Lanthanide-Encoded Polystyrene Microspheres for Mass Cytometry-Based Bioassays

LANTHANIDE-ENCODED POLYSTYRENE MICROSPHERES FOR MASS CYTOMETRY-BASED BIOASSAYS

Ahmed I. Abdelrahman

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Graduate Department of Chemistry University of Toronto

Abstract

Lanthanide-Encoded Polystyrene Microspheres for Mass Cytometry-Based Bioassays

Ahmed I. Abdelrahman, Ph.D. Thesis (2011)

Department of Chemistry, University of Toronto

This thesis describes the synthesis and characterization of metal-encoded polystyrene microspheres with a narrow size distribution designed for mass cytometry-based immuno- and oligonucleotide-assays. These particles were prepared by multiple stage dispersion polymerization techniques using polyvinylpyrrolidone (PVP) as a steric stabilizer.

As a cytometeric technique, mass cytometry necessitated metal-encoded microspheres to perform the same roles of fluorescent microspheres used in conventional flow cytometry. The first role of the microsphere was to be able to act as a platform (classifier microspheres) for bioassays. Secondly, the microspheres should be suitable for mass cytometry machine calibration as standards. To perform these roles, metal-encoded microspheres were required to have certain size, functionality and metal content criteria. Lanthanide elements were chosen as the metals for encoding the microspheres for their low natural abundance in biological systems and for their similar chemistry.

My goal was to employ two-stage dispersion polymerization, of styrene in ethanol, to introduce the lanthanide salts along with excess acrylic acid in the second stage, one hour after the initiation. Acrylic acid deemed to serve as a ligand for the lanthanide ions, through its carbonyl group, so the lanthanide ions get incorporated into the microsphere while acrylic acid is copolymerizing with styrene. Using two-stage dispersion polymerization, I could synthesize

lanthanide encoded microspheres with narrow size distribution and high lanthanide content.

However the lanthanide content distributions were unexpectedly much broader than the size distribution obtained. In addition, I could not attach biomolecules to the surface of such particles.

In an attempt to improve the characteristics of these microspheres, I employed modified

versions of multiple stage dispersion polymerization and seeded emulsion polymerization to

grow functional polymer shell on the surface of the particles prepared by dispersion

polymerization. Moreover, I coated the lanthanide encoded microspheres with silica shell which

enabled me to grow another layer of functional-silica. Consequently, I could use these particles

as classifier microspheres for mass cytometry-based immunoassays as well as fluorescence-

based oligonucleotide-assays.

iii

ACKNOWLEDGEMENTS

I would like to express my gratitude and appreciation to my supervisor Professor Mitchell A. Winnik for his support, encouragement and guidance throughout this research work. I am also grateful for the opportunity he gave me to be a part of the exciting lanthanide tags project. I gratefully acknowledge and thank Professor Vladimir Baranov, Dr. Dmitry Bandura, Dr. Olga Ornatsky and Professor Scott Tanner for the valuable suggestions, discussions and collaborations.

I would like to thank my friends and colleagues in Winnik group. Particularly, I want to thank Dr. Sheng Dai for giving me a lot of help when I started my research project in Toronto, and Dr. Gerald Guerin and Dr. Conrad Siegers for the very helpful discussions. I owe a debt of gratitude to Dr. Stuart Thickett, Dr. Dirk Weinrich, Yi (Sally) Liang and Wanjuan (Betty) Lin for the fruitful collaborations. It was very enjoyable working with them. I especially thank Stuart for his time and valuable critique of my thesis. Many thanks go to Sally for her important contribution to the success of our project. To my desk-neighbors, Daniel Majonis, Peng Liu and Dr. Yi Hou, thanks for interesting discussions on science.

I feel a deep appreciation and gratefulness for my family members in Egypt (my mother, my brother Hany and two sisters Amel and Sara) for their encouragement and unconditional love. This thesis was made possible by their full support and encouragement and is dedicated to them. To my beloved mother: without your prayers, I would not be able to get this thesis done.

I would like to express my love and gratitude to my wife Noha, who has always lovingly been by my side through times of frustration and joy. Heartfelt thanks to my lovely kids Ziad, Rofaida and Jana.

Before and after all, To Allah Almighty, who endowed me with the strength, enthusiasm and perseverance necessary to see this project through, and to whose glory this work is dedicated.

Not even the weight of an atom

or less than that or greater,

escapes from His Knowledge

in the heavens or in the earth.

………………………………Holy Quraan 34:3

Dedicated to the soul of my dad

TABLE OF CONTENT

ABSTRACT ii ACKNOWLEDGMENTS iv TABLE OF CONTENT vi LIST OF TABLES x LIST OF FIGURES AND SCHEMES xii LIST OF APPENDICES xix ABBREVIATIONS xx

1 GENERAL INTRODUCTION 1

1.1 Polymer Particles by Dispersion Polymerization 4 1.1.1 Dispersion Polymerization: Mechanism and Kinetics 5 1.1.2 Different Processes for Dispersion Polymerization 8 1.1.2.1 Dispersion Polymerization by The Batch Process 8 1.1.2.2 Semi-Continuous, Continuous and Seeded Dispersion Polymerizations 9 1.1.2.4 Two-Stage Dispersion Polymerization 11 1.1.3 Functional Microspheres by Dispersion Polymerization 12 1.1.3.1 Microspheres with Functional Groups 13 1.1.3.2 Bioconjugation Reactions as Post-Functionalization for Microspheres 15 1.1.4 Fluorescent Dye- and Metal-Containing Microspheres by Dispersion Polymerization 18 1.1.4.1 Fluorescent Dye-Labeled Microspheres 18 1.1.4.2 Precious Metal-Containing Microspheres 19 1.1.4.3 Magnetic Polymer Microspheres 20 1.1.4.4 Polymer Microspheres Functionalized with Silica 20

1.2 Flow Cytometry 21

1.3 Lanthanides: Physical and Chemical properties 23

1.4 Research objectives 25

1.5 Thesis outline 25

References: 27

2 INSTRUMENTAL METHODS AND EXPERIMENTAL DETAILS 38

2.1 Instrumental Methods 38 2.1.1 Inductively Coupled Plasma Mass Spectrometer 38 2.1.1.1 Introduction 38 2.1.1.2 Plasma Source 39

2.1.1.3 Mass Analyzers 40 2.1.1.4 Sample Preparation for ICP-MS 40 2.1.2 Mass Cytometry 41 2.1.2.1 Introduction 41 2.1.2.2 Instrument Design 42 2.1.2.2 Sample Preparation for Mass Cytometry 44

2.2 Materials 45 2.2.1 Solvents 45 2.2.2 Reagents for Particles’ Synthesis and Characterization 45 2.2.2 Reagents for Bioconjugation 46

2.3 Synthetic Procedures 47 2.3.1 Microspheres’ Synthesis by Dispersion Polymerization 47 2.3.2 PVP pyrrolidone ring opening (PVP activation). 51 2.3.3 Dispersion polymerization using modified PVP. 52 2.3.4 Seeded Emulsion Polymerization with Methacrylic Acid (MAA). 53 2.3.5 Seeded Emulsion Polymerization with Glycidyl Methacrylate (GMA). 53

2.4 Characterization Methods 54 2.4.1 Scanning Electron Microscopy (SEM) 54 2.4.2 Dynamic Light Scattering (DLS) 54 2.4.4 Gel Permeation Chromatography 54 2.4.5 Gravimetrical measurements 55 2.4.3 Titration of Acid Groups 55 2.4.6 Bioconjugation 56 2.4.7 Fluorescence Emission 57

References: 58

3 MICROSPHERES BY TWO STAGE DISPERSION POLYMERIZATION 60

3.1 Introduction 60

3.2 Designing the Synthesis of the Microspheres 62 3.2.1 Techniques to Fulfill Size Requirements 62 3.2.2 Metal-Content Requirements 65

3.3 Results and Discussion 66 3.3.1 Synthesis of the Microspheres 66 3.3.2 Determining Lanthanide Content by Mass Cytometry 69 3.3.3 Microsphere Encoding Protocols: Variability and Dimensionality 79 3.3.4. Lanthanide Incorporation, Particle-to-Particle Variability and Surface Functionality 84

3.4 Covalent Attachment of Proteins to the Surface of the Microspheres 87

3.5 Summary 91

vii

References 92

4 MICROSPHERES BY THREE STAGE DISPERSION POLYMERIZATION 94

4.1 Introduction 94

4.2 Three Stage Dispersion Polymerization 96

4.3 Factorial design 99 4.2.1 Design of the Factorial Experiments 100 4.2.2 Factorial Experiments: Results and Discussions 101

4.3 Testing Ion Leakage 110

4.4 Synthesis of Particles with Higher Variability 111

4.5 Bioconjugation 117

4.6 Summary 119

References: 121

5 SURFACE FUNCTIONALIZATION OF LANTHANIDE-ENCODED MICROSPHERES 123

5.1 Introduction 123

5.2 Results and Discussion 125

5.3 SUMMARY 143

References: 146

6 SILICA-COATED PARTICLES 147

6.1 Surface Functionalization with Silica Shell 147

6.2 Immunoassays Based on the Amino-Functionalized Particles 153 6.2.1 Streptavidin -Coated Lanthanide-Encoded Particles via Biotin- Streptavidin Sandwich (Fluorescence Assay) 153 6.2.2 Streptavidin -Coated Lanthanide-Encoded Particles via Biotin- Streptavidin Sandwich (Mass Cytometry Assay) 157 6.2.3 Streptavidin-Coated Lanthanide-Encoded Particles via covalent attachment of Streptavidin (Mass Cytometry Assay) 160

6.3 Oligonucleotide Assays-Based on the Amino-functionalized Particles 161

viii

6.5 Conclusions 166

References: 168

7 ANALYTICAL ASPECTS 170

7.1 Introduction 170

7.2 Results and Discussion 172 7.2.1. Microsphere Synthesis and Metal Incorporation 175 7.2.2. 3-Stage Dispersion Polymerization 179 7.2.3. Ion-release behavior and mass cytometry calibration standard 181 7.2.4. Lanthanide-Containing Microspheres as Internal Standards for Cell Samples. 183 7.2.5. Reproducibility of the Synthesis of Lanthanide-Containing Microspheres. 188

7.3 Conclusions 193

References 194

APPENDIX 1 198

LIST OF TABLES

1 General Introduction 1 Table 1-1: Different functional monomers used in the synthesis of polymer particles (by dispersion polymerization). 13 Table 1-2: Configuration of outer electrons of Ln atoms and Ln3+ ions 24

2 Instrumental Methods and Experimental Details 38 Table 2-1. Recipe for the synthesis of microsphere sample AA0069 as an example for 2-stage dispersion polymerization (2-DisP) of styrene with PVP55 as a dispersant in ethanol. 50 Table 2-2. Recipe for the synthesis of microsphere sample AA0105 as an example for 3-stage dispersion polymerization (2-DisP) of styrene with PVP55 as a dispersant in ethanol. 51 Table 2-3. Recipe for AA139 synthesized by dispersion polymerization of styrene with activated PVP. 52

3 Microspheres by Two Stage Dispersion Polymerization 60

Table 3-1. Recipe for the synthesis of microsphere sample AA088 as an example for 2-stage dispersion polymerization (2-DisP) of styrene with PVP55 as a dispersant in ethanol. 67 Table 3-2. Lanthanide content a of the reaction mixture for synthesis of PS-PAA microspheres (2 wt % AA/styrene) 68 Table 3-3 Particle size, size distribution, and the variation of Tm intensities for some PS-PAA microspheres synthesized in the presence of LnCl3.6H2O. 81

4 Microspheres by Three Stage Dispersion Polymerization 94 Table 4-1. The recipe for the synthesis of AA105 particle sample by 3-stage dispersion polymerization (3-DisP) of styrene with PVP55 as a dispersant in ethanol 96 Table 4.2: Size and metal content results for Samples AA105 and AA122 – AA129 synthesized by 3- DisP. 98 Table 4-3: 23 factorial design of the third stage in the multiple stage dispersion polymerizations. All other conditions for the dispersion polymerization reaction (like 1st and 2nd stage ingredients and reaction temperature) were kept constant. 100 Table 4-4. An example of a factorial design experiment: the recipe for the synthesis of AA122 by 3-stage dispersion polymerization (3-DisP) of styrene with PVP55 as a dispersant in ethanol 102 Table 4.5. The recipe for the synthesis of particle samples AA134 and AA135 by 3-stage dispersion polymerization (3-DisP) of styrene with PVP55 as a dispersant in ethanol. 114

5 Surface Functionalization of Lanthanide-coded Microspheres 123 Table 5.1. Recipe for the synthesis of functional monomer coatings onto the surface of the PS seed latex D1 by seeded emulsion polymerization. Experiments were performed with methacrylic acid (D4(MAA)) or with glycidyl methacrylate in the presence of excess surfactant (D5(GMA-2)) 133

Table 5.2. Mass cytometry results for the bioconjugation of NeutraAvidin to the surface of PS dispersion polymerization particles before further polymerization with MAA probed with a Lu-biotin reporter. 136 Table 5.2. Mass cytometry results for the bioconjugation of NeutraAvidin to the surface of PS dispersion polymerization particles before further polymerization with MAA probed with a Lu-biotin reporter. 139 Table 5.3. Reproducibility of Surface Coating Experiments Using an Excess of GMA a Seed particles (davg = 1.6 µm) synthesized by two stage dispersion polymerization. D1 refers back to sample AA087 in Chapter 3 141 Table 5.4 Mass cytometry results for bioconjugation of NeutraAvidin to the surface of PS dispersion polymerization particles before and after seeded emulsion polymerization with GMA. Binding was probed using an Lu-biotin reporter. 142

7 Analytical Aspects 170 Table 7.1. Recipes for the dispersion polymerization of styrene with PVP55 and acrylic acid in ethanol 174 Table 7.2. Particle size, size distribution, and the variation of Tm and Eu content for some PS-PAA microspheres synthesized in the presence of LnCl3.6H2O. 175 Table 7.3. Microwave digestion system: digestion program. This is table is reproduced from ref. 31 with permission. 181 Table 7.4. Averages of the integrated ion intensities over the transient signals for individual KG1a , U937 cells, and AA120-Eu microspheres. 12 different clones of CD34 antibodies: K (KG1a cells stained with 169Ir intercalator and Tm-labeled CD34) and U (U937 cells stained only with 169Ir intercalator). Each sample was examined in a 100:1 mixture with metal-containing microspheres AA120-Eu. The Eu content of the AA120-Eu gated without cells is reported on the last column. 187 Table 7.5. The recipe for the synthesis of microsphere samples AA137 A-E by 3-stage dispersion polymerization (3-DisP) of styrene with PVP55 as a dispersant in ethanol. 188 Table 7.6. Average volume of the microsphere samples AA137 A–E compared to their average number of 169Tm atoms per microsphere. Note that the standard deviation of number of 169Tm content of the different synthesis (for AA137B-E and excluding AA137A) is 2.6 x 105 Tm atoms per microsphere and standard deviation of Tm/Volume ratios (for AA137B-E and excluding AA137A) is 7.9 x 104 Tm atoms per unit volume. 191

LIST OF FIGURES AND SCHEMES

1 General Introduction 1 Scheme 1-1 Cartoon schematic of THE mass cytometer instrument. Metal-labeled particles are introduced into the ICP torch on an individual basis, whereby they are burned and the atomic composition of each particle is determined by TOF-MS. Mass spectra are recorded on the order of every 20 ms. The size of the ion cloud corresponding to each particle as it passes through the TOF chamber allows for 20 to 30 mass spectra to be collected per particle. 3 Scheme 1-2: Mechanism of dispersion polymerization copied from Yasuda et al. [79] with permission. 7 Figure 1.1: Carboxylate particles can be coupled to amine-containing molecules using an aqueous two- step coupling process with EDC and NHS or sulfo-NHS to form an amide bond with a amine containing molecules. It proceeds through an intermediate (sulfo)NHS ester, which has better reactivity for coupling amines than the EDC-reactive ester. 16 Figure 1-2: Hydroxyl-containing particles can be activated for coupling ligands using a number of strategies, which involve either aqueous or nonaqueous reactions. and the Carbonyldiimidazole and Disuccinimidyl carbonate methods provide reactive groups for amine coupling. 17 Figure 1.3: Aldehyde-particles can be reacted with amine-containing proteins or other molecules to form intermediate Schiff bases, which can be stabilized by reduction with sodium cyanoborohydride. 17 Figure 1-4: Illustration of the preparation of PEI-stabilized PS microspheres by dispersion polymerization and subsequent silica shell formation.Reproduced from Ref. [141] with permission. 21

2 Instrumental Methods and Experimental Details 38 Scheme 2-1: Schematics of the prototype CyTOF mass cytometer 43 Figure 2-1: Effect of the stirring blade depNo table of figures entries found.th on the particle size distribution 48 Figure 2-3. Potentiometeric and conductometric titrations of AA087 microspheres synthesized by 2-DisP in presence of AA (2.0 wt%/styrene). The highlighted area corresponds to 1.6 µmol of HCl 56

3 Microspheres by Two Stage Dispersion Polymerization 60 Scheme 3-1. Particle-forming polymerizations and size of resulting particles 62 Figure 3-1. SEM images for PS microsphere samples AA068, AA069 and AA050 synthesized in the presence of 1 % EuCl3 added in the second stage with (A) AAEM: 8 wt %/styrene (d = 2.7 μm, CVd =

1.5%), (B) AA: 2 wt %/styrene (d = 2.7 μm, CVd = 1.4%). and (C) AAEM: 4 wt %/styrene (d = 2.7 μm,

CVd = 1.5%). The white horizontal bars in the microphotographs represent 5 µm. 68 Figure 3-2. Screen captures for PS microsphere samples (A) AA068 and (B) AA069. AA068 was

synthesized in the presence of 1 % EuCl3 added in the second stage with AAEM: 8 wt %/styrene and

AA069 was synthesized in the presence of 1 % EuCl3 added in the second stage with AA: 2 wt %/styrene. 70 Figure 3-3: Distribution of mass cytometry signal intensity for AA050 PS microspheres prepared in 151 153 presence of EuCl3 (1.0 wt%/styrene) plus AAEM (4.0 wt%/styrene) (A): Eu and (B): Eu Intensity distributions. 71

xii

Figure 3-4: Distribution of mass cytometry signal intensity for AA181 PS microspheres prepared in presence of mixture of lanthanides (total 1.15 wt%/styrene, Table3-2) plus AA (2.0 wt%/styrene) (A-D): 169Tm and (E-H): 175Lu Intensity distributions. A and E: 105 particles per mL, B and F: 106 particles per mL, and C and G: 107 particles per mL. The dashed lines in D and H overlay plots refer to the constant Tm intensities of the M region with different particle concentrations. The arrows in D and H overlay plots are showing the increase of the doublets (H region) with increasing particle concentration. 74 Figure 3-5: 151Eu/153Eu bi-variant plots of mass cytometric results for AA050. PS microspheres (107

particles / ml). AA050 microspheres were prepared by 2-DisP in presence of EuCl3 (1.0 wt%/styrene) plus AAEM (4.0 wt%/styrene). 75 Figure 3-6: Gated distribution of mass cytometry signal intensity for AA050 PS microspheres prepared in presence of EuCl3 (1.0 wt%/styrene) plus AAEM (4.0 wt%/styrene). Gating was done by excluding the L and H regions in Figure 3-3A (A) and in Figure 3-5 according (B). 76 Figure 3-7. Distribution of mass cytometry signal intensity for three different populations of PS microspheres prepared in presence of EuCl3 (1.0 wt%/styrene) plus (AA050) AAEM (4.0 wt%/styrene, 151Eu Intensity = 6900), (AA068) AAEM (8.0 % wt%/styrene, 151Eu Intensity = 15100) and (AA069) AA (2.0 wt%/styrene, 151Eu Intensity = 90600). 78 Figure 3-8. Mass Cytometry screen captures for the analysis of three types of PS microspheres, each containing similar amounts of different Ln metals. (A) sample AA088: AA 2.0 + LaCl3 0.1 + TbCl3 0.1 +

HoCl3 0.1; (B) sample AA089: AA 2.0 + TbCl3 0.1 + HoCl3 0.1 + TmCl3 0.1 wt%/styrene. The numbers following each species refers to the wt%/styrene using in the particle synthesis. (C) Sample AA110: a

control experiment in which TmCl3 0.1 wt%/styrene and no AA was added in the particle synthesis. 80 Figure 3-9.. Distribution of gated signal intensity for encoding elements for four different populations of microspheres. The microspheres were encoded with four elements (La, Ho, Tb, and Tm) and five levels of concentration (coded from “0” to “4”). For this system of encoding the variability is equal to 624. The label “1” refers to the concentration level of 0.02 wt% LnCl3/sty. Thus, the label “2” refers to the

concentration level of 0.05 wt% LnCl3/sty, the label “3” refers to the concentration level of 0.1 wt%

LnCl3/sty and the label “4” refers to the concentration level of 0.2 wt% LnCl3/sty. 83 Figure 3-10. Tm ion release into the aqueous phase from colloidal suspensions of two Tm-containing PS microsphere samples in three different buffer solutions. AA089 microsphere sample contain 260 ppm Tm ion (w/w styrene). The pH 10.6, pH 7.0 and pH 3.0 buffer solutions are 200 mM sodium carbonate/bicarbonate, 10 mM ammonium acetate, and 50 mM sodium acetate, respectively. 86 Figure 3-11. Schematic representation of antigen capture and detection using metal-encoded microspheres and FC-MS. Carboxylated 151Eu and 153Eu-encoded PS microspheres were conjugated to a mouse IgG using carbodiimide chemistry. Microspheres were then washed and incubated with anti-mouse antigen that is labeled with a Pr-containing polymer tag (anti-mouse-IgG-Pr) to identify the presence of captured antigen on the particle. Stringently washed microspheres were analyzed for concomitant signals of 151Eu and 153Eu, and 141Pr as an indication of a successful immunoreactions. At the right, we present the chemical structure of the polymer before attachment to the antibody. X4 was reacted with a bismaleimide coupling agent and then covalently attached to the antibody via reaction with –SH groups produced by selective reduction of a disulfide bond in the hinge region of the antibody. Details are given in Ref. 21. Each polymer carries ca. 30 Pr3+ ions. 89 Figure 3-12. Mass cytometry analysis (141Pr intensities) of the interaction of anti-mouse-IgG-X4-Pr with bare AA069 microparticles and with samples of these particles subjected to conjugation conditions with BSA and with mouse IgG. 90

xiii

4 Microspheres by Three Stage Dispersion Polymerization 94 Figure 4.1. SEM image of unwashed sample AA105 (unwashed) microspheres synthesized in the presence of TmCl3 (0.1 wt%/styrene) and LaCl3 (0.1 wt%/styrene). d = 2.2 μm, CVd = 1.1%. 97 Figure 4.2. Mass cytometry measurements for AA105 169Tm/139La bi-variant plot (A) and gated intensity distribution for 139La (B) and 169Tm (C). Gating applied was highlighted by the red oval in (A). These

gated signals (B and C) were characterized by CVLa = 13 % and CVTm = 11%. 97 Figure 4.3.: SEM image and size distribution histogram of unwashed AA122 synthesized with EGDMA (4.0 wt.% / styrene) and AA (0.5 wt.% / styrene) added to the dispersion polymerization reaction 1.0 h after the 2nd stage. Note the raspberry texture for this sample which has high concentration of EGDMA. 103 Figure 4.4.: SEM image and size distribution histogram of unwashed AA123 synthesized with EGDMA (4.0 wt.% / styrene) and AA (2.0 wt.% / styrene) added to the dispersion polymerization reaction 1.0 h after the 2nd stage. Note the raspberry texture for this sample which has high concentration of EGDMA. 103 Figure 4.5.: SEM image and size distribution histogram of unwashed AA124 synthesized with EGDMA (0.5 wt.% / styrene) and AA (2.0 wt.% / styrene) added to the dispersion polymerization reaction 1.0 h after the 2nd stage. Note the presence of a few small particles, suggesting some secondary nucleation in the sample. 103 Figure 4.6.: SEM image and size distribution histogram of unwashed AA125 synthesized with EGDMA (4.0 wt.% / styrene) and AA (2.0 wt.% / styrene) added to the dispersion polymerization reaction 8.0 h after the 2nd stage. Note the raspberry texture for this sample which has high concentration of EGDMA. 104 Figure 4.7.: SEM image and size distribution histogram of unwashed AA126 synthesized with EGDMA (4.0 wt.% / styrene) and AA (0.5 wt.% / styrene) added to the dispersion polymerization reaction 8.0 h after the 2nd stage. Note the raspberry texture for this sample which has high concentration of EGDMA. 104 Figure 4.8. SEM image and size distribution histogram of AA127 synthesized with EGDMA (0.5 wt.% / styrene) and AA (0.5 wt.% / styrene) added to the dispersion polymerization reaction 8.0 h after the 2nd stage before washing (A) and after three cycles of washing (B). Note the presence of many smaller particles in (A) that makes this sample excluded from the discussion. 104 Figure 4.9.: SEM image and size distribution histogram of unwashed AA128 synthesized with EGDMA (0.5 wt.% / styrene) and AA (2.0 wt.% / styrene) added to the dispersion polymerization reaction 8.0 h after the 2nd stage. Microspheres are monodisperse in size. 105 Figure 4.10.: SEM image and size distribution histogram of unwashed AA129 synthesized with EGDMA (0.5 wt.% / styrene) and AA (0.5 wt.% / styrene) added to the dispersion polymerization reaction 1.0 h after the 2nd stage. Note the presence of some ca. 1.2 µm particles. 105 Figure 4.11. Tm content and volume of microsphere samples AA122 – AA129 synthesized by 3- DisP (A) and Tm content versus volume of microsphere same samples (B). 105 Figure 4.12. Average number of 159Tb atoms (A), normalized number of 159Tb atoms per unit volume per particle (B), average number of 169Tm atoms per particle (C) and normalized number of 169Tm atoms per unit volume for different microsphere samples (AA122 – AA129) prepared in presence of AA (0.5 – 2.0 wt.% / styrene) and EGDMA (0.5 – 4.0 wt.% / styrene). AA and EGDMA were added in the 3rd stage, 1.0 hrs (red circles) or 8.0 hrs (white circles) after the second stage. 109 Figure 4.13. Tm ion release into the aqueous phase from colloidal suspensions of two Tm-containing PS microsphere samples in three different buffer solutions. (A) AA089, synthesized by 2-stage DisP and (B) AA105 synthesized by 3-stage DisP. Both microsphere samples contain 260 ppm Tm ion (w/w styrene). The pH 10.6 is a 200 mM sodium carbonate/bicarbonate buffer solution, pH 7.0 is a 10 mM ammonium acetate buffer solution and pH 3.0 is a 50 mM sodium acetate buffer solution. 110

xiv

Figure 4.14. SEM images and particle distribution histograms for PS particle samples AA135 synthesized

in the presence of LnCl3.6H2O (where Ln = La, Nd, Eu, Tb, Dy, Ho and Tm each at 0.1 wt. % / styrene) added in the second stage with AA: 2 wt %/styrene without washing (A and C) (d = 2.1 μm, CVd = 6.6%, the arrow point to a particle that represented the small particle population) and after 3 cycles of washing

(B and D) (d = 2.2 μm, CVd = 3.2%). 115 Figure 4.15. SEM images with different magnifications for unwashed PS particle samples AA134 synthesized in the presence of LnCl3.6H2O (where Ln = La, Nd, Sm, Gd, Eu, Tb, Dy, Ho and Tm each at

0.1 wt. % / styrene) added in the second stage with AA: 2 wt %/styrene (d = 2.0 μm, CVd = 2.3%).Scale bars are 5.0 µm. 115 Figure 4.16. Screen captures of the mass cytometry results for (a) AA134 (4 particles were shown each of them has the 9 elements present); and (b) AA135 (2 particles were shown each of them has the 7 elements present). Both samples were synthesized in the presence of mixture of lanthanide salts (see Table 6-1) as examples of PS particles synthesized for the highly-encoded particles. 116 Figure 4.17: (a)Number of Tb ion per particle distributions measured by mass cytometry for a population of PS particles AA134 (red) and AA135 (blue) prepared by 3-DisP (b) A bi-variant plot of mass cytometry results for the Ho and Tm content of AA134 particles. 116 Figure 4.18. Log intensity values for 139La, 169Tm, and 141Pr from mass cytometry measurements for: Unmodified AA105 microspheres encoded with La and Tm. Pr signal here represents the nonspecific binding (first column set from left). BSA-modified AA105 microspheres encoded with La and Tm. Pr signal here represents the nonspecific binding of the anti-mouse-IgG to the BSA on the surface of the microspheres i.e. negative control (second column set from left). Mouse IgG-modified AA105 microspheres encoded with La and Tm. To which anti-mouse-IgG-X4-Pr was covalently attached. Pr signal here represents extent of specific binding of the anti-mouse-IgG to the microspheres (second column set from right). Carboxylated PS microspheress (Bangs Labs) modified with mouse IgG and to which anti-mouse-IgG- X4-Pr was covalently attached. Pr signal here represents the positive control (first column set from right). These data come from the analysis of approximately 20,000 particles for each sample. 118

5 Surface Functionalization of Lanthanide-coded Microspheres 123 Scheme 5.1. The bioconjugation conditions used to attach NeutraAvidin to particles’ surface. 126 Figure 5.1: The structure of the Lu biotin-conjugated reporter tag [8], which consists of a small peptide (Gly-Ser-Ala-Tyr-Gly-Lys-Arg-Lys) and a spacer (βAla- βAla-βAla-βAla) with a biotin molecule attached at one end and DTPA-Lu at the other end. 127 Figure 5-2. SEM images (A and C) and particle distribution histograms (B and D) for unwashed particles synthesized by 2-DisP. D2(PVP1.6) (A and B) and D3(PVP3.2) (C and D) were prepared in presence of 1.6 and 3.2 0 wt.% / styrene, respectively. 129 Figure 5.3 Mass cytometry measurements for particles prepared by 2-stage dispersion polymerization in 165 159 165 presence of activated PVP. (A) Ungated Ho/ Tb bi-variant plot for D2(PVP1.6). (B) gated Ho distribution for D2(PVP1.6) (blue curves CVHo = 20.6) and D3(PVP3.2) (red curves CVHo = 27.1 %). (C) 159 gated Tb distribution for D2(PVP1.6) (blue curves CVTb = 21.7) and D3(PVP3.2) (red curves CVTb = 30.1 %). Gating applied was 130

Figure 5.4: Mass cytometry results for bioconjugation trial of NeutrAvidin to D2(PVP1.6) (green) and

D3(PVP3.2) (blue) samples synthesized by two stage dispersion polymerization in presence of activated PVP. The sample were tested by the incubation of 109 particles with 100 μL of 500 nM of Lu-reporter in 1 % BSA, then washed once with PBS and finally redispersed in a 0.26 % w/w NaCl solution in 10 mM tris buffer. 132

Figure 5.5 SEM images of (A) washed D1 seed particles (davg = 2.11 µm and CVd = 1.3 %) synthesized by dispersion polymerization and (B) washed sample D4(MAA) (davg = 2.1 µm and CVd = 1.6 %) after seeded emulsion polymerization of D1 with MAA in a 2:1 ratio. 133 Figure 5.6 Surface acid titration results of D1 (black circles, seed latex) and D4(MAA) (open circles, after polymerization with MAA) after neutralization with excess NaOH followed by back titration with 0.01 M HCl performed under identical conditions. 135 Scheme 5.3. Proposed polymerization mechanism whereby agglomeration of small PGMA particles takes place onto the surface of the larger PS particles. 137

Figure 5.7 SEM image of PS particles (sample code D5(GMA-2), davg = 2.23 µm and CVd = 1.4 %) with a shell of glycidyl methacrylate (GMA) by seeded emulsion polymerization of D1 in the presence of excess surfactant (for SEM image of the seed latex D1, see Figure 5.5A). 137

Figure 5.8: SEM image of PS particles (sample code D5(GMA-1), davg = 2.67 µm and CVd = 3.6 %) with a shell of glycidyl methacrylate (GMA) by seeded emulsion polymerization of D1 in the presence of excess surfactant (for SEM image of the seed latex, see Figure 5.5A). 138 Figure 5.9 Mass cytometry data for the analysis of sample D5(GMA-2), which consists of a PGMA shell on D1 seed particles. Shown are (A) Tb-Ho isotopic dot-dot diagram; (B) Tb content distribution; (C) Lu content distribution for the detection of the Lu-biotin reporter, which measures the extent of protein binding to the surface. Red curve: non-specific binding of the Lu-biotin reporter to sample D5(GMA-2)-

BSA (covalent attachment of BSA, CVLu = 65.8 %). Blue curve: binding of the Lu-biotin reporter to

sample D5(GMA-2)-NA (covalent attachment of NeutraAvidin without EDC, CVLu = 32.5 %). Green curve: binding of the Lu-biotin reporter to sample D5(GMA-2)-EDC-NA (covalent attachment of

NeutraAvidin after surface activation with EDC, CVLu = 27.8 %). 140

6 Silica-Coated Particles 147 Figure 6.1: Structures of TEOS and APTS used to coat the PS particles. 148 Scheme 6-1: The procedure for the synthesis of silica coated PS particles 149 Figure 6.2 (a) and (c) SEM image and particle distribution histogram for sample AA167 (silica-coated

AA135 particles, unwashed) (d = 2.5 μm, CVd = 1.5%, after excluding the distorted spherical particles pointed by the arrows in (a) and to the right of the dashed line in (c)). (b) and (d) SEM image and particle

distribution histogram for sample X67 (APTS-coated AA167 particles, unwashed) (d = 2.5 μm, CVd = 1.5%, after excluding the silica nanoparticles pointed by the arrow in (b) and to the left of the dashed line in (d)). 150 Figure 6.3: Mass cytometry measurements of sample X67 prepared (a) Screen captures of the mass cytometry results for X67 (3 particles were shown each of them has the 7 elements present). (b) Ungated dot-dot plot of the Tm-Ho content of the particles. Gating is represented by the dashed oval. (c) Number of Tm ion per particle gated distribution measured by mass cytometry for a population of PS particles X67. 152 Figure 6.4: Linker X50, M = 588.67 (purchased from Pierce (EZ-Link NHS-PEG4-Biotin)) 153

xvi

Figure 6.5: (a) Workflow for particle sample X51. Silica-coated, amine-functionalized PS particles X67 were modified with biotin using active-ester linker X50 to obtain biotinylated-X67 particles. Following blocking against unspecific protein adsorption with BSA, biotin, on the biotinylated-X67 particles, was detected with Cy5-labeled streptavidin and CFM. The negative control was processed analogously, but was not biotinylated with linker X50. (b) CFM images of particles X51. Left: Cy5 channel, right: overlay of Cy5 channel and optical transmission image. Particles show homogeneous Cy5 fluorescence. Red fluorescence on the particle surface indicates binding of SAv-Cy5. 154 Figure 6.6 CFM images of particles X52. Left and middle: overlay of Cy5 channel and optical transmission image, right: Cy5 channel. Particles show inhomogeneous coverage with SAv-Cy5 155 Figure 6.7 (a) Workflow for particle samples X53a/b. Streptavidin-coated particles X53a/b were obtained by incubation of biotinylated particles with streptavidin in two different concentrations (X53a: 100 nM, X53b: 500 nM). Bound streptavidin was detected with biotin-fluorescein. (b) CFM images of particles X53a/b. Green fluorescence indicates binding of biotin-FITC (overlay: overlay of optical transmission image and FITC channel). 156 Figure 6.8 (a) Workflow for particle samples X55. Streptavidin-coated particles were incubated with Pr- loaded MCP X54 in three different amounts. Bound MCP was detected via CyTOF analysis (negative control: no biotin/streptavidin). (b) CyTOF result of particles X55. 157 Figure 6.9. (a) Workflow for particle samples X56a/b. After biotinylation, free amine groups on the particle surface were blocked with either DGA (particles X56a) or SA (particles X56b). Streptavidin particles were then created by incubation with streptavidin. Following binding to biotinylated, Pr-loaded MCP X54, bound MCP was detected via CyTOF analysis (negative control: no biotin/streptavidin). (b) CyTOF result of particles X56a/b. 159 Figure 6.10 (a) Workflow for particle samples X58. Particle surface amine groups were converted to carboxylic acids using DGA, which were activated with EDC/NHS and reacted with streptavidin to obtain streptavidin-coated particles X58 (BSA for negative control). Unreacted, activated acid groups were then capped with 6-aminocaproic acid. Following incubation with Pr-loaded MCP X54, analysis was carried out by CyTOF (b) CyTOF result of particles X58. 161 Figure 6.11 (a) Workflow for particle samples X59. Streptavidin particles (negative control: BSA particles) were blocked against unspecific DNA adsorption with herring sperm DNA. By incubation with biotinylated oligonucleotide A (X60), oligonucleotide-coated particles X59 were created. Particles X59 were subsequently reacted with either the Cy5-labeled anti-A probe X61 (antisense to oligo X60) or the Cy3-labeled anti-B probe X62, followed by CFM analysis. (b) CFM fluorescence images of particles X59. Cy5 fluorescence on particles S-RA indicates capture of oligo anti-A (X61). Cy3 fluorescence on particles S-RB and NC-RB is almost non-detectable showing absence of oligo anti-B (X62) (overlay: overlay of optical transmission image and Cy5 channel). 163 Figure 6.12 (a) Workflow for particle samples X65. Streptavidin particles (negative control: BSA particles) were blocked against unspecific DNA adsorption with herring sperm DNA. By incubation with biotinylated oligonucleotide A (X60: 5’-Biotin-AGCGGATAACAATTTCACACAGGA-3’) and B (X63: 5’-Biotin-CTGAGGTAGGTAGATCACTTGAGGT-3’) two types of oligonucleotide-coated particles X65 were created. Particles X65 were subsequently reacted separately (b) or combined (b) with a RAB mixture [Cy5-labeled anti-A probe X61 (5’-Cy5-TCCTGTGTGAAATTGTTATCCGCT-3’, antisense to oligo X60) and the Cy3-labeled anti-B probe X62 (-Cy3-ACCTCAAGTGATCTACCTACCTCAG-3’, antisense to oligo X63)] followed by CFM analysis. 164

xvii

Figure 6.13 CFM images obtained for particles X65. After incubation with fluorescently labeled anti-A and anti-B probes, particles carrying oligonucleotide A (X60) show only Cy5 fluorescence (top) while particles with oligonucleotide B (X63) show only Cy3 fluorescence (middle). This indicated that anti-A and anti-B probes bind only to their antisense particle-bound oligonucleotides without cross-reactivity. (Cy3/Cy5 overlay: overlay of Cy3 channel and Cy5 channel). In a mixture of both particle types (carrying oligonucleotides A or B), either Cy3 or Cy5 fluorescence was detected (bottom). This showed multiplex capability of the approach. 165

7 Analytical Aspects 170 Figure 7.1. SEM images for PS microsphere samples AA070-Tm synthesized in the presence of 1 %

TmCl3 added in the second stage with AA: 2 wt %/styrene (d = 2.1 μm, CVd = 1.8%). 176 Figure 7.3. Distribution of mass cytometry intensity signal for a population of PS microspheres (AA086-

Tm) prepared by 2-DisP in presence of TmCl3 (0.1 wt%/styrene) and AA (2.0 wt%/styrene) 179 Figure 7.4. SEM image for PS microsphere samples AA120-Eu synthesized by 3-DisP in the presence of

0.1 % EuCl3 added in the second stage with AA: 4 wt %/styrene (d = 2.2 μm, CVd = 1.4%). 180 Figure 7.5. Distribution of mass cytometry signal intensity for a population of PS microspheres (AA120-

Eu) prepared by 3-DisP in presence of EuCl3 (0.1 wt%/styrene) and AA (4.0 wt%/styrene). CVEu = 14 %. 180 Figure 7.7. Examples of bi-variant plots of mass cytometric results for (A) KG1a free cells stained with CD34-169Tm and Ir-interchelator and (B) free AA120-Eu microspheres. 183 Figure 7.8. A bi-variant plot of mass cytometry results for: 100:1 mixture of cells and microspheres. Colored points represent the Ir- and Eu-positive events. 185 Figure 7.9. Comparison between the 193Ir (from DNA intercalator) and 169Tm (from CD34-169Tm) averages of integrated ion intensities over the transient signals for KG1a cells gated alone (x-axis) and in a mixture with AA120-Eu microspheres (y-axis). 186 Figure 7.10. SEM images for samples AA137A-E of PS microspheres synthesized for reproducibility study. Samples were synthesized by 3-DisP in the presence of 0.05 % LaCl3 TbCl3 EuCl3 HoCl3 TmCl3 added in the second stage with AA: 2 wt %/styrene. Scale bar is 5.0 µm. 190 Figure 7.11. Screen captures of the mass cytometry results for AA137C, synthesized in the presence of

LaCl3 TbCl3 EuCl3 HoCl3 and TmCl3 (0.05 wt %/styrene for each), as an example of PS microspheres synthesized for the reproducibility study. The image shows 3 distinct microspheres each of them has the 5 elements present. 190 Figure 7.12. 169Tm intensity distribution from the mass cytometry measurements of AA137A-E (average number of Tm atoms per microsphere = Tm intensity x 3.95 x 104). Note the obvious increase in the Tm intensity of AA137A relative to the other samples. 191 Figure 7.13. Average volume of the microsphere samples AA137 A–E compared to their average 165Ho content (A) and average 169Tm content (B). 192

xviii

LIST OF APPENDICES Appendix 1 198

xix

ABBREVIATIONS

2-DisP two-stage dispersion polymerization

3-DisP three stage dispersion polymerization

AA acrylic acid

AAEM acetoacetylethyl methacrylate

AMBN 2, 2'-azobis(2-methylbutyronitrile)

CV coefficient of variation

DLS dynamic light scattering

DVB divinylbenzene

DyCl3 dysprosium(III) chloride hexahydrate

EDC 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride

EGDMA ethylene glycol dimethacrylate

EuCl3 europium(III) chloride hexahydrate

FITC-Avidin Avidin labeled with fluorescein isothiocyanate

GdCl3 gadolinium(III) chloride hexahydrate

GMA glycidyl methacrylate

GPC gel permeation chromatography

HMA hydroxyethyl methacrylate

HoCl3 holmium(III) chloride hexahydrate

HPC hydroxyl propyl cellulose

ICP-MS inductively coupled plasma mass spectroscopy

KG1a Human monocyte cell line

KPS potassium persulfate

LaCl3 lanthanum(III) chloride hydrate

MAA methacrylic acid

MMA methyl methacrylate

Mw molecular weight

NaHCO3 sodium bicarbonate

NdCl3 neodymium (III) chloride hexahydrate

NHS N-hydroxysuccinimide

PAA polyacrylic acid

PGMA polyglycidyl methacrylate

PMMA polymethyl methacrylate

PrCl3 praseodymium(III) chloride hexahydrate

PS polystyrene

PVP polyvinylpyrrolidone

PVP360 polyvinylpyrrolidone (Mw=360,000)

PVP40 polyvinylpyrrolidone (Mw=40,000)

PVP55 polyvinylpyrrolidone (Mw=55,000)

PVP-g-PS polyvinylpyrrolidone graft polystyrene

SDS sodium dodecyl sulfate

SEM scanning electronic microscopy

SmCl3 samarium(III) chloride hexahydrate

St styrene

TbCl3 terbium(III) chloride hexahydrate

THF tetrahydrofuran

TmCl3 thulium (III) chloride hexahydrate

TX305 Triton-X305

xxi

1 General Introduction

1 GENERAL INTRODUCTION One of the most significant challenges in contemporary biotechnology is the simultaneous detection and quantitative determination of multiple biomarkers in a single assay. The goal of these highly multiplexed assays is to be able to extract large amounts of data from smaller samples with increasing efficiency [1-8]. A variety of different formats has been proposed for these high-throughput approaches. These include multi-well microtiter plates, modified polymer surfaces (chips), and micrometer-sized polymer beads. Multiplexed bead-based arrays are an attractive option for supporting surface chemistries of immuno-[9] and gene expression assays [10]. In a manner similar to microtiter plates, various compositions, coatings or conjugation groups can be constructed or added to the microspheres to provide the requisite surface chemistry. These beads are then analyzed individually, often by flow cytometry. Cytometric fluorescent bead-based assays have demonstrated the increased sensitivity, specificity and dynamic range obtainable over standard enzyme immunoassays [11-14]. Traditional flow cytometry is based upon fluorescence or photoluminescence detection [4]. Fluorescence refers to the photo-excited emission from typical organic dyes, whereas the more general term photoluminescence incorporates emission from quantum dots and the phosphorescence-like emission from lanthanide chelates. Cytometric assays require two types of markers. The bead itself carries one or more dyes in various levels of concentration that acts as a code for the type of biomolecule attached to its surface. This type of marker is often referred to as a classifier tag, which is the identification marker within the microspheres to indicate its type. In addition, one needs a reporter tag to indicate successful binding of analytes to the particle surface. The reporter tag (also a fluorescent dye or quantum dot) is attached either to the analyte itself, or more commonly, to a secondary reagent, such as an antibody, peptide or other type of biomolecule to provide a signal associated with a successful binding event. For example, the Luminex system [15] employs classifier beads containing two dyes at ten levels of concentration, which theoretically allows 100 analytes to be identified by this bead set in one sample. The instrument is a flow cytometer equipped with two lasers, a 635-nm diode laser to excite the red and infrared dyes embedded in the beads, and a 523-nm Nd:YAG laser to excite the orange reporter, pycoerythrin (PE) attached to the reporter molecules.

2 General Introduction

Using such systems, many successful immuno- and gene expression assays have been reported. For example, Yang et al.[16] could quantify gene expression at the level of RNA transcripts by demonstrating the multiplexing of 20 genes with a lower detection limit of 100 attomole. A recently published paper describes the use of a color-coded bead mixture for testing antibody specificity [17]. A powerful high-throughput multiplex immunobead assay was used to test simultaneously 29 cytokines, chemokines, angiogenic as well as growth factors, and soluble receptors in the sera of patients diagnosed with high-risk melanoma [18]. One of the limitations of photoluminescence-based assays is the limited number of different dyes and different emission intensities that can be read simultaneously. The analysis is complicated because different dyes often have to be excited at different wavelengths. There is also a finite bandwidth to the emission that limits the number of dyes that can be examined simultaneously. Some of these problems can be mitigated by using quantum dots with a very narrow size distribution, leading to narrower emission bands. Quantitative analysis of analyte concentration in a sample is difficult, and it is also a challenge to measure simultaneously intensity ratios that differ by more than a factor of ten. A much larger amount of information can be coded if one employs different metal atoms or isotopes as labels. The type of metal and the amount present can be measured with high sensitivity by inductively coupled plasma mass spectrometry (ICP-MS) [19]. In ICP-MS, samples are burned in Ar plasma torch at ca. 7000 K, and in the process, metals are atomized and ionized with quantitative efficiency. Metal ions are detected with unit mass resolution and with a very wide dynamic range employing a quadrupole mass analyzer. Bioassays that are based on metal encoded beads and elemental reporter tags need a detection system capable of registering a fast transient signal individually. The fast data acquisition system should also be able to register at least hundreds transient events per second. These goals can be met with a novel instrument which injects cells or beads stochastically and analyzes them one-by-one. Here the beads are nebulized and injected into the plasma torch of the ICP ion source linked with time-of-flight (TOF) detection [20-22]. In this thesis, I refer to this technique as mass cytometry (Scheme 1-1).

3 General Introduction

Scheme 1-1 Cartoon schematic of the mass cytometer instrument. Metal-labeled particles are introduced into the ICP torch on an individual basis, whereby they are burned and the atomic composition of each particle is determined by TOF-MS. Mass spectra are recorded on the order of every 20 ms. The size of the ion cloud corresponding to each particle as it passes through the TOF chamber allows for 20 to 30 mass spectra to be collected per particle.

Polymer particles for biological assays must satisfy several important criteria. In particular, the particles must possess a monodisperse particle size distribution, in order to minimize variability during analysis. Additionally, the particles must possess a surface that permits bioconjugation of biomolecules such as proteins and oligopeptides, with a minimum amount of non-specific adsorption. These criteria are universal for bio-assay applications, however for the development of an assay based on the mass spectrometric detection, additional criteria must be satisfied: the particles must be sufficiently large (> 500 nm diameter) to allow for the incorporation of a high loading of metal ions, and there must be very low variability of metal content from particle to particle. The achievement of satisfying all of these criteria is a non- trivial challenge for the synthetic chemist. There are many potential techniques for the synthesis of polymer microspheres that can be used for biological assays. A variety of different heterogeneous polymerization methods exist for the synthesis of sub-micron and micron-sized polymer particles, such as emulsion [23-26], microemulsion [27-30], seeded emulsion [31-34], suspension [35, 36] and dispersion polymerization. [37-44] . In order to synthesize sub-micron polymer particles, emulsion polymerization is arguably the most suitable method, with the ability to control the resultant particle size and particle density by changing experimental parameters. Larger particles in the

4 General Introduction

range of 1.0 – 10 µm diameter are more readily produced by alternative methods such as dispersion polymerization, however the incorporation of desirable surface functionality and other additives have proven difficult [41]. In this project, dispersion polymerization was investigated with respect to its applicability for the synthesis of metal-labeled polymer particles to act as classifier particles for biological assays. In this introductory chapter, I will review dispersion polymerization and its many aspects as a potential methodology for the synthesis of metal-labeled polymer microspheres for mass cytometry-based biological assays. Additionally, the analytical method of flow cytometry will be discussed in order to understand the importance of the preparation of classifier particles and the necessary criteria that such particles must satisfy. Finally, the physical and chemical properties of the lanthanide metals and their compounds will be briefly covered due to their applicability as metal tags in this project.

1.1 Polymer Particles by Dispersion Polymerization

Dispersion polymerization is one of the many methods of heterogeneous polymerization, and is routinely used in the synthesis of monodisperse polymer particles larger than 1 micrometre in diameter. Dispersion polymerization readily allows the ability to synthesize particles in the range of 1.0 – 10 µm. While the label “dispersion polymerization” was first introduced to the literature by Barrett and Thomas at ICI in 1969 [37], it was not until 1981 when Almog et al. [38, 45] synthesized “monodisperse” particles in polar media. They used partially hydrolyzed polyvinyl alcohol (PVA) as a colloidal stabilizer and a quaternary ammonium salt as an electrostatic costabilizer. Since that time, dispersion polymerization has been used extensively in

the synthesis of polymer particles for numerous research and industrial applications [46-57]. The dispersion polymerization process, like all ab initio heterogeneous polymerization methods, consists of a particle nucleation stage and a particle growth stage. Understanding these two processes provides the ability to create functional particles (by incorporating two or more monomers, surface functional groups or particle labels such as metal ions into the particle structure) that are also monodisperse. In this section, I will highlight the critical kinetic and mechanistic aspects of dispersion polymerization, as well as modified dispersion polymerization

5 General Introduction

methods such as the ‘two stage’ approach (denoted 2-DisP) developed by Song et al [58]. As the particles synthesized in this project are for use in biological assays, the ability to create surface functional particles that permit bioconjugation to proteins and peptides is also discussed. Finally, a review is provided on the strategies that exist for the incorporation of fluorophores, metal nanoparticles into polymer particles both during and after the dispersion polymerization process.

1.1.1 Dispersion Polymerization: Mechanism and Kinetics

Dispersion polymerization [44, 59, 60] is a type of precipitation polymerization carried out in the presence of a polymeric stabilizer that is soluble in the reaction medium. The first examples were reported by Barrett and Thomas at ICI in 1969 [37]. In dispersion polymerization, monomer(s) and the steric stabilizer are soluble in the solvent; however the polymer that is formed is insoluble. Lok and Ober [39, 61-63] pioneered the synthesis of polystyrene particles via dispersion polymerization in a variety of different alcoholic solvents using cellulosic derivatives as a steric stabilizer. In the absence of an electrostatic costabilizer, Lok and Ober made polystyrene particles with very uniform sizes. They investigated the effect of the solvent, monomer and initiator concentration on the particle size and particle size distribution. Under favorable conditions, they could synthesize particles ranging from 1.0–10 µm in diameter, often of very narrow size distribution [39, 61-63]. The mechanism of dispersion polymerization of methyl methacrylate (MMA) in methanol with polyvinylpyrrolidone (denoted PVP40, with Mw – 40000 Da) as the steric stabilizer was studied by Shen et al. [64]. Upon heating, the AIBN initiator decomposes and generates free radicals, which initiate chain growth by the addition of monomers. The growing chains remain soluble in solution until they reach a critical chain length, at which point they precipitate. The precipitated chains aggregate to form nano-sized nuclei, which are unstable. These nuclei further aggregate into larger particles, and upon absorbing additional stabilizer from the medium, these large particles become stabilized against further aggregation. These particles grow by capturing small nuclei and oligomeric radicals from the continuous phase, and by polymerizing monomer located in the particle interior. In the literature, there is an agreement that nuclei are formed throughout the dispersion polymerization process, while the number of particles is determined at an early stage. After this

6 General Introduction

early stage, the number of particles does not change but the particles increase in size as the polymerization proceeds [64-66]. According to Lok et al. [61] and other researchers [40, 64, 67], the aggregation rate of nuclei and the stabilization kinetics determine the number of stable particles. These stable particles contain oligomeric and polymeric chains. The role of the polymer stabilizer in the particle formation process during dispersion polymerization is of great importance. In the absence of a polymer stabilizer, nucleation occurs, but coalescence cannot be controlled, and stable particles cannot form. As a consequence, the polymer precipitates from the solution. It is possible to state that while nucleation kinetics determines the initial number of particles, the polymer stabilizer exercises control over the particle size and size distribution [42, 68]. Hydrophilic polymers such as polyvinylpyrrolidone (PVP) [40, 64, 66, 69], polyvinyl alcohol (PVA) [38, 45], polyacrylic acid (PAA) [70-72], poly(2-ethyl-2-oxazoline) [73] and cellulosic derivatives [61] are examples of polymer dispersants that can be used in dispersion polymerization. The selection of an effective stabilizer is critical; not only be soluble in the medium throughout the reaction, but also it must be able to cover the particle surface sufficiently to prevent coagulation. According to the mechanistic model by Paine et al., [42, 69] the true stabilizer is not the homopolymer but the polymer-grafted stabilizer. In addition, physically adsorbed chains also contribute to the particle stabilization [64]. It is convenient to divide the dispersion polymerization process into two major stages, namely, nucleation and growth [69, 74-78]. Nucleation is the first stage in which the formation of nuclei (nano-size particles) is predominant. In the nucleation stage, the nuclei are not stable and aggregation of these nuclei occurs to form larger particles. On the other hand, the particle growth stage is a later stage in which the particle growth is predominant. The new nuclei formed during the particle growth stage do not form new stable particles; instead they aggregate with pre-existing stable particles. The number of stable particles does not increase during the particle growth stage. For the dispersion polymerization of styrene in ethanol using the initiator AMBN at 70 C, the particle growth stage takes much longer (usually more than 20 h) than the very fast nucleation stage (ca. 5 min) [58, 79, 80].

7 General Introduction

Scheme 1-2: Mechanism of dispersion polymerization copied from Yasuda et al. [79] with permission.

Yasuda et a1. [79] developed a model for the particle nucleation stage in the dispersion polymerization of styrene in ethanol in the presence of PVP stabilizer as depicted by Scheme 1- 2. Similar to Paine [69], they assumed that the particles were stabilized by the PVP-g-PS molecules formed in the reaction. Thus they defined the amount of the PVP-g-PS produced in the ethanol phase as the "graft available", and they defined the minimum amount of the PVP-g-PS required to prevent particle aggregation as "graft required". They also defined a stabilization time

(tstab) as the time at which the value of the "graft available" reaches that of the "graft required".

At tstab, all particle surfaces are covered with sufficient PVP-g-PS molecules and no more particle aggregation occurs. The number of particles decreased with time until tstab and thereafter became constant. Paine [69] established that the monodispersity of the particle distribution was lost when particle aggregation continued after tstab. In addition, Yasuda et a1. [79] established that the total

number of particles in the reaction remained constant at reaction times from 2 to 24 hr. They also compared the particle numbers predicted by their model with their experimental results. The

8 General Introduction

theoretical prediction of the particle concentration under various monomer concentrations and stirring speeds agreed well with the experimental results measured at 24 hr, indicating that their model could quantitatively predict the particle formation stage in dispersion polymerization. A second model was developed by Yasuda et al. [80] to describe the particle growth stage in the dispersion polymerization. In this model, experimental data obtained at the reaction time of 2 hr (i.e., well after the conclusion of the nucleation stage) were used as the initial values for the simulation. Experimental data were in good agreement with the theoretical prediction for both the particle diameter and monomer conversion as a function of time calculated with this model. In addition, they compared the evolution of monomer conversion in the dispersion polymerization of styrene in ethanol with that of the solution polymerization of styrene in cyclohexane at 70°C. The rate of dispersion polymerization was higher than that of the solution polymerization, suggesting a significant gel effect in the dispersion polymerization. They also found that the experimental molecular weight (Mw) increased with increasing the reaction time in dispersion polymerization, and was higher than the molecular weight in the solution polymerization experiment after 30 minutes of polymerization. These results suggested that the termination rate constant within the particles decreased due to the gel effect.

1.1.2 Different Processes for Dispersion Polymerization

1.1.2.1 Dispersion Polymerization by The Batch Process

A batch dispersion polymerization process involves dissolving all of the monomer and stabilizer in the desired solvent before the initiation of the reaction. Almost all of the dispersion polymerization processes reported in the literature have been performed by the batch dispersion polymerization method. The general process involves deoxygenation of the solution, raising the solution to the reaction temperature and then adding the initiator. Many researchers [65, 66, 81-92] have studied the effects of different parameters on dispersion polymerization reactions. Paine et al. [66] studied the dispersion polymerization of styrene in alcoholic media, and examined the influence of the major reaction variables on the size of PVP-stabilized particles. They found that the variable with the biggest impact was changing the solvency of the reaction medium. Solvency affected the behavior of the PVP-PS

9 General Introduction

graft polymers. Increasing the solubility of the grafted PVP-PS led to increasing the length of time those species stayed in the continuous media, rendering them less effective as a stabilizer. As a consequence, larger particles were formed. Paine et al. [66] confirmed the concept of increasing the particle size by increasing the solubility of grafted PVP-PS via three different experiments. In the first experiment, they increased the solvency for the PS and grafted PVP-PS by the addition of toluene to the reaction. They observed that increasing toluene concentration led to the formation of larger particles. Decreasing the solvency for the PS and grafted PVP-PS (by the addition of water) yielded particles with smaller diameters. In another experiment, they increased the monomer concentration, which also increased the solubility of grafted PVP-PS. They again observed an increase in the particle size. In their third experiment, a higher concentration of AIBN initiator was used, which also led to an increase in the particle size. They ascribed this behavior to the formation of lower molecular weight (i.e. more soluble) polystyrene when high initiator concentrations were employed. The process of batch dispersion polymerization has been used in the synthesis of a variety of polymer particles. However, far fewer reports describe utilization of comonomers to synthesize copolymer particles by batch dispersion polymerization [40, 41, 93, 94]. The comonomer ratio was found to affect the course of the reaction [41]. Usually, each new comonomer ratio must be treated as a new polymerization system due to the complex nature of comonomer mixtures. In most of cases reported, using a comonomer led to the formation of particles with broad particle size distributions, particularly if the different monomers polymerize at different rates or partition differently in the system.

1.1.2.2 Semi-Continuous, Continuous and Seeded Dispersion Polymerizations

The continuous polymerization reaction is a type of polymerization in which the monomer is continuously fed to a reactor and the polymer is continuously removed. In semi-continuous (or semi-batch) polymerizations, monomer is usually added incrementally over certain time period; then the reaction continues for a longer time with no further monomer addition. Two recent reports describe dispersion polymerization reaction run in semi-continuous and continuous manners. Huang et al. reported attempts to synthesize cross-linked PS particles by a semi-

10 General Introduction

continuous process [95]. The stabilizer (in this case, PVP) was dissolved in the dispersion medium (an ethanol-heptane mixture), and the solution was heated to the reaction temperature. A mixture of styrene, comonomer (2,2'- oxybis-bisethanol diacrylate) and AIBN initiator was added slowly into the flask over 5 hrs. The reaction was continued until the conversion was more than 90%. The resulting particles had a larger particle diameter (4.32 µm) and a much broader

size distribution (CVd: 1.25 %) compared to the batch dispersion polymerization of the same

recipe (3.35 µm, CVd of 1.08 %). Giaconia et al. examined the continuous dispersion

polymerization of MMA using supercritical carbon dioxide (scCO2) as a solvent [96]. The polymerization was performed at 65 C and 25 MPa with 2,2`-azobisisobutyronitrile (AMBN) as the initiator, and a reactive polysiloxane macromonomer (PDMS) as the polymeric stabilizer. In a continuously stirred tank reactor, the CO2, MMA, surfactant, and AIBN were introduced continuously, and the influence of the residence time on particle size and particle size distribution (among other parameters) was studied. The resulting particles had a larger particle diameter ( 3.7 µm) and a much broader size distribution compared to the batch dispersion polymerization of the same recipe (< 2.6 µm). Seeded polymerization is a reaction where a monomer or a monomer mixture is polymerized in the presence of particles (seeds) prior to the commencement of polymerization. Seeded emulsion polymerization is commonly employed as complementary reaction for the emulsion polymerization to post-functionalize polymer particles. However, seeded dispersion polymerization is more challenging and thus is not frequently used. The first report regarding seeded dispersion polymerization was by Okubo et al. [97], who used the reaction to decorate preformed polystyrene particles with chloromethyl groups. In their experiment, the seed polystyrene particles were synthesized by batch dispersion polymerization. Since this time, seeded dispersion polymerization has been used in some heterogeneous polymerization systems for the purpose of better control of particle size, morphology and surface functionality, in addition to fundamental investigation and mechanistic studies. For example, in order to study the mechanism of dispersion polymerization, Shen et al. [64] carried out the dispersion polymerization of MMA using PVP as the stabilizer in the presence of PMMA seed particles (3.2 µm, prepared by dispersion polymerization in methanol).

11 General Introduction

Because of the potential flexibility of the method, seeded dispersion polymerization is attractive for the creation and functionalization of polymer microparticles. The major drawback of seeded dispersion polymerization, however, is the large amount of reaction optimization required in order to create monodisperse particles in a reproducible manner. Alternative approaches for the incorporation of specific functionalities into polymer particles, such as two- (or even multi-) stage dispersion polymerization, have become increasingly common as a result. The two-stage dispersion polymerization process is discussed in detail in the following section.

1.1.2.4 Two-Stage Dispersion Polymerization

It has previously been shown [40, 61, 66, 68, 79, 98, 99] that when additives such as hydrophilic monomers, cross-linking agents or functional molecules (such as dyes) were added to a batch dispersion of styrene, poor results were obtained. These poor results consisted of particle coagulation and the loss of particle monodispersity. In order to combat these problems, a methodology known as ‘two stage dispersion polymerization’ (2-DisP) was developed in our research group by Dr Jing-She Song [58, 100-104]. In the 2-DisP process, the comonomers (or crosslinking agents or dyes) were not added to the polymerization medium until the particle nucleation stage was complete. In essence, the 2-DisP process is a type of seeded dispersion polymerization, however the seed particles are formed in situ. As discussed in the previous section, the dispersion polymerization process consists of two dominant stages – the nucleation stage, followed by the particle growth stage. The particle nucleation process is short but sensitive to many system parameters, in particular the presence of other species (comonomers or otherwise) in the reaction medium. Slight changes to the conditions where nucleation takes place greatly affects the particle size, particle density and dispersity of the particle size distribution. By adding additional reagents to the dispersion polymerization medium after the nucleation stage is complete, incorporation of desired functionality can be achieved while retaining a monodisperse particle size distribution. Traditionally, the particle nucleation stage is considered to be finished when the number of particles in the reaction medium is constant [69, 79, 80]. The monitoring of the particle number in a dispersion polymerization reaction, however, is extremely difficult. Alternatively, Dr Song relied on mathematical simulations by Yasuda et al. [79] and experiments by Kim et al. [99].

12 General Introduction

These results suggested that the nucleation stage was complete at less than 1% monomer conversion [79, 99]. From an operational point of view, Dr Song chose 2 to 3 % monomer conversion as the point to add the problematic reagents. Many comonomers added after this point became incorporated into the particles without disturbing the final particle size and size distribution [58, 100-104]. In this manner, Dr Song was able to prepare dye-labeled or functional group-containing particles with a very narrow particle size distribution, typically 1 micrometre or greater in diameter. By varying the amount of styrene monomer added in the second stage, he was able to control the final particle diameter precisely without changing the narrow size distribution. Most importantly, he found that this synthetic strategy useful to prepare cross-linked particles containing up to 3 mol % cross-linking agent (EGDMA). One of the aims of my project was to use the two-stage dispersion polymerization methodology to incorporate metal ions into polystyrene particles through the addition of the metal salt(s) with a comonomer in the second stage. Ideally, the comonomer used should have two characteristics. Firstly, the comonomer should readily copolymerize with styrene; secondly, the comonomer should be able to chelate to the metal ions added to the reaction mixture. These two characteristics will ensure the complete incorporation of the metal ions into the particle interior. Additionally, the comonomer(s) used should provide a functional group (such as

–COOH and –NH2 groups) that can permit the attachment of biomolecules to the particle surface for subsequent bioconjugation applications.

1.1.3 Functional Microspheres by Dispersion Polymerization

Using dispersion polymerization processes, researchers have synthesized a wide variety of functional microspheres (Table 1-1). Many of these microspheres have manifested suitability for biological applications. Table 1-1 summarizes some of these polymer functional microspheres based on the functional comonomer used. In this section, I describe the synthesis of functionalized particles (with carboxylic acid, hydroxyl, aldehyde and amine groups) prepared by dispersion polymerization; in addition to some bioconjugation reactions that are used for microspheres post-functionalization with biomolecules.

13 General Introduction

Table 1-1: Different functional monomers used in the synthesis of polymer particles (by dispersion polymerization).

Functional Group Monomer Reference Acrylic acid [100] -COOH Methacrylic acid [40, 105] Itaconic acid [106]

Glycidyl methacrylate [107] -CHO Acrolein [108, 109]

-NH2 Amino methylstyrene [110] -OH 2-hydroxypropyl methacrylate [111]

-CH2Cl chloromethylstyrene [97, 112]

1.1.3.1 Microspheres with Functional Groups

Dispersion polymerization has been used to synthesize polymer microspheres with carboxylic acid functionality on the surface through monomers such as acrylic acid (AA), methacrylic acid (MAA) or itaconic acid (IA). For example, Tseng et al. [40] used methacrylic acid (1.0 wt % based on styrene) and PVP as stabilizer to prepare PS-PMAA particles (d = 3.3 µm) with narrow size distributions. However, these particles were much bigger than the particles prepared without MAA (d = 2.5 µm). Yang et al. [105] studied the dispersion copolymerization of styrene with acrylic acid (AA). They used larger amounts of AA (3.3 - 6.7 wt % based on styrene) than Tseng et al. [40] and found that with increasing amounts of AA in the monomer mixture, the resulting particle size increased and the particle size distribution became broader (d = 1.5 µm and CV = 12 % for PS particles; d = 2.3 µm and CV = 19 % for PS-PAA particles with AA: 6.7 wt % based on styrene). When they used higher AA concentrations (more than 6.7 wt % based on styrene), gel-like polymer was obtained. Itaconic acid (IA) as a functional comonomer was also used to synthesize carboxyl- functionalized polystyrene particles. Song et al. reported the synthesis of PS-PIA particles with diameters in the range of 600-2100 nm by two-stage dispersion polymerization in ethanol/water media using PVP as the stabilizer [106]. Among other factors, the authors discussed the effect of IA concentration on the particle size and size distribution. They found that increasing the amount

14 General Introduction

of IA increased particle size and size distribution (d = 1.8 µm and PDI = 1.03 % for PS particles; d = 2.4 µm and PDI = 1.30 % for PS-PIA particles with IA/Sty = 0.07 w/w). Hydroxyl-functionalized particles were synthesized by dispersion polymerization through monomers such as 2-hydroxypropyl methacrylate (HPMA) and 2-hydroxyethyl methacrylate (HEMA). For example, Ali et al. prepared poly(2-hydroxypropyl methacrylate) (PHPMA) particles by dispersion polymerization using poly(N-vinylpyrrolidone) as a steric stabilizer in water. Because there was a low level of a dimethacrylate impurity in the HPMA monomer, the PHPMA particles were lightly cross-linked and exhibited microgel character. They synthesized different PHPMA particles with a diameters range of 0.5 – 0.8 µm. Narrow size distribution (PDI) of 0.03 was obtained when high PVP concentration (20 wt. % to HEMA) was used [111]. Polymeric microspheres containing aldehyde groups are usually produced by copolymerization with acrolein (aldehyde monomer). Almost all of procedures for the syntheses of polyacrolein containing particles were developed for polymerizations in aqueous media [113- 116]. However, poly(acrolein-co-styrene) copolymer microspheres were also obtained via a dispersion copolymerization reaction carried out in ethanol as reported by Basinska and Slomkowski [108, 109]. Their reaction mixture contained freshly distilled acrolein, styrene, AIBN and polyvinylpyrrolidone (PVP, Mw 40K) in ethanol using a batch process. The microspheres they obtained had diameters larger than 1.0 µm. They found that the diameter of the microspheres decreased with the increasing fraction of acrolein in the monomer feed. Amine-functionalized polymer microspheres have been synthesized by dispersion polymerization. For example, Ma et al. [117] and Wang et al. [118] reported very similar methods to prepare micron sized monodisperse polyglycidyl methacrylate (PGMA) particles with functional amino groups. In both reports, the process involved two steps: first was the preparation of parent monodisperse PGMA particles by the dispersion polymerization of GMA in ethanol/water mixture and PVP stabilizer. Then chemical modification of the PGMA particles with ethylenediamine (EDA) in water at 80 C was done to yield amino groups on the surface of the microspheres The parent PGMA particles, prepared by Ma et al. [117], had average diameter of 2.25 µm (PDI 1.02) while Wang et al. [118] particles were smaller (1.42 µm, CV 3.8 %).

15 General Introduction

Recently Itoh et al. [119] synthesized amine-functionalized polystyrene microspheres, by dispersion polymerization in ethanol, through employing polystyrene-b-poly(aminomethyl styrene) (PS-b-PAMS) as a stabilizer. Among other conditions, they investigated the effect of the PS-b-PAMS stabilizer concentration on the particle size and particle size distribution. With an increase in the stabilizer concentration from 0.05 wt% to 2.0 wt%, a decrease in particle size from 1.32 µm to 0.43 µm was observed. Although the resulting PS particles showed a narrow size distribution with PSD = 1.01 at 0.05 and 0.5 wt% of the stabilizer, the increase in the PS-b- PAMS concentration to 1.0 and 2.0 wt% led to some broadening of the PSD values to 1.06 and 1.09, respectively. They concluded, for their reaction, that narrow size distribution could be achieved by the use of 0.05–0.5 wt% PS-b-PAMS, while 1.0 wt% and above of the copolymer led to broad distributions.

1.1.3.2 Bioconjugation Reactions as Post-Functionalization for Microspheres

Many of the functionalized microspheres mentioned above have been used to covalently couple biomolecules through the appropriate reaction conditions. For each type of these functionalized particles, I describe in this section a method for their post-functionalization reaction. By far the most common reaction strategy for coupling proteins and other amine-containing molecules to carboxylate particles is through an aqueous, carbodiimide-mediated process using EDC (1-ethyl-3-(3-dimethylaminopropyl)carbodiimide), either in a single-step coupling reaction or in a two-step reaction that employs the addition of N-hydroxysuccinimide (NHS) or sulfo- NHS. Of all the crosslinking methods used in bioresearch applications today, this relatively simple coupling process is the most frequent conjugation reaction done with proteins. Using this EDC reaction, carboxylate particles first are activated with the water-soluble carbodiimide EDC to create an intermediate ester. This ester is reacted directly with amines on proteins, but it also can be used with the addition of NHS or sulfo-NHS, which results in the formation of another intermediate, the NHS ester or sulfo-NHS ester. The formation of this second ester results in a more stable intermediate in aqueous solution than the one formed with EDC, therefore the secondary coupling reaction with proteins proceeds with higher yields than with the use of EDC alone (Figure1.1). In addition, by forming the secondary (sulfo)NHS ester, excess EDC can be

16 General Introduction

removed from the particles before adding protein, thus preventing carbodiimide-mediated protein polymerization due to the presence of both amines and carboxylates on most proteins [120, 121].

Figure 1.1: Carboxylate particles can be coupled to amine-containing molecules using an aqueous two-step coupling process with EDC and NHS or sulfo-NHS to form an amide bond with a amine containing molecules. It proceeds through an intermediate (sulfo)NHS ester, which has better reactivity for coupling amines than the EDC-reactive ester. Hydroxyls are not spontaneously reactive toward functional groups on biomolecules. However, they can be activated for covalent coupling by a number of known reaction mechanisms. A convenient method of activation is to form a reactive carbonyl group on the hydroxyl particle using compounds such as carbonyldiimidazole [122] or disuccinimidyl carbonate [123]. These activating agents create imidazole carbamates or NHS-carbonates on the particle surface, which then are spontaneously reactive toward amines groups on biomolecules (Figure1-2). Aldehyde particles are spontaneously reactive with hydrazine or hydrazide derivatives, forming hydrazone linkages upon Schiff base formation. Reactions with amine-containing molecules (Figure 1.3), such as proteins, can be done through a reductive amination process using sodium cyanoborohydride.

17 General Introduction

Figure 1-2: Hydroxyl-containing particles can be activated for coupling ligands using a number of strategies, which involve either aqueous or nonaqueous reactions. and the Carbonyldiimidazole and Disuccinimidyl carbonate methods provide reactive groups for amine coupling.

Primary amine particles may be used for covalent immobilization using a number of reaction routes. A carbodiimide-mediated coupling process may be used, as described above for carboxylate particles, but this time it is done by activation of a carboxylate group on the ligand and subsequent amide bond formation with the amines on the particles. A single step EDC reaction strategy will work well for small biomolecules containing one or more carboxylates with no amines. However, for carbodiimide-mediated protein coupling to amine particles, the protein first should be activated with EDC and sulfo-NHS according to the method of Grabarek et al. [124] and then the activated intermediate added to the amine particles for conjugation.

Figure 1.3: Aldehyde-particles can be reacted with amine-containing proteins or other molecules to form intermediate Schiff bases, which can be stabilized by reduction with sodium cyanoborohydride.

18 General Introduction

1.1.4 Fluorescent Dye- and Metal-Containing Microspheres by Dispersion Polymerization

The use of polymer particles that contain fluorescent dyes, metal ions, or metallic nanoparticles is very common in biological assays. In this section, I review some approaches to synthesize fluorescent dye- and metal-containing polymer microsphere using dispersion polymerization.

1.1.4.1 Fluorescent Dye-Labeled Microspheres

Polymer particles provide a matrix that can be swollen for embedding organic dyes or fluorescent molecules in their core. These dyes and molecules form highly fluorescent microparticles when incorporated into the microspheres. Such fluorescent or dye-labeled particles are useful in multiplexed detection systems using suspension arrays (e.g., Luminex technology [125, 126]). In this application, dyes are incorporated within the microspheres to create a series of different colored microspheres that are distinguished based on their fluorescence emission. In addition, changing the amount of dye molecules within the microspheres or blending two or more dyes in a single particle population can form a gradient of different color compositions. Particle subpopulations of a particular color or emission wavelength and intensity then can be used to identify the type of target being measured in a suspension assay. Mixing together in a single solution such particle subpopulations having different colors permits multiple targets to be assayed simultaneously, wherein each particle type is identified by its color and correlated to the analyte being targeted. Flow-cytometry-based instruments then are used to detect and measure the particle color and assay result (see section 1.2 for flow cytometry technique). Dye-labeled particles also are commonly used in diagnostic lateral flow tests (like the common home pregnancy test), as the colors can be seen with the eye without the need for special detectors. In this type of assay, antibodies or antigens are coupled to the dye-labled particles and a sample solution applied to the test strip carries them along within a membrane. The particles then are captured at points in the membrane that represent either a control or a positive sample result. Large numbers of color particles docking at these points within the membrane create the visual lines associated with these disposable tests.

19 General Introduction

Dispersion polymerization has been used to prepare dye-labeled microspheres. For example, Bosma et al. reported the synthesis of fluorescently-labeled poly(methyl methacrylate) particles by dispersion polymerization using poly(12-hydroxystearic acid) as a stabilizer. They used rhodamine isothiocyanate-aminostyrene and 4-methylaminoethylmethacrylate-7-nitrobenzo-2- oxa-1,3-diazol fluorescent dyes and were able to obtain homogeneous fluorescent particles. The particles prepared had an average diameter of 690 nm and were monodisperse (CV 3.9 %) [127]. However, they found that the dye slowly diffuses throughout the entire particle, thus rendering homogeneous fluorescent spheres. Using 2-DisP, Song et al. [101] were able to synthesize monodisperse controlled micron- sized colored polystyrene particles with polymerizable dyes. These polymerizable dyes have a methacrylate moiety for copolymerization with styrene and suitable spacers between the dye nucleus and the methacrylate function to promote the solubility of the dye. As explained in section 1.1.2.4, in 2-DisP the addition of the dye comonomer was deferred until the nucleation stage was complete [58]. These dyes became incorporated into the particles without disturbing the final particle size and size distribution. Using GPC, they showed that all of the dye was covalently incorporated into the PS particles, and the UV-Vis spectra showed that the covalently- incorporated dyes possessed the same spectra as the free dye.

1.1.4.2 Precious Metal-Containing Microspheres

Lee et al. [128] synthesized PS beads by dispersion polymerization using PVP stabilizer in ethanol water mixture. They obtained particles with a diameter of 2.7 µm and narrow size distribution. These PS beads were sulfonated by using chlorosulfonic acid. Surface-sulfonated microspheres were then metalized by a cationic gold ligand (dichlorophenanthrolinegold(III) chloride) to form a 1-4 nm metallic gold layer. This fabrication method showed a great deal of potential to control the gold loading on the PS bead surface with a uniform coverage in a simple procedure. Chen et al. [129] used dispersion polymerization of styrene and a poly(N- isopropylacrylamide) macromonomer in ethanol−water media in the presence of silver nitrate to synthesize silver nanoparticle-coated microspheres. The PS microspheres had diameters ranging from 530 to 1250 nm and PDI of 1.01 - 1.02. Well-dispersed silver nanoparticles (ca. 15 nm)

20 General Introduction

were formed in situ on the surface of poly(N-isopropylacrylamide)-coated PS microspheres and over 95.8% of the silver ions were converted into zero-valent metal and immobilized on the microspheres. The authors claimed that the surface-grafted PNIPAAm chains had two functions. First, they serve as steric stabilizers to prevent the flocculation of the polystyrene particles. In addition, PNIPAAm chains adsorb the silver nanoparticles onto the surface of the microspheres.

1.1.4.3 Magnetic Polymer Microspheres

Because of the ease and simplicity of magnetic separation technique compared to other separation techniques such as centrifugal separation, magnetic polymer microspheres have been used as platforms in various biological applications like cell labeling, cell separation, immune- and oligonucleotide-assays [87, 130-138]. In addition, these magnetic polymer particles are used for small-scale affinity separations, especially for cell separations followed by flow cytometry analysis or fluorescence-activated cell sorting (FACS). Magnetic polymer particles are typically acrylate- or styrene-based particles that contain superparamagnetic iron oxide. Different methods have been used in the production of magnetic particles. For example, Horak et al. [87] prepared uniform magnetic iron oxide nanoparticles, coated with poly(ethylene oxide) or poly(vinylpyrrolidone) with ca. 10 nm diameter. The encapsulation of magnetite nanoparticles in poly(2-hydroxyethyl methacrylate) (PHEMA) or poly(2-hydroxyethyl methacrylate-co-glycidyl methacrylate) P(HEMA-co-GMA) microparticles was achieved by dispersion polymerization in toluene/2-methylpropan-1-ol and PVP as a stabilizer. The produced microspheres had up to 20 wt. % iron without aggregation and diameter range from 0.3 to 4.0 µm with a PDI of ca. 1.03.

1.1.4.4 Polymer Microspheres Functionalized with Silica

Silica coating on polystyrene particles provides a simple method for the functionalization of their surfaces. The silica coating can then be used for a further functionalization of the particle surfaces with other functional groups such as amine and aldehyde groups. Using dispersion polymerization, the reverse strategy where silica particles were coated with functional polymer shell was also reported [139, 140].

21 General Introduction

Figure 1-4: Illustration of the preparation of PEI-stabilized PS microspheres by dispersion polymerization and subsequent silica shell formation. Reproduced from Ref. [141] with permission.

Hong et al. synthesized polystyrene (PS) microspheres via dispersion polymerization in alcoholic media [141]. A cationic polyelectrolyte, polyethyleneimine (PEI) was used as a steric stabilizer. It was found that PEI served as an effective stabilizer. The PS microspheres synthesized had a diameter range of 0.5 – 1.34 m and a CVd of 22 – 14 %. These PEI-stabilized PS particles were then used as seeds for growing a silica shell using tetraethyl orthosilicate (TEOS), as illustrated by Figure 1-4. The interaction between PEI and TEOS enabled the formation of a robust silica layer on the PS microspheres.

1.2 Flow Cytometry

Flow cytometry is a technique that permits rapid measurements on particles or cells as they flow in a fluid stream one by one (in a cytometric way) through a sensing point. The important feature of flow cytometric analysis is that measurements are made separately on each particle within the suspension in turn, and not just as average values for the whole population. The ability of flow cytometry to measure multiple cell parameters led to increasingly widespread use of it in biological and medicinal research and applications. For example, commercially available Luminex instruments [4] are flow cytometers equipped with two lasers and were used in molecular biology, pathology, immunology, plant biology and marine biology as well as broad application in medicine, especially in transplantation, hematology, tumor immunology and chemotherapy [142-152].

22 General Introduction

By loading different encoding fluorescent dyes onto different polymer classifier particles,

distinguishable barcodes can be produced. The variability (VR) is defined as the number of distinguishable encoded particles that can be made by one encoding system. The variability depends upon two variables: the number of encoding elements (fluorescent dyes, N) that can be loaded into the particles and the number of concentration levels (K) of each fluorescent dye that can be distinguished by the fluorescence detector [153]. The variability (VR) can be calculated by equation 1-1:

N (VR) = K −1 1-1 In an immunoassay, antibodies are used as an affinity reagent due to their high binding affinity and specificity for the analyte of interest (antigen) in the presence of many other substances in a sample. In a “multiplexed” immunoassay, more than one analyte (antigen) is determined simultaneously in a sample [154, 155]. For example, the flow cytometry system of Luminex [4], which is dedicated to the task of identifying a bead set using Luminex 100 flow cytometery. Luminex manufactures different bead sets with unique ratios of red and infrared dyes (N = 2). Each of the dyes can be at any of ten levels of concentration (K = 10). Theoretically, according to Equation 1-1, 99 distinguishable beads are available to simultaneously identify one hundred analytes in one sample. My goal was to synthesize polymer particles encoded with large numbers of metal ions (as encoding elements) to be used as classifier beads for mass cytometry-based immune- and gene-assays. I took guidance from Luminex beads in the encoding protocol for making the classifier particles with different combination of lanthanide metals at different concentration levels. In the mass cytometry technique [20], a mass spectrometer is used to read the barcode instead of the fluorometer used for conventional flow cytometers. In this project we use elements from the lanthanide series as the metal coding elements. In the following section, I give some general physical and chemical properties of the lanthanide elements. In Chapter 3, I discuss the reasons for choosing the lanthanide series as metal codes.

23 General Introduction

1.3 Lanthanides: Physical and Chemical properties

Lanthanide series (Ln’s) comprise lanthanum (La, atomic number 57), to lutetium (Lu, atomic number 71). The set of Ln’s may be divided into two groups: the light lanthanide elements (La, Ce, Pr, Nd, Pm, Sm, Eu) and the heavy rare earth elements (Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu) [156]. In particular the lanthanides (La through Lu) possess similar physical and chemical properties and are therefore difficult to separate. Their common properties are associated with their particular atomic structure. Each has two outer electrons and eight or nine in the next inner shell [157]. The configurations of outer electrons for lanthanide atoms and lanthanide 3+ ions are given in Table 1-2. The addition of electrons to the inner shells makes little change in many of their physical and chemical properties. All Lanthanides are relatively soft and malleable as pure metals, and have a bright silver lustre. When finely divided, the metal powder oxidizes quickly when exposed to air. When present in solid lumps, the metal will not undergo spontaneous combustion, but will oxidize slowly, comparable to the rusting of other metals. Because of the similarities in the chemistry of lanthanides, many authors discuss the characteristics of the lanthanides as a group. Within this group, however, there are substantial differences among the individual Lanthanides and their compounds. Some of the Lanthanides exhibit more than one oxidation state, but all commonly favour the (III) oxidation state. Most lanthanides form lanthanide 3+ ions by losing two electrons from the 6s shell and one electron from the 4f shell. However, for La, Ce, Gd, and Lu the third electron resides in the 5d shell and these lanthanides therefore have a significantly lower third ionization potential. In nature, all lanthanides strictly occur in the (III) oxidation state, except Ce and Eu, which also occur in the (IV) and (II) oxidation state, respectively [157]. The 4f valence electrons of the lanthanides are situated in the inner orbitals and are not directly available for bonding. The lanthanides form essentially electrostatic bonds and thus their complexes are labile. Hence it is difficult to predict the conformation of a lanthanide complex, and it is difficult to isolate optical isomers of a lanthanide complex. The lanthanides have a wide range of coordination numbers, e.g. 3 to 12. The stoichiometry and the conformation of a lanthanide complex often depends on the method of preparation. The lanthanide contraction can affect the stoichiometry and structure of the complexes of light and heavy lanthanides. In some

24 General Introduction

cases isolation of complexes of lanthanides is possible either with members in the beginning or at the end of the lanthanide series.

Table 1-2: Configuration of outer electrons of Ln atoms and Ln3+ ions.

Lanthanide Z1 Configuration of outer electrons Ion radius Atom Ion (3+) 4f 5s 5p 5d 5f 6s 4f 5s 5p Å La 57 2 6 1 2 2 6 1.06 Ce 58 1 2 6 1 2 1 2 6 1.03 Pr 59 3 2 6 2 2 2 6 1.01 Nd 60 4 2 6 2 3 2 6 0.995 Pm 61 5 2 6 2 4 2 6 0979 Sm 62 6 2 6 2 5 2 6 0.964 Eu 63 7 2 6 2 6 2 6 0.950 Gd 64 7 2 6 1 2 7 2 6 0.938 Tb 65 9 2 6 2 8 2 6 0.924 Dy 66 10 2 6 2 9 2 6 0.908 Ho 67 11 2 6 2 10 2 6 0.894 Er 68 12 2 6 2 11 2 6 0.881 Tm 69 13 2 6 2 12 2 6 0.869 Yb 70 14 2 6 2 13 2 6 0.858 Lu 71 14 2 6 1 2 14 2 6 0.848

Z1: atomic number

25 General Introduction

1.4 Research objectives

When I started my Ph. D. research in January 2007, I was asked to prepare functional monodisperse micron-size polymer particles for a lanthanide tagging project. I started with the miniemulsion polymerization technique. I found that I could synthesize Ln-encoded particles with miniemulsion but the size was too small for the project requirements. I was obliged either to modify the old process or develop a new process, which should be simple, robust, and suitable for making various microspheres at a low cost. Once I discovered the suitability of two-stage dispersion polymerization for synthesizing Ln-encoded polystyrene particles, the objective of my PhD research became to improve the lanthanide content and modify the particle surface for bioconjugation. I chose to focus on the following topics: [1] To synthesize surface functionalized monodisperse polystyrene microspheres. [2] To synthesize a library of lanthanide-encoded particles. [3] To investigate the stability of the particles and ability to maintain their Ln content. [4] To explore possible ways to biofunctionalize the lanthanide-encoded particles.

1.5 Thesis outline

The thesis contains eight chapters including this chapter. Chapter 2 details materials and experimental methods used to carry out the experiments. Chapter 3 describes the two-stage dispersion polymerization for the synthesis of Ln-encoded polystyrene particles and the encoding protocols for the metal tags. Chapter 4 reports the results for the addition of a 3rd stage to dispersion polymerization as a trial to improve the size and lanthanide content variation as well as functionalizing the surface of the particles. Chapter 5 concerns the application of two different techniques to modify the surface of the particles with functional shell. In Chapter 6, I present the research done regarding the usage of Si-coated particles in immune- and gene-assays. Finally, Chapter 7 deals with some analytical aspects pertaining the creation of mass cytometry standards

26 General Introduction

using Ln-encoded PS particles. Portions of this thesis have been published and presented in scientific conferences. I also list manuscripts submitted or in preparation: 1. Ahmed I. Abdelrahman, Stuart C. Thickett, Olga Ornatsky, Vladimir Baranov and Mitchell A. Winnik “Surface Functionalization Methods to Enhance Bioconjugation in Metal- Labeled Polystyrene Particles for Immunoassays” Macromolecules, 10.1021/ma200582q.

2. Ahmed I. Abdelrahman, Olga Ornatsky, Dmitry Bandura, Vladimir Baranov, Robert Kinach, Sheng Dai, Stuart C. Thickett, Scott Tanner and Mitchell A. Winnik, “Metal- containing polystyrene beads as standards for mass cytometry” J. Anal. At. Spectrom., 2010, 25, 260 – 268.

3. Stuart C. Thickett, Ahmed I. Abdelrahman, Olga Ornatsky, Dmitry Bandura, Vladimir Baranov and Mitchell A. Winnik, “Bio-functional, lanthanide-labeled polymer particles by seeded emulsion polymerization and their characterization by novel ICP-MS detection” J. Anal. At. Spectrom., 2010, 25, 269 – 281

4. Ahmed I. Abdelrahman, Sheng Dai, Stuart C. Thickett, Olga Ornatsky, Dmitry Bandura, Vladimir Baranov, and Mitchell A. Winnik, “Lanthanide-Containing Polymer Microspheres by Multiple-Stage Dispersion Polymerization for Highly Multiplexed Bioassays” J. Am. Chem. Soc., 2009, 131, 15276 – 15283.

5. Jai Il Park, Zhihong Nie, Alexander Kumachev, Ahmed I. Abdelrahman, Bernard P. Binks, Howard A. Stone, Eugenia Kumacheva, “A Microfluidic Approach to Chemically Driven Assembly of Colloidal Particles at Gas-Liquid Interfaces”, Angew. Chem. Int. Ed. 2009, 48, 1 – 6.

6. Cedric Vancaeyzeele, Olga Ornatsky, Vladimir Baranov, Lei Shen, Ahmed Abdelrahman, and Mitchell A. Winnik, “Lanthanide-Containing Polymer Nanoparticles for Biological Tagging Applications: Nonspecific Endocytosis and Cell Adhesion” J. Am. Chem. Soc., 2007, 129, 13653 – 3660.

7. Wanjuan Lin, Xiaomei Ma, Jieshu Qian, Ahmed I. Abdelrahman, Adrienne Halupa, Vladimir Baranov, Andrij Pich, Mitchell A. Winnik, “Synthesis and Mass Cytometric Analysis of Lanthanide-encoded Polyelectrolyte Microgels” Langmuir

8. Ahmed I. Abdelrahman, Dirk Weinrich, Yi Liang, Olga Ornatsky, Vladimir Baranov and Mitchell A. Winnik, “Mass Cytometry based Gene-assays” In preparation.

9. Sheng Dai, Ahmed I. Abdelrahman, Daniel Majonis, Olga Ornatsky, Vladimir Baranov, Mark Nitz and Mitchell A. Winnik, “Lanthanide-Containing Functional Photoluminescence bar-coded resins towards bead-based biodiagnosis” In preparation.

27 General Introduction

10. Yi Liang, Ahmed I. Abdelrahman, Vladimir Baranov and Mitchell A. Winnik, “Lanthanide-Encoded Poly(Styrene-co-Methacrylic Acid) Microspheres: Synthesis and Metal Ion Release Study” In preparation.

11. Wanjuan Lin, Ahmed I. Abdelrahman,Yi Hou, Jieshu Qian ,Dirk Weinrich, Arienne Halupa, Vladimir Baranov, Mitchell A. Winnik, “Lanthanide Encoded Microgels for Highly Multiplexed Mass Spectrometry Based Bioassays” In preparation.

References:

1. Zheng W, He L: Label-Free, Real-Time Multiplexed DNA Detection Using Fluorescent Conjugated Polymers. 2009, 131(10):3432-3433. 2. Hardenbol P, Yu F, Belmont J, MacKenzie J, Bruckner C, Brundage T, Boudreau A, Chow S, Eberle J, Erbilgin A et al: Highly multiplexed molecular inversion probe genotyping: Over 10,000 targeted SNPs genotyped in a single tube assay. Genome Research 2005, 15(2):269-275. 3. Goldman ER, Clapp AR, Anderson GP, Uyeda HT, Mauro JM, Medintz IL, Mattoussi H: Multiplexed Toxin Analysis Using Four Colors of Quantum Dot Fluororeagents. Anal Chem 2004, 76(3):684-688. 4. Dunbar SA, Vander Zee CA, Oliver KG, Karem KL, Jacobson JW: Quantitative, multiplexed detection of bacterial pathogens: DNA and protein applications of the Luminex LabMAPâ„¢ system. J Microbiol Methods 2003, 53(2):245-252. 5. Nolan JP, Sklar LA: Suspension array technology: evolution of the flat-array paradigm. Trends Biotechnol 2002, 20(1):9-12. 6. Han M, Gao X, Su JZ, Nie S: Quantum-dot-tagged microbeads for multiplexed optical coding of biomolecules. Nat Biotechnol 2001, 19(7):631-635. 7. Cook EB, Stahl JL, Lowe L, Chen R, Morgan E, Wilson J, Varro R, Chan A, Graziano FM, Barney NP: Simultaneous measurement of six cytokines in a single sample of human tears using microparticle-based flow cytometry: Allergics vs. non-allergics. Journal of Immunological Methods 2001, 254(1-2):109-118. 8. Brenner S, Johnson M, Bridgham J, Golda G, Lloyd DH, Johnson D, Luo S, McCurdy S, Foy M, Ewan M et al: Gene expression analysis by massively parallel signature sequencing (MPSS) on microbead arrays. Nat Biotechnol 2000, 18(6):630-634. 9. Bailey RC, Kwong GA, Radu CG, Witte ON, Heath JR: DNA-Encoded Antibody Libraries: A Unified Platform for Multiplexed Cell Sorting and Detection of Genes and Proteins. 2007, 129(7):1959-1967. 10. Epstein JR, Ferguson JA, Lee KH, Walt DR: Combinatorial Decoding: An Approach for Universal DNA Array Fabrication. J Am Chem Soc 2003, 125(45):13753-13759. 11. Harma H, Soukka T, Lovgren T: Europium nanoparticles and time-resolved fluorescence for ultrasensitive detection of prostate-specific antigen. Clin Chem 2001, 47(3):561-568.

28 General Introduction

12. Morgan E, Varro R, Sepulveda H, Ember JA, Apgar J, Wilson J, Lowe L, Chen R, Shivraj L, Agadir A et al: Cytometric bead array: A multiplexed assay platform with applications in various areas of biology. Clinical Immunology 2004, 110(3):252-266. 13. Na K, Kim S, Park K, Kim K, Dae GW, Ick CK, Chung HM, Park KH: Heparin/poly(l- lysine) nanoparticle-coated polymeric microspheres for stem-cell therapy. J Am Chem Soc 2007, 129(18):5788-5789. 14. Nam JM, Thaxton CS, Mirkin CA: Nanoparticle-based bio-bar codes for the ultrasensitive detection of proteins. Science 2003, 301(5641):1884-1886. 15. Luminex Corp http://wwwluminexcorpcom/products 2009. 16. Yang L, Tran DK, Wang X: BADGE, BeadsArray for the Detection of Gene Expression, a high-throughput diagnostic bioassay. Genome Research 2001, 11(11):1888-1898. 17. Schwenk JM, Lindberg J, Sundberg M, Uhlén M, Nilsson P: Determination of binding specificities in highly multiplexed bead-based assays for antibody proteomics. Molecular and Cellular Proteomics 2007, 6(1):125-132. 18. Yurkovetsky ZR, Kirkwood JM, Edington HD, Marrangoni AM, Velikokhatnaya L, Winans MT, Gorelik E, Lokshin AE: Multiplex analysis of serum cytokines in melanoma patients treated with interferon-alpha 2b. Clinical Cancer Research 2007, 13(8):2422-2428. 19. Vancaeyzeele C, Ornatsky O, Baranov V, Shen L, Abdelrahman A, Winnik MA: Lanthanide-containing polymer nanoparticles for biological tagging applications: Nonspecific endocytosis and cell adhesion. J Am Chem Soc 2007, 129(44):13653- 13660. 20. Bandura DR, Baranov VI, Ornatsky OI, Antonov A, Kinach R, Lou X, Pavlov S, Vorobiev S, Dick JE, Tanner SD: Mass cytometry: Technique for real time single cell multitarget immunoassay based on inductively coupled plasma time-of-flight mass spectrometry. Anal Chem 2009, 81(16):6813-6822. 21. Ornatsky OI, Lou X, Nitz M, Schafer S, Sheldrick WS, Baranov VI, Bandura DR, Tanner SD: Study of cell antigens and intracellular DNA by identification of element- containing labels and metallointercalators using inductively coupled plasma mass spectrometry. Anal Chem 2008, 80(7):2539-2547. 22. Tanner SD, Bandura DR, Ornatsky O, Baranov VI, Nitz M, Winnik MA: Flow cytometer with mass spectrometer detection for massively multiplexed single-cell biomarker assay. Pure Appl Chem 2008, 80(12):2627-2641. 23. Harkins WD: A general theory of the reaction loci in emulsion polymerization. J Chem Phys 1945, 13(9):381-382. 24. Harkins WD: A general theory of the mechanism of emulsion polymerization. J Am Chem Soc 1947, 69(6):1428-1444. 25. Smith WV, Ewart RH: Kinetics of emulsion polymerization. J Chem Phys 1948, 16(6):592-599. 26. Smith WV: The kinetics of styrene emulsion polymerization. J Am Chem Soc 1948, 70(11):3695-3702. 27. Antonietti M, Bremser W, Muschenborn D, Rosenauer C, Schupp B, Schmidt M: Synthesis and size control of polystyrene lattices via polymerization in microemulsion. Macromolecules 1991, 24(25):6636-6643.

29 General Introduction

28. Deng Y, Wang L, Yang W, Fu S, Elaissari A: Preparation of magnetic polymeric particles via inverse microemulsion polymerization process. J Magn Magn Mater 2003, 257(1):69-78. 29. Guo JS, Elaasser MS, Vanderhoff JW: Microemulsion polymerization of styrene. Journal of Polymer Science Part a-Polymer Chemistry 1989, 27(2):691-710. 30. Guo JS, Sudol ED, Vanderhoff JW, Elaasser MS: Particle nucleation and monomer partitioning in styrene o/w microemulsion polymerization. Journal of Polymer Science Part a-Polymer Chemistry 1992, 30(5):691-702. 31. Cho I, Lee KW: Morphology of latex-particles formed by poly(methyl methacrylate)- seeded emulsion polymerization of styrene. J Appl Polym Sci 1985, 30(5):1903-1926. 32. Hawkett BS, Napper DH, Gilbert RG: Seeded emulsion polymerization of styrene. Journal of the Chemical Society-Faraday Transactions I 1980, 76:1323-1343. 33. Liu ZG, Han Y, Zhou C, Zhang MY, Li WM, Zhang HX, Liu FQ, Liu WJ: Seeded Emulsion Polymerization of Butyl Acrylate Using a Redox Initiator System: Kinetics and Mechanism. Industrial & Engineering Chemistry Research 2010, 49(16):7152-7158. 34. Wang JS, Deng W, Chen YS, Kan CY: Preparation of sulphonate-containing core/shell latex particles via seeded emulsion copolymerization. Chin Chem Lett 2010, 21(10):1259-1262. 35. Yuan HG, Kalfas G, Ray WH: Suspension polymerization. J Macromol Sci, Rev Macromol Chem Phys 1991, C31(2-3):215-299. 36. Brooks BW: Suspension Polymerization Processes. Chemical Engineering & Technology 2010, 33(11):1737-1744. 37. Barrett KEJ, Thomas HR: Kinetics of dispersion polymerization of soluble monomers .I. Methyl methacrylate. Journal of Polymer Science Part a-1-Polymer Chemistry 1969, 7(9PA1):2621. 38. Almog Y, Levy M: Dispersion polymerization of styrene: Effect of surfactant. Journal of Polymer Science: Polymer Chemistry Edition 1981, 19(1):115-126. 39. Ober CK, Lok KP, Hair ML: Monodispersed, micron-sized polystyrene particles by dispersion polymerization. Journal of Polymer Science Part C-Polymer Letters 1985, 23(2):103-108. 40. Tseng CM, Lu YY, Elaasser MS, Vanderhoff JW: Uniform polymer particles by dispersion polymerization in alcohol. Journal of Polymer Science Part a-Polymer Chemistry 1986, 24(11):2995-3007. 41. Ober CK, Lok KP: Formation of large monodisperse copolymer particles by dispersion polymerization. Macromolecules 1987, 20(2):268-273. 42. Paine AJ: Dispersion polymerization of styrene in polar-solvents .1. Grafting mechanism of stabilization by hydroxypropyl cellulose.J Colloid Interface Sci 1990, 138(1):157-169. 43. Paine AJ: Dispersion polymerization of styrene in polar-solvents .4. Solvency control of particle-size from hydroxypropyl cellulose stabilized polymerizations. Journal of Polymer Science Part a-Polymer Chemistry 1990, 28(9):2485-2500. 44. Kawaguchi S, Ito K: Dispersion polymerization. Adv Polym Sci 2005, 175:299-328.

30 General Introduction

45. Almog Y, Levy M: Studies of particle size distribution and adsorption of stabilizers in dispersion polymerization. Journal of Polymer Science: Polymer Chemistry Edition 1982, 20(2):417-429. 46. Bromberg L: Polyether-modified poly(acrylic acid): Synthesis and applications. Ind Eng Chem Res 1998, 37(11):4267-4274. 47. Kawaguchi H: Functional polymer microspheres. Progress in Polymer Science (Oxford) 2000, 25(8):1171-1210. 48. Stejskal J: Colloidal dispersions of conducting polymers. J Polym Mater 2001, 18(3):225-258. 49. Leobandung W, Ichikawa H, Fukumori Y, Peppas NA: Monodisperse nanoparticles of poly(ethylene glycol) macromers and N-isopropyl acrylamide for biomedical applications. J Appl Polym Sci 2002, 87(10):1678-1684. 50. Torres-Lugo M, Peppas NA: Preparation and characterization of P(MAA-g-EG) nanospheres for protein delivery applications. J Nanopart Res 2002, 4(1-2):73-81. 51. Zhang S, Huang X, Yao N, Horváth C: Preparation of monodisperse porous polymethacrylate microspheres and their application in the capillary electrochromatography of macrolide antibiotics. J Chromatogr A 2002, 948(1-2):193- 201. 52. Riess G, Labbe C: Block Copolymers in Emulsion and Dispersion Polymerization. Macromol Rapid Commun 2004, 25(2):401-435. 53. S panová A, Horák D, Soudková E, Rittich B: Magnetic hydrophilic methacrylate- based polymer microspheres designed for polymerase chain reactions applications. Journal of Chromatography B: Analytical Technologies in the Biomedical and Life Sciences 2004, 800(1-2):27-32. 54. Choi HJ, Park SJ, Kim ST, Jhon MS: Electrorheological application of polyaniline/multi-walled carbon nanotube composites. Diamond Relat Mater 2005, 14(3-7):766-769. 55. Jang J, Nam Y, Yoon H: Fabrication of polypyrrole-poly(N-vinylcarbazole) core-shell nanoparticles with excellent electrical and optical properties. Adv Mater 2005, 17(11):1382-1386. 56. Nagarale RK, Gohil GS, Shahi VK, Rangarajan R: Preparation of organic-inorganic composite anion-exchange membranes via aqueous dispersion polymerization and their characterization. J Colloid Interface Sci 2005, 287(1):198-206. 57. Yang SC, Yang DJ, Kim J, Hong JM, Kim HG, Kim ID, Lee H: Hollow TiO2 hemispheres obtained by colloidal templating for application in dye-sensitized solar cells. Adv Mater 2008, 20(5):1059-1064. 58. Song JS, Tronc F, Winnik MA: Two-stage dispersion polymerization toward monodisperse, controlled micrometer-sized copolymer particles. J Am Chem Soc 2004, 126(21):6562-6563. 59. Odian G: Principles of Polymerization third edition edn: John Wiley and Sons; 1991. 60. Horák D: Uniform polymer beads of micrometer size. Acta Polym 1996, 47(1):20-28. 61. Lok KP, Ober CK: Particle-size control in dispersion polymerization of polystyrene. Canadian Journal of Chemistry-Revue Canadienne De Chimie 1985, 63(1):209-216.

31 General Introduction

62. Ober CK, Lok KP, Hair ML: Monodispersed, micron-sized polymer particles - preparation and physicochemical properties. Abstracts of Papers of the American Chemical Society 1985, 189(APR-):104-COLL. 63. Ober CK, Vangrunsven F, McGrath M, Hair ML: Partitioning of monomer during dispersion polymerization. Colloids and Surfaces 1986, 21:347-354. 64. Shen S, Sudol ED, Elaasser MS: Dispersion polymerization of methyl-methacrylate - mechanism of particle formation. Journal of Polymer Science Part a-Polymer Chemistry 1994, 32(6):1087-1100. 65. Bamnolker H, Margel S: Dispersion polymerization of styrene in polar solvents: Effect of reaction parameters on microsphere surface composition and surface properties, size and size distribution, and molecular weight. Journal of Polymer Science Part a-Polymer Chemistry 1996, 34(10):1857-1871. 66. Paine AJ, Luymes W, McNulty J: Dispersion polymerization of styrene in polar- solvents .6. Influence of reaction parameters on particle-size and molecular-weight in poly(n-vinylpyrrolidone)-stabilized reactions. Macromolecules 1990, 23(12):3104- 3109. 67. Deslandes Y, Mitchell DF, Paine AJ: X-ray Photoelectron Spectroscopy and static time-of-Flight Secondary Ion Mass Spectrometry study of dispersion polymerized polystyrene latexes. Langmuir 1993, 9(6):1468-1472. 68. Horak D, Svec F, Frechet JMJ: Preparation and control of surface-properties of monodisperse micrometer size beads by dispersion copolymerization of styrene and butyl methacrylate in polar media. Journal of Polymer Science Part a-Polymer Chemistry 1995, 33(14):2329-2338. 69. Paine AJ: Dispersion polymerization of styrene in polar-solvents .7. A simple mechanistic model to predict particle-size. Macromolecules 1990, 23(12):3109-3117. 70. Ugelstad J, Berge A, Ellingsen T, Schmid R, Nilsen TN, Mork PC, Stenstad P, Hornes E, Olsvik O: Preparation and application of new monosized polymer particles. Prog Polym Sci 1992, 17(1):87-161. 71. Tuncel A, Kahraman R, Piskin E: Monosize polystyrene microbeads by dispersion polymerization. J Appl Polym Sci 1993, 50(2):303-319. 72. Tuncel A, Kahraman R, Piskin E: Monosize polystyrene latices carrying functional- groups on their surfaces. J Appl Polym Sci 1994, 51(8):1485-1498. 73. Kobayashi S, Uyama H, Lee SW, Matsumoto Y: Preparation of micron-size monodisperse polymer particles by dispersion copolymerization of styrene with poly(2-oxazoline) macromonomer. Journal of Polymer Science Part a-Polymer Chemistry 1993, 31(12):3133-3139. 74. Ahmed SF, Poehlein GW: Kinetics of dispersion polymerization of styrene in ethanol .1. Model development. Industrial & Engineering Chemistry Research 1997, 36(7):2597-2604. 75. Ahmed SF, Poehlein GW: Kinetics of dispersion polymerization of styrene in ethanol .2. Model validation. Industrial & Engineering Chemistry Research 1997, 36(7):2605- 2615. 76. Cao K, Li BF, Huang Y, Li BG, Pan ZR: Mechanism and model of dispersion polymerization using homopolymer as dispersant in polar media. Macromolecular Symposia 2000, 150:187-194.

32 General Introduction

77. Chernyshev AV, Soini AE, Maltsev VP, Soini E: A model of complete classical treatment of dispersion radical polymerization kinetics. Macromolecules 1998, 31(19):6455-6460. 78. Quintero-Ortega IA, Garcia-Moran PR, Jaramillo-Soto G, Vivaldo-Lima E, Luna- Barcenas G: Evaluation of the Performance of a Kinetic Model for Free-Radical Copolymerization of Vinyl/Divinyl Monomers in Supercritical Carbon Dioxide. Macromolecular Symposia 2009, 283-84:103-109. 79. Yasuda M, Seki H, Yokoyama H, Ogino H, Ishimi K, Ishikawa H: Simulation of a particle formation stage in the dispersion polymerization of styrene. Macromolecules 2001, 34(10):3261-3270. 80. Yasuda M, Yokoyama H, Seki H, Ogino H, Ishimi K, Ishikawa H: Simulation of particle growth in the dispersion polymerization of styrene: The termination rate constant in particles. Macromol Theory Simul 2001, 10(1):54-62. 81. Takahashi K, Miyamori S: Effects of polymerization conditions on particle-size in dispersion polymerization of methyl-methacrylate in mixed hydrophilic media. Kobunshi Ronbunshu 1993, 50(4):251-257. 82. Kim OK, Lee K, Kim K, Lee BH, Choe S: Optimum conditions for preparing micron- sized PMMA beads in the dispersion polymerization using PVA. Colloid Polym Sci 2006, 284(8):909-915. 83. Awan MA, Dimonie VL, ElAasser MS: Anionic dispersion polymerization of styrene .1. Investigation of parameters for preparation of uniform micron-size polystyrene particles with narrow molecular weight distribution. Journal of Polymer Science Part a-Polymer Chemistry 1996, 34(13):2633-2649. 84. Chen Y, Yang HW: Hydroxypropyl cellulose (hpc)-stabilized dispersion polymerization of styrene in polar-solvents - effect of reaction parameters. Journal of Polymer Science Part a-Polymer Chemistry 1992, 30(13):2765-2772. 85. Horak D: Effect of reaction parameters on the particle size in the dispersion polymerization of 2-hydroxyethyl methacrylate. Journal of Polymer Science Part a- Polymer Chemistry 1999, 37(20):3785-3792. 86. Horak D, Krystufek M, Spevacek J: Effect of reaction parameters on time dispersion polymerization of 1-vinyl-2-pyrrolidone. Journal of Polymer Science Part a-Polymer Chemistry 2000, 38(3):653-663. 87. Horak D, Semenyuk N, Lednicky F: Effect of the reaction parameters on the particle size in the dispersion polymerization of 2-hydroxyethyl and glycidyl methacrylate in the presence of a ferrofluid. Journal of Polymer Science Part a-Polymer Chemistry 2003, 41(12):1848-1863. 88. Lee KC, Her JH, Kim KJ: Effects of Reaction Parameters on Particle Size, Molecular Weight, and Kinetics for Dispersion Polymerization of n-Butyl Acrylate. Int J Polym Anal Charact 2009, 14(7):600-616. 89. Jayachandran KNN, Chatterji PR: Preparation of linear and crosslinked polymer microspheres by dispersion polymerization. Journal of Macromolecular Science- Polymer Reviews 2001, C41(1-2):79-94. 90. Mackova H, Horak D: Effects of the reaction parameters on the properties of thermosensitive poly(N-isopropylacrylamide) microspheres prepared by

33 General Introduction

precipitation and dispersion polymerization. Journal of Polymer Science Part a- Polymer Chemistry 2006, 44(2):968-982. 91. Saikia PJ, Lee JM, Lee K, Choe S: Reaction parameters in the RAFT mediated dispersion polymerization of styrene. Journal of Polymer Science Part a-Polymer Chemistry 2008, 46(3):872-885. 92. Uyama H, Kobayashi S: Dispersion polymerization of styrene in aqueous alcohol solution - effects of reaction parameters on the polymer particle formation. Polym Int 1994, 34(3):339-344. 93. Saenz JM, Asua JM: Kinetics of the dispersion copolymerization of styrene and butyl acrylate. Macromolecules 1998, 31(16):5215-5222. 94. Saenz JM, Asua JM: Dispersion copolymerization of styrene and butyl acrylate in polar solvents. Journal of Polymer Science, Part A: Polymer Chemistry 1996, 34(10):1977-1992. 95. Huang JX, Zhang HT, Hou JL, Jiang PP: Preparation of micron-size crosslinked microspheres by dispersion copolymerization of polystyrene/2,2 '-oxy-bisethanol diacrylate. Reactive & Functional Polymers 2002, 53(1):1-9. 96. Giaconia A, Filardo G, Scialdone O, Galia A: Continuous reaction system to investigate the dispersion polymerization of vinyl monomers in supercritical carbon dioxide. Journal of Polymer Science Part a-Polymer Chemistry 2006, 44(13):4122-4135. 97. Okubo M, Ikegami K, Yamamoto Y: Preparation of micron-size monodisperse polymer microspheres having chloromethyl group. Colloid Polym Sci 1989, 267(3):193-200. 98. Ober CK: Dispersion copolymerization in nonaqueous media. Makromolekulare Chemie-Macromolecular Symposia 1990, 35-6:87-104. 99. Kim JW, Kim BS, Suh KD: Monodisperse micron-sized cross-linked polystyrene particles. VI. Understanding of nucleated particle formation and particle growth. Colloid Polym Sci 2000, 278(6):591-594. 100. Song JS, Chagal L, Winnik MA: Monodisperse micrometer-size carboxyl- functionalized polystyrene particles obtained by two-stage dispersion polymerization. Macromolecules 2006, 39(17):5729-5737. 101. Song JS, Tronc F, Winnik MA: Monodisperse, controlled micron-size dye-labeled polystyrene particles by two-stage dispersion polymerization. Polymer 2006, 47(3):817-825. 102. Song JS, Winnik MA: Cross-linked, monodisperse, micron-sized polystyrene particles by two-stage dispersion polymerization. Macromolecules 2005, 38(20):8300- 8307. 103. Song JS, Winnik MA: Monodisperse, micron-sized reactive low molar mass polymer microspheres by two-stage living radical dispersion polymerization of styrene. Macromolecules 2006, 39(24):8318-8325. 104. Song JS, Winnik MA: Monodisperse, micrometer-sized low molar mass polystyrene particles by two-stage dispersion polymerization. Polymer 2006, 47(13):4557-4563. 105. Yang W, Yang D, Hu J, Wang C, Fu S: Dispersion copolymerization of styrene and other vinyl monomers in polar solvents. J Polym Sci, Part A: Polym Chem 2001, 39(4):555-561.

34 General Introduction

106. Song SN, Zhang W, Hu ZQ, Zhang ZC: Monodisperse micrometer-size carboxyl- functionalized polystyrene particles obtained by two-stage radiation-induced dispersion polymerization. Colloids and Surfaces a-Physicochemical and Engineering Aspects 2009, 348(1-3):1-8. 107. Prokopov NI, Gritskova IA, Cherkasov VR, Chalykh AE: Synthesis of monodisperse functional polymeric microspheres for immunoassay. Russian Chemical Reviews 1996, 65(2):167-180. 108. Basinska T: Poly(styrene/acrolein) microspheres for diagnostic applications. PhD Thesis. lodz; 1995. 109. Slomkowski S: Polyacrolein containing microspheres: Synthesis, properties and possible medical applications. Prog Polym Sci 1998, 23(5):815-874. 110. Fukutani K, Itoh T, Ihara E, Inoue K: Dispersion polymerization with amphiphilic block copolymer containing amino group. In: 2006; 2006: 379. 111. Ali AMI, Pareek P, Sewell L, Schmid A, Fujii S, Armes SP, Shirley IM: Synthesis of poly(2-hydroxypropyl methacrylate) latex particles via aqueous dispersion polymerization. Soft Matter 2007, 3(8):1003-1013. 112. Li SF, Yang XL, Huang WQ: Preparation of monodisperse crosslinked polymer microspheres having chloromethyl group by distillation-precipitation polymerization. Chin J Polym Sci 2005, 23(2):197-202. 113. Rembaum A, Chang M, Richards G, Li M: Structure and immunological properties of polyacrolein formed by means of ionizing radiation and base catalysis. Journal of Polymer Science: Polymer Chemistry Edition 1984, 22(3):609-621. 114. Margel S: Polyacrolein microspheres. Methods Enzymol 1985, VOL. 112:164-175. 115. Margel S, Beitler U, Ofarim M: Polyacrolein microspheres as a new tool in cell biology. J Cell Sci 1982, Vol. 56:157-175. 116. Margel S, Wiesel E: Acrolein polymerization: monodisperse, homo, and hybrido microspheres, synthesis, mechanism, and reactions. Journal of polymer science Part A-1, Polymer chemistry 1984, 22(1):145-158. 117. Ma Z, Guan Y, Liu H: Synthesis and characterization of micron-sized monodisperse superparamagnetic polymer particles with amino groups. Journal of Polymer Science, Part A: Polymer Chemistry 2005, 43(15):3433-3439. 118. Wang W, Zhang Q, Zhang B, Li D, Dong X, Zhang L, Chang J: Preparation of monodisperse, superparamagnetic, luminescent, and multifunctional PGMA microspheres with amino-groups. Chin Sci Bull 2008, 53(8):1165-1170. 119. Itoh T, Fukutani K, Hino M, Ihara E, Inoue K: Effects of polystyrene-b- poly(aminomethyl styrene)s as stabilizers on dispersion polymerization of styrene in alcoholic media. J Colloid Interface Sci 2009, 330(2):292-297. 120. Jennings TL, Rahman KS, Fournier-Bidoz S, Chan WCW: Effects of microbead surface chemistry on DNA loading and hybridization efficiency. Anal Chem 2008, 80(8):2849-2856. 121. Raez J, Biais DR, Zhang Y, Alvarez-Puebla RA, Bravo-Vasquez JP, Pezacki JP, Fenniri H: Spectroscopically encoded microspheres for antigen biosensing. Langmuir 2007, 23(12):6482-6485. 122. Bethell GS, Ayers JS, Hancock WS, Hearn MTW: Novel method of activation of crosslinked agaroses with 1,1'-carbonyldiimidazole which gives a matrix for affinity

35 General Introduction

chromatography devoid of additional charged groups. J Biol Chem 1979, 254(8):2572-2574. 123. Miron T, Wilchek M: A simplified method for the preparation of succinimidyl carbonate polyethylene glycol for coupling to proteins. Bioconjugate Chem 1993, 4(6):568-569. 124. Grabarek Z, Gergely J: Zero-length crosslinking procedure with the use of active esters. Anal Biochem 1990, 185(1):131-135. 125. Armstrong B, Stewart M, Mazumder A: Suspension arrays for high throughput, multiplexed single nucleotide polymorphism genotyping. Cytometry 2000, 40(2):102- 108. 126. . In: Luminex Corp http://wwwluminexcorpcom/products. 2009. 127. Bosma G, Pathmamanoharan C, de Hoog EHA, Kegel WK, van Blaaderen A, Lekkerkerker HNW: Preparation of Monodisperse, Fluorescent PMMA-Latex Colloids by Dispersion Polymerization. J Colloid Interface Sci 2002, 245(2):292-300. 128. Lee JH, Kim DO, Song GS, Lee Y, Jung SB, Nam JD: Direct metallization of gold nanoparticles on a polystyrene bead surface using cationic gold ligands. Macromol Rapid Commun 2007, 28(5):634-640. 129. Chen C-W, Serizawa T, Akashi M: Synthesis and Characterization of Poly(N- isopropylacrylamide)-Coated Polystyrene Microspheres with Silver Nanoparticles on Their Surfaces. Langmuir 1999, 15(23):7998-8006. 130. Fan ZH, Mangru S, Granzow R, Heaney P, Ho W, Dong QP, Kumar R: Dynamic DNA hybridization on a chip using paramagnetic beads. Anal Chem 1999, 71(21):4851- 4859. 131. Aytur T, Foley J, Anwar M, Boser B, Harris E, Beatty PR: A novel magnetic bead bioassay platform using a microchip-based sensor for infectious disease diagnosis. Journal of Immunological Methods 2006, 314(1-2):21-29. 132. Deng Y, Qi D, Deng C, Zhang X, Zhao D: Superparamagnetic high-magnetization microspheres with an Fe 3O4@SiO2 core and perpendicularly aligned mesoporous SiO2 shell for removal of microcystins. J Am Chem Soc 2008, 130(1):28-29. 133. Elaissari A: Magnetic Latex Particles in Nanobiotechnologies for Biomedical Diagnostic Applications: State of the Art. Macromolecular Symposia 2009, 281:14-19. 134. Fan AP, Lau CW, Lu JZ: Magnetic bead-based chemiluminescent metal immunoassay with a colloidal gold label. Anal Chem 2005, 77(10):3238-3242. 135. Omer-Mizrahi M, Margel S: Synthesis and characterization of magnetic and non- magnetic core-shell polyepoxide micrometer-sized particles of narrow size distribution. J Colloid Interface Sci 2009, 329(2):228-234. 136. Tang DP, Yu YL, Niessner R, Miro M, Knopp D: Magnetic bead-based fluorescence immunoassay for aflatoxin B-1 in food using biofunctionalized rhodamine B-doped silica nanoparticles. Analyst 2010, 135(10):2661-2667. 137. Voigt A, Buske N, Sukhorukov GB, Antipov AA, Leporatti S, Lichtenfeld H, Baumler H, Donath E, Mohwald H: Novel polyelectrolyte multilayer micro- and nanocapsules as magnetic carriers. J Magn Magn Mater 2001, 225(1-2):59-66. 138. Zhang M, Wang ZF, Zhang H, Dai SJ, Qiu GM, Hiroshi O: Study on magnetic responsibility of rare earth ferrite/polyacrylamide magnetic microsphere. Journal of Rare Earths 2005, 23(3):285-287.

36 General Introduction

139. Bourgeat-Lami E, Lang J: Encapsulation of inorganic particles by dispersion polymerization in polar media - 1. Silica nanoparticles encapsulated by polystyrene. J Colloid Interface Sci 1998, 197(2):293-308. 140. Bourgeat-Lami E, Lang J: Encapsulation of inorganic particles by dispersion polymerization in polar media 2. Effect of silica size and concentration on the morphology of silica-polystyrene composite particles. J Colloid Interface Sci 1999, 210(2):281-289. 141. Hong J, Lee J, Rhym YM, Kim DH, Shim SE: Polyelectrolyte-assisted synthesis of polystyrene microspheres by dispersion polymerization and the subsequent formation of silica shell. J Colloid Interface Sci 2010, 344(2):410-416. 142. Amann RI, Binder BJ, Olson RJ, Chisholm SW, Devereux R, Stahl DA: Combination of 16s ribosomal-rna-targeted oligonucleotide probes with flow-cytometry for analyzing mixed microbial-populations. Applied and Environmental Microbiology 1990, 56(6):1919-1925. 143. Clark GM, Dressler LG, Owens MA, Pounds G, Oldaker T, McGuire WL: Prediction of relapse or survival in patients with node-negative breast-cancer by dna flow- cytometry. New England Journal of Medicine 1989, 320(10):627-633. 144. Darzynkiewicz Z, Bruno S, Delbino G, Gorczyca W, Hotz MA, Lassota P, Traganos F: Features of apoptotic cells measured by flow-cytometry. Cytometry 1992, 13(8):795- 808. 145. Takahira J, Cousin A, Nelson MN, Cowling WA: Improvement in efficiency of microspore culture to produce doubled haploid canola (Brassica napus L.) by flow cytometry. Plant Cell Tissue and Organ Culture 2011, 104(1):51-59. 146. Hedley DW, Friedlander ML, Taylor IW, Rugg CA, Musgrove EA: Method for analysis of cellular dna content of paraffin-embedded pathological material using flow- cytometry. Journal of Histochemistry & Cytochemistry 1983, 31(11):1333-1335. 147. Jung T, Schauer U, Heusser C, Neumann C, Rieger C: Detection of intracellular cytokines by flow-cytometry. Journal of Immunological Methods 1993, 159(1-2):197- 207. 148. Zhao M, Kanegane H, Kobayashi C, Nakazawa Y, Ishii E, Kasai M, Terui K, Gocho Y, Imai K, Kiyasu J et al: Early and Rapid Detection of X-Linked Lymphoproliferative Syndrome with SH2D1A Mutations by Flow Cytometry. Cytometry Part B-Clinical Cytometry 2011, 80B(1):8-13. 149. Lyons AB, Parish CR: Determination of lymphocyte division by flow-cytometry. Journal of Immunological Methods 1994, 171(1):131-137. 150. Nicoletti I, Migliorati G, Pagliacci MC, Grignani F, Riccardi C: A rapid and simple method for measuring thymocyte apoptosis by propidium iodide staining and flow- cytometry. Journal of Immunological Methods 1991, 139(2):271-279. 151. Shattil SJ, Cunningham M, Hoxie JA: Detection of activated platelets in whole-blood using activation-dependent monoclonal-antibodies and flow-cytometry. Blood 1987, 70(1):307-315. 152. Siena S, Bregni M, Brando B, Belli N, Ravagnani F, Gandola L, Stern AC, Lansdorp PM, Bonadonna G, Gianni AM: Flow-cytometry for clinical estimation of circulating hematopoietic progenitors for autologous transplantation in cancer-patients. Blood 1991, 77(2):400-409.

37 General Introduction

153. Vo-Dinh T, Cullum B: Biosensors and biochips: Advances in biological and medical diagnostics. Fresenius' Journal of Analytical Chemistry 2000, 366(6-7):540-551. 154. Diamandis E, Christopoulos T: Immunoassay: San Diego: Academic Press; 1996. 155. Zhao Y, Zhao X, Sun C, Li J, Zhu R, Gu Z: Encoded silica colloidal crystal beads as supports for potential multiplex immunoassay. Anal Chem 2008, 80(5):1598-1605. 156. Henderson P: Rare earth element geochemistry. In: developments in geochemistry. New York: Elsevier; 1984. 157. Schijf J: Aqueous geochemistry of the rare earth elements in marine anoxic basins. PhD. Rijksuniversiteit te Utrecht; 1992.

38 Instrumental Methods and Experimental Details

2 Instrumental Methods and Experimental Details

In this chapter, instrumental methods, materials, experimental procedures and characterization methods employed in this thesis are described. In section 2.1, particular attention is given to a key analytical methods of this work, inductively coupled plasma mass spectroscopy (ICP-MS) and mass cytometry; there protocols, data analysis and interpretation are described in detail. In section 2.2, details about all materials utilized in this work are given. Synthetic procedures for the preparation of polymer microspheres are described in section 2.3 with a focus on dispersion polymerization techniques. Finally, in section 2.4, characterization methods, other than ICP-MS and mass cytometry, employed in this work are explained. Unless otherwise stated, all of the synthesis, experiments and characterizations presented in this thesis were done by me. Bioconjugation protocols and all experiments involving cell cultures were done by Dr. Olga Ornatsky and Dr. Dirk Weinrich. Seeded emulsion polymerization was done by Dr Stuart Thickett using the microspheres I synthesized by dispersion polymerization.

2.1 Instrumental Methods

2.1.1 Inductively Coupled Plasma Mass Spectrometer

The lanthanide ion content of solutions was measured by inductively coupled plasma mass spectrometer (ICP-MS). ICP-MS was also used to determine the lanthanide ion content of the digested microparticles. Most importantly, ICP-MS is an essential part of the mass cytometry instrument; it constitutes the analyzer and the detector in that instrument. Because of the significant importance of the ICP-MS as an analytical tool in my projects, I will start with a brief introduction to the technique before going into the details of sample preparation and measurement protocols.

2.1.1.1 Introduction

ICP-MS is a type of mass spectrometry that is highly sensitive and capable of the determination of a range of metals and several non-metals at concentrations below one part in 1012. It is based on an inductively coupled plasma as a method of producing ions (ionization) and

39 Instrumental Methods and Experimental Details

a mass spectrometer as a method of separating and detecting the ions. ICP-MS is also capable of monitoring isotopic species and ratios for the elements of choice.

2.1.1.2 Plasma Source

An inductively coupled plasma is a plasma that contains a sufficient concentration of ions and electrons to make the gas electrically conductive. The plasmas used in spectrochemical analysis are essentially electrically neutral, with each positive charge on an ion balanced by a free electron. In these plasmas the positive ions are almost all singly-charged, so there are nearly equal amounts of ions and electrons in each unit volume of the plasma. ICP for spectrometry is sustained in a torch that consists of three concentric tubes, usually made of quartz. The end of this torch is placed inside an induction coil supplied with a radio-frequency electric current. A flow of argon gas (usually 14 to 18 liters per minute) is introduced between the two outermost tubes of the torch and an electrical spark is applied for a short time to introduce free electrons into the gas stream. These electrons interact with the radio-frequency magnetic field of the induction coil and are accelerated first in one direction, then the other, as the field changes at high frequency (usually 27.12 MHz). The accelerated electrons collide with argon atoms, and sometimes a collision causes an argon atom to part with one of its electrons. The released electron is in turn accelerated by the rapidly-changing magnetic field. The process continues until the rate of release of new electrons in collisions is balanced by the rate of recombination of electrons with argon ions (atoms that have lost an electron). This produces a ‘fireball’ that consists mostly of argon atoms with a rather small fraction of free electrons and argon ions. The temperature of the plasma is very high, of the order of 10,000 K. The ICP can be retained in the quartz torch because the flow of gas between the two outermost tubes keeps the plasma away from the walls of the torch. A second flow of argon (around 1.0 L.min-1) is usually introduced between the central tube and the intermediate tube to keep the plasma away from the end of the central tube. A third flow (again usually around 1.0 L.min-1) of gas is introduced into the central tube of the torch. This gas flow passes through the centre of the plasma, where it forms a channel that is cooler than the surrounding plasma but still much hotter than a chemical flame. Samples to be analyzed are introduced into this central channel, usually as a mist of liquid formed by passing the liquid sample into a nebulizer. As a droplet of nebulized sample enters the central channel of the ICP, it evaporates and any solids

40 Instrumental Methods and Experimental Details

that were dissolved in the liquid vaporize and then break down into atoms. At the temperatures prevailing in the plasma, a significant proportion of the atoms of many chemical elements are ionized, each atom losing its most loosely-bound electron to form a singly charged ion.

2.1.1.3 Mass Analyzers

For coupling to mass spectrometry, the ions from the plasma are extracted through a series of cones into a mass spectrometer, usually a quadrupole. The ions are separated on the basis of their mass-to-charge ratio and a detector receives an ion signal proportional to the concentration. The concentration of a sample can be determined through calibration with certified reference materials such as single or multi-element reference standards. Other mass analyzers coupled to ICP systems like time of flight systems (TOF) can also be used. To determine the lanthanide ion content in solution, I used the homogenous “traditional” ICP-MS with a quadrupole mass analyzer. On the other hand, to determine the lanthanide ion content of individual microparticles, mass cytmometry was employed. Mass cytometry utilizes TOF mass analyzer.

2.1.1.4 Sample Preparation for ICP-MS

Sample analyses by traditional ICP-MS employed an ELAN DRCPlus instrument (Perkin- Elmer SCIEX). Typical operating conditions of the instrument are based on a stable Ar plasma optimized to provide less than a 3% CeO+/Ce+ ratio in a 1 ppb standard multielement solution diluted in 10% HCl. This requirement was achieved by applying 1400 W forward plasma power, 17 L/min Ar plasma gas flow, 1.2 L/min auxiliary Ar flow, and 0.95 L/min nebulizer (Burgener Micromist or MicroFlow PFA-ST concentric nebulizers) Ar flow. Under these operating conditions, the typical sensitivity is 4 × 104 cps for 1 ppb Ir standard solution in 10% HCl. The detection limits for lanthanide elements were less than 1 ppt. The sample uptake rate was adjusted depending on the particular experiment and sample size, typically 100 μL/min. Experiments were performed using an autosampler (Perkin-Elmer AS 93) modified for operation with Eppendorf 1.5 mL tubes. Sample sizes varied from 150 to 300 μL. Standards were prepared from 1000 μg/mL PE Pure Single-Element Standard solutions (Perkin-Elmer, Shelton, CT) by sequential dilution with high-purity deionized water (DIW) produced using an Elix/Gradient (Millipore, Bedford, MA) water purification system.

41 Instrumental Methods and Experimental Details

For supernatant analysis, an aliquot (100 µL) from as-prepared particle dispersion was mixed with deionized water (900 µL) and spun down by three cycles of centrifugation (5000 rpm, 30 min) immediately to separate the particles from the supernatant. Then, the supernatant

was collected and diluted 10 times with HNO3 aqueous solution (3.0 %) and then analyzed by ICP-MS.

2.1.2 Mass Cytometry

The lanthanide ion content of polymer microparticles was measured by mass cytometry: flow cytometry-time-of-flight inductively coupled plasma mass spectrometry (ICP-MS). Because of the extreme importance of the measurements done by mass cytometry in this thesis, I will present detailed design considerations and the analytical characteristics of this novel mass cytometery technique. Afterward, I will explain sample preparation and handling. The description and the details of the prototype CyTOF mass cytometry used in this research are all adapted from references [1, 2]

2.1.2.1 Introduction

Mass cytometry was developed as a new technique for the detection of proteins and other biomolecules attached to individual microparticles or in individual cells. The technique is based on attaching elemental tags to affinity products (antibodies or aptamers), in place of fluorescent labels that are used in traditional flow cytometry. The elemental tags are specially designed polymers or polymer particles that are loaded with metals. Mass cytometry is designed to benefit from the high resolution, sensitivity, and speed of analysis of inductively coupled plasma time- of-flight mass spectrometry. Because of the wide availability of stable isotopes that can be used in the tags, many proteins and gene transcripts can potentially be simultaneously detected in individual cells. The detection is done through the quantification of stable isotope tags bound to target biomarkers. Baranov et al. [3-5] was the first to suggest and to demonstrate the method of simultaneous detection by ICP-MS of multiple proteins in homogeneous biological samples using element- tagged antibodies. This idea was further explored and developed by other groups [6-11]. However, this methodology cannot be directly applied to multitarget individual cell analysis. That is because of the settling time the quadrupole mass analyzers have. The settling time of a

42 Instrumental Methods and Experimental Details

mass analyzer is the time required for stabilization of the mass filter between individual isotope measurements which is ca. 50 − 200 μs for the scanning analyzers like the quadrupole mass analyzers. This settling time is longer than the duration of the ion cloud produced in ICP from an individual microparticle ( 100 μs fwhm [12]). Thus, with the quadrupole mass analyzers, it is almost impossible to perform the measurement of two or more isotopes during a transient event of such short duration. Obviously, a simultaneous mass analyzer, such as a time-of-flight analyzer with an array detector was the candidate of choice.

2.1.2.2 Instrument Design

The schematic of the new instrument (built by Dmitry Bandura, Vladimir Baranov and Scott Tanner) is shown in Scheme 2-1. In this section, details of the mass cytometry instrumental design are quoted from Bandura et. al. [1]: “Microparticles are introduced in the form of a liquid suspension by the syringe pump (Pump 22, Harvard Apparatus Canada, Saint-Laurent, Quebec, Canada) and aspirated by a concentric nebulizer (TQ-30-A1, Meinhard Glass Products, Golden, CO). The nebulizer is connected to a custom-made heated spray chamber, to which a makeup Ar gas flow (typically at 5 L/min) is supplied via a mass flow controller. This high flow of heated makeup gas is needed to partially dry the larger droplets that contain particles and to provide adequate confinement of the high inertia larger droplets in the gas stream. A cyclonic spray chamber (PN 300-19MS, Precision Glassblowing, Centennial, CO), which allows only a small-diameter fraction of the aerosol through, is used with either a peristaltic pump (Minipuls 3, Gilson, Inc., Middleton, WI) or the syringe pump. For this prototype instrument, the plasma generator is adapted from the ELAN 6000 ICPMS (Perkin-Elmer- SCIEX, Concord, ON, Canada) and comprises a free-running (nominal 40 MHz) radiofrequency generator and an rf-balanced load coil arrangement.

43 Instrumental Methods and Experimental Details

Scheme 2-1: Schematics of the prototype CyTOF mass cytometer

The torch assembly is also from the ELAN and comprises a demountable torch and a 2 mm i.d. quartz injector and to provide adequate confinement of the high inertia larger droplets in the gas stream. A cyclonic spray chamber (PN 300-19MS, Precision Glassblowing, Centennial, CO), which allows only a small-diameter fraction of the aerosol through, is used with either a peristaltic pump (Minipuls 3, Gilson, Inc., Middleton, WI) or the syringe pump. For this prototype instrument, the plasma generator is adapted from the ELAN 6000 ICPMS (Perkin-Elmer-SCIEX, Concord, ON, Canada) and comprises a free- running (nominal 40 MHz) radiofrequency generator and an rf-balanced load coil

44 Instrumental Methods and Experimental Details

arrangement. The torch assembly is also from the ELAN and comprises a demountable torch and a 2 mm i.d. quartz injector. To transfer the plasma from the high temperature and ambient pressure of the ICP torch to the room temperature and the 0.3 μTorr-vacuum of TOF sector, the plasma is sampled through an interface which has 3 apertures: sampler (1.1 mm orifice diameter), skimmer (1 mm diameter), and reducer (1.2 mm diameter). The analyzer is operated at 76.8 kHz spectra generation frequency. A fast TOF ion detector (model 14882, ETP Electron Multipliers, SGE International Pty. Ltd., Ringwood, Victoria, Australia) is used for ion detection. The output signal of the detector is amplified by a preamplifier (FTA420, ORTEC Products Group, Oak Ridge, TN) and digitized by the analog-to-digital conversion (ADC) based 8-bit 1 GHz signal digitizer (PDA1000, Signatec, Inc., Newport Beach, CA). A trigger delay (9000 ns) and the recording segment length (3072 ns) are set to allow digitization of the segment of the signal that corresponds to m/z = 125−215, with 1 ns sampling resolution. The figures of merit of the instrument are measured under standard ICP operating conditions (< 3 % oxide ratio). At mass resolution (full width at half maximum) for m/z = 159 M/ΔM > 900, sensitivity with a standard sample aspiration is 1.4 × 108 ion counts per second per mgL-1 of Tb typically, and abundance sensitivity 6×10-4 – 1.4×10-3 (trailing and leading masses, respectively). The mass range (variable, but fixed at m/z = 125 – 215 for this work) and the abundance sensitivity are sufficient for elemental encoding with up to 60 distinct isotopes.” More details about this instrument can be found in references [1, 2, 13]. Data reported in this thesis refer to the analysis of 10,000 to 20,000 individual particles unless otherwise stated.

2.1.2.2 Sample Preparation for Mass Cytometry

The microparticles were washed by several cycles of centrifugation (3000 rpm for 20 min) and resuspension in DIW, and the resultant slurry (ca. 106 microspheres/mL) was nebulized into the mass cytometer sample introduction system, which in turn delivered microspheres individually but stochastically into the inductively coupled plasma torch. The high temperature of the plasma (> 6000 K) is sufficient to atomize and then ionize the microspheres and the Ln ions embedded in them. The ion stream was then introduced into the time-of-flight mass

45 Instrumental Methods and Experimental Details

analyzer. The transient signals corresponding to each microsphere ionization event were recorded by the detector and stored.

2.2 Materials

2.2.1 Solvents

Absolute ethanol (ACS grade), 95% ethanol (commercial grade), tetrahydrofuran (THF, ACS grade), dichloromethane (DCM, ACS grade) were used without further purification. Water was purified through a MilliQ purification system with minimum resistivity of 10 MΩ ⋅cm. All solutions for ICP-MS and mass cytometry bioassays were prepared in deionized water (Elix/Gradient water purification system, Millipore) with minimum resistivity of 18 MΩ⋅cm.

2.2.2 Reagents for Particles’ Synthesis and Characterization

Styrene (Aldrich), methyl methaacrylate (MMA, Aldrich), acrylic acid (AA, Aldrich), methacrylic acid (MAA, Aldrich), acetoacetylethyl methacrylate (AAEM, Aldrich), ethylene glycol dimethacrylate PVP (EGDMA, Aldrich), divinylbenzene (DVB, Aldrich), polyvinylpyrrolidone () (Aldrich, PVP55, M = 55,000, PVP360, M = 360,000 and PVP40, M = 40,000), 2,2'-azobis(2-methylbutyronitrile) (AMBN, Dupont USA), potassium persulfate (KPS, Aldrich), Triton-X305 (TX305, 70% solution in water, Aldrich), sodium dodecyl sulfate (SDS, Aldrich), lauryl methacrylate (Aldrich), 4,4,4-Trifluoro-1-(2-naphthyl-1,3-butanedione) (TNB,

Aldrich), and sodium bicarbonate (NaHCO3, Aldrich), were used as received.

Lanthanum(III) chloride hydrate (LaCl3, Fluka), praseodymium(III) chloride hexahydrate

(PrCl3, Aldrich), neodymium (III) chloride hexahydrate (NdCl3, Aldrich), samarium(III)

chloride hexahydrate (SmCl3, Aldrich), europium(III) chloride hexahydrate (EuCl3, Aldrich),

gadolinium(III) chloride hexahydrate (GdCl3, Aldrich), terbium(III) chloride hexahydrate

(TbCl3, Aldrich), dysprosium(III) chloride hexahydrate (DyCl3, Aldrich), holmium(III) chloride

hexahydrate (HoCl3, Aldrich), and thulium (III) chloride hexahydrate (TmCl3, Aldrich) were used as lanthanide source for most of the synthesis, unless otherwise stated. All salts used are with an assay of 99.999% trace metals basis.

High purity HCl and HNO3 for ICP-MS analysis were purchased from Seastar Chemical Inc. Phosphate-buffered saline with calcium and magnesium (PBS; 150 mM NaCl, 1.2 mM Ca2+,

46 Instrumental Methods and Experimental Details

0.8 mM Mg2+, 20 mM sodium phosphate, pH 7.4), 37% formaldehyde (Sigma), Ir3+ (iridium) diluted from stock 1000 μg mL-1 solutions to 1 ng mL-1 in 3.3 – 3.6 % HCl.

2.2.2 Reagents for Bioconjugation

Primary mouse monoclonal antibodies (mAb) anti-CD34 mAb, 0.25 mg/mL (BD Biosciences), were used at 1:100 dilution; Antibodies were labeled with the prototype MAXPAR reagents (DVS Sciences Inc., Richmond Hill, Ontario, Canada; www.DVSsciences.com), based on metal-labeled polymer tags described in detail by Lou et al [14]. Avidin labeled with fluorescein isothiocyanate (FITC-Avidin), 1-ethyl-3-(3- dimethylaminopropyl) carbodiimide hydrochloride (EDC) and N-hydroxysuccinimide (NHS) were purchased from Fluka, USA and used as received. Human monocyte cell line KG1a, a model human acute myelogenous leukemia cell line, with high CD34 antigen expression (approximately 100 000 copies per cell), as well as U937 human leukemic monocyte lymphoma cell line (that does not express CD34), were obtained from the American Type Culture Collection (Manassas, VA). Cells were propagated in MEM synthetic media, supplemented with 10% FBS (HyClone) and 2 mM L-glutamine (Invitrogen), in a humidified incubator at 37 °C and 5% CO2. Cells were split every 3-4 days, and viability was checked with trypan blue (>90% viable). The Ir-containing metallointercalator (pentamethylcyclopentadienyl)-Ir(III)- dipyridophenazine, Ir-intercalator, is the sample described in reference [15]. This Ir-intercalator is indefinitely stable in the solid state and in aqueous solution. This Ir-intercalator complex was 6 -1 found to have an equilibrium constant Kb of 2.6 x 10 M for binding to double-stranded DNA. It was used here to stain the cell DNA.

47 Instrumental Methods and Experimental Details

2.3 Synthetic Procedures

In this section the methods of synthesizing lanthanide-containing polymer particles are explained in detail. Throughout this thesis, microspheres of 1.5 µm in diameter and larger were synthesized by two-stage dispersion polymerization (2-DisP) or three stage dispersion polymerization (3-DisP). Narrow size distribution is an important common requirement in the both methodologies; we imagine narrow-size-distributed particles will have a similarly narrow distribution of lanthanide ions per particle

2.3.1 Microspheres’ Synthesis by Dispersion Polymerization

For mass cytometry-based bioassays, particles should have certain size and surface characteristics that will be discussed in detail in Chapter 3 and 4. In this section, I explain the experimental details of preparing functionalized polymer microspheres. I used dispersion polymerization because of the advantages it has over other techniques of synthesizing polymer microspheres due to its superiority in the production of very monodisperse particles and the facility to be implemented on a large scale.[16] However, dispersion polymerization is extremely sensitive to small changes in reaction conditions. For example, stirring blade depth inside the reaction flask has an unexpectedly strong effect on the particle size and particle size distribution. Figure 2-1 A shows the set up when the blade was placed deep in the reaction flask so that it touches the bottom or the wall of the flask occasionally during the course of the reaction. Figure 2-1 C shows the SEM micrograph for PS particles prepared using the set up illustrated in Figure 2-1 A. In this case, polydisperese particles were formed with aggregates comprising about 20 to 30 wt% of the total products. On the other hand, using the same recipe, colloidally-stable monodisperse PS particle (Figure 2-1 D) were obtained when I moved the stirring blade far away from the bottom and fixed at a height at which the blade was just immersed in the reaction mixture (Figure 2-1 B).

48 Instrumental Methods and Experimental Details

Figure 2-1: Effect of the stirring blade depth on the particle size distribution C D In my work, I used “two-stage” dispersion polymerization (2-DisP) method which was developed by Song et al.[17] I also examined an advanced version, three-stage dispersion (3- DisP) polymerization with ethanol or ethanol/water as the solvent. In 2-DisP, as developed by Song et al., the addition of the comonomer was deferred until the nucleation stage of the reaction was complete (e.g. after 5% styrene conversion). To my reactions, to incorporate Ln metals in the polymer particles, I deferred the addition AA and the lanthanide chloride(s) to a second stage, one hour after the initiation of the reaction. In few cases (will be mentioned in place) I used acetoacetylethyl methacrylate (AAEM) instead of AA. The reaction set-up for my dispersion polymerization is shown in Figure 2-B. A three-neck round bottom glass flask, equipped with a mechanical stirrer, a condenser, and a nitrogen inlet, were used for the reaction. An oil bath, placed on a hot plate stirrer equipped with a temperature- control regulator, was used for heating the reactor.

49 Instrumental Methods and Experimental Details

A sample recipe for the two-stage dispersion polymerization (2-DisP) of styrene with comonomer in ethanol is listed in Table 2-1. The following procedure was used: All of the stabilizer (PVP55), the co-stabilizer (TX305), initiator (AMBN), and the styrene monomer and half of the ethanol were added to the three-neck round bottom glass flask. After a homogeneous solution formed at room temperature, the solution was deoxygenated by purging with ethanol- saturated nitrogen gas at room temperature for 30 min. The flask was then placed in a 70 °C oil bath and stirred mechanically at 100 rpm with an overhead stirring bar connected to a (PTFE) half-moon stirring paddle. The comonomer (AA or AAEM) and the lanthanide salt(s) (LnCl3) were dissolved in the remaining ethanol at 70 °C under nitrogen. After the polymerization reaction had run for 1 h, the hot solution containing LnCl3 plus the comonomer solution in the remaining ethanol was added to the reaction flask. The reaction was continued for 24 h. Final conversion was approximately 95-99% as determined gravimetrically. The amount of lanthanide salts added for each batch was determined based on the targeted number of Ln ions per microsphere. The calculation for the number of Ln ions per microsphere is done using the following equation:

total number of Ln ions added to the reaction number of atoms per microsphere = total number of particles produced by the reaction (2-1) where the total number of particles produced by the reaction is calculated based on the following equation:

total mass produced total number of particles produced by the reaction = mass of a single microsphere (2-2) where the mass of the microsphere is calculated based on the measured volume by SEM and PS density of 1.05 g/cm3. In 3-stage DisP, another hot mixture of additional comonomer and a crosslinking agent, EGDMA, in ethanol (10 ml) was added to the reaction after the second stage addition. The time of adding the 3rd stage was varied from 1.0 – 18.0 hr after the 2nd stage. An example recipe for the three-stage dispersion polymerization (3-DisP) is given in Table 2-2.

50 Instrumental Methods and Experimental Details

Table 2-1. Recipe for the synthesis of microsphere sample AA0069 as an example for 2-stage dispersion polymerization (2-DisP) of styrene with PVP55 as a dispersant in ethanol.

Two-stage reactions Materials (grams added) 1st stage 2nd stage Styrene 6.25 --a PVP55 a 1.0 -- TX305 b 0.35 -- AMBN c 0.25 -- Ethanol 18.75 18.75 Comonomerd -- 0.125e f EuCl3 -- 0.063 a polyvinylpyrrolidone, Mw ≈ 55 kDa. b Triton X-305 c azomethylbutyronitrile d AA or AAEM e In the synthesis of some samples 0.25 g of comonomer (4 wt%/sty) or 0.50 g of comonomer (8 wt%/sty) was used. f EuCl3 is EuCl3.6H20. For other microsphere samples, other combinations of LnCl3.6H20 were used (where Ln = La, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho and/or Tm).

51 Instrumental Methods and Experimental Details

Table 2-2. Recipe for the synthesis of microsphere sample AA0105 as an example for 3-stage dispersion polymerization (2-DisP) of styrene with PVP55 as a dispersant in ethanol. Three-stage reactions Materials (grams added) 1st stage 2nd stage 3rd stage Styrene 6.25 -- -- PVP55 1.0 -- -- TX305 0.35 -- -- AMBN 0.25 -- -- Ethanol 18.75 10.0 10.0 Comonomer a -- 0.125 0.25 b LaCl3 -- 0.006 -- b TmCl3 -- 0.006 -- EGDMA -- -- 0.063 a AA or AAEM b LaCl3 and TmCl3 are LaCl3.6H20 and LaCl3.6H20, respectively. For other microsphere

samples, other combinations of LnCl3.6H20 were used (where Ln = La, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho and/or Tm).

2.3.2 PVP pyrrolidone ring opening (PVP activation).

In order to activate the PVP chains, I performed basic hydrolysis for PVP55 prior to using it in the dispersion polymerization. The method I used is adopted from the method of von Specht et al.[18] in which 1.0 g PVP55 was dissolved in 50 ml 0.1 N NaOH and the solution was heated for 36 h in an autoclave at 140 °C. In order to methylate the γ-amino butyric of the opened pyrrolidone ring and to prevent its closing, 2.86 ml of 35% formaldehyde solution was added, the pH was adjusted to 9 and the solution cooled down to 0 °C; 1.5 sodium tetrahydroborate was added, and the solution stirred for 45 min. The amount of ring opening obtained by von Specht et al. was 15 %. This extent of ring opening was determined based on a method described by Frank et al.[19] in which an excess of 1.5 ml 1N HCl is added to a 10 ml portion, and the amino acid formed is titrated in ethanol with 39.4 mM ethanolic KOH.

52 Instrumental Methods and Experimental Details

2.3.3 Dispersion polymerization using modified PVP.

The recipe for the two-stage dispersion polymerization (2-DisP) of styrene with AA in ethanol is listed in Table 2-3. The procedure used is similar to the one described in ref. 29, in which the PS-co-PAA beads were obtained after 24 h reaction under ethanol-saturated nitrogen. Instead of using untreated PVP as a dispersant, we doped the PVP with modified PVP at two different levels, 1.6 and 3.2 wt.% / styrene. In these reactions, AA and the lanthanide chloride salt were added in the second stage, one hour after initiating the polymerization.

Table 2-3. Recipe for AA139 synthesized by dispersion polymerization of styrene with activated PVP.

st nd a Materials (g added) 1 stage 2 stage

Styrene 6.25 -- PVP 0.9 b --

Activated PVP c 0.1 d --

TX305 0.35 --

AMBN 0.25 --

Ethanol 18.75 18.75

Acrylic acid -- 0.125 e Lanthanide salts -- 0.0156 a 1.0 hr after the first stage b 0.8 g were used in AA139 sample c prepared by the basic hydrolysis of PVP55. d 0.2 g were used in AA139 sample e LaCl3.6H20 = EuCl3.6H20 = TbCl3.6H20 = HoCl3.6H20 = TmCl3.6H20 = 3.1 mg

53 Instrumental Methods and Experimental Details

2.3.4 Seeded Emulsion Polymerization with Methacrylic Acid (MAA).

The method I presented in this thesis is analogous to that of Alam et al [20] and was performed by Dr Stuart Thickett. A sample of PS-co-PAA particles made by 2-DisP[13]

(denoted D1, 2.25 g, solids content 8.71 % w/w, davg = 2.1 µm, CVd = 1.8 %) was redispersed into water by the method described above. These particles were used as a seed for emulsion polymerization with MAA with a 2:1 core:shell ratio by mass. 0.1 g of MAA was added to the particles, along with additional water (8.33 g). In the work of Alam et al., a mixture of hydrophilic monomers were used however in this case only MAA was chosen – no additional hydrophobic monomer was added. Polymerization was initiated by KPS (5 mg, corresponding to an initiator concentration of 0.7 mM) and was allowed to proceed for 24 hours under magnetic stirring.

2.3.5 Seeded Emulsion Polymerization with Glycidyl Methacrylate (GMA).

The method presented here represents conditions chosen from the work of Omer-Mizrahi et al [21] and was performed by Dr Stuart Thickett to create poly(glycidyl methacrylate) (PGMA) shells on the surface of PS particles made by dispersion polymerization. A variety of analyses were performed in this work to coat a range of different Ln-labeled particles. A representative method is as follows: A sample of D1 was taken (equivalent to 0.15 g of solid polymer based on the solids content of the dispersion) and redispersed into pure water as described above. The aqueous dispersion was then placed in a small round bottom flask, and GMA (0.6 g) was added to the system. The weight ratio of GMA (shell) to PS (core) was close to 4:1. The latex volume was made up to a total 10 mL with Milli-Q water, and 0.1 g of SDS (approximately 36 mM, CMC = 8.2 mM in water at 25°C) was then added. The system was agitated and allowed to stir magnetically for 30 minutes to allow for swelling and establishment of monomer-swollen micelles. The system was then degassed using high purity N2 prior to the sample being raised to reaction temperature (343 K) and the injection of initiator (potassium persulfate, 0.012 g) in a small volume of water. Polymerization took place for 6 hours. Unless otherwise mentioned, certain ratios were kept constant in this work: the volume fraction of GMA in the system was 6 %; the ratio of SDS to

water was 1 % w/v; and the KPS to GMA ratio was 2 % w/vmonomer. Conversion was measured by gravimetry and typically close to 100 %. A 4 mL (2 × 2 mL) sample of the latex was cleaned

54 Instrumental Methods and Experimental Details

by three centrifugation and redispersion cycles into water. Another sample of PS particles (denoted D1’) was used as seeds for growing PGMA shells (sample D5(GMA-4)). D1’ sample

was synthesized also by 2-DisP and has an average diameter (davg) of 1.6 µm and a size

distribution (CVd) of 2.9 %.[13]

2.4 Characterization Methods

2.4.1 Scanning Electron Microscopy (SEM)

For particles with diameters on the order of 1.0 to 3.0 µm, particle diameter and diameter distributions were measured by Hitachi S-5200 FE-scanning electron microscope (SEM). For imaging, the sample was prepared by placing a drop of diluted suspension on a Formvar/carbon coated-300 mesh copper grid. There was no need to purify the particles to obtain clear images. SEM image analysis was done by ImageJ software. A particle-size histogram was constructed from measurements of at least 200 individual particles for each sample.

2.4.2 Dynamic Light Scattering (DLS)

The diameters of latex submicron particles were measured on Brookhaven Instruments model BI-90 Particle Sizer at a fixed scattering angle of 90° or with a wide angle light scattering photometer from ALV. The light source was a JDS Uniphase He-Ne laser (λ0) 632.8 nm, 35 mW) emitting vertically polarized light. The scattered light was detected by a Dual ALV-High Q.E. APD avalanche photodiode module. This detector was interfaced to the ALV-5000/EPP multiple time delay digital correlation function in real time. All measurements were carried out at a scattering angle of 90°. For sample preparation, an aliquot of latex samples was diluted at a concentration of ca. 2 wt % with deionized water and then submitted to measurements. Measurments using the ALV system were done by Dr. Gerald Guerin and Jishu Qian.

2.4.4 Gel Permeation Chromatography

The molar mass distribution was measured by gel permeation chromatography (GPC) system equipped with a Viscotek VE 3219 UV/vis detector (set to 310 nm), VE 3580 RI detector, a Polymer Laboratories gel 5 μm Mixed-D (300*7.5 mm) column, and a gel 5 μm guard column (both at room temperature). The flow rate was maintained at 0.6 mL/min using a Waters

55 Instrumental Methods and Experimental Details

515 HPLC Pump. Tetrahydrofuran (THF) was used as the eluent, and the system was calibrated with polystyrene standards. The polymer samples for GPC analysis were prepared as follows: The particles were centrifuged at 3000 rpm for 20 min. The precipitated particles were redispersed into ethanol and centrifuged again. This process was repeated three times. After the particles were dried, they were dissolved in THF, and the solution was passed through a 0.45 µm filter before injecting it into the GPC column

2.4.5 Gravimetrical measurements

The solids contents (S.C.) of the particle dispersions were determined by gravimetry. About

0.1 g of particle dispersion (WL) was tared into 20-ml glass vial, and then placed in a pre-heated oven at 120 ◦C for 60 min. After the samples cooled at room temperature, the remaining solids

(WS) were measured. The solids content (%) was calculated from equation 2-3. The monomer conversion, the weight fraction of monomer polymerized, was calculated by taking the ratio of solids content from equation 2-3 to that calculated from the recipe.

S.C. (%) = WS / WL x 100 (2-3)

2.4.3 Titration of Acid Groups

For surface acid group titration, particles were first washed by three cycles of centrifugation and resuspension in water. Titrations were monitored with a conductivity meter (EcoMet ConductivityITemp C65) and pH meter (EcoMet, Rose Scintific LTD, Edmonton, AB). All data was acquired by back (acid added to base) titration. Freshly prepared 0.1 M NaOH (200 μl) was added to an aliquot (0.500 g, 5 wt % solids content) of the washed particles 50 ml water and titrated with freshly prepared 0.025 M HCl. Quantitative information was acquired from the titration plots using the standard extrapolation/intersection method to determine the titration endpoints. Figure 2-3 illustrates an example of the Potentiometeric and conductometric titrations for microsphere sample AA087. The highlighted vertical gray band represents the concentration of HCl (in µmol) equivalent to the concentration of –COOH groups on or near the surface of the microspheres.

56 Instrumental Methods and Experimental Details

12 90 85 10 80

75 8 70

6 65

pH 60 4 55

50 2

45 Conductivity (μs/cm)

0 40 0 5 10 15 20 HCl (µ mol)

Figure 2-3. Potentiometeric and conductometric titrations of AA087 microspheres synthesized by 2-DisP in presence of AA (2.0 wt%/styrene). The highlighted area corresponds to 1.6 µmol of HCl

2.4.6 Bioconjugation

Dr. Olga Ornatsky employed standard conditions to conjugate mouse IgG to the bead surface using water-soluble carbodiimide chemistry.[22] The lanthanide encoded microparticles were first washed twice in cold Polylink Coupling buffer [50 mM 2-(N- morpholino)ethanesulfonic acid (MES), pH 5.2] by low speed centrifugation (1000 x g, 10 min). Freshly prepared heterobifunctional cross-linker EDC solution (0.2 g/ml 1-Ethyl-3-(3- dimethylaminopropyl)carbodiimide hydrochloride in MES) was then mixed with the bead suspension, followed by gentle vortexing. EDC-activated microparticles were then split into two tubes labeled “IgG” and “BSA”. An aliquot of mouse IgG/PBS (2 μL, 18 μg) was added to the tube labeled “IgG”, and an aliquot of 0.5% BSA/PBS (4 μL, 20 μg) was added to the tube labeled “BSA”. Both sets of samples were mixed gently and left to react on a shaker for 2 h at room temperature and then washed. Microparticles were then blocked in 0.5% BSA/PBS buffer for 1 h prior to addition of 100 μg/ml anti-mouse IgG-X4-Pr as the test-antigen reporter to be captured by the mouse IgG coated microspheres. Non-specific binding was tested by adding

57 Instrumental Methods and Experimental Details

similar amounts of anti-mouse IgG-X4-Pr to PS microspheres without bioconjugation or to BSA coated PS microparticles. After several washes, the microspheres were subjected to bead-by- bead mass cytometry analysis for their Ln content as an indication of a successful immunoreaction. Between 30,000 and 60,000 microspheres were analyzed.

2.4.7 Fluorescence Emission

To demonstrate the feasibility of bioconjugation to the surface of particles prepared by dispersion polymerization, I treated a microsphere sample with FITC-avidin under EDC coupling conditions (details are given in the section 2.4.6 above). After 6 centrifugation-resuspension washing cycles with phosphate-buffered saline (PBS), the particle suspension (ca 0.3 mg/mL)

was examined for fluorescence (λex 485 nm) using the “front-face” geometry of a SPEX Fluorolog 3 spectrometer (Jobin Yvon/SPEX, Edison, New Jersey) in a 1 cm quartz cuvette.

58 Instrumental Methods and Experimental Details

References:

1. Bandura DR, Baranov VI, Ornatsky OI, Antonov A, Kinach R, Lou X, Pavlov S, Vorobiev S, Dick JE, Tanner SD: Mass cytometry: Technique for real time single cell multitarget immunoassay based on inductively coupled plasma time-of-flight mass spectrometry. Anal Chem 2009, 81(16):6813-6822. 2. Tanner SD, Bandura DR, Ornatsky O, Baranov VI, Nitz M, Winnik MA: Flow cytometer with mass spectrometer detection for massively multiplexed single-cell biomarker assay. Pure Appl Chem 2008, 80(12):2627-2641. 3. Baranov V, Tanner S, Bandura D, Quinn Z: Elemental Analysis of Tagged Biologically Active Materials. 2006. 4. Baranov VI, Bandura DR, Tanner SD: European Winter Conference on Plasma Spectrochemistry 2001:85. 5. Baranov VI, Quinn Z, Bandura DR, Tanner SD: A sensitive and quantitative element- tagged immunoassay with ICPMS detection. Analytical Chemistry 2002, 74(7):1629- 1636. 6. Careri M, Elviri L, Maffini M, Mangia A, Mucchino C, Terenghi M: Determination of peanut allergens in cereal-chocolate-based snacks: Metal-tag inductively coupled plasma mass spectrometry immunoassay versus liquid chromatography/electrospray ionization tandem mass spectrometry. Rapid Communications in Mass Spectrometry 2008, 22(6):807-811. 7. Hutchinson RW, Ma R, McLeod CW, Milford-Ward A, Lee D: Immunoassay with FI- ICP-MS detection - Measurement of free and total prostate specific antigen in human serum. Canadian Journal of Analytical Sciences and Spectroscopy 2004, 49(6):429-435. 8. Ornatsky O, Baranov VI, Bandura DR, Tanner SD, Dick J: Multiple cellular antigen detection by ICP-MS. Journal of Immunological Methods 2006, 308(1-2):68-76. 9. Zhang C, Wu F, Zhang Y, Wang X, Zhang X: A novel combination of immunoreaction and ICP-MS as a hyphenated technique for the determination of thyroid- stimulating hormone (TSH) in human serum. Journal of Analytical Atomic Spectrometry 2001, 16(12):1393-1396. 10. Zhang C, Zhang Z, Yu B, Shi J, Zhang X: Application of the biological conjugate between antibody and colloid Au nanoparticles as analyte to inductively coupled plasma mass spectrometry. Analytical Chemistry 2002, 74(1):96-99. 11. Zhang S, Zhang C, Xing Z, Zhang X: Simultaneous determination of α-fetoprotein and free β-human chorionic gonadotropin by element-tagged immunoassay with detection by inductively coupled plasma mass spectrometry. Clin Chem 2004, 50(7):1214-1221. 12. Stewart II, Olesik JW: Time-resolved measurements with single droplet introduction to investigate space-charge effects in plasma mass spectrometry. Journal of the American Society for Mass Spectrometry 1999, 10(2):159-174. 13. Abdelrahman AI, Dai S, Thickett SC, Ornatsky O, Bandura D, Baranov V, Winnik MA: Lanthanide-containing polymer microspheres by multiple-stage dispersion polymerization for highly multiplexed bioassays. J Am Chem Soc 2009, 131(42):15276-15283.

59 Instrumental Methods and Experimental Details

14. Lou X, Zhang G, Herrera I, Kinach R, Ornatsky O, Baranov V, Nitz M, Winnik MA: Polymer-based elemental tags for sensitive bioassays. Angewandte Chemie - International Edition 2007, 46(32):6111-6114. 15. Ornatsky OI, Lou X, Nitz M, Schafer S, Sheldrick WS, Baranov VI, Bandura DR, Tanner SD: Study of cell antigens and intracellular DNA by identification of element- containing labels and metallointercalators using inductively coupled plasma mass spectrometry. Anal Chem 2008, 80(7):2539-2547. 16. Horák D: Uniform polymer beads of micrometer size. Acta Polym 1996, 47(1):20-28. 17. Song JS, Tronc F, Winnik MA: Two-stage dispersion polymerization toward monodisperse, controlled micrometer-sized copolymer particles. J Am Chem Soc 2004, 126(21):6562-6563. 18. von Specht BU, Seinfeld H, Brendel W: Polyvinylpyrrolidone as a soluble carrier of proteins. Hoppe Seylers Z Physiol Chem 1973, 354(12):1659-1660. 19. Frank HP: The lactam-amino acid equilibria for ethylpyrrolidone and polyvinylpyrrolidone. Journal of Polymer Science 1954, 12(67):565-576. 20. Alam MA, Miah MAJ, Ahmad H: Synthesis and characterization of dual-responsive micrometer-sized core-shell composite polymer particles. Polym Adv Technol 2008, 19(3):181-185. 21. Omer-Mizrahi M, Margel S: Synthesis and characterization of magnetic and non- magnetic core-shell polyepoxide micrometer-sized particles of narrow size distribution. J Colloid Interface Sci 2009, 329(2):228-234. 22. Jennings TL, Rahman KS, Fournier-Bidoz S, Chan WCW: Effects of microbead surface chemistry on DNA loading and hybridization efficiency. Anal Chem 2008, 80(8):2849-2856.

60 Microspheres by Two Stage Dispersion Polymerization

3 Microspheres by Two Stage Dispersion Polymerization In this chapter, I describe the synthesis and characterization of metal-encoded polystyrene microspheres. These particles were designed as a platform for high throughput and highly multiplexed bioassays based upon analysis by mass cytometry. Most of the results given in this chapter were published in Ref. [1]

3.1 Introduction

For mass cytometry-based bioassays, metal encoded microspheres should have three major characteristics: 1- Diameter in the range of 0.8 – 3.0 µm 2- Metal content in the mass cytometry detection range 3- Surface functionality for bioconjugation

First, the microspheres must be large enough to be easily injected into the mass cytometer on a microsphere-by-microsphere basis. They must contain sufficient numbers of metal ions for our target range of codes, but be small enough to guarantee complete burning and ionization of the microspheres in the inductively coupled plasma torch. Polystyrene microspheres with diameters in the range of 0.8 to 3.0 µm satisfy these requirements. Microspheres of this size are also convenient to manipulate in terms of washing and redispersing. Larger particles may not burn completely in the ICP ion source, although we do not have any information on this point. I characterize particle size in terms of the mean particle diameter (d) obtained from the analysis of SEM images. These metal-containing microspheres should also have a very narrow size distribution. I characterize the size distribution in terms of coefficient of variation of the particle diameter (CVd).

1 1 n 2 CV = (3-1) d ∑i=1(Di −Dav ) Dav n −1

61 Microspheres by Two Stage Dispersion Polymerization

th where Dav is the number average diameter of all microspheres, Di is the diameter of the i microsphere, and n is the total number of particles counted in the analysis. Second, each microsphere must have a metal-content within the detection range of the mass cytometer. In addition, the metal-containing microspheres must have a very small microsphere-to-microsphere variation in their metal-content. I characterized the metal content by determining the average number of metal atoms per microsphere. Metal variation from microsphere-to-microsphere was evaluated by the magnitude of the coefficient of variation of metal content (CVLn), which was obtained from the mass cytometry measurements.

1 n 2 ∑ = − n −1 i 1(Lni Lnav ) CVLn = (3-2) Lnav

where Lnav is the number average metal ion content of all microspheres, Lni is the metal content of the ith microsphere, and n is the total number of microspheres counted in the analysis. Assuming that all the microspheres have the same density, the metal ion content of each microsphere should be directly proportional to its volume. The coefficient of variation

in volume (CVV) is defined by an expression analogous to that in eq (3-1). We expect that

CVLn ≥ CVV. Finally, metal-containing microspheres should have surface functionality that is suitable to attach affinity products like antibodies or aptamers. Carboxylic acid groups are a good candidate for surface functionality because of the availability of many acrylate or methacrylate monomers that can copolymerize with styrene to provide the carboxylic acid groups at the particle surface. Moreover, the carboxylic acid group facilitates a variety of bioconjugation reactions that can be employed in order to attach biomolecules to the surface of the microspheres. In this chapter, I will show how I designed the microsphere synthesis to meet the size requirements and how the metal ions can be incorporated into the microspheres. In addition, I will discuss metal content determination and briefly examine the surface functionality requirement. A detailed study of the metal concentration range that can be used in mass

62 Microspheres by Two Stage Dispersion Polymerization

cytometry will be given in Chapter 7. The employment of lanthanide-encoded microspheres in gene- and immune-assays will be discussed in Chapter 6.

3.2 Designing the Synthesis of the Microspheres

3.2.1 Techniques to Fulfill Size Requirements

There are a variety of heterogeneous (or particle forming) polymerization methodologies for the syntheses of polymer microspheres, for example, suspension polymerization, soap- free emulsion polymerization, seeded emulsion polymerization, dispersion polymerization and precipitation polymerization. Different polymerization methods produce particles having different size ranges as shown in Scheme 3 - 1. Horák. [2] have reviewed various methods that have been developed to produce polymer particles with diameters larger than 1 µm.

Scheme 3-1. Particle-forming polymerizations and size of resulting particles

In suspension polymerization,[3] the initiator is soluble in the monomer, and these two are insoluble in the polymerization medium. The monomer phase is, suspended in the medium in the form of small droplets (microdroplets) by means of a stirrer and a suitable droplet stabilizer (suspension agent). The polymerization is then initiated at the desired temperature (20-100 °C) and is usually allowed to proceed to completion (~ 100%). Under these conditions, the monomer "microdroplets" are converted directly to the corresponding

63 Microspheres by Two Stage Dispersion Polymerization

polymer "microspheres" of approximately the same size. The average size of microspheres and the width of the size distribution depend mainly, among other reasons, on the size and design of polymerization equipment (reactor, stirring, etc.). Typically, suspension polymerization can be used to synthesize microsphere with sizes range from 10 µm to 5 mm. This simple technique, however, affords microspheres that have very broad size distribution. Surfactant-free emulsion polymerization (SFEP) [4, 5], involving no added surfactant, is a useful approach to prepare surface-clean polymer particles. The process uses an initiator yielding initiator radicals that impart surface-active properties to the polymer particles. Persulfates (like KPS) are useful initiators for this purpose. Particles synthesized via the surfactant-free technique are stabilized by chemically bound sulfate groups of the initiating species, , derived from persulfate ion. Since the surface-active groups are chemically bound, the latexes can be purified (freed of unreacted monomer, initiator, etc.) without loss of stability. The particles prepared by SFEP system are usually very narrow dispersed. However, there are two main limitations for using this method, first, the solid content should be maintained below 5 % to be able to produce narrow size distribution [6]. Second, using SFEP, one cannot get particles larger than 1.0 µm in diameter. Polymeric microspheres can be prepared by seeded emulsion polymerization from latexes of smaller size either by Vanderhoff’s successive seeded polymerization method [7] or by Ugelstad’s two-step activated swelling and suspension polymerization method [8]. In our group, Thickett et al. [9] synthesized polymer microspheres using a combination of SFEP and seeded emulsion polymerization. The method of seeded emulsion polymerization provided many technical advantages for that work – large particle sizes (ca. 1.0 µm), monodisperse particle size distributions and the presence of an extensive number of surface acid groups per particle using acid-functional azo initiator (4,4’-azobis(4-cyanovaleric acid), ACVA). The main limitation of seeded emulsion polymerization method is the complication imposed by using multiple stapes (i.e. seed syntheses and growth steps). In addition, seeded emulsion polymerization is difficult to implement on a large scale. Precipitation polymerization [2, 10] begins initially as a homogeneous system in the continuous phase, where the monomer and initiator are completely soluble. Initiation and

64 Microspheres by Two Stage Dispersion Polymerization

polymerization take place largely in the homogeneous medium. Upon initiation, the polymer formed is insoluble and precipitates as unstable nuclei. The continuous nucleation and the coagulation of the resulting nuclei yield larger and larger particles. After precipitation, the polymerization proceeds by absorption of monomer and initiator into the polymer particles. Precipitation polymerization normally produces irregularly shaped and polydisperese particles in a very wide size range. Dispersion polymerization [2, 10, 11] can be classified as a type of precipitation polymerization in which one carries out the polymerization of a monomer in the presence of a suitable polymeric stabilizer soluble in the reaction medium. A good solvent for both the monomer and the steric stabilizer polymers is always used. However, the solvent should be also a non-solvent for the polymer being formed. Dispersion polymerization, therefore, involves a homogeneous solution of monomer(s) with initiator and dispersant. Sterically stabilized polymer particles are formed by the precipitation of the resulting polymers. Under favorable circumstances, dispersion polymerization can be used to synthesize polymer particles of 1.0–10 µm in diameter, often of very narrow size distribution. Dispersion polymerization has advantages over other techniques due to its superiority in the production of very monodisperse particles and the facility to be implemented on a large scale. However, dispersion polymerization is extremely sensitive to small changes in reaction conditions. For example, stirring blade depth inside the reaction flask has an unexpectedly strong effect on the particle size and size distribution as described in Chapter 2. Until recently it was not possible, using dispersion polymerization, to prepare particles bearing functional groups, crosslinked particles, or other comonomer particles, without loss of control over particle size and the resultant size distribution.[12, 13] A major advance was done by Song et al.[14] who showed that the nucleation step of particle formation is much more sensitive to the presence of other monomers or crosslinking agents than the subsequent particle growth stage. By deferring the addition of “problematic reagents” until the nucleation stage of the reaction was complete (e.g. after 5% styrene conversion), Song et al.[15-17] synthesized a variety of functional particles. The process is referred to as “two-stage” dispersion polymerization (2- DisP). In this chapter, I use 2-DisP

65 Microspheres by Two Stage Dispersion Polymerization

with ethanol as the solvent to synthesize microspheres with a diameter of ca 2.0 µm. To incorporate lanthanide (Ln) metals in the polymer particles, lanthanide chloride(s) were introduced in the second stage along with a comonomer like AA or AAEM.

3.2.2 Metal-Content Requirements

The lanthanide (Ln) series of elements have been selected as metal tags for the mass spectroscopy-based bioassays because of the large number of commercially available stable isotopes (24 elements and 54 individual stable isotopes) and similar chemistry that facilitate their incorporation into the microspheres using a common synthesis methodology.[18] In addition, because of the low natural abundance of Ln elements, low background signal in the biological assays is anticipated. The challenge was to synthesize microspheres with large and uniform numbers of Ln ions per microsphere. Each microsphere must contain a sufficient number of ions for robust detection using counting statistics as a criterion. The transmission efficiency of mass cytometry is on the order of 10-4, which means that for every 104 ions generated in the plasma only one ion reaches the detector. This inefficiency is balanced by a noise level of only 0.01 counts per second, meaning that a signal of several counts per microsphere recorded over a time frame less than 1 ms is significant. Thus the target for ion incorporation is between several times 104 to 108 Ln ions per microsphere. To synthesize microspheres with 104 to 108 Ln ions per microsphere, I calculated the average weight of one microsphere based on the size obtained from SEM micrographs and the polystyrene density (1.05 g/cm3). I introduced the corresponding weight of lanthanide salt assuming the 100 % incorporation of Ln ions into the microspheres. Further discussion about the range of Ln ion concentration that can be used and detected by mass cytometry will be covered in Chapter 7.

66 Microspheres by Two Stage Dispersion Polymerization

3.3 Results and Discussion

3.3.1 Synthesis of the Microspheres

To synthesize polymer microspheres meeting the above-mentioned characteristics, I chose two-stage dispersion polymerization of styrene in ethanol in the presence of polyvinylpyrrolidone (PVP).[19] It is widely recognized that dispersion polymerization of styrene in ethanol leads to particles of the proper size range with an exceptionally narrow

size distribution. It is relatively easy to obtain particles with CVd ≈ 3% (corresponding to

CVV ≈ 10%), and with care, even narrower distributions can be obtained.[20] One of the difficulties with this method is that the introduction of functional co-monomers (“problematic reagents”) can lead to loss of control over particle size and size distribution.[12, 13] Several years ago, Song et al.[14] showed that many of these problems could be avoided if the addition of the co-monomers was delayed until the particle nucleation step was complete (a few percent monomer conversion). He referred to this method as “two- stage” dispersion polymerization (which I denote as 2-DisP). Thus in my design, I initiated dispersion polymerization of styrene in ethanol, and after approximately 10% monomer conversion, added known amounts of LnCl3 in ethanol in the presence of an excess of a comonomer that can serve as a ligand for Ln3+. I examined two different ligands: acetoacetylethyl methacrylate (AAEM), a β-ketoester, and acrylic acid (AA). AAEM was previously used as a chelating ligand for zirconium propoxide to incorporate it into hybrid nanoparticles through a combination of a sol–gel reaction and emulsifier free emulsion polymerization.[21-24] The carboxylate group of AA is known to interact strongly with lanthanide ions. The reagents used in these 2-DisP reactions are presented in Table 3-1, and the type and amounts of lanthanide salts employed are collected in Table 3-2. Experimental details are described in Chapter 2. The characteristics of some of microsphere samples synthesized are collected in Table 3-3. The reactions were clean, and yielded particles in high gravimetric yield, with no coagulum. Figure 3-1A shows a scanning electron microscope (SEM) image for sample

67 Microspheres by Two Stage Dispersion Polymerization

AA068 in which 1.0 wt%/sty EuCl3 and 8.0 wt%/sty AAEM were added in the second stage. For comparison, Figure 3-1B presents an SEM image for sample AA069 in which 1.0

wt%/sty EuCl3 and 2.0 wt%/sty AA were added in the second stage and Figure 3-1C presents

an SEM image for sample AA069 in which 1.0 wt%/sty EuCl3 and 2.0 wt%/sty AA were added in the second stage. The two samples have nearly identical mean diameters (d = 2.7

μm) with a very narrow size distribution (CVd = 1.5 %). Similarly, monodisperse

microspheres were obtained when 1.0 wt%/sty EuCl3 and 4.0 wt%/sty AAEM were added in the second stage, as shown in Figure 3-2. We conclude that the 2-Disp reaction works well to provide particles of the proper dimensions and appropriately narrow size distribution.

Table 3-1. Recipe for the synthesis of microsphere sample AA088 as an example for 2-stage dispersion polymerization (2-DisP) of styrene with PVP55 as a dispersant in ethanol.

Materials (grams added) 1st stage 2nd stage Styrene 6.25 --a PVP55 a 1.0 -- TX305 b 0.35 -- AMBN c 0.25 -- Ethanol 18.75 18.75 Comonomerd -- 0.125e f EuCl3 -- 0.063

a polyvinylpyrrolidone, Mw ≈ 55 kDa.

b Triton X-305 c azomethylbutyronitrile d AA was used as a comonomer for AA088 and all the other samples except for AA050 and AA068 where AAEM was used as a comonomer. e In the synthesis of AA068 samples, 0.25 g of AAEM (4 wt%/sty) were used and in sample AA050, 0.50 g of AAEM (8 wt%/sty) were used.

f EuCl3 is EuCl3.6H20. For other microsphere samples, other combinations of

LnCl3.6H20 were used (where Ln = La, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho and/or Tm).

68 Microspheres by Two Stage Dispersion Polymerization

Table 3-2. Lanthanide content a of the reaction mixture for synthesis of PS-PAA microspheres (2 wt % AA/styrene) a a a a a a a a Sample La Tm Ho Tb Eu Pr Lu Ln total AA050 ------1.0 -- -- 1.0 AA068 ------1.0 -- -- 1.0 AA069 ------1.0 -- -- 1.0 AA087 0.10 0.10 0.10 0.10 ------0.40 b AA088 0.10 -- 0.10 0.10 ------0.30 c AA089 -- 0.10 0.10 0.10 ------0.30 AA093 0.2 0.2 0.2 0.2 ------0.8

AA099 0.05 0.05 0.05 0.05 ------0.20 AA100 0.02 0.02 0.02 0.02 ------0.08 AA181 0. 5 0.05 -- 0.05 -- 0.5 0.05 1.15

a wt % relative to styrene of LnCl3.6H2O; b d = 1.8 ± 0.032 μm, CVD = 1.2 % c d = 1.8 ± 0.041 μm, CVD= 1.3 %

Figure 3-1. SEM images for PS microsphere samples AA068, AA069 and AA050 synthesized in the presence of 1 % EuCl3 added in the second stage with (A) AAEM: 8 wt %/styrene (d = 2.7 μm, CVd = 1.5%), (B) AA: 2 wt %/styrene (d = 2.7 μm, CVd = 1.4%). and (C) AAEM: 4 wt %/styrene (d = 2.7 μm, CVd = 1.5%). The white horizontal bars in the microphotographs represent 5 µm.

69 Microspheres by Two Stage Dispersion Polymerization

3.3.2 Determining Lanthanide Content by Mass Cytometry

In this section, I describe how mass cytometry is used to determine the microsphere-by- microsphere Ln ion content. The microspheres were washed by several cycles of centrifugation and resuspension in water, and the resultant slurry (ca. 106 microspheres/mL) was nebulized into the mass cytometer sample introduction system, which in turn delivered microspheres individually but stochastically into the inductively coupled plasma torch. The high temperature of the plasma (ca. 7000 K) was found to be sufficient to atomize and then ionize the microspheres and the Ln ions embedded in them. The ion stream was then introduced into the time-of-flight mass analyzer. The transient signals corresponding to each microsphere ionization event were recorded by the detector and stored. I first introduce the raw data to illustrate how the system works. To visualize the data stream, we use a presentation, sequentially stacked mass spectra, employed in LC-MS. Figure 3-2A shows about 5 ms of data following injection of a sample of AA068 (AAEM). Corresponding information for sample AA069 (AA) is shown in Figure 3-2B. The instrument captures sequential mass spectra taken at 13 µs intervals. Approximately 1000 microspheres per second were analyzed in this way. The y-axis reports the start time of acquisition of each mass spectrum. The x-axis corresponds to the arrival time of ions and is related to the mass of the ions. Columns that correspond to Eu isotopes (151Eu and 153Eu) are highlighted by the vertical lines in the middle of each figure. An encoded particle is registered as a signal in 10 - 30 consecutive spectra over (0.1 - 0.4 ms), and is characterized by simultaneous transient signals for all encoding elements or isotopes. In Figure 3-3A and B the ion signals that correspond to individual microsphere ionization events are highlighted by horizontal bars. Each sample (AA068 and AA069) shows the signature of two individual particles. The qualitative information presented in Figures 3-2 is used for visualization of the transient signals, for their algorithmic bracketing and decoding. In addition, if free elements were present in the solution, caused, for example, by ion leakage from the particles, a continuous signal very different from the transient signal from individual particles would be observed.

70 Microspheres by Two Stage Dispersion Polymerization

Figure 3-2. Screen captures for PS microsphere samples (A) AA068 and (B) AA069. AA068 was synthesized in the presence of 1 % EuCl3 added in the second stage with AAEM: 8 wt %/styrene and AA069 was synthesized in the presence of 1 % EuCl3 added in the second stage with AA: 2 wt %/styrene.

Quantitative information about the particle population can be drawn from a histogram representation of the frequency distribution of signal intensities for individual microspheres. In Figure 3-3A, the population distribution is presented for the 151Eu ion signal collected for 3 min (ca. 4 x 105 microspheres) for AA050 particle sample, which was synthesized in the

presence of AAEM (4 wt. % / styrene) and EuCl3 (1.0 wt. % / styrene, Table 3-1). The x-axis of this plot is the “151Eu intensity” analog output of the time of flight (TOF) detector and is

71 Microspheres by Two Stage Dispersion Polymerization

considered here as a relative number. The overall mean value of 151Eu intensity is 9100 with

overall Eu-intensity distribution (CVEu) of 103 %.

Figure 3-3: Distribution of mass cytometry signal intensity for AA050 PS microspheres prepared 151 153 in presence of EuCl3 (1.0 wt%/styrene) plus AAEM (4.0 wt%/styrene) (A): Eu and (B): Eu Intensity distributions.

In Figure 3-3A, one can easily identify three regions according to the 151Eu intensity for AA050 particles. The middle region (denoted region M), which is represented by the most intense peak in the figure, had 78 % of the whole particle population. This sharp peak in the 151 middle region had a mean value of Eu of 6,900 and CVEu of 18 %. To the left of region M, there was a low-Eu intensity region (region L) that had an average 151Eu intensity of 1100. This low-Eu intensity region formed 4 % of the particle population. The third region was the

high-Eu intensity region (region H) that can be subdivided into three sub-regions (H2, H3 and 151 H4 sub-regions). H2 sub-region had an average Eu intensity of 15000. This region 151 represented 14 % of the particle population while H3 sub-region had an average Eu intensity of 22000 and represented 3 % of the particle population. The very high-Eu intensity 151 region (H4) had an average Eu intensity of 78000 and represented only 1 % of the whole particle population. I found similar regions in the distribution of 153Eu signal intensity in sample AA050 (Figure 3-3B).

72 Microspheres by Two Stage Dispersion Polymerization

In this section, I propose the assignment of each region, in Figure 3-3A, to certain particle population. Because it represented the highest population and narrowest Eu-intensity distribution (CVEu), I saw region M as a representation of the mass cytometry signals obtained from single particle events (not debris or 2 or more particles reaching to the detector as one event). I perceived Region L as the mass cytometry signal of the Eu ions in the

continuous media or some particle debris. Because the mean value of H2 sub-region is about

the double of that of the middle major population, I assigned this H2 population to the events in which 2 particles arriving the detector together as one particle. This kind of event (in region H2) are thus attributed to “doublets”. H3 and H4 sub-regions were seen as representation of events in which three or more particles arriving the detector together as one particle. To understand the effect of the sample concentration (number of particles per mL of water injected to the mass cytometry) on the distribution of signal in these regions and to confirm my proposed assignment mentioned above, I prepared three solutions of different particle concentrations (105, 106, and 107 particles per mL) of the same sample and examined them by mass cytometry. For this experiment, I synthesized sample AA181 in the presence of AA (2 wt. % / styrene) and a mixture of lanthanide salts (total 1.15 wt. % / styrene, Table 3- 2). Figure 3-4A-C show the 169Tm signal distributions collected for 2 min for three different concentrations of AA181 sample. The three different concentrations 105, 106, and 107 particles per mL had Tm average intensities of 189, 196 and 268, respectively. Similar to sample AA050, when particle concentration of 106 particles per mL was used (Figure 3-4B), three main regions could be seen in the distribution of Tm signals. The average Tm intensities of the L, M and H regions were 38, 190 and 385, respectively. However, the representation of each region was quite different from AA050. For this sample (with particle concentration of 106 particles per mL), region M represented 93 % of the whole particle population and Regions L and H represented 4 % and 7 % of the particle population, respectively. In addition, the H sub-regions could not be resolved (Figure 3-4B). Increasing the particle concentration to 107 particles per mL (Figure 3-4C) did not affect the L region

73 Microspheres by Two Stage Dispersion Polymerization

(Tm average intensity of 38 and 4 % of the particle population) but resulted into a much bigger H region (30% of the whole particle population). Moreover, two H sub-regions could be resolved with average Tm intensities of 400 and 720. Using a lower concentration (ca. 105 particles per mL, Figure 3-4A) resulted into similar L region representation and very low H region representation (1 %, Tm average of 380). The number of particles reach to the detector of the mass cytometer (in the 2 min data acquisition time) is proportional to the concentration of particles in the analyzed samples. In other words, the particle count rate increase for samples with higher particle concentration. For the mass cytometry measurements of sample AA181, particle concentrations of 105, 106, and 107 particles per mL resulted into 1,400, 11,000 and 35,000 counts, respectively. The three different concentrations (105, 106, and 107 particles per mL) examined for sample AA181, had average Tm intensities ranged from 190 to 268. However, considering only M regions, the three particle concentrations had the same Tm content (190) and similar

CVTm (27 - 28 %). See the dashed line in the overlay plot (Figure 3-4D). In parallel, increasing the particle concentration led to the enlargement of the H region representation (the doublets). See the arrow in the overlay plot (Figure 3-4D). These results suggest that increasing the sample concentration increase the probability of particle aggregation during the mass cytometry measurements, hence higher representation of the H region. For the Lu content of sample AA181, Figure 3-4E-H shows that the same trend (of increasing H region representation) is observed with increasing the particle concentration. I conclude that, in the mass cytometry measurements of my PS samples, it was possible to minimize the extent of particle aggregation (represented by region H), if I used lower sample concentration (lower than 106 particles per mL). However, decreasing the sample concentration would result in longer measurement time to be able to obtain statistically representative data (> 10,000 particles).

74 Microspheres by Two Stage Dispersion Polymerization

Figure 3-4: Distribution of mass cytometry signal intensity for AA181 PS microspheres prepared in presence of mixture of lanthanides (total 1.15 wt%/styrene, Table3-2) plus AA (2.0 wt%/styrene) (A-D): 169Tm and (E-H): 175Lu Intensity distributions. A and E: 105 particles per mL, B and F: 106 particles per mL, and C and G: 107 particles per mL. The dashed lines in D and H overlay plots refer to the constant Tm intensities of the M region with different particle concentrations. The arrows in D and H overlay plots are showing the increase of the doublets (H region) with increasing particle concentration. 75 Microspheres by Two Stage Dispersion Polymerization

The bivariate plot is a method to show the relationship between two isotopic concentrations that have been measured on a single microsphere. Such plots permit us to see at a glance the degree and pattern of the relation between the two isotopes in the sample. Most importantly, using bivariate plots, one can group the populations of the microspheres that have similar contents of different isotopes. On a bivariate plot, the x- and y-axes can represent the concentration of any two isotopes of interest. Each point on the plot shows the x and y isotopic-content for a single microsphere.

Figure 3-5: 151Eu/153Eu bi-variant plots of mass cytometric results for AA050. PS microspheres 7 (10 particles / mL). AA050 microspheres were prepared by 2-DisP in presence of EuCl3 (1.0 wt%/styrene) plus AAEM (4.0 wt%/styrene).

Figure 3-5 shows a logarithmic 151Eu/153Eu bivariate plot for the AA050 sample. Obviously, particles that had high 151Eu-content, had high 153Eu-content as well. This proves that both Eu isotopes had the same tendency to be incorporated into the PS particles. The overall average intensity of 151Eu-content was 9100 while 153Eu had an overall average intensity of 9970. Thus the ratio of the mass cytometry signal 151Eu/153Eu was 0.91 which agrees with the isotopic natural abundance. Similar to the population distribution histogram of 151Eu for AA050 sample (Figure 3-3A), the 151Eu/153Eu bivariate plot, in Figure 3-5, shows the three populations of particles (L, M and H regions highlighted by three circles) in terms of their Eu content.

76 Microspheres by Two Stage Dispersion Polymerization

Figure 3-6: Gated distribution of mass cytometry signal intensity for AA050 PS microspheres prepared in presence of EuCl3 (1.0 wt%/styrene) plus AAEM (4.0 wt%/styrene). Gating was done by excluding the L and H regions in Figure 3-3A (A) and in Figure 3-5 according (B).

In order to obtain analytically reproducible mass cytometry data for the same sample when it is measured several times, the data obtained from the low- and high-lanthanide intensity regions (L and H regions) should be disregarded. In this thesis, I consider only the middle region (M region) for the calculation of the lanthanide content and content distribution. This process of disregarding the L and H regions is called “gating”. i.e. to gate out the unrepresentative signals during the calculation of the metal content. This approximation (gating) can be acceptable, only if the M region represents the predominant population of the signals obtained. In this thesis, I consider the gating process is acceptable, if more than 75% of the particles population fall into the middle region. There are two ways to manually apply gating on the mass cytometry signal using the FlowJo software. The first method employed the frequency distribution histogram of one isotope and gated out regions to the right and to the left of the main intense peak. The two vertical lines (defining region M in Figure 3-3A) represent the gating applied using the first

77 Microspheres by Two Stage Dispersion Polymerization

method. These lines are the boundaries of M region. They are drawn “manually” to pass through the local minima at the beginning and at the end of the intense peak that represent the M region and have the global maximum of the intensity distribution plot. Figure 3-6A represents the gated histogram produced by this first method. In the second method, I used the bivariate plot of any two isotopes presented in the sample. In this method, gating is done by selecting the middle region (M region, defined by the middle oval in Figure 3-5) and excluding (gating out) the signals outside it. In this bivariate plot gating, the oval is drawn manually as well. Applying this method of gating resulted in the gated bivariate plot in Figure 3-6B. Moreover, one can also use the same set of gated data represented in Figure 3- 6B to create a frequency distribution histogram of one isotope similar to Figure 3-6A. In this thesis, I usually use the bivariate plot gating (the second method) unless the sample contains only one isotope; in this case, I use the first method of gating that based on frequency distribution histogram. In Figure 3-7, the gated population distribution (using the bivariate plot gating) is presented for the 151Eu ion signal for three different samples, two containing AAEM and one containing AA. The mean 151Eu intensities are 6900 for AA050, 15100 for AA068 and 90600 for AA069. These results indicate an increase of Eu ion content of the microspheres with increasing concentration of the AAEM co-monomer, and more importantly, a much higher Eu content for the AA-containing microspheres. The higher Eu ion content for the AA- containing microspheres implies that, under our reaction conditions, carboxyl groups are more effective ligands than the acetoacetyl group for incorporating lanthanide ions into PS particles. As a consequence, for subsequent experiments described in this thesis, I focus on particles synthesized in the presence of acrylic acid.

78 Microspheres by Two Stage Dispersion Polymerization

AA050 4.0 % AAEM

AA068 8.0 % AAEM

AA069 2.0 % AA Distribution

151Eu Intensity

Figure 3-7. Distribution of mass cytometry signal intensity for three different populations of PS microspheres prepared in presence of EuCl3 (1.0 wt%/styrene) plus (AA050) AAEM (4.0 wt%/styrene, 151Eu Intensity = 6900), (AA068) AAEM (8.0 % wt%/styrene, 151Eu Intensity = 15100) and (AA069) AA (2.0 wt%/styrene, 151Eu Intensity = 90600).

The average number of metal ions per microsphere can be calculated from the mean intensity values as follow:

× F = (3-3) 𝐼𝐼 𝐼𝐼 where I is the mean intensity measured by𝑁𝑁 the TOF𝑇𝑇 detector; IF is the analog-to-count conversion factor for a particular ion (related to the mass response); and T is the transmission coefficient of the entire instrument. The transmission coefficient (the number of ions that reach the detector per number of ions injected) depends on tuning and the mass response of the instrument.

151 -5 For the case of AA050, the value of IF was 0.0451 for ( Eu) and T was 3.86 x 10 . The average 151Eu content of AA050 was calculated to be 8.09 x 106 151Eu ions per microsphere. Similarly, the 151Eu content of AA068 and AA069 were calculated to be 1.76 x 107 and 1.06 x 108 151Eu ions per microsphere, respectively.

79 Microspheres by Two Stage Dispersion Polymerization

3.3.3 Microsphere Encoding Protocols: Variability and Dimensionality

In this section, I consider encoding strategies for polymer particles based upon the idea that one can control lanthanide ion content per particle for a variety of lanthanide elements. There are several encoding strategies and for the purpose of this thesis, I discuss only the binary and enumeration protocols. Binary encoding represents the simplest system. The encoding elements are introduced in two concentrations (levels): 0 and 1. Thus binary encoding of particles with one or more of 10 encoding elements leads to a variability of 210 − 3 N 1 > 10 distinguishable particles. The general formula is VR = 2 −1, where VR is the variability of a binary encoding system employing N elements. Enumeration encoding is a logical extension of the binary code. In this system, elements are introduced at several N concentrations (K levels) with total variability VR = K −1. This is the highest possible variability for any encoding strategy. For 10 elements and 5 levels of concentration, the 6 variability is VR >> 10 . I begin with an examination of binary encoding. Here the task is to synthesize a series of microspheres containing one or more lanthanide elements and to examine the ability of mass cytometry to detect and resolve these elements. One set of examples is presented in Table 3-2 (AA087, AA088 and AA089). Each of the samples was synthesized with acrylic acid (2 wt%/sty) in the second stage, along with different combinations of LnCl3 salts. These particles are similar in size (d ≈ 2 μm, CVd = 1.0 - 2.0 %) with a narrow size distribution.

80 Microspheres by Two Stage Dispersion Polymerization

9000 9300 9600 9900 10200 10500 10800 0 La TbTb Ho

1.3

2.6 BBaa AA088 3.9 (A) AA088 0

Tb HoTm 1.3

2.6

3.9 AA089 (B) AA089 Transit time, ms time, Transit Transit time, ms time, Transit 0

1.3 Tm

2.6

3.9 AA110 (C) AA110

Figure 3-8. Mass Cytometry screen captures for the analysis of three types of PS microspheres, each containing similar amounts of different Ln metals. (A) sample AA088: AA 2.0 + LaCl3 0.1 + TbCl3 0.1 + HoCl3 0.1; (B) sample AA089: AA 2.0 + TbCl3 0.1 + HoCl3 0.1 + TmCl3 0.1 wt%/styrene. The numbers following each species refers to the wt%/styrene using in the particle synthesis. (C) Sample AA110: a control experiment in which TmCl3 0.1 wt%/styrene and no AA was added in the particle synthesis.

81 Microspheres by Two Stage Dispersion Polymerization

Table 3-3 Particle size, size distribution, and the variation of Tm intensities for some PS-PAA microspheres synthesized in the presence of LnCl3.6H2O.

a b c c 3 d d e f g Sample Ln Tm d ± σd (μm) CVd% V, σV(μm ) CVV% CVTm% % incorp. –COOH

AA087 0.40 0.10 2.1 ± 1.8 4.9 ± 0.3 5.5 25 89 % 2.1 x 108 AA099 0.20 0.05 1.8 ± 1.1 3.1 ± 0.1 3.5 42 88 % 1.9 x 108 AA100 0.08 0.02 1.6 ± 2.9 2.3 ± 0.3 9.2 54 87 % 2.0 x 108

a wt %/styrene of LnCl3.6H2O.

b wt % of TmCl3.6H2O based upon styrene used in the particle synthesis, see Table 3- 2.

c d = mean diameter; σd = one standard deviation; CVd = coefficient of variation of the diameter.

d V = mean particle volume; σV = one standard deviation; CVV refers to the microsphere volume. e CV of 169Tm intensity measured by mass cytometry. f % of Tm in the reaction that is incorporated into the microspheres. g Number of –COOH groups/microsphere by titration.

To demonstrate the ability of mass cytometry to resolve the presence of multiple Ln ions in a single microsphere, I present data in Figures 3-8A and B, which represent screen captures of signals from samples AA088 (La, Tb, Ho) and AA089 (Tb, Ho, Tm). The dense vertical lines labeled “La,” “Tb” and “Ho” in Figure 3-8A (AA088) refer to signals from multiple mass spectra taken during the transit of a single microsphere through the plasma torch. The dark vertical line labeled “Ba” is due to trace amounts of barium in the buffer solution used for the introduction of this sample into the mass cytometry instrument. Figure 3-8 B shows signals from two successive microspheres of Sample AA089. One can see clear and distinct signals of the elements Tb, Ho, and Tm, but La is not detected. To test whether Ln ions associate or undergo non-specific binding to the PVP corona surrounding the

82 Microspheres by Two Stage Dispersion Polymerization

particles, I synthesized a sample (AA110) in which TmCl3 but no acrylic acid was added in the second stage of the 2-DisP reaction. No signals were detected in the mass cytometry analysis of this sample (Figure 3-8C) that could be associated with individual particles. One can see only a thin streak corresponding to weak traces 169Tm3+ ions present in the solution. These results demonstrate that particles suitable for binary encoding can be synthesized by 2- DisP and read by mass cytometry. The next objective in building a multiplexed assay by mass cytometry detection is to move to enumeration encoding which incorporates a variety of metals and a range of different levels of concentration. To meet this goal, we have to demonstrate the ability of mass cytometry to detect and quantify different levels of metal ion concentration. To test this idea, I synthesized four sets of PS microsphere samples using in the synthesis five levels of Ln concentration (0, 0.02, 0.05, 0.10 and 0.20 wt%/sty) introduced in the second stage of the dispersion polymerization reaction. In these reactions, I used four elements (La, Tb, Ho, and Tm). Data for four sets of particles are presented in Figure 3-9. Here the y-axis represents the normalized distribution of each element. The x-axis reports on the signal intensity on a log scale associated with each element. The panels are grouped by element.

83 Microspheres by Two Stage Dispersion Polymerization

Figure 3-9.. Distribution of gated signal intensity for encoding elements for four different populations of microspheres. The microspheres were encoded with four elements (La, Ho, Tb, and Tm) and five levels of concentration (coded from “0” to “4”). For this system of encoding the variability is equal to 624. The label “1” refers to the concentration level of 0.02 wt% LnCl3/sty. Thus, the label “2” refers to the concentration level of 0.05 wt% LnCl3/sty, the label “3” refers to the concentration level of 0.1 wt% LnCl3/sty and the label “4” refers to the concentration level of 0.2 wt% LnCl3/sty.

84 Microspheres by Two Stage Dispersion Polymerization

The normalized peaks in each panel are labeled “1”, “2”, “3” and “4”, respectively, to indicate that they originate from PS particles prepared in the presence of 0.02, 0.05, 0.10 and

0.20 wt% LnCl3/sty, respectively. The label “0” refers to the absence of a Ln element, and, of course, there is no peak in the mass spectrum corresponding to concentration level “0”. The label “1”, for example, on any peak refers to the concentration level of 0.02 wt%

LnCl3/sty. Thus, the label “2” on any peak refers to the concentration level of 0.05 wt%

LnCl3/sty, the label “3” on any peak refers to the concentration level of 0.1 wt% LnCl3/sty

and the label “4” on any peak refers to the concentration level of 0.2 wt% LnCl3/sty. The data in Figure 3-9, collected for several thousand microspheres, demonstrate essentially baseline resolution in the ability to detect 169Tm, 159Tb, 165Ho, and 139La over a concentration range of three orders of magnitude. Thus the enumeration encoded microspheres can be distinguished successfully with minimal overlap. This example of enumeration encoded microspheres has a variability of 624 (i.e. 54 − 1) and with a larger range of elements or isotopes, opens the possibility to resolve an extremely large number of unique biomarkers.

3.3.4. Lanthanide Incorporation, Particle-to-Particle Variability and Surface Functionality

The widths of the peaks in Figure 3-9 provide information on the microsphere-to- microsphere variability of the Ln content of the microspheres. It appears that the greatest variation is in microsphere sample AA100 (distributions labeled with “1” in Figure 3-9). The AA100 microsphere sample was synthesized with the lowest concentration of Ln ions (0.05

wt% LnCl3/sty) relative to the other two samples (AA087 and AA099). More quantitative information about this variability is presented in Table 3-3, where I list values for the coefficient of variation of the Tm content (CVTm) for four PS microsphere samples. They range from 25% to 54%, and are substantially larger than the variation in volume of the

particle samples (CVV). The other elements present in these particles show similar variability. Broadening in the Ln content distribution of the microspheres relative to their size distribution is not well understood. This unexpected broadening might be caused by

85 Microspheres by Two Stage Dispersion Polymerization

synthesis-related factor that perturb the expected Poisson distribution of lanthanide ions per microsphere during the polymerization. In addition, the statistical error associated with mass cytometry measurements may also contribute to the broadening of the lanthanide distribution. In Chapter 4, where I present some examples of microsphere samples with very narrow size

distribution, the minimum CVLn I could obtain was 12 to 15 %. It is possible that both synthetic and instrument factors contribute to the broadening in CVLn relative to CVd. Although the general broadening in the Ln content distribution of the microspheres is not well understood, the relatively greater broadening for microsphere samples that have lower lanthanide concentration (like AA100) can be explained. When the measured concentration decreases, the signal-to-noise ratio also decreases, especially when the same background matrix (same polystyrene sample and dispersing media) is employed. Accordingly, the percentage of the error (standard deviation) increases and that manifests itself in broader distributions. Other information about these metal-encoded polymer microspheres is also presented in Table 3-3. The microspheres prepared by 2-DisP, have approximately 2 x 108 titratable carboxyl groups per particle, which I assume are on or near the particle surface. These groups are intended for covalent attachment of biomolecules such as antibodies, and this level is comparable to that of commercial particles intended for bioassays. For example, carboxylated PS microspheres from Bangs Labs, # PC04N have an average of 8.4 x 108 –COOH group per microsphere. To examine the efficiency of using acrylic acid in the second stage of 2-DisP synthetic methodology to anchor (or incorporate) the lanthanide ions into the polystyrene microspheres, I use the term “incorporation efficiency”. The incorporation efficiency is the ratio between the amount of lanthanide ions measured in the microspheres to the amount added in the synthesis. The amount of the lanthanide ions in the microspheres was measured indirectly by measuring the lanthanide ions in the serum (the continuous phase) with convensional ICP-MS. The particles were first separated from the serum. Basically, 1.00 mL of the “as-prepared” particle dispersion was sedimented with a high speed centrifuge at 14,000 rpm for 30 minutes. The supernatant was collected, diluted with 3 % high purity

86 Microspheres by Two Stage Dispersion Polymerization

HNO3 solution and analyzed by ICP-MS. The values of the incorporation efficiency of Thulium into PS microsphere synthesized by 2-DisP are given in Table 3-3. 88% of the Tm3+

ions introduced as TmCl3 in the reaction were incorporated into the polymer particles. ICP- MS measurements confirmed that the rest of Tm ions were found to remain (or to be released during or immediately after the reaction) in the continuous ethanolic medium. Similar results were obtained for all the Ln salts used in the synthesis.

6 pH 10.6 4

ppb pH 7 2

pH 3 0

10 1000 100000

Log (Time / min)

Figure 3-10. Tm ion release into the aqueous phase from colloidal suspensions of two Tm-containing PS microsphere samples in three different buffer solutions. AA089 microsphere sample contain 260 ppm Tm ion (w/w styrene). The pH 10.6, pH 7.0 and pH 3.0 buffer solutions are 200 mM sodium carbonate/bicarbonate, 10 mM ammonium acetate, and 50 mM sodium acetate, respectively.

One of the important characteristic for microspheres that need to be assessed is the stability of the particles toward leakage of the Ln ions into the aqueous medium, particularly under the experimental conditions associated with attaching biomolecules to the particles. We used traditional ICP-MS to follow the loss of lanthanide ions into the aqueous medium as a function of time. Experiments were carried out on AA089 microsphere sample that was

prepared by 2-DisP and has 0.3 wt% LnCl3/sty (Ln = Tm, Ho and Tb, see Table 3-2). An aliquot of AA089 microspheres was first washed by three cycles of centrifugation- redispersion in water. In the last cycle, particles were split into three portions and redispersed

87 Microspheres by Two Stage Dispersion Polymerization

in three different buffer solutions, namely 200 mM sodium carbonate/bicarbonate (pH 10.6), 10 mM ammonium acetate (pH 7) and 50 mM sodium acetate (pH 3), to make the solids content of 0.5 % and kept stirring continuously. Aliquots (100 µL) were taken at different time intervals over a three-week period. Each sample was mixed with deionized water (900 µL) and spun down by centrifugation (5000 rpm, 30 min) immediately to separate the particles from the supernatant. At the end, the supernatant was collected and diluted 10 times

with HNO3 aqueous solution (3.0 %) and then analyzed by ICP-MS. Figure 3-10, at pH 3, there is essentially no detectable leakage of Tm3+ in the sample over the 3 weeks of the experiment. At pH 7, the particles leak a small amount of Tm3+ (less than 3.0 ppb equivalent to 0.06 %) when placed in buffer, but the amount present in the water does not increase over time. Leakage of Tm3+ ions is more significant at high pH, and again does not increase over time. Note that the time scale in Figure 3-10 is logarithmic, and the times sampled are very long. After 3 weeks at pH 10.6, sample AA089 had lost only 0.1% of its Tm content. Thus we conclude that leakage of embedded Ln ions into the aqueous medium is unlikely to be a source of problems in using or functionalizing these particles for bioassays.

3.4 Covalent Attachment of Proteins to the Surface of the Microspheres

The major objective of synthesizing these metal encoded microspheres is to use them in sandwich-type immunoassays. As a first step in this direction, we attempted proof-of- principle experiments involving covalent attachment of proteins to the surface of these carboxylated microspheres. These experiments were carried out by Dr Olga Ornatsky. Our experimental design is depicted in Figure 3-11. In this example, the microspheres are encoded with 151Eu and 153Eu ions. These microspheres are configured to capture antigens of interest with antibodies immobilized on the particle (here mouse IgG). The analyte (anti- mouse IgG) is labeled with a metal-chelating polymer (MCP) to act as a reporter tag. In the example, the MCP carries multiple copies of Pr3+ ions. Analysis of a successful binding event by mass cytometry would detect the 151Eu and 153Eu characteristic of the PS microsphere as well as the 141Pr signal from the analyte. In our experiments, we use a MCP closely related to

88 Microspheres by Two Stage Dispersion Polymerization

that described by Lou et. al. [25] with different ligands as pendant groups (see X4 in Figure 3-11. Identical chemistry was used for attachment of the polymer to the antibody. An aliquot of microsphere sample AA069 was divided into three portions. The first portion was kept as untreated microspheres without any activation by EDC nor any modification with bovine serum albumin (BSA) or mouse IgG (labeled AA069). The other two portions were activated by EDC and then one portion was treated with BSA in PBS buffer as a control for non-specific reporter tag binding (labeled AA069-BSA). The third portion was treated with mouse IgG after EDC activation (labeled AA069-IgG). These later two portions (AA069-BSA and AA069-IgG) were mixed gently, allowed to react on a shaker for 2 h at 23 °C and then washed. After blocking the microspheres with 0.5% BSA/PBS buffer for 1 h, these two portions (AA069-BSA and AA069-IgG) and untreated AA069 (total 3 samples) were all incubated with anti-mouse-IgG-X4-Pr (100 µg/mL). After several washes, the samples were analyzed by mass cytometry. Between 30,000 and 60,000 microspheres were analyzed for each sample. Figure 3-12 compares the 141Pr intensity signals for the three samples (AA069, AA069- BSA and AA069-IgG). The 141Pr signals were weak and indistinguishable for all three cases. This result indicates unsuccessful (or insufficient) attachment of IgG molecules to the surface of the microspheres. The difficulty in attaching IgG molecules to –COOH groups at the surface of the PS microspheres can be attributed to the presence of the PVP corona that surrounds the microsphere. It seems that the PVP corona hinders the access of the biomolecules to the -COOH groups at the surface of the particles.

89 Microspheres by Two Stage Dispersion Polymerization

Figure 3-11. Schematic representation of antigen capture and detection using metal-encoded microspheres and FC-MS. Carboxylated 151Eu and 153Eu-encoded PS microspheres were conjugated to a mouse IgG using carbodiimide chemistry. Microspheres were then washed and incubated with anti-mouse antigen that is labeled with a Pr-containing polymer tag (anti- mouse-IgG-Pr) to identify the presence of captured antigen on the particle. Stringently washed microspheres were analyzed for concomitant signals of 151Eu and 153Eu, and 141Pr as an indication of a successful immunoreactions. At the right, we present the chemical structure of the polymer before attachment to the antibody. X4 was reacted with a bismaleimide coupling agent and then covalently attached to the antibody via reaction with –SH groups produced by selective reduction of a disulfide bond in the hinge region of the antibody. Details are given in Ref. 21. Each polymer carries ca. 30 Pr3+ ions.

90 Microspheres by Two Stage Dispersion Polymerization

Intensity 8

Pr 141 4

0 AA069 AA069 AA069

BSA IgG

141 Figure 3-12. Mass cytometry analysis ( Pr intensities) of the interaction of anti-mouse-IgG-X4- Pr with bare AA069 microparticles and with samples of these particles subjected to conjugation conditions with BSA and with mouse IgG.

91 Microspheres by Two Stage Dispersion Polymerization

3.5 Summary

In this chapter, I describe the synthesis of a series of lanthanide-containing polystyrene (PS) microspheres with a very narrow size distribution. The microspheres were designed for highly multiplexed bioassays based upon a mass cytometer comprising a time-of-flight (TOF) inductively coupled plasma mass spectrometer. This instrument carries out microsphere-by-microsphere analysis at a rate of about 1000 microspheres per second. Upon entry into the plasma torch, individual microspheres are atomized in an Ar plasma, ionized, and then analyzed by TOF mass spectrometry. Approximately 10 to 30 mass spectra are taken during the transit time of a particle through the plasma. The PS microsphere samples reported here were synthesized by dispersion polymerization of styrene in ethanol. They contain covalently grafted polyvinylpyrrolidone (PVP) chains at the surface to provide colloidal stability, and carboxylic acid groups for the attachment of biomolecules. They also contain up to 108 Ln ions per particle. These ions were introduced at different levels of concentration to meet the needs of an enumeration encoding formalism. For enumeration encoding with metal atoms or isotopes, the variability (VR) of the encoding depends upon the number of different clearly distinguishable levels of concentration (K) of N different elements or isotopes that can be introduced synthetically and detected by mass cytometry. N The magnitude of the variability is given by the expression VR = K – 1. The data in Figure 3-6 show that four levels of finite concentration, varying over three orders of magnitude, for four different elements (La, Tb, Ho, Tm) can be distinguished and analyzed quantitatively. Since encoding encompasses zero concentration of each element, K = 5 and N = 4 for this sample set, and V = 624. Mass cytometry is capable of analyzing 20 or more lanthanide ions or discrete isotopes simultaneously. Thus extending this methodology to N = 10 and K = 5 should be straight forward (V > 106). Values of K are limited by our ability to control the microsphere-to-

microsphere variation of Ln ions. We obtained values of CVLn of about 20% to 25% for particles synthesized by two-stage dispersion polymerization. Carboxylated PS microspheres

92 Microspheres by Two Stage Dispersion Polymerization

could not be functionalized with biomolecules using EDC chemistry. This behavior can be attributed to the limited access to the microsphere functional surface.

References

1. Abdelrahman AI, Dai S, Thickett SC, Ornatsky O, Bandura D, Baranov V, Winnik MA: Lanthanide-containing polymer microspheres by multiple-stage dispersion polymerization for highly multiplexed bioassays. J Am Chem Soc 2009, 131(42):15276-15283. 2. Horák D: Uniform polymer beads of micrometer size. Acta Polym 1996, 47(1):20- 28. 3. Yuan HG, Kalfas G, Ray WH: Suspension polymerization. J Macromol Sci, Rev Macromol Chem Phys 1991, C31(2-3):215-299. 4. Shoaf GL, Poehlein GW: Kinetics of emulsion copolymerization with acrylic acids. J Appl Polym Sci 1991, 42(5):1213-1237. 5. Song Z, Poehlein GW: Kinetics of emulsifier-free emulsion polymerization of styrene. Journal of Polymer Science, Part A: Polymer Chemistry 1990, 28(9):2359- 2392. 6. Arshady R: Suspension, emulsion, and dispersion polymerization: A methodological survey. Colloid & Polymer Science 1992, 270(8):717-732. 7. Bradford EB, Vanderhoff JW: Electron microscopy of monodisperse latexes. J Appl Phys 1955, 26(7):864-871. 8. Ugelstad J, Mórk PC, Kaggerud KH, Ellingsen T, Berge A: Swelling of oligomer- polymer particles. New methods of preparation. Adv Colloid Interface Sci 1980, 13(1-2):101-140. 9. Thickett SC, Abdelrahman AI, Ornatsky O, Bandura D, Baranov V, Winnik MA: Bio-functional, lanthanide-labeled polymer particles by seeded emulsion polymerization and their characterization by novel ICP-MS detection. J Anal At Spectrom 2010, 25(3):269-281. 10. Odian G: Principles of Polymerization third edition edn: John Wiley and Sons; 1991. 11. Kawaguchi S, Ito K: Dispersion polymerization. Adv Polym Sci 2005, 175:299-328. 12. Kim JW, Kim BS, Suh KD: Monodisperse micron-sized cross-linked polystyrene particles. VI. Understanding of nucleated particle formation and particle growth. Colloid Polym Sci 2000, 278(6):591-594. 13. Yasuda M, Seki H, Yokoyama H, Ogino H, Ishimi K, Ishikawa H: Simulation of a particle formation stage in the dispersion polymerization of styrene. Macromolecules 2001, 34(10):3261-3270.

93 Microspheres by Two Stage Dispersion Polymerization

14. Song JS, Tronc F, Winnik MA: Two-stage dispersion polymerization toward monodisperse, controlled micrometer-sized copolymer particles. J Am Chem Soc 2004, 126(21):6562-6563. 15. Song JS, Chagal L, Winnik MA: Monodisperse micrometer-size carboxyl- functionalized polystyrene particles obtained by two-stage dispersion polymerization. Macromolecules 2006, 39(17):5729-5737. 16. Song JS, Tronc F, Winnik MA: Monodisperse, controlled micron-size dye-labeled polystyrene particles by two-stage dispersion polymerization. Polymer 2006, 47(3):817-825. 17. Song JS, Winnik MA: Cross-linked, monodisperse, micron-sized polystyrene particles by two-stage dispersion polymerization. Macromolecules 2005, 38(20):8300-8307. 18. Kagan HB: Introduction: Frontiers in lanthanide chemistry. Chem Rev 2002, 102(6):1805-1806. 19. Ober CK, Lok KP, Hair ML: Monodispersed, micron-sized polystyrene particles by dispersion polymerization. Journal of polymer science Polymer letters edition 1985, 23(2):103-108. 20. Tseng CM, Lu YY, El-Aasser MS, Vanderhoff JW: Uniform polymer particles by dispersion polymerization in alcohol. Journal of Polymer Science, Part A: Polymer Chemistry 1986, 24(11):2995-3007. 21. In M, Gerardin C, Lambard J, Sanchez C: Transition metal based hybrid organic- inorganic copolymers. J Sol-Gel Sci Technol 1995, 5(2):101-114. 22. Pich A, Zhang F, Shen L, Berger S, Ornatsky O, Baranov V, Winnik MA: Biocompatible hybrid nanogels. Small 2008, 4(12):2171-2175. 23. Sanchez C, In M: Molecular design of alkoxide precursors for the synthesis of hybrid organic-inorganic gels. J Non-Cryst Solids 1992, 147-148(C):1-12. 24. Wang J, Shi TJ, Jiang XC: Synthesis and characterization of core-shell ZrO2/PAAEM/PS nanoparticles. Nanoscale Research Letters 2009, 4(3):240-246. 25. Lou X, Zhang G, Herrera I, Kinach R, Ornatsky O, Baranov V, Nitz M, Winnik MA: Polymer-based elemental tags for sensitive bioassays. Angewandte Chemie - International Edition 2007, 46(32):6111-6114.

94 Microspheres by Three Stage Dispersion Polymerization

4 Microspheres by Three Stage Dispersion Polymerization In this chapter, I describe the synthesis of lanthanide-containing microspheres by three-stage dispersion polymerization (3-DisP). This synthetic methodology was developed in an attempt to prepare microspheres that meet the size, metal-content and, most importantly, the surface functionality requirements to be used as a platform for mass cytometry-based bioassays and to avoid the shortcomings of the microspheres synthesized by two-stage dispersion polymerization (2-DisP).

4.1 Introduction

I have explained in Chapter 3 that the carboxyl group is our choice to functionalize the surface of microspheres because this group is one of the most commonly used functional groups in biomedical and biodiagnostic applications. For instance, the carboxyl group can be used in conjugating biomolecules to the surface of microspheres employing standard EDC (1-ethyl-3-(3- dimethylaminopropyl) carbodiimide)) coupling conditions. [1, 2] To be able to synthesize microspheres with a –COOH containing monomer like acrylic acid (AA) without distorting the particle size distribution, I used the two stage dispersion polymerization method (2-DisP) developed in our group by Song et. al. [3], in which the addition of AA to the microsphere synthesis reaction was delayed until the nucleation stage was over. Prior to the work of Song et al., the synthesis of crosslinked microspheres was considered to be one of the biggest challenges for dispersion polymerization. The literature has many references [4-13] that demonstrate that dispersion polymerization failed when cross-linking agents were included in the reaction. The most common result was flocculation or coagulation of the final product. In the few cases where the reaction appeared to succeed, one finds odd-shaped particles and a broad size distribution. Tseng et al. [14] reported the consequences of using a cross-linking agent in a dispersion polymerization reaction. In their experiments, they mixed small amounts of divinylbenzene (DVB) with styrene for dispersion polymerization in ethanol. When they added DVB at the concentration of only 0.3 wt % based on the total monomer, they obtained a broad size distribution. Higher levels of DVB concentration led to coagulation of the dispersion. Horák [15] investigated dispersion polymerization as a means of synthesizing poly(2- hydroxyethyl methacrylate) (PHEMA) microparticles in toluene/alcohol mixtures for biomedical

95 Microspheres by Three Stage Dispersion Polymerization

applications. Although the particles they obtained had a relatively narrow size distribution, their size distribution was broader than that of particles prepared by dispersion polymerization of styrene in ethanol. The need for cross-linked particles prompted a detailed investigation reaction conditions that led to incorporation of ethylene glycol dimethacrylate (EGDMA) as a crosslinking agent. They found that delaying addition of the EGDMA until the reaction had run for 2 hr led to optimal reaction with minimal perturbation of the particle size distribution [16]. This report predates the work of Song et. al. [17] who independently showed that one could synthesize crosslinked polystyrene microsphere without losing control over the particle size distribution by delaying the addition of the cross-linking agent until the end of nucleation stage, In Chapter 3, I showed how it was possible to synthesize poly(styrene-co-acrylic acid) microspheres with a narrow-size distribution by 2-DisP. Using 2 wt.% acrylic acid, microspheres were loaded with different lanthanide ions. The size and Ln-content of the microspheres were suitable for mass cytometry analysis. However, these particles suffered two main drawbacks. First was their relatively wide lanthanide content distribution (compared to their size distribution). Secondly, and most importantly, these microspheres had a limited ability to be biofunctionalized. There are many ways one can think of to overcome these problems. One potential solution to this problem is to activate the inert PVP layer that covers the microspheres. Alternatively, one can grow a shell of functional polymer on the surface of these particles, to ‘cover up’ the PVP coating and allow bioconjugation to take place by providing appropriate surface functionality. In Chapter 5, I will discuss utilization of partially-modified PVP as a dispersant for dispersion polymerization. In addition, seeded emulsion polymerization will be investigated as a method to functionalize the surface of microspheres synthesized by 2- Disp. In this chapter, I propose a methodology to overcome the Ln distribution and bioconjugation issues found in the microspheres synthesized by 2-DisP. I explored the effect of adding more acrylic acid and a crosslinking agent, in a third stage, during the synthesis of the microsphere by dispersion polymerization. I call this process three-stage dispersion polymerization (3-DisP). In 3-DisP, the crosslinking agent was expected to reduce the release of Ln ions from the growing microspheres into the continuous medium; hence it would help the microspheres to maintain their Ln content and consequently improve the lanthanide content distribution (CVLn). I chose ethylene glycol dimethacrylate (EGDMA) as the crosslinker for its suitability to crosslink

96 Microspheres by Three Stage Dispersion Polymerization

styrene-based systems. At the same time, an extra amount of acrylic acid (AA) was added to increase the number of carboxylic groups on the surface available for bioconjugation. Moreover, higher AA concentration was expected to provide more binding sites within the microspheres to enhance Ln ions incorporation efficiency.

4.2 Three Stage Dispersion Polymerization

As described earlier, it was difficult to attach representative biomolecules to the surface of particles prepared by 2-DisP. As a step to overcome this problem, and also in an attempt to obtain particles with a narrower range of Ln ions per particle, I examined the idea of adding a third stage to the dispersion polymerization reaction. The 3-DisP reaction was carried out as for 2-DisP (described in Chapter 3), but at approximately 60% styrene conversion, additional acrylic acid (2.0 wt.% / styrene) and a small amount of EGDMA cross-linking agent (2.0 wt.% / styrene) were added to the reaction as a warm solution in ethanol. Experimental details of 3-DisP were presented in Chapter 2.

Table 4-1. The recipe for the synthesis of AA105 particle sample by 3-stage dispersion polymerization (3-DisP) of styrene with PVP55 as a dispersant in ethanol Materials (grams added) 1st stage 2nd stage a 3rd stage b Styrene 6.25 -- -- PVP55 1.0 -- -- TX305 0.35 -- -- AMBN 0.25 -- -- Ethanol 18.75 10.0 10.0 Acrylic Acid -- 0.125 0.125

LaCl3.6H20 -- 0.0063 --

TmCl3.6H20 -- 0.0063 -- EGDMA -- -- 0.125

a Added 1.0 hr after the 1st stage. b Added 8.0 hr after the 2nd stage.

97 Microspheres by Three Stage Dispersion Polymerization

The recipe is presented in Table 4-1 and the characteristics of sample AA105 prepared in this way are listed as the first entry in Table 4-2. A scanning electron microscope image of these particles is presented as Figure 4-1. The AA105 particles prepared by 3-DisP had an overall

particle size of 2.2 µm, with a very narrow distribution of sizes (CVd = 1.1 %). Because of the acrylic acid added in the third stage of the reaction, the number of titratable -COOH groups per microsphere increased by nearly a factor of 2 (4.3 × 108-COOH groups/particle) compared to microspheres prepared by 2-DisP. In addition, the Ln incorporation efficiency is somewhat improved (92% compared to ca. 88% in 2-DisP, measured by conventional ICP-MS). Mass cytometry measurements were carried out on sample AA105 to examine the lanthanide content of AA105. Figure 4.2A shows a 169Tm/139La bi-variant plot for the AA105 sample. The mass cytometry signals were gated as described in Chapter 3 to exclude the high intensity region (ca. 8 % of the whole population). The dotted oval in Figure 4.2A shows the data retained after gating. Figure 4-2B and C show the gated intensity distributions for 139La and 169Tm, respectively. Obviously, there was a substantial improvement in the variation in lanthanide content per particle compared to particles synthesized by 2-DisP. Here, for sample

AA105, CVTm and CVLa are 11% and 13 %, respectively. Using equation 3-3, I calculated the Tm content of AA105 and found 3.2 x 107 ions per particle.

Figure 4.1. SEM image of unwashed sample AA105 (unwashed) microspheres synthesized in the presence of TmCl3 (0.1 wt%/styrene) and LaCl3 (0.1 wt%/styrene). d = 2.2 μm, CVd = 1.1%.

Figure 4.2. Mass cytometry measurements for AA105 169Tm/139La bi-variant plot (A) and gated intensity distribution for 139La (B) and 169Tm (C). Gating applied was highlighted by the red oval in (A). These gated signals (B and C) were characterized by CVLa = 13 % and CVTm = 11%.

98 Microspheres by Three Stage Dispersion Polymerization

Table 4.2: Size and metal content results for Samples AA105 and AA122 – AA129 synthesized by 3- DisP.

d a CV b V c Tm d Tm / V e CV f Tb d Tb / V e CV f d Tm Tb (µm) (%) (µm3) (x106) (x106) (%) (x106) (x106) (%) AA105 2.2 1.1 5.7 30.2 5.3 11 ------AA122 2.5 11 8.2 30.1 3.7 34 39.5 4.8 33 AA123 2.1 4 4.9 17.7 3.7 23 24.1 5.0 21 AA124 2.7 11 10.3 39.3 3.8 35 49.3 4.8 33 AA125 2.2 5 5.6 16.7 3.0 26 22.9 4.1 22 AA126 2.6 3 9.2 29.7 3.2 19 36.8 4.0 19 AA127 2.2 31 5.6 27.3 4.9 58 34.0 6.1 51 AA128 2.3 2 6.4 29.7 4.7 13 37.6 5.9 12 AA129 2 6 4.2 21.0 5.0 29 27.0 6.4 27

a Average particle diameter from SEM image b Particle size distribution calculated based on equation 3-1:

1 1 n 2 CV = d ∑i=1(Di −Dav ) Dav n −1 c Average particle diameter d Average number of atoms per particle measured by mass cytometry e Number of atoms per unit volume f Lanthanide content distribution calculated based on equation 3-2:

1 n 2 ∑ = − n −1 i 1(Lni Lnav ) CVLn = Lnav

99 Microspheres by Three Stage Dispersion Polymerization

4.3 Factorial design

The size and metal content characteristics of particles prepared by 3-DisP, explained in the previous section, were superior to those prepared by 2-DisP. In this section, I explore the possibility of further improving the quality of the particles prepared by 3-DisP through changing the 3rd stage conditions (concentration and the time of addition of the reactants). Thinking about the addition of AA and EGDMA in a third stage, I had three main synthetic parameters (factors) that I wished to examine for their effect(s) on the synthesized microspheres: 1- The amount of the crosslinking agent, EGDMA. 2- The amount of AA. 3- The timing of the addition of EGDMA and AA to the polymerization reaction. Each combination of these factors will be considered as an experiment. To determine the extent of success of each experiment, one can use different kinds of measurement(s) or response(s). For my experiments, the responses can be one or a combination of these measurements: 1- Lanthanide content and lanthanide content distribution 2- Size and size distribution 3- Extent of bioconjugation In experiments that involve several factors, in which it is necessary to study the joint effect of these factors on a response or multiple responses, factorial designs are widely used [18]. A complete factorial design means that, in the experiments, all possible combinations of the levels of the factors are investigated [18]. Applying a complete set of factorial design experiments for optimizing the amount of reagents to be added in the third stage and the time of their addition is impossible because of the unlimited number of levels I have for each factor. To simplify the problem, I decided to use the simplest factorial design method, the 2k factorial design, where k is the number of factors in the experiments (k = 3 in our design), and the number 2 means that we will only try two levels for each factor, for example high and low concentration levels. Thus, the size of my 2k factorial design is 23 = 8 experiments. Table 4-3 shows a set of randomly ordered sequence of experiments.

100 Microspheres by Three Stage Dispersion Polymerization

4.2.1 Design of the Factorial Experiments

To perform factorial experiments examining the extreme conditions for the 3rd stage, I had to determine the high and low levels for each of the synthetic factor: EGDMA concentration, AA concentration and time of the 3rd stage addition. The only condition for choosing these levels was that it should result in a stable and coagulum free latex. For EGDMA, I arbitrarily chose 0.5 wt.% / styrene and 4.0 wt.% / styrene as the low and high levels of EGDMA, respectively. Regarding the second factor, the AA low concentration level in the third stage was set to be 0.5 wt.% / styrene and the high concentration level to be 2.0 wt.% / styrene. Finally, the time interval between the second and third stage additions was varied between 1.0 and 8.0 hours as low and high time levels, respectively. These values were chosen to bracket these of sample AA105 and tested to ensure that they yielded stable microsphere dispersions.

Table 4-3: 23 factorial design of the third stage in the multiple stage dispersion polymerizations. All other conditions for the dispersion polymerization reaction (like 1st and 2nd stage ingredients and reaction temperature) were kept constant.

a b c Experiment # Sample # EGDMA Acrylic Acid Time 1 AA122 High Low Low

2 AA123 High High Low

3 AA124 Low High Low

4 AA125 High High High

5 AA126 High Low High

6 AA127 Low Low High

7 AA128 Low High High

8 AA129 Low Low Low

a Amount of ethylene glycol dimethacrylate added in the 3rd stage: Low is 0.5 wt.% / styrene and High is 4.0 wt.% / styrene b Amount of acrylic acid added in the 3rd stage: Low is 0.5 wt.% / styrene and High is 2.0 wt.% / styrene c The time between 2nd and 3rd stage: Low is 1.0 hr after the second stage and High 8.0 hr

101 Microspheres by Three Stage Dispersion Polymerization

Based on this factorial design, I synthesized a series of eight microspheres samples (AA122 – AA129). This factorial design experiments aimed at the understanding of how the extremes of the 3rd stage conditions affect the size, size distribution, metal content and metal content distribution of the produced microspheres. These microsphere samples all had the same initial polymerization mixture and the same amount of acrylic acid (2.0 wt.% / styrene) and lanthanide

salts (TbCl3.6H20 = TmCl3.6H20 = 0.1 wt.% / styrene) in the second stage. The differences in the synthesis of samples AA122-AA129 were only in the amount of the reactants as well as the timing of the 3rd stage. See Table 4-4 for the ingredients used in the synthesis of AA122 – AA129 microspheres.

4.2.2 Factorial Experiments: Results and Discussions

All samples resulted in stable and coagulum-free dispersions with a gravimetric conversion of 95 – 99 %. SEM images for each sample (“as synthesized” without washing or purification) along with their size distribution histogram are given in Figures 4.3. – 4.10. Table 4.2 shows the size, and size distribution measurements for the microspheres synthesized by 3-DisP. The smallest average particle diameter obtained was 2.0 µm and the highest was 2.7 µm. Microsphere Sample AA127 was synthesized in presence of low AA and low EGDMA concentration levels (each at 0.5 wt.% / styrene). In this sample, the third stage was added 8.0 hours after the second stage. AA127 microsphere sample had a mean diameter of 2.2 µm. As seen in by Figure 4.8.A, this sample had a substantial number of submicron particles (300-500 nm) that lead to unexpected broadening in the particle size distribution (CVD = 31%). This big population of submicron particles was never noticed with samples synthesized by 2-DisP and other samples prepared by 3-DisP. I washed Sample AA127 as a trial to get rid of the small particles. The washing was done by spinning the sample dispersion down using slow–speed centrifugation (3000 rpm for 20 min) then the supernatant was removed and the pelleted microspheres were redispersed in water. After 3 cycles of washing, AA127 sample showed only a very small population of submicron particles (Figure 4.8.C). Consequently, the mean particles

diameter increased (d = 2.4 µm) and the size distribution was significantly decreased (CVd = 2.6 %). As there is no clear explanation for why this submicron particles’ population was formed at the first place, this experiment was excluded from the following discussions.

102 Microspheres by Three Stage Dispersion Polymerization

Table 4-4. An example of a factorial design experiment: the recipe for the synthesis of AA122 by 3- stage dispersion polymerization (3-DisP) of styrene with PVP55 as a dispersant in ethanol Materials (grams added) 1st stage a 2nd stage a 3rd stage b Styrene 6.25 -- -- PVP55 1.0 -- -- TX305 0.35 -- -- AMBN 0.25 -- -- Ethanol 18.75 10.0 10.0 Acrylic Acid -- 0.125 0.125 c

TbCl3.6H20 -- 0.0063 --

TmCl3.6H20 -- 0.0063 -- EGDMA -- -- 0.031 c a The same concentrations were used for all the other 7 samples (AA123 – AA129) b Added 1.0 hr after the 2nd stage. See Table 4-1 for the time of addition used in the other 7 samples (AA123 – AA129) c See Table 4-1 for details of the amount of AA and EGDMA used for the synthesis of the other 7 samples (AA123 – AA129)

103 Microspheres by Three Stage Dispersion Polymerization

Figure 4.3.: SEM image and size distribution histogram of unwashed AA122 synthesized with EGDMA (4.0 wt.% / styrene) and AA (0.5 wt.% / styrene) added to the dispersion polymerization reaction 1.0 h after the 2nd stage. Note the raspberry texture for this sample which has high concentration of EGDMA.

Figure 4.4.: SEM image and size distribution histogram of unwashed AA123 synthesized with EGDMA (4.0 wt.% / styrene) and AA (2.0 wt.% / styrene) added to the dispersion polymerization reaction 1.0 h after the 2nd stage. Note the raspberry texture for this sample which has high concentration of EGDMA.

Figure 4.5.: SEM image and size distribution histogram of unwashed AA124 synthesized with EGDMA (0.5 wt.% / styrene) and AA (2.0 wt.% / styrene) added to the dispersion polymerization reaction 1.0 h after the 2nd stage. Note the presence of a few small particles, suggesting some secondary nucleation in the sample.

104 Microspheres by Three Stage Dispersion Polymerization

Figure 4.6.: SEM image and size distribution histogram of unwashed AA125 synthesized with EGDMA (4.0 wt.% / styrene) and AA (2.0 wt.% / styrene) added to the dispersion polymerization reaction 8.0 h after the 2nd stage. Note the raspberry texture for this sample which has high concentration of EGDMA.

Figure 4.7.: SEM image and size distribution histogram of unwashed AA126 synthesized with EGDMA (4.0 wt.% / styrene) and AA (0.5 wt.% / styrene) added to the dispersion polymerization reaction 8.0 h after the 2nd stage. Note the raspberry texture for this sample which has high concentration of EGDMA.

Figure 4.8. SEM image and size distribution histogram of AA127 synthesized with EGDMA (0.5 wt.% / styrene) and AA (0.5 wt.% / styrene) added to the dispersion polymerization reaction 8.0 h after the 2nd stage before washing (A) and after three cycles of washing (B). Note the presence of many smaller particles in (A) that makes this sample excluded from the discussion.

105 Microspheres by Three Stage Dispersion Polymerization

Figure 4.9.: SEM image and size distribution histogram of unwashed AA128 synthesized with EGDMA (0.5 wt.% / styrene) and AA (2.0 wt.% / styrene) added to the dispersion polymerization nd reaction 8.0 h after the 2 stage. Microspheres are monodisperse in size.

Figure 4.10.: SEM image and size distribution histogram of unwashed AA129 synthesized with EGDMA (0.5 wt.% / styrene) and AA (0.5 wt.% / styrene) added to the dispersion polymerization reaction 1.0 h after the 2nd stage. Note the presence of some ca. 1.2 µm particles.

Figure 4.11. Tm content and volume of microsphere samples AA122 – AA129 synthesized by 3- DisP (A) and Tm content versus volume of microsphere same samples (B).

106 Microspheres by Three Stage Dispersion Polymerization

Figure 4.11.A. illustrates the relation between the Tm content and the microsphere volume of each sample. Interestingly, regardless of the synthetic factors employed (AA and EGDMA concentrations used and the time of their addition), the Tm content of the microspheres was correlated to their volumes; which means that the average Tm content of the microspheres increased with the increase of the average diameter of the microsphere sample. This dependence is only clear for large volume changes (for example, AA122 vs. AA123 and AA123 vs. AA124). On the other hand, the dependence of the Tm content of the microspheres on their volume was negligible in the case of small volume changes (e.g., AA123 vs. AA125 and AA125 vs. AA129),. To further understand the relation between the Tm content of the microspheres and their volume, I plotted Tm content (y-axis) against the microsphere volume (x-axis) in Figure 4.9.B. Although the figure shows an overall increase of the Tm content for larger microspheres, this general trend has many exceptions (at least 3, AA123, AA125 and AA126). In other words, this correlation between the Tm content of the microspheres and the average volume of the microspheres is not a direct relation. I conclude that the particle volume is not the only factor that affects the Tm content of the microspheres (see the Tm/V column in Table 4.3). I found that the Tb content of these samples (AA122-AA129) show similar behavior, when compared with the volume of the microsphere. The observations mentioned above led to three main conclusions. First, due to the obvious effect of the microsphere size on the lanthanide content of the microspheres, I “normalized” the lanthanide content to account for the microsphere volume change. Normalized content refers to the average number of lanthanide atoms per particle divided by its number average volume (in µm3). The second conclusion is that there is a need to study the reproducibility of the microsphere synthesis, i.e. if the same recipe were repeated for more than once under exactly the same condition, would it produce similar size of microspheres? It is worth recalling that the nucleation stage in dispersion polymerization is very delicate. Even for simple polymerization of styrene in ethanol, successive reactions can give different-size particles, each with a narrow size distribution [9, 10, 19, 20]. This experiment will be examined again in Chapter 7 as part of a discussion on the analytical aspects of particle synthesis by dispersion polymerization. Finally, although the microsphere size seems to be very effective parameter on the lanthanide content of the microspheres, it is not the only factor that affects the lanthanide content; rather, the other

107 Microspheres by Three Stage Dispersion Polymerization

synthetic factors, such as AA and EGDMA concentration, may also have their effects on the lanthanide content of the microspheres. Figure 4.12A and C show three-dimensional plots for the 159Tb and 169Tm content of samples AA122 – AA129, determined by mass cytometry, as function of EGDMA and AA concentrations. Mass cytometry measurements reveal that the average number of 159Tb atoms can be as low as 1.7 x 107 Tb atoms per particles and as high as 3.9 x 107 Tb atoms per particles. A similar range was found for the number of 169Tm atoms per particles (2.3 x 107 – 4.9 x 107). Figure 4.12B and D show three-dimensional plots for the “normalized” content for these samples. As explained earlier, the normalized content was obtained by dividing the average number of lanthanide ions per particle over its average volume. The microsphere sample AA129 had the highest normalized Tb content (6.4 x 106 Tb atoms per µm3) while sample AA126 had the lowest normalized Tb content (4.0 x 106 Tb atoms per µm3). See Table 4.2. Although there is no clear trend of particle size, particle size distribution and lanthanide content responses to the change of the synthetic factors, I will mention some of the observations obtained by analyzing the SEM images and mass cytometry measurements of these eight samples: 1. The most significant observation in the SEM images was that when a high concentration level of EGDMA (4.0 wt.% / styrene) was used, particles with a lumpy surface akin to a ‘raspberry-like’ structure were almost always obtained (Figure 4.3., 4.4., 4.6. and 4.7.). This is in contrast to particles made with low concentration level of EGDMA (0.5 wt.% / styrene) which have smooth surfaces similar to those obtained when I synthesized lanthanide-containing particles by 2-DisP. 2. All samples had narrower lanthanide content distribution than that of particles synthesized by 2-DisP, even for the microsphere samples whose size distribution were wider than those of the microspheres synthesized with 2-DisP (compare the result in Table 4-2 with results in Table 3-4 obtained for microspheres prepared by 2-DisP). 3. The narrowest particle size distributions were obtained for two samples (AA128 and

AA126 have CVd of 2 % and 3 %, respectively). Both had their third stage added 8.0 hours after the second stage. In contrast, the broadest size distributions (excluding

AA127) were obtained for samples AA122 and AA124 (CVd of 11 % for both). Both

108 Microspheres by Three Stage Dispersion Polymerization

samples had their third stages added only 1.0 h after the second stage. The broadening in the size distribution when additional AA and EGDMA were added quickly after the 2nd stage can be ascribed to the insufficient time given to the 2nd stage content (especially acrylic acid) to be fully consumed. This result suggests that one should employ longer waiting time between the second and the 3rd stage. 4. No matter when the 3rd stage was added to the polymerization mixture, using high concentrations of AA (2.0 wt.% / styrene) and EGDMA (4.0 wt.% / styrene) yielded the lowest number of metal per particle as depicted by Figures 4.12A and C. 5. Figures 4.12A and C show that AA124, AA122 and AA128 samples had the highest number of metal atoms per particles. Unfortunately there is no pattern that can be used to link these three samples together. 6. As revealed by Figures 4.12B and D, regardless the time of addition of the 3rd stage and the AA concentration used in the preparation of the microspheres by 3-DisP, the normalized Tb and Tm content of the microspheres were the highest when the low concentration of EGDMA (0.5 wt.% / styrene) was used. In contrast, when the high concentration of EGDMA (4.0 wt.% / styrene) was used, the normalized Tb and Tm content of the microspheres were the lowest. 7. When the 3rd stage was added 8.0 hr after the 2nd stage, the normalized Tb and Tm content of the microspheres appeared to be inversely proportional to the EGDMA concentration and were no affected by the AA concentration. On the other hand, when the 3rd stage was introduced 1.0 hr after the 2nd stage, the normalized Tb and Tm content of the microspheres were not affected by either the AA concentration or the EGDMA concentration. 8. Figures 4.12B and D show that AA129 and AA128 samples were found to have the highest metal content per unit volume in all the samples prepared by 3-DisP.

None of these samples prepared by the factorial design experiments had the beneficial characteristics of sample AA105, as revealed by the SEM and mass cytometry results. Sample AA128 was the sample with the narrowest size and lanthanide content distributions as well as the highest lanthanide content among the AA122-AA129 samples. Actually, the 3rd stage conditions

109 Microspheres by Three Stage Dispersion Polymerization of AA128 sample were very similar to AA105 except for EGDMA concentration. I concluded that no improvement can be attained by changing the 3rd stage conditions of AA105 and, consequently, this sample was selected be tried for further analysis like lanthanide ion leakage and bioconjugation experiments.

Figure 4.12. Average number of 159Tb atoms (A), normalized number of 159Tb atoms per unit volume per particle (B), average number of 169Tm atoms per particle (C) and normalized number of 169Tm atoms per unit volume for different microsphere samples (AA122 – AA129) prepared in presence of AA (0.5 – 2.0 wt.% / styrene) and EGDMA (0.5 – 4.0 wt.% / styrene). AA and EGDMA were added in the 3rd stage, 1.0 hrs (red circles) or 8.0 hrs (white circles) after the second stage.

110 Microspheres by Three Stage Dispersion Polymerization

4.3 Testing Ion Leakage

Before turning my attention to the protein functionalization of the particles, I needed to assess the stability of the particles prepared by 3-DisP toward leakage of the Ln ions into the aqueous medium, particularly under the experimental conditions associated with attaching biomolecules to the particles. I used traditional inductively coupled plasma-mass spectroscopy (ICP-MS) to follow the loss of 169Tm ion into the aqueous medium as a function of time. Experiments were carried out on two samples, AA089 prepared by 2-DisP (Chapter 3), and sample AA105 prepared by 3-DisP. The results are presented in Figure 4.13 for particles suspended in three aqueous solutions buffered at pH 3.0 (50 mM sodium acetate solution), 7.0 (10 mM ammonium acetate solution) and 10.6 (200mM sodium carbonate/bicarbonate solution). At pH 3, there was essentially no detectable leakage of Tm3+ in either sample for the 3 weeks of this experiment. At pH 7, the particles prepared by 2-DisP leaked a small amount of Tm3+ when placed in buffer, but the amount present in the water did not increase over time (middle curve in Figure 4.13.A). This loss of Tm3+ was significantly reduced for the particles prepared by 3-DisP. Leakage of Tm3+ ions was more significant at high pH, and again was substantially reduced for

Figure 4.13. Tm ion release into the aqueous phase from colloidal suspensions of two Tm- containing PS microsphere samples in three different buffer solutions. (A) AA089, synthesized by 2- stage DisP and (B) AA105 synthesized by 3-stage DisP. Both microsphere samples contain 260 ppm Tm ion (w/w styrene). The pH 10.6 is a 200 mM sodium carbonate/bicarbonate buffer solution, pH 7.0 is a 10 mM ammonium acetate buffer solution and pH 3.0 is a 50 mM sodium acetate buffer solution.

111 Microspheres by Three Stage Dispersion Polymerization

sample AA105. Note that the time scale in Figure 4.13 is logarithmic, and the times sampled were very long. After 3 weeks at pH 10.6, sample AA089 had lost only 0.1% of its Tm content, and sample AA105 had lost even less (0.02% of its Tm content). Thus I concluded that leakage of embedded Ln ions into the aqueous medium was unlikely to be a source of problems in using or functionalizing these particles for bioassays.

4.4 Synthesis of Particles with Higher Variability

One of my objectives was to expand the variability of the enumeration encoded library mentioned in Chapter 3 in order to make possible the use of the lanthanide encoded particles in highly multiplexed assays. Particle samples AA134 and AA135 were synthesized with a formulation similar to that of AA105. AA134 sample was prepared in the presence of 9 lanthanide salts while 7 lanthanides were used to synthesize sample AA135. A recipe for the synthesis of AA134 particle sample, by 3-DisP of styrene and acrylic acid in ethanol, is listed in Table 4-5. The following procedure was used for sample AA134: the stabilizer, the co-stabilizer, the initiator, styrene and ethanol were mixed and deoxygenated at room temperature. After 30 min, the polymerization was initiated by elevating the reaction temperature to 70°C. Well after

the nucleation stage was over (1.0 hr from initiation), acrylic acid and a mixture of LnCl3.6H2O (where Ln = La, Nd, Sm, Gd, Eu, Tb, Dy, Ho and Tm each at 0.1 wt. % / styrene) in warm ethanol were added to the polymerization mixture as the second stage. In the third stage, more acrylic acid and EGDMA as a crosslinking agent were added, 8.0 hrs after the second stage. The AA135 particle sample was prepared by the same procedure except that in the second stage, the Sm and the Gd salts were omitted i.e. this reaction had only 7 lanthanides (see Table 4-5). Stable, coagulum-free dispersions were obtained in both experiments after 24 hours of polymerization with monomer conversions of ca. 98 %. SEM was used to determine the particle size and particle size distribution of sample AA134 and AA135 (Figure 4-14 and 4-15). The particles in sample AA134 had an average diameter of 2.0 µm (Figure 4-14), which is similar to that of sample AA105 [21]. AA134 particles had a size

distribution (CVd) of 2.3 %, Figure 4-15. Surprisingly, sample AA135, which was prepared in the presence of only 7 lanthanide salts and had an average diameter of 2.1 µm, had a population of small particles (d = 1.5 – 1.7 µm, Figure 4-15A and C). Even though this small-particle

112 Microspheres by Three Stage Dispersion Polymerization

population represented only ca. 5 % of the sample measured by SEM), it had a significant effect

on the particle size distribution (CVd =6.6 %). I have no explanation for why I obtained this kind of small particle population in this particular synthesis. However, after 3 cycles of washing by sedimentation at 2000 rpm and redispersion in water, I was able to remove almost all of the small

particles. The purified sample was chacterized by d = 2.2 µm and CVd = 3.2 % (Figure 4-15B and D). A comparison between the histograms of particle size distributions of sample AA135 before and after washing is presented in Figure 4.15C and D. Both histograms look very similar except for the disappearance of the small particle population after washing. The metal content for these particles was measured by mass cytometry. To demonstrate the ability of mass cytometry to resolve the presence of multiple Ln ions in a single particle, I present data in Figure 4.16A and B which represent screen captures of signals from samples AA134 (La, Nd, Sm, Gd, Eu, Tb, Dy, Ho, Tm) and AA135 (La, Nd, Eu, Tb, Dy, Ho, Tm). The dense vertical lines in Figure 4.17A, for sample AA134, refer to signals from multiple mass spectra taken during the transit of a single particle through the plasma torch. Figure 4.16A shows signals from four successive particles of sample AA134. Figure 4.16B shows signals from two successive particles of Sample AA135. One can see clear and distinct signals of the different isotopes (like Tb, Ho, and Tm) but Sm and Gd were not detected as they were not included in the synthesis. Figure 4.17A shows the 159Tb gated distribution (88 % of the sample population) in samples AA134 and AA135 as examples for the lanthanides added to both samples. The average Tm atom content of sample AA134 was 2.5 x 106 159Tb atoms per particle and the coefficient of

variation of the Tb content distribution (CVTb) was 27.1 %. Sample AA135 was found to contain 6 159 2.1 x 10 Tb atoms per particle, with CVTb equal to 25.5 %. Bivariate “dot-dot” plots are another way of presenting the mass cytometry results. These bivariate plots show the relationship between two variables (isotopic concentrations) that have been measured on a single particle. Such plots permit us to see at a glance the degree and pattern of the relation between the two isotopes in the sample. Most importantly, using bivariate plots, one can group the populations of the particles that have similar contents of different isotopes. On a bivariate plot, the x- and y-axes can represent the intensities of any two isotopes of interest. Each point on the plot shows the x and y isotopic-content for a single particle. Figure 4.17B shows the two-dimensional projections (a bivariate plot) [22] of some multidimensional data sets

113 Microspheres by Three Stage Dispersion Polymerization obtained as a result of the mass cytometry experiment. Figure 4.17B presents a gated logarithmic 165Ho/169Tm bivariate plot for the AA134 sample. Although the particles show different amounts of Ho and Tm content ranging from ca. 6 x 105 up to 107 for both lanthanides, the vast majority (> 85%) of the particles exhibit a very tight distribution of lanthanide content ca, 2 x 106 and

CVLn of ca. 25%. This behavior in the bivariate plot of AA134 sample reflects the characteristic feature of particles prepared by 3-Disp, which is their relatively lower particle-to-particle variability of the lanthanide content compared to particles prepared by 2-DisP (see Chapter 3). The results illustrated in this section demonstrate the ability to synthesize PS particles loaded with up to 9 lanthanide metals with no significant effect on the particle size and particle size distribution. The metal content of such particles was very well resolved. Combining these results with the ability of baseline resolution for 5 levels of concentrations (Chapter 3), the encoding variability for such particles would increase dramatically, if the enumeration encoding protocol were used. Almost 2 x 106 (i.e. 59 − 1) uniquely encoded particles could be produced. This opens the possibility to resolve an extremely large number of unique biomarkers through the use of such particles as a platform for bioassays.

114 Microspheres by Three Stage Dispersion Polymerization

Table 4.5. The recipe for the synthesis of particle samples AA134 and AA135 by 3-stage dispersion polymerization (3-DisP) of styrene with PVP55 as a dispersant in ethanol. Materials (grams added) 1st stage 2nd stage a 3rd stage b Styrene 6.25 -- -- PVP55 1.0 -- -- TX305 0.35 -- -- AMBN 0.25 -- -- Ethanol 18.75 10.0 10.0 Acrylic Acid -- 0.125 0.125

LaCl3.6H20 -- 0.0032 --

NdCl3.6H20 -- 0.0032 --

SmCl3.6H20 -- 0.0032 -- c GdCl3.6H20 -- 0.0032 -- c EuCl3.6H20 -- 0.0032 --

TbCl3.6H20 -- 0.0032 --

DyCl3.6H20 -- 0.0032 --

HoCl3.6H20 -- 0.0032 --

TmCl3.6H20 -- 0.0032 -- EGDMA -- -- 0.031

a Added 1.0 hr after the initiation. b Added 8.0 hr after the 2nd stage. c Not used in the synthesis of AA135

115 Microspheres by Three Stage Dispersion Polymerization

Figure 4.14. SEM images with different magnifications for unwashed PS particle samples AA134 synthesized in the presence of LnCl3.6H2O (where Ln = La, Nd, Sm, Gd, Eu, Tb, Dy, Ho and Tm each at 0.1 wt. % / styrene) added in the second stage with AA: 2 wt %/styrene (d = 2.0 μm, CVd = 2.3%).Scale bars are 5.0 µm.

Figure 4.15. SEM images and particle distribution histograms for PS particle samples AA135 synthesized in the presence of LnCl3.6H2O (where Ln = La, Nd, Eu, Tb, Dy, Ho and Tm each at 0.1 wt. % / styrene) added in the second stage with AA: 2 wt %/styrene without washing (A and C) (d = 2.1 μm, CVd = 6.6%, the arrow point to a particle that represented the small particle population) and after 3 cycles of washing (B and D) (d = 2.2 μm, CVd = 3.2%).

116 Microspheres by Three Stage Dispersion Polymerization

Figure 4.16. Screen captures of the mass cytometry results for (a) AA134 (4 particles were shown each of them has the 9 elements present); and (b) AA135 (2 particles were shown each of them has the 7 elements present). Both samples were synthesized in the presence of mixture of lanthanide salts (see Table 6-1) as examples of PS particles synthesized for the highly-encoded particles.

Figure 4.17: (a)Number of Tb ion per particle distributions measured by mass cytometry for a population of PS particles AA134 (red) and AA135 (blue) prepared by 3-DisP (b) A bi-variant plot of mass cytometry results for the Ho and Tm content of AA134 particles.

117 Microspheres by Three Stage Dispersion Polymerization

4.5 Bioconjugation

One application of metal encoded microspheres would be in sandwich-type immunoassays. As a first step in this direction, proof-of-principle experiments involving covalent attachment of proteins to the surface of these carboxylated microspheres were carried out. The experimental design was similar to the one used in Chapter 3 (Figure 3-11). Here, the particle sample AA105 (labeled with 139La and 169Tm) was used. These microspheres were prepared with 4.0 % AA and show more than 108 –COOH groups per particle on their surface. Thus, AA105 sample was deemed to covalently attach antibodies (mouse IgG), on the surface. Consequently, it should be able to capture antigens of interest (anti-mouse IgG). The analyte (anti-mouse IgG) is labeled with a metal-chelating polymer (MCP) to act as a reporter tag (a metal label that confirms successful bioconjugation to the microsphere surface). In our case, we used MCP that carries multiple copies of Pr3+ ions. Analysis of a successful binding event by mass cytometry would detect the 139La and 169Tm characteristic of the PS microsphere as well as the 141Pr signal from the analyte (anti-mouse IgG). In these experiments, MCP (similar to that described in Lou et al. [23] with different ligands as pendant groups) was used and identical chemistry was used for attachment of the polymer to the antibody. All the bioconjugation experiments described in this section were carried out by Dr Olga Ornatsky.

In the bioconjugation experiments, a sample of AA105 microspheres (labeled with 139La and 169Tm) was divided into 3 portions. One portion was kept untreated (named AA105). The second portion was activated by EDC and then was treated with BSA in PBS buffer as control for non- specific reporter tag binding (named AA105-BSA). The third portion was activated by EDC and then was treated with mouse IgG after EDC activation (named AA105-IgG). In parallel, a sample of commercial carboxylated microspheres (1.1 µm PS beads from Bangs Labs, # PC04N) were treated with mouse IgG as a positive control (named PS-COOH). This PS-COOH sample was considered to be a “positive control” for its well known ability to attach biomolecules through EDC chemistry.[24] These 3 samples (AA105-BSA, AA105-IgG and PS-COOH-IgG) were blocked with 0.5% BSA/PBS buffer for 1 h. Then the 3 samples and the untreated AA105 (total 4 samples) were incubated, separately, with anti-mouse-IgG-X4-Pr (100 µg/mL). After several washes (3 cycles of centrifugation and redispersion in PBS buffer), the samples were analyzed by mass cytometry. Between 30,000 and 60,000 microspheres were analyzed for each sample.

118 Microspheres by Three Stage Dispersion Polymerization

Figure 4.18. Log intensity values for 139La, 169Tm, and 141Pr from mass cytometry measurements for: Unmodified AA105 microspheres encoded with La and Tm. Pr signal here represents the nonspecific binding (first column set from left).

BSA-modified AA105 microspheres encoded with La and Tm. Pr signal here represents the nonspecific binding of the anti-mouse-IgG to the BSA on the surface of the microspheres i.e. negative control (second column set from left).

Mouse IgG-modified AA105 microspheres encoded with La and Tm. To which anti-mouse-IgG- X4-Pr was covalently attached. Pr signal here represents extent of specific binding of the anti-mouse- IgG to the microspheres (second column set from right).

Carboxylated PS microspheress (Bangs Labs) modified with mouse IgG and to which anti-mouse- IgG-X4 -Pr was covalently attached. Pr signal here represents the positive control (first column set from right).

These data come from the analysis of approximately 20,000 particles for each sample.

The data from these experiments are presented in Figure 4.18. The left-hand set of columns shows the result for a control experiment in which untreated AA105 microspheres were exposed to anti-mouse-IgG-X4-Pr (100 µg/mL) and tested for non-specific adsorption to the microsphere surface. This Pr signal was very weak and could be considered as the background noise. The second from the left set of columns reported on the interaction of anti-mouse-IgG-X4-Pr with the BSA-labeled beads. The Pr signal was higher than that of the bare AA105 microspheres. The second from the right set of columns presents the data for the interaction of anti-mouse-IgG-X4- Pr with the particles conjugated to mouse IgG. The strong signal here indicates detectable

119 Microspheres by Three Stage Dispersion Polymerization bioconjugation events, i.e. anti-mouse-IgG specifically bound to microspheres. The binding of the reporter tag to the metal-encoded beads was measured effectively by mass cytometry and yielded a signal two orders of magnitude greater than that due to non-specific adsorption of anti- mouse IgG to the particles, and one order of magnitude stronger than the non-specific interaction of the analyte with BSA conjugated to the particle surface.

In Figure 4.18, the commercial carboxylated sample functionalized with IgG (PS-COOH IgG) is represented by the right set of columns. These columns show the mass cytometry data for the interaction of anti-mouse-IgG-X4-Pr with PS-COOH IgG as a positive control sample. Even though the PS-COOH particles were much smaller than AA105, the Pr signal from the commercial carboxylated particles was way (17 times) higher than that of AA105. The strong reporter signal is crucial for constituting high-multiplexty bioassay as discussed in the Chapter 3. Even though the particles prepared with 3-DisP showed better ability to be attached to biomolecules than particles prepared with 2-DisP, the extent of surface functionalization of these particles were still not sufficient for high-multiplexty bioassays. These particles need further surface modification to enable the attachment of sufficient amount of biomolecules to the surface.

4.6 Summary

In this chapter I discussed the synthesis of lanthanide-containing poly(styrene-acrylic acid) copolymer microspheres by three stage dispersion polymerization (3-DisP) of styrene in ethanol. The PS microsphere samples were designed for highly multiplexed bioassays based upon a mass cytometer comprising a time-of-flight (TOF) inductively coupled plasma mass spectrometer. I investigated the addition of extra amounts of acrylic acid (AA) and introduced EGDMA as a crosslinking agent in a stage after the second stage to improve the surface functionality and lanthanide distribution in the microspheres. Factorial design experiments were used to understand the effect of the amount of AA and EGDMA added in the third stage as well as the time of the addition on the synthesized microspheres. 23 factorial experiments were made in random order using 2 levels for each of the three parameters (AA concentration, EGDMA concentration and time of the 3rd stage addition). I tried AA concentrations of 0.5 and 2.0 wt.% / styrene and EGDMA concentrations of 0.5 and 4.0

120 Microspheres by Three Stage Dispersion Polymerization

wt.% / styrene. In addition, I examined two different times for the addition of the 3rd stage (1.0 and 8.0 hr). Generally, microsphere samples synthesized with 3-Disp have narrower lanthanide content distribution than that of microspheres synthesized by 2-DisP. This was also true for the sample 3- made by 3-DisP and had a wide size distribution. When high concentration of EGDMA (4.0 wt.% / styrene) was used, the microsphere morphology was affected and the number of lanthanide atoms per unit volume of the microspheres decreased. Sample AA105 was synthesized with 2.0 wt.% / styrene of AA and 2.0 wt.% / styrene of EGDMA added 8.0 hr after the second stage. Sample AA105 was the most suitable sample for bioconjugation trial because of its combination of characteristics (lanthanide content and distribution and size distribution) compared to all the samples prepared for the factorial design experiments. Sample AA105 was tested for the leakage of ions in different buffers. The sample showed minimal release of Tm ions over the three weeks which makes it a good candidate for bioconjugation. With 3-DisP I could synthesize particles with up to 9 lanthanide elements. This means that enumeration encoded particles with a variability of almost 2 x 106 uniquely coded particles is attainable. As a test for the protein binding capabilities of AA105 microspheres, a simple model bioassay was constructed by choosing, as an analyte, anti-mouse IgG labeled with a 141Pr- containing metal chelating polymer. In mass cytometry experiments, a strong signal for Pr accompanied the signals for La and Tm associated with mouse IgG conjugated particle. The Pr signal was two orders of magnitude stronger than that due to non-specific adsorption of the metal chelating polymer onto the AA105 beads themselves and an order of magnitude stronger than that from non-specific adsorption to the BSA-functionalized beads. However, the signal from the positive control carboxylated microspheres was much stronger, suggesting that AA105 surface functionality is not optimized for high-multiplicity bioassays. As a conclusion, microspheres synthesized by 3-DisP showed a better lanthanide content distribution and a slightly-improved ability for bioconjugation compared to particles prepared by 2-DisP. In other words, 3-DisP method could be utilized to decrease the breadth of the lanthanide distribution; but it didn’t add much to the ability of surface functionalization. In the next chapters, I will examine other techniques to improve the surface functionality of the

121 Microspheres by Three Stage Dispersion Polymerization

microspheres synthesized by dispersion polymerization like using a functionalized dispersant in the dispersion polymerization or growing functional shells for already-made microspheres.

References:

1. Jennings TL, Rahman KS, Fournier-Bidoz S, Chan WCW: Effects of microbead surface chemistry on DNA loading and hybridization efficiency. Anal Chem 2008, 80(8):2849-2856. 2. Raez J, Biais DR, Zhang Y, Alvarez-Puebla RA, Bravo-Vasquez JP, Pezacki JP, Fenniri H: Spectroscopically encoded microspheres for antigen biosensing. Langmuir 2007, 23(12):6482-6485. 3. Song JS, Chagal L, Winnik MA: Monodisperse micrometer-size carboxyl- functionalized polystyrene particles obtained by two-stage dispersion polymerization. Macromolecules 2006, 39(17):5729-5737. 4. Thomson B, Rudin A, Lajoie G: Dispersion copolymerization of styrene and divinylbenzene: Synthesis of monodisperse, uniformly crosslinked particles. Journal of Polymer Science, Part A: Polymer Chemistry 1995, 33(3):345-347. 5. Thomson B, Rudin A, Lajoie G: Dispersion copolymerization of styrene and divinylbenzene. II. Effect of crosslinker on particle morphology. J Appl Polym Sci 1996, 59(13):2009-2028. 6. Hattori M, Sudol ED, El-Aasser MS: Highly crosslinked polymer particles by dispersion polymerization. J Appl Polym Sci 1993, 50(11):2027-2034. 7. Horak D, Svec F, Frechet JMJ: Preparation of colored poly(styrene-co-butyl methacrylate) micrometer size beads with narrow size distribution by dispersion polymerization in presence of dyes. Journal of Polymer Science Part a-Polymer Chemistry 1995, 33(17):2961-2968. 8. Kim JW, Kim BS, Suh KD: Monodisperse micron-sized cross-linked polystyrene particles. VI. Understanding of nucleated particle formation and particle growth. Colloid Polym Sci 2000, 278(6):591-594. 9. Lok KP, Ober CK: Particle-size control in dispersion polymerization of polystyrene. Canadian Journal of Chemistry-Revue Canadienne De Chimie 1985, 63(1):209-216. 10. Ober CK: Dispersion copolymerization in nonaqueous media. Makromolekulare Chemie-Macromolecular Symposia 1990, 35-6:87-104. 11. Shen S, Sudol ED, Elaasser MS: Dispersion polymerization of methyl-methacrylate - mechanism of particle formation. Journal of Polymer Science Part a-Polymer Chemistry 1994, 32(6):1087-1100. 12. Tseng CM, Lu YY, Elaasser MS, Vanderhoff JW: Uniform polymer particles by dispersion polymerization in alcohol. Journal of Polymer Science Part a-Polymer Chemistry 1986, 24(11):2995-3007. 13. Yasuda M, Seki H, Yokoyama H, Ogino H, Ishimi K, Ishikawa H: Simulation of a particle formation stage in the dispersion polymerization of styrene. Macromolecules 2001, 34(10):3261-3270.

122 Microspheres by Three Stage Dispersion Polymerization

14. Tseng CM, Lu YY, El-Aasser MS, Vanderhoff JW: Uniform polymer particles by dispersion polymerization in alcohol. Journal of Polymer Science, Part A: Polymer Chemistry 1986, 24(11):2995-3007. 15. Horák D: Effect of reaction parameters on the particle size in the dispersion polymerization of 2-hydroxyethyl methacrylate. Journal of Polymer Science, Part A: Polymer Chemistry 1999, 37(20):3785-3792. 16. Ahmad H, Dupin D, Armes SP, Lewis AL: Synthesis of Biocompatible Sterically- Stabilized Poly(2-(methacryloyloxy) ethyl phosphorylcholine) Latexes via Dispersion Polymerization in Alcohol/Water Mixtures. Langmuir 2009, 25(19):11442-11449. 17. Song JS, Winnik MA: Cross-linked, monodisperse, micron-sized polystyrene particles by two-stage dispersion polymerization. Macromolecules 2005, 38(20):8300- 8307. 18. Gunst RF, Mason RL: How to construct fractional factorial experiments. Milwaukee, Wis.: ASQC Quality Press; 1991. 19. Ober CK, Lok KP: Formation of large monodisperse copolymer particles by dispersion polymerization. Macromolecules 1987, 20(2):268-273. 20. Ober CK, Lok KP, Hair ML: Monodispersed, micron-sized polystyrene particles by dispersion polymerization. Journal of Polymer Science Part C-Polymer Letters 1985, 23(2):103-108. 21. Abdelrahman AI, Dai S, Thickett SC, Ornatsky O, Bandura D, Baranov V, Winnik MA: Lanthanide-containing polymer microspheres by multiple-stage dispersion polymerization for highly multiplexed bioassays. J Am Chem Soc 2009, 131(42):15276-15283. 22. Parks DR, Roederer M, Moore WA: A new "logicle" display method avoids deceptive effects of logarithmic scaling for low signals and compensated data. Cytometry Part A 2006, 69(6):541-551. 23. Lou X, Zhang G, Herrera I, Kinach R, Ornatsky O, Baranov V, Nitz M, Winnik MA: Polymer-based elemental tags for sensitive bioassays. Angewandte Chemie - International Edition 2007, 46(32):6111-6114. 24. Bandura DR, Baranov VI, Ornatsky OI, Antonov A, Kinach R, Lou X, Pavlov S, Vorobiev S, Dick JE, Tanner SD: Mass cytometry: Technique for real time single cell multitarget immunoassay based on inductively coupled plasma time-of-flight mass spectrometry. Anal Chem 2009, 81(16):6813-6822.

123 Surface Functionalization of Lanthanide-Encoded Microspheres

5 Surface Functionalization of Lanthanide-Encoded Microspheres In Chapter 3 and 4, I described the synthesis and characterization of lanthanide-encoded polystyrene microspheres by dispersion polymerization. These particles are excellent candidates for mass cytometry-based immunoassays; however they have previously shown no or little ability to conjugate biomolecules to the particle surface. I present here three approaches to post- functionalize these particles to enhance the covalent attachment of proteins. My first approach used partially hydrolyzed poly (N-vinylpyrrolidone) as a dispersion polymerization stabilizer to synthesize particles with a higher concentration of -COOH groups on the particle surface. An alternative strategy to provide -COOH functionality to the lanthanide-encoded particles, employed seeded emulsion polymerization to graft poly(methacrylic acid) (PMAA) chains onto the surface of these particles. In the third approach, seeded emulsion polymerization was used as a method to grow a functional polymer shell (in this case, poly (glycidyl methacrylate) (PGMA)) onto the surface of seed particles prepared by dispersion polymerization. All the seeded emulsion polymerization reactions were carried out by Dr Stuart Thickett using, as seeds, particles I synthesized by dispersion polymerization. Bioconjugation experiments were performed by Dr Olga Ornatsky. Most of the results presented in this chapter were published in Ref.[1].

5.1 Introduction

The PS particles prepared by 2-DisP and 3-DisP described in Chapter 3 and 4 had narrow size distributions and sufficient lanthanide content for mass cytometry detection. We encountered difficulty, however, in attaching significant numbers of biomolecules to the particles for use in model bioassays. For example, only small amounts of NeutrAvidin could be attached to the surface of these particles in spite of a large number of surface –COOH groups determined by titration. These particles were sterically stabilized by a corona of poly(N-vinylpyrrolidone) (PVP). While this polymer is useful in that it is effective at suppressing non-specific protein adsorption to the particles,[2] these chains also appear to interfere with bioconjugation to the carboxyl groups under the corona at the particle surface. In order to use particles synthesized by the 2-DisP method in immunoassays, the issue of the protein-repellent PVP surface layer must be addressed. There are some potential solutions to this problem, for example, modifying the PVP surface layer by opening (hydrolyzing) the

124 Surface Functionalization of Lanthanide-Encoded Microspheres

pyrrolidone ring (potentially yielding numerous carboxyl and amino groups within the corona), or creating a functional polymer shell over the pre-existing surface of the particles. There are examples in the literature of building polymer shells onto dispersion polymerization particles. Zhang et al.[3] reported seeded dispersion polymerization using a PS particle seed, with a mixture of styrene and methacrylic acid (MAA) in the second stage, to form carboxylated particles in ethanol. In their work, seed particles of diameter 1.9 µm were grown to a final size of 2.2 µm with a narrow particle size distribution (PDI = 1.02) with a very high accessible acid content by titration (up to 97 mg of acid monomer per gram of polymer). However, the reaction appeared to be delicate, and very specific conditions were required to obtain monodisperse particles in a coagulum-free system. There are other examples in the literature describing particles made by dispersion polymerization in ethanol and subsequently used as seed latexes for emulsion polymerization after redispersion (via centrifugation) into water. Three publications are of particular interest. Omer-Mizrahi et al.[4] synthesized a shell of poly(glycidyl methacrylate) (PGMA) onto PS particles in water. The shell was extremely thick (approximately 360 nm) and the particle morphology was raspberry-like, but the particle size distribution remained narrow. This synthesis employed a concentration of sodium dodecyl sulfate well above its critical micelle concentration, with the resultant morphology suggesting that the nucleated PGMA particles coagulate and stick to the surface of the significantly larger PS particles, forming a shell with functional epoxide groups. Alam et al.[5] synthesized pH-responsive particles by seeded emulsion polymerization of 1.7 µm diameter PS seed particles with a mixture of MAA, N-isopropylacrylamide (NIPAM) and a cross-linking agent in a 2:1 core:shell ratio. Finally, Ahmad et al.[6] reported seeded emulsion copolymerization of styrene and hydroxyethyl methacrylate (HEMA) with a PS seed at different comonomer ratios to form functionalized particles. All of these methods are potential candidates for functionalization of my metal-encoded dispersion polymerization particles. In this chapter, I present three approaches to synthesize lanthanide-encoded polymer particles with functional surfaces suitable for further bioconjugation, while retaining the narrow particle size and lanthanide content distributions only attainable with particles synthesized by dispersion polymerization. In the first approach, I pre-activated the PVP via basic hydrolysis[7] to provide functional groups on the PVP polymer, and then used the “activated” PVP as a component of the dispersants in 2-DisP copolymerization of styrene and acrylic acid. For the

125 Surface Functionalization of Lanthanide-Encoded Microspheres

second approach, Dr Stuart Thickett used the particles obtained by 2-DisP in the presence of PVP as seeds for second stage polymerization of methacrylic acid (MAA) with the idea of generating PMAA grafts on the PVP corona of the particles. In the third approach, Dr Thickett used the dispersion particles as seeds for second-stage polymerization with GMA to grow an epoxy-rich shell of PGMA over the PVP corona of the particles. I examined the particle size and particle size distribution of these particles by scanning electron microscopy (SEM). I also measured the metal content of these particles by mass cytometry. Their stability was monitored in different aqueous media over different time intervals. Results of both techniques were evaluated with respect to satisfying the criteria necessary for reproducible bioassays as determined by mass cytometry.

5.2 Results and Discussion

The objective of the experiments described here is to explore methods to modify the surface of metal-encoded polystyrene microparticles prepared by two-stage dispersion polymerization to improve the attachment of bioaffinity agents to the particle surface. As mentioned in chapter 3, particles prepared by dispersion polymerization in the presence of poly(N-vinylpyrrolidone) (PVP) as a polymeric stabilizer are protected against non-specific protein adsorption by a solvent-swollen corona of PVP chains. This corona also interferes with the attachment of biomolecules such as NeutrAvidin to the surface of the particles, in spite of a substantial concentration of carboxylic acid groups (ca. 109 per particle) detected by titration. I will use the effectiveness of NeutrAvidin conjugation as a measure of the ease of attachment of bioaffinity agents to the particles following surface modification.

126 Surface Functionalization of Lanthanide-Encoded Microspheres

The bioconjugation conditions used to attach NeutraAvidin to particles’ Scheme 5.1. surface.

127 Surface Functionalization of Lanthanide-Encoded Microspheres

Our approach used in bioconjugation is summarized in the lower half of Scheme 5.1. The particles were treated with an appropriate reagent (here we used EDC to activate surface –COOH groups) and then we attempted to couple NeutrAvidin to the particles. In a second step, the modified particles were exposed to a biotin-polypeptide conjugate reporter group (whose structure is shown in Figure 5.1) containing a single DTPA-Lu ion at the end opposite to biotin [8] as well as a spacer (βAla- βAla-βAla-βAla) and a small peptide (Gly-Ser-Ala-Tyr-Gly-Lys- Arg-Lys). The particle suspension was subsequently analyzed by mass cytometry. The high binding affinity of biotin with NeutrAvidin (K ~ 1015 M-1)[9-11] ensures that each surface-bound NeutrAvidin unit will be ‘labeled’ with 3 or 4 Lu ions, which can be detected by their mass. In our design, successful biofunctionalization would lead to attachment of a minimum of 105 Lu ions per particle.

Figure 5.1: The structure of the Lu biotin-conjugated reporter tag [8], which consists of a small peptide (Gly-Ser-Ala-Tyr-Gly-Lys-Arg-Lys) and a spacer (βAla- βAla-βAla-βAla) with a biotin molecule attached at one end and DTPA-Lu at the other end. The molecular weight of the reporter is 1783.1 g/mol, and was synthesized at the University of Toronto. This reporter will bind to any Neutravidin that is conjugated to the particle surface.

128 Surface Functionalization of Lanthanide-Encoded Microspheres

As a negative control, parallel experiments were carried out in which the EDC chemistry was used to attach bovine serum albumin (BSA) instead of NeutrAvidin to the particle surface. BSA-modified particles were to measure the extent of non-specific reporter group binding to the surface of the particles relative to the protein (NeutrAvidin) of interest. Following the addition of the Lu-biotin reporter and analysis by mass cytometry, the measured Lu intensity of the BSA modified particles would indicate the amount of non-specific reporter group binding to the particles. For the NeutrAvidin-labeled particles, I report the mean number of Lu atoms per particle, but for the BSA-labeled particles, I report the Neut/BSA ratio, the ratio of the mean number of Lu atoms per NeutrAvidin particle to the mean number of Lu atoms per BSA particle (see Table 5.2). High values of the Neut/BSA ratio indicate a particle population with very low amounts of non-specific reporter-group binding. Conversely, a Neut/BSA ratio that is close to 1 indicates extensive non-specific binding to the particle surface and the inability to differentiate the mass cytometric signal of a target protein compared to background adsorption events.

Scheme 2. Ring opening of PVP by basic hydrolysis

5.2.1 Dispersion Polymerization with activated PVP. I initially examined the use of activated PVP as a dispersant in 2-stage dispersion polymerization (2-DisP) experiments to synthesize Ln-labeled polymer particles. To activate chains of PVP55 (PVP with an average molecular weight of 55 kDa), the polymer was treated with 1.0 N KOH at 140 °C for 36 h (see Scheme 5.2). The amino group of the opened

pyrrolidone ring was then methylated with formaldehyde in the presence of NaBH4 to prevent its re-closing in solution. Back titration against KOH, using the method described in Frank et al.,[12] indicated that 18 % of the pyrrolidone rings were opened under these reaction conditions. Experimental details are presented in Chapter 2.

129 Surface Functionalization of Lanthanide-Encoded Microspheres

The activated PVP (in a mixture with un-activated PVP55) was then used as a dispersant in the preparation of lanthanide-encoded PS-co-PAA particles by 2-DisP in absolute ethanol (see the recipe in Chapter 2). I prepared two samples, denoted D2(PVP1.6) and D3(PVP3.2). Both samples were synthesized in the presence of five Ln chloride salts (La, Eu, Tb, Ho and Tm), each at the concentration level of 0.05 wt.% / styrene. The salts were dissolved in the ethanolic solution of AA (2.0 wt.% / styrene) and added to the reaction one hour after initiating the

polymerization. D2(PVP1.6) was prepared in the presence of activated PVP (1.6 wt.% / styrene)

while D3(PVP3.2) sample was prepared at a higher activated PVP concentration (3.2 wt.% / styrene). In both reactions, the total PVP content of the reaction mixture was 16 wt.% / styrene.

Figure 5-2. SEM images (A and C) and particle distribution histograms (B and D) for unwashed particles

synthesized by 2-DisP. D2(PVP1.6) (A and B) and D3(PVP3.2) (C and D) were prepared in presence of 1.6 and 3.2 0 wt.% / styrene, respectively.

Stable, coagulum-free dispersions were obtained in both experiments after 24 hours of polymerization. Figure 5.2 shows representative SEM images and histograms of the particle size distribution for unwashed particle samples D2(PVP1.6) and D3(PVP3.2). Sample D2(PVP1.6) had an average diameter of 1.60 µm, which is smaller than the diameter (ca. 2.0 µm) of particles

130 Surface Functionalization of Lanthanide-Encoded Microspheres

prepared with a similar recipe in the presence of only unactivated PVP55 [13]. The modified reaction maintained a narrow size distribution of the particles (CVd = 3.5 %). Sample

D3(PVP3.2), which was prepared in the presence of a higher concentration of the activated PVP,

had a much smaller average diameter (1.33 µm) and a broader size distribution (CVd = 4.4 %). I explain the small particle size produced in the presence of the modified PVP in terms of the role of grafting of the dispersant to PS in the dispersion polymerization reaction. The grafting of polystyrene from chains of PVP is essential for the anchoring of PVP onto the growing particles and its ability to act as a dispersant. If more PVP-PS grafting is possible, newly nucleated particles will be more effectively stabilized, resulting in more initial particles and a smaller overall final particle size. I now recognize that the activated PVP has sites with more labile hydrogen atoms along the polymer backbone (labeled with the asterisk in Scheme 5.2) due to the higher electron density on the nitrogen atom, which can more effectively stabilize the tertiary radical. The higher electron density on the nitrogen atom is a consequence of the replacement of the electron withdrawing amide group, in the PVP55, with the electron donating methyl group in the activated PVP.

Figure 5.3 Mass cytometry measurements for particles prepared by 2-stage dispersion polymerization in 165 159 165 presence of activated PVP. (A) Ungated Ho/ Tb bi-variant plot for D2(PVP1.6). (B) gated Ho distribution for D2(PVP1.6) (blue curves CVHo = 20.6) and D3(PVP3.2) (red curves CVHo = 27.1 %). (C) \159\\\\ gated Tb distribution for D2(PVP1.6) (blue curves CVTb = 21.7) and D3(PVP3.2) (red curves CVTb = 30.1 %). Gating applied was The metal content for these particles was measured by mass cytometry. Figure 5.3A shows 165 159 the ungated Ho/ Tb bi-variant plot for D2(PVP1.6) samples. In Figure 5.3A, the x- and y-axes represent the number of metal ions per particle, calculated from the metal ion intensities using Equation 3-3. As manifested by the red oval in Figure 5.3A, I gated out the low intensity signals

131 Surface Functionalization of Lanthanide-Encoded Microspheres

165 159 and the very high intensity signals in the Ho/ Tb bi-variant plot for D2(PVP1.6) sample. The process of gating was described in Chapter 3. The same gating method was applied to the data 165 from the D3(PVP3.2) sample. Gated Ho distribution histograms of D2(PVP1.6) and D3(PVP1.6) sample are presented in Figure 5.3B and gated 159Tb distribution histograms of the data fot

D2(PVP1.6) and D3(PVP1.6) samples are shown in Figure 5.3C. The average Ho atom content of 6 sample D2(PVP1.6) was 2.2 x 10 Ho atoms per particle and the coefficient of variation of the Ho

content distribution (CVHo) was 20.6 %, while sample D3(PVP3.2) was found to contain 1.24 x 6 10 Ho atoms per particle, with CVHo equal to 27.1 %. (Figure 5.3B). The ratio of Ho content

between the two samples [D2(PVP1.6) to D3(PVP3.2)] is 1.8, which is nearly identical to the ratios found for the other four metals incorporated into the interior of these particles (see Figure

5.3C for additional histograms of the Tb content of D2(PVP1.6) and D3(PVP3.2). I noticed that

this ratio (1.8) is also very close to the ratio of particle volumes of D2(PVP1.6) and D3(PVP3.2) (1.7). Consequently, the difference in the metal content could be ascribed to the difference in particle volume, and the use of differing amounts of activated PVP as stabilizers did not affect the ability to incorporate Ln ions into the particle interior during the synthesis. The ability of particles synthesized in the presence of activated PVP to undergo bioconjugation was then examined by attempted covalent attachment of NeutrAvidin to the surface of these particles using EDC chemistry. Aqueous samples of 109 polymer particles per mL were washed in Polylink Coupling buffer, and MES buffer was then added. A solution of EDC in MES buffer was prepared and added to the particles. The sample was then added to a tube containing NeutrAvidin. After mixing and reacting overnight at 4 °C, the extent of binding was measured through the use of the Lu biotin-conjugated reporter tag with subsequent measurement by mass cytometry. Figure 5.4 shows the Lu reporter signal distribution obtained

from the attempted covalent attachment of NeutrAvidin to D2(PVP1.6) and D3(PVP3.2) samples. Unfortunately, both samples have not shown any distinguished bead events that have detectable Lu signal. The signal shown in Figure 5.4 represents only the free Lu in the continuous media.

132 Surface Functionalization of Lanthanide-Encoded Microspheres

Figure 5.4: Mass cytometry results for bioconjugation trial of NeutrAvidin to D2(PVP1.6) (green) and D3(PVP3.2) (blue) samples synthesized by two stage dispersion polymerization in presence of activated PVP. The sample were tested by the incubation of 109 particles with 100 μL of 500 nM of Lu-reporter [8] in 1 % BSA, then washed once with PBS and finally redispersed in a 0.26 % w/w NaCl solution in 10 mM tris buffer.

In this paragraph, I attempt to explain the low level of Lu signal measured, from which I infer that little NeutrAvidin was covalently bound to the particle surface. From the decrease in particle size, I know that the modified PVP is selectively incorporated into the PS particles. Thus the problem with limited NeutrAvidin coupling is not due to the absence of modified PVP on the particle surface. Instead, the problem must be with the location of the carboxyl groups that were added to the particle. If grafting of PS to activated PVP takes place at the ring-opened sites, then these sites will be located close to or at the surface of PS core of the particle. From this perspective, these new carboxyl groups are hidden by the PVP corona from functionalization by proteins. It is also possible that the additional concentration of carboxyl groups generated by this method is insufficient to make an appreciable difference to the measured extent of bioconjugation. Given these results, no further analysis of this system was considered.

133 Surface Functionalization of Lanthanide-Encoded Microspheres

5.2.2 Seeded Emulsion Polymerization with Methacrylic Acid In our second approach, samples of PS particles synthesized by dispersion polymerization in the presence of PVP55 [13] were sedimented by centrifugation, redispersed in water, and used as seeds for subsequent modification by polymerization with methacrylic acid. The experimental design was based on the work of Alam et al.[5]. In their work, PS seed particles prepared by dispersion polymerization were transferred to water and subsequently reacted in a seeded emulsion polymerization process with a mixture of monomers, MAA, N-isopropylacrylamide, and an acrylamide-based crosslinker.

Table 5.1. Recipe for the synthesis of functional monomer coatings onto the surface of the PS seed latex D1 by seeded emulsion polymerization. Experiments were performed with methacrylic acid (D4(MAA)) or with glycidyl methacrylate in the presence of excess surfactant (D5(GMA-2))

Materials Amount added (g) D4(MAA) D5(GMA-2) Methacrylic acid 0.1 - Glycidyl methacrylate - 0.67 Seed Latex D1 2.25 1.71 (as solid polymer) (0.2) (0.15) Water 8.33 9.48 Sodium dodecyl sulfate - 0.1 Potassium persulfate 0.005 0.012

134 Surface Functionalization of Lanthanide-Encoded Microspheres

MAA was used as the sole second stage monomer with the idea of incorporating chains of the moderately water-soluble PMAA into the seed particles. In this way, we hoped to generate particles with an increased acid content at the surface for subsequent bioconjugation experiments. The ratio (by mass) of polystyrene to MAA was chosen to be 2:1. Seed particles were swollen overnight at room temperature with the added MAA under constant stirring prior to polymerization. The experimental parameters used for these syntheses are shown in Table 5.1. Experimental details are presented in Chapter 2. The seed latex utilized was the redispersed sample D1, which is shown in Figure 5.5A. SEM analysis of the particle population (denoted D4(MAA)) after polymerization and three subsequent centrifugation and redispersion cycles into pure water indicated that the particles

remained smooth and monodisperse (davg = 2.1 µm, CVd = 1.6 %, see Figure 5.5B.). There was evidence in the SEM micrograph in Figure 5.5B that some particles were slightly deformed after polymerization with MAA, in contrast to the uniform PS spheres in the seed latex (Figure 5.5A), but the origin of this effect is unclear. The fractional conversion of MAA in D4(MAA) was determined by gravimetry to be 70 %, however no increase in average particle size was observed relative to the seed latex D1. This curious observation was similar to the results reported by Alam et al.,[5] who saw little increase in particle size after polymerization, but verified the presence of acid groups on the surface of their particles by demonstrating a strong particle size dependence on the pH of the continuous phase. Upon centrifugation and redispersion of the particles into pure water, we found that approximately 48 % of the newly formed solids had been washed away in the supernatant. The large reduction in the solids content of the latex was attributed to aqueous-phase polymerization of MAA occurring at the same time as grafting of PMAA chains onto the particles. Despite the large amount of aqueous-phase PMAA formed during this process, the mass fraction of acidic polymer incorporated into the particles increased significantly relative to the D1 seed particles: in sample D4(MAA), the mass of PMAA was ca. 13 % of the total mass of polymer, while in D1, the mass of incorporated acrylic acid comonomer was only 1.6 % of the total polymer mass.

135 Surface Functionalization of Lanthanide-Encoded Microspheres

Figure 5.6 Surface acid titration results of D1 (black circles, seed latex) and D4(MAA) (open circles, after polymerization with MAA) after neutralization with excess NaOH followed by back titration with 0.01 M HCl performed under identical conditions.

The concentration of surface acid groups was measured for the particles D1 and D4(MAA) by conductometric and potentiometric surface acid titration. While gravimetric analysis of D4(MAA) demonstrated an increase in the acid content of the particles, the titrations were needed to determine the location of these additional acid groups. From the synthesis protocol (pre-swelling the particles with MAA prior to polymerization), it was likely that some of the added PMAA was buried within the particle interior where it would not be detected by titration [14]. An increase was observed in titratable acid groups by approximately a factor of two (from 1.0 × 109 to 2.1 × 109 acid groups per particle) after polymerization with MAA and purification of the particles. Under identical dilution conditions, both equivalence points (representing the titration of free base and surface carboxylate groups respectively) occur at the same pH (see Figure 5.6) while a greater number of titratable acid groups was clearly evident on D4(MAA) (open circles) compared to D1 (filled circles). Based on the particle number density (i.e. the number of polymer particles per unit volume of latex) and the PMAA content of my particles by gravimetry, we predict 2.4 × 109 acid groups per particle from the added PMAA. If we assume that the increase in titratable surface acid groups (i.e. an additional 1.1 × 109 acid groups per particle) was solely from the PMAA now present in the system, we can conclude that nearly 50 % of the PMAA present in D4(MAA) is titratable and hence resides at the particle surface.

136 Surface Functionalization of Lanthanide-Encoded Microspheres

To test the effectiveness with which the PMAA particles (D4(MAA)) could be biofunctionalized, EDC chemistry was used in an attempt to attach NeutrAvidin to the particles. In parallel, identical experiments on the precursor D1 particles were carried out. After washing the particles by successive centrifugation-redispersion steps, the particles were treated with an excess of the biotin-polypeptide-Lu reagent and analyzed by mass cytometry.[8] As a negative control we used similar chemistry to bind BSA to both D1 and D4(MAA) particles and, after washing, treated these particles with an excess of the biotin-polypeptide-Lu reagent. The results for these experiments are presented in Table 5.2. In the middle column, I report the mean number of Lu ions per particle detected by mass cytometry for the D1 particles and for the D4(MAA) particles following the surface binding of NeutrAvidin and treatment with the Lu reporter as described above. The Lu signal from the D4(MAA) particles was measured to be approximately 50% higher (1.0 x 105 vs. 6.7 x 104 Lu ions per particle) than the D1 particles, suggesting a slight increase in the attachment of NeutrAvidin to the modified particles. The data presented in the middle column of Table 5.2 also consists of a contribution that was due to non-specific adsorption of the reporter group to the particles. As a control, similar experiments on both particles treated with EDC were carried out and then reacted with BSA prior to exposure to the Lu-reporter group. The right-hand column of Table 5.2 is the ‘Neut/BSA ratio’ for the two sets of particles studied. The data in Table 5.2 indicates there was only a 10% increase in the Lu signal above that measured for non-specific binding to the D1 seed particles, and only a 40% increase for binding to D4(MAA). I infer from these results that introducing grafted PMAA chains onto the surface of my seed particles did not greatly improve the amount of specific protein binding, with the Neut/BSA ratio being only slightly above unity. These disappointing results mentioned above can be contrasted with other experiments in our laboratory, with the PS particles synthesized by a combination of surfactant free emulsion polymerization and seeded emulsion polymerization, by Dr. Stuart Thickett, to obtain Ln- encoded PS particles. This synthesis employed azo-cyanopentanoic acid as an acid-containing initiator and gave much more encouraging results in terms of the signal from the biotin- polypeptide reporter group. In those experiments, they found a typical Lu reporter signal of over 2.8 x 105 Lu ions per particle, and a Neut/BSA ratio of over 100 under bioconjugation conditions identical to those reported here [15]. Those particles had other problems (a tendency to aggregate in solution and broader particle-to-particle distribution of loaded Ln ions compared to particles

137 Surface Functionalization of Lanthanide-Encoded Microspheres made by dispersion polymerization) that led us to carry out the experiments reported in this chapter. However, as the grafting of PMAA to D1 seed particles did not lead to a sufficient improvement in the amount of bioconjugation to the particles, we turned to other approaches to meet our goals. 5.2.3 Seeded Emulsion Polymerization with Glycidyl Methacrylate In a second approach, the methodology described by Omer-Mizrahi et al. [4] for the growth of a poly(glycidyl methacrylate) (PGMA) shell onto the surface of PS particles by seeded emulsion polymerization was followed. This procedure is unusual compared to most seeded emulsion polymerization experiments in that the polymerization was performed with a surfactant concentration well above its CMC. These authors obtained particles with a raspberry-like surface morphology. We infer from this result that small, SDS-stabilized PGMA particles (of the order of 50 nm) are nucleated in the reaction and then coalesce onto the surface of the much larger PS particles (of the order of 2 µm) to build a PGMA shell as the polymerization proceeds (see Scheme 5.3). Such a coating of PGMA would potentially cover up the protein-repellent PVP layer, and the presence of reactive epoxide groups on the surface should permit subsequent bioconjugation.

Scheme 5.3. Proposed polymerization mechanism whereby agglomeration of small PGMA particles takes place onto the surface of the larger PS particles.

138 Surface Functionalization of Lanthanide-Encoded Microspheres

In our experiments, the D1 PS seed particles were transferred from ethanol to water via three centrifugation and redispersion cycles. The ratio of GMA to PS was set at approximately 4:1 by mass, and the recipe used to prepare sample D5(GMA-2) is presented in Table 5.1. A number of experiments were performed in order to determine the ability to form PGMA shells on the surface of the D1 PS particles (see Table 5.3). The SEM image shown in Figure 5.7 demonstrates an increase in particle size compared to the smooth D1 PS seed particles (Figure 5.5A.), with a final roughened spherical morphology. A large number of very small PGMA particles, with diameters in the range of 50 – 100 nm, were observed in the background of these SEM images. A sample of the supernatant fluid (after centrifugation of the larger PS particles) was examined by dynamic light scattering and showed the presence of particles with a mean diameter of 52 nm. In the larger PS particles, the PGMA shell was approximately 60 – 100 nm thick, with only a small percentage of the added GMA actually grafted to the particle surface. Washed, redispersed and dried particles were analyzed by FTIR, confirming the presence of PGMA grafting from characteristic epoxide peaks at 845 and 910 cm-1. The previously reported ‘raspberry’ morphology was seen in only a few experiments (see Figure 5.8 for the SEM image of sample D5(GMA-1)), suggesting that the resultant particle morphology is sensitive to system parameters such as stirring effects and the concentration of added surfactant. This is possible given the unusual polymerization mechanism employed in these experiments.

139 Surface Functionalization of Lanthanide-Encoded Microspheres

I examined the lanthanide (Ln) content distribution of PGMA-coated particles by mass cytometry and compared the results to those obtained with the D1 precursor particles. Recall that the D1 seed latex contained La, Tb, Ho and Tm (particle sample AA087 described in Chapter 3 after 3 cycles of washing and redispersion in water) and that the particle-to-particle variation

(CVTb) of these elements was 15.1 %. Mass cytometry data for sample D5(GMA-2) containing a PGMA shell is presented as an isotopic ‘dot-plot’ diagram in Figure 5.9A. The color intensity in Figure 5.9A represents a 2-dimensional map of the distribution of both the 159Tb and 165Ho isotopes. A one-dimensional projection of Figure 5.9A onto the 159Tb axis (after gating) is shown in Figure 5.9B as a single isotope distribution. This distribution is both monomodal and narrow. The narrow variation in number of ions found for each isotope demonstrates that the construction of the PGMA shell on the particle surface preserves the narrow distribution of metal ions in the particle interior.

Sample Number of Lu ions per particle Neut/BSA a D1 (seed PS particles) 6.7 x 104 1.1 b 5 D4(MAA) 1.0 x 10 1.4

a Neut/BSA refers to the ratio of the average number of Lu ions per particle of the sample with

NeutraAvidin bound to the surface compared to the equivalent experiment where BSA was

attached to the surface.

b Synthesized by seed emulsion polymerization with MAA using D1 seeds.

140 Surface Functionalization of Lanthanide-Encoded Microspheres

Figure 5.9 Mass cytometry data for the analysis of sample D5(GMA-2), which consists of a PGMA shell on D1 seed particles. Shown are (A) Tb-Ho isotopic dot-dot diagram; (B) Tb content distribution; (C) Lu content distribution for the detection of the Lu-biotin reporter, which measures the extent of protein binding to the surface. Red curve: non-specific binding of the Lu-biotin reporter to sample D5(GMA-2)-BSA (covalent attachment of BSA, CVLu = 65.8 %). Blue curve: binding of the Lu- biotin reporter to sample D5(GMA-2)-NA (covalent attachment of NeutraAvidin without EDC, CVLu = 32.5 %). Green curve: binding of the Lu-biotin reporter to sample D5(GMA-2)-EDC-NA (covalent attachment of NeutraAvidin after surface activation with EDC, CVLu = 27.8 %).

141 Surface Functionalization of Lanthanide-Encoded Microspheres

Table 5.3. Reproducibility of Surface Coating Experiments Using an Excess of GMA

Sample GMA: PS davg initial, CVd initial, Shell % GMA Comment

Code Ratio davg final CVd final Thickness Incorporated (precursor) (µm) (%) (nm) D5(GMA -1) 3.87 2.11, 2.67 1.3, 3.6 280 30 Raspberry (D1) -like D5(GMA-2) 4.41 2.11, 2.23 1.3, 1.4 60 6 Rough,

(D1) spherical

D5(GMA-3) 5.21 2.11, 2.24 1.3, 3.9 65 6 Rough,

(D1) spherical D5(GMA -4) 3.96 1.61, 1.79 1.8, 2.6 90 29 Rough, (D1’) a spherical

a Seed particles (davg = 1.6 µm) synthesized by two stage dispersion polymerization (see Ref. 1). D1 refers back to sample AA087 in Chapter 3

Following the formation of a PGMA shell on the PS seed latex, we examined the covalent attachment of NeutrAvidin to the surface of these particles. The amount of NeutrAvidin bound to the particle surface was measured by mass cytometry via the analysis of the binding of the Lu-tagged biotin-polypeptide reporter. The first set of experiments involved direct reaction of NeutrAvidin with the PGMA-coated particles (sample D5(GMA-2)). Azizi et al.[16] reported that epoxides readily react with aliphatic and aromatic amines in water at room temperature, and based on this precedent we examined the possibility that lysine amine groups of NeutrAvidin would react directly with the epoxide groups present on the surface of my particles. This experiment is denoted D5(GMA-2)-NA in Table 5.4, we observe that there is a significant increase (approximately eight times greater) in the number of reporter Lu ions per particle for bioconjugation to D5(GMA-2) relative to the seed particles (D1). In addition, the amount of non-specific reporter-group binding to D5(GMA-2) decreased significantly (4.7 x 104 Lu ions were detected per particle for non-specific binding to D5(GMA-2) compared to 6.7 x 104 Lu ions

142 Surface Functionalization of Lanthanide-Encoded Microspheres

for D1). This increase led to an increase in the Neut/BSA ratio from 1.1 to 10.4. The distribution of Lu ions per particle is shown in Figure 5.9C for D5(GMA-2)-NA (blue curve). This peak is much narrower and appears at greater Lu ion values than the signal for the negative control D5(GMA-2)-BSA (red curve). The Lu signal from the negative control measures the tendency of the Lu-biotin polypeptide to bind non-specifically to the BSA-labeled particles. These results demonstrate a significant improvement in the extent of bioconjugation to my particles following the construction of a PGMA shell.

Table 5.4 Mass cytometry results for bioconjugation of NeutraAvidin to the surface of PS dispersion polymerization particles before and after seeded emulsion polymerization with GMA. Binding was probed using an Lu-biotin reporter.

Sample Number of Lu ions (reporters) Neut/BSA a

D1 (seed PS particles) 6.7 x 104 1.1

D5(GMA-2)- NAb 4.9 x 105 10.4 D5(GMA -2)-EDC-NAc 7.1 x 105 15.1 ST120d 3.2 x 105 11.5 a Neut/BSA refers to the ratio of the average number of Lu ions per particle of the sample with NeutraAvidin bound to the surface compared to the equivalent experiment where BSA was attached to the surface. b NeutraAvidin was conjugated to PGMA-coated particles without EDC activation. c PGMA -coated particles was activated with EDC prior to NeutraAvidin attachment to the surface. d PS particles made by seeded emulsion polymerization (see Ref. 3) In a second set of experiments, the PGMA-coated particles were pretreated with EDC prior to exposure to NeutrAvidin. I denote these particles as D5(GMA-2)-EDC-NA in Table 5.4. While this experiment was designed as a kind of control, since one would not expect the hydrolysis of surface epoxy groups [17] in water to form carboxylic acids, we in fact found a 50% increase the measured number of reporter Lu ions per particle. The data in Table 5.4 not only shows the 50% increase in Lu signal compared to the particles reacted with NeutrAvidin without treatment with EDC, but that the Neut/BSA ratio increased from 10.4 to 15.1, indicating a further reduction in non-specific protein binding (2.8 x 104 Lu ions were detected per particle

143 Surface Functionalization of Lanthanide-Encoded Microspheres

for non-specific binding to D5(GMA-2)). The Lu content distribution for D5(GMA-2)-EDC-NA

is shown in Figure 5.9C (green curve); the distribution is monomodal, narrow (CVLu = 27%) with a significantly greater number of Lu ions per particle than seen in the data for non-specific reporter-group binding (red curve). These results suggest that there are carboxylic acid groups present on the surface of the PGMA-coated particles, which are accessible to activation with EDC and subsequent reaction with NeutrAvidin. While we have no unambiguous explanation for this result, it may be that some of the –COOH groups present on the surface of the seed D1 particles migrated to the surface of the particles during the polymerization with GMA. I compare the results for bioconjugation of NeutrAvidin to PGMA-coated particles presented here to the results obtained with the Ln-encoded PS particles mentioned above that were prepared by a combination of surfactant-free emulsion polymerization and seeded emulsion polymerization.[15] The acid groups in these particles come from the ACVA initiator (4, 4'- Azobis(4-cyanovaleric acid)), which generates particles with an ever higher titratable surface acid concentration than present in D1. In Table 5.4 I compare bioconjugation data from Ref. 30 for the covalent attachment of NeutrAvidin to these particles (denoted ST120) with data obtained here for PGMA-coated particles. Here we found more than two-fold higher conjugation of NeutrAvidin to the surface of the PGMA-coated particles (D5(GMA-2)-EDC-NA) than that found for sample ST120, at an equivalent Neut/BSA ratio. The data in Table 5.4 show that an average of 7 × 105 Lu ions were detected per particle for D5(GMA-2)-EDC-NA sample, indicating effective reporter group binding and effective covalent attachment of NeturAvidin to the particles. Thus, sample D5(GMA-2)-EDC-NA represents the best demonstration of surface bioconjugation to date.

5.3 SUMMARY

Previous experiments reported by our laboratory described the synthesis of polystyrene microparticles encoded with lanthanide ions, designed to be used in multiplexed bead-based assays with detection by mass cytometry. The PS particles were synthesized by two-stage dispersion copolymerization of styrene, acrylic acid, and lanthanide salts in ethanol in the presence of poly(N-vinylpyrrolidone) (PVP) as a polymeric stabilizer. These particles were characterized by a narrow particle size distribution and a controllable number, ranging from 105

144 Surface Functionalization of Lanthanide-Encoded Microspheres to 108, of Ln ions per particle. Although surface titration of these particles in water showed the presence of substantial numbers of surface carboxylic acid groups (ca. 109 per particle), covalent coupling of bioaffinity agents to the particle surface was inefficient. It appears that the water- swollen PVP corona grafted to the PS particle surface, which likely extends a few nm into the aqueous medium, screens the surface against the approach by molecules as large as antibodies and NeutrAvidin to the PS surface, where the –COOH groups are located. Thus only low levels of biofunctionalization could be detected. In this chapter, I describe three strategies designed to overcome this problem. In the first strategy, I carried out new particle syntheses using a mixture of PVP and partially hydrolyzed PVP as the polymeric stabilizer. Ring-opening hydrolysis of PVP, followed by methylation of the secondary amino group produces a polymer with pendant carboxylic acid groups. If these –COOH groups were located at the edge of the corona, they should be accessible for attaching biomolecules such as NeutrAvidin. The particle syntheses were successful in that PS particles with a narrow size distribution were obtained. The decrease in particle size in the presence of the partially hydrolyzed polymer indicates that these polymer molecules are more reactive toward grafting than PVP itself: more nuclei were formed leading to a larger number of smaller particles. Bioconjugation experiments, however, demonstrated that the attachment of NeutrAvidin was only marginally improved by the presence of reactive stabilizers, This result suggested that the –COOH groups on the partially hydrolyzed PVP are located close to the inner PS surface of the particles and are not accessible for the covalent attachment of biomolecules. In the second approach, Ln-encoded PS microparticles obtained by 2-DisP were used as seeds for second-stage emulsion polymerization with methacrylic acid (MAA). The particles were first swollen with MAA and then polymerization was initiated with potassium persulfate. After purification by centrifugation-resuspension to remove PMAA homopolymer, the particles remained narrow in size distribution, increased in mass by 13%, and the number of titratable acid groups had doubled to 2.1 × 109 acid groups per particle. Nevertheless, we had difficulty in coupling significant amounts of NeutrAvidin to the particle surface via EDC chemistry. In the third approach, the same Ln-encoded PS particles were used as seeds for emulsion polymerization of glycidyl methacrylate (GMA), with the idea of coating the particles with a layer of PGMA. Polymerization in the presence of excess surfactant led to a roughened particle morphology in which a shell of PGMA ca. 60 – 100 nm thick was deposited onto the surface of

145 Surface Functionalization of Lanthanide-Encoded Microspheres

the seed particles. These PGMA particles were stable to centrifugation and redispersion into water. The PGMA provided reactive sites to attach NeutrAvidin directly to the particle surface via the reaction of the epoxide group with primary amine residues. As a negative control, BSA was used in place of NeutrAvidin. The extent of coupling was tested with a biotin-polypeptide-Lu reporter probe and monitored by mass cytometry. For the direct labeling of the particles with NeutrAvidin, I detected an average of 4.9 x 105 ions per particle, more than 8 times greater than for the seed particles themselves. More importantly, the enhanced signal was accompanied by a substantial reduction in non-specific binding. Taking the Lu signal for the BSA-functionalized beads as a measure of non-specific binding, I found a Neut/BSA ratio of 10 for the GMA particles, nearly an order of magnitude improvement on the seed particles themselves. Further increases to the amount of bioconjugation was achieved when EDC was used as an activating agent, indicating that carboxylic acid groups are also present at the particle surface. For these particles, I found an average of 7 × 105 Lu ions were per particle, with a Neut/BSA ratio of 15.

In the current state of the art for the development of Ln encoded polymer microparticles for particle-based assays by mass cytometry, I have shown that two-stage dispersion copolymerization for the synthesis of PS particles provides excellent control over particle size, size distribution and Ln ion content. Improvements were needed to enhance the biofunctionalization of these particles. In this chapter I show that the formation of PGMA shells by seeded emulsion polymerization is a very promising approach for solving this problem. The incorporation of PGMA onto the surface of my particles increased the amount of protein conjugation by a factor of 10 or more relative to the original particles, while retaining particle colloidal stability and a narrow particle size distribution. Measurement of the Ln content distribution of particles by mass cytometry was also unaffected by the presence of the PGMA shell. Given the ability to resolve numerous metal tags simultaneously by mass cytometry, I believe that this methodology is a viable route for the development of particles for use in multiplexed bioassays based on mass cytometry detection.

146 Surface Functionalization of Lanthanide-Encoded Microspheres

References:

1. Abdelrahman AI, Thickett SC, Liang Y, Ornatsky O, Vladimir Baranov, Mitchell A. Winnik: Surface Functionalization Methods to Enhance Bioconjugation in Metal-Labeled Polystyrene Particles for Immunoassays. Macromolecules 2011. 2. Wu D, Zhao B, Dai Z, Qin J, Lin B: Grafting epoxy-modified hydrophilic polymers onto poly(dimethylsiloxane) microfluidic chip to resist nonspecific protein adsorption. Lab on a Chip - Miniaturisation for Chemistry and Biology 2006, 6(7):942-947. 3. Zhang H, Huang H, Sun R, Huang J-X: Preparation of Micron-Size Monodispersed PS/P(St/MAA) Microspheres by Seeded Dispersion Polymerization. J Appl Polym Sci 2006, 99:3586-3591. 4. Omer-Mizrahi M, Margel S: Synthesis and characterization of magnetic and non-magnetic core-shell polyepoxide micrometer-sized particles of narrow size distribution. J Colloid Interface Sci 2009, 329(2):228-234. 5. Alam MA, Miah MAJ, Ahmad H: Synthesis and characterization of dual-responsive micrometer-sized core-shell composite polymer particles. Polym Adv Technol 2008, 19(3):181-185. 6. Ahmad H, Miah MAJ, Rahman MM: Preparation of micron-sized composite polymer particles containing hydrophilic 2-hydroxyethyl methacrylate and their biomedical applications. Colloid Polym Sci 2003, 281:988-992. 7. Von Specht BU, Seinfeld H, Brendel W: Polyvinylpyrrolidone as a soluble carrier of proteins. Hoppe-Seyler's Zphysiolchem 1973, 354(12):1659-1660. 8. Lathia US, Ornatsky O, Baranov V, Nitz M: Development of inductively coupled plasma-mass spectrometry-based protease assays. Anal Biochem 2010, 398(1):93-98. 9. Hong CY, Pan CY: Direct synthesis of biotinylated stimuli-responsive polymer and diblock copolymer by RAFT polymerization using biotinylated trithiocarbonate as RAFT agent. Macromolecules 2006, 39(10):3517-3524. 10. Sankaran NB, Rys AZ, Nassif R, Nayak MK, Metera K, Chen BZ, Bazzi HS, Sleiman HF: Ring- Opening Metathesis Polymers for Biodetection and Signal Amplification: Synthesis and Self-Assembly. Macromolecules 2010, 43(13):5530-5537. 11. Weber PC, Ohlendorf DH, Wendoloski JJ, Salemme FR: STRUCTURAL ORIGINS OF HIGH-AFFINITY BIOTIN BINDING TO STREPTAVIDIN. Science 1989, 243(4887):85-88. 12. Frank HP: The lactam-amino acid equilibria for ethylpyrrolidone and polyvinylpyrrolidone. Journal of Polymer Science 1954, 12(67):565-576. 13. Abdelrahman AI, Dai S, Thickett SC, Ornatsky O, Bandura D, Baranov V, Winnik MA: Lanthanide-containing polymer microspheres by multiple-stage dispersion polymerization for highly multiplexed bioassays. J Am Chem Soc 2009, 131(42):15276-15283. 14. Rios L, Hidalgo M, Cavaille JY, Guillot J, Guyot A, Pichot C: Polystyrene (1) poly(butyl acrylate-methacrylic acid)(2) core-shell emulsion polymers .1. Synthesis and colloidal characterization. Colloid Polym Sci 1991, 269(8):812-824. 15. Thickett SC, Abdelrahman AI, Ornatsky O, Bandura D, Baranov V, Winnik MA: Bio-functional, lanthanide-labeled polymer particles by seeded emulsion polymerization and their characterization by novel ICP-MS detection. J Anal At Spectrom 2010, 25(3):269-281. 16. Azizi N, Saidi MR: Highly chemoselective addition of amines to epoxides in water. Org Lett 2005, 7(17):3649-3651. 17. Kalal J, Svec F, Marousek V: Reactions of epoxide groups of glycidyl methacrylate copolymers. Journal of Polymer Science Part C-Polymer Symposium 1974(47):155-166.

147 SILICA-COATED PARTICLES

6 SILICA-COATED PARTICLES In Chapter 3, I described the synthesis of lanthanide-encoded polystyrene particles using two-stage dispersion polymerization (2-DisP) and in Chapter 4, I described three-stage dispersion polymerization (3-DisP) as a method for preparing lanthanide-encoded polystyrene particles. These particles were loaded with up to 9 different lanthanide metals and some of them had 5 levels of concentration. The enumeration encoding of such particles, as defined by 6 9 Equation 1-1, had a variability (VR) of 2 x 10 (i.e. 5 − 1). The combination (9 lanthanides and 5 concentration levels) would construct a library of nearly 2 x 106 uniquely encoded particles that could be used in bio-analytical applications like immunoassays and oligonucleotide assays. In Chapter 5, I demonstrated the utilization of seeded emulsion polymerization technique to grow functional shells using particles produced by dispersion polymerization as seeds. The objective of those experiments was to enable covalent attachment of biological molecules to the surface of the particles. In this chapter, I describe alternative method to functionalize the surface of the Ln- encoded particles (made by 3-DisP) through growing functionalized silica layers. All the bioconjugation experiments included in this chapter were performed by Dr Dirk Weinrich.

6.1 Surface Functionalization with Silica Shell

One widely used method to modify the surfaces of different types of polymer particles is silica-coating [1-11]. Silica-coating can be used to introduce a variety of functionalities to the surface of the particles. Various coating procedures have been developed for that purpose. These include Stöber method [12], aqueous deposition [13, 14], miniemulsion [10] and using silane coupling agents as surface primers [15]. The most common method to coat polystyrene particles with silica is through Stöber growth [12] of silica shells by addition of tetraethoxysilane to solutions of seed particles in an ethanol/ammonia mixture. To coat polystyrene particles with silica, one has to overcome the incompatibility between the hydrophilic silica and the hydrophobic polystyrene particle surface. In other word, an interfacial affinity between the substrate surface (polystyrene particles) and forming silica has to be achieved. This is usually attained through chemical or physical surface treatment of the

148 SILICA-COATED PARTICLES

polystyrene particles. For example, Yang et al. [16] modified amino-functionalized polystyrene particles (PSt-NH2, 2.1 µm in diameter and PDI of 1.05) with 0.1 N of NaNO2 to obtain + + diazonium-functionalized particles (PSt-N2 ). After the incubation of PSt-N2 particles with 35

nm SiO2 nanoparticles at 0 °C, Si-coated microspheres were obtained via ionic linkage. Due to the diazonium group sensitivity to heat, Yang et al. could convert this ionic bond into a true chemical bond upon thermal treatment (60 °C , 1.0 h). These Si-coated particle were very stable toward ultrasonication as well as toward etching against polar solvent. Because of their simplicity, physical surface treatments are very common methods to prepare the polystyrene particle surface for silica coating. In these methods, an amphiphilic polymer is adsorbed onto the particle surface. When a Si monomer (e.g. tetraethoxysilane (TEOS, Figure 6-1)) is added, the amphiphilic polymer on the particle surface hosts the silica shell formation. The most common polymer used in this physical modification method is poly(vinylpyrrolidone) (PVP). This is due to the PVP amphiphilic character that enables it to be adsorbed onto many different surfaces. The amphiphilic characteristics of PVP can be attributed to the presence of apolar methylene and methine groups in the ring and along the polymer backbone, as well as the highly polar amide group in the pyrrolidone ring [17]. Using physically adsorbed PVP, various nanostructured particulate materials have been coated by the silica layer, including gold, silver, bohemite rods and most importantly polystyrene [5].

tetraethoxysilane (TEOS) 3-(Aminopropyl)triethoxysilane (APTS)

Figure 6.1: Structures of TEOS and APTS used to coat the PS particles.

149 SILICA-COATED PARTICLES

Hong et al. [18] synthesized polystyrene particles by dispersion polymerization using PVP stabilizer. The particles were 2.3 µm in diameter and had CV of 8.4 %. Since these particles contain PVP chains on their surface, the formation of silica shell from TEOS in the presence of these particles was facilitated to give core/shell structure. The silica shell had a thickness of 50 – 70 nm. Similar to this method reported by Hong et al. [18], my method of coating the polystyrene particles relied on the presence substantial amount of PVP already resides on the surface of these particles. In this section, I describe the coating of AA135 samples (synthesized by 3-DisP in presence of 7 lanthanide elements and described in Chapter 4), with a silica shell using TEOS though the Stöber method. Then the silica coated particles were functionalized by growing another layer of

(3-Aminopropyl) triethoxysilane (APTS, Figure 6-1) to provide –NH2 group functionality to the particles surface (See scheme 6-1).

Scheme 6-1: The procedure for the synthesis of silica coated PS particles

Sample AA167 refers to the silica-coated particles of AA135. This sample was prepared by washing AA135 particles in ethanol (three cycles of sedimentation by centrifugation at 4000 rpm for 15 min and redispersion) to obtain 3 ml of sample at 1.0 wt. % solids content. Then ammonia

(29.3 wt % NH3 in water) was added to obtain a 4.2 vol % ethanolic ammonia solution. Immediately afterwards, a 0.5 ml TEOS solution (10 vol % in ethanol) was added under stirring (600 rpm), and the reaction mixture was stirred for another 12 h to obtain AA167 particles.

150 SILICA-COATED PARTICLES

Figure 6.2 (a) and (c) SEM image and particle distribution histogram for sample AA167 (silica- coated AA135 particles, unwashed) (d = 2.5 μm, CVd = 1.5%, after excluding the distorted spherical particles pointed by the arrows in (a) and to the right of the dashed line in (c)). (b) and (d) SEM image and particle distribution histogram for sample X67 (APTS-coated AA167 particles, unwashed) (d = 2.5 μm, CVd = 1.5%, after excluding the silica nanoparticles pointed by the arrow in (b) and to the left of the dashed line in (d)).

The SEM image for sample AA167 is shown in Figures 6.2 (a) and the particle size distribution histogram of sample AA167 is shown in Figures 6.2 (c). The coating of AA135 particles with silica was confirmed by the obvious increase in the particle size and the decrease in the particle size distribution (d = 2.5 µm and CVd = 1.5 %) compared to AA135 (d = 2.2 µm

and CVd = 3.2 %, see Chapter 4) suggesting the formation of ca. 300 nm-thick silica shell. In the SEM image (Figure 6.2 (a)), one can notice a few distorted spherical particles with a slightly larger diameter (2.7 - 2.8 µm, pointed by the arrows). The formation of this slightly larger particle population can be attributed to the excess strain on the seed particles caused by the formation of the silica shell. If these few distorted spherical particles were included in the

151 SILICA-COATED PARTICLES

calculation of the particle diameter of sample AA167, the average particle diameter would be 2.6

µm with CVd of 2.2 %. In a manner very similar to that used to prepare AA167 sample, the amine-coated particles (sample X67) were synthesized by growing APTS shell on AA167 particles. An aliquot of AA167 sample were washed in ethanol (three cycles of sedimentation by centrifugation at 4000 rpm for 15 min and redispersion) to obtain 3 ml of 1.0% S.C. Then ammonia (29.3 wt % NH3 in water) was added to obtain a 4.2 vol % ethanolic ammonia solution. Immediately afterwards, a 0.5 ml APTS solution (10 vol % in ethanol) was added under stirring (600 rpm), and the reaction mixture was stirred for another 12 h to obtain X67 amino-particles. The SEM image and the particle size distribution histogram of sample X67 are shown in Figures 6.2 (b) and (d). X67 sample has a population of silica nanoparticles and clusters of approximately 100 - 300 nm (pointed by the arrow in Figure 6.2 (b) and to the left from the dashed line in Figure 6.2 (d)). If this population was included in the calculation of the particle

diameter of sample X67, the average particle diameter would be 2.3 µm with very broad CVd of 29.7 %. More realistic numbers were obtained, when I discarded this silica nanoparticle

population (d = 2.6 µm and CVd = 1.5 %). This suggests the formation of ca. 100 nm-thick amine-silica shell. Discarding the small silica nanoparticles is reasonable because of two reasons. As explained earlier, the objective of synthesizing these silica-coated particles is to employ them in mass cytometry-based assays. Since these silica nanoparticles contain no lanthanide, they can be eliminated easily during the mass cytometry analysis. The other reason is that all the bioconjugation reactions include washing steps (cycles of centrifugation, supernatant removal and redispersion). Because of the large weight difference between these nanoparticles and the microspheres, I expect the removal of these nanoparticles with the supernatant during washing.

152 SILICA-COATED PARTICLES

Figure 6.3: Mass cytometry measurements of sample X67 prepared (a) Screen captures of the mass cytometry results for X67 (3 particles were shown each of them has the 7 elements present). (b) Ungated dot- dot plot of the Tm-Ho content of the particles. Gating is represented by the dashed oval. (c) Number of Tm ion per particle gated distribution measured by mass cytometry for a population of PS particles X67.

The metal content for X67 particles was measured by mass cytometry and compared to that of AA135 seed particles. In Figure 6.3(a), I present an image of a screen capture of signals from sample X67 that contains La, Nd, Eu, Tb, Dy, Ho and Tm ions. The dense vertical lines in Figure 6.3(a) refer to signals from multiple mass spectra taken during the transit of a single particle through the plasma torch. This figure shows signals from three successive particles of sample X67. The ungated dot-dot plot for the 169Tm and 165Ho content of the particles is shown in Figure 6.3(b). After gating (represented by the dashed oval in Figure 6.3(b)), the 169Tm distribution of sample X67 was obtained as an example for the lanthanides presentenced in the particle (Figure

153 SILICA-COATED PARTICLES

6.3(c)). The average Tm atom content of sample X67 was 1.8 x 106 169Tm atoms per particle and the coefficient of variation of the Tm content distribution (CVTm) was 26 %. These results are very similar to the mass cytometry results obtained for the AA135 sample (the original seeds of 6 169 X67 which has 1.9 x 10 Tm and CVTm of 25 %, Chapter 4). This confirms that lanthanide- encoded polystyrene particles maintained their lanthanide content after the silica coating process. Another important observation here is that the mass cytometry response for the lanthanide content of the particles was not affected by the silica shell; from analytical chemistry point of view, one might worry that plasma temperature of ICP torch would be affected by the silica content of the particles. If plasma temperature fluctuates, the number of ions released by each particle would change and consequently the response of the mass cytometry to the metal content would have altered.

Figure 6.4: Linker X50, Mw = 588.67 (purchased from Pierce (EZ-Link NHS-PEG4-Biotin))

6.2 Immunoassays Based on the Amino-Functionalized Particles

6.2.1 Streptavidin -Coated Lanthanide-Encoded Particles via Biotin- Streptavidin Sandwich (Fluorescence Assay)

In this section I describe the utilization of X67 amino-particles as Ln-encoded particles for immunoassays. The experiments were carried out by Dr Dirk Weinrich. Commercially available, activated biotin linker X50 (Pierce, NHS-PEG4-Biotin Mw = 588.67, Figure 6.4) was attached to the particles by reacting X67 particles (2.5 mg) with X50 (4.9 mg) in 200 mM phosphate buffer

154 SILICA-COATED PARTICLES

at pH 8.5 for 30 min under stirring in a glass or Eppendorf vial with a magnetic stirring bar (particles X51, Fig. 6.5(a)).

The number of amine groups on sample particles X67 was unknown, but for the purpose of estimating the reaction stoichiometry, it was assumed to be ten times higher than the number of carboxylic acid groups present in the precursor particles (108 per particle, Chapter 2). This estimation was done based on the higher expected density of amine functional groups in the silica coating. Based on the average diameter calculated from SEM images (d = 2.6 µm), we calculated the average mass of X67 particle to be ca. 13x10-12 g per particle assuming the PS density of 1.05

155 SILICA-COATED PARTICLES

and silica density of 2.1 g.cm-3. Accordingly, an amine concentration of 332 µmol/g particles was calculated. Based on this value, 4.9 mg X50 corresponding to ten equivalents per amine group was used to obtain biotinylated-particles (biotinylated-X67, Figure. 6.5(a)). Following washing (by sedimentation at 13,000 rpm, discarding the supernatant and resuspension) in PB buffer, the biotinylated-X67 particles were blocked against non-specific protein adsorption with 1 % BSA. The particles were then incubated with 100 nM Cy5-labelled Streptavidin (SAv-Cy5). The resulted particles were denoted X51 (Figure 6.5(a)). Figure 6.5(b) shows the confocal fluorescence microscopy (CFM) images obtained for X51. The CFM images show homogeneous Cy5 coating that demonstrates the streptavidin binding capability of the biotinylated-X67 particles surface. We found that simple incubation of particles with linker X50 in an Eppendorf vial without stirring led to CFM images showing intensity gradients and different intensities from particle to particle (particles X52, Figure 6.6). This was attributed to inhomogeneous coverage of particles with biotin groups.

Figure 6.6 CFM images of particles X52. Left and middle: overlay of Cy5 channel and optical transmission image, right: Cy5 channel. Particles show inhomogeneous coverage with SAv-Cy5

According to Dr Weinrich, handling of silica-coated PS particles, for bioconjugation reactions, was found to work best in 2.0 ml Eppendorf vials with a slanted bottom. Spinning the particle suspension for 5 seconds at 13,000 rpm was found to pellet particles effectively. The supernatant was then removed with an Eppendorf pipette (200 µl tip). In order to ensure no particle losses, careful monitoring of the withdrawn supernatant was required. A particle

156 SILICA-COATED PARTICLES

suspension volume of 200 µl led to optimal shape of the particle pellet for supernatant removal following centrifugation. Streptavidin has four binding sites. We assume that one or two of these sites will bind to the biotin groups on the particles, which lead to two to three binding sites remaining unoccupied and facing away from the particle surface. This property of SAv enables the assembly of biotin-SAv- biotin sandwiches [20]. Figure 6.7(a) shows the steps for the formation of biotin-SAv-biotin sandwich on the surface of X67 particles. First, SAv particles (X53a/b) were prepared from X67 as described earlier in this section. The biotin binding capability of SAv particles X53a/b was investigated by incubation with a fluorescein-labeled biotin probe. Figure 6.7 (b) shows the formation of fluorescein on the particle surface as revealed by CFM. This indicates successful biotin binding on the surface of SAv particles (X53). Incubation with two different concentrations of SAv (particles X53a: 100 nM and particles X53b: 500 nM) led to an increase in fluorescence signal, suggesting a larger number of biotin binding sites on the particle surface at higher SAv concentration.

Figure 6.7 (a) Workflow for particle samples X53a/b. Streptavidin-coated particles X53a/b were obtained by incubation of biotinylated particles with streptavidin in two different concentrations (X53a: 100 nM, X53b: 500 nM). Bound streptavidin was detected with biotin-fluorescein. (b) CFM images of particles X53a/b. Green fluorescence indicates binding of biotin-FITC (overlay: overlay of optical transmission image and FITC channel).

157 SILICA-COATED PARTICLES

6.2.2 Streptavidin -Coated Lanthanide-Encoded Particles via Biotin- Streptavidin Sandwich (Mass Cytometry Assay)

To determine applicability of the SAv-coated particles (X53) for mass cytometry-based immunoassays, X53 particles were incubated with a metal-containing polymer (MCP) (Figure 6- 8(a)). The MCP used for this experiment was Pr-loaded X54 synthesized by Yijie Lu. Three different amounts of X45 polymer (0.018 mg, 0.0018 mg, 0.0002 mg) were used to examine which concentration would give a better specific binding to the SAv-coated particles (X55 particles, Figure 6-8(a)). For each of the three concentration levels, a negative control sample (with no biotin functionalization) was prepared.

Figure 6.8 (a) Workflow for particle samples X55. Streptavidin-coated particles were incubated with Pr-loaded MCP X54 in three different amounts. Bound MCP was detected via CyTOF analysis (negative control: no biotin/streptavidin). (b) CyTOF result of particles X55.

158 SILICA-COATED PARTICLES

The six samples of X55 particles (3 levels of X45 concentration and their negative controls) were analyzed by mass cytometry to detect the extent of MCP binding to the surface of the particles by monitoring the Pr signal (Figure 6-10(b)). For the highest concentration level of MCP (X54 = 0.018 mg), the Pr signal intensity was 30 counts (correspond to ca. 105 Pr per particle) which means that there were detectable amount of X45 polymer attached to the surface of the X55 particles. Decreasing intensities were observed for the lower MCP concentrations (X54 = 0.0018 mg and 0.0002 mg, Figure 6-8(b)). However, a three- to five-fold higher mass cytometry signal intensities were observed for all three negative controls. The very high signals obtained for negative control samples could be caused by non-specific binding of negatively charged MCP to free, unreacted amine groups on the particle surface, which are positively charged under the conditions of the particle assay (aqueous buffer, pH around 7). A reduction in non-specific MCP adsorption could be achieved by capping of unreacted amine groups on the particle surface. Amine-coated particles (X67) were functionalized with the activated biotin linker as before (as depicted in Figure 6-5 (a)). Then the biotinylated X67 particles were reacted with a large excess of either diglycolic anhydride (DGA, particles X56a) or succinic anhydride (SA, particles X56b) as blocking agents for the excess -NH2 groups on the surface of the biotinylated X67 particles.

The reaction of blocking excess -NH2 was done using DGA (ca. 3.3 M) or SA (ca. 4 M). 50 µl aliquots of 5.0 M NaOH were added under stirring until pH was 12-13 then 50 µl aliquots of 1.0 M HCl was added to pH 5-7. (Figure 6-9(a)). Both SA and DGA were dissolved in DMF prior to addition to the aqueous particle suspension as anhydride hydrolysis occurs on the order of seconds in aqueous solution [21]. Following washing (by sedimentation at 13,000 rpm, discarding the supernatant and resuspension) in HEPES buffer, particles were blocked with 1 % BSA in HEPES buffer. The particle were washed again (by sedimentation at 13,000 rpm, discarding the supernatant and resuspension) in HEPES buffer then functionalized with SAv (500 nM in HEPES buffer). The particles were then incubated with 0.005 mg Pr-loaded MCP X54 (Figure 6.9 (a)) for 18 h at 4 °C. After three cycles of washing (by sedimentation at 13,000 rpm, discarding the supernatant and resuspension) in 200 mM PB buffer pH 7.5, the particles

159 SILICA-COATED PARTICLES

were subjected to mass cytometry analysis. Negative controls for this experiment were done by skipping the functionalization with the activated biotin linker (X50) and following the same subsequent steps (reaction with DGA or SA, blocking with BSA, functionalization with SAv and incubation with MCP X54).

Figure 6.9. (a) Workflow for particle samples X56a/b. After biotinylation, free amine groups on the particle surface were blocked with either DGA (particles X56a) or SA (particles X56b). Streptavidin particles were then created by incubation with streptavidin. Following binding to biotinylated, Pr-loaded MCP X54, bound MCP was detected via CyTOF analysis (negative control: no biotin/streptavidin). (b) CyTOF result of particles X56a/b.

Mass cytometry analysis of particles X56a, X56b and their negative controls are shown in Figure 6.9(b). The X56a sample, prepared with DGA, showed 6-fold Pr signal higher than its negative control while X56b sample, prepared with SA, showed 4-fold Pr signal higher than its negative control. Hence, this blocking strategy, using diglycolic anhydride or succinic anhydride, was found to be successful. A commercially available, negatively charged polymer, poly(acrylic acid) (PAA, M.wt. 4771), was also investigated for its blocking capability against negatively charged MCP X54. DGA-blocked SAv particles were synthesized as for particles X56. Then particles (12.5 µl of 5

160 SILICA-COATED PARTICLES

% w/w) were incubated in PAA solution (85 mg/ml in 50 µl of 0.5 M HEPES buffer) prior to incubation with 0.005 mg Pr-loaded MCP X54 (particles X57). CyTOF analysis showed almost no binding of MCP X54 to the SAv-coated particles. Additionally, no binding of MCP X54 to BSA coated particles (negative control) was detected via CyTOF. A possible explanation is coverage of the particle surface by PAA, which could prevent binding of MCP X54. Hence this strategy was judged to be unsuitable for blocking the unreacted amine groups on the particle surface against negatively charged MCP.

6.2.3 Streptavidin-Coated Lanthanide-Encoded Particles via covalent attachment of Streptavidin (Mass Cytometry Assay)

An alternative strategy to generate SAv-coated particles from amine-functionalized silica- coated particles X67 proceeded via direct covalent binding of SAv to the particle surface without prior functionalization with biotin (Figure 6-10(a)). Amine groups on the particle surface were converted to acid groups using DGA as for particles X56 (particles X58, Figure 6.10(a)). Following activation of the resulting acid groups with EDC/NHS in a large excess and washing (by sedimentation at 13,000 rpm, discarding the supernatant and resuspension) of the activated particles with HEPES buffer (50 mM HEPES, pH7.4) particles were incubated with either SAv (or BSA in the negative control). Unreacted, activated acid groups were subsequently blocked with a large excess of 6-aminocaproic acid. After incubation with 0.005 mg Pr-loaded MCP X54 were submitted for mass cytometry analysis. Figure 6.10(b) shows the results of mass cytometry analysis for X58 sample and its negative control; a very high Pr signal (362 counts) was obtained for X58 particles with only a signal of 15 counts for the negative control sample (ca. 24 fold lower than X58). This was a superior result compared to what had been obtained for other protocols tried such as the one used to prepare particles X56a/b based on a biotin-SAv sandwich. Thus, I concluded using such a protocol (of converting –NH2 group to –COOH, direct SAv attachment via EDC and blocking excess –COOH with 6-aminocaproic acid) is the best suited bioconjugation protocol for my lanthanide encoded particles to be used in immunoassays.

161 SILICA-COATED PARTICLES

Figure 6.10 (a) Workflow for particle samples X58. Particle surface amine groups were converted to carboxylic acids using DGA, which were activated with EDC/NHS and reacted with streptavidin to obtain streptavidin-coated particles X58 (BSA for negative control). Unreacted, activated acid groups were then capped with 6-aminocaproic acid. Following incubation with Pr-loaded MCP X54, analysis was carried out by CyTOF (b) CyTOF result of particles X58.

6.3 Oligonucleotide Assays-Based on the Amino-functionalized Particles

In this section, I describe the possibility of expanding the utilization of the NH2-silica coated particles, which carries a lanthanide code, in other types of biological assays. I decided to try these particles as platform for oligonucleotide assay because such an assay is one of the biological assays that demands higher multiplexity [22], which can be offered by my lanthanide encoded particles. To investigate oligonucleotide binding, X67 particles with covalently bound SAv (or BSA in negative control particles), were synthesized as for particles X58. X67 particles were blocked

162 SILICA-COATED PARTICLES

with 1.0 g/l random herring sperm DNA: 5’-AGG CGC CCA ATA CGA AAA AAA AAA AAA AAA (Biobasic Inc.). This sperm DNA contains a random strand as well as A15) against non- specific DNA adsorption. The word “random” here means that it has a sequence that does not match to any of the probes used in this experiment. Then, the particles were incubated with biotinylated oligonucleotide A (X60: 5’-Biotin-AGCGGATAACAATTTCACACAGGA-3’, which was adapted from Niemeyer et al. [23]) (particles X59, Figure 6.11(a)). After washing (by sedimentation at 13,000 rpm, discarding the supernatant and resuspension) in TBS-E buffer (20 mM TRIS pH 7.5, 150 mM NaCl, 5 mM EDTA), the particles were incubated with either a Cy5- labeled anti-A probe X61 (5’-Cy5-TCCTGTGTGAAATTGTTATCCGCT-3’, the antisense of X60) against oligo A, or a Cy3-labeled anti-B probe X62 (-Cy3- ACCTCAAGTGATCTACCTACCTCAG-3’, the antisense of X63) against a different biotinylated oligonucleotide B (X63, 5’-Biotin-CTGAGGTAGGTAGATCACTTGAGGT-3’), which was not present in this assay (both 1 nM). In Figure 6-11(b), CFM showed only the presence of the Cy5-labeled anti-A probe on the particle surface. This result indicated successful binding of biotinylated oligonucleotide A by the SAv particles, as well as specific detection with the fluorescently labeled anti-A oligo, whereas no non-specific adsorption of the non-matched anti-B probe was observed. The negative control (BSA particles) showed no fluorescence signal at all. This result demonstrated that binding of anti-A probe X61 occurred due to base pairing and not due to non-specific adsorption of the oligonucleotide.

163 SILICA-COATED PARTICLES

Figure 6.11 (a) Workflow for particle samples X59. Streptavidin particles (negative control: BSA particles) were blocked against unspecific DNA adsorption with herring sperm DNA. By incubation with biotinylated oligonucleotide A (X60), oligonucleotide-coated particles X59 were created. Particles X59 were subsequently reacted with either the Cy5-labeled anti-A probe X61 (antisense to oligo X60) or the Cy3-labeled anti-B probe X62, followed by CFM analysis. (b) CFM fluorescence images of particles X59. Cy5 fluorescence on particles S-RA indicates capture of oligo anti-A (X61). Cy3 fluorescence on particles S-RB and NC-RB is almost non-detectable showing absence of oligo anti-B (X62) (overlay: overlay of optical transmission image and Cy5 channel).

Increasing the concentration of the fluorescently labeled probe oligonucleotides from 1 nM to 10, 100 or 1000 nM led to an increase in fluorescence signal (data not shown). Non-specific adsorption of labeled antisense probes was minimal, but could be further suppressed by blocking of the particles with herring sperm DNA prior to incubation with oligonucleotides. In a multiplexing experiment, as depicted by Figure 6.12(a), SAv particles were surface- functionalized with either biotinylated oligonucleotide A (X60, to obtain X65-SA coded particles) or oligonucleotide B (X63, to obtain X65-SB coded particles). Following blocking with herring sperm DNA, each particle type (X65-SA or X65-SB) particles (and their negative controls) were subsequently incubated with a mixture of two fluorescently labeled antisense probes (the mixture was denoted RAB). Namely, this RAB mixture contained X61 (Cy5-labeled anti-A probe, antisense to oligo X60) and X62 (Cy3-labeled anti-B probe antisense to oligo X63) as shown in Figure 6.12(b). X65-SA-RAB and X65-SA-RAB particles were obtained from

164 SILICA-COATED PARTICLES

incubating RAB mixture with X65-SA and X65-SB, respectively. Furthermore, as illustrated in Figure 6.12(c), a mixture of both particle types (denoted X65-SAB which contained X65-SA and X65-SB particles) was incubated with both probes (RAB mixture) to further examine the selective affinity of both particles (X65-SA and X65-SB). The resultant mixture of particles was denoted X65-SAB-RAB.

Figure 6.12 (a) Workflow for particle samples X65. Streptavidin particles (negative control: BSA particles) were blocked against unspecific DNA adsorption with herring sperm DNA. By incubation with biotinylated oligonucleotide A (X60: 5’-Biotin-AGCGGATAACAATTTCACACAGGA-3’) and B (X63: 5’- Biotin-CTGAGGTAGGTAGATCACTTGAGGT-3’) two types of oligonucleotide-coated particles X65 were created. Particles X65 were subsequently reacted separately (b) or combined (b) with a RAB mixture [Cy5-labeled anti-A probe X61 (5’-Cy5-TCCTGTGTGAAATTGTTATCCGCT-3’, antisense to oligo X60) and the Cy3-labeled anti-B probe X62 (-Cy3-ACCTCAAGTGATCTACCTACCTCAG-3’, antisense to oligo X63)] followed by CFM analysis.

165 SILICA-COATED PARTICLES

Figure 6.13 CFM images obtained for particles X65. After incubation with fluorescently labeled anti- A and anti-B probes, particles carrying oligonucleotide A (X60) show only Cy5 fluorescence (top) while particles with oligonucleotide B (X63) show only Cy3 fluorescence (middle). This indicated that anti-A and anti-B probes bind only to their antisense particle-bound oligonucleotides without cross-reactivity. (Cy3/Cy5 overlay: overlay of Cy3 channel and Cy5 channel). In a mixture of both particle types (carrying oligonucleotides A or B), either Cy3 or Cy5 fluorescence was detected (bottom). This showed multiplex capability of the approach.

166 SILICA-COATED PARTICLES

I used confocal fluorescence microscopy (CFM) to examine the attachment of fluorescently labeled antisense probes (RAB mixture: X61 and X63) to the oligonucleotide-functionalized particles (X65SA, X65SB and X65SAB). Figure 6.13 shows the CFM images (right panels) and their corresponding optical images (left panels) for samples X65SA-RAB, X65SB-RAB and X65SAB-RAB. In each of the three samples, the fluorescently labeled antisense probes bound only to their respective particle type with no unspecific adsorption being observed. Negative control BSA particles showed no fluorescence at all indicating no unspecific adsorption of any oligonucleotide.

6.5 Conclusions

I report the surface functionalization of lanthanide-encoded polystyrene microspheres designed for highly multiplexed bioassays like immunoassays and oligonucleotide assays. The PS microsphere samples reported here were synthesized by three stage dispersion polymerization (3-DisP) of styrene in ethanol. They contain a mixture of seven lanthanide ions. An average of more than 106 Ln ions per particle was found for each of the lanthanide metals used with a distribution of 26%. The lanthanide-encoded particles were coated with ca. 300 nm-thick silica shell using the Stöber growth method in ammonia/ethanol solution. Silica-coated particles were further coated with APTS to provide -NH2 functionality on the surface. SEM shows that the -NH2 containing shell has a thickness of about 100 nm.

The -NH2 functionalized particles were examined for different biofunctionalization

strategies. First, using activated biotin linker, Cy5-labelled Streptavidin was attached to the -NH2 functionalized particles. Assuming that one (or two) of the SAv four binding sites would bind to the biotin on the particle surface, three (or two) binding sites would remain unoccupied and facing away from the particle surface. This property of SAv enabled the assembly of biotin-SAv- biotin sandwiches. This sandwich assembly was confirmed by the ability of SAv particles to capture fluorescein-labeled biotin probe as manifested by CFM images. To determine the applicability of SAv-coated particles for mass cytometry based bioassays, conjugation of these SAv particles to biotinylated-MCP were attempted. When Pr-loaded MCP

167 SILICA-COATED PARTICLES

was conjugated to SAv particles, mass cytometry Pr signal intensities were three- to five-fold lower than that observed for the negative controls. This was ascribed to non-specific binding of negatively charged MCP to free, unreacted amine groups on the particle surface. To reduce the non-specific adsorption of MCP to the particle surface, unreacted amine groups on the particle surface were blocked (capped) by excess amount of either diglycolic anhydride (DGA) or succinic anhydride (SA). Mass cytometry analysis of particles prepared with DGA, showed 6-fold Pr signal higher than its negative control. Hence, this blocking strategy was found to be successful. Alternatively, another strategy to generate SAv-coated particles from amine-functionalized silica-coated particles was examined. The strategy proceeded via covalent binding of SAv directly to the particle surface (without biotin linker) through converting amine groups on the particle surface to acid groups using DGA. The EDC/NHS-activated acid groups were then coupled directly with SAv to obtain SAv particles. The Pr-loaded biotinylated-MCP was then attached to SAv particles. Mass cytometry analysis showed a very high Pr signal (362 counts) with only a very weak negative control signal (15 counts). This was a superior result compared to what had been obtained for particles based on a biotin-SAv sandwich. The SAv-coated particles were also used in a CFM-based oligonucleotide assay. The SAv particles were functionalized with either biotinylated oligonucleotide A (5’-Biotin- AGCGGATAACAATTTCACACAGGA-3’) or biotinylated oligonucleotide B (5’-Biotin- CTGAGGTAGGTAGATCACTTGAGGT-3’). Then the ability of these oligonucleotide- functionalized particles to capture fluorescence-labeled oligonucleotide probes was examined. Using a Cy5-labeled anti-A probe (5’-Cy5-TCCTGTGTGAAATTGTTATCCGCT-3’) against oligonucleotide A and a Cy3-labeled anti-B probe (-Cy3- ACCTCAAGTGATCTACCTACCTCAG-3’) against oligonucleotide B, strong specific binding and relatively low non-specific binding were observed with CFM. In order to utilize the novel and rapidly improving mass cytometry technique as an analytical tool for bioassays, future experiments were envisaged to employ lanthanide-encoded oligonucleotide probes for mass cytometry analysis of oligonucleotide binding, eventually resulting in a mass cytometry-based mRNA assay.

168 SILICA-COATED PARTICLES

References:

1. Correa-Duarte MA, Giersig M, Liz-Marzan LM: Stabilization of CdS semiconductor nanoparticles against photodegradation by a silica coating procedure. Chem Phys Lett 1998, 286(5-6):497-501. 2. Ung T, Liz-Marzan LM, Mulvaney P: Controlled method for silica coating of silver colloids. Influence of coating on the rate of chemical reactions. Langmuir 1998, 14(14):3740-3748. 3. Gerion D, Pinaud F, Williams SC, Parak WJ, Zanchet D, Weiss S, Alivisatos AP: Synthesis and properties of biocompatible water-soluble silica-coated CdSe/ZnS semiconductor quantum dots. J Phys Chem B 2001, 105(37):8861-8871. 4. Lu Y, Yin YD, Mayers BT, Xia YN: Modifying the surface properties of superparamagnetic iron oxide nanoparticles through a sol-gel approach. Nano Lett 2002, 2(3):183-186. 5. Graf C, Vossen DLJ, Imhof A, van Blaaderen A: A general method to coat colloidal particles with silica. Langmuir 2003, 19(17):6693-6700. 6. Ahmad T, Rhee I, Hong S, Chang Y, Lee J: Silica-coated Iron-oxide Nanoparticles Synthesized as a T-2 Contrast Agent for Magnetic Resonance Imaging by Using the Reverse Micelle Method. Journal of the Korean Physical Society 2010, 57(6):1545- 1549. 7. Jackson AC, Bartelt JA, Marczewski K, Sottos NR, Braun PV: Silica-Protected Micron and Sub-Micron Capsules and Particles for Self-Healing at the Microscale. Macromol Rapid Commun 2011, 32(1):82-87. 8. Liu YL, Ai KL, Yuan QH, Lu LH: Fluorescence-enhanced gadolinium-doped zinc oxide quantum dots for magnetic resonance and fluorescence imaging. Biomaterials 2011, 32(4):1185-1192. 9. Chen M, Zhou SX, Wu LM, Xie SH, Chen Y: Preparation of silica-coated polystyrene hybrid spherical colloids. Macromol Chem Phys 2005, 206(18):1896-1902. 10. Zhang YH, Chen H, Zou QC: Anionic surfactant for silica-coated polystyrene composite microspheres prepared with miniemulsion polymerization. Colloid Polym Sci 2009, 287(10):1221-1227. 11. Hwang DR, Hong J, Hong CK, Shim SE: Synthesis of Positively Charged Silica- Coated Polystyrene Microspheres via Dispersion Polymerization Initiated with Amphoteric Initiator. J Dispersion Sci Technol 2010, 31(2):155-161. 12. Stober W, Fink A, Bohn E: Controlled growth of monodisperse silica spheres in micron size range. J Colloid Interface Sci 1968, 26(1):62-&. 13. Iler RK: US Patent. In. 2885366; 1959. 14. Iler RK: The chemistry of silica: Wiley: New York; 1979. 15. Liz-Marzan LM, Giersig M, Mulvaney P: Homogeneous silica coating of vitreophobic colloids. Chem Commun 1996(6):731-732. 16. Yang Z, Cong H, Cao W: Narrowly dispersed micrometer-sized composite spheres based on diazonium–polystyrene. J Polym Sci, Part A: Polym Chem 2004, 42(17):4284-4288.

169 SILICA-COATED PARTICLES

17. Smith JN, Meadows J, Williams PA: Adsorption of Polyvinylpyrrolidone onto Polystyrene Latices and the Effect on Colloid Stability. Langmuir 1996, 12(16):3773- 3778. 18. Hong J, Han H, Hong CK, Shim SE: A direct preparation of silica shell on polystyrene microspheres prepared by dispersion polymerization with polyvinylpyrrolidone. J Polym Sci, Part A: Polym Chem 2008, 46(8):2884-2890. 19. Van Blaaderen A, Van Geest J, Vrij A: Monodisperse colloidal silica spheres from tetraalkoxysilanes: Particle formation and growth mechanism. J Colloid Interface Sci 1992, 154(2):481-501. 20. Muller W, Ringsdorf H, Rump E, Wildburg G, Zhang X, Angermaier L, Knoll W, Liley M, Spinke J: Attempts to mimic docking processes of the immune system: recognition-induced formation of protein multilayers. Science 1993, 262(5140):1706- 1708. 21. Bunton CA, Fuller NA, Perry SG, Shiner VJ: The hydrolysis of carboxylic anhydrides. Part III.* Reactions in initially neutral solution. Journal of the Chemical Society (Resumed) 1963:3028-3036. 22. Eastman PS, Ruan W, Doctolero M, Nuttall R, de Feo G, Park JS, Chu JSF, Cooke P, Gray JW, Li S et al: Qdot Nanobarcodes for Multiplexed Gene Expression Analysis. Nano Lett 2006, 6(5):1059-1064. 23. Wacker R, Niemeyer CM: Synthesis of Covalent Oligonucleotide-Streptavidin Conjugates and Their Application in DNA-Directed Immobilization (DDI) of Proteins. Current Protocols in Nucleic Acid Chemistry 2005:UNIT 12.17.

170 Analytical Aspects

7 Analytical Aspects In this chapter, I examine the suitability of metal-containing polystyrene microspheres for the calibration of a mass cytometery instrument, a single particle analyzer based on an inductively coupled plasma ion source and a time-of-flight mass spectrometer. These metal- containing microspheres are also verified for their use as internal standards for this instrument. These microspheres were synthesized by multiple-stage dispersion polymerization with acrylic acid as a comonomer. Acrylic acid acts as a ligand to anchor the metal ions within the interior of the microspheres. The reproducibility of this synthetic methodology is also examined. It is important to mention that this chapter does not contain any calibration method or internal standardization procedure. Rather, in this chapter, I examine the suitability of these microspheres, for such analytical purposes. Most of the results given in this chapter were published in Ref [1].

7.1 Introduction

Immunophenotyping is a cellular analysis methodology for the identification of biomarkers using fluorescently conjugated affinity reagents. This is one of the key advances in medicinal research, which employs antibodies that are reactive against cell antigens to distinguish specific subsets within a heterogeneous mixture of cells. In such an assay, the quantification of a subset of interest can be readily accomplished on a cellular level by the use of a flow cytometer. To detect the presence of a cell-bound monoclonal antibody by flow cytometry, the antibody must be coupled either directly or indirectly to a fluorescent tag (e.g. fluorescent dyes or quantum dots).[2-9] The differentiation of fluorophores can be achieved by detection of their different emission spectra. Quantification of these fluorescence signals provides the determination of the amount of fluorophore bound to a particular cell. With the use of appropriate standards, the amount of detected fluorophore can be used to estimate the number of antigens targeted by the fluorescence-tagged antibody. In many cases, fluorescence standards are calibration microspheres which contain fluorophore(s) with sharp signals that can be used to generate a calibration curve.[10-14] Much

171 Analytical Aspects

less frequently, these microspheres can be mixed with the sample of interest for use as an internal standard.[15, 16] The main limitation of fluorescence-based immunoassays, however, is the low variability in both the type and number of different fluorophores with emission intensities that can be adequately resolved to allow for simultaneous detection. In addition to this limitation, the broadness of emission spectra and the different excitation wavelengths of different fluorophores further complicate such measurements. Some of these problems can be mitigated by using quantum dots with very narrow size distributions, which give significantly narrower emission bands.[17] However, the limited linear dynamic range of fluorescence-based assays makes quantitative analysis using flow cytometry challenging. This challenge is particularly apparent during the analysis of samples that consist of analytes that differ in concentration by more than an order of magnitude. A much larger amount of information can be obtained using different metal atoms or isotopes as labels, coupled with their detection by atomic mass spectrometry.[18, 19] Metal- encoded microspheres coupled with inductively coupled plasma mass spectrometry detection opens the door to multiplexed analyses that can differentiate over an order of magnitude more unique labels than what is possible with assays based upon luminescence detection.[20-22] Expanding on the concept of the fluorochrome microsphere assays of conventional flow cytometry, I have incorporated different metals into polymer microspheres at different levels of concentration,[23] with microsphere-by-microsphere detection and readout based upon mass cytometry instrumentation.[24-26] The quantification of data possible via mass cytometry provides an absolute determination of the number of ions in the sample, rather than the relative or semi-quantitative processes used in conventional flow cytometers. It is worth noting that the behavior of different elements in an Ar plasma torch depends on their physical properties, such as atomic mass and ionization efficiency but these variations are constant for a robust plasma. In order to effectively correct for temporal variations in signal intensity (usually caused by the instrument tuning and drift), the use of standard calibration microspheres is important. The calibration will account for the variations between different mass cytometer instruments and for the slight signal drift during mass spectra acquisition. Internal standard microspheres which are added directly to the samples in known quantity should have known concentrations of metal(s) with atomic masses and ionization efficiencies close to those of the metal tag.[27] Also, the

172 Analytical Aspects

efficiency of the sample introduction system can be assessed by the number of registered microsphere events per acquisition time. This number can be used to estimate the number of cells in the sample if the efficiency of the sample introduction system is less than perfect. In this chapter, I describe the synthesis and the possibility of utilizing of metal-containing polymer microspheres as standard microspheres for mass cytometry instrument calibration and as internal standards for cell analysis. I use synthetic methods similar to those that I have described in Chapter 3 and Chapter 4 for the synthesis of lanthanide-encoded microspheres for mass cytometry-based immunoassays. The focus here is on exploring the upper detection limit of mass cytometry, as well as optimizing the lanthanide content of the microspheres to fall in the detection range of the instrument. In addition, I show proof of principle experiments in which microspheres are mixed with different cell lines and subjected to mass cytometry. Finally, I examine the reproducibility of the synthesis of lanthanide encoded microspheres by dispersion polymerization.

7.2 Results and Discussion

For metal-containing microspheres to be used as internal standards or calibration standards in mass cytometry, four major requirements must be met. Two of these requirements are similar to the requirements, mentioned in Chapter 3, for the PS microspheres to be used as platforms for mass cytometry based bioassays. First, the microspheres must be large enough to be easily injected into the mass cytometer on a microsphere-by-microsphere basis, but small enough to guarantee complete atomization and ionization of the microspheres in the ICP torch. Polystyrene microspheres with diameters in the range of 0.8 to 3.0 µm satisfy these requirements. Microspheres of this size are also convenient to manipulate in terms of washing and redispersing. I characterize particle size in terms of the mean particle diameter (d) obtained from the analysis of SEM images. These metal-containing microspheres should also have a very narrow size distribution. I characterize the size distribution in terms of coefficient of variation of the particle diameter (CVd).

1 1 2 = =1( i av ) (7-1) av 1 𝑛𝑛 𝐶𝐶𝐶𝐶𝑑𝑑 𝐷𝐷 �𝑛𝑛− ∑𝑖𝑖 𝐷𝐷 − 𝐷𝐷

173 Analytical Aspects

where Dav is the number average diameter of all particles, Di is the diameter of the i'th particle, and n is the total number of particles counted in the analysis. Second, the metal-containing microspheres must have a very small microsphere-to- microsphere variation in lanthanide content. I characterize the lanthanide content by determining the average number of Ln atoms per microsphere. Ln variation from microsphere-to-microsphere

will be evaluated by the magnitude of the Ln coefficient of variation (CVLn), which is obtained from the mass cytometry measurement. This value is defined by an expression analogous to that

in eq (1). I expect that CVLn ≥ CVV, the coefficient of variation of the particle volume. At this moment I do not have a clear way to differentiate between measurement and content contributions to CVLn. There are two other requirements for Ln-encoded PS microspheres to be used as internal standards or calibration standards in mass cytometry technique. First, each microsphere must have a lanthanide-content close to the middle of the detection range of the mass cytometer. I will show how I define this concentration and how I designed the polymer microsphere synthesis to meet this requirement. Second, metal-containing microspheres to be used as calibration standards should retain their lanthanide content in fluid media during prolonged storage as well as during experimental assays. This will be gauged by determining Ln release behaviour from the particle interior as a function of time.

174 Analytical Aspects

Table 7.1. Recipes for the dispersion polymerization of styrene with PVP55 and acrylic acid in ethanol

Two-stage reactions Three-stage reaction

Materials (g added) 1st stage 2nd stage 1st stage 2nd stage 3rd stage Styrene 6.25 -- 6.25 -- --

PVP55 a 1.0 -- 1.0 -- --

b TX305 0.35 -- 0.35 -- --

c AMBN 0.25 -- 0.25 -- -- Ethanol 18.75 18.75 18.75 10.0 10.0

Acrylic acid -- 0.125 -- 0.125 0.125

d d LnCl3 ------

EGDMA ------0.063

a polyvinylpyrrolidone, M ≈ 55 kDa. b Triton X-305 c azomethylbutyronitrile d AA070-Tm: 6.3 mg TmCl3.6H20; AA086-Tm: 0.63 mg TmCl3.6H20, and AA120-Eu: 0.63 mg EuCl3.6H20.

175 Analytical Aspects

Table 7.2. Particle size, size distribution, and the variation of Tm and Eu content for some PS-PAA microspheres synthesized in the presence of LnCl3.6H2O.

a a b Incorporation d e Sample d ± σd (μm) CVd% Ln loading efficiency (%)c Ln measured CVLn% AA070-Tm 2.1 ± 0.052 1.8 1.00 102 8.0 x 107 25 AA086-Tm 1.9 ± 0.066 1.9 0.10 129 1.1 x 107 37 AA120-Euf 2.2 ± 0.035 1.4 0.10 119 1.2 x 107 14 AA137 Af 1.99 ± 0.042 2.1 0.25 126 5.5 x 106 18 AA137 Bf 1.65 ± 0.033 2.0 0.25 122 2.2 x 106 17 AA137 Cf 1.61 ± 0.025 1.6 0.25 117 2.0 x 106 15 AA137 Df 1.65 ± 0.041 2.5 0.25 125 2.6 x 106 22 AA137 Ef 1.65 ± 0.041 2.5 0.25 124 2.4 x 106 23

a d = mean diameter; σd = one standard deviation; CVd = coefficient of variation of the diameter. b Ln loading: wt %/styrene of LnCl3.6H2O based upon styrene used in the particle synthesis, see Table 1. c Incorporation efficiency: % of Tm (Eu in AA120-Eu) atoms in the reaction that is incorporated into the microspheres calculated from mass cytometry measurements. d The number of Tm (Eu in AA120-Eu) atoms per microsphere from the mass cytometry intensity and calculated by Equation 7-2. e CV of 169Tm and 153Eu intensity measured by mass cytometry. f Synthesized by 3-stage dispersion polymerization.

7.2.1. Microsphere Synthesis and Metal Incorporation

For the synthesis of lanthanide-containing polymer microspheres, I employed the same technique described in Chapter 3 and Chapter 4 for the synthesis of monodisperse PS-co-PAA microspheres. The synthesis is based on the dispersion polymerization of styrene in ethanol in the presence of polyvinylpyrrolidone (PVP).[23] In this synthesis strategy, the addition of the acrylic acid is delayed until the particle nucleation step is complete (a few percent monomer conversion). I refer to this method as “two-stage” dispersion polymerization (denoted 2-DisP). Thus in my design, I initiate dispersion polymerization of styrene in ethanol, and after approximately 10% monomer conversion, add known amounts of LnCl3 in ethanol in the presence of an excess of acrylic acid (AA) that serves as a ligand for Ln3+. The carboxylate group of AA is known to interact strongly with lanthanide ions. The reagents employed in these

2-DisP reactions and the amounts of LnCl3 salt used are presented in Table 7-1. The

176 Analytical Aspects

characteristics of the particle samples synthesized are collected in Table 7-2. The reactions were clean, and yielded particles in 90 – 95 % gravimetric yield, with no coagulum. I was first interested in testing the range of lanthanide ion concentrations that can be studied by mass cytometry The current research prototype mass cytometry instrument [24, 28] provides an ion transmission efficiency of T ≈ 5x10-5, which yields a sensitivity on the order of 200 counts per second per part per trillion. To see if high loadings of metal ions would saturate the detector, I designed a synthesis that would yield microspheres containing ca. 108 copies of an individual isotope. I selected 169Tm as a mono-isotopic element, and synthesized the AA070-Tm sample,

described in Table 7-1, in which 1.0 wt%/sty TmCl3 and 2.0 wt%/sty AA were added in the second stage. The incorporation efficiency of the lanthanide ions is the percentage of lanthanide ions added in the reaction that were incorporated into the microspheres. According to the high lanthanide ion incorporation efficiency obtained in the syntheses described in Chapter 3 and Chapter 4, I based my calculations on the assumption of complete encapsulation of Ln ions into the polymer microspheres. Figure 7-1 shows a scanning electron microscope (SEM) image of sample AA070-Tm. The sample has mean a diameter d = 2.1 μm, with a narrow size distribution

(CVd = 1.7 %).

5 µm

Figure 7.1. SEM images for PS microsphere samples AA070-Tm synthesized in the presence of 1 % TmCl3 added in the second stage with AA: 2 wt %/styrene (d = 2.1 μm, CVd = 1.8%).

177 Analytical Aspects

To examine the metal ion (169Tm) content of sample AA070-Tm, this microsphere dispersion was washed by three cycles of centrifugation and resuspension in water. The resultant slurry (ca. 106 microspheres/mL) was nebulised into the mass cytometer sample introduction system, which in turn delivered microspheres individually but stochastically into the inductively coupled plasma torch. The high temperature of the plasma was sufficient to vaporize, atomize and then ionize the microspheres and the Tm ions embedded in them. The ion stream was then introduced into the time-of-flight mass analyzer. The transient signals corresponding to each microsphere ionization event were recorded by the detector and stored. The histogram representation of the frequency distribution of the integrated ion intensity over the transient signal for individual microspheres gives quantitative information about the metal ion content of the microsphere population. In Figure 7.2, the population distribution is presented for the 169Tm ion signal collected for 3 min (ca. 4 x 105 microspheres) for AA070-Tm. The x-axis of this plot is the intensity analog output of the TOF detector and is considered here as a relative number. The mean 169Tm intensity is 81700. The average number of metal ions per microsphere can be calculated from the mean intensity values as follow:

× = F (7-2) 𝐼𝐼 𝐼𝐼 where I is the mean intensity measured by the𝑁𝑁 TOF detector;𝑇𝑇 IF is the analog-to-count conversion factor for a particular ion (related to the mass response); and T is the transmission coefficient of the entire instrument and corresponds to the number of ions that reach the detector per number of ions injected. T value depends on tuning and the mass response of the instrument. The tuning of mass cytometry instrument is done daily using a standard solution that contains different lanthanide ions that covers the lanthanide series mass range (La, Tb and Tm at 0.5 ppb w/w). Using the mass response (number of counts from the mass cytometry detector) to the known concentration of the ions, the transmission coefficient (T) is then calculated for each ion in the standard solution. For the lanthanide ions that are not included in the standard solution (like Eu ions in AA120-Eu sample), T is assumed to be the same as that for Tb, the nearest ion that exists in the standard solution. As introduced in Chapter 3, the incorporation efficiency is the ratio between the amount of lanthanide ions measured in the microspheres to the amount added

178 Analytical Aspects

in the synthesis. Table 7.2 shows the incorporation efficiency values for some of the lanthanide- encoded samples described in this chapter. The incorporation efficiency here is calculated from the mass cytometry measurements and equation 7.2 (rather than by traditional ICP-MS described in Chapter 3). One can notice the too high incorporation efficiency values obtained for most of the samples. This result indicates that there is a problem in the way the transmission coefficient is obtained for measurements carried out by polymer beads. This problem will be discussed later in this chapter.

Figure 7.2. Distribution of mass cytometry intensity signal for a population of PS microspheres (AA070-Tm) prepared by 2-DisP in presence of TmCl3 (1.0 wt%/styrene) and AA (2.0 wt%/styrene). The dashed-line indicates the sharp cutoff at ca. 105 169Tm signal intensity that this is likely due to detector saturation.

169 -5 For the case of AA070-Tm, the value of IF was 0.04 for ( Tm) and T was 3.9 x 10 . The average Tm content of AA070-Tm was calculated to be 0.8 x 108 169Tm ions per microsphere. This level of loading is close to the saturation (upper limit) of the detector as manifested by the sharp drop in the Tm integrated intensity distribution (the right side of the distribution in Figure

7-2). The width of the Tm intensity histogram showed CVTm > 25%. These results are similar to the ones I obtained previously for Eu-containing microspheres.[23] In fact, detector saturation in 169 the case of sample AA070-Tm indicates that the Tm measured intensity and CVTm values are in error and likely different from the actual values. I conclude that lanthanide-containing polymer microspheres, synthesized at this high level of Ln ion content, do not meet the requirements for standard microspheres.

179 Analytical Aspects

The next logical step was to decrease the amount of TmCl3 salt added in the second stage of the 2-DisP. Sample AA086-Tm was prepared in the presence of 2.0 wt%/sty AA and 0.1 wt%/sty

of TmCl3, a factor of ten lower than that employed in the synthesis of sample AA070-Tm. The

lanthanide-containing polymer microspheres produced were also monodisperse (CVd = 1.9 %) with a size similar to that of AA070-Tm (d = 1.9 µm). Figure 7.3 shows the distribution of the integrated signal intensity for sample AA086-Tm. From the distribution average value, AA086- Tm microspheres were found to have ca. 1.1 x 107 169Tm atoms per microsphere. This number of

atoms is adequate for mass cytometry measurements. However, CVTm was even higher (ca. 37 %) than AA070-Tm and these values are too large for these microspheres to be used as mass cytometry standards. It is known from my previous experiments Chapter 3 that the Ln content and Ln distribution values are independent of the type of Ln metal used in the 2-DisP microsphere synthesis. For example, very similar results were obtained when 0.1 wt%/sty of 7 159 TbCl3 was used instead of TmCl3 (0.9 x 10 Tb ions per microsphere and CVTb of ca. 41 %). See Figure.S2 in the Supporting Information for the mass cytometry measurement of AA083-Tb.

Figure 7.3. Distribution of mass cytometry intensity signal for a population of PS microspheres (AA086-Tm) prepared by 2-DisP in presence of TmCl3 (0.1 wt%/styrene) and AA (2.0 wt%/styrene)

7.2.2. 3-Stage Dispersion Polymerization

In Chapter 4, I showed that adding more acrylic acid plus a small amount of cross-linking agent to a dispersion polymerization reaction after more than half of the styrene was consumed

180 Analytical Aspects

led to particles with a narrower distribution of Ln ions per microsphere. I refer to this synthesis strategy as 3-stage dispersion polymerization (3-DisP). I used a similar protocol here in an attempt to synthesize microspheres suitable for mass cytometry calibration. The recipe is presented in Table 7-1, and the characteristics of sample AA120-Eu prepared in this way are listed as the third entry in Table 7.2. A scanning electron microscope image of these particles is presented in Figure 7-4. The key result for this methodology is that the overall particle size is similar to that of sample AA070-Tm, and with a very narrow size distribution. Figure 7-5 shows the mass cytometry distribution of AA120-Eu. Although the Eu atom content of AA120-Eu (calculated for 153Eu from Figure 7-5, ca. 1.1 x 107) is similar to Tm content of AA86-Tm, there is a substantial improvement in the microsphere-to-microsphere variation in lanthanide content per particle. For this sample, CVEu is 14%. This value is very similar to values for La and Tm found for sample AA105 whose synthesis was reported in Chapter 4. For the mass cytometry measurement of sample AA105, see Figure 4-13 in Chapter 4.

Figure 7.4. SEM image for PS microsphere samples AA120-Eu synthesized by 3-DisP in the presence of 0.1 % EuCl3 added in the second stage with AA: 4 wt %/styrene (d = 2.2 μm, CVd = 1.4%) The scale bar is 5 µm.

Figure 7.5. Distribution of mass cytometry signal intensity for a population of PS microspheres (AA120-Eu) prepared by 3-DisP in presence of EuCl3 (0.1 wt%/styrene) and AA (4.0 wt%/styrene). CVEu = 14 %.

181 Analytical Aspects

In Table 7.2, I noted a problem in the amounts of lanthanide ions present in the samples as calculated from mass cytometry measurements. For example, sample AA120-Eu was calculated to have too high Eu ions per particle (1.2 x 107 Eu ions per particle, which correspond to 119 % incorporation efficiency). To investigate this problem, the AA120-Eu microspheres as well as their supernatant were examined by ICP-MS for their Eu content. Freeze-dried microspheres was microwave-digested in concentrated HNO3 according to a reported protocol [29]. This protocol is described in Chapter 2 and summarized in Table 7.3. The detailed supernatant collection steps and sample preparation for ICP-MS analysis are also described in Chapter 2. The ICP-MS results show no detectable Eu in the supernatant of sample AA120-Eu. For the digested AA120-Eu microspheres, 2.9 µmol Eu / g polymer were found [correspond to 7.6 x 106 (± 15 %) ions per microsphere]. This Eu-content value is about 25 % less than that obtained by mass cytometry [1.2 x 107 (± 14 %) ions per microsphere]. This means that the transmission factor T in equation 7.2 is underestimated for the mass cytometry calculations. That explains the too high incorporation efficiencies reported in Table 7.2. We do not yet understand the reason for the difference in T values between standard solutions and polymer bead samples. It may be related to differences in the behavior of the tuning homogeneous solution and the heterogeneous bead suspension in the plasma. Investigating the origin of this difference is one of the interesting points for future research.

Table 7.3. Microwave digestion system: digestion program. This is table is reproduced from ref. 31 with permission. STEP Time (min) Power (W) 1 3 400 2 1 0 3 5 400 4 1 0 5 2 400 6 1 0 7 10 400 8 30 Allow to cool to room temperature

7.2.3. Ion-release behavior and mass cytometry calibration standard

Metal-containing microspheres synthesized by 3-DisP appear to retain their lanthanide content in aqueous solutions of different pH values, as explained in Chapter 4. That chapter

182 Analytical Aspects

described experiments carried out over a three-week time frame. For metal-containing microspheres to be suitable for mass cytometry instrument calibration, the metal content of the microspheres should be examined over longer storage times. Consequently, I decided to repeat the Ln ion-release study on sample AA120-Eu over a six-month period, which I believe is sufficient to judge their Ln-release behavior. I examined the stability of the microspheres in a range of aqueous media that are commonly utilized during cell analysis. I used conventional inductively coupled plasma-mass spectrometry solution analysis to follow the leakage of Ln ions (153Eu in case of AA120-Eu) into the aqueous medium as a function of time. Experiments were performed on sample-Eu 120 synthesized by 3- DisP. This sample contains 260 ppm Eu ion (w/w based on polystyrene). The results presented in Figure 7.6 are for 0.5 wt% microspheres in suspension in three aqueous solutions buffered at pH 3.0 (50 mM sodium acetate), 7.0 (10 mM ammonium acetate) and 10.6 (200mM sodium carbonate/bicarbonate). At pH 7, there is no measurable release of Eu3+ from the AA120-Eu microspheres over the entire period of study.

Figure 7.6. 153Eu ion release into the aqueous phase from colloidal suspensions of PS microsphere sample (0.5 % solids content of AA120-Eu, synthesized by 3-stage DisP) in three different buffer solutions. Microspheres contain 260 µg/L Eu ion (w/w styrene). pH 10.6 buffer solution: 200mM sodium carbonate/bicarbonate, pH 7.0 buffer solution: 10 mM ammonium acetate and pH 3.0 buffer solution: 50 mM sodium acetate. The right-hand y-axis represents the percentage of 153Eu ion released into the aqueous phase to the number of 153Eu-content of the microspheres

183 Analytical Aspects

After 6 months at pH 3, sample AA120-Eu showed a minimal increase of the Eu3+ ions in the medium (ca. 0.3 µg/L). At pH 10.6, a small amount of Eu3 ions was released into the continuous medium over the first two days to reach an Eu3+ concentration of 1.1 µg/L. This is about four times higher than that of pH 3, but corresponds to less than 0.01% of the Eu content of the microspheres. The release profile leveled off after two days, and no further loss of ions to the aqueous phase could be detected over the next six-months. Thus, I conclude that leakage of embedded Ln ions into the aqueous medium is unlikely to be a source of problems in using my lanthanide-containing microspheres for the calibration of the mass cytometry instrument.

7.2.4. Lanthanide-Containing Microspheres as Internal Standards for Cell Samples.

Here I consider another way of presenting the mass cytometry results; through bivariate or scatter plots. The bivariate plots graph the relationship between two variables (isotopic concentrations) that have been measured on a single microsphere. Such plots permit us to see at a glance the degree and pattern of the relation between the two isotopes in the sample. Most importantly, using bivariate plots, one can group the populations of the microspheres that have similar contents of different isotopes. On a bivariate plot, the x- and y-axes can represent the concentration of any two isotopes of interest. Each point on the plot shows the x and y isotopic- content for a single microsphere.

Figure 7.7. Examples of bi-variant plots of mass cytometric results for (A) KG1a free cells stained with CD34-169Tm and Ir-interchelator and (B) free AA120-Eu microspheres.

184 Analytical Aspects

Figure 7.7 shows the two-dimensional projections (a bivariate plot)[30] of some multidimensional data sets obtained as a result of the mass cytometry experiment. Figure 7.7A. presents a logarithmic 151Eu/153Eu bivariate plot for the AA120-Eu sample. This is another way of presenting the same data used to construct Figure 7.5. Although the microspheres show different intensities of Eu content ranging from ca. 103 up to 3 x 104 for both isotopes, the vast majority (> 90%) of the microspheres exhibit a very tight distribution of intensities around ca, 8 x 103. This behaviour in the bivariate plot of AA120-Eu sample reflects the characteristic feature of microspheres prepared by 3-Disp, which is their low microsphere-to-microsphere variability of the metal (Eu) content. For comparison, I examined, by mass cytometry, the metal content of KG1a, a model human acute myelogenous leukemia cell line. These cells were fixed and then treated with an iridium intercalating agent that is taken up in amounts that reflect the DNA content of the cells. In Figure 7.7B I present the 191Ir/193Ir bivariate plot for these cells. The broad area associated with the 191Ir/193Ir projection manifests the wide distribution of Ir-content of the KG1a cells. This type of result is expected from any cell line due to the variability of DNA content in a given cell sample. The main goal for this work was to examine the usefulness of lanthanide-containing polymer microspheres as internal standards for cell line measurements by mass cytometry. Accordingly, I need to test the effect of mixing microspheres and cells on the mass cytometry response. In a proof of concept experiment, a sample of AA120-Eu microspheres was mixed with a suspension of KG1a cells. This cell line is well known to have a high level of CD34 antigen expression. Therefore, we separately stained its surface with different clones of CD34 monoclonal antibodies, which in turn were independently tagged with169Tm in the form of ions bound to a metal chelating polymer. [27] I denote these tagged antibodies uniformly as CD34-169Tm because the observed difference in the staining efficiency is outside of scope of this thesis. In addition, the KG1a cells were fixed and stained with the Ir-intercalator.[22] The 193Ir/153Eu bivariate projection for mass cytometry data obtained for a 100:1 mixture of KG1a cells and AA120-Eu microspheres is shown in Figure 7.8. The cell-like events are Ir- positive and Eu-negative and appear in the upper left corner of the plot. The microsphere-only events are Eu-positive and Ir-negative and appear in the lower right corner of the plot in Figure

185 Analytical Aspects

7.8. In addition, there is a small population of events (ca. 2.2%) that show a Eu-positive and Ir- positive response in the upper right corner of the plot. This signal must correspond to microspheres and cells that interact or pass simultaneously through the plasma torch of the ICP mass spectrometer. This result is unexpected because the metal-containing PS microspheres are coated with a corona of polyvinylpyrrolidone that suppresses protein adsorption to the microspheres.[31]

Figure 7.8. A bi-variant plot of mass cytometry results for: 100:1 mixture of cells and microspheres. Colored points represent the Ir- and Eu-positive events.

To check if the CD34-169Tm antibody was responsible for this interaction, I mixed AA120- Eu microspheres with a different cell line, U937 human leukemic monocyte lymphoma cells that have no CD34 antigen and were fixed and stained only with the Ir-intercalator. I observed very similar behaviour with the U937 cell line. This behaviour can be seen as a kind of “association” between the cells and metal-containing microspheres that occurred either during sample preparation or because of association as the samples entered the plasma torch. I do not yet have a clear explanation for this behaviour. Nevertheless, this interaction affects only a small fraction of the cells, and there is no significant difference between the metal-content of cells measured in the

186 Analytical Aspects

presence and absence of metal-containing microspheres, as can be estimated roughly by comparing the 193Ir intensities in Figure 7.7B and Figure 7.8. More mass cytometry results for 12 replicates of each KG1a (K1-K12) and U937 (U1-U12) cell lines are presented in Table 7.4. In this table, each row represents one sample whose 169Tm- and 193Ir-contents were determined from two distributions. First, the data for a cell population was gated in the absence of AA120-Eu microspheres (column head: free cells). Second, data were obtained for another population gated together with AA120-Eu microspheres (column head: cells and microspheres). The column on the right-hand side of Table 7-4 represent the signal intensity associated with the 153Eu-content of the AA120-Eu microspheres gated in the absence of cells. The data in Table 7-4 show that the 169Tm- and 193Ir-content of both KG1a and U937 cells were independent of the presence of the metal-containing microspheres. Figure 7.9 shows a plot of some of the mass cytometry data from Table 7-4. The y-axis presents signal intensities for 169Tm- and 193Ir for KG1a cell lines measured by mass cytometry in the presence of AA120-Eu microspheres. These values are plotted against the signal intensities for 169Tm- and 193Ir for KG1 cells measured in the absence of AA120-Eu microspheres. The best- fit line drawn through the data has a slope of 1.0 and is passes through the origin. Accordingly, I conclude that the presence of the AA120-Eu microspheres does not affect the mass cytometry response to the metal content of the cells. Thus the metal-containing microspheres can be used as an internal standard in mass cytometry measurements on cell samples.

Figure 7.9. Comparison between the 193Ir (from DNA intercalator) and 169Tm (from 169 CD34- Tm) averages of integrated ion intensities over the transient signals for KG1a cells gated alone (x-axis) and in a mixture with AA120-Eu microspheres (y-axis). 187 Analytical Aspects

Table 7.4. Averages of the integrated ion intensities over the transient signals for individual KG1a , U937 cells, and AA120-Eu microspheres. Free Cells Cells and microspheres Free Samples 169Tm 193Ir 169Tm 193Ir 153Eu ih153Eu

K1 33000 49000 33000 46000 8000 7000 K2 62000 44000 63000 43000 7000 7000 K3 71000 44000 72000 49000 7000 7000

K4 17000 48000 17000 47000 8000 7000 Tm Ir interchelatorIr 169

- K5 42000 48000 41000 48000 7000 7000 169 K6 59000 48000 60000 50000 7000 7000 K8 36000 49000 36000 48000 8000 7000

and CD34 K9 44000 47000 39000 52000 8000 7000 K10 16000 51000 16000 50000 8000 7000 K11 28000 51000 27000 50000 8000 7000 KG1a stained with with stained KG1a K12 30000 48000 30000 52000 8000 8000

U1 160 51000 166 51000 10000 9000 U2 190 51000 201 49000 9000 9000 U3 280 50000 278 53000 9000 9000 U4 460 52000 454 51000 9000 9000

Ir interchelator U5 3300 51000 3697 56000 9000 9000 169 U6 1900 55000 434 50000 9000 9000 U7 530 50000 772 50000 9000 9000 U8 590 50000 531 50000 10000 9000 U9 1700 50000 1787 50000 10000 9000 U10 390 51000 411 54000 9000 9000 U11 280 53000 324 53000 9000 8000

U937 stainedU937 with only U12 460 52000 771 46000 9000 8000 12 different clones of CD34 antibodies: K (KG1a cells stained with 169Ir intercalator and Tm-labeled CD34) and U (U937 cells stained only with 169Ir intercalator). Each sample was examined in a 100:1 mixture with metal-containing microspheres AA120-Eu. The Eu content of the AA120-Eu gated without cells is reported on the last column.

188 Analytical Aspects

7.2.5. Reproducibility of the Synthesis of Lanthanide-Containing Microspheres.

In this section, I examine the reproducibility of 3-DisP synthetic mythology in terms of size, and metal content of the synthesized microspheres. For this sake, I synthesized 5 samples with exactly the same formula (AA137 A-E) as explained in Table 7.5. Each of these samples was loaded with five lanthanide salts (La, Tb, Eu, Ho and Tm each at the concentration of 0.05 wt% / styrene). The lanthanide salts were added in the second stage of the dispersion polymerization, one hour after initiation, along with acrylic acid (2.0 wt% / styrene). In the third stage, eight hours after the second stage, I added more acrylic acid (2.0 wt% / styrene) as well as EGDMA (1.0 wt% / styrene) as a crosslinker.

Table 7.5. The recipe for the synthesis of microsphere samples AA137 A-E by 3-stage dispersion polymerization (3-DisP) of styrene with PVP55 as a dispersant in ethanol. Materials (grams added) 1st stage 2nd stage a 3rd stage b Styrene 6.25 -- -- PVP55 1.0 -- -- TX305 0.35 -- -- AMBN 0.25 -- -- Ethanol 18.75 10.0 10.0 Acrylic Acid -- 0.125 0.125

LaCl3.6H20 -- 0.0032 --

TbCl3.6H20 -- 0.0032 --

EuCl3.6H20 -- 0.0032 --

HoCl3.6H20 -- 0.0032 --

TmCl3.6H20 -- 0.0032 -- EGDMA -- -- 0.031

a Added 1.0 hr after the initiation. b Added 8.0 hr after the 2nd stage.

189 Analytical Aspects

All five samples resulted in stable and coagulum-free dispersions. Figure 7.10 shows SEM images of microsphere samples AA137A-E prepared to examine the reproducibility of 3-DisP.

All of the samples have a narrow size distribution (CVd = 1.6 – 2.5 %). However, the metal ion content of each microsphere is expected to be proportional to its volume rather than its diameter.

I used the coefficient of variation in microspheres' volume (CVV, Equation 7-3) to describe the width of distribution in microspheres’ volume.

1 1 n 2 CVv = (7-3) ∑i=1(Vi −Vav ) Vav n −1 th where Vav is the number average volume of all microspheres, Vi is the volume of the i microsphere, and n is the total number of particles counted in the analysis. Screen capture image of mass cytometry measurement of AA137C is given in Figure 7.11. The image shows the lanthanide code of three microspheres with very minimal amount of lanthanide in the continuous media. Although all metals were introduced with the same concentration, one can notice that the Europium isotopes have less bright spots in the three microspheres presented in the figure. This is attributed to the fact that Eu is the only “di- isotopic” element in the lanthanide code used here; Eu has two isotopes 151Eu and 153Eu with natural abundance of 52.2 % and 47.8 % respectively. Fig 7.12 illustrates 169Tm intensity

distribution of the 5 samples. All samples have narrow lanthanide distribution (CVLn = 15 – 23 %). Volume and Tm content of AA137A-E microsphere samples are listed in Table 7-6. A more detailed table that contains a complete set of the lanthanide content measured by mass cytometry is given in Appendix 1. Microsphere sample AA137A has an exceptionally high volume and lanthanide content relative to the other 4 samples despite all of the five samples were prepared with the same formulation and reaction conditions. In light of the results obtained in Chapter 4, where lanthanide content of the microspheres demonstrated a strong correlation with their volume, I studied the change of the lanthanide content of the microsphere with the microsphere volume for AA137A-E samples. The relation between the microsphere volume and their 165Ho content are shown in Figure 7.13.A (similar relation for 169Tm is shown in Figure 7.13.B). As expected AA137A sample had the highest volume compared to the other samples. Figure 7.13.A

190 Analytical Aspects

and B reveal the same strong dependency of the lanthanide content of the microspheres on their volume.

Figure 7.10. SEM images for samples AA137A-E of PS microspheres synthesized for reproducibility study. Samples were synthesized by 3-DisP in the presence of 0.05 % LaCl3 TbCl3 EuCl3 HoCl3 TmCl3 added in the second stage with AA: 2 wt %/styrene. Scale bar is 5.0 µm.

sit time 3

Tran Figure 7.11. Screen captures of the mass cytometry results for AA137C, synthesized in the presence of LaCl3 TbCl3 EuCl3 HoCl3 and TmCl3 (0.05 wt %/styrene for each), as an example of PS microspheres synthesized for the reproducibility study. The image shows 3 distinct microspheres each of them has the 5 elements present.

191 Analytical Aspects

Table 7.6. Average volume of the microsphere samples AA137 A–E compared to their average number of 169Tm atoms per microsphere. Note that the standard deviation of number of 169Tm content of the different synthesis (for AA137B-E and excluding AA137A) is 2.6 x 105 Tm atoms per microsphere and standard deviation of Tm/Volume ratios (for AA137B-E and excluding AA137A) is 7.9 x 104 Tm atoms per unit volume.

a b c d Sample Volume (µm) CVV (%) Tm Tm/Volume

AA137A 4.12 6.4 5.45 1.33

AA137B 2.35 5.9 2.23 0.95

AA137C 2.17 4.7 2.00 0.92

AA137D 2.38 7.5 2.62 1.10

AA137E 2.38 7.5 2.39 1.00

a Average microsphere volume from SEM images b Calculated by Equation 7-3 c Average number of Tm atoms per microsphere from mass cytometry measurements (x 106) d Normalized Tm content = number of Tm atoms per unit volume ( x 106)

Figure 7.12. 169Tm intensity distribution from the mass cytometry measurements of AA137A-E (average number of Tm atoms per microsphere = Tm intensity x 3.95 x 104). Note the obvious increase in the Tm intensity of AA137A relative to the other samples.

192 Analytical Aspects

Figure 7.13. Average volume of the microsphere samples AA137 A–E compared to their average 165Ho content (A) and average 169Tm content (B).

To understand the reasons that may lead to this substantial variation in the microsphere volume, which lead to a variation in the lanthanide content of the microspheres, I describe here the mechanism of particle formation in dispersion polymerization. According to Ober et al.[32, 33] and Paine,[34-37] the particle formation process of dispersion polymerization is divided into two major stages, that is, an early stage (particle formation or nucleation stage) in which the formation of particles or nuclei and aggregation between them are predominant and a later stage (particle growth stage) in which the particle growth is predominant. The nucleation stage is very short (in order of 5 - 15 min in case of styrene polymerization in ethanol) and extremely sensitive to changes in the polymerization reaction conditions.

193 Analytical Aspects

There is a general agreement that nuclei are formed throughout the polymerization process, while the number of particles is determined in the early stages. After the particle formation stage is complete, the number of particles does not change, but the particles increase in size.[34] [38- 40] The number of nuclei formed in the sensitive nucleation stage may vary for minor changes in the reaction condition. This is a key point in the dispersion polymerization that greatly affect the reproducibility of the produced particle size and make the results obtained from AA137 microspheres sample acceptable. That said, if AA137A is excluded, the other four samples are reproducible (< 10 % error). For example, the normalized Tm content (number of Tm atom per unit volume) for microsphere samples AA137 B-E has an average of 1.0 x 106 Tm atoms per unit volume with a standard deviation of 7.9 x 104 Tm atoms per unit volume (7.9 % error).

7.3 Conclusions

I report the synthesis of a series of metal-containing polystyrene microspheres with a very narrow size distribution, designed for the calibration of mass cytometry instruments. The PS microsphere samples reported here were synthesized by dispersion polymerization of styrene in ethanol. They contain up to 108 Ln ions per microsphere but were optimized to contain about 107 Ln ions, a concentration that falls in the detection range of the mass cytometer and is suitable for microspheres to be employed as instrument standards. In general, the metal-containing microspheres synthesized by 2-stage dispersion polymerization in the presence of 0.1% LnCl3 and 2% acrylic acid met the size, size distribution, and Ln-content requirements for the microspheres to be used as standard microspheres for mass cytometry measurements. A shortcoming of these microspheres was a large microsphere-to-microsphere variation in their lanthanide ion content. By modifying the particle synthesis strategy to add additional acrylic acid and a small amount of cross-linking reagent later in the reaction (3-stage dispersion polymerization, 3-DisP), a much lower microsphere-to-microsphere variation in the metal content was obtained. Lanthanide-containing microspheres, prepared by 3-DisP, are stable when stored in buffer at pH values ranging from 3 to 10.6, with no significant leaching of their embedded Ln into

194 Analytical Aspects

aqueous media. When these microspheres were mixed with cell suspensions of two different cell lines (KG1a and U937), only about 2% of the total signal came from cells that associated with microspheres. There was no influence of the presence of the microspheres on the metal content of the cells determined by mass cytometry. I conclude that microspheres prepared by 3-DisP are well suited to be used as mass cytometry instrument calibration standards, and as internal standard microspheres for the measurement of the metal content of cells by mass cytometry. Mass cytometry is a novel and rapidly improving analytical tool for phenotyping bioassays. I hope that the standard microspheres presented here will offer a way to obtain more consistent results and will help in improving the quality of the measurements using this technique. I found it difficult to reproduce the synthesis of lanthanide encoded microspheres by three stage dispersion polymerization. For five replicates, I got one microsphere sample that has larger size and higher Ln content relative to the other four samples. However, for the rest of the samples (4 microsphere samples), the synthesis was reproducible and the variation was within 10 %.

References

1. Abdelrahman AI, Ornatsky O, Bandura D, Baranov V, Kinach R, Dai S, Thickett SC, Tanner S, Winnik MA: Metal-containing polystyrene beads as standards for mass cytometry. J Anal At Spectrom 2010, 25(3):260-268. 2. Baumgarth N, Roederer M: A practical approach to multicolor flow cytometry for immunophenotyping. Journal of Immunological Methods 2000, 243(1-2):77-97. 3. Belov L, De la Vega O, Dos Remedios CG, Mulligan SP, Christopherson RI: Immunophenotyping of leukemias using a cluster of differentiation antibody microarray. Cancer Research 2001, 61(11):4483-4489. 4. Borowitz MJ, Guenther KL, Shults KE, Stelzer GT: Immunophenotyping of acute leukemia by flow cytometric analysis: Use of CD45 and right-angle light scatter to gate on leukemic blasts in three-color analysis. American Journal of Clinical Pathology 1993, 100(5):534-540. 5. Cazzola M: Flow cytometry immunophenotyping for diagnosis of myelodysplastic syndrome. Haematologica 2009, 94(8):1041-1043. 6. Chattopadhyay PK, Price DA, Harper TF, Betts MR, Yu J, Gostick E, Perfetto SP, Goepfert P, Koup RA, De Rosa SC et al: Quantum dot semiconductor nanocrystals for immunophenotyping by polychromatic flow cytometry. Nature Medicine 2006, 12(8):972-977.

195 Analytical Aspects

7. Comans-Bitter WM, De Groot R, Van den Beemd R, Neijens HJ, Hop WCJ, Groeneveld K, Hooijkaas H, Van Dongen JJM: Immunophenotyping of blood lymphocytes in childhood: Reference values for lymphocyte subpopulations. Journal of Pediatrics 1997, 130(3):388-393. 8. Rothe G, Schmitz G: Consensus protocol for the flow cytometric immunophenotyping of hematopoietic malignancies. Leukemia 1996, 10(5):877-895. 9. Truong F, Smith BR, Stachurski D, Cerny J, Medeiros LJ, Woda BA, Wang SA: The utility of flow cytometric immunophenotyping in cytopenic patients with a non- diagnostic bone marrow: A prospective study. Leukemia Research 2009, 33(8):1039- 1046. 10. Hoffman RA: Standardization, calibration, and control in flow cytometry. Current protocols in cytometry / editorial board, J Paul Robinson, managing editor [et al] 2005, Chapter 1. 11. Wang L, Gaigalas AK, Marti G, Abbasi F, Hoffman RA: Toward quantitative fluorescence measurements with multicolor flow cytometry. Cytometry Part A 2008, 73(4):279-288. 12. Schwartz A, Fernandez Repollet E, Vogt R, Gratama JW: Standardizing flow cytometry: Construction of a standardized fluorescence calibration plot using matching spectral calibrators. Communications in Clinical Cytometry 1996, 26(1):22- 31. 13. Vogt Jr RF, Whitfield WE, Henderson LO, Hannon WH: Fluorescence intensity calibration for immunophenotyping by flow cytometry. Methods 2000, 21(3):289- 296. 14. Wu Y, Campos SK, Lopez GP, Ozbun MA, Sklar LA, Buranda T: The development of quantum dot calibration beads and quantitative multicolor bioassays in flow cytometry and microscopy. Anal Biochem 2007, 364(2):180-192. 15. Hansen S, Dahl R, Hoffmann HJ: Determination of blood leukocyte concentration with constant volume acquisition on a flow cytometer is comparable to individualized single platform testing with beads as internal reference standard. Journal of Immunological Methods 2008, 338(1-2):58-62. 16. Jensen IM, Kristensen JS, Thomsen M, Ellegaard J, Hokland P: Optimization of a method for simultaneous analysis of cell surface phenotype and cell cycle in human bone marrow and peripheral blood leukocytes by flow cytometry: The value of both internal and external standards. Analytical Cellular Pathology 1993, 5(4):213-223. 17. Han M, Gao X, Su JZ, Nie S: Quantum-dot-tagged microbeads for multiplexed optical coding of biomolecules. Nat Biotechnol 2001, 19(7):631-635. 18. Baranov VI, Quinn ZA, Bandura DR, Tanner SD: The potential for elemental analysis in biotechnology. J Anal At Spectrom 2002, 17(9):1148-1152. 19. Quinn ZA, Baranov VI, Tanner SD, Wrana JL: Simultaneous determination of proteins using an element-tagged immunoassay coupled with ICP-MS detection. J Anal At Spectrom 2002, 17(8):892-896. 20. Ornatsky O, Baranov VI, Bandura DR, Tanner SD, Dick J: Multiple cellular antigen detection by ICP-MS. Journal of Immunological Methods 2006, 308(1-2):68-76.

196 Analytical Aspects

21. Ornatsky OI, Kinach R, Bandura DR, Lou X, Tanner SD, Baranov VI, Nitz M, Winnik MA: Development of analytical methods for multiplex bio-assay with inductively coupled plasma mass spectrometry. J Anal At Spectrom 2008, 23(4):463-469. 22. Ornatsky OI, Lou X, Nitz M, Schafer S, Sheldrick WS, Baranov VI, Bandura DR, Tanner SD: Study of cell antigens and intracellular DNA by identification of element- containing labels and metallointercalators using inductively coupled plasma mass spectrometry. Anal Chem 2008, 80(7):2539-2547. 23. Abdelrahman AI, Dai S, Thickett SC, Ornatsky O, Bandura D, Baranov V, Winnik MA: Lanthanide-containing polymer microspheres by multiple-stage dispersion polymerization for highly multiplexed bioassays. J Am Chem Soc 2009, 131(42):15276-15283. 24. Tanner SD, Bandura DR, Ornatsky O, Baranov VI, Nitz M, Winnik MA: Flow cytometer with mass spectrometer detection for massively multiplexed single-cell biomarker assay. Pure Appl Chem 2008, 80(12):2627-2641. 25. Tanner SD, Ornatsky O, Bandura DR, Baranov VI: Multiplex bio-assay with inductively coupled plasma mass spectrometry: Towards a massively multivariate single-cell technology. Spectrochimica Acta - Part B Atomic Spectroscopy 2007, 62(3 SPEC. ISS.):188-195. 26. Thickett SC, Abdelrahman AI, Ornatsky O, Bandura D, Baranov V, Winnik MA: Bio- functional, lanthanide-labeled polymer particles by seeded emulsion polymerization and their characterization by novel ICP-MS detection. J Anal At Spectrom 2010, 25(3):269-281. 27. Lou X, Zhang G, Herrera I, Kinach R, Ornatsky O, Baranov V, Nitz M, Winnik MA: Polymer-based elemental tags for sensitive bioassays. Angewandte Chemie - International Edition 2007, 46(32):6111-6114. 28. Bandura DR, Baranov VI, Ornatsky OI, Antonov A, Kinach R, Lou X, Pavlov S, Vorobiev S, Dick JE, Tanner SD: Mass cytometry: Technique for real time single cell multitarget immunoassay based on inductively coupled plasma time-of-flight mass spectrometry. Anal Chem 2009, 81(16):6813-6822. 29. Perring L, Alonso MI, Andrey D, Bourqui B, Zbinden P: An evaluation of analytical techniques for determination of lead, cadmium, chromium, and mercury in food- packaging materials. Analytical and Bioanalytical Chemistry 2001, 370(1):76-81. 30. Parks DR, Roederer M, Moore WA: A new "logicle" display method avoids deceptive effects of logarithmic scaling for low signals and compensated data. Cytometry Part A 2006, 69(6):541-551. 31. Wu D, Zhao B, Dai Z, Qin J, Lin B: Grafting epoxy-modified hydrophilic polymers onto poly(dimethylsiloxane) microfluidic chip to resist nonspecific protein adsorption. Lab on a Chip - Miniaturisation for Chemistry and Biology 2006, 6(7):942- 947. 32. Ober CK, Lok KP: Formation of large monodisperse copolymer particles by dispersion polymerization. Macromolecules 1987, 20(2):268-273. 33. Ober CK, Lok KP, Hair ML: Monodispersed, micron-sized polystyrene particles by dispersion polymerization. Journal of polymer science Polymer letters edition 1985, 23(2):103-108.

197 Analytical Aspects

34. Paine AJ, Luymes W, McNulty J: Dispersion polymerization of styrene in polar- solvents .6. Influence of reaction parameters on particle-size and molecular-weight in poly(n-vinylpyrrolidone)-stabilized reactions. Macromolecules 1990, 23(12):3104- 3109. 35. Paine AJ: Dispersion polymerization of styrene in polar-solvents .1. Grafting mechanism of stabilization by hydroxypropyl cellulose. J Colloid Interface Sci 1990, 138(1):157-169. 36. Paine AJ: Dispersion polymerization of styrene in polar-solvents .4. Solvency control of particle-size from hydroxypropyl cellulose stabilized polymerizations. Journal of Polymer Science Part a-Polymer Chemistry 1990, 28(9):2485-2500. 37. Paine AJ: Dispersion polymerization of styrene in polar-solvents .7. A simple mechanistic model to predict particle-size. Macromolecules 1990, 23(12):3109-3117. 38. Kawaguchi S, Ito K: Dispersion polymerization. In: Polymer Particles. vol. 175; 2005: 299-328. 39. Nugroho MB, Kawaguchi S, Ito K, Winnik MA: Control of particle-size in dispersion polymerization using poly(ethylene oxide) macromonomers. Journal of Macromolecular Science-Pure and Applied Chemistry 1995, A32:593-601. 40. Uchida T, Kawaguchi S, Ito K: Dispersion (co)polymerization of styrene in polymeric media to prepare polymer micro blends. Designed Monomers and Polymers 2002, 5(2- 3):285-296.

198 Analytical Aspects

Appendix 1

Average volume of the microsphere samples AA137 A–E compared to their average number of lanthanide atoms per microsphere (lanthanides = 139La, 153Eu 159Tb 165Ho and 169Tm).

Diameter Volumea 10-6 x Ln / Sample CV b (%) Ln/Volumed (µm) (µm3) V particle c La: 2.39 La: 0.58 Eu: 2.60 Eu: 0.63 AA137A 1.99 4.12 6.4 Tb: 4.78 Tb: 1.16 Ho: 6.91 Ho: 1.68 Tm: 5.45 Tm: 1.33 La: 0.98 La: 0.42 Eu: 1.06 Eu: 0.45 AA137B 1.65 2.35 5.9 Tb: 1.96 Tb: 0.83 Ho: 2.74 Ho: 1.17 Tm: 2.23 Tm: 0.95 La: 0.88 La: 0.40 Eu: 0.95 Eu: 0.44 AA137C 1.61 2.17 4.7 Tb: 1.75 Tb: 0.81 Ho: 2.41 Ho: 1.11 Tm: 2.00 Tm: 0.92 La: 1.15 La: 0.48 Eu: 1.25 Eu: 0.52 AA137D 1.65 2.35 7.5 Tb: 2.30 Tb: 0.97 Ho: 3.18 Ho: 1.34 Tm: 2.62 Tm: 1.10 La: 1.05 La: 0.19 Eu: 1.14 Eu: 0.48 AA137E 1.65 2.35 7.5 Tb: 2.10 Tb: 0.88 Ho: 2.94 Ho: 1.24 Tm: 2.39 Tm: 1.00

a Average microsphere volume from SEM images b Calculated by Equation 7-3 c Average number of lanthanide atoms per microsphere from mass cytometry measurements d Normalized lanthanide content = number of lanthanide atoms per µm3