Particle Motion Analysis in Pancreatic Islets Isolation using a  
Quadrupole Magnetic Flow Sorter  
 
by  
 
Venkata Sunil Kumar Sajja  
 
 
 
 
A dissertation submitted to the Graduate Faculty of  
Auburn University  
in partial fulfillment of the  
requirements for the Degree of  
Doctor of Philosophy  
 
Auburn, Alabama  
May 09, 2011  
 
 
 
 
Keywords: Islet Isolation, Diabetes, Quadrupole Magnetic Sorter,  
Particle Tracking Velocimetry 
 
 
 
Copyright 2010 by Venkata Sunil Kumar Sajja 
 
  
Approved by  
 
Thomas R. Hanley, Chair, Professor of Chemical Engineering 
Paul W. Todd, Chief Scientist, Techshot Inc. 
James F. Leary, Professor of Biomedical Engineering, Purdue University 
Robert P. Chambers, Professor of Chemical Engineering 
Elizabeth Lipke, Assistant Professor of Chemical Engineering 
 
 
 
 
 ii
Abstract 
 
Pancreatic islet transplantation offers a viable option to achieve permanent metabolic 
control in Type 1 diabetes patients.  However, large quantities of pure viable donor islet cells are 
necessary for transplantation.  Using currently available islet isolation methods multiple donor 
organs are required to achieve successful transplantation, and there is a demand for an isolation 
method with high islet yield and viability.  Additionally, with porcine xeno-islet cell 
transplantation providing much hope, improving the porcine islet isolation process has become a 
worthwhile endeavor.  This dissertation is the summery of the work aimed to develop a 
Quadrupole Magnetic Sorter to isolate pancreatic islets from exocrine tissue. 
Computational Fluid Dynamics (CFD) simulations were used (Chapter 2) to predict the 
flow patterns, pressure drop and nonspecific crossover in a newly designed QMS flow channel 
for the isolation of pancreatic islets of Langerhans.  Simulation results were compared with the 
theoretically and experimentally determined results to validate the CFD model.  CFD simulations 
were employed to compare performance of two models of QMS flow channels with differing 
splitter positions.  Results of the simulations were used to show that one design gives up to 10% 
less nonspecific crossover than another and this model can be used to optimize the flow channel 
design to achieve maximum purity of magnetic particles.  
 Magnetic isolation is a promising method for separating and concentrating pancreatic 
islets for transplantation in Type 1 Diabetes patients.  Continuous magnetic islet sorter was 
designed to overcome the restrictions of current purification methods that result in limited yield, 
viability and purity of the isolated islets.  The performance of the islet sorter depends on the 
 iii 
resulting speed of the islets in an applied magnetic field, a property known as magnetophoretic 
mobility.  Essential to the design and operation of the magnetic sorter is a method to measure the 
magnetophoretic mobilities of magnetically infused islets.  Magnetic particle tracking 
velocimeter (MPTV) was developed to measure (Chapter 3) the magnetophoretic mobility of 
particles up to 1000 microns in diameter.  Velocity measurements are performed in a well-
characterized isokinetic magnetic energy gradient using video imaging followed by analysis of 
the video images using a computer algorithm that produces histogram of absolute mobilities.  
Mobility distributions obtained indicated that magnetized islets have sufficient mobility to be 
captured by the proposed sorting method, with this result confirmed in test isolations of 
magnetized islets. 
To achieve islet isolation with high purity and yield Quadrupole Magnetic Sorting 
(QMS), a single cell separation method, is being modified for the isolation of pancreatic islets 
(Chapter 4).  Islets are infused with 4.6?m Dynabeads? and separated continuously with QMS.  
Results from 10 porcine pancreas isolations indicated possibility of infusing islets with magnetic 
beads and isolating them continuously by reducing the exposure time of islets to enzymes.  QMS 
isolated islets showed good morphology compared to standard COBE isolated islets and the 
Oxygen Consumption Rate (OCR) per DNA measurements confirmed the viability of the islets 
after isolation.  QMS isolation can save the culture time and help to eliminate the mechanical 
stress due to centrifugation on the islets.  Nude mice transplantation results confirmed 
Dynabeads do not affect the functionality of the islets. 
 
 
 
 
 
 
 vi
Acknowledgements 
 
 
I would like to express my sincere gratitude towards Dr. Thomas R. Hanley for giving me 
an opportunity to work with him on this research. I also would like to sincerely thank him for his 
consistent support, warm encouragement I would like to extend my thanks to Dr. Paul W. Todd 
for his support, guidance, thoughtful ideas, and also for serving on my committee. I would like to 
acknowledge Dr. James F. Leary, Dr. Robert P. Chambers and Dr. Elizabeth Lipke for serving 
on my committee and university reader Dr. Christopher J. Easley. 
I would like to express my sincere thanks to David J. Kennedy and Byron Guernsey from 
IKOTECH, LLC for their help and support. I would like to thank Mark Deuser and John 
Vellinger for giving me opportunity to work at Techshot, IN.  I express my earnest appreciation 
toward the members of Techshot for their kind help, especially Alan Constance and Bill Johnson.  
I also want to thank Dr. Klearchos Papas and his group from University of Minnesota, 
MN, Mike Taylor and his group at Cell Tissue Systems and Robert McCarthy and his group at 
Vitacyte, IN group for providing facilities and help with isolations.   
I would like to express deep gratitude and gratefulness to my father Mr. Seshagiri Rao 
Sajja for being a constant source of inspiration and motivation, mother Mrs. Nagamani Sajja for 
her enduring love and immense moral support and family members Sunanda, Picheswara Rao 
and Sai Sreeja. My most special thanks belong to my wife Saroja for her love, sacrifice and 
subject discussions. 
 vi
Table of Contents 
 
 
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  ii 
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv 
List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix 
List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  xiii 
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 
1.1 Diabetes and Transplantation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 
1.2 Recordi Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 
1.3 Purification of Islets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 
1.4 Density Gradient Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 
1.5 Magnetic Separations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 
1.6 Quadrupole Magnetic Sorting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9 
1.7 Magnetic Particle Tracking Velocimetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 
1.8 Magnetic Beads. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 
1.9 Research Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 
1.10 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16 
2. Computational Fluid Dynamics Simulation of a Quadrupole Magnetic Sorter 
Flow Channel: Effect of Splitter Position on Nonspecific Crossover . . . . . . . . . . . . . . . 23 
2.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23 
2.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  24 
2.3 Materials and Methods . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
 vi
2.3.1 QMS and Separation Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  . . . 25 
2.3.2 Experimental Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  . . .  29 
2.3.3 Computational Fluid Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  . . . 29 
2.3.4 Governing Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  . . 30 
2.3.5 Discretization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  34 
2.3.6 Solvers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 
2.3.7 Pressure-Velocity Coupling Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 
2.3.8 Geometric Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .   37 
2.3.9 Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  41 
2.3.10 Simulation Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  41 
2.4 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .41 
2.4.1 CFD Simulations: Flow Analysis and Pressure Drop Predictions . . . . . . .  41 
2.4.2 Nonspecific Crossover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  44 
2.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  51 
2.6 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  51 
2.7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  52 
3. Application of Magnetic Particle Tracking Velocimetry to Quadrupole Magnetic           
Sorting of Porcine Pancreatic Islets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 
3.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  55 
3.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  56 
3.3   Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .58 
3.4   Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 
3.4.1 Particles and Viscous liquid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  61 
3.4.2 QMS system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 
 vii
3.4.3 MPTV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 
3.4.4 Magnet Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 
3.4.5 Flow system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  63 
3.4.6 Imaging System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  63 
3.4.7 Analysis of video data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 
3.4.8 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 
3.5 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .66 
3.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .78 
3.7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .80 
4. QMS Isolation of Porcine Islets of Langerhans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 
4.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .83 
4.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .84 
4.3 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 
4.3.1 QMS and separation theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  86 
4.3.2 Magnetophoretic Mobility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  88 
4.3.3 Pancreas Procurement and labeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  88 
4.3.4 MRI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 
4.3.5 Histopathology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  90 
4.3.6 Pancreas Digestion and Islets Isolation  . . . . . . . . . . . . . . . . . . . . . . . . . . .  90 
4.4 Islet Quality Assessment  . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 
4.4.1 Oxygen Consumption Rate (OCR) Assay . . . . . . . . . . . . . . . . . . . . . . . . .  91 
4.4.2 DNA Quantification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  91 
4.4.3 Nude Mouse Bioassay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 
 viii 
4.5 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .92 
4.5.1 Cross-over studies of QMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .   92 
4.5.2 Isolation of Islets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .   98 
4.5.3 Magnetophoretic Mobility Measurements . . . . . . . . . . . . . . . . . . . . . . . .  100 
4.5.4 Histology of Pancreas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 
4.5.5 Quality assessment of Islets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  105 
4.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .113 
4.7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .114 
5. Summary and Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .118 
Appendix A: QMS Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  120 
Appendix B: MPTV Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  127 
Appendix C: Fluent Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  130 
 
 
 
 
 
 
 
 
 
 xiii 
List of Figures 
 
 
Figure 1.1 Ricordi chamber used for islet isolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 
Figure 1.2 Flow diagram of the islets isolation process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 
Figure 1.3 Steps involved in Islet Transplantation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  6 
Figure 1.4 Insulin independence, insulin dependence or absence of fasting           
C-peptide post last infusion islets alone recipients . . . . . . . . . . . . . . . . . . . . . . . .10 
Figure 1.5 Schematic diagram of the quadrupole magnetic cell sorter (QMS)   . . . . . . . . . . . . . 11 
Figure 1.6 Cell tracking velocimeter (CTV) showing coordinate orientation . . . . . . . . . . . . . . . 13 
Figure 1.7 Magnetophoretic Mobility Histograms for selected microbeads,            
50nm ? 10?m . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 
Figure 2.1 Schematic diagram of the quadrupole magnetic cell sorter (QMS) . . . . . . . . . . . . . . 27 
Figure 2.2 Geometry drawings of Prototype I and Prototype II flow channels . . . . . . . . . . . . . . 31 
Figure 2.3 Flow chart for segregated solver used in this study . . . . . . . . . . . . . . . . . . . . . . . . .   38 
Figure 2.4 Unstructured triangular mesh on the faces of QMS flow channel . . . . . . . . . . . . . . . 39 
Figure 2.5 Unstructured mesh along with the boundary layer on the inlet pipes           
and channel face . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 
Figure 2.6 Velocity (m/s) contours in x plane at different distance from flow            
distributor to the end point of the splitter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 
Figure 2.7 a) Velocity (m/s) magnitude contours in the flow channel at z = 1.1875?           
(flow is in positive x direction. b) Velocity profile of the             
  fluid at the middle point of the channel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 
Figure 2.8 a) Pressure (Pa) contours in the flow channel at z = 1.1875?            
b) Comparison of the pressure drop from CFD simulations and  
       empirical correlations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
 xiii 
Figure 2.9 Comparison of calculated and observed nonspecific crossover as a            
function of outlet flow ratio at a total flow rate of 400 ml/min and inlet  
  flow ratio of 0.25 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  48 
Figure 2.10 Comparison of the nonspecific crossover as a function of total  flow  
  rate at an inlet and outlet flow ratio of 0.25 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 
Figure 2.11 Nonspecific crossover as a function of particle concentration            
                 in the sample at a total flow rate of 400ml/min, inlet and  
outlet flow ratio of 0.25 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 
Figure 3.1 Schematic diagram of the quadrupole magnetic cell sorter (QMS) . . . . . . . . . . . . . . 60 
Figure 3.2 A simplified diagrammatic representation of an MCTV system . . . . . . . . . . . . . . . .  64 
Figure 3.3 Mobility histogram given by MPTV for Dynabeads . . . . . . . . . . . . . . . . . . . . . . . . .  67 
Figure 3.4 Histograms of Dynabeads in each fraction of the QMS output . . . . . . . . . . . . . . . . .  68 
Figure 3.5 Particle tracks developed by MPTV. Red dots are the particles tracked            
and black spots are disturbance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  69 
Figure 3.6 Magnetophoretic mobility histogram of BSI magnetic particles . . . . . . . . . . . . . . . . 71 
Figure 3.7 Magnetophoretic mobility histogram of BSII magnetic particles . . . . . . . . . . . . . . .  72 
Figure 3.8 Magnetophoretic mobility histogram of BSIII magnetic particles . . . . . . . . . . . . . .   73 
Figure 3.9 Comparison of MPTV predicted fractional recovery of BSI particles           
     in the three outlet fractions of the QMS with experimental results . . . . . . . . . . . 75 
Figure 3.10 Comparison of MPTV predicted fractional recovery of BSII particles           
       in the three outlet fractions of the QMS with experimental results . . . . . . . . . . . 76 
Figure 3.11 Comparison of MPTV predicted fractional recovery of BSI             
particles in the three outlet fractions of the QMS with experimental results . . . . 77 
Figure 3.12 a) Magnetophoretic mobility histogram of pancreatic islets isolated           
with QMS. b) MPTV predicted b fraction c) a fraction d) wall fraction of  
the islets at a total flow rate of 400ml/min and Ra? = 0.25 and Ra = 0.6 . . . . . . .79 
Figure 4.1 Schematic diagram of the quadrupole magnetic cell sorter (QMS) . . . . . . . . . . . . . . 87 
Figure 4.2 Crossover of the acinar tissue at different outlet flow ratios for            
                     constant inlet flow ratio and total flow rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 
 xiv
Figure 4.3 Crossover of the acinar tissue at different inlet flow ratios for constant          
outlet flow ratio and total flow rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .96 
Figure 4.4 Crossover of the acinar tissue at different tissue concentrations at          
constant inlet flow ratio, outlet flow ratio and total flow rate . . . . . . . . . . . . . . . 97 
Figure 4.5 Magnetophoretic mobility histogram of digested tissue sample taken         
before sending it through QMS for isolation . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 
Figure 4.6 Magnetophoretic mobility histogram of digested tissue sample taken         
from Negative (a) fraction of QMS isolation . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 
Figure 4.7 Magnetophoretic mobility histogram of digested tissue sample taken        
from Positive (b) fraction of QMS isolation . . . . . . . . . . . . . . . . . . . . . . . . . . . .103 
Figure 4.8 Magnetophoretic mobility histogram of digested tissue sample taken         
from wall fraction of QMS isolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .104 
Figure 4.9 MRI of the control connecting/duodenal lobe (above) and the          
experimental splenic lobe (below), in which infused MP resulted in  
well-distributed hypointense regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..106 
Figure 4.10 Representative low and high magnification micrographs of an islet           
located in the experimental splenic lobe (distal region), illustrating  
minimal accumulation of magnetic particles in the acinar tissue (Fig. 1A, H/E) 
 and significant accumulation within the islet (Fig. 1B, H/E) and magnetic 
particles within capillaries of an islet located in the proximal splenic lobe,  
near the site of infusion at the celiac trunk (Fig. 1C, insulin) . . . . . . . . . . . . . . .107 
Figure 4.11 Pictures of the islets taken under microscope at 20X magnification         
a) islets isolated with QMS b) islets isolated with COBE . . . . . . . . . . . . . . . . . 108 
Figure 4.12 Stimulation index for OCR measurements from control and infused         
islets from three separate isolations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 
Figure 4.13 Stimulation index for average OCR measurements from control          
and infused islets from three separate isolations at day0 and day7 . . . . . . . . . . 111 
Figure A.1 Photograph of the QMS setup with fluid bags . . . . . . . . . . . . .  . . . . . . . . . . . . . . . 120 
Figure A.2 Components of the QMS. Boxed labels are permanent QMS components,          
unboxed labels are consumable items replaced after one use . . . . . . . . . . . . . . .121 
Figure A.3 Screnshot of the QMS software used to control the QMS operation . . . . . . . . . . . .122 
Figure A.4 QMS set up for pancreas isolation at University of Minnesota laboratory . . . . . . . 123 
 xv
Figure A.5 Absorbance detector traces for P810 isolation run.  The QMS was turned 
on at time t=0 minutes.  The turbidity sensor data logging was activated 
at t=42 minutes [1].  The digestion process was switched to recovery 
mode and the QMS began operations, with tissue leaving the flow  
channel at t=48 minutes [2].  The tissue concentration with the Ricordi 
chamber upright peaked at approximately t=72 minutes [3].   
The Ricordi chamber was inverted to maximize the recovery at  
t=92 minutes [4].  The tissue exiting the inverted Ricordi chamber 
peaked at t=102 minutes [5].  The isolation was terminated at  
approximately t=114 minutes [6] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 
Figure B.1 Photograph of the MPTV assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 
Figure B.2 Photograph of the MPTV set up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  128 
Figure B.2 Screenshot of the IKOVISION used to run the MPTV . . . . . . . . . . . . . . . . . . . . . . 129 
Figure C.1 Summery of the particle flow simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .132 
Figure C.2 Particles trajectories calculated in Fluent simulations . . . . . . . . . . . . . . . . . . . . . . . 133 
 
 
 
 
 
 
 
 
 
 
 
 xiii 
List of Tables 
 
 
Table 4.1 Total flow rates used in crossover studies and corresponding ISS and           
OSS values.   Tissue which enters in between core and ISS has to  
cross the OSS to reach the positive fraction . . . . . . . . . . . . . . . . . . . . . . . . . . . .  94 
Table 4.2 Purity and yields of the islets from all the experiments conducted with           
four porcine pancreas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 
Table A.1 Summary of the Porcine isolation experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . .  125 
 1
1. Introduction 
1.1  Diabetes and Transplantation  
Diabetes is a group of metabolic diseases that cause hyperglycemia.  The consequences 
of diabetes include retinopathy (loss of vision), nephropathy (renal failure) and peripheral 
neuropathy with risk of amputation.  The majority of diabetes cases fall into two categories:  In 
type 1 diabetes, the cause is deficiency of insulin secretion.  Type 1 diabetes accounts for 5 to 10 
per cent of diagnosed diabetes.  In type 2 diabetes the cause is combination of resistance to 
insulin action and deficit insulin secretion (Gavin et al., 2003). 
According to the reports presented by CDC in 2008, 24 million Americans, 8 per cent of 
the population, are diabetic (CDC, 2008).  Diabetes costs the US economy $132 billion per year, 
$92 billion in direct medical costs (CDC, 2002; Paul et al., 2003).  It has been suggested that 
insulin independence can be achieved in diabetes with islet transplantation (Markmann et al., 
2003).  
Since the discovery of insulin in 1921, insulin therapy has saved many lives of type 1 
diabetics.  Insulin treatment cannot fully prevent chronic complications, and intensive insulin 
treatment to improve metabolic control has paralleled an increased risk of severe hypoglycemia 
(Berney et al., 2001; Oberholzer et al., 1999).  The English surgeon Watson Williams was the 
first to attempt transplantation of pancreatic fragments in 1893 from a sheep to a boy.  The first 
rodent pancreas transplantation was performed in 1966 along with kidney transplantation at the 
University of Minnesota (Ballinger et al., 1972).  The International Pancreas Transplant Registry 
reports 17,000 whole pancreas transplants performed in the US up to 2004.  Success after one
 2
year varied from 72 to 84 per cent. Although whole pancreas transplantation is the most reliable 
means of restoring full metabolic control, it involves complex surgery.  With only 5,000 cadaver 
organs available in any given year, transplantation is not available to treat all diabetes patients 
(Markmann et al., 2003).  Islet transplantation offers a viable option to insulin therapy for 
improved metabolic control and to full organ transplantation with lower cost and risk. 
Successful islet transplantation can result in insulin independence (Berney et al., 2001).  
The recent dramatic improvement in the success rate of the islet transplantation in humans has 
prompted considerable interest for more widespread application of this methodology (Shapiro et 
al., 2000).  The first clinical islet transplantation in humans was attempted in 1985 at Washington 
University Medical School in St. Louis (Ricordi et al., 1988).  The process involved maceration 
of the organ into cell fragment broth, from which islets were purified by a density gradient 
method as islets are slightly less dense than acinar tissue.  Islets constitute only 1-2 % of the 
pancreas volume and their isolation and purification are stressful mechanical and enzymatic 
procedures that can inflict significant damage, which may be further amplified by prolonged 
times of warm and cold ischemia with human islets (London et al., 1994; Robertson et al., 1998).  
One limiting factor to the islet transplantation is the need for the large numbers of islets, obtained 
from more than one donor pancreas per recipient. 
1.2  Recordi Method 
The most important development for islet isolation was the automated method introduced 
by Dr. Camillo Ricordi.  Ricordi?s method utilized a refined mixture of enzymes (collagenase) 
produced by the bacterium Clostridium hystoliticum.  Ricordi disconnected the organ from the 
duodenum, and a canula was inserted into the main pancreatic duct which serves as the outlet for 
digestive enzymes produced by the pancreatic exocrine.  With all the accessory ducts clamped, 
 3
the organ was distended by collagenase injected into the main duct.  The dilated organ was 
placed in a conical chamber, the ?Ricordi Chamber (Figure 1.1)?, through which the heated 
enzyme solution was circulated.  A ball mill using either glass or ceramic marbles was used to 
liberate islets from the tissue (Ricordi et al., 1988).  Flow diagram of the Ricordi digestion 
process is shown in figure 1.2. 
The output from the Ricardi Chamber contains islets with some exocrine tissue.  Large 
volumes of contaminated acinar tissue within islet auto grafts have been associated with portal 
hypertension, hepatic infarction, splenic bleeding and death.  Therefore, it is necessary to purify 
islets before transplantation (Sulaiman et al., 2006).  A diagram of the steps involved in the 
isolation and transplantation of pancreatic islets is shown in figure 1.3. 
1.3  Purification of Islets 
Insulin independence can be achieved in type 1 diabetes by transplanting 10,000 Islet 
Equivalents (IEQ) per kilogram recipient weight of purified islets (Sakuma et al., 2008).  To 
achieve high yields with required purities, many islet purification techniques have been 
developed for isolation of islets from exocrine tissue either by targeting islets or by targeting 
exocrine tissue, e.g. hand picking of islets, fluorescence-activated sorting by staining islets with 
neutral red (Gray et al., 1989; Jindal et al., 1994), or destruction of non-islet tissue by laser 
energy (Brunicardi et al., 1994).  All these techniques were successfully employed to purify rat 
islets but failed to scale up to the human tissue digest volume. 
 Islets or acinar tissue has been labeled with magnetic beads coated with monoclonal 
antibodies and then isolated with bar magnet (Muller-Ruchhoiltz et al., 1987; Winoto-Morbach 
et al., 1989; Winoto-Morbach et al., 1989a; Soon-Shiong et al., 1990; Winoto-Morbach et al., 
1994; Davies et al., 1994; Davies et al., 1996).   Since islets make up only one to two per cent of  
 4
 
 
Figure 1.1: Ricordi chamber used for islet isolation (www.biorep.com). 
 
 
 
 
 
 
 
 
 
 
 5
 
 
Figure 1.2: Flow diagram of the islets isolation process (Ricordi, 1992). 
 
 
 
 
 
 
 
 
 
 
 6
 
Figure 1.3:  Steps involved in Islet Transplantation (Sulaiman et al., 2006). 
 
 
 
 
 
 
 
 
 
 
 
 7
pancreas, negative selection is difficult, requiring effective labeling of 98 per cent of the tissue.  
In addition, negative selection removes only the acinar cells, leaving ductal and other 
unidentified cell types remaining after pancreatic digestion.  Positive selection of the islets offers 
the advantage over negative selection and the presence of beads in islets does not alter islet 
functionality (Nandigala et al., 1997).  
1.4  Density Gradient Method 
 While several methods of the islet isolation have been developed, the most successful 
method is density gradient centrifugation.  In this method, tissue digestate is centrifuged and re-
suspended in a dense medium.  Less dense media are layered on top and the gradient is fed to a 
COBE 2991 cell processor.  After several minutes of centrifugation, islets are separated from the 
interface between two discontinuous gradient layers (Samejima et al., 1998).  
The current clinical practice using on density gradient centrifugation does not achieve the 
necessary yield and purity of islets as this process relies on the small density difference between 
islets and acinar tissue.  Density gradient cannot be standardized because tissue density and size 
distribution vary with each pancreas isolation.  Tissue size is a function of the degree of 
enzymatic digestion, over which there is no direct control.  In addition, density gradients have a 
limited ineffectiveness for digests in which islets have not been completely disconnected from 
the surrounding acinar tissue (London et al., 1998; Soon-Shiong et al., 1989). 
Exocrine tissue present in purified islets, such as lymph nodes, vascular tissue and ductal 
fragments, are major stimuli for induction of the immune response that results in acute rejection 
(Mitsukazu et al., 1986).  Although some studies indicate that islets should be transplanted 
unpurified, transplantation of purified islets should be preferred to achieve appropriate 
revascularization of transplants (Heuser et al., 2000). 
 8
1.5  Magnetic Separations 
 There is a major disparity between the number of available organ donors and recipients in 
need.  Using current protocols 1 to 4 pancreas may be required to restore normal blood sugars in 
a single recipient.  This is in large part due to variations in islet recovery and potency and the 
substantial losses in islet yield during organ procurement and storage as well as islet isolation 
and purification (Hering et al., 2002).  Additionally, with porcine xeno-islet cell transplantation 
providing much hope (Hering et al., 2006), improving the porcine islet isolation process has 
become a worthwhile endeavor.  Figure 1.4 shows the insulin independence achived in the 
patients who received islets transplantation in last decade (Michael et al., 2009). 
Immunomagnetic cell separation provides a highly attractive alternative to density?dependent 
methods for islet purification.  The versatility of magnetic fields makes them a useful candidate 
for biological separations.  Magnetic separations were initially developed to isolate glomuruli 
from murine kidneys by infusing iron oxide particles (Joyce Gauthier et al., 1988).  Since islets 
of Langerhans have a similar angioarchitecture to mouse glumeruli, islets can be purified using 
magnetic force (Pinkse et al., 2004).  
The use of Dynabeads for islet purification was first reported in 1989.  Magnetic particles 
coated with antibodies impart mobility to selected cells and allow separation by a magnetic field.  
Magnetic fields have been used to select islets and separate them from the acinar tissue (Winoto-
Morbach et al., 1989b; Davies et al., 1995).  The degree of purification achieved in rats nearly 
reached hand selection quality (Muller-Ruchholtz et al., 1987).  By positive selection, many 
unwanted tissue elements, such as lymph nodes, arteries and other ductal fragments, would not 
contaminate the purified islets.  Unlike the density gradient method, magnetic separation subjects 
islets to little mechanical stress and requires less time for separation.  Magnetic separations are 
 9
effective in any medium and are easily scalable.  The use of quadrupole magnetic separation 
increases the purity and yield of islets compared to use of simple magnet (Davies et al., 1994).  
1.6 - Quadrupole Magnetic Sorting 
The QMS technology is based on a process known as split-flow lateral transport thin 
separation, also referred to by the acronym SPLITT, which is a subset of Field Flow 
Fractionation (FFF) technology. SPLITT was introduced by J.C. Giddings (Giddings et al., 1985) 
to extend the capabilities of FFF to the separation of colloids, macromolecules, and particulate 
materials (Martin et al., 1992).  Quadrupole magnetic sorting (QMS) has been extensively 
modeled and tested and proved successful for single cell separation.  QMS is a high-throughput, 
high-gradient, continuous magnetic cell separation system.  QMS was initially designed for the 
positive separation of immunomagnetically labeled single cells from nonmagnetic cell 
population.  Figure 1.5 provides a diagrammatic view of the QMS sorting mechanism.  QMS 
employs three subsystems for operation:  a fluid flow channel, a magnetic field for particle 
separation and pumps to regulate flow into and out of the flow channel.  A sample containing 
labeled and unlabeled cells enters the flow channel (a?) with carrier buffer (b?).  Separated 
labeled cells (b) and unlabeled cells (a) leave the flow channel at the bottom.  Separation is a 
function of many factors including the entering flow rates, the total flow rate and the degree of 
magnetization of cells.  Cell labeling may be quantified by the magnetophoretic mobility (Sun et 
al., 1998; Williams et al., 1999; Hoyos et al., 2000; McCloskey et al., 2003; McCloskey et al., 
2003a; Moore et al., 2004; Tong et al., 2007; Jing et al., 2007).  
 
 
 10
 
 
Figure 1.4:  Insulin independence, insulin dependence or absence of Fasting C-peptide post last 
infusion of islets alone recipients (Michael et al., 2009). 
 
 11
 
 
Figure 1.5:  Schematic diagram of the quadrupole magnetic cell sorter (QMS). 
 
 
 
Qb?
Qa?
Qa
Qb
MA
GN
ET
MA
GN
ET
Tr
an
sp
ort
 La
mi
na
Tr
an
sp
ort
 La
mi
na
ri
ro
riss
ross
MA
GN
ET
MA
GN
ET
Tr
an
sp
ort
 La
mi
na
Tr
an
sp
ort
 La
mi
na
 12
1.7 ? Magnetic Particle Tracking Velocimetry 
Magnetophoretic mobility of magnetically labeled cells can be measured by a Magnetic 
Particle tracking velocimetry (MPTV).  Figure 1.6 illustrates the MPTV system MPTV leverages 
current technologies in video microscopy, computer processing speed and finite element analysis 
for magnetic fields to measure the induced motion of cells and particles in the highly 
characterized magnetic and gravity field.  The motion of the particles in the known field is then 
translated into a characteristic parameter called magnetophoretic mobility in a magnetic field and 
sedimentation rate in a gravity field.  MPTV measurements involve videotaping the movement of 
immunomagnetically labeled particles through a known viscosity medium and magnetic 
susceptibility in a well-defined magnetic energy density gradient.  The velocity of each particle 
along with its location within the magnetic energy gradient is recorded.  From this information, 
magnetophoretic mobility of each particle is calculated.  A significant advantage of this method 
over other techniques is that larger numbers of cells can be processed in very little time (Moore 
et al., 2004; Chalmers et al., 1999; Zhang et al., 2002; Chalmers et al., 1999a).   
 
 
 
 13
 
Figure 1.6: Magnetic Particle Tracking Velocimeter (MPTV) showing coordinate orientation. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Computer 
Sample
Buffer
Observation Channel
Magnet Pole Pieces
Lens
Camera
Vacuum pump
 14
 
1.8  Magnetic Beads 
 Several manufactures like SpherotechTM, MiltenyiTM, InvotrogenTM produce magnetic 
beads from 50 nm to 10 ?m in size for biological separation applications.  Some of the magnetic 
beads available for separations are tested for their magnetophoretic mobility.  After considering 
the differences in the vasculature of the islets and acinar tissue Dynabeads? were selected to use 
for pancreas infusion to isolate islets from acinar tissue.  Figure 1.7 shows the histograms of the 
magnetophoretic mobilities of different magnetic beads from 50nm to 10?m in size and 
Dynabeads shows the highest mobility among the all beads tested. 
 
1.9  Research Objectives 
 The application of the QMS system for isolation of islets can lead to greater efficiencies 
and lower costs in the islet transplantation for Type 1 diabetics.  The overall objective of this 
research is to design and characterize a flow channel and develop optimized conditions for islet 
isolation using QMS.  The specific objectives are as follows: 
? To develop a reliable QMS process for isolation of human islets from acinar tissue for 
transplantation. 
? To design and operate an improved MPTV to determine magnetophoretic mobilities of 
islets. 
? To optimize flow channel operation using computational fluid dynamics to simulate the 
QMS flow channel to study the flow patterns and pressure variations.   
 
 
 
 15
 
 
Figure 1.7:  Magnetophoretic Mobility Histograms for selected microbeads, 50nm ? 10?m (data provided by David J. Kennedy of 
IKOTECH, LLC)
 16
1.10  References 
 
Ballinger WF, Lacy PE. 1972. Transplantation of intact pancreatic islets in rats. Surgery 72:175-
186. 
Berney T, Buhler L, Caulfield A, Oberholzer J, Toso CH, Alejandro R, Cooper DKC, Ricordi C, 
Morel PH. 2001. Transplantation of islets of Langerhans New developments. Swiss Med 
Wkly 47-48:671-680. 
Brunicardi FC, Oh Y, Shevlin L, Suh E, Kleinman R, Stein E, Lipaz G, Plant DV, Imagawa D, 
Fetterman HR. 1994. Laser destruction of human nonislet pancreatic tissue. Transplant 
Proc 26(6):3354-3355. 
CDC. 2008. Number of people with diabetes increases to 24 million, in Press Release, D.o.h.a.h. 
services, Editor, Center for Disease Control. 
CDC. 2002. National Health Interview Survey. Centers for Disease Control and Prevention, 
National Center for Chronic Disease Prevention and Health Promotion, Division of 
Diabetes Translation. 
Chalmers JJ, Zhao Y, Nakamura M, Melnik K, Lasky L, Moore L, Zborowski M. 1999.  An 
Instrument to Determine the Magnetophoretic Mobility of Labeled, Biological Cells and 
Paramagnetic Particles.  Journal of Magnetism and Magnetic Materials 194:231-241. 
Chalmers JJ, Haam S, Zhao Y, McCloskey K, Moore L, Zborowski M, Williams PS. 1999a. 
Quantification of Cellular Properties from External Fields and Resulting Induced 
Velocity: Cellular Hydrodynamic Diameter.  Biotechnology and Bioengineering 64:509-
518. 
 17
Davies JE, Robertson GS, Swift S, Chamberlain J, Bell PR, James RF, London NJ. 1994. Use of 
a quadripole magnet significantly improves immunomagnetic islet purification. 
Transplant Proc 26(2):649-650. 
Davies JE, Winoto-Morbach S, Ulrichs K, James RF, Robertson GS. 1996. A comparison of the 
use of two immunomagnetic microspheres for secondary purification of pancreatic islets. 
Transpntation 62(9):1301-1306. 
Davies JE, James RF, London NJ, Robertson GS. 1995. Optimization of the magnetic field used 
for immunomagnetic islet purification. Transplantation 59(5):767-771.  
Gavin JR, Alberti K, Davidson M, DeFronzo R, Drash A, Gabbe S, Genuth S, Harris M, Kahn R, 
Keen H, Knowler W, Lebovitz H, Maclaren N, Palmer J, Raskin P, Rizza R, Stern M. 
2003. Report of the expert committee on the diagnosis and classification of diabetes 
mellitus: Committee Report. Diabetes Care 26(S1):S5-S20. 
Giddings JC. 1985. A System Based on Split-Flow Lateral-Transport Thin (SPLITT) Separation 
Cells for Rapid and Continuous Particle Fractionation. Sep Sci Technol 20:749-768. 
Gray DW, G?hde W, Carter N, Heiden T, Morris PJ. 1989. Separation of Pancreatic islets by 
Fluorescence-Activated Sorting, Diabetes 38(suppl 1):133-135. 
Hering BJ, Matsumoto I, Sawada T, Nakano M, Sakai T, Kandaswamy M, Sutherland DER. 
2002. Impact of two-layer pancreas preservation on islet isolation and transplantation. 
Transplantation 74:1813-1816. 
Hering BJ, Wijkstrom M, Graham ML, et al. Prolonged diabetes reversal after intraportal 
xenotransplantation of wild-type porcine islets in immunosuppressed nonhuman 
primates. Nat Med 2006; 12 (3): 301. 
 18
Heuser M, Wolf B, Vollmar B, Menger MD. 2000. Exocrine contamination of isolated islets of 
Langerhans deteriorates the process of revascularization after free transplantation. 
Transplantation 69(5):756-761. 
Hoyos M, Moore LR, McCloskey KE, Margel S, Zuberi M, Chalmers JJ, Zborowski M. 2000. 
Study of magnetic particles pulsed-injected into an annular SPLITT-like channel inside a 
quadrupole magnetic field. J Chromatogr A 903(1-2):99-116. 
http://www.biorep.com/RicordiChamber-1#images/ptos/diabetes/ricordi-chamber-gal2.jpg 
(09/17/2010). 
Jindal RM, McShane P, Gray DWR, Morris PJ. Isolation and Purification of Pancreatic 1994. 
Islets by Fluorescence Activates Islet Sorter, Transplantation Proceedings 26(2):653. 
Jing Y, Chalmers JJ, Zborowski M. 2007. Blood progenitor cell separation from linical 
leukapheresis product by magnetic nanoparticle binding and agnetophoresis. Biotechnol 
Bioeng 96(6):1139-1154. 
Joyce Gauthier V, Mart Mannik. 1988. A method for isolation of mouse glomeruli for 
quantitation of immune deposits. Kidney International 33:897-899. 
London NJM, Swift SM, Clayton HA. 1998. Isolation, culture and functional evaluation of islets 
of Langerhans. Diabetes & Metabolism 23:200-207. 
London NJ, Robertson GS, Chadwick DR, Johnson PR James RF, Bell PR. 1994. Human 
pancreatic islet isolation and transplantation. Clin Transplant 8(5):421-459. 
Markmann JF, Deng SP, Huang X, Desai NM, Velidedeoglu EH, Lui CY, Frank A, Markmann 
E, Palanjian M, Brayman K, Wolf B, Bell E, Vitamaniuk M, Doliba N, Matschinsky F, 
Barker CF, Naji A. 2003. Insulin independence following isolated islet transplantation 
and single islet infusions. Annals of Surgery 237(6):741-751. 
 19
Martin M, Williams PS. 1992. Theoretical Advancement in Chromatography and Related 
Separation Techniques; Dondi, F., Guiochon, G., Eds.; Kluwer Academic Publishers: 
Dordrecht, The Netherlands 51:513- 580. 
McCloskey KE, Chalmers JJ, Zborowski M. 2003. Magnetic cell separation: Characterization of 
magnetophoretic mobility. Analytical Chemistry 75(24):6868-6874. 
McCloskey KM, Moore LR, Hoyos M, Rodrigues A, Chalmers JJ, Zborowski M. 2003a. 
Magnetic cell separation is a function of antibody binding capacity (ABC). Biotechnol 
Prog 19(3):899-907. 
Michael A, Hering B. 2009. CITR annual report. 
Mitsukazu Gotoh, Takashi Maki, Susumu Satomi, Janis Porter, Anthany P Monaco. 1986. 
Immunological characteristics of purified pancreatic islet grafts, Transplantation 
42(4):387-390. 
Moore LR, Milliron S, Williams PS, Chalmers JJ, Margel S, Zborowski M.  2004. Control of 
Magnetophoretic Mobility by Susceptibility-Modified Solutions as Evaluated by Cell 
Tracking Velocimetry and Continuous Magnetic Sorting.  Analytical Chemistry 76:3899-
3907. 
Muller-Ruchholtz W, Leyhausen G, Petersen P, Schubert G, Ulrichs K. 1987. A simple 
methodological principle for large scale extraction and purification of collagenase-
digested islets. Transplant Proc 19(1):911-915. 
Nandigala P, Chen TH, Yang C, Hsu WH, Heath C. 1997. Immunomagnetic isolation of islets 
from rat pancreas. Biotechnol Prog 13:844-848. 
Oberholzer J, Triponez F, Lou J, Morel P. 1999. Clinical islet transplantation: A review. Annals 
of the New York Academy of Sciences 875:189-199. 
 20
Paul H, Tim D, Plamen N. 2003. Economic costs of Diabetes in the U.S. in 2002, Diabetes Care 
26:917-932. 
Pinkse GGM, Steenvoorde E, Hogendoorn S, Noteborn M, Terpstra OT, Bruitin JA, De Heer E. 
2004. Stable transplantation results of magnetically retracted islets: a novel method. 
Diabetologia 47:55-61. 
Ricordi C, SCHARP dw, Lacy PE. 1988. Reversal of diabetes in nude mice after transplantation 
of fresh and 7 days cultures (24oC) human pancreatic islets. Transplantation 45: 994-996. 
Ricordi C, Lacy PE, Finke EH, Olack BJ, Scharp DW. 1988. Automated method for isolation 
of human pancreatic islets. Diabetes 37:413-420. 
Ricordi C. The automated method for islet isolation. In Pancreatic Islet Cell Transplantation 
1892-1991: One Century of Transplantation for Diabetes. Ricordi C, Ed. Austin, TX, 
R.G. Landes Company 1992, p.99-112. 
Robertson GS, Dennison AR, Johnson PR, London NJ. 1998. A review of pancreatic islet 
autotransplantation. hepatogastroenterology 45(19):226-235. 
Sakuma Y, Ricordi C, Miki A, Yamamoto T, Pileggi A, Khan A, Alejandro R, Inverardi L, Ichii 
H. 2008. Factors That Affect Human Islet Isolation, Transplantation Proc 40:343-345. 
Samejima T, Yamaguchi K, Iwata H, Morkawa N, Ikada Y. 1998. Gelatin density gradient for 
isolation of islets of Langerhans. Cell Transplant 7(1):37-45. 
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 
glucocorticoid-free immunosuppressive regimen. N Engl J Meg 343(4) 289-290. 
 21
Soon-Shiong P, Heintz R, Terasaki P. 1989. Identification of novel blood groupreactive 
monoclonal antibodies cytotoxic to human acinar cells but not islets. Transplantation 
Proceedings 21(1):2622-2623. 
Soon-Shiong P, Fujioka T, Terasaki P, Heintz R, Lanza RP. 1990. Islet purification by a novel 
immunomicrosphere cell depletion technique. Transplant Proc 22(2):780-781. 
Sulaiman AN, James Shapiro AM. 2006. Advances in pancreatic islet transplantation in humans, 
Diabetes, Obesity and Metabolism 8:15-25. 
Sun L, Zborowski M, Moore L, Chalmers JJ. 1998. Continuous, Flow-Through 
Immunomagnetic Cell Separation in a Quadrupole Field. Cytometry 33:469-475. 
Tong X, Xiong Y, Zborowski M, Farag SS, Chalmers JJ. 2007. A Novel High Throughput 
Immunomagnetic Cell Sorting System for Potential Clinical Scale Depletion of T Cells 
for Allogeneic Stem Cell Transplantation. Exp Hematol 35(10)1613-1622. 
Williams PS, Zborowski M, Chalmers JJ. 1999. Flow Rate optimization for the Quadrupole 
Magnetic Cell Sorter. Anal Chem 71:3799-3807. 
Winoto-Morbach S, Ulrichs K, Leyhausen G, Muller-Ruchholtz W. 1989. New principle for 
large-scale preparation of purified human pancreas islets. Diabetes 38(Supp 1):146-149. 
Winoto-Morbach S, Leyhausen G, Schunke M, Ulrichs K, Muller-Ruchholtz W. 1989a. 
Magnetic microspheres (MMS) coupled to selective lectins: a new tool for large-scale 
extraction and purification of human pancreatic islets. Transplant Proc 21(1):2628-2630. 
Winoto-Morbach S, Krout OS, Heiser A, Ulrichs K, Muller-Ruchholtz W. 1994. Lectin binding 
to acinar tissue for complete magnetophoretic purification of porcine pancreatic islets 
depends on the composition and pH of the incubation medium. Transplant Proc 
26(2):646-648. 
 22
Winoto-Morbach, Ulrichs K, Hering BJ, Leyhausen G, Muller-Ruchholtz W. 1989b. Methods in 
Islet transplantation research, Hormone and Metabolic Research, Supplement 25:51-54. 
Zhang H, Nakamura M, Comella K, Moore L, Zborowski M, Chalmers J.  2002. 
Characterization/Quantification of the Factors Involved in the Imparting a 
Magnetophoretic Mobility on Cells and Particles.  European Cells and Materials 3(2):34-
36. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 23
2. Computational Fluid Dynamics Simulation of a Quadrupole Magnetic 
Sorter Flow Channel: Effect of Splitter Position on Nonspecific Crossover 
2.1  Abstract 
 In the Quadrupole Magnetic Sorter (QMS) magnetic particles enter a vertical flow 
annulus and are separated from non-magnetic particles by radial deflection into an outer annulus 
where the purified magnetic particles are collected via a flow splitter.  The purity of magnetically 
isolated particles in QMS is affected by the migration of nonmagnetic particles across transport 
lamina in the annular flow channel.  Computational Fluid Dynamics (CFD) simulations were 
used to predict the flow patterns, pressure drop and nonspecific crossover in a newly designed 
QMS flow channel for the isolation of pancreatic islets of Langerhans.  Simulation results were 
compared with the theoretically and experimentally determined results to validate the CFD 
model.  CFD simulations were employed to compare performance of two models of QMS flow 
channels with differing splitter positions.  Results of the simulations were used to show that one 
design gives up to 10% less nonspecific crossover than another and this model can be used to 
optimize the flow channel design to achieve maximum purity of magnetic particles.  
 
Keywords:  Quadrupole Magnetic Sorter, Computational Fluid Dynamics, Annulus Flow, 
Nonspecific Crossover  
 24
2.2  Introduction 
Insulin independence can be achieved in type 1 diabetes patients by transplanting purified 
pancreatic islets of Langerhans (Sakuma et al., 2008).  Isolation of the islets is the important step 
in preparing the islets for the transplantation.  To achieve high yields with required purities, 
various islet purification techniques have been developed for isolation of islets from exocrine 
pancreas tissue either by targeting islets or by targeting exocrine tissue (Soon-Shiong et al., 
1990; Winoto-Morbach et al., 1989), e.g. hand picking of islets, fluorescence-activated sorting 
by staining islets with neutral red (Gray et al., 1989; Jindal et al., 1994), or destruction of non-
islet tissue by laser energy (Brunicardi et al., 1994).  All of these techniques were successfully 
employed to purify rat islets but failed to scale up to the human tissue digest volume.  Present 
human islet isolation processes rely on the density gradient centrifugation method which depends 
on small density difference between islets and acinar tissue.  The density gradient centrifugation 
process, however, has been identified as a potential source of islet mass loss and islet damage 
due to mechanical stress associated with centrifugation and prolonged exposure to proteolytic 
enzymes (Samejima et al., 1998; Pinkse et al., 2004).  
Quadrupole magnetic sorting (QMS) is a well-designed and tested technology for 
magnetic separations of single cells, e.g., stem cells (Moore et al., 2001; Nakamura et al., 2001; 
Lara et al., 2002) and has been successfully applied to several cell separations and purifications 
(Sun et al., 1998; Tong et al., 2007; Jing et al., 2007).  When QMS, developed for single cell 
separation, was used for isolation of islets, very low purity was achieved with good yields.  Low 
purity is due to the size difference between single cells and islets (Shenkman et al., 2009).  Islet 
diameters are distributed between 50 to 500 microns whereas single cell size is around 1 to 10 
microns.  It has been observed that the mechanical stress applied in QMS does not affect islet 
 25
functionality after isolation (Shenkman et al., 2009).  To overcome limitations associated with 
the commonly used density gradient method and to develop a technique that yields sufficient 
number of islets to reduce the donor to recipient ratio to 1:1, we redesigned (Kennedy et al., 
2007) and tested high capacity QMS to isolate porcine islets of Langerhans.  
In QMS isolation procedures particles of interest labeled with magnetic micro particles 
are isolated from unlabeled particles in a magnetic field while flowing through an annular flow 
channel.  Labeled particles migrate radially outward in a quadrupole magnetic field and leave the 
channel on its outer periphery as a purified fraction.  The presence of any unlabeled particles in 
isolated magnetically labeled particles is known as nonspecific crossover.  The occurrence of 
nonspecific crossover is of particular importance in the isolation of islets as the islets constitute 
only one to two percent of the entire pancreas.  Thus, even one percent of nonspecific crossover 
could reduce the purity of islets by 50 percent.  There are many factors that contribute to 
nonspecific crossover in QMS, such as hydrodynamic forces, particle concentration in the 
sample, total flow rate, flow ratios, splitter imperfections (Williams et al., 2003; Williams et al., 
2003a; Williams et al., 2008) and splitter position.  The effects of splitter position, sample 
concentration, total flow rate and flow ratios were tested in this study.  Computational fluid 
dynamics software was used to predict the fluid flow pattern throughout the flow channel and 
nonspecific crossover. 
2.3  Materials and Methods 
2.3.1  QMS and Separation Theory 
The QMS technology is based on a process known as split-flow lateral transport thin 
separation, which is a subset of Field Flow Fractionation (FFF) technology that separates 
immunomagnetically labeled islets based on their magnetophoretic mobility (m).     
 26
Magnetophoretic mobility is defined as the velocity of a particle per unit magnetic energy 
gradient.  A QMS system consists of three essential components: flow channel, magnet and 
pumps.  Separation takes place in the flow channel which consists of a cylindrical core that is 
concentric with an external cylindrical shell (Figure 2.1).  The sample, consisting of magnetically 
labeled islets along with unlabelled tissue particles, enters at a? close to the core through a 
specially designed inlet: carrier fluid, enters the channel at b? adjacent to the outer wall.  Both 
flows merge at the end of the inlet splitter and continue to flow at each side of a virtual surface, 
known as inlet splitting surface (ISS) at distance riss from the center of the core.  
ISS can be seen in the Figure 2.1 close to the core and remaining at the same distance 
along the fully developed laminar flow in the annulus region of the channel.  The outer flow rate 
Qb? is maintained higher than the inner flow rate Qa? to keep the sample flow near the core.  A 
second virtual surface known as the outer splitting surface (OSS) at a distance ross from the 
center of the core separates the two outlets which are separated by the outlet splitter.  The 
distance between ISS and OSS, ross-riss, is known as the Transport Lamina. Magnetically labeled 
particles in the sample migrate radially and cross the transport lamina and leave in the positive 
fraction b.  The unlabeled and weakly labeled particles which cannot cross the OSS will exit in 
the negative fraction a.   
For an ideal separation in QMS with laminar flow, the transport lamina thickness should 
be zero as the unlabeled particles would not be expected to cross the transport lamina.  However, 
magnetically labeled islets are generally not uniform in magnetization.  A range of magnetization 
is exhibited due to variation in the number of magnetic particles that enter the vascular structure 
of islets.  To completely isolate pure labeled particles, it is necessary to create conditions that 
allow the least mobile particles to migrate across the thickness of the transport lamina.  The  
 27
 
Figure 2.1 : Schematic diagram of the quadrupole magnetic cell sorter (QMS). 
 
 
 
Qb?
Qa?
Qa
Qb
MA
GN
ET
MA
GN
ET
Tr
an
sp
ort
 La
mi
na
ri
ro
riss
ross
Non-Magnetic 
Particle
Magnetic 
Particle
MA
GN
ET
MA
GN
ET
Tr
an
sp
ort
 La
mi
na
 28
thickness of the transport lamina changes with changes in inlet and outlet flow rate ratios and 
change in splitter position.  
The particle trajectory in the annulus is described by the integral equation;  
?? +=
k
k m
s
z
dkku vkvdz
1
)(
)(
0
         (1) 
where )(kv  is the velocity profile of the fluid in the annulus, sv is the Stokes sedimentation 
velocity of the particle in the fluid, um(k) is the radial magnetophoretic velocity of the particle 
and k  is defined as 
 
0r
rk =                   (2) 
Here  is defined as the radial coordinate from the center of the flow channel core and 0r is the 
radius from the center to the wall. Upon the integration of eq. (1) and simplification we can get 
1
0
1
0
2
1
11
00
)ln(
)1(
),(.
2 ASm
k
kvr
kA
kkI
mSr
Qz
mm
s
imm
+?= pi                   (3) 
where mm is called the magnetophoretic mobility of the particle, 0mS  is the magnetic force at the 
inner surface of the outside wall of the flow channel and sv is sedimentation velocity of the 
particle.  1A , 2A and ),( 11 kkI  are functions used to simplify the calculations and are given by 
)1( 221 AkA i ?+=          (4) 
i
i
k
kA
1ln
1 2
2
?=           (5) 
k
kkAkkkkI 1])(ln22ln4[),(
2
2
2
11 +?=      (6) 
 29
The radial position of the ISS, KISS is calculated using the following integration factor (Kennedy 
et al., 2007): 
ISS
i
k
kISSi kAkkAkkkkI ]ln22[),(
2
2
2
2
42
2 ?+?=     (7) 
Similarly, the position of the OSS can be calculated using eq. (7) by replacing kISS with kOSS. 
2.3.2  Experimental Measurements 
The flow channel used for the experiments was called ?prototype I?.  Nonspecific 
crossover experiments were performed with nonmagnetic particles, Cultisphers?, macro-porous 
gelatin-coated micro-carrier beads with diameter range of 200 to 380 ?m, similar in size and 
density to islets and fragments of exocrine tissue in pancreas digests.  The sample consisting of 
cultisphers was introduced into the a? inlet flow stream and carrier buffer was introduced into the 
b? inlet flow stream by dual head Watson-Marlow peristaltic pumps, and the positive fraction 
outlet was controlled by a peristaltic pump while leaving the negative fraction outlet to exit at 
atmospheric pressure to maintain equilibrium of flow in the flow channel.  
 Experiments were carried out at total flow rates of 250, 300 and 400 ml/min with inlet 
flow ratio (Qa?/Q) of 0.25 and at different outlet flow ratios (Qa/Q) 0.25, 0.3, 0.5 and 0.7.  
Experiments were conducted at different sample concentrations to study the effect of particle 
concentration in the sample on nonspecific crossover.  Turbidity sensors are connected to 
positive and negative fraction outlets to detect absorbance.  Nonspecific crossover was calculated 
from the areas of the absorbance peaks. 
2.3.3  Computational Fluid Dynamics  
The commercial CFD code, FLUENT, was used for the simulations, and FLUENT?s 
preprocessor, GAMBIT, was used to generate flow-field meshes.  An unstructured mesh 
consisting of tetrahedral volumetric elements in 3D is used in the entire domain.  Velocity inlet 
 30
boundary conditions with velocities normal to the plane of the channel inlets with fluid 
properties corresponding to water at 298K are set at the inlets, outflow boundary conditions 
corresponding to fully developed flow were used at the outlets and the no-slip condition is 
applied to all walls. 
The geometry used for the first case is based on the quadrupole magnetic separation 
channel for islets, ?prototype I?.  The length of this flow channel is 254 mm with inner diameter 
of the outer wall of 6.03 mm and core diameter is 5.08 mm.  Splitters with radius 5.588 mm and 
thickness 0.14 mm are used at inlet and outlet.  A specially designed flow distributor with 
diameter 5.96 mm is placed in the b inlet.  For the second case (prototype II), the channel length 
and outer wall diameter are the same as for Prototype I, but the core diameter is 5.08 mm and 
inlet splitter diameter 5.58 mm with thickness 0.14 mm and outlet splitter diameter 5.84 mm with 
thickness 0.14 mm.  The two QMS separation flow channels analyzed in the study are shown in 
Figure 2.2.  
2.3.4  Governing Equations 
For flow in the QMS flow channel, FLUENT, a commercial CFD package, is utilized to 
solve conservation equations for mass and momentum.  FLUENT is based on a set of partial 
differential equations maintaining conservation of mass and momentum.  These equations 
describe the convective motion of the fluid.  The law of conservation of mass applied to an 
infinitesimal fixed control volume generates the continuity equation in differential conservation 
form: 
 31
a? b?
a b
a? b?
a b
Prototype I Prototype II
INLET 
SPLITTER
CORE
FLOW 
ANNULUS
OUTLET 
SPLITTER
 
Figure 2.2: Geometry drawings of Prototype I and Prototype II flow channels.   
 
 
 
 
 
 
 
 
 32
0. =?+??
?
Vt ??         (10) 
The first term in the Equation 10 represents the rate of increase in the density and the 
second term represents the rate of mass flux passing out of the control volume.  For a steady 
incompressible flow, ? = constant and 0=?? t? which reduces Equation 10 to 
00. =??+??+??=?
?
z
u
y
u
x
uorV        (11) 
Newton's second law applied to a fluid passing through an infinitesimal control volume 
yields the momentum equation: 
ijVVVt pi?? .).()( ?=?+?
? ???        (12) 
The first term in the Equation 12 represents the rate of increase of momentum per unit volume in 
the control volume.  The second term represents the rate of momentum lost by convection 
through control surface.  The forces applied by external stresses such as normal and shearing 
stresses which can be represented by the stress tensor ?ij.  The stress tensor can be written as 
k
k
ij
i
j
j
i
ijij x
u
x
u
x
up
?
??
?
?+
?
?+?= ???pi
3
2)[(      (13) 
Substituting equation 13 into equation 12 and simplifying produces the Navier - Stokes 
equations: 
???
?
???
?
?
??
?
?+
?
?
?
?+??=
?
k
k
ij
i
j
j
i
j
ij x
u
x
u
x
u
xvpDt
VD ????
3
2)(     (14) 
For a Cartesian coordinate system, Equation 14 can be separated into three scalar equations.  The 
conservation of momentum in the x-direction is given by 
 33
zyxx
p
Dt
Du zxyxxx
?
?+
?
?+
?
?+
?
??= ????       (15) 
The conservation of momentum in the y-direction is given by 
zyxy
p
Dt
Dv zyyyxy
?
?+
?
?+
?
?+
?
??= ????       (16) 
The conservation of momentum in the z-direction is given by 
zyxz
p
Dt
Dw zzyzxz
?
?+
?
?+
?
?+
?
??= ????       (17) 
The viscous normal stresses are given by 
???
?
???
?
?
??
?
??
?
?=
z
w
y
v
x
u
xx 23
2 ??        (18) 
???
?
???
?
?
??
?
??
?
?=
z
w
x
u
y
v
yy 23
2 ??        (19) 
???
?
???
?
?
??
?
??
?
?=
x
u
y
v
z
w
zz 23
2 ??        (20) 
and the viscous shear stresses are given by 
???
?
???
?
?
?+
?
?==
y
u
x
v
yxxy ???        (21) 
?????? ??+??== zuxwzxxz ???        (22) 
???
?
???
?
?
?+
?
?==
z
v
y
w
zyyz ???        (23) 
The first law of thermodynamics applied to a fluid passing through an infinitesimal, fixed 
control volume yields the following energy equation: 
 
 34
( ) ( ) ( ) ( )wpzvpyupxTVkhVht ????????????????=???????+?? ???? .. ??  
( ) ( ) ( )zxyxxx uzuyux ??? ??+??+??+  
( ) ( ) ( )zyyyxy uzuyux ??? ??+??+??+  
( ) ( ) ( )zzyzxz uzuyux ??? ??+??+??+                                      (24) 
For most problems in gas dynamics, it is possible to assume an ideal gas, defined as a gas 
whose intermolecular forces are negligible.  An ideal gas obeys the ideal gas equation of state: 
RTp ?=          (25) 
In this study, the ideal gas equation of state is used to calculate the density for the 
compressible gas flow with change in pressure. 
The continuity, x-momentum, y-momentum, z-momentum and energy equations are 
solved using FLUENT to simulate incompressible, laminar, steady-state flow. 
2.3.5  Discretization 
The process of converting partial differential equations into a set of algebraic relations 
that can be solved using a computer is called discretization (Anderson, 1999).  Discretization 
involves two main steps:  (1) converting the continuum partial differential equations into 
algebraic relations and (2) converting the continuous physical domain into nodes, volumes, or 
elements where the algebraic equations will be solved.  Discretization of partial differential 
equations can be accomplished in many different ways.  Among those, the two types of 
discretization used to solve problems in fluid mechanics and heat transfer are finite differences 
(discretization of partial differential form of equations) and finite volumes (discretization of 
equations in the integral form). 
 35
In the finite difference approach, the continuous problem or domain is discretized so that 
the dependent variables are considered to exist only at discrete points.  In the finite volume 
method, the conservation principles are applied to a fixed region or space known as the control 
volume.  Using the finite volume approach, either control volumes are established first and grid 
points are placed at the center of the volumes (cell-centered method) or grid points are 
established first and then the boundaries of the control volume are fixed (vertex-centered 
method) (Tannehill et al., 2004).  The finite volume method using the cell-centered approach is 
used in this study to convert the governing equations into an algebraic form that is then solved 
numerically. 
A variety of methods are used for discretizing the governing equations using the finite 
volume method.  In this study, first-order and second-order accurate upwind methods are used.  
The cell-centered method is used, solving the flow properties at the cell centers; the values at the 
interfaces (used to obtain cell fluxes) are obtained by interpolating the cell-centered values with 
an upstream direction bias. This method is called an upwinding scheme. 
2.3.6  Solvers 
FLUENT was used in this study to simulate flows through the QMS flow channel.  This 
section discusses the FLUENT solver options that are used in this study.  Governing equations 
for conservation of mass, momentum, energy, and chemical species are solved using cell-
centered control volume based segregated solver. 
 Using a segregated solver, the governing equations are solved sequentially (i.e., 
segregated from one another).  Several iterations of the solution loop are performed before 
obtaining a converged solution.  The segregated solver basic steps, shown in Figure 2.3, are: 
 36
1. Initial guess of pressure, velocity and species transport quantities, such as mass flux, is 
made. 
2. The discretized Navier-Stokes equations (i. e., momentum equations) are solved using the 
guessed values of P*, u*, v*, etc.  In this step coefficient to determine the fluxes through 
the cell/boundary faces by conduction/convection etc.  The flow variables (p, u, v, w, 
etc.) for the entire domain (for each grid node) are obtained at the end of this step. 
3. Once the flow properties are calculated, the continuity equation is next verified.  If the 
continuity equation is not satisfied, pressure and velocity values (u, v, w) are corrected 
using pressure correction equation.  SIMPLE algorithm is used for pressure-velocity 
coupling. 
4. Once the flow properties are obtained, other discretized transport equations are solved.  
The initial guess for the transport quantities are depicted in step 1.  After solving the 
transport equation, we get the transport properties at each node 
5. In this step the flow and transport variables, obtained in step 3 and 4, at each node are 
compared with values of previous iteration and residuals are calculated using L2 Norm.  
If the residuals are less than prescribed tolerance limit, the solution is final.  If the 
residual is higher than the tolerance limit, steps 1 through 5 are repeated until the 
convergence criteria are satisfied.  The initial guess values are replaced with values of 
flow and transport variables obtained from the current iteration.  
2.3.7  Pressure-Velocity Coupling Method 
A pressure-velocity coupling is used to derive an equation for pressure from the discrete 
continuity equation.  The SIMPLE algorithm is used to obtain a relationship between velocity 
and pressure corrections to enforce mass conservation and thus obtain the pressure field. 
 37
2.3.8  Geometric Modeling 
GAMBIT, a commercial pre-processor, is used to generate geometric models used in this 
study.  FLUENT is used to solve the discretized governing equations.  For post-processing the 
results obtained from FLUENT, TECPLOT are used. GAMBIT is used to create a volume mesh 
in the fluid domain.  An unstructured mesh consisting of tetrahedral volumetric elements in three 
dimensions is used in the entire domain except at the walls, where a boundary layer mesh 
consisting of prismatic volumetric elements is used. 
Figure 2.4 shows a sample mesh consisting of triangular elements on the splitter and 
sample inlet.  Figure 2.5 shows the unstructured mesh on the top inlet face along with the 
prismatic boundary layer mesh on the inlet pipe faces.  The effect of using a boundary layer 
mesh around the fibers and particles is also examined by considering a case with an unstructured 
mesh in the entire domain.  Pressure drops on the structured and unstructured mesh around the 
fibers and the particles varied by up to two to eight per cent; moreover, the unstructured mesh 
results had significantly larger numerical error as estimated by a grid refinement study.  The 
accuracy of the fully tetrahedral mesh results could be increased by refining the mesh, but the 
additional refinement is expensive in both computation cost and solution time.  As a result, a 
prismatic boundary layer mesh is used around the fibers and the particles to reduce the 
computation cost and to improve numerical accuracy. 
 
 38
 
Figure 2.3:  Flow chart for segregated solver used in this study. 
 
 39
 
Figure 2.4:  Unstructured triangular mesh on the faces of QMS flow channel. 
 40
 
 
Figure 2.5:  Unstructured mesh along with the boundary layer on the inlet pipes 
        and channel face. 
 
 
 
 
 
 
 
 
 41
2.3.9  Boundary Conditions 
 Three different boundary conditions used in this study are no-slip, velocity inlet, and 
outflow.  For laminar flows with zero velocity at the wall, the no-slip boundary condition is 
enforced at the walls.  Velocity inlet boundary conditions are used to define the flow velocity 
along with all flow properties at the inlet and in some cases at outlet.  Outflow boundary 
conditions are used to model flow exits where the details of the flow velocity and pressure are 
not known prior to solution of the flow problem.  At outflow boundaries, all of the necessary 
boundary information is extrapolated from the interior. 
2.3.10  Simulation Procedure 
An iterative solution approach is employed where the solution is advanced in pseudo-
time until the steady state equations are satisfied to a specified tolerance.  At each iteration the 
governing equations are solved using a segregated solver.  A finite volume technique is used to 
solve the discrete form of the governing equation on the computational grid.  The standard 
SIMPLE algorithm is used in which the momentum equations are first solved for the velocity 
components.  A pressure correction is then determined that drives the velocity field towards 
satisfying the mass conservation equation.  Iterative convergence is assessed by monitoring the 
L2 norms of the steady-state residuals.  The convergence criterion is set to 10-7 for all 
conservation equations. 
2.4  Results and Discussion 
2.4.1  CFD Simulations: Flow Analysis and Pressure Drop Predictions 
 One important feature of the flow channel is the design of the flow distributor placed at 
the carrier buffer inlet (b?) to obtain circumferentially uniform flow in the flow channel.  
Uniform flow around the core is important to maintain laminar flow in the annulus to yield 
 42
minimum crossover and maximum isolation. Improvements made to the flow distributor design 
were verified by CFD analysis.  Figure 2.6 shows the flow distributor design with flow profile at 
different distances downstream from the distributor.  Incoming fluid enters the distributor from 
two branches. Direct flow through the distributor from these branches is prevented by the lack of 
notches at the two inflow locations.  The simulations show the uniform flow developed by the 
time fluid reaches the splitter end point.   
CFD simulations are performed using the generated flow-field meshes in order to predict 
the pressure drop and analyze the details of the flow in the flow channel.  A plane is extracted at 
the center of the domain (i.e., Z=3.01625 mm) and displayed as a colored ring for the case with 
total flow rate 400 ml/min with inlet and outlet flow ratio 0.25.  The velocity magnitude contours 
are shown in Figure 2.7a for sample (a?) velocity 0.05 m/s and buffer (b?) velocity 0.15 m/s.  
Careful examination of the calculated fluid velocity profiles revealed that the flow was fully 
developed within a relatively short distance of the splitter edge for every ratio of inlet and outlet 
flows.   
Figure 2.7b shows the velocity profile at the middle of the channel.  Pressure magnitude 
contours are shown in Figure 2.8 along with the pressure gradient found from CFD simulations 
plotted against total flow rates compared with the theoretically calculated pressure drop using the 
Haigen-Poiseuille equation for annular pipes.  Figure 2.8b shows that the pressure drop predicted 
using CFD is in good agreement with theoretical calculations with pressure drop increasing with 
the increase in total flow rate. 
 43
       
Figure 2.6: Velocity (m/s) contours in x plane at different distance from flow distributor to the end point of the splitter.  
(m/s)
 44
2.4.2  Nonspecific Crossover 
The new design of the flow channel is tested for nonspecific crossover at different total 
flow rates ranging from 250 ml/min to 400 ml/min, input flow ratios (Qa?/Q), outlet flow ratios 
(Qa/Q) ranging from 0.25 to 0.7.  When flow rates less than 250ml/min were used, cultisphers 
settled in the tubing and clogging of the inlet was observed.  The effect of the particle 
concentrations in the sample and total flow rate on the crossover was examined.  Some of these 
experiments were also conducted with pig pancreas (results not presented) and the same results 
as with cultisphers were observed. CFD simulations were conducted for all of these cases using 
the discrete phase model and compared with the experimental results.  Nonspecific crossover 
(Sb) is calculated as the ratio of non-magnetic particles leaving in positive collection (Nb) to the 
particles leaving in both negative (Na) and positive collection. 
ba
b
b NN
NS
+=        (26) 
Figure 2.9 shows the nonspecific crossover of Cultisphers at different outlet flow ratios 
and at fixed inlet flow ratio of 0.25 and total flow rate of 400ml/min.  These experiments were 
conducted with the flow channel design prototype I, in which the inlet and outlet splitters have 
the same 5.58mm diameter.  A clear decrease in nonspecific crossover with increase in outlet 
flow ratio is observed because the increase in outlet flow ratio increases the transport lamina 
thickness.  Figure 2.9 also presents the comparison of nonspecific crossover obtained from 
experiments with prototype I and CFD simulation-predicted crossover for prototype II.  Good 
 45
Figure 2.7: a) Velocity (m/s) magnitude contours in the flow channel at z = 1.1875? (flow is in positive x direction. b) Velocity profile 
of the fluid at the middle point of the channel. 
 46
Figure 2.8: a) Pressure (Pa) contours in the flow channel at z = 1.1875? b) Comparison of the pressure drop from CFD simulations and 
empirical correlations.
 47
agreement is found between experimental results and CFD simulation-predicted values with 
same prototype flow channel with less than 10% deviation.  This deviation might be due to the 
negligence of particle-particle interaction in CFD modeling.  CFD-predicted crossover for 
prototype II is compared with that of CFD predictions and experimental values with Prototype I.  
Crossover is slightly less for prototype II when compared to prototype I. The diameter of the 
outlet splitter is increased in prototype II which increases the transport lamina thickness and 
reduces the number of nonmagnetic particle that leave in the positive fraction.  Straight line in 
the figure 2.9 represents the ideal flow of the fluid without solid particles.   
Figure 2.10 shows the crossover of nonmagnetic particles at different total flow rates in 
the Prototype I channel at a fixed inlet and outlet flow ratio of 0.25.  Experimental crossover 
values increase with decreasing total flow rates as the hydrodynamic lift forces move the 
particles away from wall at low flow rates which helps particles to cross the transport lamina and 
leave in the positive fraction.  Crossover obtained from CFD simulations for different total flow 
rates are under predicted by 5 to 10 percent for prototype I.  Comparison of calculated crossover 
values at different total flow rates for the two prototypes shows a decrease in crossover for 
prototype II due to the increase in transport lamina thickness resulting from the outlet splitter 
position.        
Figure 2.11 shows crossover values at different sample concentrations at a fixed total 
flow rate of 400ml/min.  Outlet flow ratio is maintained equal to inlet flow ratio at 0.25. 
Increasing cultisphers concentration increases the particle-particle interaction which results in 
more nonspecific crossover.  Higher concentrations of particles during experiments also clogged 
the flow paths in the sample inlets.  CFD simulations predict decreased crossover with the new 
design prototype II when compared with the prototype I.  
 48
            
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
0.0
0.1
0.2
0.3
0.4
0.5
0.6
S b
Qa/Q
 Experimental-Prototype I
 CFD - Prototype-I
 CFD - Prototype-II
 Ideal 
 
Figure 2.9: Comparison of calculated and observed nonspecific crossover as a function of outlet 
flow ratio at a total flow rate of 400 ml/min and inlet flow ratio of 0.25. 
 
 
 
 
 49
            
300 320 340 360 380 400
0.06
0.07
0.08
0.09
0.10
0.11
0.12
0.13
0.14
0.15
S b
Q, ml/min
 Experimental - Prototype I
 CFD - Prototype I
 CFD - Prototpe II
 
Figure 2.10: Comparison of the nonspecific crossover as a function of total flow rate at an inlet 
and outlet flow ratio of 0.25.              
 
 
 
 
 
 50
             
0.10 0.15 0.20 0.25 0.30
0.05
0.06
0.07
0.08
0.09
0.10
0.11
S b
Cultispheres Concentration, g/L
 Experimental - Prototype I
 CFD - Prototype I
 CFD - Prototype II
 
Figure 2.11: Nonspecific crossover as a function of particle concentration in the sample at a total 
flow rate of 400ml/min, inlet and outlet flow ratio of 0.25.            
 
 
 
 
 
 
 
 
 
 
 
 
   
 51
2.5  Conclusions 
 CFD simulations of flow pattern performed on the new design of QMS flow channel 
confirm circumferentially uniform flow development around the annular channel.  Quantitative 
agreement between experimental measurements of nonspecific crossover and prediction based on 
CFD modeling of the fluid flow was shown.  For all flow conditions, crossover predicted by 
CFD simulations was found to be slightly lower than experimentally observed results.  This 
difference may be due to the contribution of crossover from other factors such as particle lift and 
particle interactions. Diffusion was not considered in the CFD modeling.  The good agreement 
between experimental and CFD predicted results allowed the performance of simulations with 
different channel prototype models to develop a design to minimize nonspecific crossover.  
Though increasing the diameter of the outlet splitter decreases nonspecific crossover the 
diameter needs to be optimized based on the yield of magnetically labeled particles with change 
in splitter diameter.  Therefore future CFD simulations must include the migration of the 
magnetic particles. 
 
 
 
2.6  Acknowledgements 
This research was funded in part by the U.S. Department of Health and Human Services under 
SBIR grant 5R44DK072647-03 from the National Institute and Digestive and Kidney Research 
(NIDDK) awarded to Techshot, Inc., Greenville, Indiana, USA.  
 
 
 52
2.7  References 
Anderson JD. Computational Fluid Dynamics: The Basics with Applications. McGraw Hill, Inc., 
1995. ISBN 0-07-001685-2. 
Brunicardi, FC, Oh Y, Shevlin L, Suh E, Kleinman R, Stein E, Lipaz G, Plant DV, Imagawa D, 
Fetterman HR. 1994. Laser destruction of human nonislet pancreatic tissue. Transplant 
Proc 26(6):3354-3355. 
Gray DWR, Gohde W, Carter N, Heiden T, Morris PJ. 1989. Separation of Pancreatic islets by 
Fluorescence-Activated Sorting. Diabetes 38(suppl 1):33-135. 
Jindal RM, McShane P, Gray DWR, Morris PJ. 1994. Isolation and Purification of Pancreatic 
Islets by Fluorescence Activates Islet Sorter. Transplant Proc 26(2):653. 
Jing Y, Chalmers JJ, Zborowski M. 2007. Blood progenitor cell separation from clinical 
leukapheresis product by magnetic nanoparticle binding and magnetophoresis. 
Biotechnol Bioeng 96(6):1139. 
Kennedy DJ, Todd P, Logan S, Becker M, Papas KK, Moore LR. 2007. Engineering quadrupole 
magnetic flow sorting for the isolation of pancreatic islets. Journal of Magnetism and 
Magnetic Materials 311:388-395. 
Lara O, Nakamura M, Zborowski M and Chalmers JJ. 2002. Negative depletion cell sorting 
using a quadrupole magnetic cell sorter. Eur. Cells Mater. 3:62-64. 
Moore LR, Rodriguez AR, Williams PS, McCloskey K, Bolwell BJ, Nakamura M, Chalmers JJ 
and Zborowski M. 2001. Progenitor cell isolation with a high capacity quadrupole 
magnetic flow sorter. J. Magn. Magn. Mater. 225:277-284. 
 53
Nakamura M, Decker K, Chosy J, Comella K, Melnik K, Moore LR, Lasky LC, Zborowski M 
and Chalmers JJ. 2001. Separation of Breast Cancel Cell Line from Human Blood Using 
a Quadrupole Magnetic Flow Sorter. Biotechnol. Prog. 17:1145-1155. 
Pinkse G, Steenvoorde E, Hogendoorn S, Noteborn M, Terpstra OT, Bruijn JA, and De Heer E. 
2004. Stable transplantation results of magnetically retracted islets: a novel method. 
Diabetologia 47:55. 
Sakuma Y, Ricordi C, Miki A, Yamamoto T, Pileggi A, Khan A, Alejandro R, Inverardi L, Ichii 
H. 2008. Factors That Affect Human Islet Isolation. Transplant Proc 40:343-345. 
Samejima T, Yamaguchi K, Iwata H, Morkawa N, Ikada Y. 1998. Gelatin density gradient for 
isolation of islets of Langerhans. Cell Transplant 7(1):37- 45. 
Shenkman RM, Chalmers JJ, Hering BJ, Kirchhof N and Papas KK. 2009. Quadrupole Magnetic 
Sorting of Porcine Islets of langerhans, Tissue Engineering: Part C 15(2):147-156. 
Shenkman RM, Godoy-Silva R, Papas KK and Chalmers JJ. 2009. Effects of Energy Dissipation 
Rate on Islets of Langerhans: Implications for isolation and Transplantation, 
Biotechnology and Bioengineering 103(2):413-423. 
Soon-Shiong P, Fujioka T, Terasaki P, Heintz R, Lanza RP. 1990. Islet purification by a novel 
immunomicrosphere cell depletion technique. Transplant Proc. 22(2):780-781. 
Sun L, Zborowski M, Moore LR, Chalmers JJ. 1998. Continuous, Flow-Through 
Immunomagnetic Cell Separation in a Quadrupole Field. Cytometry 33:469. 
Tannehill JC, Anderson DA, Pletcher RH. 2004. Computational Fluid Mechanics and Heat 
Transfer. Taylor and Francis, second edition, ISBN 1-56032-046-X. 
 54
Tong X, Xiong Y, Zborowski M, Farag SS and Chalmers JJ. 2007. A Novel High Throughput 
Immunomagnetic Cell Sorting System for Potential Clinical Scale Depletion of T Cells 
for Allogeneic Stem Cell Transplantation. Exp Hematol 35(10):1613. 
Williams PS, Moore LR, Chalmers JJ and Zborowski M. 2003. Splitter Imperfections in Annular 
Split-Flow Thin Separation Channels: Effect on Nonspecific Crossover. Anal. Chem. 
75:1365-1373. 
Williams PS, Decker K, Nakamura M, Chalmers JJ, Moore LR and Zborowski M. 2003a. 
Splitter Imperfections in Annular Split-Flow Thin Separation Channels: Experimental 
Study of Nonspecific Crossover. Anal. Chem. 75: 6687-6695. 
Williams PS, Hoyos M, Kurowski P, Salhi D, Moore LR and Zborowski M. 2008. 
Characterization of Nonspecific Crossover in Split-Flow Thin Channel fractionation. 
Anal. Chem. 80:7105-7115. 
Winoto-Morbach S, Ulrichs K, Leyhausen G and Muller-Ruchholtz W. 1989. New principle for 
large-scale preparation of purified human pancreas islets. Diabetes. 38(Supp 1):146-149. 
  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 55
3. Application of Magnetic Particle Tracking Velocimetry to Quadrupole 
Magnetic Sorting of Porcine Pancreatic Islets 
3.1  Abstract 
 Magnetic isolation is a promising method for separating and concentrating pancreatic 
islets of Langerhans for transplantation in Type 1 Diabetes patients.  We are developing a 
continuous magnetic islet sorter to overcome the restrictions of current purification methods that 
result in limited yield, viability and purity of the isolated islets.  In Quadrupole Magnetic Sorting 
(QMS) islets are magnetized by infusing superparamagnetic microbeads into islets? vasculature 
via arteries that serve the pancreas.  The performance of the islet sorter depends on the resulting 
speed of the islets in an applied magnetic field, a property known as magnetophoretic mobility.  
Essential to the design and successful operation of the QMS is a method to measure the 
magnetophoretic mobilities of magnetically infused islets. We have adapted a magnetic particle 
tracking velocimeter (MPTV) to measure the magnetophoretic mobility of particles up to 
1000?m in diameter.  Velocity measurements are performed in a well-characterized uniform 
magnetic energy gradient using video imaging followed by analysis of the video images with a 
computer algorithm that produces a histogram of absolute mobilities.  MPTV was validated 
using surrogate magnetic agarose beads and subjecting them to QMS. Mobility distributions of 
labeled porcine islets indicated that magnetized islets have sufficient mobility to be captured by 
the proposed sorting method, with this result confirmed in test isolations of magnetized islets. 
Keywords: Particle tracking velocimetry, magnetic flow sorter, pancreatic islets isolation, 
magnetic particles. 
 56
3.2  Introduction 
 
 Magnetic isolation of islets is gaining in popularity due to its ease of use, speed and 
selectivity. Current islet isolation procedures depend on centrifugation in density gradients.  
They are limited in their throughput and separation efficiency, and they subject islets to 
potentially detrimental physical stresses (London et al., 1998; Soon-Shiong et al., 1989).  To 
address the problems with throughput and efficiency, a continuous flow quadrupole magnetic 
islet separator was developed (Kennedy et al., 2007).  Essential to the design and operation of 
such a device is a method to measure the magnetophoretic mobilities of magnetically infused 
islets. 
 The magnetophoretic mobility of a particle indicates how responsive the particle is to an 
applied magnetic field.  Paramagnetic particles have positive susceptibility and move toward 
increasing magnetic field intensity.  Islets are rendered paramagnetic by infusing the donor 
pancreas with paramagnetic microspheres such as Dynabeads?. Beads capture efficiency varies 
with the age of the donor and pancreas weight and affects the magnetophoretic mobility of the 
islets. 
Several different techniques have been used to measure the magnetophoretic mobility of 
paramagnetic particles, including the use of visual microscopic observation of the movement of 
the labeled particle with stopwatch and magneto-cytometry (Gill et al., 1960; Davis et al., 1993).  
Magnetic Particle Tracking Velocimetry (MPTV) measures cell surface antigen counts through 
the quantification of antibody binding capacities (ABC) (McCloskey et al., 2001; Melnik et al., 
2001; McCloskey et al., 2000; McCloskey et al., 2001a) and can be used to characterizes 
magnetophoretic mobilities of nano and micro-particles (Haefeli et al., 2002; Zhang et al., 2002; 
Chalmers et al., 1999) and living cells (Chalmers et al., 1999).  Two important features of the 
 57
MPTV apparatus set it apart from other magnetic tracking technologies: 1) the use of an 
isodynamic (i.e. constant-force) magnetic field as indicated by ?Sm = constant?, and 2) capacity 
to track a large number of individual particles, up to 10,000, in a fraction of an hour. Both 
features are essential for high accuracy and precision of the MPTV analysis (Sridhra Redy et al., 
1996; Moore et al., 2004; Zborowski et al., 2003).  Those features are at the basis of the 
competitive advantage of the MPTV over other magnetic motion analyzers described in the 
literature that are based on single-particle analysis in high-gradient magnetic fields which are 
difficult to measure(Watarai et al., 2001; Wanabe et al., 2004; Suwa et al., 2004). 
MPTV leverages current technologies in video microscopy, computer processing speed 
and finite element analysis for magnetic fields to measure the induced motion of cells and 
particles in the highly characterized magnetic and gravity field.  The motion of the particles in 
the known field is then translated into a characteristic parameter called magnetophoretic mobility 
in a magnetic field and sedimentation rate in a gravity field.  MPTV measurements involve 
videotaping the movement of immunomagnetically labeled particles through a known viscosity 
medium and magnetic susceptibility in a well-defined magnetic energy density gradient.  The 
velocity of each particle along with its location within the magnetic energy gradient is recorded.  
From this information, the magnetophoretic mobility of each particle is calculated.  A significant 
advantage of this method over other techniques is that large numbers of cells or particles can be 
processed in a very short time. 
Present MPTV devices developed to measure magnetophoretic mobility of cells have 
problems with flow cells and data analysis software. They also have fixed amount of magnetic 
field strength and are only able to measure mobilities of single cells.  The MPTV described here 
in measures magnetophoretic mobilities and sedimentation rates of cells and large particles like 
 58
islets with reliable software and variable magnetic field strength.  The software developed for 
MPTV also gives the optimum parameters to use with QMS to achieve isolation with maximum 
purity and yields.  This paper explains the features of a refined MPTV system with results 
obtained with large particles including porcine pancreatic islets of Langerhans.  
3.3  Theory 
 The following analysis applies to the use of MPTV applied to the isolation of 
magnetically labeled pancreatic islets of Langerhans by quadrupole magnetic sorting. The 
magnetophoretic mobility (?m) of a particle is defined as the ratio of the velocity of the particle 
in the magnetic field, uf, to the magnetic field energy gradient, Sm.  
m
f
m S
u=?      (1) 
Any population of magnetically labeled particles exhibits a statistical distribution of ?m based on 
several characteristics. Mobility of the pancreatic islets depends on the number of magnetic 
Dynabeads? infused into each islet.  The following equation relates the Dynabeads and islet 
properties to the magnetophoretic mobility of an immunomagnetically labeled islet. 
i
iD
m D
Vn
pi?
??
3
?=     (2) 
Where nD is the number of Dynabeads, ?? is the difference in magnetic susceptibility between 
the medium and the beads, Vi is the volume of a bead, ? is the viscosity of the medium, Di is the 
diameter of the islet. 
 The magnetophoretic mobility distribution measured using MPTV is used to 
predict the QMS output fractions based on the defined flow rate parameters.  QMS is a split flow 
type continuous magnetic sorter developed to separate single cells by labeling specific cells with 
magnetic beads.  QMS was improved to isolate Dynabeads infused islets from exocrine tissue 
 59
based on their magnetophoretic mobility (?m) (Kennedy et al., 2007).  Figure 3.1 describes the 
basic QMS mechanism for isolation of magnetic cells from nonmagnetic cell populations.  QMS 
can be operated continuously with labeled islets with unlabeled tissue entering at inlet a? 
(sample) and carrier buffer at inlet b? (buffer) and isolated islets exiting the system at outlet b 
(positive), unlabeled tissue at outlet a (negative), some of the tissue which are labeled high 
enough to reach the wall of the flow channel will stick to the wall and can be collected at end of 
the isolation process and is considered as Wall (W) fraction.  The boundary between inlet flows 
is called as Inner Splitting Surface (ISS) and the boundary between outlet flows is called as 
Outer Splitting Surface (OSS).  Distance between ISS and OSS is called the Transport Lamina.  
The first critical mobility whereby a particle entering the flow channel at the ISS just reaches the 
OSS and is eluted into the b (positive) fraction is identified as ?m0 and is defined by 
0
0
2
1
1
00
0
)ln(
)1(
),(
2 m
ISSOSSs
i
OSSISS
m
m LS
vr
A
I
LSr
Q ll
l
ll +
?= pi?    (3) 
Where il  is the radial position given by ii rr=l and OSSISSo lll ,,  correspond to wall, ISS and 
OSS.  L is distance between two splitters.  1A , 2A and ),( 11 llI  are functions used to simplify the 
calculations and are given by 
)1( 221 AA i ?+= l          (4) 
i
iA
l
l
1ln
1 2
2
?=           (5) 
l
llllll 1])(ln22ln4[),(
2
2
2
11 AI +?=      (6) 
The radial position of the ISS, ISSl is calculated using the following integration factor (Kennedy 
et al., 2007): 
 60
                      
Figure 3.1: Schematic diagram of the Quadrupole Magnetic cell Sorter (QMS) 
 
 
 
 
 
 
 
Sample Inlet (a?) Buffer Inlet (b?)
Positive 
Fraction (b) NegativeFraction (a)
Magnet
Non-Magnetic 
Particle
Magnetic 
ParticleISS
OSS
Magnetic 
ForceT
ra
ns
po
rt 
La
mi
na
Splitter
Tr
an
sp
or
t L
am
ina
 61
ISS
i
AAI ISSi lllllllll ]ln22[),( 2222422 ?+?=      (7) 
Similarly, the position of the OSS can be calculated using eq. (7) by replacing ISSl  with OSSl . 
The second critical mobility whereby a particle entering at the wall of the core reaches the OSS 
and exit with the b fraction is identified as ?m1and is defined by  
0
0
2
1
1
00
1
)ln(
)1(
),(
2 m
iOSSs
i
OSSi
m
m LS
vr
A
I
LSr
Q ll
l
ll +
?= pi?    (8) 
The third critical mobility whereby a particle entering at the ISS reaches the wall of the shell and 
may be trapped in the flow channel is identified as ?m2 and is defined by 
0
0
2
1
1
00
2
)ln(
)1(
),(
2 m
ISSos
i
oISS
m
m LS
vr
A
I
LSr
Q ll
l
ll +
?= pi?    (9) 
The final critical mobility whereby a particle entering at the core wall reaches the wall of the 
shell and remains on the wall is identified as ?m3 and is defined by 
0
0
200
3
)ln(
2 m
ois
m
m LS
vr
ALSr
Q ll+=
pi?     (10) 
All the particles with m<m0 exit in a fraction, ?m0? ?m? ?m1 will exit either in a or b fraction, 
?m1? ?m? ?m2 will exit in b fraction, ?m2? ?m? ?m3 will exit in b fraction or become trapped on 
the flow channel wall and ?m? ?m3 with be trapped on the flow channel wall. 
3.4  Materials and Methods 
3.4.1  Particles and Viscous Liquid 
 Three types of magnetic particles with similar size ranges and different magnetization 
were used for this study. Agarose magnetic beads (Bioscience Beads, RI) with diameters 250-
350 ?m and magnetite loading of 0.5%, 1% and 6% by volume were used for testing.  As the 
particles used for testing settled very rapidly in the aqueous solution, high-viscosity liquid was 
 62
prepared by dissolving Ficoll? (400,000 MW) (Sigma-Aldrich) to maintain the particles in 
suspension while measuring their mobilities.  
3.4.2  QMS System 
QMS is a split flow type continuous magnetic sorter developed to separate single cells by 
labeling specific cells with magnetic beads.  QMS was improved to isolate Dynabeads infused 
islets from exocrine tissue based on their magnetophoretic mobility (?m) (Kennedy et al., 2007).  
Figure 3.1 describes the basic QMS mechanism for isolation of magnetic cells from nonmagnetic 
cell populations.  QMS can be operated continuously with labeled islets with unlabeled tissue 
entering at inlet a? and carrier buffer at inlet b? and isolated islets exiting the system at outlet b 
and unlabeled tissue at outlet a.  
3.4.3  Magnetic Particle Tracking Velocimetry 
The MPTV technology is comprised mainly of four components (Figure 3.2): a sample 
channel containing the suspension of cells and particles, a magnet capable of providing a 
constant magnetic force in the sample channel zone, a pump for introducing fluid and sample, a 
video microscope capable of imaging the sample with various degrees of magnification, and a 
computer capable of capturing and processing the video images for particle mobility analysis.   
3.4.4  Magnet Assembly 
The custom designed magnet assembly is comprised of a base plate, two neodymium-
iron-boron (NeFeB) magnets and two 1018 carbon steel pole pieces.  The pole pieces are shaped 
to match a very specific modified hyperbolic profile.  The magnetic force is perpendicular to the 
direction of gravitational force so magnetophoretic mobility measurements are independent of 
sedimentation velocity, which can be used independently to estimate particle dimensions. 
 
 63
3.4.5  Fluid System 
The stopped flow channel consists of a borosilicate glass channel with square (2 mm) 
cross-section.  One end of the 6 cm long flow channel is connected by a pair of solenoid pinch 
valves to a disposable 50 ml syringe for sample injection and a 50ml syringe for priming buffer 
while the outlet end connects into a waste vessel. The observation channel is positioned within 
the magnet assembly and is sandwiched between the video lens and the backlight.  Vacuum 
pump and pinch valves were used to control the sample flow into and out of the channel. 
3.4.6  Imaging System  
The selection of the video microscope system components and their operational settings 
is critical, as the pixel size, field of view, magnification, and frame capture speed dictate the 
maximum and minimum velocities and particle sizes that can be accurately analyzed.  Particle 
movement in the flow channel was recorded with a grasshopper 2.0MP B&W, 1394b, and 1/1.8 
inch CCD camera (Point Gray, AZ) operating at 30 frames per second with 1x objective 
(Edmund Optics, NJ) mounted on a ? inch Manual translation stage (Thor Labs, NJ). Light was 
supplied by dark field illumination using a white light source (Edmond Optics, NJ). 
3.4.7  Analysis of Video Data 
Image processing software converts the video data into useful velocity data and 
corresponding property data such as hydrodynamic radius and magnetophoretic mobility.  A 
program named ?Cytotest? first thresholds the images into binary black or white images based on 
a user-defined gray-level threshold.  This eliminates pixels that are below certain brightness, 
leaving the light-reflecting particles visible in the image records. It then identifies particles or 
cells in each frame of the video file using a program named ?OpenCV? in which cells & particles 
outside a specific size range are rejected.  It then searches successive frames in the video record  
 64
 
 
Figure 3.2:  A simplified diagrammatic representation of an MPTV system 
 
 
 
 
 
 
 
 
 
 
 
 
 
Computer 
Sample
Buffer
Observation Channel
Magnet Pole Pieces
Lens
Camera
Vacuum pump
 65
 
to identify particle movement by connecting the tracks of each identified particle to its match in a  
successive frame.  Proprietary predictive and adaptive algorithms are employed to improve the 
accuracy of particle matching across frames and eliminate bad tracks.  The track lengths are then 
converted to a rate of travel in pixels per millisecond, and that rate is converted to a 
magnetophoretic mobility measurement based on the magnet strength and image pixel size.  That 
value is recorded in a histogram with other tracks delimited by logarithmic bin sizes and plotted 
in the software. 
3.4.8  Experimental 
 
The camera lens was adjusted to visualize the central region of the sample cell at constant 
magnetic energy gradient using the MPTV translation stage.  Particle suspensions were injected 
into the sample cell using a syringe after priming the cell with buffer using a vacuum pump.  
Fixed volumes of the particle suspension typically 1ml to 5ml were injected into the field of view 
to observe the magnetic deflection of the particles.  Particle samples were pumped in a direction 
opposite to the direction in which the magnetic energy gradient operates on the particles then 
stopped. Particle deflection was observed with no fluid velocity so that particle motion was only 
due to magnetic and gravitational force.  Successive volumes of the suspensions were injected to 
obtain several series of deflection images at a rate of 30 frames per second.  After analyzing the 
samples in MPTV, the optimized parameters given by software were used to run QMS with 
magnetic particles. Outlets from QMS were counted for particle concentrations by microscope 
and compared with MPTV predictions.  
 
 
 66
3.5  Results and Discussion 
 Several improvements were made on cell tracking velocimetry to facilitate measurement 
of the magnetophoretic mobility of much larger particles such as pancreatic islets and to obtain 
essential data to set the parameters of the QMS to isolate the magnetic fraction (islets) from non-
magnetic fraction in QMS. A newly built MPTV was tested with Dynabeads? to evaluate 
performance.  Figure 3.3 shows a histogram of the magnetophoretic mobility of Dynabeads? 
developed by MPTV.  The magnetophoretic mobility was obtained by dividing the MPTV 
determined particle velocity by the magnitude of the magnetic energy gradient, Sm, 6.275 
TA/mm2. Particles with mobility less than 10-16 m3/TAs are considered as underflows.  Figure 3.5 
is the tracks developed by MPTV for one set of Dynabeads moving magnetically in the sample 
cell.  Dark spots in the Figure 3.5 are the disturbances on the sample cell wall, for example, were 
neglected when tracking by thresholding the images.  Any of these disturbances measured as 
tracks were characterized as ?underflows? (particles with zero mobility).  These are seen in 
Figure 3.3 as a vertical line at 10-16 on mobility axis. 
The important development made to the MPTV is the ability to predict the QMS flow 
parameters required to get the desired isolation based on the mobilities measured using MPTV.  
Figure 3.4 shows the histograms developed by MPTV for the three outlet fractions from QMS 
for Dynabeads based on the mobility predictions for the total flow rate in the QMS of 400 
ml/min, inlet flow ratio of 0.2 and outlet flow ratio of 0.4.  Underflows are neglected while 
predicting the flow parameters.  
 67
 
 
Figure 3.3: Mobility histogram given by MPTV for Dynabeads  
 68
 
Figure 3.4: Histograms of Dynabeads in each fraction of the QMS output. 
 
 69
 
Figure 3.5: Particle tracks developed by MPTV. Red dots are the particles tracked and black spots are disturbance. 
 70
The expression of particle magnetization from their motion in a magnetic field is a 
compound quantity depending on particle size, magnetic velocity and field gradient.  Field 
gradient in the measurement region is assumed to be constant.  Size of the three different 
magnetic particles used in this study is maintained similar in size. So the variation in magnetic 
mobility is due to the magnetite loading in the particles.  Variation in the magnetite concentration 
changes the velocity of the particle in the magnetic filed, which is measured by the MPTV.  
Three different magnetic particles with magnetite loading 0.5%, 1% and 6% were used for 
MPTV studies. 
Figure 3.6 shows the histogram of the magnetic micro beads BSI.  The x-axis represents 
the magnetophoretic mobility values of the magnetic bead on a log scale and y-axis represents 
the fraction of the beads with specific magnetophoretic mobility.  Mobilities were measured in a 
high viscous Ficoll? medium to avoid settling of the micro beads while measuring the velocity 
due to magnetic field.  Histograms are plotted with mobilities calculated for water based on the 
viscosities.  Mobilities are measured for the beads with minimum size of 250 ?m and maximum 
size of 350 ?m by omitting the other smaller and bigger beads.  Total 126 beads are tracked with 
2030 tracks. Beads with mobility less than 10-16 are grouped as under flows.  
Figure 3.7 and 3.8 shows the mobility histograms for magnetic micro beads BSII and 
BSIII. Results generated with 147 beads with 2672 tracks for BSII and 186 beads with 3276 
tracks for BSIII were plotted.  MPTV can analyze large number of particles in small amount of 
time on a particle-by-particle basis.  As the iron oxide concentration varies in beads BSI, BSII 
and BSIII, which changes the density of the beads, measurements were done with Ficoll? 
solution with appropriate density to avoid the settling of the particles in the sample cell. The 
required density of the medium was 1.36 g/cm3 corresponding to Ficoll? 400 having viscosity of  
 71
1E-15 1E-14 1E-13 1E-12 1E-11 1E-10 1E-9 1E-8 1E-7
0.0
0.5
1.0
1.5
2.0
2.5
3.0
Fr
ac
tio
n o
f P
ar
tic
les
Mobility, m3/T-A-s
 
 
 
  
Figure 3.6: Magnetophoretic mobility histogram of BSI magnetic particles. 
 
 
 
 72
 
1E-15 1E-14 1E-13 1E-12 1E-11 1E-10 1E-9 1E-8 1E-7
0.0
0.5
1.0
1.5
2.0
2.5
3.0
Fr
ac
tio
n o
f P
ar
tic
les
Mobility, m3/T-A-s
 
 
 
Figure 3.7: Magnetophoretic mobility histogram of BSII magnetic particles. 
 
 73
1E-14 1E-13 1E-12 1E-11 1E-10 1E-9 1E-8 1E-7
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
Fr
ac
tio
n o
f P
ar
tic
les
Mobility, m3/T-A-s
 
 
 
Figure 3.8: Magnetophoretic mobility histogram of BSIII magnetic particles. 
 
 
 
 
 
 
 
 
 
 
 
 74
10.8 Pa s. Particle velocities were corrected for viscosity when calculating magnetophoretic 
mobilities. Mobility histograms show the increase in the average and maximum mobility of the 
beads with the increase in the magnetite concentration from beads BSI to beads BSIII and these 
histograms indicate that significant differences in magnetophoretic mobility can be detected 
between magnetic beads.  
Magnetophoretic mobility measurements and flow rate parameters prediction made by 
MPTV software were tested by conducting the experiments in QMS using the pure magnetic 
beads.  Figure 3.8 shows the comparison of the predictions and experimental results for the three 
outlet fractions from QMS for magnetic particles BSI. The QMS isolation process was controlled 
by three flow parameters: total flow rate (Q), inlet flow ratio (Ri=Qa?/Q), Outlet flow ratio 
(Ro=Qa/Q). These flow rates were fixed based on the measured magnetophoretic mobility of the 
magnetic beads.  The critical mobilities ?m0 and ?m2 were selected to maximize the amount of 
beads in the b fraction.  The flow rates predicted were Q = 400 ml/min, Ri = 0.25 and Ro = 0.4 
corresponding to critical mobilities for beads BSI of ?m0 = 14.79X10-12 m3/TAs and ?m2 = 
1.355X10-11 m3/TAs.  
Figure 3.9 is the recovery of the particles in all fractions.  Total recovery of the magnetic 
particles during the experimental isolation with QMS was 98%.  The loss of some magnetic 
particles was due to the settling of the particles in the inlet tubes and at the bottom of the flow 
channel.  MPTV predicted ?a? fraction recovery was high compared to the experimental isolation 
as the MPTV predicted ?a? fraction also contains some disturbance tracks.  Predicted and 
experimental ?b? fraction recovery is very low because the magnet strength of the QMS is high 
enough to capture most of the magnetic particles on to the wall of the flow channel.  The 
difference in the wall fraction recoveries is due to the difference in the ?a? fractions which  
 75
a b W
0
20
40
60
80
100
%
 re
co
ve
ry
Fraction
 MPTV Predicted
 Experimental
 
 
 
Figure 3.9: Comparison of MPTV predicted fractional recovery of BSI particles in the three 
outlet fractions of the QMS with experimental results. 
 
 76
a b W
0
20
40
60
80
100
%
 re
co
ve
ry
Fraction
 MPTV predicted
 Experimental
 
 
 
Figure 3.10: Comparison of MPTV predicted fractional recovery of BSII particles in the three 
outlet fractions of the QMS with experimental results. 
 
 
 77
a b W
0
20
40
60
80
100
%
 re
co
ve
ry
Fraction
 MPTV predicted
 Experimental
 
 
 
Figure 3.11: Comparison of MPTV predicted fractional recovery of BSI particles in the three 
outlet fractions of the QMS with experimental results. 
 
 
 
 
 
 
 
 
 
 
 78
affected the total recovery. Figures 3.10 and 3.11 show the experimental outlet fractions 
compared with MPTV predictions and recoveries for beads BSII and BSIII.   
Figure 3.12a shows the histogram of magnetophoretic mobilities of the porcine islets 
isolated from acinar tissue with QMS (Chapter 4).  These are the islets from the b fraction of the 
QMS operated at a total flow rate of 400 ml/min and inlet ratio of 0.4 and outlet ratio of 0.6.  
Islets were infused with Dynabeads.  After MPTV developed the mobility histogram, QMS flow 
parameters for the islets isolation were set to Q= 400 ml/min, Qa?/Q = 0.4 and Qa/Q = 0.6 and 
fractional recovery histograms were generated.  Figure 3.12b, 3.12c and 3.12d shows the MPTV 
predicted fractional recovery of each outlet of QMS.  These histograms show that more than 90% 
of the islets used to measure the mobility exit in the b fraction. Islets that are in the ?a? fraction 
histogram show the nonspecific crossover in the islet isolation experiments.  This also confirmed 
the efficiency of the MPTV to predict the QMS parameters based on the measured mobilities.  
3.6  Conclusions 
 The purpose of this study was to develop a MPTV to measure the mobility of magnetic 
particles up to 1000 microns in size and to predict optimized flow parameters based on the 
magnetophoretic mobilities to isolate magnetic particles from nonmagnetic particles using QMS.  
The capability of newly developed MPTV in measuring magnetophoretic mobility was 
confirmed with the measurements of the mobilities of the standard Dynabeads?.  MPTV was 
successfully used to measure the mobility of magnetic particles up to 500 microns in size with 
different mobilities.  MPTV?s ability to predict the optimized flow parameters was also tested 
successfully with magnetic particles and the isolated islets of Langerhans as it is the final 
application of the MPTV.  MPTV can be used online to analyze the mobilities of the 
 79
 
Figure 3.12: a) Magnetophoretic mobility histogram of pancreatic islets isolated with QMS. b) MPTV predicted b fraction c) a fraction 
d) wall fraction of the islets at a total flow rate of 400ml/min and Ri = 0.25 and Ro = 0.6.
1 E -1 6 1 E -1 5 1 E -1 4 1 E -1 3 1 E -12 1 E -1 1 1 E -1 0 1 E -9 1 E -8
0 .0
0 .5
1 .0
1 .5
2 .0
2 .5
3 .0
3 .5
4 .0
4 .5
5 .0
F r
a c
t i o
n  o
f  I s
l e t
s
M o b ility, m 3/T -A -s
1E -16 1E -15 1E -14 1E -13 1E -12 1E -11 1E -10 1E -9 1E -8
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
F r
a c
t i o
n  o
f  I
s l e
t s
M o b ility, m 3/T -A -s
1 E -1 6 1 E -1 5 1 E -1 4 1 E -1 3 1 E -1 2 1 E -1 1 1E -1 0 1E -9 1 E -8
0 .0
0 .5
1 .0
1 .5
2 .0
2 .5
3 .0
3 .5
4 .0
4 .5
5 .0
F r
a c
t i o
n  o
f  I
s l e
t s
M o b ility, m 3/T -A -s
1E-16 1E-15 1E-14 1E-13 1E-12 1E-11 1E-10 1E-9 1E-8
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
F r
a c
t i o
n  o
f  I s
l e t
s
M obility, m 3/T-A-s
a
d
b
c
F r
a c
t i o
n  o
f  I s
l e t
s
F r
a c
t i o
n  o
f  I
s l e
t s
F r
a c
t i o
n  o
f  I
s l e
t s
F r
a c
t i o
n  o
f  I s
l e t
s
 80
magnetically infused islets and to predict the flow parameters before sending tissue through 
magnetic field to isolate using QMS. 
3.7  References 
Chalmers JJ, Zhao Y, Nakamura M, Melnik K, Lasky L, Moore L, Zborowski M. 1999.  An 
Instrument to Determine the Magnetophoretic Mobility of Labeled, Biological Cells and 
Paramagnetic Particles.  Journal of Magnetism and Magnetic Materials 194:231-241. 
Chalmers JJ, Haam S, Zhao Y, McCloskey K, Moore L, Zborowski M, Williams PS.  1999. 
Quantification of Cellular Properties from External Fields and Resulting Induced 
Velocity: Cellular Hydrodynamic Diameter.  Biotechnology and Bioengineering 64:509-
518. 
Davis RS. 1993. New method to measure magnetic susceptibility, Measurement Science and 
Technology 4(2):141-147. 
Gill SJ, Malone CJ, Downing M. 1960. Magnetic susceptibility of single small particles. The 
review of scientific instruments 31(12):1299-1303. 
H?feli UO, Ciocan R, Dailey JP. 2002. Characterization of Magnetic Particles and Microspheres 
and Their Magnetophoretic Mobility Using a Digital Microscopy Method.  European 
Cells and Materials 3(2):24-27. 
Kennedy DJ, Todd P, Logan S, Becker M, Papas KK, Moore LR. 2007. Engineering quadrupole 
magnetic flow sorting for the isolation of pancreatic islets. Journal of Magnetism and 
Magnetic Materials 311:388-395. 
London NJM, Swift SM, Clayton HA. 1998. Isolation, culture and functional evaluation of islets 
of Langerhans. Diabetes & Metabolism 23:200-207. 
 81
McCloskey KE, Zborowski M, Chalmers JJ. 2001. Measurement of CD2 Expression Levels of 
IFN-?-Treated Fibrosarcomas Using Cell Tracking Velocimetry.  Cytometry 44:137-147. 
McCloskey KE, Chalmers JJ, Zborowski M. 2000. Magnetophoretic Mobilities Correlate to 
Antibody Binding Capacity.  Cytometry 40:307-315. 
McCloskey KE, Comella K, Chalmers JJ, Margel S, Zborowski M. 2001a. Mobility 
Measurements of Immunomagnetically Labeled Cells Allow Quantitation of Secondary 
Antibody Binding Amplification.  Biotechnology and Bioengineering 75(6):642-655. 
Melnik K, Nakamura M, Comella K, Lasky LC, Zborowski M, Chalmers JJ. 2001. Evaluation of 
Eluents from Separations of CD34+ Cells from Human Cord Blood Using a Commercial, 
Immunomagnetic Cell Separation System.  Biotechnology 17:907-916. 
Moore LR, Milliron S, Williams PS, Chalmers JJ, Margel S, Zborowski M.  2004. Control of 
Magnetophoretic Mobility by Susceptibility-Modified Solutions as Evaluated by Cell 
Tracking Velocimetry and Continuous Magnetic Sorting.  Analytical Chemistry 76:3899-
3907. 
Nakamura M, Zborowski M, Lasky LC, Margel S, Chalmers JJ.  2001. Theoretical and 
Experimental Analysis of the Accuracy and Reproducibility of Cell Tracking 
Velocimetry.  Experiments in Fluids 30:371-380. 
Soon-Shiong P, Heintz R, Terasaki P. 1989. Identification of novel blood group reactive 
monoclonal antibodies cytotoxic to human acinar cells but not islets. Transplantation 
Proceedings 21(1):2622-2623. 
Sridhar Reddy, Lee R Moore, Liping Sun, Maciej Zborowski, J J Chalmers. 1996. 
Determionation of the magnetic susceptibility of labeled particles by video imaging, 
Chemical Engineering Science, 51(6):947-956. 
 82
Suwa M, Watarai H. 2004. Magnetophoretic Detection of Photo-Induced Spin Transition.  
Chemistry Community 1656-1657. 
Watanabe K, Suwa M, Watarai H. 2004. New Principles of Magnetophoretic Velocity Mass 
Analysis.  Analytical Sciences 20:1483-1485. 
Watarai H, Namba M. 2001. Magnetophoretic Behavior of Single Polystyrene Particles in 
Aqueous Manganese (II) Chloride.  Analytical Sciences 17:1233-1236. 
Zborowski M, Ostera GR, Moore LR, Milliron S, Chalmers JJ, Schechter AN.  2003. Red Blood 
Cell Magnetophoresis.  Biophysics Journal 84:2638-2645. 
Zhang H, Nakamura M, Comella K, Moore L, Zborowski M, Chalmers J.  2002. 
Characterization/Quantification of the Factors Involved in the Imparting a 
Magnetophoretic Mobility on Cells and Particles.  European Cells and Materials 3(2): 34-
36. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 83
4. QMS Isolation of Porcine Islets of Langerhans 
4.1  Abstract 
Pancreatic islet transplantation offers a viable option to achieve permanent metabolic 
control in Type 1 diabetes patients.  However, large quantities of pure viable donor islet cells are 
necessary for transplantation.  Using currently available islet isolation methods multiple donor 
organs are required to achieve successful transplantation, and there is a demand for an isolation 
method with high islet yield and viability. Additionally, with porcine xeno-islet cell 
transplantation providing much hope, improving the porcine islet isolation process has become a 
worthwhile endeavor. To achieve isolation with high purity and yield Quadrupole Magnetic 
Sorting (QMS), a single cell separation method, is being modified for the isolation of pancreatic 
islets. Islets are infused with magnetic4.6?m Dynabeads? and separated continuously with 
QMS.  Results from 10 porcine pancreas isolations indicated the possibility of infusing islets 
with magnetic beads and isolating them continuously, thereby reducing the exposure time of 
islets to digestive enzymes. QMS isolated islets showed good morphology compared to those 
isolated by the standard COBE centrifugation method. Oxygen consumption rate (OCR) per 
DNA measurements confirmed the viability of the islets after isolation.  The QMS isolation 
method can save post-isolation culture time and helps to eliminate the mechanical stress due to 
centrifugation of the islets. Restoration of euglycemia in diabetic nude mice transplantation 
experiments results confirmed Dynabeads do not affect the functionality of the islets. 
Keywords: magnetic flow sorter, pancreatic islets, magnetic particles, diabetes, islets isolation
 84
4.2  Introduction 
 Whole pancreas transplants or transplants of isolated pancreatic islets can achieve insulin 
independence in type 1 diabetes (Shapiro et al., 2000; Hering et al., 2005; Frank et al., 2004; 
Froud et al., 2005).  Whole Pancreas transplant includes complex surgery and involves 
significant morbidity and mortality.  Even 20 years after the first successful clinical islet 
transplantation the goal of donor-to-recipient ratio of 1:1 is still elusive.  While significant 
progress has been made towards achieving this objective, most transplants require two, three, and 
in some cases even up to four infusions of pancreatic islets from separate donors before insulin 
independence is realized. 
 Islets constitute only 1-2% of the pancreas volume, and their isolation and purification 
are stressful mechanical and enzymatic procedures that can inflict significant damage, which 
may be further amplified by prolonged times of warm and cold ischemia (London et al., 1994). 
Current clinical human islet purification protocols rely on a density gradient centrifugation 
method which depends on the small density differences between islets and exocrine tissue.  As 
the density of the exocrine tissue depends on the age of the donor and progress of enzymatic 
digestion, over which there is no direct control, purity of islets attainable low and inconsistent.  It 
is generally accepted that there are substantial losses in islet yield and potency during digestion 
and purification (Hering et al., 2002). During the process of enzymatic digestion and subsequent 
purification, liberated islets encounter immense stresses caused by, among other factors, 
hypoxia, sudden and repeated changes in temperature, shear forces, disruption of cell matrix 
interactions, removal of critical growth factors, high concentrations of reactive oxygen 
intermediates, proinflammatory cytokines and exposure to hyperosmolar density gradients. 
Density gradient method also has undesirable effects on islets like exposure to high mechanical 
stress, toxic chemicals and prolonged exposure to proteolytic enzymes (Samejima et al., 1998; 
 85
Pinkse et al., 2004). Additionally, with porcine xeno-islet transplantation providing much hope ( 
Hering et al., 2006), improving the porcine islet isolation process has become a worthwhile 
endeavor. 
 Magnetic micro-particles coated with antibodies were used to successfully isolate islets 
from rat pancreas by targeting exocrine tissue (Davies et al., 1995; Winto-Morbach et al., 1989; 
Davies et al., 1994; Davies et al., 1997).  The major objective in these studies was to enhance 
secondary purification following isolation and primary purification with standard density 
gradients and centrifugation. Positive selection of islets is more desirable than negative selection, 
as positive selection would also allow subsequent magnetic manipulation of islets (Nandigala et 
al., 1997).  These techniques used bar magnets to isolate islets in a batch process.  These 
techniques fail to get high purities as inter-islet binding formed capture nets and retained 
significant amounts of acinar tissue, and monoclonal antibodies were found to easily shear off of 
tissue under gentle handling. 
 Magnetic separation of glomeruli from kidney tissue was achieved by directly infusing 
the iron oxide particles into renal vasculature (Gauthier et al., 1988).  It had been observed that 
pancreatic islets and renal glomeruli have similar angioarchitechtures.  Pinkse et al. (2004) 
infused unlabeled paramagnetic microbeads directly into the vascularization of rat pancreata 
prior to enzymatic digestion (Pinkse et al., 2004). Results showed that the microbeads were 
preferentially captured in the islet vasculature, that the islets were easily purified, and that the 
purified islets could reverse insulin dependence in rodents.  This process was successfully 
applied to human islets to achieve maximum of 80% yields but only 40% purity.  Iron oxide 
particles were replaced by Dynal? D-450 Dynabeads? to label only islets by taking the 
advantage of the hypervascularity of islets compared with acinar tissue (Tons et al., 2008). 
 86
 Quadrupole Magnetic Sorting (QMS) technology developed for stem cell separation was 
successfully applied to purify porcine islets of Langerhans with high yields but with very low 
purity ( Shenkman et al., 2009).  It has been observed that mechanical stress applied in QMS or 
the magnetic beads present in the islets do not affect islets functionality (Shenkman et al., 
2009a).  To overcome the limitations of the density gradient method and to develop a technique 
which makes donor to recipient ratio to 1:1, we redesigned and tested QMS (Kennedy et al., 
2007) to isolate porcine islets of langerhans as soon as they are liberated from the pancreatic 
tissue mass.  
4.3  MATERIALS AND METHODS 
4.3.1  QMS and separation theory 
 QMS is a type of split-flow thin channel continuous isolation device (Sun et al., 1998; 
Tong et al., 2007; Jing et al., 2007; Hoyos et al., 2003). Figure 4.1 shows the design and the 
mechanism of the QMS for isolating magnetic islets from non-magnetic tissue. Briefly, the 
system is composed of an annular column with a concentric inlet splitter to separate the islet 
suspension inlet (a?) from the carrier buffer inlet (b?) and an outlet splitter to separate the outlet 
for isolated islets (Positive Fraction, b) from that for nonmagnetic tissue (Negative fraction, a). 
The column is centered within the bore of a strong Halbach quadrupole magnet and by the effect 
of magnetic force, islets previously infused with Dynabeads are deflected outward into the 
positive fraction or onto the wall. Unlabeled tissue is collected in the inner, negative fraction. 
The magnetic field is provided by four neodymium-iron-boron permanent magnets and four pole 
pieces with a maximum magnetic field strength of the quadrupole, B0, of 0.714 T.  
 87
 
Figure 4.1: Schematic diagram of the quadrupole magnetic cell sorter (QMS) showing separation 
of magnetic from non-magnetic particles. 
 
 
 
 
 
Splitter
ISS
OSS
Buffer 
Inlet (b?)
Sample
Inlet (a?)
Positive
Fraction (b)NegativeFraction (a)
Magnetic
Particle
Non-magnetic
Particle
 88
The fluid trajectory in the annulus is well documented and is described by the relation: 
?? =
?
?
???
1
)(
)(
0
duvdz
m
z
      (4.1) 
Where um is the velocity of the magnetic particle in the magnetic field, ? is the radial position, 
)(?v  is the velocity profile of fluid flow. As the density of the islets is high enough to sediment 
in the flow channel, adding the Stokes? sedimentation velocity )( sv  to the velocity relation gives: 
?? +=
?
?
???
1
)(
)(
0
du vvdz
m
s
z
      (4.2) 
4.3.2  Magnetophoretic Mobility 
The QMS system?s ability to isolate magnetic bead-infused islets depends on the 
magnetphoretic mobility of islets, which indicates the degree to which the islets are infused with 
Dynabeads. The magnetophoretic mobility, m? , can be calculated by measuring the velocity of 
the magnetic particle in the constant magnetic energy gradient. 
m
m
m S
v=?           (4.3) 
mv  is the magnetically induced velocity and mS  is the magnetic energy gradient. The 
magnetophoretic mobility can be measured using Magnetic Particle Tracking Velocimetry 
(MPTV). MPTV is the improved model of the Cell tracking velocimeter to measure the mobility 
of the large magnetized  particles such as islets.  
 
4.3.3  Pancreas Procurement and labeling  
All procedures involving animals were approved by the Institutional Animal Care and 
Use Committee (IACUC) and performed at the Schulze Diabetes Institute at the University of 
 89
Minnesota. The porcine procurement was performed by standard protocol used at the Schulze 
Diabetes Institute (Ferrer et al., 2008). Pancreata was harvested from adult Landrace pigs 
following cardiac death using an en bloc dissection technique. The posterior aorta was identified, 
longitudinally divided and both the celiac trunk (CT) and superior mesenteric artery (SMA) were 
cannulated. Distal splenic, gastric and hepatic vessels were clamped and 1 L of cold preservation 
solution (CPS, cold storage/purification stock solution containing 2% Pentastarch, Mediatech, 
Inc, Herndon, VA) was flushed into both the CT and SMA, simultaneously. During the flush, the 
main pancreatic duct was identified and cannulated just proximal to its insertion into the 
duodenum. Approximately 60 mL of CPS was then slowly infused over two minutes by hand-
syringe into the main pancreatic duct. Following the flush, the pancreas was excised and divided 
into the combined connecting and duodenal lobes and splenic lobe, all while taking care to 
preserve native vasculature. To compare infused versus uninfused islets obtained from the same 
pancreas, the splenic pancreatic lobe alone was infused with magnetic beads. Paramagnetic 
Magnetic Particles (4.5 ?m diameter, Dynabead M450, Invitrogen, Carlsbad, CA) suspended in 1 
L of CPS were then infused into the splenic lobe only, via the CT and SMA by either hanging 
bag or hand-syringe infusion. Infusion of MP suspension was immediately followed by a 1 L 
flush with CPS. All lobes were then submerged into CPS and stored at 0-4?C for transportation 
from the procurement facility to the laboratory.  A procedure requiring less than 30 minutes.   
4.3.4  MRI 
Following complete formalin fixation (? 24 hours), pancreata underwent MRI. MRI was 
performed at 1.5 T with an APOLLO spectrometer (Tecmag Inc., Houston, TX, USA) and a 
custom-built 16-leg low-pass birdcage resonator (24 cm diameter, 20 cm in length). Images were 
obtained with a T2*-weighted true three-dimensional gradient echo sequence with one 
 90
dimension frequency-encoded and the other two dimensions phase-encoded. A nonselective 
pulse was used to nutate the nuclei approximately 25 degrees during the scan. Typical 
acquisition parameters included a 9.6 msec acquisition time, 6.5 msec echo time, 120 msec 
repetition time, and a 20 cm x 10 cm x 5 cm field of view. 
4.3.5  Histopathology  
After MRI, biopsies were collected from several regions of the pancreas, embedded in 
paraffin and sectioned at 4 ?m. Sections were examined using hematoxylin and eosin (H/E) and 
insulin immunohistochemical stains. Sections were evaluated by an experienced histopathologist 
(Dr. Thomas D.O?Brien) to estimate the fraction of magnetic particle incorporation within islet 
and acinar tissues.  
4.3.6  Pancreas Digestion and Islets Isolation 
Tissue samples were collected from pancreas before digestion for histology evaluation 
for beads. Then the infused pancreas was digested for islets isolation using the Ricordi method 
(Ricordi et al., 1998). Before performing an actual isolation using QMS, the QMS column was 
filled with Hanks? balanced salts solution (HBSS, Sigma, St.Louis, MO) with serum solution. 
Once the digested tissue collection started from the Ricordi chamber, the collection tube from the 
Ricordi chamber was connected to a reservoir bag, partially filled with HBSS with 10% porcine 
serum and placed in line between the QMS flow channel a? inlet and the Ricordi chamber.  
Tissue arrived at the reservoir at a flow rate equal to the QMS inlet flow rate. A reservoir bag is 
essential to avoid any pressure problems due to differences in the flow rates of QMS inlet and 
Ricordi chamber collection pumps and also helps to avoid tissue concentration problems. Outlet 
fractions from QMS were collected and centrifuged to collect samples for analysis. After sending 
all digested tissue through the QMS, the flow channel was carefully removed and islets attached 
 91
to the wall were recovered for counts and quality analysis studies. All tubing and flow channel 
used are disposable and sterilized.  All flow channels were recycled in this research. 
4.4  Islet Quality Assessment 
4.4.1  Oxygen Consumption Rate (OCR) Assay 
OCR was measured on COBE (standard method) and QMS purified islets as described 
previously (Papas et al., 2007) . Islet samples were rinsed and suspended in Dulbecco?s modified 
Eagle?s medium (Mediatech, Herndon, VA) containing 4.5 g/L L-glutamine and supplemented 
with 100 U/ml penicillin, 100?g/ml streptomycin, 10 mM HEPES without serum. Then each 
islet suspension was placed into three or more 200-?L titanium chambers. The chambers were 
sealed and maintained at 37oC. The time-dependent oxygen partial pressures (pO2) within the 
chambers were recorded over time using a fluorescent-based fiber optic oxygen sensor (Micro 
Oxygen Uptake System, FO=SYSZ-P250; Instech Laboratories, Plymouth Meeting, PA). The 
initial tracings at the highest pO2 were then fit by a straight line and OCR was calculated from 
the following equation: OCR = Vch?(?pO2), where ?pO2=Dt is the slope of the line, Vch is the 
chamber volume, and ? is the Bunsen solubility coefficient of oxygen in an aqueous medium at 
37oC (1.27 nmol mm Hg-1 mL-1). 
4.4.2  DNA Quantification 
 Islet samples analyzed by the OCR assays were subsequently sampled for quantification 
of DNA content. To measure DNA content, islet samples were diluted in an aqueous solution of 
1M ammonium hydroxide and 3.4mM Triton X-100 and sonicated. DNA content was 
determined using the Quant-iT PicoGreen dsDNA kit (Molecular Probes, Eugene, OR) as per 
manufacturer?s instructions. Fluorescence readings were taken on a SpectraMax M5 microplate 
reader (Molecular Devices, Sunnyvale, CA) 
 92
4.4.3  Nude Mouse Bioassay 
 Adult mice were rendered diabetic by intraperitoneal streptozotocin injection. Once 
hyperglycemia was established, 2,000 IEQ of islets from Dynabeads infused and un-infused 
pancreas were transplanted into the renal subcapsular space of diabetic nude mice and observed 
for 30 or more days. After 30 days the mice were nephrectomized to ensure that blood glucose 
returned to diabetic levels, after which the mice where euthanized. 
4.5  Results and Discussion 
4.5.1  Cross-over studies of QMS 
 The presence of unlabelled magnetic particles in the positive fraction (b) of the output is 
called non-specific cross over.  Hydrodynamics at the edges of splitters, channel or splitter 
imperfections, hydrodynamic instabilities and shear-induced diffusion are some phenomena that 
contribute to non specific crossover (Williams et al., 2003; Williams et al., 2003a; Williams et 
al., 2008).  As the islets count for only 2% of whole pancreas, a small amount of crossover can 
obviously change the purity of isolated islets.  Crossover studies were conducted using only 
nonmagnetic acinar tissue. Acinar tissue from a pig pancreas, ?CTS1?, collected from the 
negative QMS fraction was centrifuged and re-suspended into 4 liters HBSS buffer and 500ml of 
sample was used for each run. The effect of different total flow rates, inlet and outlet flow ratios 
and concentration of tissue in the sample was tested. Peristaltic pumps (Watson-Marlow) were 
used to pump the sample and carrier liquid through the flow channel and to control positive 
outlet flow rate.  The negative (a) outlet was held open at atmospheric pressure to balance the 
flow rates.  Turbidity sensors connected to positive and negative outlets were used to measure 
the quantity of tissue in each outlet.  From the resulting absorbance data non specific crossover 
was calculated using the following equation. 
 93
ba
b
b NN
NS
+=      (4.4) 
Where Nb is the amount of tissue in the b outlet and Na is the amount of tissue in the a outlet. The 
amount of the tissue in the negative fraction was calculated with the following relation 
ba
a
a NN
NS
+= .     (4.5) 
The total volumetric flow rates used for the crossover experiments are presented in Table 4.1.  
Experiments were conducted with different inlet and outlet flow rate ratios and to test the effect 
of the amount of the tissue in the sample at constant total flow rate of 400ml/min. Figure 4.2 
shows the relative amount of tissue in the negative and positive fractions at different outlet flow 
ratios for constant inlet flow ratio of 0.25 at a total flow rate of 400ml/min. It shows the decrease 
in the crossover with the increase in outlet flow ratio. Increasing the outlet flow ratio increases 
the transport lamina thickness (Figure 4.1) which helps to reduce the crossover. Figure 4.3 shows 
the crossover of the acinar tissue at different inlet flow rate ratios at constant outlet flow ratio 
and constant total flow rate. Crossover also increases with the increase in the inlet flow ratio. 
Increasing the inlet flow ratio at constant total flow rate moves the inner splitting surface away 
from the core towards outer wall which moves the outer splitting surface away from the splitter 
causing more crossover. Figure 4.4 shows the crossover at constant total flow rate, inlet and 
outlet flow ratios with the change in the amount of the tissue in the sample. Increasing sample 
concentration results in an increase in the amount of tissue that enters in between the core and 
ISS and has to cross the OSS to reach the positive fraction. crossover with the increase in the 
concentration of the tissue in the sample, interactions between tissue fragments plays a role in the 
crossover.   
 
 94
Table 4.1: Total flow rates used in crossover studies and corresponding ISS and OSS values.    
Flow rate Q 
(ml/min) 
Rin Rout 
Inner Splitting Surface 
(rISS, cm) 
Outer Splitting Surface 
(rOSS, cm) 
400 
400 
400 
400 
400 
400 
350 
0.25 
0.25 
0.25 
0.25 
0.3 
0.4 
0.25 
0.4 
0.5 
0.6 
0.7 
0.6 
0.6 
0.6 
2.698 
2.698 
2.698 
2.698 
2.716 
2.649 
2.698 
2.749 
2.781 
2.813 
2.846 
2.813 
2.813 
2.813 
 
 
 
 
 
 
 95
0.3 0.4 0.5 0.6 0.7
0.0
0.2
0.4
0.6
0.8
1.0
S b
, S
a
Qa/Q
 Sb
 Sa
 
 
 
Figure 4.2: Crossover of the acinar tissue at different outlet flow ratios for constant inlet flow 
ratio and total flow rate.  The relative amount of tissue (ordinate) was calculated using equations 
(4.4) and (4.5).  See Table 4.1. 
 
 
 
 96
0.20 0.25 0.30 0.35 0.40 0.45
0.0
0.2
0.4
0.6
0.8
1.0
1.2
S b
, S
a
Qa/Q
 Sb
 Sa
 
 
 
Figure 4.3: Crossover of the acinar tissue at different inlet flow ratios for constant outlet flow 
ratio and total flow rate. 
 
 
 
 
 97
4 6 8 10 12 14 16
0.0
0.2
0.4
0.6
0.8
1.0
1.2
S b
, S
a
Tissue Concentration, gm/liter
 Sb 
 Sa  
 
 
Figure 4.4: Crossover of the acinar tissue at different tissue concentrations at constant inlet flow 
ratio, outlet flow ratio and total flow rate. 
 
 
 
 
 
 
 
 
 
 
 
 
 98
4.5.2  Isolation of Islets 
A total of five pig pancreata were infused with Dynabeads and processed through QMS 
for purification of islets. In the case of the first three pancreata, P808, P810 and P685, only the 
splenic lobe was infused with the Dynabeads and for the other two pancreata, P686 and P687, the 
whole pancreas was infused with Dynabeads. In all cases splenic lobe was digested for QMS 
isolation and the other two (Duodenal and Connecting) lobes were separately digested for 
standard COBE isolation. After digesting the P808 pancreas, tissue was centrifuged and re-
suspended into 1L solution and split into two halves and processed at different flow rates. Tissue 
from the other four pancreata was send through QMS straight from the Ricordi chamber by 
maintaining the flow rates of QMS and Ricordi chamber collection pump equal. Total flow rate, 
inlet flow ratio and outlet flow ratios were changed for each run to study the effect on recovery 
and purities. Table 4.2 presents the flow rates and flow ratios studied. Even though QMS was 
designed to collect the magnetic islets in the positive (?b?) fraction, Dynabeads infused into the 
islets made them sufficiently magnetic so that infused islets reached the wall of the flow channel. 
Islets were collected from the flow channel wall after sending total tissue through the QMS by 
carefully removing the flow channel from the magnet assembly.  Manual islet counts using 
dithizone staining was used to asses the purity, yield and total IEQ of the islets for QMS and 
COBE isolated fractions. Table 4.2 shows the results from all isolations. This shows the 
differences between COBE and QMS isolated islets based on the IE. 
 
 
 
 
 99
Table 4.2: Purity and yields of the islets from all the experiments conducted with four porcine 
pancreata at the University of Minnesota. 
 
Isolation Condition IE / gm (yield) Total IE based on DNA 
P685 Conn/Duo COBE 1,772 26,760 
P685 Splenic QMS 479 421,942 
P686 Conn/Duo COBE 804 33,625 
P686 Splenic QMS 277 85,274 
P687 Conn/Duo COBE 1,256 60,913 
P687 Splenic QMS 160 226,713 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 100
4.5.3  Magnetophoretic Mobility Measurements 
Samples from the digested tissue and all three fraction from QMS were run through 
MPTV to check the magnetophoretic mobilities of the magnetic bead infused islets. These tests 
helped to confirm the presence of magnetic tissue in the negative fractions or the presence of 
nonmagnetic tissue in the isolated wall fraction.  Figure 4.5 shows the histograms developed by    
MPTV for tissue sample from Ricordi chamber before sending it through QMS for isolation. 
Critical mobilities required for islets to reach the positive fraction of the QMS output are m0 = 
2.91 x 10-12 m3/T-A-s (minimum value of mobility required for islets to cross transport lamina) 
to m1 = 1.257X10-11 m3/T-A-s (maximum value of mobility required for islets to cross transport 
lamina). Islets with mobility less than m1 will end up in the negative fraction and the islets with 
mobility more than m1 will reach and stick to the wall during isolation. Peaks at magnetophoretic 
mobility less than 1e-12 are the result of the sedimentation of tissue in the MPTV cell. The 
mobility histogram shows small quantities of tissue with mobilities greater than m1 which is due 
to pancreas consists of around 98% of unlabeled acinar tissue.  Figure 4.6 is the mobility 
histogram for the negative fraction and confirms the absence of magnetic tissue in the negative 
outlet fraction.  Figure 4.7 is the histogram for the positive fraction and shows islets with small 
mobility, which is not enough to reach the wall of the flow channel. Figure 4.8 is the histogram 
for the wall fraction with 65% purity and shows higher mobility values. MPTV can not 
differentiate the islets and acinar tissue but it helped to confirm the amount of magnetic tissue in 
the samples from all fractions. 
 
 
 101
1E-15 1E-14 1E-13 1E-12 1E-11 1E-10 1E-9 1E-8
0
1
2
3
4
Ti
ssu
e f
ra
cti
on
Mobility, mm3/T-A-s
 
 
 
Figure 4.5: Magnetophoretic mobility histogram of digested tissue sample taken before sending 
it through QMS for isolation.                         
 
 
 
 
 102
1E-15 1E-14 1E-13 1E-12 1E-11 1E-10 1E-9 1E-8
0
1
2
3
4
Ti
ssu
e F
ra
cti
on
Mobility, mm3/T-A-s
 
 
 
Figure 4.6: Magnetophoretic mobility histogram of digested tissue sample taken from the 
Negative (a) fraction of a QMS isolation. 
 
 
 
 
 103
1E-15 1E-14 1E-13 1E-12 1E-11 1E-10 1E-9 1E-8
0
1
2
3
4
5
6
7
Ti
ssu
e F
ra
cti
on
Mobility, mm3/T-A-s
 
 
 
Figure 4.7: Magnetophoretic mobility histogram of digested tissue sample taken from the 
Positive (b) fraction of a QMS isolation. 
 
 
 
 
 104
1E-15 1E-14 1E-13 1E-12 1E-11 1E-10 1E-9 1E-8
0
1
2
3
4
5
Ti
ssu
e F
ra
cti
on
Mobility, mm3/T-A-s
 
 
 
Figure 4.8: Magnetophoretic mobility histogram of digested tissue sample taken from wall 
fraction of QMS isolation. 
 
 
 
 
 
 
 
 
 
 
 
 
 105
4.5.4  Histology of Pancreas 
Hanging bag or hand-syringe infusion techniques were used to infuse the beads into the 
pancreas with varying concentration of magnetic particles. T2*-weighted MRI illustrated a 
uniform distribution of hypointense regions, indicating the presence of magnetic particles 
throughout the experimental splenic lobe in all organs (Figure 4.9). Histologic analysis confirms 
that magnetic particles were found predominantly within islet microvasculature, with few present 
in surrounding acinar tissue as distinguished by hematoxylin and eosin (H/E) staining (Figure 
4.10). Hand-syringe infusion technique achieved good infusion of magnetic particles into the 
islets.  Further details are given in a separate publication (ref. Rizzari paper). 
4.5.5  Quality assessment of Islets 
Quality assessment studies were conducted on QMS and COBE isolated islets. Purity and 
yield of the islets from P808 isolation was too low to allow quality tests, so quality assessment 
was done for the other four isolations. Figure 4.11 shows the pictures of the islets taken 
immediately after isolation using QMS and COBE method. Islets isolated with COBE method 
appear irregular in shape and fragile, whereas the islets isolated using QMS appear very solid wit 
smooth surfaces. Pressure applied to the islets and prolonged exposure to the digestion enzymes 
during the COBE separation damage the small islets causing them to fragment into small tissue 
while the smooth (low-shear) flow of islets through the QMS flow channel helps islets to 
maintain the solid structure and minimizes the time islets are exposed to digestion enzymes. 
 
 106
  
Figure 4.9: MRI of the control connecting/duodenal lobes (above) and the experimental splenic 
lobe (below), in which infused magnetic beads resulted in well-distributed hypointense regions. 
 
 
 
 
 
 
 
Connecting/Duodenal Lobes 
Splenic Lobe 
 
 
 107
 
Figure 4.10: Representative low and high magnification micrographs of an islet located in the 
experimental splenic lobe (distal region), illustrating minimal accumulation of magnetic particles 
in the acinar tissue (Fig. 4.10A, H/E) and significant accumulation within the islet (Fig. 4.10B, 
H/E) and magnetic particles within capillaries of an islet located in the proximal splenic lobe, 
near the site of infusion at the celiac trunk (Fig. 4.10C, insulin stain). 
 
 
 
 
 
 
 
 
 
B 
 M 
C 
 
M
 
20 
A Acinar tissue 
 
Islet 
60 20   
 
   
  
 
 108
 
Figure 4.11: Pictures of the dithizone-stained islets taken under microscope at 20X magnification 
a) islets isolated with QMS b) islets isolated with COBE. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a b 
 109
Figure 4.12 shows the OCR per DNA (nmol min-1 mg-1 DNA) for COBE and QMS 
isolated islets. These results for day 0 clearly show the difference between QMS and COBE 
isolated islets. OCR per DNA for COBE isolated islets on day 0 was less than the standard 
required value of 150 nmol min-1 mg-1 DNA to transplant into the mouse for in vivo testing 
whereas the QMS isolated islets have high enough OCR to transplant. COBE and QMS isolated 
islets were cultured for 7 days and checked for OCR and the resulting increased OCR values for 
COBE and QMS isolated islets both exceeded the required standard values. Figure 4.13 shows 
the average values of the OCR/DNA for COBE and QMS isolated islets at day 0 and day 7. 
QMS isolation can reduce or eliminate culture time and helps to eliminate the mechanical stress 
due to centrifugation on the islets.  
Islets were evaluated to check their ability to reverse diabetes in mice. These studies help 
to study the effect of the presence of Dynabeads on the islets? functionality. Immunodeficient 
diabetic mice were transplanted with 2000IEQ of islets from COBE and QMS isolation from the 
same pancreas. Histological evaluation confirmed no inflammatory reaction in the murine 
kidneys due to the presence of the Dynabeads and diabetes reversal was achieved in the mice 
with Dynabead infused islets. 
 
 
 
 
 110
P865 P866 P867
0
50
100
150
200
250
300
OC
R/
DN
A
 COBE
 QMS
 
 
 
Figure 4.12: Stimulation index for OCR measurements from control and infused islets from three 
separate isolations. 
 
 
 
 
 111
day0 day7
0
50
100
150
200
250
300
OC
R/
DN
A
 COBE
 QMS
 
 
 
Figure 4.13: Stimulation index for average OCR measurements from control and infused islets 
from three separate isolations at day0 and day7. 
 
 
 
 
 
 
 
 
 
 
 
 112
Based on the MPTV tests and visual observations it is confirmed that the infusion of the 
Dynabeads into the islets was not 100% successful. Even though Dynabeads were selected based 
on their size and the vascular structure differences between islets and acinar tissue, not all islets 
were infused with beads, and some of the acinar tissue also received magnetic beads. Obtaining 
homogeneity of bead infusion is important to isolate islets using QMS, and more studies are 
ongoing at University of Minnesota to further the understanding of the anatomy of porcine 
pancreata, improvement of procurement, flushing and homogeneous perfusion of the organ 
through the vasculature for optimal bead infusion, effect of pressure of infusion, different bead 
concentration and also considering the option of other beads with less magnetophoretic mobility 
and optimal size.  
When digested tissue was first centrifuged and suspended in buffer prior to running 
through QMS during the P808 isolation, the tissue formed large clumps and clogged the QMS 
inlet port and reduced the purity of islets produced by a great percentage. By sending the tissue 
straight from the Ricordi chamber to QMS these problems were eliminated, and the time islets 
were exposed to digestion enzymes was greatly reduced thereby avoiding over digestion and 
fragmentation problems. When islets stick to the wall of the QMS flow channel they are exposed 
to fresh buffer with serum and not the digestion fluid. Even though the QMS flow channel was 
designed to collect the bead infused islets in the positive fraction, the higher magnetophoretic 
mobilities of the Dynabeads forced us to collect the infused islets on the wall of the flow channel 
and collect them after sending all tissue through QMS. Some acinar tissue sedimented at the 
bottom of the flow channel and it reduced the purity of the islets collected from wall. Some small 
modifications in the flow channel design can also increase the purity of the islets.  
 
 113
4.6  Conclusions 
QMS successfully isolated the Dynabead infused islets immediately after islets were 
liberated from the pancreas in a continuous process. Visual observations through microscope and 
quality assessment of the isolated islets confirmed the advantage of the QMS isolation over the 
standard COBE method. Achieving the homogeneous infusion of the magnetic beads into the 
islets and by implementing the small changes in the flow channel design it is possible to make 
the QMS a successful instrument to isolate superior quality porcine or human islets in less time 
than required by the available standard methods.   
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 114
4.7  References 
Davies JE, James RF, London NJ, Robertson GS. 1995. Optimization of the magnetic field used 
for immunomagnetic islet purification. Transplantation  59(5):767. 
Davies JE, Robertson GS, Swift S, Chamberlin J, Bell PR, James RF, London NJ. 1994. 
Transplatation Proceedings 26(2):649-650. 
Davies JE, chamberlain JC, Swift S, James RF, London NJ, Robertson GS. 1997. The use of 
immunomagnetic separation for secondary purification of pancreatic islets. A comparison 
of different magnetic fields in the rat. Adv Exp Med Biol 426:435-440. 
Ferrer J, Scott WE, 3rd, Weegman BP, Suszynski TM, Sutherland DE, Hering BJ. 2008. Pig 
pancreas anatomy: implications for pancreas procurement, preservation, and islet 
isolation. Transplantation 86(11):1503-1510. 
Frank A, Deng S, Huang X, Velidedeoglu E, Bae YS, Liu C, Abt P, Stephenson R, Mohiuddin 
M, Thambipillai T, Markmann E, Palanjian M, Sellers M, Naji A, Barker CF, Markmann 
JF. 2004. Transplantation for type I diabetes: comparison of vascularized whole organ 
pancreas with isolated pancreatic islets. Ann Surg. 240:631. 
Froud T, Ricordi C, Baidal DA, Hafiz MM, Ponte G, Cure P, Pileggi A, Poggioli R, Ichii H, 
Khan A, Ferreira JV, Pugliese A, Esquenazi VV, Kenyon NS, Alejandro R. 2005. Islet 
transplantation in type 1 diabetes mellitus using cultured islets and steroid-free 
immunosupression: Miami experience. Am J Transplant. 5:2037. 
Gauthier VJAMM. 1988. A method for isolation of mouse glomeruli for quantitation of immune 
deposits. Kidney Int. 33:897. 
 115
Hering BJ, Kandaswamy R, Ansite JD, Eckman PM, Nakano M, Sawada T, Matsumoto I, Ihm 
SH, Zhang HJ, Parkey J, Hunter DW, Sutherland DE. 2005. Single donor, marginal-dose 
islet transplantation in patients with type 1 diabetes. JAMA 293:830. 
Hering BJ, Matsumoto I, Sawada T, Nakano M, Sakai T, Kandaswamy R, Sutherland DER. 
2002. Impact of two-layer pancreas preservation on islet isolation and transplantation 
Transplantation 74:1813-1816.  
Hering BJ, Wijkstrom M, Graham ML, H?rdstedt M, Aasheim TC, Jie T, Ansite JD, Nakano M, 
Cheng J, Li W, Moran K, Christians U, Finnegan C, Mills CD, Sutherland DE, Bansal-
Pakala P, Murtaugh MP, Kirchhof N, Schuurman HJ. 2006. Prolonged diabetes reversal 
after intraportal xenotransplantation of wild-type porcine islets in immunosuppressed 
nonhuman primates. Nat Med 12 (3):301-303. 
Hoyos PS, Decker K, Nakamura M, Chalmers JJ, Moore LR, Zborowski M. 2003. Splitter 
imperfections in annular split-flow thin separation channels: experimental study of 
nonspecific crossover. Anal Chem 75:6687. 
Jing Y, Chalmers JJ, Zborowski M. 2007. Blood progenitor cell separation from clinical 
leukapheresis product by magnetic nanoparticle binding and magnetophoresis. 
Biotechnol Bioeng 96(6):1139. 
Kennedy DJ, Todd P, Logan S, Becker M, Papas KK, Moore LR. 2007. Engineering quadrupole 
magnetic flow sorting for the isolation of pancreatic islets. Journal of Magnetism and 
Magnetic Materials 311:388-395. 
London NJ, Robertson GS, Chadwick DR, Johnson PR, James RF, Bell PR. 1994.Human 
pancreatic islet isolation and transplantation. Clin transplant 8(5):421-459. 
 116
Mccloskey KE, Moore LR, Hoyos M, Rodriguez A, Chalmers JJ, Zborowski M. 2003. 
Magnetophoretic cell sorting is a function of antibody binding capacity. Biotechnol Prog 
19:899. 
Nandigala P, Chen TH, Yang C, Hsu WH, Heath C. 1997. Immunomagnetic isolation of islets 
from rat pancreas. Biotechnol Prog. 13:844-848. 
Papas KK, Colton CK, Nelson RA, Rozak PR, Avgoustiniatos ES, Scott WE, Wildey GM, 
Pisania A, Weir GC, Hering BJ. 2007. Human islet oxygen consumption rate and DNA 
measurements predict diabetes reversal in nude mice. Am J Transplant 2007;7(3):707-
713. 
Pinkse G, Steenvoorde E, Hogendoorn S, Noteborn M, Terpstra OT, Bruijn JA, De Heer E. 
2004. Stable transplantation results of magnetically retracted islets: a novel method. 
Diabetologia 47:55. 
Ricordi C, Lacy PE, Finke EH, Olack BJ, Scharp DW. 1998. Automated method for isolation of 
human pancreatic islets. Diabetes 37:413. 
Samejima T, Yamaguchi K, Iwata H, Morkawa N, Ikada Y. 1998. Gelatin density gradient for 
isolation of islets of Langerhans. Cell Transplant 7(1):37- 45. 
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):289-290. 
Shenkman RM, Chalmers JJ, Hering BJ, Kirchhof N and Papas KK. 2009. Quadrupole Magnetic 
Sorting of Porcine Islets of langerhans, Tissue Engineering: Part C 15(2):147-156. 
 117
Shenkman RM, Godoy-Silva R, Papas KK and Chalmers JJ. 2009a. Effects of Energy 
Dissipation Rate on Islets of Langerhans: Implications for isolation and Transplantation, 
Biotechnol Bioeng 103(2):413-423. 
Sun L, Zborowski M, Moore LR, Chalmers JJ. 1998. Continuous, Flow-Through 
Immunomagnetic Cell Separation in a Quadrupole Field. Cytometry 33:469. 
Tong X, Xiong Y, Zborowski M, Farag SS and Chalmers JJ. 2007. A Novel High Throughput 
Immunomagnetic Cell Sorting System for Potential Clinical Scale Depletion of T Cells 
for Allogeneic Stem Cell Transplantation. Exp Hematol 35(10):1613. 
T?ns HAM, Baranski AG, Terpstra OT, Bouwman E. 2008. Isolation of the Islets of Langerhans 
from the human pancreas with magnetic rectraction. Transplant Proc. 40:413. 
Williams PS, Zborowski M, Chalmers JJ. 1999. Flow Rate optimization for the Quadrupole 
Magnetic Cell Sorter. Anal Chem 71:3799. 
Williams PS, Moore LR, Chalmers JJ, Zborowski M. 2003. Splitter Imperfections in Annular 
Split-Flow Thin Separation Channels: Effect on Nonspecific Crossover. Anal Chem 
75:1365-1373. 
Williams PS, Decker K, Nakamura M, Chalmers JJ, Moore LR, Zborowski M. 2003a. Splitter 
Imperfections in Annular Split-Flow Thin Separation Channels: Experimental Study of 
Nonspecific Crossover. Anal Chem 75:6687-6695. 
Williams PS, Hoyos M, Kurowski P, Salhi D, Moore LR, Zborowski M. 2008. Characterization 
of Nonspecific Crossover in Split-Flow Thin Channel fractionation. Anal Chem 80:7105-
7115. 
Winoto-Morbach, Ulrichs K, Hering BJ, Leyhausen G, Muller-Ruchholz W. 1989. Methods in 
Islet Transplantation Research. Hormone and Metabolic Research supplement 25:51-54. 
 118
5. Summary and Future Work 
The dissertation has successfully demonstrated that Quadrupole Magnetic Flow Sorter 
can be used for isolation of pancreatic islets from exocrine tissue using magnetic beads infusion 
into islets. Differences in the vascular structure of islets from acinar tissue can be used for 
successful infusion of Dynabeads? only into islets. QMS isolated islets are better in quality 
because of the less stress applied by QMS and the reduction of the exposure time of the islets to 
the digestive enzymes. CFD simulations can be used to test the performance of the QMS flow 
channel and can be used as a tool to improve the efficiency of the flow channel to achieve good 
isolation with high purity. MPTV developed to quantify the magnetophoretic mobility of the 
labeled tissue can be used as a tool to optimize the parameters of the QMS to isolate labeled 
islets from exocrine tissue. 
CFD simulations of flow pattern performed on the new design of QMS flow channel 
confirm circumferentially uniform flow development around the annular channel.  Quantitative 
agreement between experimental measurements of nonspecific crossover and prediction based on 
CFD modeling of the fluid flow was shown.  For all flow conditions, crossover predicted by 
CFD simulations was found to be slightly lower than experimentally observed results.  This 
difference may be due to the contribution of crossover from other factors such as particle lift and 
particle interactions. Diffusion was not considered in the CFD modeling.  The good agreement 
between experimental and CFD predicted results allowed the performance of simulations with 
different channel prototype models to develop a design to minimize nonspecific crossover. 
 119
The capability of newly developed MPTV in measuring magnetophoretic mobility was 
confirmed with the measurements of the mobility of the standard Dynabeads. MPTV was 
successfully used to measure the mobility of magnetic particles up to 500 microns in size with 
different mobilities. MPTV?s ability to predict the optimized flow parameters was also tested 
successfully with magnetic particles and the isolated islets of Langerhans as it is the final 
application of the MPTV.   
The present research has opened up several avenues for continuing the fundamental and 
development research works on magnetic islet isolations. Following research area may be further 
explored to support the future islet transplantation studies. 
1. Computational Fluid Dynamics study of the flow of the magnetic particles in the Quadrupole 
Magnetic Sorter flow channel using magnetic models in the FLUENT, which along with 
nonspecific crossover studies helps to design the magnet and flow channel assembly for 
QMS to achieve the best isolation purity with the collection of acinar tissue in the negative 
fraction and pure islets in positive fraction. Flow channel can be modified to avoid the 
clogging and settling of the tissue in the flow channel. 
2. Development of the MPTV further to track only the moving particles in the flow channel by 
omitting the disturbances on the flow channel and make it possible to connect the MPTV 
online to the flow process of the Ricordi digestion process to get the QMS parameters for 
successful isolation of islets. 
3. Further study of the magnetic beads infusion process into the islets to achieve uniform and 
enough beads infusion in the whole pancreas.  
   
 120
Appendix A: QMS Operation 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
  
 
Figure A.1: Photograph of the QMS setup with fluid bags.
 121
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
   
 
     
 
 
 
 
Figure A.2: Components of the QMS. Boxed labels are permanent QMS components, unboxed 
labels are consumable items replaced after one use 
 
 
 
 
Negative Collection  
Priming Buffer  
Bottom Waste  
Positive Islet Collection  
Infused Islet Suspension Top Waste Carrier Solution 
Upper Tubing Set 
Lower Tubing Set 
Flow Channel  
with Integral Tubing Set 
Carrier Solution Pump  
Sample Pump 
Positive Collection Pump  
Power ON/OFF 
(button and status light) 
Emergency OFF Button 
WARNING Quadrupole Magnet  
Enclosure 
Pinch Shutoff Valves 
(9 total) 
Flow Transmittance Sensors 
Pump Flow Direction 
Indicator Light 
Caution Light and Alarm 
Leveling Feet (4 places) Bubble Level 
Internal Tubing 
Replaced After One 
Use 
Internal Tubing 
Replaced After One 
Use 
Internal Tubing 
Replaced After One 
Use 
 122
 
 
Figure A.3: Screnshot of the QMS software used to control the QMS operation. 
 
 
 
 
 
 
 
 
 
 123
 
 
 
Figure A.4: QMS set up for pancreas isolation at University of Minnesota laboratory. 
 
 
 
 
 
 
 
 
 124
  
 
Figure A.5: Absorbance detector traces for P810 isolation run.  The QMS was turned on at time 
t=0 minutes.  The turbidity sensor data logging was activated at t=42 minutes [1].  The digestion 
process was switched to recovery mode and the QMS began operations, with tissue leaving the 
flow channel at t=48 minutes [2].  The tissue concentration with the Ricordi chamber upright 
peaked at approximately t=72 minutes [3].  The Ricordi chamber was inverted to maximize the 
recovery at t=92 minutes [4].  The tissue exiting the inverted Ricordi chamber peaked at t=102 
minutes [5].  The isolation was terminated at approximately t=114 minutes [6].  
 125
 
 
 
 
 
 
 
             
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Table A.1: Summery of the Porcine isolation experiments. 
Date Exp no Q Qa'/Q Qa/Q Condition 
2/26/2009 VC-02-1 224 0.25 0.25 flow Channel outside the magnet, no bead infusion 
2/26/2009 VC-02-1A 224 0.25 0.25 flow Channel outside the magnet, no bead infusion 
2/26/2009 VC-02-2 224 0.25 0.35 flow Channel outside the magnet, no bead infusion 
2/26/2009 VC-02-2A 224 0.25 0.35 flow Channel outside the magnet, no bead infusion 
2/26/2009 VC-02-3 224 0.25 0.45 flow Channel outside the magnet, no bead infusion 
2/26/2009 VC-02-3A 224 0.25 0.45 flow Channel outside the magnet, no bead infusion 
4/24/2009 VC-04-1 400 0.25 0.5 Flow channel in the magnet, no bead infusion 
4/24/2009 VC-04-2 400 0.25 0.6 Flow channel in the magnet, no bead infusion 
4/24/2009 VC-04-3 400 0.25 0.7 Flow channel in the magnet, no bead infusion 
4/24/2009 VC-04-4 400 0.25 0.7 Flow channel in the magnet, no bead infusion 
6/11/2009 VC-06-1 400 0.25 0.7 2ml dynabeads infused into pancreas 
1/5/2010 P808 - T1 400 0.25 0.5 4ml of dynabeads in splenic lobe 
1/5/2010 P808 - T2 400 0.25 0.7 splenic lobe tissue resuspended into 1L 
1/5/2010 P808 - C1 400 0.25 0.5 
Control lobes(duodenal and connectiong lobe without 
beads) 
1/5/2010 P808 - C2 400 0.25 0.7 only for cross over studies, tissue obtained from cobe  
1/7/2010 P810 400 0.25 0.7 4ml of dynabeads in splenic lobe 
          
Tissue was send to QMS staright from Ricordi 
chamber 
5/25/2010 P865 400 0.25 0.6 4ml dynabeads infused into splenic 
5/26/2010 P866 400 0.4 0.6 4ml dynabeads infused into pancreas 
5/27/2010 P867 400 0.43 0.6 4ml dynabeads infused into pancreas 
6/7/2010 CTS-1 400 0.43 0.6 4ml dynabeads infused into whole pancreas 
6/9/2010 CTS-2 400 0.3 0.6 4ml dynabeads infused into whole pancreas 
 126
A.1  QMS Operation Procedure 
 
1. Install pump tubing, tubing set and flow channel. 
2. Connect fluid and empty bags to the tube set. 
3. Prime the flow channel and tubes.  
4. Calibrate the pumps. 
5. Set the flow rates using QMS islets software. 
6. Start turbidity sensors to get the absorbance data. 
7. Use the QMS software to run QMS in auto mode. 
8. Collect the positive and negative fractions. 
9. Collect wall fraction by removing flow channel from magnet. 
10. Drain the tubing and clean. 
 
 
 
 
 
 
 
 
 
 
 127
Appendix B: MPTV Operation 
 
 
Figure B.1: Photograph of the MPTV assembly. 
 128
 
Figure B.2: Photograph of the MPTV set up. 
 
 
 
 
 
 
 
 129
 
Figure B.3: Screeshot of the IKOVISION used to run the MPTV.
 130
Appendix C: Fluent Simulations 
C.1  Fluent Simulation Procedure  
?  Start the program from the command line by typing ?fluent 3ddp?. 
?  Open the .msh file with the pull-down menus File barb2right Read barb2right Case.  
?  Save the loaded .msh file as a .cas file by the pull-down menus: File barb2right Write barb2right 
Case. 
?  Under Grid barb2right Scale make sure that the domain extents in meters are correct. If not 
specify that the grid was created in inches. 
?  Under Define barb2right Materials define the ?Fluid Material? as liquid water by loading it 
from the database or finding it in the pull-down menu.  
?  Under Define barb2right Boundary Conditions set Fluid to liquid water, Outlet to zero gauge 
pressure, and Inlet to the appropriate Velocity Magnitude (calculated from the volumetric 
flow rate and cross-sectional area at the channel inlet). In this case 00 ml/min gave a 
velocity magnitude of 0.42 m/s. 
? Under Define barb2right Injections create ?Surface? inert injections released from the channel 
Inlet.  
?  Under Define barb2right Units use the SI default. 
?  Accept the default Define barb2right Solver options. 
?  Under Injection barb2right Define Inert material properties. 
?  Under Define barb2right Models barb2right Viscous make sure the regime is Laminar.
 131
? Under Define barb2right Models barb2right Discrete Model. 
? Under Solve barb2right Monitors barb2right Residual set the Convergence Criterion to at least 1E-6 
and the Iteration Storage to at least several thousand. Make sure the Residual Monitor 
appears on the screen. 
? Under Solve barb2right Initialize set the X velocity to the same value as the calculated 
inlet X-Velocity Magnitude. 
? Under Solver barb2right Controls barb2right Solution use the SIMPLE Pressure-Velocity Coupling, 
Standard Pressure Discretization, and Second Order Upwind Momentum Discretization. 
The Under-Relaxation Factors (except Momentum) can be left at the default values. 
?  Start the solution from Solver barb2right Iterate with at least several thousand iterations 
and reporting after every ten. 
?  A simulation can be expected to run for several hours. If after several hours the Residuals 
remain flat you may halt the iteration and reduce the Momentum Relaxation Factor from 
the Solver barb2right Controls barb2right Solution menu then restart the iteration. The Momentum 
Relaxation Factor may have to be reduced several times. 
?  When the residuals have converged, run the Discrete model using display particle 
trajectories and select summery. 
?  Save the data from the File barb2right Write barb2right Case and Data menu.  
?  The results may be seen from the many option in the Display pull-down menu. 
 
 132
 
Figure C.1: Summery of the particle flow simulation.
 133
 
Figure C.2: Particles trajectories calculated in Fluent simulations.