 1  3.238x10' 0.029  3  3  1  0.018  EGF*bFGF  2.522x10"  3  1  VEGF*KGF VEGF*HGF VEGF*bFGF  4.836x10" 9.969x10-  3  1  4  1  2.522x10" 4.836x10" 9.969x10-  EGF*HGF  KGF*HGF KGF*bFGF HGF*bFGF Error Total  0.018  0.019 0.012 0.027 4.342x10' 0.023 0.20  3  1 1 1 1 3 18  86  1.54 0.87 0.70 3.17 1.95 1.86 0.42 3.72 2.31  3  3  4  0.019 0.012 0.027 4.342x10" 7.744x10'  3  3  0.33 0.62 0.13 2.46 1.58 3.48 0.56  Table 7.15: ANOVA for Model in Table 4.6 Source Model EGF VEGF KGF HGF bFGF EGF*VEGF EGF*KGF EGF*HGF EGF*bFGF VEGF*KGF VEGF*HGF VEGF*bFGF KGF*HGF KGF*bFGF HGF*bFGF Curvature Error Total  Sum of Squares 6.980x10 1.225x10 4.556x10 3.600x10 1.600x10 3.192x10 1.600x10 4.830x10 1.600x10 6.561x10 6.250x10 1.266x10 1.243x10 3.136x10 7.290x10 3.782x10 1.711x10 2.195x10 8.910x10  1U  9  9  7  10  9  7  9  7  9  8  10  10  9  8  9  10  9  10  Degrees of Freedom 15 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 3 19  Mean Squares 4.653x10 1.225x10 4.556x10 3.600x10 1.600x10 3.192x10 1.600x10 4.830x10 1.600x10 6.561x10 6.250x10 1.266x10 1.243x10 3.136x10 7.290x10 3.782x10 1.711x10 7.317x10  F-Value 6.36 1.67 6.23 0.049 21.87 4.36 0.022 6.60 0.022 8.97 0.85 17.30 16.99 4.29 1.00 5.17 23.39  M  9 9  7  10  9  7 9  7  9  8  10  10  9  8  9  10  8  Table 7.16: ANOVA for Model in Table 4.7 Source Model EGF VEGF KGF Residual Lack of Fit Pure Error Corr Total  Sum of Squares 5.02x10 9.76x10 2.41x10 2.01x10 1.30x10 2.37x10 2.16x10 2.19x10  1U  9  10  10  10  10  10  9  Degrees of Freedom 3 1 1 1 1 14 11 3  87  Mean Squares 1.675x10 9.761x10 2.418x10 2.017x10 1.301x10 1.700x10 1.964x10 7.317x10  F-Value  1U  9  10  10  10  9.86 5.74 14.22 11.87 7.65  9  9  8  2.68  Table 7.17: ANOVA for Model in  Table 4.8 Source  Sum of Squares  Degrees of Freedom  Mean Squares  F-Value  Model EGF VEGF KGF  23310.96 41.72 675.52  15 1  1554.06 41.72 675.52  28.90  HGF bFGF EGF*VEGF EGF*KGF EGF*HGF EGF*bFGF VEGF*KGF VEGF*HGF VEGF*bFGF KGF*HGF KGF*bFGF HGF*bFGF Curvature Error Total  1350.69 1187.14  1 1 1 1 1 1 1  1767.66 93.62 4652.75 167.32 1107.12 1582.16 2293.76  1 1 1 1 1 1 1  5534.08 161.30 29006.33  3 19  81.20 414.69 32.06 7863.54  1  81.20 414.69 32.06 7863.54 1350.69 1187.14 1767.66 93.62 4652.75 167.32 1107.12 1582.16 2293.76 5534.08 53.77  0.78 12.56 • 1.51 7.71 0.60 146.26 25.12 22.08 32.88 1.74 86.54 3.11 20.59 29.43 42.66 102.93  Table 7.18: ANOVA for Model in Table 4.9 Source Model EGF VEGF KGF  Sum of Squares  Degrees of Freedom  Mean Squares  321.72  15 1 1 1  21.45 27.09 0.28 6.73  27.09  1.48  1  7.51 43.56  17.96  37.39  1  37.39  15.42  0.29  1  21.81 8.29  1  0.29 21.81  0.12 8.99  8.29 128.60 11.94  3.42 53.02  0.28 6.73 1.48  bFGF  7.51  1 1  EGF'VEGF  43.56  EGF*KGF  VEGF*KGF VEGF*HGF VEGF*bFGF KGF*HGF KGF*bFGF HGF*bFGF Curvature Error Total  8.84 11.17 0.12 2.78 0.61  HGF  EGF*HGF EGF*bFGF  F-Value  128.60 11.94 0.013 26.32 0.42 3.26 7.28 332.25  1 1 1 1 1 1 1 3 19  88  3.10  0.013 26.32  4.92 0.001 10.85  0.42 3.26  0.17 1.34  2.43  Table 7.19: ANOVA for Model in Table 4.10 Sum of  Source  Squares 300.77  Model  Degrees of Freedom  Mean  F-Value  Squares  10 1  30.08  27.09  27.09  7.74  0.28 6.73  1 1  0.28 6.73  EGF*VEGF EGF*KGF . EGF*bFGF VEGF*HGF  1.48 7.51 43.56 37.39 21.81 128.60  1 1 1 1 1 1  1.48 7.51 43.56 37.39 21.81 128.60  0.080 1.92 0.42  KGF*bFGF  26.32  1  26.32  7.52  Residual Lack of Fit Pure Error  31.49 24.21  9 6 3 19  3.50 4.03 2.43  1.66  EGF VEGF KGF HGF bFGF  Corr Total  7.28 332.25  8.60  2.15 12.45 10.69 6.23 36.76  Table 7.20: ANOVA for Model in Table 4.11 Source Model EGF VEGF KGF  Sum of  Degrees of  Mean  Squares  Freedom  Squares  721.26 1.04  15 1 1  48.08 1.04  19.13 25.90  HGF bFGF EGF*VEGF EGF*KGF  16.25 4.36 1.59 1.53  EGF*HGF EGF*bFGF VEGF*KGF  42.57 7.87 54.81  1 1 1 1 1 1 1 1  VEGF*HGF  307.38  1  VEGF*bFGF  134.01  1  KGF*HGF  16.64  KGF*bFGF  1.86 86.34  1 1  HGF*bFGF. Curvature Error Total  ' 12.70 1.57 735.53  1 1 3 19  89  19.13 25.90 16.25 4.36 1.59 1.53 42.57 7.87 54.81 307.38 134.01 16.64 1.86 86.34 12.70 0.52  F-Value 92.01 1.99 36.60 49.57 31.09 8.34 3.03 2.92 81.45 15.06 104.87 588.14 256.41 31.85 3.56 165.20 24.30  Table 7.21: ANOVA for Model in Table 4.12 Source  Sum of Squares  Degrees of Freedom  Mean Squares  Model EGF VEGF KGF HGF  1.497x10" 3.321x10  15 1  9.980x10 3.321x10  2.464x10 7.131x10 1.100x10  1 1 1  2.464x10 7.131x10  bFGF EGF*VEGF EGF*KGF EGF*HGF EGF*bFGF  1.188x10 3.001x10 1.474x10 1.805x10 6.491x10  VEGF*KGF  2.054x10  VEGF*HGF VEGF*bFGF KGF*HGF  3.490x10 2.052x10 4.264x10  KGF*bFGF HGF'bFGF Curvature Error Total  7.454x10 2.116x10 6.416x10 2.141x10 2.160x10  8  5  s  10  s  10  10  9  s  9  10  10  9  9  10  10  9  11  F-Value  M  s  s  s  1 1 1 1 1  1.100x10 1.188x10 3.001x10 1.474x10 1.805x10 6.491x10  1  2.054x10  1 1 1 1 1 1  3.490x10 2.052x10 4.264x10 7.454x10 2.116x10 6.416x10 7.136x10  3 19  10  8  10  10  9  s  13.99 0.47 3.453x101.00 15.41 0.17 42.05 20.66 2.53 0.91  9  2.88  10  48.91 28.75 5.98 10.45 29.65 89.91  10  9  9  10  10  8  Table 7.22: ANOVA for Model in Table 4.13 Source  Sum of Squares  Degrees of Freedom  Model  1.33x10"  EGF VEGF KGF HGF bFGF EGF*VEGF EGF*KGF  3.32x10 2.46x10 7.13x10 1.10x10 1.19x10 3.00x10 1.47x10  VEGF*HGF  3.49x10  10  VEGF*bFGF  2.05x10  10  HGF*bFGF  2.12x10  10  Curvature  6.42x10  10  Residual Lack of Fit Pure Error Corr Total  1.84x10 1.62x10 2,14x10 2.16x10"  s  s  8  10  10 1 1 1 1 1  Mean Squares  F-Value  1.33x10  lu  3.32x10 2.46x10 7.13x10 1.10x10  s  s  s  10  5.81 0.14 0.00 0.31 4.79 0.05 13.07 6.42  1 1  1.19x10 3.00x10 1.47x10  1  3.49x10  10  15.20  1 1  2.05x10  10  8.94  2.12x10  10  9.22  1  6.42x10  10  27.94  10  8  10  5 3 19  2.30x10 3.25x10 7.14x10  8  10  10  9  90  8  10  10  9  9  s  4.55  4  Table 7.23: Prediction for CK7+ Cell Concentration and % Proliferating CK7+ Cells  Predictions are based upon the models given in Table 4.11 (excluding insignificant interactions) and Table 4.13. Growth Factor Concentration (ng/mL) EGF  VEGF  KGF  HGF  bFGF  0  0 0  0  0  0  0  0  20 20 0 0  0 0 20 20 20  0 0 0 0 0  0 0  0 0 20 20 0 0 20 20  20 0 0 0 0 20 20 20 20  0 16.7 16.7 16.7  0 0 0 0  16.7 16.7 16.7 16.7  0 0 0 0  0 0 20  0 0 0  16.7 0  0 20 20  20 0 0 20 20 0 0 20 20  0 20  20 0 20 0 20 0 20 0 20 0 20 0 20 0 20 0 20 0 20 0  20 20  0 0  0 0 16.7 16.7 16.7 16.7  % CK7+ Cells that are Proliferating  3.0 -5.6  -4.3 -5.3  3.6 -12.1 -5.3 8.5 -13.2  6.1 10.7 -4.4 0.2 4.2  -16.7  8.9 -0.6 -2.4 -2.0  4  0 0 0 0  20 20 20 20 20 20 20 20 20 20  0 0 0  Concentration of CK7+ (10 cells/mL)  -1.9 -0.2 23.3 7.6 0.1 13.9 6.5  -3.9 5.0 3.1 -3.9 -5.7  3.0 -0.5 1.2 10.3  -8.1 -6.3 -3.6  -8.0 15.4 -0.2  -1.8 -1.2 0.7 -4.1 -2.2 10.6 5.9 -2.5 -7.1  -5.3 1.4 15.3 -6.4 -9.9 -9.7  0  20 20 20 0 0 0 0 20  16.7  20  -7.8  ir7t5  20  0  20  16.7  20  6.1  12.8  0  20  20  16.7  20  -1.3  -2.9  20  20  20  16.7  20  -4.8  -7.6  20 0 20 0 20 0 20 0  Table 7.24: ANOVA for Model in Table 4.14 Sum of  Degrees  Mean  F  Source  Squares  Of Freedom  Square  Value  Model  1.33X10"  7  1.900X10"*  2.00  J  0  1  0  VEGF  3.20X10"  5  1  3.200X10"  5  0.34  KGF  2.20X10  A  1  2.205X10"  4  2.32  HGF  5.00X10"  7  1  5.000X10'  7  0  E G F * V E G F (aliased with K G F * H G F )  9.68X10"  4  1  9.680X10"  4  10.19  E G F * K G F (aliased with V E G F * H G F )  2.45X10"  5  1  2.450X10"  5  0.26  E G F * H G F (aliased with V E G F * K G F )  8.45X10"  5  1  8.450X10"  5  0.89  1  0 9.500X10"  EGF  Curvature  0  0  Pure Error  2.85X10"*  3  Cor Total  1.61X10"  11  3  91  0 5  Table 7.25: ANOVA for Model in Table 4.15  Source  Sum of  Degrees  Mean  F  Squares  Of Freedom  Square  Value  Model  3.990x10"  EGF  2.450x10"  7  5.699x1  1  2.450x10"  VEGF  0.000  1  0.000  KGF  3.920x10"  4  HGF  9.800x10"  5  1  3.920x10"  4  1  9.800x10'  5  E G F * V E G F (aliased with K G F * H G F )  2.450x10"  5  1.45  1  2.450x10  -5  0.36  E G F * K G F (aliased with V E G F ' H G F )  2.645x10"  4  E G F * H G F (aliased with V E G F * K G F )  7.605x10"  4  1  2.645x10  -4  3.91  1  7.605x10"  4  Curvature  3.227x10"  4  11.25  1  3.227x10"  4  4.77  Pure Error  2.027x10^  3  6.758x10"  5  Cor Total  4.515x10"  11  J  3  3  92  8.43 3  36.25 0.000 5.80  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0058660/manifest

Comment

Related Items