USF Libraries
USF Digital Collections

Using multiwavelength UV-visible spectroscopy for the characterization of red blood cells

MISSING IMAGE

Material Information

Title:
Using multiwavelength UV-visible spectroscopy for the characterization of red blood cells an investigation of hypochromism
Physical Description:
Book
Language:
English
Creator:
Nonoyama, Akihisa
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:

Subjects

Subjects / Keywords:
multiwavelength UV-visible spectrophotometry
hemoglobin
red blood cell
Mie theory
light scattering
hypochromism
Dissertations, Academic -- Chemistry -- Doctoral -- USF
Genre:
government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: Particle analysis using multiwavelength UV-visible spectroscopy provides the potential for extracting quantitative red blood cell information, such as hemoglobin concentration, cell size, and cell count. However, if there is a significant presence of hypochromism as a result of the concentrated hemoglobin (physiological value of 33%), successful quantification of red cell values would require a correction. Hypochromism has been traditionally defined as a decrease in absorption relative to the values expected from the Beer-Lambert Law due to electronic interactions of chromophores residing in close proximity of one another. This phenomenon has been suggested to be present in macroscopic systems composed of strong chromophores such as nucleic acids, chlorophyll, and hemoglobin.The study presented in this dissertation examines the presence of hypochromism in red blood cells as a part of a larger goal to qualitatively and quantatively characterize red blood cells using multiwavelength UV-visible spectroscopy. The strategy of the study was three-fold: 1) to determine the instrumental configuration that would provide the most complete information in the acquired spectra, 2) to develop an experimental model system in which the hemoglobin content in red blood cells could be modified to various concentrations, and 3) to implement an interpretation model based on light scattering theory (which accounts for both the scattering and absorption components of the optical density spectrum) to provide quantitative information for the experimental system.By this process, hypochromicity was redefined into two categories with molecular hypochromicity representing the traditional definition and macroscopic hypochromicity being an attenuation of the absorption component due to a scattering-related effect. Successful simulations of experimental red cell spectra containing various amounts of hemoglobin were obtained using the theoretical model. Furthermore, successful quantitative interpretation of the red blood cell spectra was achieved in the context of corpuscular hemoglobin concentration, corpuscular volume, and cell count solely by accounting for the scattering and absorption effects of the particle, indicating that molecular hypochromicity was insignificant in this macroscopic system.
Thesis:
Thesis (Ph.D.)--University of South Florida, 2004.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Akihisa Nonoyama.
General Note:
Includes vita.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 256 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001498115
oclc - 57714957
notis - AJU6710
usfldc doi - E14-SFE0000508
usfldc handle - e14.508
System ID:
SFS0025199:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001498115
003 fts
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 041209s2004 flua sbm s000|0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0000508
035
(OCoLC)57714957
9
AJU6710
b SE
SFE0000508
040
FHM
c FHM
090
QD31.2 (ONLINE)
1 100
Nonoyama, Akihisa.
0 245
Using multiwavelength UV-visible spectroscopy for the characterization of red blood cells
h [electronic resource] :
an investigation of hypochromism /
by Akihisa Nonoyama.
260
[Tampa, Fla.] :
University of South Florida,
2004.
502
Thesis (Ph.D.)--University of South Florida, 2004.
504
Includes bibliographical references.
500
Includes vita.
516
Text (Electronic thesis) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 256 pages.
520
ABSTRACT: Particle analysis using multiwavelength UV-visible spectroscopy provides the potential for extracting quantitative red blood cell information, such as hemoglobin concentration, cell size, and cell count. However, if there is a significant presence of hypochromism as a result of the concentrated hemoglobin (physiological value of 33%), successful quantification of red cell values would require a correction. Hypochromism has been traditionally defined as a decrease in absorption relative to the values expected from the Beer-Lambert Law due to electronic interactions of chromophores residing in close proximity of one another. This phenomenon has been suggested to be present in macroscopic systems composed of strong chromophores such as nucleic acids, chlorophyll, and hemoglobin.The study presented in this dissertation examines the presence of hypochromism in red blood cells as a part of a larger goal to qualitatively and quantatively characterize red blood cells using multiwavelength UV-visible spectroscopy. The strategy of the study was three-fold: 1) to determine the instrumental configuration that would provide the most complete information in the acquired spectra, 2) to develop an experimental model system in which the hemoglobin content in red blood cells could be modified to various concentrations, and 3) to implement an interpretation model based on light scattering theory (which accounts for both the scattering and absorption components of the optical density spectrum) to provide quantitative information for the experimental system.By this process, hypochromicity was redefined into two categories with molecular hypochromicity representing the traditional definition and macroscopic hypochromicity being an attenuation of the absorption component due to a scattering-related effect. Successful simulations of experimental red cell spectra containing various amounts of hemoglobin were obtained using the theoretical model. Furthermore, successful quantitative interpretation of the red blood cell spectra was achieved in the context of corpuscular hemoglobin concentration, corpuscular volume, and cell count solely by accounting for the scattering and absorption effects of the particle, indicating that molecular hypochromicity was insignificant in this macroscopic system.
590
Adviser: Potter, Robert L.
Co-adviser: Garcia-Rubio, Luis H.
653
multiwavelength UV-visible spectrophotometry.
hemoglobin.
red blood cell.
Mie theory.
light scattering.
hypochromism.
690
Dissertations, Academic
z USF
x Chemistry
Doctoral.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.508



PAGE 1

Using Multiwavelength UV-Visi ble Spectroscopy for the Ch aracterization of Red Blood Cells: An Investigation of Hypochromism by Akihisa Nonoyama A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Chemistry College of Arts and Sciences University of South Florida Co-Major Professor: Robert L. Potter, Ph.D. Co-Major Professor: Luis H. Garcia-Rubio, Ph.D. German F. Leparc, M.D. Li-June Ming, Ph.D. Date of Approval: November 5, 2004 Keywords: multiwavelength UV-visible spectrophotometry, hypochromism, light scattering, Mie theory, re d blood cell, hemoglobin Copyright 2004 Akihisa Nonoyama

PAGE 2

Dedication This dissertation is dedicated to all of my family members who have supported me throughout my life and allowed me to achieve my goals: to my late father Meihan, who helped in making my educational decisions and sent me on the path to my graduate career; to my aunt, Michi ko, for providing me with financial and moral support throughout my stay in graduate school; to my mother Keiko for raising me into an honest person and for providing financial and mo ral support; to my wife Tina for her encouragement and support since we met in 1994; to my son Jacob who gave me a more of a reason to succeed at my endeavors; and to any future child…I just don’t know who you are yet.

PAGE 3

Acknowledgements I would like to acknowledge the guidance and support of my co-major professors, Dr. Robert L. Potter and Dr. Luis H. Garc ia-Rubio. I would also like to thank Dr. German F. Leparc, Chief Medical Officer of the Florida Blood Services, for his invaluable advice in hematology and for providing us with samples and laboratory space without which this research would not have b een possible. Additionally, I want to thank everyone that has offered help or encouragemen t in the laboratories of Dr. Potter and Dr. Garcia-Rubio. Finally, I tha nk University of South Flor ida and the Department of Chemistry for providing me with the opportunity to achieve my goal.

PAGE 4

i Table of Contents List of Tables…………………………………………………………………………….. iv List of Figures……………………………………………………………………………. vi List of Abbreviations…………………………………………………………………… xiv List of Symbols………………………………………………………………………… xvi Abstract………………………………………………………………………………… xvii Chapter 1: Introduction…………………………………………………………………... 1 1.1 Blood…………………………………………………………………………. 1 1.2 Theoretical Modeling………………………………………………………… 5 Chapter 2: Whole Blood, Red Blood Cells and Hemoglobin……………………………. 6 2.1 Whole blood……………………………………….......................................... 6 Leukocytes ………………………………………………………………... 7 Thrombocytes ……………………………………………………………. 13 Erythrocytes …………………………………………………………….. 14 Plasma ……………………………………………………………………14 Serum Albumin ………………………………………………………….. 15 2.2 The Erythrocyte……………………………………………………………...16 The Erythrocyte Membrane and Cytoskeleton ………………………….. 18 The Erythrocyte Surface ………………………………………………… 23 The Life of an Erythrocyte ………………………………………………. 24 Metabolic Pathways in the Erythrocyte …………………………………. 27 Morphology of the Red Cell …………………………………………….. 28 2.3 Hemoglobin………………………………………………………………….29 Hemoglobin Structure and Function ……………………………………..29 Hemoglobin Cooperactivity ……………………………………………... 36 Hemoglobin Derivatives ………………………………………………… 37 2.4 Current Methods of Red Cell Analysis……………………………………... 38 2.5 Erythrocytes and Multiwavelength UV-Visible Analysis………………….. 40 Chapter 3: Multiwavelength Spectrophotom etry and Its Application for the Characterization of Red Blood Cells…………………………….................. 43

PAGE 5

ii 3.1 Features of Spectrophotometry……………………………………………... 43 Absorption Component ………………………………………………….. 44 Absorption in Proteins …………………………………………………... 48 Scattering Component ……………………………………………………50 3.2 Light Scattering Theory…………………………………………………….. 50 Scattering Trends ………………………………………………………... 50 Mie Theory ………………………………………………………………. 55 Optical Properties ………………………………………………………. 58 Mie Theory Considerations ………………………………………………59 3.3 Spectroscopy of Blood and Erythrocytes…………………………………… 60 Spectroscopy of Whole Blood …………………………………………... 60 Spectroscopy of Erythrocytes …………………………………………… 61 Spectroscopy of Hemoglobin ……………………………………………. 64 Hypochromism …………………………………………………………... 65 Considerations of Instrumental Configuration …………………………. 72 Defining the Problems …………………………………………………... 74 Experimental and Theoretical Approach ………………………………... 76 Chapter 4: Spectrophotometry of Puri fied and Modified Red Blood Cells…………….. 79 4.1 Hypotonic Modification of Erythrocytes…………………………………… 79 4.2 Materials and Methods……………………………………………………… 82 The UV-Visible Spectrophotometer ……………………………………... 82 Types of Spectrophotometers Used ……………………………………… 84 Care and Preparation of the Cuvette …………………………………… 85 Preparation for Spectrophotometry …………………………………….. 85 Spectrum of Deionized Water …………………………………………… 86 Background Correction …………………………………………………..87 Hematology Analyzer …………………………………………………… 90 Drabkin’s Hemoglobin Determination Assay …………………………… 93 Calibration of Conductivity Me ter for Ionic Strength Measurements …………………………………………………………….96 Sample Preparation for Resealing Experiments ………………………… 97 Spectrophotometry of Leuko-Reduced Red Cells ………………………. 101 Calculations for Resealing Experiments ………………………………. 102 Red Cell Resealing Experiment …………………………………………104 4.3 Results and Discussion……………………………………………………..106 Effects of Instrumental Setup on the Spectra of Erythrocytes and Hemoglobin ………………………………………………………... 106 Spectral Evaluation of Adsol ……………………………………………112 Washed and Leuko-Reduced Cells ……………………………………... 114 Spectrum of Resealed Red Cells ………………………………………...120 Reproducibility of the Resealed Cells ………………………………….. 124 Microscopy ……………………………………………………………... 126 Comments on the Acquired Experimental Data ……………………….. 131

PAGE 6

iii Chapter 5: Implementation and Validation of the Mie Theory………………………... 134 5.1 Sensitivity Analysis of the Mi e Theory: A test of the model……………...134 Simulation of Erythrocytes ……………………………………………... 138 Effects of Size on Erythrocyte Spectra …………………………………. 140 Effect of varying MCHCs on Red Cell Spectra ………………………… 145 Combined Effect of MCHC and MCV Changes ………………………... 147 5.2 Qualitative Analysis of Experimental Spectra…………………………….. 149 5.3 Conclusions………………………………………………………………... 153 Chapter 6: Application of the Interpretation Model for the Quantitative Analysis of Erythrocytes…………………………………………………... 154 6.1 The Interpretation Model………………………………………………….. 155 Variants of the Interpretation Model …………………………………... 158 Corroboration of the Estimated Parameters …………………………... 159 Testing the Interpretation Model ………………………………………. 162 6.2 Quantitative Results of the Interpretation Model………………………….. 168 Quantitative Results Obtained with RBCHb01a ……………………….. 170 6.3 Discussion …………………………………………………………………..177 Chapter 7: Conclusion………………………………………………………………….180 Future Work ……………………………………………………………. 183 References……………………………………………………………………………… 185 Appendices……………………………………………………………………………...196 Appendix A: Mie Theory Formulae……………………………………………197 Appendix B: Concentration-Rela ted Instrumental Limitations of Hemoglobin and Erythrocytes……………………………………………… 199 Appendix C: Spectral Characterizati on of a Liposome Model System………. 202 Appendix D: RBC Swelling…………………………………………………... 211 Appendix E: Spiking Red Cell Suspensions with Hemoglobin………………. 214 Appendix F: Detailed Examination of Experimental and Simulated Spectra of Resealed Cells……………………………………………………… 219 Appendix G: Alternate Versions of the Interpretation Model………………… 226 Appendix H: UV-Visible Analysis of Molecular Aggregates………………… 230 About the Author………………………………………………………………… End Page

PAGE 7

iv List of Tables Table 2.1: Normal values of major blood components………………………………….. 7 Table 2.2: Percent abundances of leukocyte sub-types…………………………………...8 Table 2.3: Protein conten t of human erythrocytes….........................................................16 Table 4.1: Linearity ranges for some important parameters of the SeronoBaker hematology analyzer…………………………………………………..92 Table 4.2: Error of the parameters of Serono-Baker hematology analyzer outputs as defined by the Para 12 Multi-parameter Assayed hematology control…………………………………………………………...92 Table 4.3: Error of the parameters of Serono-Baker hematology analyzer outputs calculat ed from 6 replicates of the same sample…………………… 92 Table 4.4: Recipes for the mixing propor tions of each tube to obtain the concentrations in the Blood Hemoglobin column…………………………... 94 Table 4.5: Values obtained by the Serono-Ba ker hematology analyzer of whole blood and washed cells……………………………………………... 116 Table 4.6: Values obtained from the he matology analyzer for the different stages of red cell purification……………………………………………… 118 Table 4.7: Values obtained for five experi mental replicates of resealed cells for the parameters HGB, MCHC, and MCV………………………………126 Table 5.1: Physiological volume range of red cells in 5 fl increments and their equivalent sphere diameters…………………………………………..140 Table 6.1: The three versions of the interpretation models and the variations in their functions…………………………………………………158

PAGE 8

v Table 6.2: Table of estimated and measured values (g/dl) for representa tive erythrocyte sample s containing various amounts of hemoglobin……………………………………………………. 170 Table 6.3: Comparison of cell counts of resealed cells between measured (Coulter Z2 and Serono-Baker 9110) and calculated values………………. 176 Table 6.4: Comparison of the equivale nt sphere diameter ( m) of resealed cells between measured (Coulter Z2 and Serono-Baker 9110) and calculated values…………………………………………………………… 176

PAGE 9

vi List of Figures Figure 1.1: Blood components and their spectral contributions to the whole blood spectrum………………………………………………………………. 2 Figure 1.2: Three important pathlengths to consider in the spectral analysis of red cells…………………………………………………………………... 4 Figure 2.1: An electron microscopy picture of a mature neutrophil using Peroxidase staining…………………………………………………………. 9 Figure 2.2: An electron microscopy pict ure of an eosinophil using peroxidase staining…………………………………………………………. 10 Figure 2.3: An electron microscopy picture of a basophil using peroxidase staining………………………………………………………… 10 Figure 2.4: An electron microscopy pict ure of a lymphocyte………………………….. 11 Figure 2.5: An electron microscopy pi cture of a monocyte……………………………. 12 Figure 2.6: The figure depicts the organization of the membrane and membrane skeleton………………………………………………………… 18 Figure 2.7: A cartoon and micrograph repr esentations of the erythrocyte…………….. 17 Figure 2.8: The general chemical structur e of a phosphoglyceride…………………….. 19 Figure 2.9: R3 groups for major types of phosphoglycerides…………………………... 19 Figure 2.10: A structural formula of sphingomyelin showing the ceremide and phosphocholine moieties……………………………………………. 20 Figure 2.11: A structural fo rmula of cholesterol………………………………………. 21 Figure 2.12: The figure depicts the organization of the membrane and membrane skeleton……………………………………………………….. 23

PAGE 10

vii Figure 2.13: The figure illustrates the cell lineage of each of the blood cells gene rated from the pluripotenti al stem cell (CFU-S)……………… 25 Figure 2.14: An electron micrograp h of a discocyte (A) and an echinocyte (B)……………………………………………………………. 28 Figure 2.15: An electron micrograph of a drepanocyte………………………………… 29 Figure 2.16: The chain of hemoglobin shows the globular tertiary structure consisting of eight -helices……………………………………………… 31 Figure 2.17: A structural depic tion of the heme molecule……………………………… 31 Figure 2.18: An illustration of the heme pocket in the deoxygenated state……………. 32 Figure 2.19: When deoxygenated, the iron is pulled out of the heme plane ~0.6 wh eras the iron resides in the plane in the presence of a ligand…………………………………………………………………33 Figure 2.20: This figure illustrates the dynamics in the interface between the T and R forms of hemoglobin………………………………………… 35 Figure 2.21: Structural formula fo r 2,3-diphosphoglycerate (2,3-DPG)……………….. 35 Figure 2.22: Oxygen binding curves of hemoglobin and myoglobin, with the hem oglobin curve showing a sigm oidal shape denoting its allosteric behavior………………………………………………………… 37 Figure 3.1: Diagram of electromagnetic radiat ion in the UV, visible, and IR regions…………………………………………………………………...43 Figure 3.2: A depiction of a wave of electr omagnetic radiation where the magnetic field (B ) and electric field (E) are perpendicular to each other as well as to the direc tion of propagation…………………………… 45 Figure 3.3: A representation of an n electronic transition in a carbonyl moiety……………………………………………………………..46 Figure 3.4: A representation of an electronic transition in a carbonyl moiety……………………………………………………………. 46 Figure 3.5: Chemical structures of the chromophoric amino acids that cont ribute to the absorban ces of proteins…………………………………... 49

PAGE 11

viii Figure 3.6: Scattering patterns for a spherical particle that is small relative to the wavelength (R ayleigh scattering) where A is scattering due to a polarized incident light perpendicular to th e scattering plane, B is unpolarized incident light, and C is polarized incident light parallel to the scattering plane………………………………………………………. 51 Figure 3.7: Scattering profile s of large particles (~1.24 m) as a sphere, a nd ellipses in various orientations………………………………………… 52 Figure 3.8: Calculated scattering patterns of spherical particles increasing in si ze generated using th e Mie theory, where = 500 nm, the medium is water, an d the optical properties ar e that of bacteria…………………….. 53 Figure 3.9: Angular scatteri ng prediction for an equiva lent sphere red cell…………... 55 Figure 3.10: Extinction spectra of oxyhemoglobin and methemoglobin……………….. 65 Figure 3.11: An example of molecular hypochromism exhibited by DNA in a double helical configuration……………………………………………….67 Figure 3.12: The effect of chain length of a polyadenine oligonucleotide on optical density, where the x-axis is the chain length and the y-axis is the optical density……………………………………………………….68 Figure 3.13: Theoretical vs. experimental sp ectra of poly-ad enosine stacks……………69 Figure 3.14: Schematic representati on of the angle of acceptance……………………... 73 Figure 4.1: Illustrated representation of the hypotonic permeabilization of red blood cells……………………………………………………………... 80 Figure 4.2: The front view of a Hewlett Packard/Agilent 8453 Spectrophotometer………………………………………………………… 82 Figure 4.3: Cuvette slot of an HP/Agile nt 8452 spectrophotometer with a levered locking mechanism………………………………………………. 77 Figure 4.4: The optical machinery of the HP/Agilent 8453 Spectrophotometer…………………………………………………………. 78 Figure 4.5: A representative blank spectrum used to correct for ambient conditions prio r to sample measurement, where the X-axis is the wavelength in nanometers, and the Y-ax is represents optical density units………………………………………………………………... 80

PAGE 12

ix Figure 4.6: Representative spectrum of deionized water……………………………….. 82 Figure 4.7: Representative spectru m of isotonic (0.9%) PBS (pH 7.0-7.2)……………..83 Figure 4.8: A sample re adout obtained from the Serono-Baker hematology Analyzer……………………………………………………………………. 85 Figure 4.9: A stardard curve of conduc tivity vs. PBS concentration (%)……………… 90 Figure 4.10: Representative comparison spectra of purified red cells and hemoglobin in solution acquired from an Agilent 8453 spectrophotometer with an acceptance angle of 2o……………………… 107 Figure 4.11: Representative comparison spectra of purified red cells and hemoglobin in solution acquired from a Perkin-Elmer Lambda 900 spectrophotometer with an acceptance angle >2o................ .108 Figure 4.12: Representative comparis on spectra of purified red cells and hemoglobin in solution acquired from a Perkin-Elmer Lambda 18 sp ectrophotometer fitted with an integrating sphere………... 109 Figure 4.13: Optical density spectra of Adsol………………………………………….113 Figure 4.14: Normalized (by area under curv e) optical density spectra of whole blood and washed cells…………………………………………… 115 Figure 4.15: Stages in the purificat ion of red cells……………………………………. 117 Figure 4.16: Comparison of red cell suspen sion and supernatant (I)…………………..118 Figure 4.17: Comparison of red cell suspensi on and supernatant (II)………………… 119 Figure 4.18: Optical density spectra of a resealed red cell suspension (MCHC: 0.084) and the supernatant…………………………………….. 121 Figure 4.19: Optical density spectra of a resealed red cell suspension (MCHC: 0.218, 0.186) and the supernatant……………………………… 122 Figure 4.20: A compilation of raw spectral data of resealed cells with diffe rent encapsulated hemoglobin concentrations……………………… 123 Figure 4.21: Compilation spectra normalized using the area unde r the curve method…………………………………………………………………... 124

PAGE 13

x Figure 4.22: Normalized resealed cell spectra of five replicates……………………… 125 Figure 4.23: A phase contrast microscopy picture of erythrocytes in their native state magnified 400x……………………………………………... 127 Figure 4.24: A phase contrast microscope image of non-viable red cells…………….. 128 Figure 4.25: A phase contrast microscope image (1000x, oil immersion) of hypotoni cally resealed red cells containing 21.4% (w/v) hemoglobin……………………………………………………………….129 Figure 4.26: A phase contrast microscope image (1000x, oil immersion) of hypot onically resealed red cells containing 5.6% (w/v) hemoglobin………………………………………………………………130 Figure 5.1: Diagram of the model used to ca lculate spectra and its inputs…………….135 Figure 5.2: Plot of the contents of an optical properties file for oxyhemogl obin over the wavelengt h range of 190-1100 nm……………...137 Figure 5.3: Simulated spectrum of red blood cells under physiological conditions………………………………………………………………… 139 Figure 5.4: Illustration of cell suspension wher e the size is increased by the hematocrit and weight-based cell concentra tion remains constant.………. 141 Figure 5.5: Simulated spectra of varying MCV at constant MCHC (0.05 mass fraction)………………………………………………………..143 Figure 5.6: Simulated spectra of varying MCV at constant MCHC (0.20 mass fraction)………………………………………………………..144 Figure 5.7: Simulated spectra of erythrocytes at varying MCV at constant MCHC (0.33 mass fraction)………………………………………………. 145 Figure 5.8: Simulated spectra of erythroc ytes at varying MCHC and constant MCV…………………………………………………………….. 147 Figure 5.9: Combined spectral eff ect of changes in RBC hemoglobin concentration and volume………………………………………………… 149 Figure 5.10: Experimental spectra of resealed cells with varying MCHC and MCV values………………………………………………………… 151

PAGE 14

xi Figure 5.11: Simulated spectra of resealed cells using the experimental MCHC and MCV values in Figure 5.20………………………………… 152 Figure 6.1: Schematic diagram of the interpretation model……………………………156 Figure 6.2: Comparison of hemoglobin con centration (HGB) values (g/dl) measured by the Serono-Baker and the Drabkin’s assay………….. 160 Figure 6.3: Comparison of the mean corpuscu lar hemoglobin concentration (MCHC) values (mass fraction) measured by the Serono-Baker and the Drabkin’s assay…………………………………………………... 161 Figure 6.4: Output of the spectral estimat e by the RBCHb01a interpretation model for low range MCHC……………………………………………… 164 Figure 6.5: A representation of the algorithmic process of the interpretation model……………………………………………………………………... 165 Figure 6.6: Output of the spectral estimat e by the RBCHb01a interpretation model for medium range MCHC…………………………………………. 166 Figure 6.7: Output of the spectral estimat e by the RBCHb01a interpretation model for physiological range MCHC……………………………………. 167 Figure 6.8: Comparison of the calculated MC HC to the MCHC values measured by the hematology analyzer and the manual Drabkin’s assay…………………………………………………………… 169 Figure 6.9: Comparison of the calculated hemoglobin concentration (HGB) to the HGB values measured by the hematology analyzer and the manual Drabkin’s assay…………………………………………..171 Figure 6.10: Calculated vs measured equiva lent sphere diameters…………………… 173 Figure 6.11: Calculated vs measured cell volumes……………………………………. 174 Figure 6.12: Calculated vs measured cell counts……………………………………… 175 Figure B1: Assessment of the optical density linearity limits of the Agilent 8453 spectrophotometer as a function of he moglobin (solution) concentration……………………………………… 200

PAGE 15

xii Figure B2: Serial dilutions of whole blood for the examination of multiple scattering……………………………………………………… 201 Figure C1: Diagram of a liposome……………………………………………………. 203 Figure C2: Flowchart of the method for he moglobin harvesting prior to the production of hemoglobin-encapsulated liposomes (hemosomes)….…….. 205 Figure C3: Protocol for the generation of hemoglobin-encapsulated liposomes (hemosomes)…………………………………………………… 206 Figure C4: Normalized spectra of crude and filtered albumin liposomes compared to a spectrum of free albumin in solution……………………… 207 Figure C5: Normalized spectra of crude and filtered hemosomes compared to a spectrum of free hemoglobin in solution………………… 208 Figure C6: Light microscope picture of albumin liposomes magnified to 400x…………………………………………………………. 210 Figure D1: Spectra of red cells in varying tonicities of PBS media………………….. 212 Figure E1: Spectrum of the control PBS sp iked with two volumes of free hemoglobin…………………………………………………………… 215 Figure E2: Spectrum of a purified red cel l suspension spiked with free hemoglobin…………………………………………………………… 216 Figure E3: Spectrum of a resealed cell suspension spiked with free hemoglobin…………………………………………………………… 217 Figure F1: Experimental data set of reseal ed red cells in the low MCHC range with varying MCV values………………………………………….. 219 Figure F2: Simulated spectra of the low MCHC range data set normalized on a per cell basis…………………………………………………………. 220 Figure F3: data set of resealed red cells in the medium MCHC range with varying MCV values……………………………………………………… 221 Figure F4: Simulated spectra of the medium MCHC range data set normalized on a per cell basis…………………………………………….. 222

PAGE 16

xiii Figure F5: Experimental data set of reseal ed red cells in the high MCHC range with varying MCV values………………………………………….. 223 Figure F6: Simulated spectra of the high MCHC range data set normalized on a per cell basis…………………………………………………………. 224 Figure G1: Resealed cell data in terpreted using RBCHb01a interpretation model………………………………………………………. 227 Figure G2: Data of same resealed sample as Figure A1 interpreted with RBCHb01b, accounting for the presence of methemoglobin……………. 228 Figure G3: An RBCHb02a interpretation of a purified red cell suspension spiked with free hemoglobin……………………………………………… 229 Figure H1: Ammonium sulfate pr ecipitation of bovine serum albumin……………… 233 Figure H2: Ammonium sulfate precipitation of 0.10 mg/ml hemoglobin……………. 234

PAGE 17

xiv List of Abbreviations UV-vis ultraviolet-visible RBC red blood cells WBC white blood cells Hb hemoglobin HCT hematocrit HGB total hemoglobin concentration MCV mean corpuscular volume MCH mean corpuscular hemoglobin MCHC mean corpuscular hemoglobin concentration RDW red cell distribution width PLT platelet MPV mean platelet volume PBS phosphate buffered saline RPM revolutions per minute EDTA ethylenediamine tetra-acetic acid OD optical density HSA human serum albumin BSA bovine serum albumin

PAGE 18

xv 2,3-BPG 2,3-bisphosphoglycerate PSD particle size distribution RBD Rayleigh-Gans-Debye LR leukocyte reduced DTS diffuse transmission spectrum SATS small angle transmission spectrum

PAGE 19

xvi List of Symbols A( ) absorption at a given wavelength ( ) Beer’s law extinction coeffi cient at a given wavelength wavelength I0 intensity of the incident light It intensity of the transmitted light m( ) complex refractive index n( ) refractive index ( ) absorption coefficient turbidity Np number of particles D particle diameter f(D) particle size distribution Qext( ,m( )) Mie efficiency coefficient size parameter approximation of errors H covariance matrix regularization parameter

PAGE 20

xvii Using Multiwavelength UV-Visi ble Spectroscopy for the Ch aracterization of Red Blood Cells: An Investigation of Hypochromism Akihisa Nonoyama ABSTRACT Particle analysis using multiwavelengt h UV-visible spectroscopy provides the potential for extracting quantitative red blood cell information, such as hemoglobin concentration, cell size, and cell count. Howeve r, if there is a significant presence of hypochromism as a result of the concentrated hemoglobin (physiologi cal value of 33%), successful quantification of red cell va lues would require a correction. Hypochromism has been traditionally defined as a decrease in absorption relative to the values expected from the Beer-Lambert Law due to electronic interactions of chromophores residing in close proximity of one another. This phenomenon has been suggested to be present in macroscopic sy stems composed of st rong chromophores such as nucleic acids, chlorophyll, and hemoglobin. The study pres ented in this dissertation examines the presence of hypochromism in red blood cells as a part of a larger goal to qualitatively and quantatively characteri ze red blood cells using multiwavelength UVvisible spectroscopy. The strategy of the study was three-fo ld: 1) to determine the instrumental configuration that would provide the most complete information in the acquired spectra,

PAGE 21

xviii 2) to develop an experimental model syst em in which the hemoglobin content in red blood cells could be modified to various concentrations, and 3) to implement an interpretation model based on light scatte ring theory (which accounts for both the scattering and absorption components of th e optical density spectrum) to provide quantitative information for the experimental system. By this process, hypochromicity was redefined into two categories with molecular hypochromicity representing the traditional definition and macroscopic hypochromicity being an attenuation of the absorption component due to a scattering-re lated effect. Successful simulations of experimental red cell spectr a containing various amounts of hemoglobin were obtained using the theoretical model. Furthermore, successful quantitative inte rpretation of the red blood cell spectra was achieved in the context of corpuscu lar hemoglobin concentration, corpuscular volume, and cell count solely by accounting for the scattering and absorption effects of the particle, indicat ing that molecular hypochromicity was insignificant in this macroscopic system.

PAGE 22

1 Chapter 1: Introduction Multiwavelength ultraviolet-visible (UVvis) spectrophotometry is a powerful tool for the characterization of particles in suspension. With the acquisition of one spectrum, it is possible to obtain information on parameters such as particle count, size, shape, and chemical composition.1,2 Application of this tech nology coupled with spectral interpretation using the theory of light scattering allows for the analysis of particles in a large range of sizes (10-9 – 10-6 m). The method proves particularly useful in the examination of micron-sized particles due to their significant s cattering properties, especially if they exhibit a hi gh optical contrast (high refracti ve index) in relation to the background medium. For such particles, light scattering theory provides the means to interpret the combined scatte ring and absorption components of the spectrum to extract a wealth of information about the suspension system. Furthermore, the opportunity to examine the spectrum over a large wavele ngth range (190 – 1100 nm) allows for redundant analysis through math ematical corroboration of al l wavelengths, providing a high level of reliability of the elucidated values. 1.1 Blood The capabilities of multiwavelength UVvisible spectrophotometry offers the potential for the characterization of whole blood. Whole blood is a complex system with

PAGE 23

2 the major components being red blood cells (RBC, or erythrocytes), white blood cells (WBC, or leukocytes), platelets, and plas ma, each making a contribution to the whole blood spectrum (Figure 1.1). Each indivi dual component exhib its unique spectral features based on their physical characteristics that impact their optical behavior. The combination of important parameters (size, shape, chemical composition) that influence the cumulative spectral attributes of the pa rticle is referred to as the joint property distribution (JPD).3,4 0 0 0 0 00 00 00 00 0 00 00 900 00 0 Wvg 0 0 0 0 0 00 00 00 0 00 0 00 90 00 00 Wvg 0000 000 000 000 000 000 000 000 000 9000 000 00 00 00 00 00 00 00 00 00 00 Wvg 0000 00 00 00 00 00 00 00 00 00 00 00 00 0 00 00 00 0 00 00 00 Wvg x 0 00 0 00 0 00 0 00 00 00 00 00 00 00 00 00 00 00 Wvg xWhole Blood Platelets Red Blood Cells White Blood Cells Plasma Figure 1.1: Blood components and their spectral contributions to the whole blood spectrum. The spectrum of each major component (red blood cells, platelets, plasma, and white blood cells) is unique to their indivi dual physical properties such as size and chemical composition.

PAGE 24

3 The objective of the study reported in this dissertation is to obtain a multiwavelength characterizati on of red blood cells. In w hole blood, red cells dominate in number and in its contribution to the overa ll spectrum; hence it is important to achieve an accurate spectral depiction of the red cells as a necessary step to obtaining information on the other major components of whole blood such as white blood cells, platelets, and plasma. In the context of spectral analys is, the red cell suspen sion can be broken down into three pathlengths, each of which offers different levels of particle information as indicated in Figure 1.2. Each pathlength denotes a different type of hypochromism, with observed hypochromism representing the most general definition: a decrease in the optical density (and extinction coefficient) as a result of the increase in concentration of a strong chromophore relative to the value expected by the Beer-Lambert Law.5 Macroscopic hypochromism is a type of observed hypochromism where a decrease in the optical density (relative to B eer-Lambert) is seen due to the combined effects of absorption and scattering of chromophores in aggregated and encapsulated systems. Moreover, molecular hypochromism is another category of observed hypochromism where molecular charge in teraction among closely packed chromophores causes a decrease in the absorption spectrum.5 Since red blood cells contain a highly absorbing species, hemoglobin, it is necessary to acc ount for the possibility of hypochromicity to attain a qualitative op tical characterization of RBCs. The study clearly establishes two subcategories of hypochromism a nd identifies the level of si gnificance of each to the spectral analysis of red cells.

PAGE 25

4 Cuvette Red Blood Cell Hemoglobin L1L2L3•particle number •observed hypochromism•macroscopic scattering proterties(size, shape, internal structure) •macroscopic hypochromism•chemical composition •absorption and molecular scattering •molecular hypochromism Figure 1.2: Three important pathlengths to consider in the spectral analysis of red cells. Each pathlength accounts for a different set of information reflected in the spectrum of the suspension. The issue of hypochromism was addressed by closely examining two paradigms: measurement variables including instrumental configuration and the combined effect of the scattering and absorption components on the red cell spectrum. The instrumental configuration of the optical arrangement of commercially available spectrophotometers may reflect differences in the design of the optics, particularly in the angle of acceptance, which can result in different spectral features that can mislead the interpretation of the optical density spectrum. Some studies th at have disregarded the importance of the detection angle on the spectrum have attempted to explain the observed hypochromism in ways other than to consider the instrumental configuration.6,7

PAGE 26

5 The second method of investigating hypochr omism in red cells was to examine the absorption and scattering components of the re d cell spectrum in an effort to interpret the spectrum using the theory of light scattering To this end, any hypochromic effect present in the spectrum would have to be accounted for to achiev e a good interpretation of the particle parameters (eg. size, number and chemical composition). The investigation was extended beyond physiologica l red cell values by developing a protocol to modify the hemoglobin composition. A procedure using hypotonic shock (with varying incubation times) was developed to control the final cellular hemoglobin concentration. Observing cells with va rying hemoglobin content allowed for the inspection of spectral featur es and trends in relation to their changing optical characteristics. 1.2 Theoretical Modeling Theoretical modeling and inte rpretation of the red cells were performed using the Mie theory, which provides anal ysis for spherical particles.8 This approach has already proven successful in offering good spectral inte rpretation for particle systems such as polymers,1 molecular aggregates,2 platelets,9 and microorganisms.10 In the case of the red cells, the capabilities of the Mie theory were extended to obtain good values for RBC number, size, and hemoglobin concentration. The success of this interpretation model helped to identify the two cat egories of hypochromism and thei r influence, if any, on the transmission spectrum of purifie d and modified red blood cells.

PAGE 27

6 Chapter 2: Whole Blood, Red Blood Cells and Hemoglobin 2.1 Whole Blood Whole blood is a complex system of re generating cells and physiologically essential molecules which work to sustain lif e in various ways. It is responsible for transporting nutrition, oxygenating tissues, removing waste products, controlling body temperature, and maintaining hemostasis.11,12 Whole blood is composed of four major components: plasma, thrombocytes (plate lets), leukocytes (white blood cells), and erythrocytes (red blood cells). It has a viscosity approximate ly 1.5 times that of water, maintains a specific gravity of 1.050 – 1.060, and has a physiologic pH of 7.35 – 7.45.12 The cell-to-plasma ratio is on the order of 45%:55% and a normal adult sustains 70 – 75 milliliters of blood per kilogram of body weight which constitutes 7 – 8% of total body weight.11 The components of blood have been well characterized and baseline normal values have been established for each consti tuent (Table 2.1). Such accepted values are typically represented as ranges or statistical averages due to the fact that the physiological abundances of cells and proteins vary accord ing to age, sex, race, geographic locations, and pathologies. Typically, 95% of all normal individuals fall within + 2 standard

PAGE 28

7 deviations of the reported reference valu es, with only the remaining 5% normal individuals being outsi de of this range.11 Blood Component Size ( m) Volume (fl) Count (number/ l) Role Erythrocytes 7 – 8 80 – 100 Male: 4.7 6.1x106 Female: 4.2 5.4x106 O2 and CO2 transport Leukocytes 4 – 11x103 Granulocytes Neutrophils 10 – 15 -4,420 Phagocytose small organisms Basophils 10 – 15 -40 Mediate inflammatory events Eosinophils 10 – 15 -200 Allergic inflammatory response Monocytes 15 – 22 -300 Phagocytose organisms Lymphocytes 10 -2,500 Cellular and humoral immunity Thrombocytes 1 – 2 7.4 10.4 1.5 4.5x105 Maintain hemostasis Plasma ---Fluid medium in blood Table 2.1:11,13,14,15,16,17 Normal values of major blood components. Leukocytes Leukocytes, or white blood cells (WBC) c ontribute only about one percent of the total blood volume, yet plays an important role in defense against infections and phagocytosis of cell debris.11,13 Leukocytes are classified under two major categories: granulocytes and mononuclear cells (non-granulated cells). As the name implies, the granulocytes exhibit a granular cytoplasm when stained and ca n be further organized into three subcategories: neutrophils, eosinoph ils, and basophils. The mononuclear cells include lymphocytes and monocytes.12 The abundance of each leukocyte type in a healthy individual is i llustrated in Table 2.2.

PAGE 29

8 Table 2.2:12 Percent abundances of leukocyte sub-types. Among the granulocytes, neutrophils are the primary line of defense owing to their functional characteristics and large nuc leus. They will destroy foreign agents mainly by means of phagocytosis, and are a majo r contributor to inflammatory responses. Neutrophils are approximately 10 to 14 m in diameter and contain a lobulated nucleus (two to five lobes) with condensed chromati n. The cytoplasm contains granules of sizes ranging from 200 to 500 nm. These granules ar e classified as per oxidase-positive and peroxidase-negative, with the la tter typically being smaller in size (~ 200 nm diameter). An electron micrograph will show on th e average, 200 to 300 granules, with approximately one third being peroxi dase-positive granules (Figure 2.1).14 Eosinophils, which are the second most a bundant of the granulocytes, have the capacity to secrete toxic granules to help eradicate invading cells. They also aid in digesting old clots. Eosinophils are usually slightly larger than neutrophils and contain a nucleus with two lobes. The granules in the cytoplasm are larger than those in the Leukocyte Type Percent Abundance Neutrophil 62.0% Eosinophil 2.3% Basophil 0.4% Lymphocyte 30.0% Monocyte 5.3%

PAGE 30

9 neutrophils and are more uniform in size. Moreover, the granules contain crystalline structures that are readil y seen in the electron micr oscope picture in Figure 2.2. Basophils, normally low in concentrati on, increase in number during the healing phase of an inflammation. Furthermore, they release histamine, heparin, bradykinin and serotonin to help mediate the inflammatory response.12 The size of the basophils are slightly smaller than the neutrophils. The nucleus is larger than those of the other granulocytes, and typically are not as segmented as they are in neutrophils. They contain large granules that are less uniform in size co mpared to those of the eosinophils (Figure 2.3).14 Figure 2.1:14 An electron microscopy picture of a mature neutrophil using peroxidase staining. The picture shows multiple lobes of nuclei (n), peroxidase-positive granules (p+), and peroxidase-negative granules (p-).

PAGE 31

10 Figure 2.2:14 An electron microscopy picture of an eosinophil using peroxidase staining. The nuclei has two lobes, and the granules (g) contains crystalline structures (arrow). Figure 2.3:14 An electron microscopy picture of a basophil using peroxidase staining. The granules stain lightly a nd are irregular in size.

PAGE 32

11 Lymphocytes compose the major non-granul ocyte component w ith a lifespan of 100-300 days and are differentiated into three types: B-lymphocytes T-lymphocytes and natural killer (NK) cells. All are principal contributors to the immune response. The Blymphocyte recognizes antigens that cause th e cells to proliferat e and release large amounts of immunoglobulins (antibodies). The T-lymphocytes aids in B-lymphocyte function as well as exhibiting cytotoxic eff ects such as lysing virus-infected cells.12,13 Upon activation, T-cells (T-lymphocytes) can prolif erate into one of four types of cells: memory T-cells (memory for previously en countered antigens), helper T-cells (help enhance the activities of lymphocytes and phagocytes a nd stimulate proliferation and maturation of B-cells), suppressor T-cells (suppress over-activity of lymphocytes and phagocytes), and cytotoxic T-cells (kill cells targeted by specific antigen).12 The NK cells provide immunity against infectious ag ents and transformed cells. Morphologically, Figure 2.4:14 An electron microscopy picture of a lymphocyte. The nucleus dominates the cellular space.

PAGE 33

12 the three types of lymphocytes share sim ilarities, with sizes ranging from 6 to 15 m (Figure 2.4). The nucleus is round or kidne y-shaped ad occupies ~ 90% of the cell.14 Figure 2.5:14 An electron microscopy picture of a monocyte. It contains a relatively large nucleus (N) along w ith visible organelles such as th e ER (thin arrow), mitochondria (thick arrow), and lysosomes (empty arrow). Monocytes ( Figure 2.5) consti tute two to eight percen t of whole blood cells and are released from the marrow as immature cells Once in tissues, they differentiate into the larger macrophages to fight infecti ons and remove debris from blood through phagocytosis. They are more resilient than neutrophils and they can also endocytose larger particles.12 In terms of size, the monocytes large blood cells with diameters ranging from 12 to 15 m. The shapes of the nucleus vary form round to oval and lobulated. The cytoplasm contains frequent ly vacuolated granules (~ 50 to 200 nm).

PAGE 34

13 Figure 2.6:14 An electron microscopy picture of macrophage. Internal contents include the nucleus (N), ER (thin arrow), Golgi z one (G), mitochondria (thick arrow), and lysosomes (empty arrow). Monocytes will differentiate into macrophages in tissue, an event characterized by an increase in cellular size up to 22 m accompanied by an increase in the granules (Figure 2.6).14 Thrombocytes Thrombocytes, or platelets, are derived from megakaryocytes in the bone marrow and play a crucial role in maintaining hemost asis. Mature megakaryocytes are released from the marrow and subsequently fragment into platelets, which survive in the vasculature for approximately ten days.12,14 Through receptor-mediated signaling, platelets converge at the site of vascular injury and activ ate, allowing for aggregation, which is part of the clotting process. Fu rthermore, they will release granules and

PAGE 35

14 cytoplasmic contents containing essential hemo static components such as clotting factors and serotonin which promote the clotting cascade.14 Erythrocytes Erythrocytes (red blood cells – RBC) are biconcave disc oid cells containing the protein hemoglobin. Its primary functi on is to transport molecular oxygen (O2) from the lungs to the tissues and to carry carbon dioxide (CO2) from the tissues to the lungs. The composition and functions of RBCs will be discussed in detail in section 2.2. Plasma Plasma is the liquid medium of w hole blood through which the blood cells are circulated. While the cells make up approxi mately 45% of the vol ume of blood, plasma constitutes the complementary 55%. It is primarily composed of water and it accounts for roughly six liters of blood in a normal human body.11 Aside from the cells, plasma also contains a multitude of proteins, ions, lipids, and carbohydrates, many of which are responsible for maintaining the integrity of the vascular sy stem. Some major proteins include albumin, -globulins, and fibrinogen, with albumin being the most abundant. Important ions include H+, Na+, K+, Mg2+, Ca2+, and Cl-.11,18,19 The plasma-enabled fluidity of whol e blood helps the heart deliver essential biological components to all body tissues with gr eat efficiency. It allows for the body to maintain hemostasis, protect against infections deliver oxygen to tissu es, transport lipids via water soluble proteins, and remo ve catabolites, among other functions.14 This

PAGE 36

15 intricate network of components is maintained in a delicate balance of functions in the vascular system of a healthy individual. Advances in clinical and research-oriented diagnoses have helped to identify the numerous abnormalities currently seen in whole blood. Serum Albumin Human serum albumin (HSA) is the most abundant protein in plasma at a concentration of 42 g/L and serves several important biological functions within the vascular system.20 Synthesized predominantly in the liver, HSA exhibits dimensions of 134 x 41 angstroms and has a molecular weig ht of 66,439 da. At physiological pH, the protein displays a net charge of approximately -17 and is composed of 585 amino acid residues.20,21 Albumin possesses the unique capacit y to bind numerous endogenous and exogenous compounds, making it a vital transp ort protein. It binds mainly water insoluble substances, such as lipids and also potentially toxic compounds such as bilirubin, giving its transport mechanism a multi-purpose role.20, 22 Its abundance also provides the vasculature with 80% of the os motic pressure needed to maintain the integrity of cell struct ure in whole blood. Human serum albumin is composed of the following chromophoric amino acid residues: 1 tryptophan, 1 cystei ne, 17 cystines (disulfide bridges), 18 tyrosines, and 31 phenylalanines.20 The structure consists of a series of three identical domains with each of the domains connected to each other by a hinge region and a disulfide bond. Each

PAGE 37

16 domain is predominantly characterized by helic es and the three domains are arranged in a trough-like tertiary structure. Bovine serum albumin has a slightly smaller number of amino acids with 583 and possesses a molecular weight of 66,267 da. It is structurally similar to the human isoform, and chromophoric residues include 19 tyrosines, 26 phenylalanines, 18 cystines and 2 tryptophans.20,23 Albumin is an extensively studied molecu le which has been well characterized in many different aspects and it has proven to be a useful model molecule in our research (See Appendices C and H). Our studies revolve around UV-visible spectroscopy and albumin has been analyzed in terms of its opt ical behavior based on its size and chemical composition.24 2.2 The Erythrocyte Erythrocytes, or red blood cells (RBC), are dark red, flex ible, biconcave cells that carry O2 and CO2 to and from tissues respectively.14 In the early 1800’s, Francois Magendie did not heed colleague William He wson’s recommendation to dilute blood in Component Concentration (mg/ml) Total protein 371.0 Hemoglobin 361.8 Non-hemoglobin protein 9.2 Insoluble stroma protein 6.3 Protein from enzymes 2.9 Table 2.3:14 Protein content of human erythrocytes

PAGE 38

17 serum when studying them. He instead used water, and through his microscopy studies, reported that red blood cells were air bubbles.25 Since then, scientific knowledge of erythrocytes has soared as a result of advances in research technology. Red blood cells generally exhibit a diameter of approximately 7.5 – 8.7 m, (Figure 2.1) and are found at a concentration of 5 x 106 cells per microliter of blood.13,14 They owe their deep red co lor to the presence of th e iron-containing molecule hemoglobin (33% m/v).14 In fact, roughly 97% of all to tal protein in erythrocytes is hemoglobin, as illustrated in Table 2.3. Th at is, approximately 95% of the total cell weight can be attributed to the chromophoric molecule.12 Furthermore, mature human erythrocytes lack a nucleus and mitoc hondria and exhibit minimal metabolism yet maintain their existence fo r an average of 100-120 days.14,26 Figure 2.7:26 A cartoon and micrograph represen tations of the erythrocyte.

PAGE 39

18 The biconcave discoid has a thickness of approximately 2 m (Figure 2.7), a surface area of 136 m2, and a volume of ~90 fl.26 Its flat shape keeps the hemoglobin molecules closer to the membrane which allo ws for a more efficient transfer of oxygen. The resourceful functionality of the cell is a ttributed to the compos ition of the membrane and the underlying cytoskeleton. The Erythrocyte Membrane and Cytoskeleton The membrane of the red blood cell is a phospholipid bilayer (75 angstroms in thickness) in which the non-pol ar, hydrophobic tails orient to form the internal region of the membrane, while the polar, hydrophilic h eads of the phospholipids face out into the aqueous plase.26,27 The density of the membrane alone is reported to be 1.15 g/cm3.27 The contents of the red cell membrane can be broken down by mass into the following: 52% protein, 40% lipid and 8% carbohydrate.27 The major types phospholipids present in the membrane include phosphatidylcholine (28% of total phospholipids), phosphatidyl serine (13%), phosphatidylethanolamine (27%), and sphingomyelin (26%).14 Known generally as glycerophospholipids or phosphoglycerides, phosphatidylchol ine, phosphatidylserine, and phosphatidylethanolamine are derivatives of glycerol-3-phosphate, with the R1 and R2 groups being fatty acids which are associated with the first two carbons of the 3-carbon backbone via an ester linkage (Figure 2.8). The R1 group typically represents C16 and C18 fatty acids while R2 is often C16-C20 unsaturated fatty acids.28 The third carbon is linked

PAGE 40

19 to the R3 group by a phosphodiester bond with the variations of R3 defining the type of phosphoglyceride (Figure 2.9). C O OCH2CH O C H2C PO O O R3O O R1R2 Figure 2.8: The general chemical structure of a phosphoglyceride. R1 and R2 represent hydrophobic fatty acid chains and R3 defines the specific type of phospholipid. C H2N(CH3)3C H2C H2C H2NH3C H2H CNH3C O O Ethanolamine Choline Se r ine Figure 2.9: R3 groups for major types of phosphoglycerides: phosphatidylcholine, phosphatidylethanolomine, and phosphatidylserine.

PAGE 41

20 Sphingomyelin is derived from a long ch ain amino alcohol sphingosine and is classified as a sphingolipid. The deriva tive contains a ceramide group which is characterized by a fatty acid linked to the second carbon by an amide bond. The head carbon is attached to a phosphocholine (Figure 2.10). H C CH H2CPO O O OH C H C H H3C(H2C)12N H C H3C(H2C)16O C H2C H2N(CH3)3O Phosphocholine Ceramide Figure 2.10: A structural formula of sphingomyelin showing the ceremide and phosphocholine moieties. Unesterified choleste rol is also found in the membra ne in a 1:1 molar ratio with the phospholipids. Cholesterol (Figure 2.11) is a bulky ring system that modulates the membrane by hindering the movements of the fa tty acid side chains and decreasing their fluidity. The presence of the cholesterol also obstructs the potential crystallization of the phospholipids if they become packed too closely.28 Furthermore, it creates openings in the membrane which allow for the pa ssive transport of certain cations.11 Hence, the diet of an individual can influence the lipid and cholesterol concentrati ons in the blood stream which in turn could affect the quality of the erythrocyte membranes.

PAGE 42

21 CH CH2H2C CH2CH CH3H3C HO CH3CH3H3C Figure 2.11: The structural formula of cholesterol. The membrane phospholipid contents exhi bit a characteristic asymmetry. Phosphatidylserine and phosphatidylethanolamin e are found in larger quantities in the inner leaflet of the bilaye r whereas phosphatidylcholine and sphingomyelin are mostly located in the outer layer. An enzyme, ami nophospholipid translocase (flipase), has been identified to execute an ATP-dependent translocation of phosphatidylserine and phosphatidylethanolamine from the outside to the inner layer. The asymmetry can be maintained on the most part due to the slow fluidity of the membrane, and the electrostatic attractions of sk eletal proteins to the phosphatid ylserine. Alterations in the phospholipids distributions may have a significant effect on vascular normalcy. For

PAGE 43

22 example, a substantial amount of phosphatidylse rine on the external layer may incite the activation of blood clotting.14 The characteristic and func tional shape of the red bl ood cell is maintained by a network of proteins in the skeletal matrix. The erythrocyte also owes its deformability to this intricate matrix and internal viscosity, allowing the cell (~8 m in diameter) to squeeze through capillary openings as small as 3 m, or increase its volume from 90 fl up to 150 fl.11,14 The supporting network of proteins (Fi gure 2.12) in the membrane skeleton consists of spectrin and actin, among others. Spectrin, the most abundant protein of the skeleton, is a heterodimer composed of two non-identical subunits ( and ) with molecular weights of 240 kD and 220 kD respectively.29 The two subunits are intertwined in a linear fashion, measuring approximately 1000 angstroms. Furthermore, the heterodimer will interact head-to-head to form a heterotetramer, which will be the core of the meshwork in the membrane skeleton.28,30 Five to eight spectrin heteromer tails come together at a junctional complex composed of a combination of actin, protein 4.1 (band 4.1), tropomyosin, adducin, and protein 4.9 (dematin).11,31,32 The spectrin tails are directly associated with actin and prot ein 4.1. The junctional complex itself is anchored down via protein 4.1 to the membra ne by its association to glycophorin C, an integral membrane glycoprotein which carri es an antigen on its extracellular domain.11 A deficiency in glycophorin C shows a deficit of protein 4.1 and a disruption in membrane stability that leads to a mild case of elliptocytosis.33 In between two junctional complexes, ankyrin (protein 2.1) attaches to the spectrin tetramer, and through the aid of

PAGE 44

23 protein 4.2 (pallidin), the comp lex is anchored to the integr al membrane protein, band 3 (protein 3).33 This elaborate meshwork of the membrane skeleton maintains the erythrocyte’s biconcave discoi d shape as well as its flexibility which allows it to pass through small openings. Figure 2.12:11 The figure depicts the organizati on of the membrane and membrane skeleton. The spectrin heterotetramer att aches to both the junctional complex (actin, tropomyosin, adducing, protein 4. 9 and protein 4.1) and the a nkyrin/protein 4.2 complex. The junctional complex is attached to the membrane by way of glycophorin C and the ankyrin/protein 4.2 is attached via band 3. The Erythrocyte Surface The outer surface of the red blood cell has an overall negative charge due to its richness in neuraminic acid residue s primarily located on the glycophorin A transmembrane protein.14 More importantly, it is the different blood group determinants on the surface that comprise about 22 diffe rent blood group systems. Two systems, the ABO and rhesus (Rh) blood groups have the largest clinical implications.28

PAGE 45

24 The antigenic determinants for the ABO blood group system are oligosaccharides of type A, B, and H attached on surface sphingoglycolipids. The three subtypes differ in sugar residues at their n on-reducing ends with all th ree sharing the same base polysaccharide sequence. The A and B variants each exhibit a unique sugar residue attached to the base whereas the H antigen has none. The H antigen is therefore considered to the precursor to the A and B t ypes, and in the absence of A or B antigens, the individual is deemed type O. A type A individual bears A antigens on the red cell surface while the blood plasma contains antibod ies to the B antigen. A type B person would show the opposite phenotype. If anti-A se rum is introduced to type A red cells, the antibodies would cause the cells to a gglutinate, illustratin g the importance of matching up blood types for clinical transfusions.28 AB individuals carry both types of antigens on the cell surface but do not express either antibody in their plasma. The Life of an Erythrocyte Hematopoiesis is the formation of cells in the vascular system and erythropoiesis is specifically the process by which red blood cells are developed. In post-birth individuals, cells are generated from the bone marrow, where all blood cells are differentiated from a pool of common stem cells.11,12,14,15 Although less than five percent of the stem cells are dividing at one time, approximately 1x1011 cells are generated per day from a stem cell pool of around 2x106 by constant division and proliferation.11 Figure 2.13 shows the lineage of the diffe rent blood cells. The stem cell (CFU-S), deemed “pluripotential” due to its capacity to differentiate, can be stimulated into

PAGE 46

25 becoming CFU-GEMM (colony-forming unit-gr anulocyte, erythrocyte, monocyte, megakaryocyte) or CFU-L (colony-formi ng unit-lymphocyte). The CFU-GEMM will then be differentiated into one of the comm itted progenitor cells (with the aid of specific hematopoietic growth factors) for neut rophils, eosinophils, basophils, monocytes, platelets, or erythrocytes. The CFU-L will either become the Bor the T-lymphocyte.11 Figure 2.13:11 The figure illustrates the cell line age of each of the blood cells generated from the pluripotential stem cell (CFU -S). The CFU-GEMM (colony-forming unitgranylocyte, erythrocyte, monocyte, megaka ryocyte) will differentiate into committed progenitor cells, eventually producing neut rophils, monocytes, eosinophils, basophils, platelets, and erythrocytes. The CFUL (colony-forming unit-lymphocyte) will differentiate into Tand B-lymphocytes.

PAGE 47

26 As indicated in figure 2.13, erythrop oiesis shares the CFU-S to CFU-GEMM lineage with the granulocytes, monocytes, a nd platelets. From the CFU-GEMM stage, however, differentiation into the early progenitor cell blast-forming unit-erythrocyte (BFU-E), the late progenitor cell col ony-forming unit-erythrocyte (CFU-E) and subsequent proliferation are initiated by co mbinations of the granulocyte-macrophage colony-stimulating factor (G M-CSF), interleukin-3 (IL-3), interleukin-4 (IL-4), and erythropoietin.11 Proliferation of CFU-E leads to maturation stages in the following order: pronormoblast (rubriblast), basophilic normoblast (prorubricyte), polychromatophilic normoblast (rubricyte), or thochromatic normoblast (metarubricyte), reticulocyte, and erythrocyte.11,15,26 The pronormoblasts and basophilic normoblasts are large in size and volumes (up to 20 m and 300-800 fl respectively) and have clumped nuclear chromatin. Polychromatophilic normoblasts show early signs of hemoglobin production (blue-green in color) and a smaller size (10-12 m). The hemoglobin content is visibly increased in the or thochromatic normoblast as demonstrated by the cell’s pink color and it contains 2/3 of the total erythr ocyte hemoglobin. The nucleus is gradually pushed out of the cell for it to become a ma rrow reticulocyte. Residual mitochondria and RNA will help synthesize the remaining hemoglobin before they are eliminated. The size of the reticulocyte continues to decrease in size and lose RNA for approximately 24 hours post release into circulation. Once the reticu locytes are released vi a the sinuses, they will mature within a day into erythrocytes.11,26 One CFU-E will produce 16-32 adult erythrocytes.26

PAGE 48

27Metabolic Pathways in the Erythrocyte Although the red cell is devoi d of many common cellular activ ities, it still exhibits metabolic pathways essential to maintaining it s viability for its lifetime in circulation. Briefly, the Embdon-Meyerhof pathway, or gl ycolysis, accounts for 90-95% of the cell’s glucose consumption. The glucose, which is harvested externally (due to the fact that erythrocytes lack glycogen reserves), is converted to lactate producing two moles of adenosine triphosphate (ATP) for every mole of glucose reacted. The ATP is then used to maintain everything from membrane integrity to active ion pumps (Na+ and Ca2+ out; K+ and Mg2+ in). Two subsidiaries of the gl ycolysis pathway are the methemoglobin reductase pathway and the Rapoport-Leuberi ng shunt. The methemoglobin reductase keeps hemoglobin iron in a reduced state (Fe2+) and maintains the oxidized form, methemoglobin (Fe3+) to about two percent. The Rapoport-Leubering shunt takes the 1,3-bisphosphoglycerate (1,3-BPG) from the glyc olysis pathway and catalyzes it to 2,3bisphosphoglycerate (2,3-BPG). 2,3-BPG is important for the release of oxygen by the hemoglobin molecule. The hexose-monophosphate shunt produces nicotinamide adenine dinucleotidephosphate (NADPH) and glutathion e (GSH) both in their reduced forms. When the sulfhydryl groups (-SH) of hemoglobin become oxidized, GSH will work to reduce these groups, and in turn will become oxidized (GSSG). GSSG will then be reduced to back GSH with the aid of NADPH Deterring the oxidation of the hemoglobin –SH groups is important since such an event would lead to denatu ration and precipitation of the oxygen-carrier. Furthermore, GSH/ NADPH are responsible for maintaining reduced sulfhydryl groups on the membrane.11,34

PAGE 49

28Morphology of the Red Cell The characteristic biconcave shape of an erythrocyte is referred to as a discocyte.14 There have been many speculations as to why and how the cell is able to maintain this unique shape. One hypothetical model showed that the discocyte shape has lowest bending energy due to the reduced surface curvature.35 From a biochemical standpoint, the shape could be explained as an effect due to the interactions between the molecules which make up the membrane and membrane skeleton. Regardless of the cause of the shape, it cannot be argued that it is an efficient design to carry out its task. Aside from its discocyte shape, erythroc ytes may take on alternate shapes, usually due to an abnormality. Such variant forms include, but are not limited to, echinocyte, elliptocyte, and drepanocyte.14 The echinocytes are spherical, crenated cells characterized by spicule proj ections, not unlike a sea ur chin (Figure 2.14). The echinocytes will form under circumstances such as prolonged storage or uremia. Elliptocytes are oval-shaped red cells and are commonly seen in types of anemias and thallassemias. Finally, the drepanocyte (Figure 2.15) is the sickling of the cell as seen in sickle cell anemia patients.14,35 Figure 2.14:35 An electron micrograph of a discoc yte (A) and an echinocyte (B).

PAGE 50

29 Figure 2.15:35 An electron micrograph of a drepanocyte. 2.3 Hemoglobin Erythrocytes can generally be consider ed as sacks of hemoglobin, the oxygen carrying molecule which gives the cell its red color. As previously mentioned, approximately 97% of the total red cell protei n is hemoglobin with a cell mass fraction of approximately 0.33.14 The protein is found in di fferent subtypes such as A1 (most common, ~95%), A2 (~2-4%), and F (fetal).36,37 It is one of the most well-studied proteins due to its elegant complexities as we ll as its availability. It can also exhibit a multitude of abnormalities, leading to various cl inical pathologies. Moreover, its optical characteristics provide a means for the spectr ophotometric analysis of blood quality. Hemoglobin Structure and Function: Simple organisms can sustain life by depending on the diffusion or oxygen, but with tissues thicker than ~1 mm, the rate is too slow to support life. Additionally, oxygen

PAGE 51

30 solubility in blood plasma is around 1 x 10-4 M, which is too low for the metabolic needs of animals. Whole blood, which contains 150 g Hb/L can carry O2 at concentrations up to 0.01 M.36 With the average man containing approximately 1 kg of hemoglobin, its major functions aside from O2 transport include CO2 transport and buffering the blood to maintain its homeostatic environment.38 It has a molecular weight of 64,458 kD and overall dimensions of 64 x 55 x 50 .28,38 Its tetrameric nature allows for its complex yet efficient oxygen binding properties. Hemoglobin (Hb) contains two distinct entities: the globul ar portion of the protein and the heme group. The apoprotein of adult hemoglobin A1 consists of two and two -subunits (22), whereas subtype A2 is designated 22 and F is 22.38 The heme group consists of the reduced form of iron (Fe2+) which binds to the molecular oxygen for transport. The discussion on Hb characteristics will focus on the most common subtype, A1. The -chain contains 141 amino acid residues and the -chain has 146 residues, with approximately 70 – 75% of both chains arranged as a series of -helices (7 in the chain and 8 in the -chain with the helices named A-H and A-G respectively starting from the amino-terminus).14,38,39 A heme (ferroprotoporphyr in IX) is noncovalently associated with each subunit in the space creat ed between helices E and F (Figure 2.16). The heme is tucked into a hydrophobic pocket, wh ich acts as the stabil izer to the binding of the molecule to the protein subunit. That is, specific hydrophobic amino acid residues are within proximity of the heme for van der Waals contact. The hydrophobicity also

PAGE 52

31 Figure 2.16:40 The chain of hemoglobin shows the globular tertiary structure consisting of eight -helices. The heme group is tucked in the crevice between helices E and F. Figure 2.17:28 A structural depiction of the he me molecule. The protoporphyrin IX consists of four pyrrole rings linked in a he terocyclic ring system. The Fe(II) center is chelated to the four nitrogen groups. The pr oximal histidine (F8) can be seen coordinated to the iron from the bottom of the plane and th e molecular oxygen binds to the iron on the opposite side.

PAGE 53

32 repels any small polar molecules away from the ligand binding site. Furthermore, a nonpolar crevice makes the oxidation of Fe2+ to Fe3+ more difficult.14 The heme consists of an iron (II) coordinate d to the nitrogens of four pyrrole rings in a heterocyclic ring system named protoporphy rin IX (Figure 2.17). When the heme is situated in a hemoglobin subunit, the iron is al so coordinated covalent ly to the imidazole Figure 2.18:28 An illustration of the heme pocket in the deoxygenated state. The proximal histidine (F8) is coor dinated to the iron center as a part of a square pyramidal geometry. In the pocket, the heme is surrounded by mostly hydrophobic residues, which keeps out small polar molecules. The distal histidine (E7) stabili zed the ligand in the oxygenated state. In the absence of a lig and, the heme adopts a domed shape.

PAGE 54

33 nitrogen of the proximal histidine (histidine F8 ) to form a square pyramidal geometry in the deoxygenated state (Figure 2.18). When a molecular oxygen associates to the Fe in an octahedral geometry, a distal histid ine (E7) will hydrogen-bond to the ligand to stabilize it.36,38 Moreover, in the deoxygenated fo rm, the iron center is pulled by the proximal histidine approximately 0.6 out of th e plane of the heme (Figure 2.19) to form a slightly domed structure. However, when an oxygen binds, the iron is pulled back into the heme plane, causing strain in the ring st ructure. The movement of the iron center repositions the proximal histid ine and alters the distance between certain residues, such as Val FG5, and the heme (they become close r). Additionally, the proximal histidine is tilted with respect to the he me plane in the deoxygenated state, but becomes more Figure 2.19:36 When deoxygenated, the iron is pulled out of the heme plane ~0.6 whereas the iron resides in the plane in the presence of a ligand. The shift affects the tertiary structure, and ultimately, the quarter nary structure of the tetrameric hemoglobin molecule.

PAGE 55

34 perpendicular when the iron is retracted into the plane (in the oxygenated state), subsequently affecting the positi oning of certain neighboring residues.41 Such shifts within the subunit will cause the globular structure to reorganize, resulting in a conformational change.36,41 A shift in one subunit will a ffect the conformation of the juxtaposed hemoglobin subunits, and such is the basis for its cooperative properties (discussed later in this chapter). The quarternary structure involves the subunits of the tetramer being held together by a combination of direct hydrogen bonds, solvent mediated hydrogen bonds, hydrophobic contacts and salt bridges.14 The deoxygenated form is typically known as the T (tense) state and the oxygenated form is considered the R (relaxed) state.36,38,41 While the R state is stabilized by ligand binding, the T state is a more rigid form and it is stabilized by a series of sa lt bridges that involve the C-terminal residues of each subunit; residues which are otherwise free.38 Furthermore, the T state has a Val E11 in the subunits positioned on the distal side of the heme, hindering ligand binding.41 The interface also shows two stable positions that switch between the states. In the T state, Tyr C7 hydrogen bonds with Asp G1, but upon oxygenation, the R state causes a shift in the interface, which is now st abilized by a hydrogen bond between Asp G1 and Asn G4 (Figure 2.20). In both conformations the hills and valleys on the surfaces of subunits complement each other well, such that any intermediate positioning is not energetically favorable.36 Another factor that favors the deoxygenat ed state of hemoglobin is the presence of the compound 2,3-diphosphoglycerate (2,3 -DPG) (Figure 2.21). 2,3-DPG binds

PAGE 56

35 Figure 2.20:36 This figure illustrates the dynamics in the interface between the T and R forms of hemoglobin. In the T form, Tyr C7 hydrogen bonds with Asp G1. The R form sees a shift in the interface and a new interaction between Asp C1 and Asn G4. The hills and valleys of the surfaces co mplement each other in both conformations such that any intermediate would be unstable. CC C H H HOPO3 2-OPO3 2-O O Figure 2.21: Structural formula for 2,3-di phosphoglycerate (2,3-DPG). tightly to the central cavity created by the f our subunits by way of el ectrostatic attractions and hydrogen bonding. The molecule has an im portant function for the following reason.

PAGE 57

36 In arterial blood, where oxygen is abundant (pO2 ~100 torr), approximately 95% of the hemoglobin is saturated with oxygen. As it ci rculates, the red cells release ~40% of the oxygen. Due to hemoglobin’s high affinity for O2, it is the presence of 2,3-DPG that allows for the efficient delivery of oxygen.28 Hemoglobin Cooperativity The complex series of conformational dynami cs that stems from the interaction of O2 to the heme and ultimately affecting the quarternary structure plays an important role in the hemoglobin’s ability to effectively de liver oxygen. This allo steric behavior of hemoglobin can be seen in an oxygen dissoci ation curve of fractional oxygen saturation (Y) as a function of partial pressure of oxygen (pO2). Figure 2.22 shows a sigmoidal curve for hemoglobin compared with myoglobin (a single subunit oxygen carrier similar in structure to one of Hb’s subunits) which s hows a hyperbolic curve re flecting its lack of allostericity. The myoglobin curve suggests that it will not release its bound oxygen until the partial pressure of O2 in the surrounding environment drops below 20 torr. Although a single hemoglobin subunit has similar affin ity to oxygen as myoglobin, the cooperative feature of the tetrameric structure of hemoglobin helps lower its attraction to O2 with every subsequent O2 it loses. As noted before, this is due to the conformational change experienced by each subunit from the loss (or gain) of a ligand.36 The sigmoidal curve of isolated hemoglobin in itself is a significan t reflection of the regul atory feature of oxygen delivery. However, in its native environmen t (in erythrocytes), the presence of such

PAGE 58

37 factors as 2,3-BPG further modifies the prof ile of the oxygen dissociation curve. Hence, red cells show a more reduced oxygen-binding capacity than hemoglobin by itself. Figure 2.22:28 Oxygen binding curves of hemogl obin and myoglobin, with the hemoglobin curve showing a sigmoidal sh ape denoting its allosteric behavior. Hemoglobin Derivatives In its primary function as an effec tive oxygen-carrier, hemoglobin houses its iron atoms in the reduced state. The molecule, how ever, can exist in small quantities with the iron in an altered oxidation state. Additiona lly, hemoglobin exhibits the capacity to bind to ligands other than molecular oxygen. Methemoglobin is the form of hemoglobi n that exists with oxidized irons.14 In this state, the molecule loses its ability to reversibly bind molecular oxygen, hence it is not the preferred physiological form. As oxyhemoglobin is converted to methemoglobin, the color changes from bright red to a dark brown color, corresponding to shifts in the

PAGE 59

38 absorption peaks at ~410 nm and 500-600 nm.14,42 Although the mostly hydrophobic pocket in which the heme resides prevents th e oxidation of the iron to a degree, it cannot prevent it completely. In vivo, methemoglobin reductase (also known as cytochrome B5 reductase) is present to reduce the meth emoglobin in order to maintain normal concentrations at <1.5% of the total hemoglobin.43,44 Methemoglobin’s affinity for cyanide (CN) has been exploited in the de velopment of a standard assay for total hemoglobin quantification. All forms of hemoglobin with the exception of sulfhemoglobin is converted into the stable cyanohemoglobin form and assayed at 540 nm with respect to a standard curve.45,46,47 Hemoglobin in its reduced state also has potential ligand al ternatives beyond O2. Carboxyhemoglobin involves carbon monoxide (CO) binding to the heme center with an affinity that is approximatel y 200 times stronger than that of oxygen. This is due to the fact that the rate of dissociation of CO is significantly slower that O2.14 The presence of 0.1% atmopheric CO would convert 50% of the hemoglobin into its carboxy derivative and prove to be fatal within one hour as a result of oxygen deprivation to the body.48 Sulfhemoglobin is a form in which sulfur is bound to the porphyrin group of the heme and lowers the iron’s affinity to oxygen ~100 times.14 It is typically caused by, but not limited to, the ingestion of drugs, such as sulfonamides and phenacetin.14 2.4 Current Methods of Red Cell Analysis Red cells are typically analyzed for transfusion viability at blood banks by monitoring such parameters as cell counts and hemoglobin concentrations. A common

PAGE 60

39 automated method for obtaining such values is electric impedance-based hematologic analyzers like the Coulter counters.49 In these instruments, th e cells are passed into an aperture and through a constant electric current. The impe dance caused by the cells as they pass through the current results in a voltage spike which is converted into the volume of the cell. Thus, the volume of the ce ll is proportional to its induction of current change as it travels through the aperture.49,50 Cell types (erythrocytes, le ukocytes, platelets) are identi fied based on size cutoffs of their known physiological distributions, hence the counts and mean cell volume are obtained. Knowing these two values, the % hematocrit, or the packed red cell volume can be calculated by the relationship: Mean corpuscular volume (MCV)51 = % hematocrit (HCT) x 10 (Eq 2.1) erythrocyte count (x106/l) The erythrocytes are then lysed and the he moglobin is allowed to disperse in free solution. Quantification of the hemoglobi n involves the standard Drabkin’s method where all forms of hemoglobin are converted to the cyanomethemoglobin derivative and compared against standards. The resulting c oncentration of the di luted hemoglobin (Hb) is reported in g/dL. The mean corpus cular hemoglobin concentration (MCHC) is calculated by the following equation. MCHC (%)51 = hemoglobin (g/dl) x 100 (Eq 2.2) % packed cell volume

PAGE 61

40 Mean corpuscular hemoglobin (MCH) is a nother interpretation of MCHC and is expressed in picograms of Hb/cell. The a bove parameters, as the names imply, are all average values and they may or may not reflect an abnormality in the donor unit. 2.5 Erythrocytes and Multiwave length UV-Visible Analysis In the context of clinical settings, the need for quality red cells for transfusion purposes is immeasurable. For this reason, blood banks must adopt the latest technology in quality control for donated bl ood to insure that transfusion recipients are getting viable blood units and not ones that ar e infected or incompatible The current blood banking system implements multiple methods and instrumentations for different analyses including cell counts, blood typi ng, and screening of infectio ns, among others. Efforts to improve upon current technologies in the area of efficiency, accuracy and cost are constantly underway. Our proposed entry into clinical whole blood analysis revolves around multiwavelength ultraviolet-visible (uv-vis) spectroscopic analysis. Based on the knowledge of the optical properties of whole blood components (erythrocytes, leukocytes, thrombocytes, plasma), it is po ssible to obtain size, shape, number, and chemical composition information from a singl e scan. The spectrograph can be modeled as a mathematical representation based on the Mie theory to corroborate with the experimental data. This method has been demonstrated with certain types of blood analyses. The first was a blood typing met hod in which blood groups were determined by changes seen in the spectra when red cells were introduced to agglutinating

PAGE 62

41 antibodies.52,53 That is, the spectra showed their se nsitivity to changes in particle size and number. Secondly, uv-vis spectroscopy was used for the quantitative an alysis for platelet quality.9 The multiwavelength analysis successfully estimated the particle size distribution (PSD) for platelet s and also showed the potenti al for monitoring platelet activation. In terms of erythrocyte analysis, the impli cations of this technology span from the routine screening of whole bl ood to diagnosis of certain ty pes of diseases to online monitoring of blood oxygenation duri ng surgeries. It is importa nt to establish a spectral baseline for normal erythrocytes so that a ny anomalous conditions could be immediately detected. A simple scan of a patient’s w hole blood could produce information about his or her cell counts, MCV, MCH, MCHC, a nd the ability to detect any morphology variations associated with red blood cells, such as sickle cell anemia (which produces sickle-shaped red cells as a resu lt of hemoglobin polymerization).14 Infections may potentially be revealed as it has been s uggested by data (Garcia-Rubio, L. H. et al., unpublished) that blood from individuals infe cted with the malaria parasite show characteristic spectrophotometric features. Furt hermore, the ability to screen transfusable units for proper leuko-depletion, or the capac ity to confirm the viability of 5-day-old blood could enhance the effectivenes s of quality control in blood banks.54,55 In this quest to spectrally characterize the norm in whole blood, it is necessary to separate the optical behavior of each component. An accurate depiction and the defining of the spectral limits of normal erythrocytes are particularly important for a few reasons: 1) it is the most abundant cell in the vascular system, 2) it contains hemoglobin, a protein with a

PAGE 63

42 high absorptivity (extinction) co efficient, and 3) the majority of the spectral fingerprint of whole blood is due to the c ontribution of red cells. In order to use this technology as a mu ltifunctional analytical tool for red blood cell quality, it is necessary to establish the optical properties of normal red cells and how they affect the absorption and scattering of light. It is well know n that erythrocytes contain hemoglobin at a physiologic concen tration of 33%. At such elevated concentrations, it has been suggested by using similarly ch romophoric compounds (chlorophyll, nucleic acids) that the absorb ance component experiences an observed hypochromism. In this context, hypochromis m refers to a decrease in the absorbance (and hence the absorption coefficient) due to an increase in the concentration of a strongly absorbing species.56 The phenomenon has been documented in free solution as well as the absorbing species being stacked in the form of micelles.57,58 It is proposed in this study that the UV-visible characterization of red blood ce lls be approached from the context of a combined scattering and absorpti on effect based on light scattering theory. In the process of using this technology to so lve the optical behavior red cells, the problem of hypochromism will be addressed.

PAGE 64

43 Chapter 3: Multiwavelength Spectropho tometry and Its Application for the Characterization of Red Blood Cells 3.1 Features of Spectrophotometry Spectrophotometry is widely used for vari ous types of chemical and biological analysis, and by its nature span s a broad range of wavelengths. This dissertation focuses on applying the significant potential of multiw avelength ultraviolet-visible (UV-vis) spectrophotometry to characterize erythroc ytes. The spectrophotometer used for experiments, the Hewlett Packard/Agilent 8453, outputs data in the wavelength range of 190 – 1100 nm, which covers the near UV (200 – 380 nm), the visible (380 – 780 nm), and part of the near infrared (IR ) region (780 – 1100 nm) (Figure 3.1).59,60 Figure 3.1:59 Diagram of electromagnetic radiation in the UV, visible, and IR regions.

PAGE 65

44 Particle analysis requires both the acquisition of spectra l data and the ability to understand and interpret the ab sorption and scattering compone nts of the optical density spectrum. The optical density is often referred to as absorbance; however this is typically under the assumption that the ab sorption dominates and that the molecule under analysis is small in size on the order of proteins (usu ally < 10 nm) with negligible scattering. Even in this size range, light scattering o ccurs in the Rayleigh regime and must be accounted for interpretation purposes.24 While the absorption component dominates in solutions or pure liquids and solids, light scattering still makes a contribution.8 Hence a more specific terminology will be established where all raw spectra will be referred to as optical density (OD) in this dissertation. Absorption Component Electromagnetic radiation is a two-component field consisting of an electric (E) and magnetic (B) fields (Figure 3.2). Both constituents are oscillating waves and are oriented perpendicular to each other as well as to the direction of propagation. The rate of propagation of the radiation in a vacuum, more commonly referred to as the speed of light (c) is equal to 3 x 108 m/s. The unit energy of the emitted radiation is a wavelengthdependent function E = hc/ (Eq 3.1) where h is Planck’s constant (6.62 x 10-34 J-s) and is the wavelength. When the radiation travels through a medium and interacts with a chromophoric molecule, the

PAGE 66

45 energy is absorbed in a quantized manner. Th at is, the light energy must equal the energy required to allow for a specific electronic transition.60 Figure 3.2:61 A depiction of a wave of electroma gnetic radiation where the magnetic field (B) and electric field (E) are perpendicu lar to each other as well as to the direction of propagation. The wavelength is the distance of one full wave oscillation. As inferred above, the energies of UV a nd visible light absorbed by a molecule result in electrons changing energy levels. No t all electrons can make such transitions as is the case with -bonds, which requires too high an exci tation energy to effectively be transitioned. The allowed transitions that arise as absorbances are n *, n *, and *, where n is a non-bonding orbital and are orbitals found in double and triple bonds. The asterisks denote anti-bonding energy levels. The n occurs with nonbonding electrons adjacent to saturated bonds an d require energies of short wavelengths typically below 200 nm. The n occurs when a non-bonding el ectron is elevated to a low lying anti-bonding orbital, such as in the case of the carbonyl moiety in Figure 3.3, and it is characterized by relatively a low absorptivity coefficient. The transitions (Figure 3.4) are usually seen in conjugated systems and show the highest absorptivity

PAGE 67

46 coefficients. In the context of absorbance spectroscopy in the UV and visible wavelength range, the n and transitions are the most relevant.59,60 CO CO Figure 3.3:60 A representation of an n electronic transition in a carbonyl moiety. CO CO Figure 3.4:60 A representation of an electronic transition in a carbonyl moiety. Inorganic chelates, such as the heme group of hemoglobin, are sources of high absorption. Metals are strongly chelated by moieties like nitrogen that possesses an unshared pair of electrons. Hence there is the possibility for the excitation of these unshared pairs of electrons. The metal at oms themselves typically have high lying energy shells and orbitals and they have the capacity for electronic transitions. Finally, the chelate is capable of a charge transfer, or the movement of electrons from the ligand to the metal, or vice versa. These transfers show intense absorbance, owing to the ability of the electrons to transition to unoccupied orbitals of the metal ions.60 When light travels through a vessel (cuvett e) containing a solution of an optically absorbing substance, the incident or source light is attenuated as it exits the vessel. In a particle suspension, the transmitted radiation is a combination of non-absorbed incident

PAGE 68

47 light and forward scattered light. In a solution, where the absorbing species are sufficiently smaller than the wavelength and the refractive index of the species compared to that of the medium is close to unity, it is categorized in the Ra yleigh scattering regime, where the contribution of the absorption dominates over that of the scattering.8 Considering the above assumption, the Beer-Lambert law is a limit of the Mie theory as the particles become infinitely small and the medium-to-particle refractive index ratio approaches one.62 Under these conditions, the sc attering is negligible and the transmitted light exclusively describes absorbance. The Beer-Lambert relationship, typically used for solutions, can be used to calculate a wavelength-dependant absorptivity (extinction) coefficient A() = ()lc (Eq 3.2) where A() is the absorbance value at a given wavelength, l is the pathlength of the cuvette (typically 1 cm or 0.1 cm), c is th e concentration of the solution (moles/L or g/ml), and () is the extinction coeffici ent at the given wavelength.63 The extinction coefficient is defined by Bouger’s Law, ) ( 4 ) ( (Eq 3.3) where ( ) is the imaginary part of the complex refractive index representing the chromophore’s absorptive characteristics (com plex refractive index is defined in the Optical Properties section of this chapter). Absorbance in this context can be further defined as A = log (I0/It) (Eq 3.4)

PAGE 69

48 where I0 is the incident light from the radiation source and It is the transmitted light seen by the detector of the spectrophotometer. The measured data obtained by the spectr ophotometer is collected first as percent (%) transmission. The relationship be tween absorbance and transmission is63 A = -log (1/T) (Eq 3.5) where T is transmission, and when converted in terms of % transmission, A = 2 – log (%T) (Eq 3.6) Absorption in Proteins Biological systems contain proteins, many of which possess multiwavelength UVvisible fingerprints from their chemical com position. This is due to the presence of chromophoric amino acids. The major contribu ting amino acids are: aromatic residues tyrosine, tryptopan, and phenylalanine, th iol containing cysteine, and its derivative cystine (Figure 3.5).64,65,66 Tryptophan, phenylalanine and tyrosine are all relatively hydrophobic residues that have conjugated ring systems and their absorption peaks correspond to the transition. Tryptophan’s high level of conjugation in its in dole side chain results in the most intensely absorbing bands. It is a bul ky residue that occurs less frequently in proteins than other residues. Phenylal anine is a chemically non-reactive aromatic residue that can participate in hydrophobic pockets. Tyrosine is slightly more polar due to the hydroxyl (-OH) group and has the cap ability of hydrogen bondi ng for stability. Furthermore, the hydroxyl group can i onize at a basic pH and change its

PAGE 70

49 spectrophotometric profile from its protonated state. Cysteine cont ains a polar thiol (-SH) group and is able to undergo a redox re action to form a disulfide bond with the sulfur of another cysteine residue to forming cy stine. This bond is integral in the folding of proteins as it stabilizes ter tiary and quarternary structures.67 HCCOO NH3C H2OHHCCOO NH3C H2HCCOO NH3C H2HCCOO NH3C H2NH SHHCCOO NH3C H2SSC H2CH NH3COOTyrosine Phenylalanine Tryptophan Cystine Cysteine Figure 3.5: Chemical structures of the chromophor ic amino acids that contribute to the absorbances of proteins.

PAGE 71

50Scattering Component In a complex heterogeneous mixtur e such as whole blood, the scattering component has a significant presence in the op tical density spectra. This is because of the abundance of particles with diameter s on the order of micrometers and having refractive indices greater than that of the medium. When in cident light encounters one of these scattering bodies, the electric component of the electromagnetic radiation induces dipole moments within different regions of the particles, exciting them into an oscillatory motion. These accelerated charges reradiate th e energy in all direc tions and it is this radiation that is referre d to as scattered light.8 Although all media scatter light to some extent through events such as density fluctuations, the scattering caused by part icles in suspension is the most profound. Understanding how to translate the data will al low the investigator to obtain size, shape, and orientation information on the particles, given that the chemical composition of the particles is known. 3.2 Light Scattering Theory Scattering Trends When the particle is small compared to the wavelength of analysis, the scattered waves are approximately in phase. This will scatter the light in a relatively uniform manner in all directions (Figure 3.6). Howeve r, as the particle increases in size, the

PAGE 72

51 scattered waves tend to dominate in the forw ard direction. The shap e and the orientation Figure 3.6:8 Scattering patterns for a spherical particle that is small relative to the wavelength (Rayleigh scattering ) where A is scattering due to a polarized incident light perpendicular to the scattering plane, B is unpolarized incident light and C is polarized incident light parallel to the scattering plan e. The source beam originates from the left (180o) and the detector resides at 0o. of the particle with respect to the incide nt beam will further determine the scattering pattern. Figure 3.7 shows scattering mode ls of a few representative shapes and orientations. The top left is a sphere of significant diameter (~1.24 m) and it is easy to see the dissymmetry of the profile along the plane perpendicular to the unpolarized incident beam. An elliptical particle of equi valent volume (top right) maintains its axis of symmetry parallel to the beam similar to the sphere, but the scatteri ng lobes contain more oscillations. When the spheroids are orient ed in various angles with respect to the incident light, the dissymmetry is evident on all planes.68 To add to the complexity, the

PAGE 73

52 chemical composition of the particle will determine its refractive index which will have a bearing on the magnitude of reradiation.8 Another important parameter that affects the direction and magn itude of scattered light is the size of the partic le. It has previously been discussed that particles small compared to the wavelength exhibit Ra yleigh scattering and produce symmetrical scattering patterns in the forwar d and backward directions. As the size of the particle increases to be larger than the wavelength of the incident light, the profile becomes biased in the direction of forward scattering.8 Figure 3.7:8,68 Scattering profiles of large particles (~1.24 m) as a sphere, and ellipses in various orientations. The refractive index (m) of the part icles is 1.05 and the wavelength is 500 nm.

PAGE 74

53 A simulation model based on angular formulations of the Mie theory was used to predict angular scattering pr ofiles of large particles. The model accepted wavelength, diameter of sphere, the real a nd imaginary part of the refrac tive index of the particle, and the refractive index of the medium as inputs. Holding all parameters constant except for size, Figure 3.8 shows the predicted scatteri ng profile for the optical properties of Figure 3.8: Calculated scattering patterns of sphe rical particles increasing in size generated using the Mie theory, where = 500 nm, the medium is water, and the optical properties are that of bacteria.69 With increasing size, the s cattering shifts in the forward direction. o2am 02 0.4 0.6 0.8 1 30 210 60 240 90 270 120 300 150 330 180 0 o2bm 02 0.4 06 08 1 30 210 60 240 90 270 120 300 150 330 180 0 o2cm 02 0.4 0.6 0.8 1 30 210 60 240 90 270 120 300 150 330 180 0 Diameter = 100 nm Diameter = 1000 nm Diameter = 500 nm

PAGE 75

54 bacteria at = 500 nm. A bacterium is considered a “soft particle” that exhibits a small refractive index close to that of the medium compared to pa rticles like erythrocytes that show a larger refractive index and a higher visu al contrast to the medium. Results show that with an increase in size, there is a noticeable change in favor of forward scattering. Analysis of particles with larger refractiv e indices, polystyrene and alumina, yield a similar trend. Understanding this phenomenon is advantage ous to our work with erythrocytes which are approximately 8 m in diameter, or ~5.5 m in diameter in terms of an equivalent sphere. Reynolds (1975) showed the forward scat tering trend to be true for erythrocytes using the Rayleigh-Gans-Debye approximation which accounts for the shape of the scattering particle.70 A log polar plot of the s cattering intensities (at 800 nm) predicts the strong forward scat tering of the erythrocytes.70 Our calculations based on the Mie theory yields similar re sults (Figure 3.9). The figure suggests that approximately 97% of the total scattered light is scatte red in the forward direction within a 5o angle. The inset shows a semi-logarithmic plot eluc idating the oscillatory features of the scattering at larger angles. It is evident by the inset that at an angle of approximately 4o, there is an intensity decrease of nearly th ree orders of magnitude. Thus we can take advantage of the forward scattering property of macroscopic partic les to effectively interpret data captured by varying acceptance angles.

PAGE 76

55 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 020406080100120140160180 Scattering AngleNormalized Intensity 10E-08 10E-07 10E-06 10E-05 10E-04 10E-03 10E-02 10E-01 1.0E+00 020406080100120140160180 Scattering Anglelog (Normaized Intensity) Figure 3.9: Angular scattering predic tion for an equivalent sphere red cell. The large plot is a linear plot and the inset is represented as a semi-loga rithmic scale. A majority of the scattered light is directed in the forward direction with in five degrees. Mie Theory First published in 1907, Mie theory provi des an exact solution to Maxwells equations for spherical particles and contains information on the refractive index and the absorption coefficient of the materials constituting the particles, as well as, on the absorption and scattering components of the transmitted light.8,71,72 The optical density (OD), sometimes referred to as turbidity, contains various information about the particulate system being characterized. Gi ven enough information about the system, Mie theory can provide a general solution for its extinction properties with the assumption that the particles in question ar e isotropic spheres. Mie theory is widely used because of its lack of limitations in two important areas: 1) it is applicable fo r any size and 2) there are no refractive index constraints.8,72 The same cannot be said for related

PAGE 77

56 approximations although they are simpler in execution. For example, Rayleigh scattering is limited to particles with diameter (D) << (more specifically, it is typically defined as being effective for particles < /20) and the refractive index of the particle to be close to that of the medium (n/n0 ~ 1).8 The Rayleigh-Debye-Gans (RDG) approximation perceives the particle as a body containing mu ltiple scattering centers and allows for the implementation of particle form factors. However, this theory limits analysis to particles where D and n/n0 ~ 1.8 The turbidity () is defined by the equation dD D f m Q D Next p) ( )) ( ( 42 0 (Eq 3.7) where Np is particle number per unit volume, D is the particle diameter, Qext(,m) is the extinction efficiency which can be evaluated using the appropriate approximation (in our case, Mie theory), is the size parameter, m is the complex refractive index, and f(D) is the particle size distribution (PSD). The size parameter is a function of wavelength () and is defined as D (Eq 3.8) and the complex refractive index (m()) is ) ( ) ( ) ( ) (0 n i n m (Eq 3.9)

PAGE 78

57 where n( ) is the refractive index of the particle, n0( ) is the refractive index of the medium, and is the absorption coefficient of the particle. (See Appendix A for further details of the Mie equations.) The Mie extinction coefficient (Qext) is an important term that defines the sum effect of the scattering and absorption components of the particle. Hence, Qext = Qsca + Qabs (Eq 3.10) where Qsca and Qabs represent the scattering and abso rption constituents respectively. Thus Equation 3.8 can be separate d into two independent terms: dD D f m Q D N dD D f m Q D Nabs p sca p) ( )) ( ( 4 ) ( )) ( ( 42 0 2 0 (Eq 3.11) Furthermore, the absorption coefficient of chromophores in solution ( ), or BeerLambert absorption coefficient is directly related to the Mie absorption coefficient ( ) through Bouger’s law ) ( 4 ) ( (Eq 3.12) enabling direct comparison of the magnitude of the absorption coefficients for encapsulated chromophores and chromophores in solution.8 Through the understanding of the turbidity equation, it is possible to use it in a constructive manner depending on the available information. If an experimental spectrum of a single population particle system is available and the optical properties are known, the size of the particles could be calcu lated. Conversely, if enough information about the system is obtained through an alte rnative analysis, it would be feasible to construct a spectrum that would correspond clos ely with the experimental. The latter

PAGE 79

58 case is considered the “direct problem”, wh ere if the particle size, shape, PSD, and composition are given, it is possible to predic t the intensity of the irradiance in a given direction. The opposite predicament or the “i nverse problem” is far more complex as it demands the elucidation of particle pa rameters in an experimental spectrum.8 Optical Properties The complex refractive index (Eq 3.9) of a pa rticle is also referre d to as its optical properties. The real part of the refractive index, n(), is the component which describes the scattering capacity of the particle. In order for it to be meaningful, it must be observed against a reference, which in this case is the refractive index of the medium (n0()). The imaginary part of m(), (), is the absorption coefficient and is prominent in strong chromophores such as hemoglobin. Consideration of the two parts in combination with complementary parameters helps paint an optical picture of the particle. For example, small spheres absorb proportionally to their volume, whereas large metallic spheres absorb only about half of the radia tion striking their surface implying that it is a strong scatterer.72 Furthermore, the ratio of the refractive i ndex to that of the medium dictates the sensitivity of the measurement. A high ratio of n()/n0() implies a high level of visual contrast of the particle. Er ythrocytes contain a significant physiological concentration of highly chromophoric hem oglobin (33%) that account for its prominent visibility under the microscope and also its intense spectra. Additionally, the medium itself is instrumental in the outcome of the spectra l data of a suspension. It has been

PAGE 80

59 demonstrated by Garcia-Rubio (1992) that a chan ge in the refractive index of the solvent will cause shifts in optical density peaks, a phenomenon that can be readily corrected for.73 As the notation indicates, the comp lex refractive index is a function of wavelength, thus it is not the same th roughout when dealing with multiwavelength spectrographs. Obtaining a full frequency range of reliable optical prop erties dictates the accuracy of the mathematical estimations. One way is to acquire dependable optical density data on the particle being characteri zed. In the case of hemoglobin, absorbance spectrum over the entire wavelength (190 – 1100 nm) is obtained experimentally and the extinction coefficient (()) is generated. Through Bouger’s Law (Eq 3.12), () can be calculated. Using the Kramers-Kronig transform, the full multiwavelength complex refractive index can be obtained. In this relationship, if either n() or () are known, the unknown parameter may be calculated.8 Mie Theory Considerations Although the Mie theory is a versatile and widely us ed approximation, it is important to understand that it is not the definitive solution to a ll particle optics. It may, however, be a powerful tool if used within its constraint s and great care is taken in making certain assumptions. One case is the seemingly limited use of the theory on particles which are strictly is otropic spheres. It has been shown, however, that when working with particle sizes in the micron ra nge, the forward scatte ring becomes dominant (Figure 3.9). Various studies with protein aggregates, platelets, and microorganisms,

PAGE 81

60 have demonstrated that using equivalent-volume spheres prod uces reliable estimations of the transmission spectra.9,74,75 The biconcave shape of red blood cells proves to be no different as the studies in this disserta tion suggest that the equivalent sphere approximation yields depe ndable calculations. Another consideration is the relationship between part icle concentration and the light interaction among multiple particles. In theory, it would be easiest to examine the optical behavior of a single particle, but in reality, this is not the case as the particles are suspended in media in numbers. The pr oblem posed by a concentrated particle suspension is presence of the multiple scatte ring phenomenon. Given that the system is crowded enough, the scattered light will be reradiated by numerous particles, complicating the interpretation of the spectrum. In the case of erythrocytes, it has been determined that diluting whole blood or a red cell suspension to approximately 4000 cells/l provides enough separation of the cells in the media that it does not give rise to the multiple scattering problem. The value of ~4000 cells/l is not an arbitrary number. It is a concentration that produ ces an optical density value ne aring the upper limit (1.2) of the linear range of the spectrophotometer. (See Appendix B for details of multiple scattering and instrumental limits in the context of our methods.) 3.3 Spectroscopy of Blood and Erythrocytes Spectroscopy of Whole Blood Since blood is one of the most important biological fluids dea lt with in clinical settings, much is known about its properties as well as methods of its evaluation. In

PAGE 82

61 terms of optics, Angstrom was the first to use spectroscopic methods to study blood characteristics in 1855, and the spectros copy of hemoglobin came soon after in 1862.76 Since this time, advancements in this field of research have brought about various ways of depicting the optical propert ies of blood components. In te rms of optical contributions by cells, erythrocytes which compose approxima tely 99% of all blood cells, is the major contributor of spectral features. This is further compounded by the fact that the hemoglobin is a strong chromophore. It is not to say, however, that other cells such as leukocytes and platelets are spectrally invisi ble. The sensitivity of multiwavelength UVvisible spectroscopy allow for the detection of more subtle changes such as leukocytedepletion. Changes in spectral features as a re sult of platelet activati on can be detected as well, with calculations reflecting the altera tions in the platelet count and size due to aggregation.9 Moreover, the plasma, which is approximately 55% of the total whole blood volume, will contribute absorbance bands in the UV range owing to its protein content. Spectroscopy of Erythrocytes The topic of interest in this dissertati on is the erythrocyte and its behavior in a field of light. The spectroscopic study of red cells is facilitated by some important properties of red cells: 1) it contains hemogl obin, a strong chromophore, 2) it exists free in a suspension unlike tissue cells hence it can be diluted into the realm of single scattering 3) it can be isolated with relative ease due to its density, 4) there are seemingly no complex internal structur es (although the formation of transient structures of

PAGE 83

62 aggregated hemoglobin cannot be completely discounted), and 5) its properties fit within the acceptable limits of the Mie theory. In or der to intelligently scrutinize red cells and to be able to make certain well-informed assumptions in relation to spectral characterization, several important parameters must be considered: the hematocrit (i.e. concentration), hemolysis, osmolarity, aggr egation, orientation, sedimentation, and deformation of the cells.76 Hemolysis in the red cell suspension would be evident as a significant hemoglobin band (417 nm) band in th e red cell spectra and the investigator must be alert to noticing such changes. Sens itivity to hemolysis would be valuable to the quality control of whole blood in blood banks. Changes in osmolarity would modify both the size and the shape of the cell. Red cells placed in hypotonic media will expand into an ellipsoid and cells in hypertonic media will shrink. A suspension of red cells has been suggested to adopt some degree of organi zed orientation and a complete optical characterization of red cells w ould require the interpretation of these orientation effects with respect to the spectral features.77 For this reason, applying mechanical force to the system in an effort to orient the cells in some uniform fashion should manifest changes in the spectra. Additionally, the density of the cells will cause erythrocytes to sediment to the bottom of the suspension ove r time. The rate of sedime ntation, however, is low and settling takes place over a course of hours.78 Finally, the flexible membrane of the red cells allows them to deform as they flow through blood vessels. In a relatively static experimental system (suspended in a medium inside of a cuvette), no such force is present to deform the shape of the cell to any significant extent. However, Brownian

PAGE 84

63 bombardment of the membrane cause flickers in the cells that may play a role in its orientation and sedimentation rates.77 Erythrocyte modeling has been shown to fit within the framework of the Mie theory. One reason for this success is that red cells do not contain any organelles hence they may be generally treated as homogeneous scatterers. Steinke and Shepherd showed by using single wavelength scattering that th e calculated scattering cross section of the RBC can predict experimental values of the cells.79 The merit of using the Mie theory has been independently confirmed, and the th eory has also been applied to two-angle light-scattering measurements for the determ inations of hemoglobi n concentration and erythrocyte volume.80,81 The difference in our approach stems primarily from the analysis of multiwavelength spectra in the UV, vi sible, and near-infrared regi ons. Our efforts exist on a grand scale as we attempt to examine pa rticle characterizatio n across the entire wavelength range of 190 – 1100 nm. In doing so, we hope to gain the knowledge to recognize qualitative spectral features relate d to hematologic parameters as well as developing the ability to reliably quantify va lues such as cell volume, cell number, and hemoglobin concentration. Generally in particle suspension analysis, the optical density of a particle suspension is directly affected by changes in the me dium, the shape of the particle, or the chemical composition of the particle.82 The result may be a shift in maxima, an increase or a decrease in the overall elevation of th e spectrum, or the masking or unmasking of absorption peaks by the scatte ring component. Red cells fall into the regime of

PAGE 85

64 macroscopic particles where scattering eclipse s much of its absorptive features, rendering its analysis to be more complex than simply taking the sum of its parts. An important contributor to the large-partic le effect of erythrocytes is the hemoglobin encapsulated within the cells. Its strongly chromophoric qua lity gives rise to w hole particle scattering which in turn is related to its high physiological concentrations. Spectroscopy of Hemoglobin In order to spectroscopically characterize red blood cells, it is important to first examine the extinction spectrum of hemoglobin. Figure 3.10 shows the multiwavelength extinction spectra of two common hem oglobin derivatives, oxyhemoglobin and methemoglobin. The spectrum of oxyhem oglobin was obtained experimentally by members of our group, using commerci ally available human hemoglobin A0 (SigmaAldrich). The spectrum of methemog lobin was mathematically acquired.69 Both spectra agree well with previously published data.83 The oxyhemoglobin derivative exhibits characteristic bands at 270, 337, 417, 547, and 575 nm. The methemoglobin shows a larger peak around 400 nm that is shifted co mpared to that of the 417 nm oxyhemoglobin peak. Moreover, the methemoglobin no longer shows that 547/575 nm doublet, characteristic of the oxyhemoglobin. This work uses oxyhemoglobin as the primary derivative in the modeling of red cell spectra. This assumption is accurate since red cells exposed to the atmosphere become almost completely oxygenated. A small amount of the oxyhemoglobin may become oxidized to methemoglobin and this will be examined

PAGE 86

65 briefly by a version of our interpretation mode l (for the quantification of hemoglobin) in Appendix G. 0 2000 4000 6000 8000 10000 12000 14000 1902903904905906907908909901090 Wavelength (nm)Extinction Coefficient (cm^2/g) oxyhemoglobin methemoglobin Figure 3.10: Extinction spectra of oxyhemoglobin and methemoglobin. Each has unique features that reflect the ch aracteristics of the heme group. Hypochromism Although erythrocytes can be describe d as translucent biconcave ‘bags’ containing a homogenous distribution of hemogl obin, this simplicity is deceptive when considering its light scatteri ng properties. Successful ch aracterization a nd modeling of red cell spectra begins with the acquisition of reliable optical properties for hemoglobin and knowledge of the constraints of our mode l system. Moreover, when dealing with a strongly absorbing protein such as hemoglobin at high concen trations, we cannot ignore

PAGE 87

66 the possibility of the significance of a hypochromic effect reported in similarly chromophoric biomolecules like chlor ophyll and deoxyribonucleic acid (DNA).84,85,86,87 Hence, observed hypochromism is defined as any decrease in the observed absorption coefficient relative to the solution spectru m of the same chromophore (Beer-Lambert). Bear in mind, that this is a sp ecific usage of the term and is different from the decrease in absorbance seen with a proportional decrea se in the chromophore concentration. Early perceptions of the hypochromic e ffect of biomolecules surfaced in the 1950s and 1960s which largely dealt with the stacking behavior of polynucleotides with respect to optical density.86,88 Moreover, chlorophyll in ch loroplasts was reported to cause permutations in transmission spectra.85 Molecular hypochromism was defined by these studies to be a decrease in absorption (and extinction) due to changes in electronic interactions from altered pr oximity between chromophores.5 The studies typically dealt with chromophores in polymeric structures where one absorbing center would shield an adjacent absorbing center such that the sec ond molecule is seeing a local field produced by the neighboring molecule, not the field from the incident light. Although more than one type of interaction influence the change in optical density, the predominant effect is thought to be the Coulombic in teraction between th e dipoles of the molecules. Lawley (1956) showed this phenomenon with DNA when he compared two different arrangements of the poly nucleotides (Figure 3.11).88 The band of greater amplitude represents DNA in its denatured state. Wh en the nucleotides are stacked in a more ordered double-helix, however, th e extinction coefficient decreases by up to 35%.

PAGE 88

67 Vekshin (1999) showed a similar effect by the empirical stacking of adenines into an oliginucleotide.89 The result is illustrated in Figure 3.12 where the X-axis is the adenine count of poly(A) and the Y-axis is the optical density. As the oligonucleotide chain was increased in length, the optical density at wave lengths of 210 nm and 260 nm Figure 3.11:88 An example of molecular hypochro mism exhibited by DNA in a double helical configuration. When denatured into a random coil, the DNA shows a higher extinction band. decreased by as much as 28%. Vekshin proposed a model that accounts for the hypochromic phenomenon not in terms of dipol ar interactions but as a probability function describing the likelihood of the kth chromophore in a stack seeing and absorbing the light that is not shielded by an adjacent chromophore.90 The probability of absorption of the first chromophore of th e stack as a function of wavelength (P1()) is defined as

PAGE 89

68 P1() = ()/S1 where () is the absorption cross-section and S1 is the effective geometric area of the chromophore. The probability of light absorption by the k-th chromophore is defined as Pk() = [1-P1()]k-1P1() and the probability of light ab sorption by the whole stack is P0 = P1() + P2() + ….. Pk() Figure 3.12:6 The effect of chain length of a polyadenine oligonucleotide on optical density, where the x-axis is the chain length a nd the y-axis is the optical density. With increased stacking, a molecular hypochromic ef fect is seen, decreasi ng the optical density by up to 28%. 210 nm 260 nm

PAGE 90

69 In Figure 3.13, Vekshin showed that hi s model was able to account for the molecular hypochromism of stacked poly(A) oligonucleotides.6 Curve 1 shows the extinction spectrum for of adenosine and curve 2 is the spectrum for penta-adenosine phosphate. The stacked nature of the pent a-adenosine properly shows the expected hypochromism. Furthermore, the calculated pr ediction of the penta-adenosine (curve 3) fits the experimental data fairly well. Figure 3.13:6 Theoretical vs. experimental spectra of poly-adenosine stacks. Curve 1 is experimental data of single adenines and curve 2 is that of penta-adenosine stacks. The molecular hypochromism is evident in the stacks. Curve 3 is a theoretical calculation of the penta-adenosine spectrum using Vekshin’s model. In terms of hemoglobin, its high extinc tion coefficient raises the question of potential molecular hypochromism. At low c oncentrations of hemoglobin on the order of spectral dilutions (< 15 mg/ml) the molecules are dispersed enough so that no significant stacking or dipolar interactions are present. With increasing concentrations approaching the physiological value of ~33% (330 mg/ml), the molecules re side significantly close to

PAGE 91

70 one another. Nonetheless, cont rary to the fixed nucleotides of a helix where the partially stacked bases are in Van der Waals contact, hemoglobin encapsulated in red blood cells presumably exist as independent proteins with their intermolecular distance estimated to be 62 – 75 36,91 (compare to hemoglobin dimensions of 64 x 55 x 50 ).28,38 Under such conditions, random collisions are prev alent, however, the effect of organized molecular hypochromism is questionable. Studies based on the Mie theory done by our group predicted a decrease in the absorption component of hemoglobin when encap sulated in erythrocytes, an event that can be accounted for by a second classi fication of observed hypochromism: macroscopic hypochromism. This is defined as a phenomenological decrease in the absorption due to the combined effects of absorption and scat tering of chromophores in aggregated and encapsulated systems. The effect is directly related to the size of the particle and the refractive index of the particle both of whic h describes whole particle scattering. Such effects could be implicated in erythrocytes, given their size and chemical composition. Mie theory converges to the Beer-Lambert law under the conditions that the molecular size approaches zero and the particle-to-medium refractive index ratio (n/n0) approaches one.73,92 This means that the optical density of the system is a direct result of the absorption component, with the scattering comp onent becoming negligible. However, as the encapsulated hemoglobin concentration is increased, our Mie-based simulations suggest that a scattering-related de crease in absorpti on transpires. Historically, spectroscopic work has been done on red blood cells with mixed results due to conflicting interp retations of models and/or ex perimental systems. Kramer

PAGE 92

71et al. (1951)93 suggested that non-hemolysed bloo d does not follow the Beer-Lambert Law due to its scattering pr operties and “by the magnifi ed absorption of light by intracellular hemoglobin due to th e refraction of light between and within erythrocytes.” The paper attributed part of the attenuation of the transmitted light to an increase in light absorption. Anderson and Se kelj (1967) observed undilute d blood in an integrating sphere spectrophotometer.94 By implementing a model correcting for the multiple scattering effects of the concentrated part iculate system, Anderson and Sekelj showed that the absorption of light was the same in both an erythrocyte and a hemoglobin solution. In contrast, our st udies using the Mie theory sugge sted that the high refractive index difference between the erythrocyte a nd the medium resulted in the dominant contribution of the scattering component w ith very little chro mophore sampling by the light. The deconvolution of spect ra by means of light scattering theory supports this idea. The question of hypochromicity bears multiple layers: 1. In an encapsulated system such as re d blood cells, do the internal chromophores exhibit molecular hypochromism as sugge sted by other studies or can it be explained by the scattering theory as macroscopic hypochromism (or a combination of the two)? 2. In studies examining the hypochromism of red cells and related macroscopic particle systems, many fail to detail th e instrumental configuration of their experiments. That is, the angle of ac ceptance of the detector is not properly defined, hence what has been perceive d as molecular hypochromism may be an

PAGE 93

72 issue of improper interpretation of data in the context of instrumental configuration. Considerations of Instrumental Configuration The interaction of light with particle s results in absorption and scattering phenomena; particles scatter light in all dire ctions and, depending on the experimental set-up, changes in the pr operties of the scattered light rela tive to the incident light may be measured as a function of the observation angle. For finite aperture detectors, the angular resolution of the scattered light depends on the acceptance angle () of the detector. Therefore the detector respons e is proportional to a weighted average of the scattered intensity over the angles defi ned by its aperture; a limiting case being measurements with an integrating sphere where essentially all sc attered light is measur ed and subtracted from the transmitted light to obtain estimates of the attenuation due to absorption. The angle of acceptance is defined in Figure 3.14 and it is related to the radius of the detector and the distance of the sample to the detector, where 0o is the direction of the incident light. The importance of the accepta nce angle stems from the inherently pronounced light scattering properties of macroscopic particles as they change in size and refractive index. Associated with such change is the angular dependence of scattering and the effectiveness of the data interpretation would be a direct correlati on to the awareness of photometer design. Latimer (1975)95 detailed the importance of considering the angle of acceptance of the detector, however hi s work has been largely ignored.

PAGE 94

73 Optical scattering analyses of erythrocytes have typically been limited to a single wavelength79 or to a small number of select wavelengths,80,81,96 while multiwavelength representations have been reported as diffuse spectra.6,7 Diffuse transmission spectra (DTS) results from the acquisition of a la rge amount of scatte red light at wide Figure 3.14: Schematic representation of the a ngle of acceptance. The size of the detector radius determines the amount of scattered light captured by the spectrophotometer. angles and is typically defined by the angle of acceptance of the detector.82 The collection of diffuse scattering plays an important role in the features of an optical density spectrum of a macro-particle system. Light collected by a small acceptance angle shows a general increase in the OD and a larger addi tion of the scattering component compared to that obtained with a larger detector angle.97 In contrast, data fo r red cells reported as a DTS shows an overall decrease in the intensity of the spectrum. This is because as scattered light is collected at more angles, th e light is averaged out and is subtracted out of the transmission spectrum. Since a small angle transmission spectrum (SATS) collects

PAGE 95

74 less of the scattered light, the weighted averag e of the uncollected sc attered light is not subtracted from the transmission spectrum. Th erefore, it is possible to infer the amount of light scattered at angles greater than and to successfully deconvolute the spectrum into its absorption and s cattering components. A misconception in the use of OD data of suspensions is that the scattering component is a confounding constituent devoid of any significant information, and thus eliminated by a simplified scattering correction, such as in the case of Leach and Sheraga (1960).98 Spectra of particle suspensions corre cted in this manner typically reveal absorption bands of decreased magnitudes and introduce a need for an empirical correction factor to provide re levant quantitative informati on. By accounting for both the absorption and scattering components of the spectrum using the light scattering theory, apparent alterations in abso rption can be justified. Defining the Problems Although the task of spectroscopically ch aracterizing red blood cells seems quite straight-forward, there are a num ber of confounding issues asso ciated with fact that the cells are significant scatte rers of light. These issues are detailed here. 1. It is necessary to determine the best instrumental configuration for the examination of our particle suspension. Since red cells scatter light, it is important to know the accep tance angle of the spectr ophotometer, and at what acceptance angle the OD spectrum contains the most complete absorption and scattering information.

PAGE 96

75 2. The examination of molecular hypochromism in a hemoglobin solution is not possible due to the limitations of the in strument and the high extinction values (see figure 3.10). The upper concentration limit for the small doublet peaks (500 – 600 nm) using a 1 mm pathlength cuvette is ~14 mg/ml. The upper limit for the larger 417 nm band is even lower at ~1.5 mg/ml (1 mm pathlength). This is not close to the physiological con centrations of ~330 mg/ml. 3. Obtaining spectra of free hemoglobin is not a good representation of its optical behavior when encapsulated within red cells. Hemoglobin molecules free in solution are small, Rayleigh scatterers a nd their spectra are represented by their dominant absorption features. However, when they are encapsulated, they affect the refractive index of the cell, contributing to the m acroscopic whole particle scattering properties of the cell. 4. An effective experimental model system needs to be implemented in order to optically characteri ze the red blood cell and to a llow for the evaluation of hypochromism. Corroborative methods for im portant particle pa rameters such as cell count, hemoglobin concentration, a nd cell size must be established to complement the experimental system. 5. A dependable theoretical model relying on li ght scattering theory is necessary to evaluate the experimental data. Mie theory, based on spherical particles of equivalent volume, has been successfully te sted on particles of smaller refractive indices (platelets, microorganisms). The theory must be tested on red blood cells for compatibility.

PAGE 97

76Experimental and Theoretical Approach With the problems clearly defined, a multi-pronged approach was devised employing experimental methods in combinati on with theoretical in terpretation. This approach is outlined as follows: 1. Red blood cell samples were examined in commercially available spectrophotometers bearing different acceptance angles to demonstrate the importance of optical confi guration on spectral features. 2. Red blood cells were permeabilized and re sealed to obtain modified hemoglobin concentrations. 3. Lipid vesicles (liposomes) were used to encapsulate varying concentrations of hemoglobin or other model pr otein. (See Appendix C) 4. A mathematical model based on the light sc attering theory was used to simulate and interpret spectra of the experimental model systems. An examination of the significance of in strumental design was performed using three spectrophotometers with different angles of acceptance. Purified red cells diluted to the same concentration under the same conditions were analyzed on all three spectrophotometers and compared. Each spect rum was also compared to a spectrum of free hemoglobin in solution with the mass held constant. The results obtained from this study allowed us to determine the optimal phot ometer design for red cell characterization and to verify the reliability of previously published studies.

PAGE 98

77 The next proposed experimental approach was to isolate red cells and manipulate the hemoglobin concentration across vari ous ranges to examine its spectral manifestations. A modified protocol was established for this purpose. The red blood cells were permeabilized in a hypotonic buffer solution to cause the cells to spill out hemoglobin in different amounts. The partially drained cells were then forced to reseal by restoring the tonicity of the solution. The result was red cells containing modified amounts of hemoglobin. It may appear that the characterization of unmodified eryt hrocytes alone is enough to answer most questions, but it will no t present the complete picture. If our knowledge is limited to normal physiological para meters, applications of this analysis cannot be extended into abnormal values. It is also in the interest of pure science to broaden the limits of the study as wide as po ssible. Hence the target of the red cell modifications was to achieve encapsulated hemoglobin concentrati ons in the medium range (MCHC ~ 15-20% w/v) and in the low range (MCHC ~ 5-10% w/v) to characterize a trend, if any, in the absorp tion and scattering com ponents of the optical density spectra. A theoretical model based on the Mie theo ry was used to analyze the spectra acquired from the modified red cells. Usi ng the particle parameters obtained from a hematology analyzer (see Chapter 4), the model was used to simulate features and trends seen in the experimental data. Moreover, the Mie theory was extended to interpret the spectral data to elucidate values for particle parameters such as pa rticle count, particle size, and hemoglobin concentration.

PAGE 99

78 Finally, the possibility of using a lipos ome model to encapsulate hemoglobin and model proteins was explored (Appendix C). Results showed, however, that the liposomes were hard to characterize optically due to their multi-lamellar nature and small size (10-9 m diameter). Although extended work on the liposome model would have resolved such issues, the modi fied red cell approach in comb ination with the theoretical model offered conclusive answers to the prob lems defined in the previous section.

PAGE 100

79 Chapter 4: Spectrophotometry of Puri fied and Modified Red Blood Cells 4.1 Hypotonic Modification of Erythrocytes Hypotonic modification of re d cells has been used for various purposes including loading the cells with isotope-labeled hemogl obin, hemoglobin S and C, dextran, ferritin and enzymes.99,100,101,102,103 The biological tool has de monstrated its usefulness in different ways, ranging from model systems of membrane behavior to delivery devices for enzyme therapy. The manipulable characte ristics of the red cell membrane proved to be advantageous in modifying the encapsulate d hemoglobin concentrations thus we have developed a protocol excl usively for the purpose of our investigation. The erythrocyte membrane under native c onditions is normally not permeable to large proteins such as hemoglobin. U pon addition of a hypotonic buffer, however, the change in the osmotic pressure on the outside environment will cause water to enter through the semi-permeable membrane and inst igate cell swelling. The osmotic pressure () can be defined by the equation =MRT where M is the molarity of the solution, R is the ideal gas constant (0.0821 L-atm/mole-K), and T is temperature in Kelvin.63 Thus in a hypotonic medium, the internal osmotic pressu re of the cell will become greater than that of the medium (int > ext), initiating the flow of water into the cell (Figure 4.1). When the internal pressure is great enough, the membrane produces holes of roughly 200

PAGE 101

80 500 allowing the passage of large molecules.101 Seeman (1967) reported that the holes were transient and exist only fo r 15 20 seconds following the hypotonic shock.104 During this time, a chemical gradient generates th e flow of material in and out of the cell. Tonicity is then restored to isotonicity (0.9% ionic strength) to assist in the resealing of the cell.100,101,113 Hb Hb Introduction of hypotonic bufferCritical step! Incubation, restoration of tonicityint> ext Figure 4.1: Illustrated representation of the hypotonic permeabiliza tion of red blood cells. The final hemoglobin concentration in th e restored cells is controlled by the time of incubation after the cells are permeabilized. According to a study by Bodemann a nd Passow (1971), the hypotonic shock procedure produces three types of cells: type I are cells that reseal relatively quickly following hemolysis, type II cells reseal after re storation of the media to its original ionic strength, and type III are cells that remain lysed and never reseal.113 Temperature conditions during the experiment aff ect cell resealing as well. At 0 oC, there are virtually

PAGE 102

81 no type I cells but mostly type II whereas experiments done at 37 oC show that the proportion shifts in favor of type I cells. Furthermore, the permeabilization step performed at 0 oC and 25 oC show the presence of approximately 38% type III cells, but cells modified at 37 oC show that 75% of the cells are type III. The purpose of our red cell modification wa s not to load the cells with foreign molecules, but to reduce the internal hemogl obin concentration for spectral interpretation (i.e., hemoglobin out, H2O in). Thus a modified protocol was established. The method had three distinct steps: 1) hypotonic shock, 2) incubation, an d 3) restoration of tonicity for resealing (Figure 4.1). It was found that the resulting hemogl obin concentration of the restored cells could be controlled by managing the vol ume of hypotonic buffer, and the incubation time. For a desired MCHC in the low range (~ 0.05 – 0.15 mass fraction), a relatively large volume of buffer was used for permeabilization and the sample was incubated for 30 minutes on ice prior to the re sealing step. In this case, the hemoglobin concentration gradient is large and the low temperature maintains the membrane holes while the gradient is allowed to equilibrate. If a higher MCHC was intended, smaller volumes of hypotonic buffer and shorter inc ubation times were used (such as a 1:1 volume ratio of packed red cell to hypotonic buffer). (See Material s & Methods section for details.) The ability to obtain enca psulated hemoglobin co ncentrations from physiological values (~ 0.33 mass fraction) down to very low mass fractions of ~ 0.05 provides a valuable opportunity for the funda mental investigation of hypochromism.

PAGE 103

824.2 Materials and Methods The UV-Visible Spectrophotometer The spectral measurements represented in this dissertation were performed predominantly on a Hewlett Packard/Agilent 8453 diode array spectrophotometer. The unit houses a single levered cuvette slot (Fig ures 4.2 and 4.3) a nd generates optical density values across a wavelength range from 190 nm to 1100 nm. The spectrophotometer is interfaced to a pers onal computer loaded with the Agilent Chemstation software which is capable of customizable spectral processing. Figure 4.2:105 The front view of a Hewlett Pack ard/Agilent 8453 Spectrophotometer Figure 4.4 illustrates the optical arrangemen t of the spectrophotometer. The radiation emissions come from two sources: a de uterium lamp and a tungsten lamp. The deuterium lamp emits light radiation over a wavelength range of 190 nm to 800 nm and the tungsten lamp contributes radiation in the higher wave lengths from 370 nm to 1100

PAGE 104

83 nm. The wavelength range of the deuterium lamp corresponds to the ultraviolet and visible regions whereas that of the tungsten represents the visible and the near-infrared regions. Figure 4.3:105 Cuvette slot of an HP/Agilent 845 2 spectrophotometer with a levered locking mechanism. As the source radiation is emitted, the source lens collimates the light aimed to pass through the sample. The shutter opens to allow the light to propagate towards the sample and close as soon as the measurement is finished. Once the incident light passes through the sample, the transmitte d light is passed through a spectrograph source lens which refocuses the light towards th e slit. The slit is a small aperture which

PAGE 105

84 limits the size of the incoming beam and prepares it for the grating, which in turn separates the light into its full array of wa velengths. The diode array detects the full spectrum of wavelengths and transmits the in formation to the inte rfaced computer where it can then be manipulated for analysis. Figure 4.4:105 The optical machinery of the HP /Agilent 8453 spectrophotometer. Types of Spectrophotometers Used The Agilent 8453 spectrophotometer with a 2o angle of acceptance was the primary instrument used for data acquisition a nd analysis. However, an important aspect of our investigation related to apparent hypochromism was su spected to be a result of instrumental configuration. We therefor e employed two altern ate spectrophotometers with varying setups. Diffuse transmission sp ectra were obtained using the Perkin-Elmer Lambda 900 (Perkin-Elmer, CT) spectrophotome ter. The sample cuvette was placed in front of an integrating sphere module and the diffuse spectra was acquired by the light being transmitted across the sphere (acceptance angle > 2o). The Perkin-Elmer Lamda

PAGE 106

85 18 fitted with an integrating sphere modul e (RSA-PE-18, Labsphere, NH) was used to collect essentially all of the light scattered by the suspension. Care and Preparation of the Cuvette The cuvettes used in this project are qu artz cuvettes (Starna Cells Inc, CA) which allows for sample reading in the ultraviolet, visible, and near-infra red wavelength ranges. Cuvette capacities are typically 3 ml and 1 ml, both with a 1 cm pathlength. The cuvettes are handled and stored with care, cleaning as often as necessary to prevent build-up of protein deposits on the walls. An effective cleaning method is sonication in dilute soap water for approximately 30 minutes followed by a methanol wash to eliminate any waterinsoluble contaminants. To ensure that this method does not lead to surfactant build-up, an occasional and brief (1 minute) treatment with chromic acid is recommended. In between experiments, the cuvettes are routinel y filled and stored with deionized water to prevent any residual proteins and other contaminants from drying and adhering strongly to the walls. Such a preventative measure facilitates the cleaning process prior to each experiment. Preparation for Spectrophotometry Prior to experiments, the spectropho tometer was switched on and allowed to warm up for approximately 20-30 minutes, whic h is a conservative time frame compared to the manual’s suggestion of 15 minutes. Th is allowed the lamps to reach their full intensity and provide reliable spectra. Once the lamps were ready, a blank was taken

PAGE 107

86 without a cuvette inserted in to the compartment to provide the spectrophoto meter with a point of reference. A good blank should show a relatively un iform baseline throughout the spectrum (Figure 4.5). Figure 4.5:105 A representative blank spectrum used to correct for ambient conditions prior to sample measurement, where the X-axis is the wavelength in nanometers, and the Y-axis represents optical density units. Spectrum of Deionized Water Prior to taking the first sample spectrum of the day, a clean cuvette was rinsed repeatedly with deionized water and then was filled to approximately two-thirds of its capacity with the deionized wate r (the level must be high en ough for the incident beam to go through the sample).

PAGE 108

87 The spectrophotometer was configured to an integration time of 15 seconds. Since the HP 8453 spectrophotometer takes te n complete multiwavelength scans over one second, a 15 second integration time produces a spectrum which is an average of 150 consecutive scans. The output file provides the mean and sta ndard deviation of the total number of scans per wavelength. The deionized water was scanned in th e spectrophotometer and the resulting spectrum served two purposes. The first was to guarantee the proper functionality of the instrument by ascertaining the consistenc y of the background water spectrum from experiment to experiment. The second pur pose was to evaluate the cleanliness of the cuvette. Figure 4.6 shows a representative spectrum of deioni zed water in a clean cuvette. The optical density value at 190 nm should typically be below 0.25 AU. The peak located at approximately 980 nm is charac teristic of water and is seen in spectra taken with a 1 cm pathlength cuvette. A 1 mm pathlength cuvette is not sufficient to show a prominent peak. Evidence of impur ities would usually show in the absorption range roughly between 200 and 500 nm. This is not to say that ab sorptions do not occur past 500 nm, however, Soret bands of chromoph oric proteins are more commonly in the UV and low visible wavelength range, and sl ight contamination on the walls of the cuvette would show up as small bumps or elev ations of the baseline in that region. Background Correction Once the cuvette was verified to be clean, the spectrum for the solvent background was established so that the bac kground correction could be implemented to

PAGE 109

88 the spectral analysis. The cuvette was washed a few times with the solvent which is most commonly isotonic (0.9%, pH 7.0-7.2) blood bank phosphate buffered saline (PBS) obtained from Nerl Diagnos tics, RI. The PBS was composed of 0.150 M sodium chloride, 0.006 M sodium diphosphate, and 0.002 M potassium phosphate. The PBS was 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1902903904905906907908909901090 Wavelength (nm)OD Figure 4.6: Representative spectrum of deionized wate r. This is an important step taken prior to any experimentation to ensure the cleanliness of the cuvette. Any contamination will show up as slight peaks, typically in the UV and low visible wavelengths. scanned for 15 seconds in the spectrophotom eter and obtained a spectrum represented by Figure 4.7. The water peak at around 980 nm wa s still evident. The ionic interactions of the sodium chloride (NaCl) produced an in tense, saturated peak below 210 nm. The saturation occurred when the absorption in tensity exceeded the linear range of the

PAGE 110

89 instrument (upper limit: 1.2 OD), hence this region lacks reliable information about the system (see Appendix B for explana tion of instrumental linearity). The saline spectrum was saved as an auxiliary file in the Hewlett Packard Chemstation software. The software was cust omized so that this background (auxiliary) spectrum is automatically subtracted from each scanned sample. The auxiliary spectrum was updated three or four times during the experiment to account for any unforeseen changes in the instrument (i.e. lamp intensity ), the cuvette, or ambient lighting in the room. The most common sign for the spectru m to be updated was when parts of the 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 1902903904905906907908909901090 Wavelength (nm)OD Figure 4.7: Representative spectrum of isotoni c (0.9%) PBS (pH 7.0-7.2). Features include the water peak around 980 nm and a saturated peak below 210 nm.

PAGE 111

90 baseline of the processed spectrum dipped belo w zero. Visual examination of the solvent in the cuvette was also conducted consistently to ensure that any s catterers such as dust particles were not present to elevate the baseline of the auxiliary spectrum. Hematology Analyzer The major method of experimental corroboration was pe rformed with the hematology analyzer model 9110+ (Serono-Baker, PA). The instrument is primarily a cell counter, giving cell number and cell volumes by means of electrical impedence. Reagent flows through an apertu re between electrodes that c onduct a constant electrical current. The homogeneity of the reagent (f ree of contaminations) keeps the current consistent and establishes a baseline. As biological cells are introduced into the system and flow through the electrical field, they act as partial insulators and temporarily increase the electrical potential. The magn itude of the circuit resistance is directly proportional to the cell volume and is compared against a calibration to give the mean cell volume. The number of cells passing thr ough the aperture per unit sample volume is also quantified to yield the cell counts. A di scriminating threshold system for size ranges is used to categorize the results into red cells white cells, and platelets, and eliminate any possible detection of debris.106 A modified Drabkin’s solution is us ed for quantification of hemoglobin.106 The red cells are lysed and the hemoglobin is converted to cyanomethemoglobin by the solution. An optical density reading is taken at 540 nm and compared against a

PAGE 112

91 calibration, where the absorban ce value is directly proportion al to the concentration of hemoglobin in the blood sample. The readout of the Serono-Baker provides the following information (Figure 4.8): Figure 4.8: A sample readout obtained from th e Serono-Baker hematology analyzer. where WBC is white blood cell count, RBC is red blood cell count, HGB is hemoglobin concentration in solution once the cell is lysed, HCT is hematocrit, MCV is mean corpuscular volume, MCH is mean corpuscula r hemoglobin, MCHC is mean corpuscular hemoglobin concentration, RDW is red cell dist ribution width, PLT is platelet count, and MPV is mean platelet volume. Although the instrument is typically used to analy ze physiological values in clinical settings, it can reliably detect values beyond normal ranges. The linear ranges of the analyzer for some important parameters ar e defined in Table 4.1. The error of the instrument is defined per parameter in th e brochure of the Para 12 Multi-parameter

PAGE 113

92 Assayed Hematology Control (Str eck Laboratories, La Vista, NE) and are shown in Table 4.2 in the context of normal physiological values In Table 4.3, a sixreplicate analysis of the same whole blood sample was obtai ned from the Serono-Baker system. Linearity Range RBC 0.45 – 7.26 mill/l HGB 0.1 – 30.0 g/dl PLT 20 – 1071 thsn/l Table 4.1: Linearity ranges for some important parameters of the SeronoBaker hematology analyzer. Mean + error RBC 4.17 + 0.020 mill/l MCV 85.6 + 4.0 fl MCHC 33.9 + 2.3 % (w/v) HGB 12.1 + 0.4 g/dl PLT 237 + 30 thsn/l Table 4.2: Error of the parameters of Ser ono-Baker hematology analyzer outputs as defined by the Para 12 Multi-parameter A ssayed Hematology Control. The control represents normal physiological values. Mean + error RBC 4.99 + 0.077 mill/l MCV 91.9 + 0.4 fl MCHC 32.3 + 0.4 % (w/v) HGB 14.8 + 0.1 g/dl PLT 216 + 3 thsn/l Table 4.3: Error of the parameters of Se rono-Baker hematology analyzer outputs calculated from 6 replicates of the same sample.

PAGE 114

93 The manual analysis typically showed smalle r error compared to those reported by the control manufacturers, with the exception of the RBC count. It should be noted, however, that six replicates of one patient sample does not necessarily represent larger test batches. Drabkin’s Hemoglobin Determination Assay A manual assay was employed throughout th e experimentation for two reasons. The first was to supplement and verify the hemoglobin concentrations obtained by the Serono-Baker hematology analyzer. Th e second was to acquire hemoglobin concentrations for steps in which use of th e hematology analyzer was not feasible. For example, the hematology analyzer only accepted samples in the form of a particle suspension. Therefore, free hemoglobin soluti ons not encapsulated in cells could only be analyzed using the manual Drabkin’s assay. The kit used for the manual Drabkin’ s hemoglobin assay (HG980) was obtained from Randox Laboratories Ltd, UK. The met hod used a colorimetric assay that was standardized by the Intern ational Committee for Standa rdization in Haematology.46,47 The principle of this assay is based on the fact that all derivatives of hemoglobin with the exception of sulfhemoglobin are capable of being converted to a stable form of cyanomethemoglobin. Since in most cases, th e sulfhemoglobin content is negligible, the method is effective in quantifyi ng the total hemoglobin concentr ation, typically in percent (%), or grams of hemoglobin per deciliter of solution (g/dL).

PAGE 115

94 The kit included concentrated Dra bkin’s reagent (52 mmol/L potassium phosphate, 30.4 mmol/L potassium ferricyan ide, and 38.4 mmol/L potassium cyanide), Brij-35 solution (detergent, 25%), and methemog lobin standard (18 g/dL). The standard was used to generate a calib ration curve to which the unknown data can be compared for quantification. The working Drabkin’s solu tion was mixed by adding the contents of one 20 ml bottle of concentrated Drabkin’s reagen t to 980 ml of deionized water and 0.5 ml of Brij-35 solution. For the calibration curve, f our tubes were prepared a nd the standard solution was mixed with Drabkin’s solution according to the specifications indicated in the table below (Table 4.4): Standard Drabkin's Blood Tube Solution Solution Hemoglobin NO. (ml) (ml) (g/dl) 1 0.0 6.0 0.0 2 2.0 4.0 6.0 3 4.0 2.0 12.0 4 6.0 0.0 18.0 Table 4.4: Recipes for the mixing proportions of each tube to obtain the concentrations in the Blood Hemoglobin column. The absorbance measurements obtained from each tube will be used to gene rate the standard curve. Each of the tubes was mixed well by invers ion and was allowed to stand at room temperature for approximately 15 minutes to ensure total conversion of the methemoglobin to cyanomethemoglobin. The samples were then scanned in the spectrophotometer, using the Dr abkin’s solution as the auxi liary spectrum for background correction. The optical density values at a wavelength of 540 nm were collected and

PAGE 116

95 these values were plotted as a function of concentration. Th e resulting extrapolated line was expected to pass through the origin. The equation of the line served to evaluate the concentrations of unknown solutions by obtaini ng the optical density value at 540 nm. The unknown samples being assayed c ould either be free hemoglobin or hemoglobin encapsulated in red cells. The latter is feasible because of the presence of the detergent (Brij-35) in the reagent. The am phiphilic detergent molecules will disrupt the membrane, forming detergent-lipid-protein micelles and bursting the cells.107 The cells spill their hemoglobin content in soluti on where it will be first oxidized to methemoglobin, then converted to cyanomet hemoglobin. For the assay, suggested volume ratios for Drabkin’s to sample are 5.00 ml : 0.02 ml, or 2.50 ml : 0.01ml. The former ratio was used because larger quanti ties will render small errors negligible. Once mixed, the sample was allowed to stand in room temperature for approximately 20 minutes and then a spectrum was obtained and background-corrected with the Drabkin’s reagent. The OD value at 540 nm was convert ed to a concentrati on using the equation for the standard curve. It was important to note that the concen tration obtained from a suspension of red cells using this assay did not directly reflect the m ean corpuscular hemoglobin concentration. The acquired value repres ented the hemoglobin concentration in the original sample if the cells were completely lysed and the hemoglobin was diluted in the media. This value was the same as the HGB parameter obtained by the hematology analyzer. Calculation of the MCHC required the relationship MCHC = (HGB x 100)/HCT (L/L volume fraction)

PAGE 117

96 where HGB is g/dl and MCHC will be reported as g/ml. The hematocrit (HCT) was acquired by the hematology analyzer, hence ma y introduce instrumental bias into the calculated MCHC. Calibration of Conductivity Meter fo r Ionic Strength Measurements For the purpose of determining the ioni c strength of prepar ed saline solutions, Conductivity Monitor Model #1670440 (BioRad, CA) was used to prepare a calibration curve using carefully prepared dilutions of stock 0.9% (determined by the company for the particular lot used) isotoni c phosphate buffered saline at ro om temperature. The stock y = 17761x + 550.29 R2 = 0.9959 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 00.10.20.30.40.50.60.70.80.91 % PBSmicromho/cm Figure 4.9: A stardard curve of conductivity vs. PBS concentration (%). The dilutions were carefully prepared from a stock 0.9% PBS and each sample was measured with the conductivity meter. They obtained valu es were converted to micromho/cm by multiplying by a factor of 35.46.

PAGE 118

97 PBS was diluted accordingly at 0.1% intervals from 0.9% to 0.1% and each sample was measured for conductivity. For each measur ement, the value obtained by the meter was multiplied by a factor of 35.46 to get conduc tivity units of micromho/cm. The unit mho represents inverse ohms which is a unit of electrical resistance. Figure 4.9 illustrates the standard curve of conductivity as a function of PBS concentr ation. The equation of the line y = 17761x + 550.29 was used to calculate the concentration of an unknown saline sample, where y is the conductivity value in mi cromho/cm and x is the % PBS. The error limits of the isotonic saline were re ported to be approximately 0.88 + 0.02 and each batch was tested by Nerl on red cells to ensure their viability (personal communication with a Nerl technical director). He nce using the isotonic value of 0.9% for the stock solution kept us within reasonable limits of the calibration curve. Sample Preparation for Resealing Experiments Cells were obtained from the Florida Bl ood Services in two forms. The first was a 5 ml whole blood sample stored in EDTA tubes and kept for screening purposes. The cells were refrigerated at -4o C. Typical samples used for the spectroscopic experiments were no older than two days unless otherwise specified. The age of the cells have been shown to have significant effect on morphol ogy (i.e. crenation) as demonstrated by another member of this project (unpublished data).108 The second type of sample was packed cells suspended in AdsolTM (Fenwal Division, Baxter Healthcare Corp, IL). AdsolTM is used as a stabilizer for blood cell units and can extend the storage period for red blood cells. As a common blood banking pr actice, a large number of transfusable

PAGE 119

98 units were packed and the plasma was replaced by AdsolTM for prolonged viability of the cells for up to 42 days. The age of the packed cells was also no more than two days old unless otherwise noted. The packed cells typi cally came in transfusable volumes of one pint and were used for the preparation of isolated erythrocytes via leukocyte-depletion and washing. Initial exploratory studies were done using the 5 ml whole blood samples. The whole blood was analyzed for cell counts and hemoglobin conten t on the Serono-Baker hematology analyzer. This allowed for the veri fication of the paramete rs to be within the normal range of human physiology. The obtaine d red blood cell count is also helpful in obtaining the proper dilution of the whole blood sample for spect ral purposes. That is, it has been previously determined that diluting the erythrocyte count to ~4000 (cells/micron) will give an optical density of approximately one.109 As noted previously, the upper linear limit of the inst rument is an optical density value of 1.2. (see Appendix B) The dilution calculation is performed as follows, using a red cell count from data obtained on March 1, 1999: RBC count obtained from Ser ono-Baker: 5,130,000 cells/mm3 where cubic millimeters is equivalent to the volume unit of microliters (l). Due to the large number of cells, it is not accura te to dilute down to ~4000 cells/l in one dilution. The routine protocol calls for a series of two dilutions, with the first being a 1:50 dilution, or 0.050 ml of whole blood suspension (mixed well by inversion) in 2.45 ml of isotonic

PAGE 120

99 0.9% phosphate buffered saline. Further cal culation is done for the second dilution, which is contingent upon the concen tration from the first dilution: First Dilution: M1V1=M2V2 (5.13 x 106 cells/l)(0.050 ml) = M2(2.50 ml) M2 = 1.03 x 105 Second Dilution: (1.03 x 105)V1 = (4000 cells/l)(3.00 ml) V1 = 0.117 ml of the 1:50 whole blood dilu tion (total dilution factor = 1:1250) Hence, 0.117 ml of the 1:50 dilution was pipe tted into 2.883 ml of isotonic PBS for a final volume of three milliliters. The 3 ml volume was the capacity used in 3.5 ml cuvettes so that the suspension may be mixed directly into the vessel. The whole blood suspension was then transf erred into a 15 ml conical tube with a cap for washing. The tube was filled to the top with isotonic PBS, capped, and carefully inverted several times to homogenize the su spension of cells. The tubes were placed evenly in a table top centrifuge (Allegra 6, Beckman Coulter, Inc.) with a swinging bucket rotor (GH-3.8A) and spun for approxi mately five minutes at 3500 RPMs (1500 x g). The spin packed the cells at the bottom of the tube and the supernatant left above the packed cells was carefully extracted with a tran sfer pipet. The cells were resuspended in PBS and washed for a total of three times at which point the cells were relatively free of platelets and plasma protein, and the leukocytes were drama tically reduced in number.

PAGE 121

100 After the third wash, the pack ed cells were suspended to a rough 1:1 volume ratio with PBS and examined in the Serono-Baker for platel et and leukocyte reduction, and to verify that the red cell volume and corpuscular hem oglobin have been maintained (i.e. that the erythrocytes had not been generally disr upted). The RBC counts are also used to determine the proper dilution (p reviously described) for a sp ectrograph. A spectrum of the washed cells will provide a qualitative assessment of the plasma reduction as well as any unforeseen changes that may have ensu ed following the preparations, such as morphological alterations or signif icant lysing of th e red cells. The AdsolTM units were obtained as whole donor units held in 450 ml (~ 1 pint) collection pouches. The AdsolTM contained 2000 mg/dl of dextrose, 750 mg/dl of mannitol, 27 of mg/dl adenine, a nd 900 mg/dl of sodium chloride.110 A spectral evaluation of AdsolTM was performed and is presented in the results section. The isolation of red cells from these donor units was more elegant. The bag of red cells was subjected to a leukocyte reducti on by physically linking the bag to a leuko-reducing filter (Sepacell Pre-Storage R-500 II Leukocyte Reduction Set for Red Cells, Baxter Healthcare Corp., Fenwal Division, IL). The leukocyte reduction efficiency was reported to give residual WBC counts of < 1.1 x 105 cells per unit with a 91% RBC recovery.111 The common clinical use for thes e filters is to reduce leukocytes for transfusion purposes. The advantages of leukocyte reduction are 1) to prevent nonhemolytic febrile reactions as a result of antibodies to leukocytes in a reci pient previously exposed to transfusions or pregnancy and 2) to minimize transmission of viral disease such as cytomegalovirus.14,112

PAGE 122

101 The filtration was performed in a cold room (4o C), using gravity to pass the sample through the filter and into a collecti on bag. The filtered blood cells were then placed in an automated cell washer where th ey were packed by centrifugation and the AdsolTM was replaced with isotonic saline. Th e cells were used w ithin 24 hours after preparation. Samples from each step were kept so that it was possible to spectrophotometrically m onitor the changes. Spectrophotometry of Leuko-Reduced Red Cells Counts were performed on the Serono-Bak er hematology analyzer to obtain RBC counts, hemoglobin concentration, and also to monitor the effec tiveness of the leukoreduction. The RBC cell counts were used to make the necessary dilutions to acquire the proper cell concentrations for spectrophotom etry as previously described. The spectra were collected from each step of the red cell isolation process to show the sensitivity of UV-visible sp ectroscopy to detect the change s in the system. After the spectra were taken, each sample was retr ieved from the cuvette and spun down for approximately 8 minutes at 1500 x g. This pelleted the suspended cells in the sample. The resulting supernatants we re scanned in the spectropho tometer to detect any free hemoglobin in solution which may have resulte d from cell lysis. The Drabkin’s assay was performed on the post-leuko-reduced, postwashed (PLR/PW) sample to verify the values obtained by the Serono-Baker for further quantificatio ns and calculations.

PAGE 123

102Calculations for Resealing Experiments The following is an example of the calcu lations performed in preparation for a typical resealing experiment. Numbers di ffered from experiment to experiment depending on the desired corpuscular hemogl obin concentration contained within the resealed cells. Two types of buffer were needed for th e permeabilization experiments (low salt and high salt phosphate buffers). The low salt (hypotonic) phosphate buffer (7mM, pH 7.2) was prepared by combining 1.44 g Na2HPO4 and 0.24 g KH2PO4. Isotonic (0.9%) phosphate buffered saline (PBS) (pH 7.0-7.2) (Nerl Diagnostics, RI) was obtained from Florida Blood Services. When a higher ionic strength PBS was necessary (eg. 2.0% PBS), NaCl was added to the isotonic PB S and the tonicity was determined by conductivity. The cells to be permeabilized were pack ed by centrifugation and mixed in a 1:1 volume ratio with the hypotonic phosphate buffer. Exact quan tities were 1 ml of packed cells in 1 ml of hypotonic buffer. The rese aling process required the determination of concentrated PBS to bring the permeabili zed system back to isotonicity. The concentration of this restoration buffer was calculated by the following process: The cells were assumed to contain the salt concentration equivalent of 0.9%, or 0.009 g/ml. Hence, 1 ml packed cells x 0.009 g/ml = 0.009 g equivalent For the 7 mM PB, approximately 1.7 g of salt wa s dissolved into one liter of water, hence the w/v concentration was 0.0017 g/ml.

PAGE 124

103 1 ml PB x 0.0017 g/ml = 0.0017 g salt So 1 ml of PB contributed 0. 0017 g salt to the mixture. Summing up the two values gave a total of 0.0107 g salts in the 2 ml mixture. Since the reaction tube held up to approximately 6 ml, it was arbitrarily determined that the total reaction volume would be 4 ml (although any other volume up to 6 ml would have been viable alternatives). Thus the final volume of 4 ml needed to be restored to 0.9%, or 0.009 g/ml x 4 ml = 0.036 g sa lt needed in final mixture. The difference between the final weight n eeded (0.036 g) and the current weight contained in the original 2 ml mixture (0.0107 g) was 0.0253 g. That is, 2 ml of solution containing 0.0253 g of salt was needed to rest ore the mixture back to isotonicity. (0.0253g/2ml)*100 = 1.3% PBS So 2 ml of 1.3% PBS was needed to bri ng the reaction mixture to a 4 ml volume of isotonic condition. The 1.3% PBS was prepared by adding NaCl to the isotonic blood bank saline and checking the tonicity with a cal ibrated conductivity meter. T ypically, a stock solution of approximately 2.0% was made and diluted to 20 ml of the desired PBS concentration as demonstrated below: M1V1 = M2V2 (2.0%)V1 = (1.3%)(20 ml) V1 = 13 ml of 2.0% PBS Add 7 ml of deionized water for final volume of 20 ml

PAGE 125

104Red Cell Resealing Experiment The procedures for hypotonic modificati on of red cells have been taken and modified from a number of published methods.99,100,101,102,103,113,114 The goal was to be able to vary the hemoglobin concentrati on in the resealed cells and successfully reproduce the data in high, medium and low con centration ranges for further analysis. As previously mentioned, the cells to be modifi ed were prepared either by washing a 5 ml sample of whole blood by centrifugation, or a large adsol-red cell unit was leuko-reduced and subjected to an automated wash. The Drabkin’s assay was administered to the prepared red cell suspension for hemoglobin quantification, and counts were obtained from the Serono-Baker hematology analyzer. The cells were permeabilized under different conditions depending on the desired outcome. For low concentrations of final MCHC, relatively large amounts of hypotonic buffer (7 mM phosphate buffer) were used in relation to the packed cell volume. For example, the ratio of packed cell volume to hypotonic buffer volume might be 1:5. The suspension was mixed carefully by inversion and incubated on ice for approximately 30 minutes. At the conclusion of the incubati on, the appropriate am ount of hypertonic phosphate buffered saline would be added to re store the tonicity of the suspension to 0.9%. The mixture was then incubated for 45 minutes at 37o C to ensure proper resealing for a majority of the cells. The result wa s resealed red cells containing low amounts of hemoglobin. For medium to large concentrations of hemoglobin in the resealed cells, the proportion of hypotonic buffer was va ried. In one representati ve experiment, the packed

PAGE 126

105 cell to hypotonic buffer volume ratio was set to 1: 1. The incubation conditions were also modified to 1 minute at room temperature. The suspension was immediately restored to isotonicity and incubated at 37o C for 45 minutes. The resealed cells contained higher amounts of hemoglobin. Once the cell restoration was comp lete, the suspension was spun down by centrifugation to sediment the cells and isol ate the supernatant. The supernatant was evaluated for hemoglobin concentration using the Drabkin’s assay so that all of the hemoglobin in the sample suspension can be accounted for. The rest of the supernatant was carefully extracted using a transfer pipe t and isotonic PBS was added for washing. The packed cells were carefully resuspended via multiple vessel inversions and packed again by centrifugation. The washes were perf ormed three times, or until the supernatant exhibited a minimal or no noticeable red tint of hemoglobin. After the final wash, the packed cells were resuspended with approximately an equal vo lume of PBS. The mixture was analyzed in the Serono-Baker and with the Drabkin’s assay. Using the obtained counts, the spectral dilution of approximately 4000 cells/l was determined and spectra were taken of the samples. Each sample being prepared in the cuvette was carefully inverted three times for consistency and the cuvette was placed in the holder in the same orientation for each scan. In between sample s, the cuvette was carefully rinsed several times with deionized water via squirt bottle, and then with saline (since saline is the sample medium). The salin e spectrum for the background co rrection was renewed a few times during each experiment to ensure that a ny deviations in the light source or slight contaminations on the cuvette walls were accounted for. Such errors were minor,

PAGE 127

106 however, and should not affect the spectra si gnificantly, especial ly at large optical densities. The samples prepared for spectra were transferred to centrifuge tubes and sedimented for the collection of the supernatan ts. The spectra of the supernatants were taken to detect any free hemoglobin which may have been present to affect the overall spectra of the samples. 4.3 Results and Discussion Effects of Instrumental Setup on the Spectra of Erythroc ytes and Hemoglobin As an important step to achieving a mean ingful interpretation of the erythrocyte suspension, it was necessary to examine the e ffects of instrumental arrangements that captured different amounts of scattered light Three instrumental constructs were examined: 1) the Agilent 8453 diode array spectrophotometer captured the transmitted light at a small (2o) acceptance angle (small angle tran smission spectra), 2) the PerkinElmer Lambda 900 spectrophotometer was arrang ed to acquire data at larger (> 2o) angles (referred to as diffuse transmission sp ectra), and 3) the Perkin-Elmer Lambda 18 spectrophotometer was fitted with an integrating sphere module to collect light scattered at all angles by the sample. Figure 4.10 s hows a solid curve representing a spectrum of purified, diluted red cells (4000 cells/l) containing a physiological hemoglobin concentration (33% w/v) taken with a sm all angle Agilent spectrophotometer. The dashed curve shows the spectrum of a fr ee hemoglobin solution obtained by lysing the same concentration of red cells in a hypot onic buffer (7mM phosphate buffer), thus the

PAGE 128

107 total mass of the hemoglobin was constant between the two samples. All the spectra presented in this section were obtained in replicates of three with good re producibility. The spectrum of the encapsulated hemoglobi n exhibits a significantly greater OD throughout the entire wavelength range compar ed to the spectrum of the hemoglobin solution, however it lacks the prominent abso rption peaks that defi ne free hemoglobin (characteristic hemoglobin peaks at 270, 337, 417, 547, and 575 nm). 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1902903904905906907908909901090 Wavelength (nm)OD Purified RBC Suspension Lysed RBCs (Oxyhemoglobin Solution) Figure 4.10: Representative comparison spectra of purified red cells and hemoglobin in solution acquired from an Agilent 8453 spectro photometer with an acceptance angle of 2o. The RBCs were purified by washing w hole blood by centrifugation and passing the suspension through a leuko-reduction filter. The RBC concentration is approximately 4000 cells/l. The spectrum of the hemoglobin so lution represents a concentration of ~0.12 mg/ml.

PAGE 129

108 Figure 4.11 illustrates spectra of the same samples prepared in the same manner as Figure 4.10 but captured as a diffuse tran smission spectrum (DTS) on a Perkin-Elmer Lambda 900 spectrophotometer with a larger ac ceptance angle. Two features are immediately apparent compared to the spectra in the previous figure. First is the presence of hemoglobin peaks in the DTS, whereas these peaks in the small angle transmission spectrum (SATS) were masked. Second, the DTS of the suspension showed an overall lower spectral intensity across the en tire wavelength range compared to the red cell spectrum taken with the smaller, 2o acceptance angle (Figur e 4.10), although it was 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1902903904905906907908909901090 Wavelength (nm)OD Purified RBC Suspension Lysed RBC (Oxyhemoglobin Solution) Figure 4.11: Representative comparison spectra of purified red cells and hemoglobin in solution acquired from a Perkin-Elmer La mbda 900 spectrophotometer with an acceptance angle >2o. The RBC concentration is approximately 4000 cells/l and the concentration for the hemoglobin in solution is ~0.12 mg/ml.

PAGE 130

109 still elevated compared to the free hemogl obin spectrum. The lowered OD for the DTS was attributed to the loss of scattering info rmation as the wider angle collected scattered light from more particles. Figure 4.12 represents spectra acquired by an integrating sphere. In this arrangement, all of the scattered light wa s collected, hence the resulting transmission spectrum reflected solely the absorption component (hence it can be referred to as an 0 0.2 0.4 0.6 0.8 1 1.2 1.4 190290390490590690790890 RBC Suspension Lysed RBCs (Oxyhemoglobin Solution) Figure 4.12: Representative comparison spectra of purified red cells and hemoglobin in solution acquired from a Perkin-Elmer Lambda 18 spectrophotometer fitted with an integrating sphere. The RBC concen tration is approximately 4000 cells/l and the concentration for the hemoglobin in solution is ~0.12 mg/ml.

PAGE 131

110 absorption spectrum). The baseline of the ab sorption spectrum per particle of the red cells overlapped that of the spectrum of the he moglobin solution, contrary to the elevated nature of the diffuse transmission spectrum in Figure 4.11. The hemoglobin peaks of the red cell absorption spectrum we re more pronounced in compar ison to the peaks of the DTS, however, the peaks at 270, 337 and 417 nm were less intense than the peaks of the hemoglobin solution. The doublet at 547 and 575 nm show ed good overlap between the whole and lysed red cell spectra. Examination of the red cell spectra in Figures 4.10 and 4. 12 revealed clear differences in the presence and absence of the scattering component, respectively. The integrating sphere data demonstrated simila rities to the hemoglobin solution absorption albeit the sphere showed decreases in sele ct peaks at the lower wavelengths. With a small acceptance angle, the scattering component masked the hemoglobin absorption peaks and caused an overall elevation in the op tical density spectrum. It can be inferred that the small amount of light scattered at wide angles not capt ured by the detector provides the differences in f eatures and information content of the three instrumental configurations presented here. Moreover, for a DTS, the instrument collects more scattered light than a sma ll angle detector, lowering th e overall optical density and elucidating the characteristic hemoglobin peaks. By capturi ng more scattered light, the DTS reflects less scattering information in th e OD data. To summarize, a configuration that collects a large amount of scattered light loses scattering information in the transmission spectrum because the intensities of the scattered light are averaged over the

PAGE 132

111 angles accepted by the detector. In contrast, small acceptance angle transmission spectra contain a better balance of absorp tion and scattering information. It is interesting to note that the sp ectra of the hemogl obin solution does not change across the three varying instrumental co nstructs, verifying that the differences in the red cell spectra are strictly related to sca ttering from large particles. Using the free hemoglobin spectrum as a reference, it is clear that the intensity of the SATS at 417 nm is higher than the hemoglobin Soret band wher eas the DTS and integrating sphere data show smaller peaks. The reduction of select hemoglobi n peaks (270, 337 and 417 nm) particularly with the DTS, had been previ ously attributed to a molecular hypochromic effect. 6,84,115 Our studies have clearly shown, how ever, that this apparent hypochromism is a matter of instrumental perspective. A smaller acceptance angle captured forward scattered light from the macroparticles w ithout diluting the in formation content by capturing too much scattered li ght from wider angles. It wa s also obvious, that with a smaller acceptance angle, the spectrum of the encapsulated system showed no “hypochromism” compared to the free hemogl obin solution since this type of spectrum contains both absorbance (Qabs) and scattering (Qsca) information. When the scattering component is eliminated from the spectru m (Figure 4.12), the attenuation of the absorption component (compared to the ab sorption spectrum of free hemoglobin) describes macroscopic hypochromism. Recall that Qabs and Qsca are additive constituents of Qext.8 Therefore, as the rela tive contribution of the scattering component increases, a concomitant decrease in the abso rption component will be observed.

PAGE 133

112 Comparing the three instruments, the small angle Agilent spectrophotometer yielded spectra that containe d the most complete informa tion about the scattering and absorption components of the red cell suspensi on. For this reason, Agilent instrument was used exclusively for the qualitative and quantitative characterization of red cells in their native and modified forms. Spectral Evaluation of AdsolTM Since a significant number of blood units used in the experiments were stored in AdsolTM for prolonged transfusion viability, it was necessary to evaluate the spectral implications of the effect of AdsolTM. Two spectra were taken: AdsolTM in its undiluted state, and “Adsol” diluted in the same manner as whole blood. The latter was prepared by first diluting the AdsolTM to 1:50 with isotonic PBS (0.050 ml AdsolTM, 2.45 ml PBS), and then following with a second dilution of proportions which are similar to that of whole blood being diluted to 4000 cells/l (0.200 ml 1:50 Adsol, 2.800 ml PBS). Figure 4.13 illustrates the outcome. Concentrated AdsolTM clearly showed prominent absorption at the wavelengths below 300 nm, a possible problem if the whole blood was spectrally processed without dilution. The peak can most likely be attributed to adenine. However, since the blood cells need to be diluted to give reliable data, such a dilution was simulated with the AdsolTM alone. The result was a spectrum that was essentially devoid of any features. The small peak close to 200 nm is a typical artifact of subtracting a large background peak from an even larger sample peak. Since both of these peaks are well beyond the linear range of the instrument, they are usually saturated (as seen in the

PAGE 134

113 concentrated AdsolTM peak) and contain little information. Hence the difference between such peaks produced a peak containing unrel iable information. The water peak at approximately 980 nm in the concentrated AdsolTM spectrum is present because no background correction was employed. A correction with -0.01 0.49 0.99 1.49 1.99 2.49 2.99 3.49 3.99 1902903904905906907908909901090 Wavelength (nm)OD Concentrated Adsol Diluted Adsol Figure 4.13: Optical density spectra of Adsol. The solid-line spectrum represents undiluted Adsol directly from the bag. The dashed-line spectrum represents Adsol when diluted in the same manner as a red cell suspension would for spectral purposes. water would produce essentially the same sp ectrum without the 980 peak. The diluted spectrum was background corrected with PB S (since the dilution was prepared with PBS), thus the absence of the 980 peak. The issue with Adsol was whether its ab sorption bands were prominent enough to interfere with the spectrum of red cells pr ior to washing and le uko-reduction. At high

PAGE 135

114 concentrations, it did in fact show high absorption in the lower wavelengths (190-400 nm). However, when it was diluted in the same manner as the red cell suspension, the Adsol was optically transparen t at wavelengths greater th an ~220 nm, eliminating the need to correct for its eff ects in the background (under red cell measurement conditions). Washed and Leuko-Reduced Cells Before the cells were modified, they we re prepared from either whole blood or Adsol-suspended blood. The whole blood samp les were prepared with manual washings by centrifugation to rid the suspension of pl asma, platelets, and some white cells. A combination of spectra and Serono-Baker coun ts show the degree of success of this method. Figure 4.14 is a representative normalized sample spectra showing both whole blood and washed cells. Normalization elim inates any concentration effects of the samples and provides a better visual compar ison of the spectral ch anges caused by other parameters such as the removal of certain blood components, or ch anges in cell size. Whole blood showed an elevated optical de nsity throughout the entire wavelength range encompassing the optical effects of each major component in this complex mixture. The most striking difference between the washed and unwashed samples is the large peak at 202 nm and the small but relatively broad peak at 280 – 290 nm, an effect accounted for by the removal of the plasma components. When spun down and the red cells were sufficiently packed, the remaining supernatant was a clear liquid w ith a slight yellow tint. Sh ifting viewing angles of the supernate against the light revealed a shimme r due to the suspended platelets which were

PAGE 136

115 less dense than the red and white cells. Dire ctly above the bottom layer of red cells was a white layer commonly referred to as the buffy coat that consisted mostly of leukocytes. The clear plasma containing platelets were eas ily extracted and removed using a transfer 0.0E+00 5.0E-04 1.0E-03 1.5E-03 2.0E-03 2.5E-03 1902903904905906907908909901090 Wavelength (nm)Normalized OD whole washedOct 16, 1998 Figure 4.14: Normalized (by area under curve) op tical density spectra of whole blood and washed cells. The cells were washed manually three times by centrifugation to reduce the amount of white cells, pl atelets, and plasma proteins. pipet. However, removal of the layer of white cells proved to be trickier and the effective elimination of the leukocytes meant taking with it some red cells from the bottom layer. After three washes, the packed red cells we re resuspended in PBS at an approximate volume ratio of 1:1.

PAGE 137

116 Table 4.5 shows the values obtained by th e Serono-Baker hematology analyzer to corroborate with the cha nges in the spectra. Consistent with the objective of the washes, there was a significant reduction seen in th e white blood cell (WBC) and platelet (PLT) numbers, albeit the removal was not complete Due to the estimated volume of the resuspension of the packed cells (approximate 1:1 volume ratio), the result of the red cell Whole blood Washed cells WBC(thsn/cu mm) 6.4 1.6 RBC (mill/cu mm) 4.29 3.81 HGB (grams/dl) 13.5 12.3 HCT (%) 43.7 38.3 MCV (cu microns) 101.9 100.6 MCH (pg) 31.5 32.3 MCHC (%) 30.9 32.1 PLT(thsn/cu mm) 276 38 MPV (cu microns) 10.1 10.8 Table 4.5: Values obtained by the Serono-Baker hematology analyzer of whole blood and washed cells. counts (RBC) varied but al ways on the order of 106/mm3 (106/l). In this case, the cell count decreased slightly with a corresponding decrease in the hemato crit value (HCT). Reduced RBC also meant a decrease in th e overall hemoglobin count (HGB) as manifested in the results. The mean corpuscular volume (MCV) values showed, however, that the washings di d not disturb the size of the red cells although it is not possible to determine how the morphology ma y have been affected. The hemoglobin concentration within the red cells (MCH, MC HC) were also undistur bed, as well as the platelet volume.

PAGE 138

117 The Adsol-suspended cells were prep ared by automated washing and leukoreduction via filtration. Figure 4.15 shows the different stages of red blood cell purification with the washing step done first. The wash eliminated the plasma and the most significant spectral difference is the peak at approximately 200-230 nm. With the removal of the leukocytes, however, a change in the overall spectral profile is evident. In corroboration with the spectral data, hemato logy analyzer counts showed values that correspond with each step (Table 4.6). For corroborative purposes, it was important to evaluate every aspect of the 0.0E+00 2.0E-04 4.0E-04 6.0E-04 8.0E-04 1.0E-03 1.2E-03 1.4E-03 20030040050060070080090010001100Wavelength (nm)Normalized OD packed cells in Adsol post wash post wash/LRMay 14,2003 Figure 4.15: Stages in the purification of red cells. The samples were first washed with isotonic phosphate buffered saline and then passed through a leukocyte filter.

PAGE 139

118 Pre-wash/LR Post-wash Post-wash/LR WBC(thsn/cu mm) 10.6 3.3 ***** RBC (mill/cu mm) 5.53 6.54 6.47 HGB (grams/dl) 18.4 21.3 21.0 HCT (%) 55.3 64.0 63.3 MCV (cu microns) 100.0 97.9 97.8 MCH (pg) 33.3 32.6 32.5 MCHC (%) 33.3 33.3 33.2 PLT(thsn/cu mm) 72 0 0 MPV (cu microns) 14.4 ----Table 4.6: Values obtained from the hematology an alyzer for the different stages of red cell purification. 0 0.2 0.4 0.6 0.8 1 1.2 1.4 20030040050060070080090010001100 Wavelength (nm)OD red cell suspension supernatantMa y 14, 2003 Figure 4.16: Comparison of red cell suspension and supernatant (I). The blue spectrum represents a red cell suspension post wash/LR. The pink spectrum of lesser magnitude is the supernatant after the cells we re spun down by centrifugation.

PAGE 140

119 system. Hence it was necessary to de termine whether the hemoglobin was all encapsulated or if some were free in soluti on. In Figure 4.16, th e solid line spectrum represents the raw data of the suspended re d cells after the washi ng and leuko-reduction. The sample was then poured out of the cuvette and into a centrifuge tube. The sample was centrifuged for approximately 8 minutes at 1500 x g to pellet the cells. The pink dashed-line spectrum that looks more lik e a spectrum of hemoglobin reflects the supernatant obtained post-cen trifugation. The amount of free hemoglobin seen in the supernatant depends on the quality of the washes. Typically, three washes were sufficient for elimination of a majority of the free hemoglobin. Visual inspection of the 0 0.2 0.4 0.6 0.8 1 1.2 1.4 20030040050060070080090010001100 Wavelength (nm)OD red cell suspension supernatantJune 3, 2003 Figure 4.17: Comparison of red cell suspension and supernatant (II). The blue spectrum represents a red cell suspension pos t wash/LR. The pink spectrum of lesser magnitude is the supernatan t after the cells were s pun down by centrifugation.

PAGE 141

120 supernatant should yield a colorless medium Washing excessively, however, was not necessarily advantageous since repeated sh ear between cells may cause some to lyse, increasing the free hemoglobin in solution. Figure 4.17 show s a different trial with a more thorough wash. The supernatant sp ectrum reflected a minute amount of hemoglobin compared to that seen in Figure 4.16. Typical washes more often yielded results similar to that of Figure 4.16. Estimated quan tification of free hemoglobin Figures 4.16 and 4.17 using the exti nction coefficient at 417 nm (7786 cm2/g) are 0.032 mg/ml and 0.0013 mg/ml respectively. Howeve r, the presence of any amount of free hemoglobin should not be a c oncern as long as a reliable spectrum is obtained for quantification. Spectrum of Resealed Red Cells The desired outcome of the resealing experi ments was to obtain, at the very least, mean corpuscular hemoglobin concentrations in the medium (~ 0.15 – 0.20 mass frac) and low range (~ 0.05 – 0.10 mass frac), si nce physiological concentrations were established in the previous sect ion. Figure 4.18 represents a resealed cell suspension with an MCHC of 0.084 (mass fraction), or 8.4% w/v. A striking cha nge in the feature compared to the spectrum of physiological MC HC (Figures 4.16, 4.17) is the decrease in overall scattering, revealing more of the prominent hemoglobin peak at 417 nm along with the slight emergence of the doublet between 500 and 600 nm. The supernatant spectrum shows a relatively sma ll amount of hemoglobin.

PAGE 142

121 -0.01 0.19 0.39 0.59 0.79 0.99 20030040050060070080090010001100 Wavelength (nm)OD MCHC: 0.084 supernateMay1, 2003 Figure 4.18: Optical density spectra of a res ealed red cell suspension (MCHC: 0.084) and the supernatant. There is less scatte ring and a more pronounced hemoglobin peak in the cell spectrum compared to that of the high MCHC (Figure 4.12). There is a small amount of free hemoglobin in th e supernatant as indicated by the supernate spectrum. Figure 4.19 illustrates resealed red ce lls with MCHCs in the medium range (0.218, 0.186 mass fractions). There is a more prominent scattering seen throughout these spectra compared to that of the lower MCHC (Figure 4.18), resulting in an increased masking of the hemoglobin peaks. However, the overall scattering is less than the physiological spectru m (Figure 4.17) and the hemoglobi n absorption features are not completely hidden. Between the mass fractio ns of 0.218 and 0.186, clear differences are seen to fit the scattering trend. The supern atant spectra again show little free hemoglobin contamination.

PAGE 143

122 0 0.2 0.4 0.6 0.8 1 1.2 20030040050060070080090010001100 Wavelength (nm)OD MCHC: 0.218 (A) MCHC: 0.186 (B) A supernatant B supernatantSept 4, 2003 Figure 4.19: Optical density spectra of a reseal ed red cell suspension (MCHC: 0.218, 0.186) and the supernatant. The degree of masking of the 417 nm hemoglobin peak by scattering is somewhere between those of th e high and low MCHC ranges (Figures 4.12, 4.13). The supernatants show small amounts of hemoglobin. Figure 4.20 shows a more complete compila tion of the spectra of resealed red cell data spanning a broad range of mean cor puscular hemoglobin concentrations. Every sample spectrum represents a cell suspen sion diluted to approximately 4000 cells/l. Hence, the intensities of the optical densities roughly correspond to a constant cell count. As the encapsulated hemoglobin concentration decreases, the scattering effect decreases, effectively reducing the amplitude of the ove rall spectrum. As previously emphasized, the reduction of scatte ring unmasks more of the features of hemoglobin. The less the encapsulated hemoglobin, the more the spectru m looks like free hemoglobin in solution.

PAGE 144

123 These changes in the features are magnified when the spectra are normalized by dividing through by the calculated area under the curve (F igure 4.21). This type or normalization eliminates concentration effects and amplifie s effects caused by only the properties of the particle.75 Here, it is abundantly evident that with decreasing MCHC, the 417:383 nm peak-to-trough ratio increases. Though this is not necessarily a linear effect, the trend is intuitive and is readily recognized. 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 20030040050060070080090010001100 Wavelength (nm)OD 0.336 0.228 0.150 0.084 0.048MCHC (mass fraction) Figure 4.20: A compilation of raw spectral data of resealed cells with different encapsulated hemoglobin concentrations.

PAGE 145

124 0.0E+00 5.0E-04 1.0E-03 1.5E-03 2.0E-03 2.5E-03 3.0E-03 3.5E-03 4.0E-03 20030040050060070080090010001100 Wavelength (nm)Normalized OD 0.336 0.228 0.150 0.084 0.048MCHC (mass fraction) Figure 4.21: Compilation spectra normalized using the area under the curve method. This normalization eliminates concentration effects to amplify the effects due to the optical properties of the particle. Reproducibility of the Resealed Cells The hypotonic modification of the red cel ls was shown to be reproducible by performing the experiment in five parallel reaction vessels under the same conditions, using red cells taken from the same stock sa mple. Figure 4.22 show the spectra for the five samples, normalized to eliminate any concentration differences as a result of pipetting errors. Table 4.7 lists the values obtained from the hema tology analyzer for the HGB, MCHC, and MCV. The spectra of the five replicates show st rong similarities in the features across the entire wavelength range. The reported values in the table further

PAGE 146

125 support the reproducibility of th e experimental method. The small differences among the replicates can be attributed to experimental error (such as pipetting errors) and the error of the hematology analyzer. A comparison of experiments done under the same conditions but with different samples ( not pictured) did not always yield high reproducibility because in this case, new va riables are being intr oduced by varying the blood sample (i.e., patient genetics, how often patient donates blood, etc). 0.0E+00 20E-04 40E-04 60E-04 80E-04 10E-03 12E-03 1.4E-03 16E-03 18E-03 1902903904905906907908909901090 Wavelength (nm)Normalized OD Replicate 1 Replicate 2 Replicate 3 Replicate 4 Replicate 5Jun303 Figure 4.22: Normalized resealed cell spectra of five replicates. The modified cells are in the range of 23% MCHC. The spectra generally reflect good re producibility with the spectra sharing common features ac ross the entire wavelength range.

PAGE 147

126 HGB (g/dl) MCHC (mass frac) MCV (fl) Replicate 1 7.4 23.8 86.2 Replicate 2 6.3 22.8 89.1 Replicate 3 7.2 23.9 86.8 Replicate 4 6.8 22.0 88.0 Replicate 5 6.1 22.8 87.8 Table 4.7: Values obtained for five experimental replicates of resealed cells for the parameters HGB, MCHC, and MCV. The va lues show good agreement indicating high reproducibility of the hypotoni c modification experiments. Microscopy With each experiment, visual corrobora tion of the red cells was established by observing smears of the samples under the microscope. On a few occasions, we gained access to a camera-mounted phase contrast mi croscope (Nikon Diaphot, Department of Biology, University of South Florida) to document photographs of the cells. Both microscopes used were equipped with a 10x ocul ar lens with objective lenses of 10x, 40x, and 100x (oil immersion) where the total magni fication was the product of the ocular and the objective magnifications. Typically, 400x and 1000x magnifications were used to assess the morphologies of the cells. A quick microscopic inspection of the whole and washed cells was useful in evaluating the quality of the cells. Although th e scope of the field of vision is small, examining a large number of random areas of the blood smear on the microscope slide gave an idea of whether or not there were any obvious morphological abnormalities in a particular sample. A high quality red cell suspension should look like Figure 4.23 where a majority of the erythrocytes assume a biconc ave shape with the dimple clearly visible in the center of the donut shape. The hash ma rks at the bottom of the photograph give a

PAGE 148

127 Figure 4.23: A phase contrast microscopy picture of erythrocytes in their native state magnified 400x. The sample was prepared by smearing a small drop of the cell suspension on a glass microscope slide. The hash marks at the bottom represent 10 m per interval. The cells show their characteris tic biconcave shape with a noticeable dimple in the center. good estimation of the overall sizes of the cells, where one interval represents 10 m. An alternate sample suspended in PBS at room temperature for a few hours is illustrated in Figure 4.24. This photograph shows crena tion in which the cells have adopted a spherical conformation with spiculated protrusi ons. If this phenomenon is seen in whole blood samples, it could be a result of aging, contam inated medium, or improper storage (temperature). It has been demonstrat ed that over time, stored red cells will lose their membrane integrity and begin to crenate. Red cells suspended in isotonic PBS will adopt the speculated morphology sooner than its whole bloo d counterpart (matter of hours for a dilute suspension). The non-crenated cells in the picture show a flatter

PAGE 149

128 Figure 4.24: A phase contrast microscope image of non-viable red cells. Some of the cells show a crenated conformation in wh ich they take on a spherical shape with protruding spicules. Other cells show a less de fined dimple. Each interval at the bottom of the picture represents 10 m. discoid shape with a less appa rent center dimple. In this case, the cells may have acquired excess water to become more ellipt ical, or indicate immi nent crenation. Likewise with the washed and leuko-re duced cells, observing the cells was a quick and easy assessment of quality. Since the washes were performed with isotonic PBS, if the ionic strength buffer was inaccurate it would have affected the cells in ways such as crenation and swelling. Moreover, both washing and leukodepleting filtration posed the possibility of disr upting the native membrane shap e due to shear force. Figure 4.25 shows an oil immersion (1000x magnification) picture of resealed red cells representing an estimated mean cor puscular hemoglobin concentration of 21.4% (compared to a physiological c oncentration of 33%). In the act of permeabilizing and resealing, the cells have lost their biconcave conformation and seemed to have adopted

PAGE 150

129 Figure 4.25: A phase contrast microscope image (1000x, oil immersion) of hypotonically resealed red cells containing 21.4% (w/v) hemo globin. The variation in sizes has increased and the morphology has ch anged significantly to distorted shapes (crenated and elliptical cells). Each interval of hash marks represents 10 m. more distorted shapes (crenated or elliptica l). The approximate diameters of the cells look to be slightly larger than their unm odified counterparts with a slightly broader distribution of sizes. Figure 4.26 is an oil immersion picture of a sample with a reported MCHC of 5.6%. Due to the low hemoglobin con centration, the contrast of the cells with respect to the background medi um is greatly reduced (n/n0 is approaching 1). The sizes seem a bit more uniform at approximately 5 m compared to the cells containing 21.4% (Figure 4.25). This may be due to the fact that with the low MCHC samples, the samples are equilibrated for 30 minutes on ice after being subjected to hypotonic shock. In contrast, the medium range MCHC samples ar e exposed to the shock for approximately a minute before being restored. Thus these latte r samples do not have time to equilibrate

PAGE 151

130 Figure 4.26: A phase contrast microscope image (1000x, oil immersion) of hypotonically resealed red cells containing 5.6% (w/v) hemogl obin. The reduction in the corpuscular hemoglobin concentrations has sign ificantly reduced the contrast between the cell and the background medium. The cell diam eters seem to be typically in the 5 m range. The shapes seem to be spherical, howev er, it is impossible to assess accurately the 3-dimensional shape. Each interv al of hash marks represents 10 m. and depending on the age of the cells present, it is possible to get more of a mixed population of resealed cells. The sizes of the resealed cell samples we re difficult to visually estimate. The diameters of the cells seemed to typically be 5 – 10 m in all the samples evaluated (3 samples documented per low, medium and high MCHC ranges). The shapes were mostly perceived to be elliptical or crenated (more spherical). However, since light microscopy did not provide a good estimate of the thre e dimensional morphology, it is not a good method to corroborate the sizes (volumes) of the cells to the values obtained by the hematology analyzer.

PAGE 152

131Comments on the Acquired Experimental Data The first significant assertion made in this dissertation is that perceived hypochromism can often be accounted for by understanding the manner in which the spectrophotometer collects the light. There ar e clear differences in the spectral features of the same sample taken with a small angle detector and an integr ating sphere. Studies that have reported spectra of macroscopic pa rticles with high refractive indices have presented diffuse scattering data and have maintained that any OD decreases of the encapsulated system compared to that of the fr ee chromophore in solu tion are a result of molecular hypochromicity.6,7 We have shown here that th is is not the case, but instead the OD decrease can be attributed to a scatte ring-related effect. In this effect, the increase in the scattering component resu lted in the attenuation of the absorption component and was defined above as macr oscopic hypochromism. Moreover, it was determined that a small angle spectrophotomet er contained the most complete scattering information useful in the qualitative and quant itative analysis of particle suspensions. Thus to summarize, observed hypochromicity has been redefined on two levels: Macroscopic hypochromicity which is due to changes in scattered light and as consequence is proportional to the size a nd refractive index of the scattering elements. Microscopic or molecular hypochromicity which is due to the electronic interactions resulting from close proxim ity of chromophoric groups and therefore a function of the chromophore concentration within the particle.

PAGE 153

132 Preliminary assessment of the red cell purification and resealed red cell data indicates the proficiency of a small angle spectrophotometer to detect changes in the composition of the suspension (purification of erythrocytes) and changes in encapsulated hemoglobin concentrations. Fo r the purification of red cells the removal of the plasma components and white blood cells reflected small changes in the spectrum. The postwashed/post-LR spectrum (Figure 4.15) show ed the largest difference in the spectral features that could be a resu lt of the leukocyte removal, or a change in morphology of the cells after being suspended in PBS (or a combination of the two). Resealed cells in the high, medium, and low MCHC ranges all showed clear differences in the spectral features. High hemoglobin concentration increased the contrast of the cells with respect to the medium by increasing its complex refractive index. The resulting increase in scattered li ght was represented by an elevation of optical density across the entire wavelength range. Light microscopy proved to be a useful tool in evaluating the quality of the cells prior to and after the resealing experiments. Each experiment was visually monitored at select steps to ensure consistency the resu lts. It was also possible to make rough estimates of the sizes of the cells. Howeve r, such approximations were too vague to make accurate quantifications of the cell volumes. The data acquired by red cell purifica tion and hypotonic modification amounted to over 50 samples. The data was assessed in terms of important parameters such as the mean corpuscular volume (MCV) and the m ean corpuscular hemoglobin concentration (MCHC) to identify spectral trends and feat ures. This analysis was necessary in our

PAGE 154

133 efforts to corroborate the expe rimental data with spectra modeled mathematically based on the Mie theory. As previously mentioned, a sample spectrum is represented by a set of parameters, each unique to the pa rticular sample being examined It is important to pay particular attention to the experimental conditions and corroborative outcomes to accurately explain the different features being manifested in the spectra. It should be noted that changes in key parameters of th e particulate systems (i.e. MCHC, MCV, etc.) are not reflected spectrally in a linear fashion and therefore are often not intuitive. This complexity stems from the fact that the spect ral changes arise from a combination of the absorption and scattering effects and it is this synergy that we are striving to understand by using a combination of wet chemis try and mathematical modeling.

PAGE 155

134 Chapter 5: Implementation and Validation of the Mie Theory As described in Chapter 3, the model for the mathematical construction of spectra is based on Mie theory, which represents th e spectra as the combined effect of the absorption and the scattering components of the cell suspension. An effective implementation of modeling has proven to be a valuable tool in providing guidance in experimental direction. Where there are in strumental limitations in obtaining empirical data, theoretical modeling sees no such restrictions. Through modeling, it is possible to quickly generate a series of spectra a llowing for the exploration of trends and manipulations of variables. Using modeled sp ectra as a guide may help raise a red flag when an unexpected feature is seen experimentally. Ultimately, these simulations help the scientist make an educated decision about the direction of their experimentation and also be used for verificati on post data-collection. 5.1 Sensitivity Analysis of Mie Th eory: A test of the model Given enough knowledge of the parameters of a cell suspension, it was possible to calculate a mathematical representation of the spectrum where the qualitative features were similar to those of the experimental spectrum. The current working implementation of the model was used to simulate the combin ed effect of cellular and plasma components of whole blood.69 The flexibility of this model allowed for the exploration of isolated

PAGE 156

135 effects of single blood components, and hence the program was used for the interpretation of red blood cells. Since the model is based on Mie theory, the cells are represented by spheres of equivalent volume. Though this may be initia lly be construed as a restriction, the model was demonstrated to give good estimations of experimental values, and effectively predicted trends with changing parameters. dDDfmQDNext p)())(,( 42 0 )( )()( )(0 n in m Figure 5.1: Diagram of the model used to calculate spectra and its inputs. The model is based on the turbidity () equation and the optical properties of the suspended cells (m()). Model Diameter Cell Volume Fraction # Protein Components Mass Fraction of Protein Components Optical Properties of Protein Components 00 02 04 06 08 10 12 14 16 18 20 200300400 00 60 700 800 00 1000 100Spectrum

PAGE 157

136 Figure 5.1 shows a block diagram of the inputs of the model. Each of the parameters serves an important purpose in describing th e joint property distri bution and it is their cohesion that gives rise to the qualitative a nd quantitative features of the calculated spectrum. The diameter, as mentioned above, is that of a sphere of equivalent volume (often referred to as the equivalent sphere ). The cell volume fraction is a variable representing the hematocrit (HCT) value. Th e model is capable of predicting the effects of multiple chromophores within the cells. Hence the number of protein components must be specified. In the case of red cel ls, hemoglobin is the so le protein component, although there is the capabil ity of including multiple derivatives of the protein (deoxyhemoglobin, methemoglobin, etc). The mass fraction or the concentration of the protein is a particularly im portant parameter due to th e strong absorptivity of hemoglobin. The complex refractive index (m()) of hemoglobin is defined by an optical properties file representing the absorption coefficient () and refractive index (n) as well as the refractive index of the medium, water. Figure 5.2 depicts the components of the optical properties file of oxygenated he moglobin. The shape of the absorption component of hemoglobin resembles that of a typical hemoglobin absorption spectrum and was derived from experimental optical de nsity spectra of hemoglobin in solution. The real part of the complex refractive index of hemoglobin (n())69 was then generated from () using the Kramers-Kronig transform men tioned previously in Chaper 3. The refractive index of the medium (wate r) was obtained from values reported by Thormaehlen et al. (1985).116

PAGE 158

137 0.0 0.1 0.2 0.3 0.4 0.5 0.6 1902903904905906907908909901090 Wavelength (nm)RI k (Hb) n (Hb) n (medium) Figure 5.2: Plot of the contents of an optical properties file for oxyhemoglobin over the wavelength range of 190-1100 nm. Plots includ e the absorption coefficient and refractive index of hemoglobin and the re fractive index of water. Although the model is stric tly used in this case for the modeling of red blood cells, its flexibility of usage extends beyond th e scope of this study. The Mie theory was adapted to build a complete spectrum of whole blood or spectra of their individual components, given that reliable optical properties existed for each component. Aside from the erythrocytes, the major components of interest included leukocytes, platelets, and plasma proteins. Spectral features of pl atelets were successfully calculated to look like those obtained experimentally.9 Conversely, informati on such as particle size distribution (PSD) and pl atelet activation states were extr acted from the features of the

PAGE 159

138 experimental spectra.9,117 Narayanan (1999) was able to accurately model the spectrum of albumin, the major plasma protein.118 Simulation of Erythrocytes The initial stage in the modeling required the progra m (Bldgen08) to demonstrate its ability to calculate a spectrum of a red cell suspension at physiological parameters. The inputs of the simulation were as follows: Wavelength range for analysis: 190 1100 Spectral resolution in nm: 1 Blood concentration (g/ml) in cuvette: 1e-3 Cell pathlength in cm: 1 Erythrocyte volume fraction: 0.45 Erythrocyte average diameter (cm): 5.56e-4 Erythrocyte density (g/ml): 1 Number protein components: 1 Mass fraction for component #1: 0.33 Property filename for component #1: ophbo2.01 Leukocyte volume fraction: 0 Platelet volume fraction: 0 Plasma volume fraction: 0 This analysis was done for the entire spectral wavelength range of 190-1100 nm at 1 nm intervals (911 total points). The blood concentration in the c uvette (1e-3 g/ml) is an arbitrary number which is typically held constant. Since this is a weight-based concentration, one must be c onscious of the fact that hold ing this number constant and varying the size of the cell woul d change the number-based concentration. Physiological values were maintained for the erythroc yte volume fraction (hematocrit), the mass fraction of hemoglobin, and the diameter of the equivalent sphere (calculated from a volume of 90 fl). Figure 5.3 represents the generated spectrum. The overall shape of the

PAGE 160

139 spectrum closely resembles that of the experime ntal data of red cells (see Figures 4.8, 4.10). The 190-600 nm range incl udes all of the absorption p eaks; hence the spectrum is rich with features resulting from the abso rptive effects of hemoglobin. The 600-1100 nm range is largely considered the “scattering ra nge” due the lack of any strong absorbance peaks. However, it would be a misconcepti on to assume that features in the 190-600 nm range are devoid of any scattering effects. In fact the scattering is typically more prominent from 190 to 600 nm than in the “sca ttering range”. The model typically does a good job fitting the scattering portion above 600 nm. However, as both the absorption and scattering effects both become impor tant below 600 nm, the model may not 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 1902903904905906907908909901090 Wavelength (nm)OD Figure 5.3: Simulated spectrum of red blood cell s under physiological conditions. The simulation was calculated by the m odel based on the Mie theory.

PAGE 161

140 necessarily provide a perfect re production of the features of the experimental spectrum. This may be due to the spherical approximati on of Mie theory. The spherical assumption eliminates potentially important particle susp ension characteristics such as particle shape and orientation which may be necessary for improved fits. Effect of Size on Erythrocyte Spectra One of the important parameters of the sp ectral characterization or erythrocytes is size. Physiologically, the reported volume limits of re d cells range between 80 – 100 fl (see Chapter 2). The simulations serve as a sensitivity analysis for the qualitative features of the red cell spectra in relation to size. Since the model takes the size input as the diameter of a sphere of equivalent vol ume, the volume values were converted using the equation of a sphere and are reported in Table 5.1. Volume (fl) Diameter (cm) 80 5.35e-4 85 5.46e-4 90 5.56e-4 95 5.66e-4 100 5.76e-4 Table 5.1: Physiological volume range of red ce lls in 5 fl increments and their equivalent sphere diameters. The simulations held the following parameters constant: Wavelength range for analysis: 190 1100 Spectral resolution in nm: 1 Blood concentration (g/ml) in cuvette: 1e-3 Cell pathlength in cm: 1 Erythrocyte volume fraction: 0.45

PAGE 162

141 Erythrocyte density (g/ml): 1 Number protein components: 1 Property filename for component #1: ophbo2.01 Leukocyte volume fraction: 0 Platelet volume fraction: 0 Plasma volume fraction: 0 Figure 5.4: Illustration of cell suspension where th e size is increased by the hematocrit and weight-based cell concentration remains constant. Under thes e conditions, the cell number changes, hence the need for furthe r corrections for a comparison on a per cell basis. The five sizes from Table 5.1 were repeat ed with hemoglobin mass fractions of 0.33, 0.20, and 0.05 (high, medium and low MCHC valu es respectively). Figure 5.4 illustrates physical implications of the system in term s of varying cell size. The simulation was initially done by keeping the hematocrit and th e weight-based cell concentration (g/ml) constant while varying only the mean corpus cular volume. The problem however, is that this changed the number of cells in the system Thus the concentration effects must be eliminated to isolate the comparison to cha nges in volume. One solution was to adjust the hematocrit and weight-based cell concentra tion values accordingly so that the cell

PAGE 163

142 number was kept constant. The second solution was to correct the discrepancy by normalizing each of the spectra by the area und er the curve. The latter method was employed. Figure 5.5 shows the effect of size on cells with a low mean corpuscular hemoglobin concentration of 0.05. Slight differences in intensities are seen in the two prominent peaks at 225 nm and 417 nm. Th e peak at 417 nm and the doublet between 500 – 600 nm represent the charact eristic peaks of hemoglobin. Observing the peaks, one can notice that the smaller sizes have hi gher optical densities This trend is counterintuitive when consid ering the absorption compone nt alone. The larger the particle, the more hemoglobin in the pathlength of the light hence it would seem that the larger cell should adopt the higher absorption pe ak. But in reality, the opposite is true and this is consistent with the theory a nd the observations of Garcia-Rubio (1987) on the effect of particle size on the absorption spectra.92 Across the entire wavelength range (at low MCHC), there is a minimal change of features as size increases. This is a dire ct result of the low he moglobin concentration within the cells. A low MCHC implies a low contrast between the refractive indices of the cell and the medium and under such conditi ons, the scattering component is relatively small compared to cells containing he moglobin at a physiological level. Figure 5.6 illustrates the effect of size on cells at a medium hemoglobin concentration (mass fraction of 0.20). Th e hemoglobin bands are present but to a lesser degree, as scattering across the entire wa velength range is more prominent. The difference in the spectra due to the increase in cell size is more noticeable as well. As

PAGE 164

143 seen with the previous simu lation, the intensity in the absorption region (190 ~ 500 nm) decreases with an increase in size. The la rger refractive index ratio of cell/medium (n/n0) accounts for a more significant contribution of the scattering component which explains the more pronounced spectral differences be tween the cell sizes. Furthermore, the medium MCHC is beginning to show signs of the flattening of the spectra, a phenomenon previously seen with size increas es due to swelling or aggregation.118 This flattening is typically marked by the reduction of the optical density in the absorption region with the increase in size and also th e decrease in the slope of the scattering region (600 – 1100 nm). The spectra shows an isosbestic point at approximately 480 nm and the trend of intensity vs. size inverts as this point is crossed coming from either direction. 0.0E+00 5.0E-04 1.0E-03 1.5E-03 2.0E-03 2.5E-03 3.0E-03 3.5E-03 4.0E-03 4.5E-03 1902903904905906907908909901090 Wavelength (nm)Normalized OD Units 80 fl 85 fl 90 fl 95 fl 100 fl MCVFigure 5.5: Simulated spectra of varying MCV at constant MCHC (0.05 mass fraction).

PAGE 165

144 0.0E+00 2.0E-04 4.0E-04 6.0E-04 8.0E-04 1.0E-03 1.2E-03 1.4E-03 1.6E-03 1.8E-03 2.0E-03 1902903904905906907908909901090 Wavelength (nm)Normalized OD Units 80 fl 85 fl 90 fl 95 fl 100 fl MCV Figure 5.6: Simulated spectra of varying MCV at constant MCHC (0.20 mass fraction). Figure 5.7 is a simulation in which bo th the size range (80 – 100 fl) and the MCHC (0.33 mass fraction held constant) are within relevant physiological parameters. The first striking feature to noti ce is that the overall shape of the spectra looks similar to experimental spectra of red cells. The n/n0 has increased to the point to where the whole particle scattering dominates the spectra. This also accounts for the fact that the effect of the size change is dominant compared to the sp ectra in the previous two simulations. The spectral flattening is evident on a larger scale as well.

PAGE 166

145 0.0E+00 2.0E-04 4.0E-04 6.0E-04 8.0E-04 1.0E-03 1.2E-03 1.4E-03 1.6E-03 1902903904905906907908909901090 Wavelength (nm)Normalized OD Units 80 fl 85 fl 90 fl 95 fl 100 fl MCV Figure 5.7: Simulated spectra of erythrocytes at varying MCV at constant MCHC (0.33 mass fraction). Effect of Varying MCHCs on Red Cell Spectra The simulation analysis of erythrocytes naturally progressed to the examination of trends due to varying mean corpuscular he moglobin concentrations while holding the size constant. In this case, the interdependent parameters such as size and hematocrit were not changing therefore an adjustment for pa rticle number was not necessary. Certain parameters were held constant as follows: Wavelength range for analysis: 190 1100 Spectral resolution in nm: 1 Blood concentration (g/ml) in cuvette: 1e-3 Cell pathlength in cm: 1 Erythrocyte volume fraction: 0.45

PAGE 167

146 Erythrocyte average diameter (cm): 5.35e-4 Erythrocyte density (g/ml): 1 Number protein components: 1 Property filename for component #1: ophbo2.01 Leukocyte volume fraction: 0 Platelet volume fraction: 0 Plasma volume fraction: 0 The sole varying parameter was the MCHC (in mass fractions) which was examined at 0.33, 0.20, 0.09, 0.5. Figure 5.8 shows the re sult with the MCV set at 80 fl. The most compelling trend to be noticed is that as the MCHC increases, there is an overall elevation of the spectra across the entire wa velength range due to the increase in light scattering. At the low MCHC values (0. 05 and 0.09), the charac teristic hemoglobin singlet (417 nm) and the doublet (500~600 nm) are visible. Even a small difference in mass fractions between 0.05 and 0.09 demonstrat es a substantial scattering increase for the higher MCHC (evidenced by the elevation of the 0.09 spectrum). At a hemoglobin mass fraction of 0.20, the shape of the spectru m is significantly different from that of 0.09 with the hemoglobin peaks becoming less prominent due to the peaks being somewhat masked by the elevation of the sc attering component. A physiological MCHC (0.33 mass fraction) shows the absence of th e hemoglobin peaks as they are completely masked by the scattered light. The increase in the internal hemogl obin content increases the n/n0 contrast which contribute s to the increase in whole particle scattering. Furthermore, the increase in hemoglobin dens ity suggests that less of the chromophores are being sampled by the incident light, a phe nomenon that would be less profound if the hemoglobin were free in solution. Similar si mulations done at constant sizes of 90 and 100 fl yielded the same trend across the four MCHCs shown in Figure 5.8.

PAGE 168

147 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 1902903904905906907908909901090 Wavelength (nm)OD MCHC = 0.33 MCHC = 0.20 MCHC = 0.09 MCHC = 0.05 blood cell conc = 1e-3 g/ml cell diameter = 5.35e-4 cm cell volume fracion = 0.45 Figure 5.8: Simulated spectra of erythrocytes at varying MCHC and constant MCV. Combined Effect of MCHC and MCV Changes Once the individual effects of MCHC a nd MCV variations had been determined by the model, it was important to examine c oncomitant changes in both parameters. The following simulation mimics a probable physio logical event where the red cells are altered in size due to changes in the osmotic environment. Associated with this change is a size-dependent alteration in concentrati on of the internal he moglobin depending upon whether the cell is swelling or sh rinking. If the cells swell, the influx of water will dilute the hemoglobin concentration and vice versa. As the cell volume (size) changed, the MCHC adjustments were calculated as follows:

PAGE 169

148 Given: MCHC = 33% at MCV = 90 fl 33% = 33 g/dl = 0.33 g/ml so, 0.33 g/ml x 90 x 10-12 ml = 2.97 x 10-11 g Hb Adjusting the MCHC in a cell volume of 95 fl: 2.97 x 10-11 g Hb/95 x 10-12 ml = 0.31 g/ml Thus the diluted MCHC mass fraction is 0.31. The table included in Figure 5.9 show s the MCV changes and the corresponding MCHCs. The hematocrit was held consta nt at 0.45 and the weight-based cell concentration in the cuvette was held at 1e-3 g/ml. Similar to a previous simulation, the result was normalized (area under the cu rve method) to eliminate any cell number variation so that each spectra was expressed on a per cell basis. The result was a direct comparison between cells of different sizes and their corresponding hemoglobin concentrations. The swelling and shrinking of red cells even within physiological size constraints produced significant qualitative changes in th e spectra. The cell with the largest volume (100 fl) contained the most d ilute hemoglobin (mass of he moglobin held constant) and vice versa for the smallest volume cell. Refe rring back to a previous simulation (Figure 5.7) that examined only changes in the MCV, the trend in this current simulation (Figure 5.9) shows an opposite effect. That is, wh en changing only the MCV (Figure 5.7), the particle with the largest si ze (100 fl) but with the same hemoglobin content experienced the largest flattening of the spectrum. Ho wever, in the case of swelling cells, the spectrum flattens when the MCV decreases and the MCHC increases. This can be explained by the dominant spectral manifest ation of increasing MC HC. The effect of

PAGE 170

149 increasing the MCHC (increase in n/n0) is evident in the drama tic changes in the spectra in Figure 5.8 and this is also responsible for the differences s een in Figure 5.9. Hence, the effects of the hemoglobin concentra tion override the effects of size. 0.0E+00 2.0E-04 4.0E-04 6.0E-04 8.0E-04 1.0E-03 1.2E-03 1.4E-03 1.6E-03 1902903904905906907908909901090 Wavelength (nm)Normalized OD Units rbc1 rbc2 rbc3 rbc4 rbc5 MCHC MCV(fl) 029 100 031 95 033 90 035 85 037 80 Figure 5.9: Combined spectral effect of cha nges in RBC hemoglobin concentration and volume. 5.2 Qualitative Analysis of Experimental Spectra In light of the theoretical studies presented in the previous secti on, the next step was to examine the outcome in an experimental sy stem to see if the predicted trends shown by Mie theory hold true when compared to e xperimental values. The concentration of hemoglobin in red blood cells was systematically altered by osmotic permeabilization, followed by restoration of the cells unde r physiologic osmolarity. A variety of

PAGE 171

150 concentrations and sizes we re obtained by controlling the incubation time following the permeabilization step. The sizes and concen trations were confirmed by using a SeronoBaker hematology analyzer. Each of thes e experimental sets was normalized using the area under the curve method from 230 – 900 nm. All of the spectra were subjected to solvent (isotonic phosphate bu ffered saline) correction, t hus subtracti ng the large, saturated saline peak below 230 nm renders th is region unreliable for data extraction. The wavelength range of 900 – 1100 nm also oc casionally showed errors in the way that the typically straight scattering region might tail downward due to instrumental uncertainty. Hence, the 230 – 900 nm range en compassed the most consistent data. The following figures test the ability of the model to accurately simulate combinations of varying MCHC and MCV. Figure 5.10 represents a group of experimental data selected over a range of MCHCs (0.048 – 0.336 mass fraction) each accompanied by varying MCV quantities. The inset shows raw data with each sample diluted to a concentration of approximately 4000 cells/l. The large plot is normalized data designed to amplify the features of the spectra to facilitate visual qualitative comparisons. As the MCHC increased, the overa ll optical density increased (inset). The differences in the MCV provided a more s ubtle contribution to the spectra. The normalized plot helped to show the emergence of the hemoglobin peaks as the encapsulated hemoglobin concentrations decr ease. The lowest MCHC sample (C4) showed the highest, most pronounced hem oglobin peaks whereas the highest MCHC sample (C1) exhibited a near absence of the characteristic peaks. The observed behavior is expected and consistent w ith the scattering theory; as th e concentration of hemoglobin

PAGE 172

151 within the cell is decreased, so is the contrast of the particle relative to the suspending medium, which results in a decrease of the scattering contribution the. Figure 5.11 presents the simulations of th e experimental spectra using the same MCHC and MCV values. The empirical trends are evident in the simulated spectra: 1) there is a flattening in the curve with increasing MCHC, 2) the hemoglobin peaks become more apparent with decreasing MCHC, and 3) the inset plots show that with a constant cell count, the increase in MCHC generates a dr amatic increase in th e intensity of the Figure 5.10: Experimental spectra of resealed cells with varying MCHC and MCV values. The MCHC is expressed in mass fractions and the MCV in fl. The inset represents the raw experimental data with each sample adjusted to a concentration of approximately 4000 cells/l. The large plot represents the normalized data where each raw data was divided through by the area under their respective curve. The normalized plot amplifies the features of the cu rves to facilitate visual comparison. 0.0E+00 5.0E-04 1.0E-03 1.5E-03 2.0E-03 2.5E-03 3.0E-03 3.5E-03 4.0E-03 4.5E-03 5.0E-03 1902903904905906907908909901090 Wavlength (nm)Normalized OD 0.336/89.9 (C1) 0.238/86.2 (C2) 0.084/71.0 (C3) 0.048/78.8 (C4) MCHC/MCV 0.0 0.2 0. 0.6 0.8 1.0 1.2 1. 190290390 905906907908909901090 Wavelength (nm)OD

PAGE 173

152 overall spectrum. The good agreements seen between the experimental and simulated spectra of the unmodified and hypotonically modified erythrocyt es indicate the effectiveness of the scattering theory in descri bing the spectral feat ures as functions of the size and the hemoglobin concentration. Figure 5.11: Simulated spectra of resealed cel ls using the experi mental MCHC and MCV values in Figure 5.20. The MCHC is expressed in mass fractions and the MCV in fl. The inset represents simulations of the raw experimental data. The large plot represents the normalized data where each raw data was divided through by the area under the respective curves. The normalized plot amplifies the features of the curves to facilitate visual comparison. 0.0E+00 5.0E-04 1.0E-03 1.5E-03 2.0E-03 2.5E-03 3.0E-03 3.5E-03 4.0E-03 4.5E-03 5.0E-03 1902903904905906907908909901090 Wavelength (nm)Normalized OD 0.336/89.9 (D1) 0.238/86.2 (D2) 0.084/71.0 (D3) 0.048/78.8 (D4) MCHC/MCV 0.0 0.2 0. 0.6 0.8 1.0 1.2 1. 190290390 905906907908909901090 Wavelength (nm)OD

PAGE 174

1535.3 Conclusions This chapter examined the effectivene ss of the Mie theory based model to simulate the spectral features and tre nds of purified and hypotonically modified erythrocytes. The acquisition of the experimental data of red cells possessing various MCHC and MCV values revealed features that were similar to the predictions made in the previous chapter. Furthe rmore, simulations of experimental data using corresponding values of hematologic parameters obtained from the hematology analyzer provided good agreement of the overall features. As tw o major parameters, MCHC and MCV, were changed, the trends of the spectra were m odeled successfully to pa rallel those of the experimental spectra. Increasing MCV (at c onstant MCHC) showed a slight decrease in the absorption region (<600 nm), however this effect was not as dramatic as changing MCHC at constant MCV. In this latter cas e, increasing the MCHC drastically increased the whole particle scattering of the cell and this resulted in an eleva tion of the entire OD spectrum accompanied by the masking of the char acteristic hemoglobin absorption peaks. Moreover, as the MCHC and MCV were both varied, the dominant parameter proved to be the MCHC in dictating the spectral trends Also, normalized spectra showed that as the cellular hemoglobin content increased, the absorption peak decreased, strongly indicating the presence of macroscopic hypochromism at higher hemoglobin concentrations. This successful attempt to qualitatively simulate the purified and modified red cells allowed us to take the next step in our investigation: to implement a Mie interpretation model for the quantification red cell suspensions.

PAGE 175

154 Chapter 6: Application of th e Interpretation Model for the Quantitative Analysis of Erythrocytes To this point, we have identifie d the presence of macroscopic hypochromicity as a scattering-related effect in the red cell spect ra. Moreover, the features and trends of experimentally modified red blood cells were successfully simulated using the mathematical model. This chapter describes how the method was further extended to an interpretation model that permits the extraction of particle information from experimental spectrum incl uding chemical composition, particle size, and particle number. The quality of resu lts from the interpre tation provide strong evidence that molecular hypochromicity is neg ligible in the UV-visible spectrum of a red blood cell system. In Chapter 3 the concept of molecular hypochromicity was introduced. Previous studies6,7 suggested the possibility of hypochromic ity in red blood cells because of its high hemoglobin content. Considering that hemoglobin is a st rong chromophore and it exists in large concentration inside the red cell, testing fo r molecular hypochromicity is not a trivial issue. If it exists, a hypochromi city correction would be required in addition to the implementation of a scattering and abso rption model for the qua ntification of total hemoglobin in a red cell suspension. This test for molecular hypochromicity was to determine whether the interpre tation model that accounts fo r scattering can provide good

PAGE 176

155 estimates of the total hem oglobin concentration. The scattering model based on Mie theory simultaneously provides estimates of MCHC, MCV, and RBC counts, all parameters that can be measured indepe ndently. The good agreement found herein between the spectroscopy-based estimates a nd the measured values provide further support to our conclusion that molecular hypoc hromicity is not significant for red blood cells. 6.1 The Interpretation Model The kernel of the interpretation model is the turbidity equation based on Mie theory (as described extensively in Chapter 3) and it provides adequate estimations of the above-mentioned parameters for a given sample spectrum. One of the big advantages to this method is its capability to efficiently analyze the spectrum across a broad wavelength range, providing a redundant mathematical ch eck that increases the reliability of the estimated parameters. This is in contrast to many other studies using one or only a few selected wavelengths for their estimations.79,80,81 The general input-output format for the model is illustrated in the schematic diagram in Figure 6.1. The two major inputs are the optical properties of the chromophore (hemoglobin) and the measured spect rum. The computation consists of an iterative process to determine the best solution set for the output parameters which correspond to a good statistical fit of the sp ectrum. The wavelength range used for analysis was 240 – 900 nm. The range from 190 nm to 239 nm was omitted due to the solvent peak interfering with the sample sp ectra. Figure 4.6 (Chapter 4) showed a

PAGE 177

156 spectrum of the solvent (0.9% PBS) that showed a large saturated peak < 240 nm. A background correction (subtracti on of solvent spectrum from sample spectrum) of the sample spectrum would render th is region unreliable for spec tral interpretation. The wavelength range of 901 1100 nm was omitted due to spectral perturbations seen in this range resulting from an inconsistent instru mental light source. Although the diagnostics indicated that the lamp was producing inte nsities in this wave length range within acceptable limits, some of the spectra showed dips below the baseline after background correction. It should be not ed, however, that this omissi on did not alter the estimated values for the following reason. The scattering only region extends from approximately 620 nm to 1100 nm, a range that is largely linear. Hence, analyzing 620 900 nm is a good representation of the en tire scattering wa velength range. Model* Optical Properties of chromophore(s): ( ), n( ) Measured Spectrum Particle Count Avg. Particle Diameter Mean Corpuscular Hemoglobin Concentration Particle Size Distribution 0 0 0 0 0 90 90 90 0 0 90 90 90 90 090 Wvg Spectral Fits dDDfmQDNext p)())(,( 42 0 Figure 6.1: Schematic diagram of th e interpretation model. I nputs consist of the optical properties of the chromophore and the experimentally measured spectrum. The outputs include the particle count, aver age particle diameter (of an equivalent sphere), the mean corpuscular hemoglobin concentration, particle size distribution, and the spectral fit.

PAGE 178

157 Due to the multiple parameters composed in a solution set, regularization methods were implemented to confine the possible es timates and arrive at a reasonable solution while also limiting solutions to a statistically good fit of the spectrum.119 To achieve this, the discretization of the integr and of the turbidity equation can be written in a matrix form as a quadrature approximation = Af + (Eq. 6.1) where represents the sum of the measurement error (m) and the quadrature error introduced in the discretization (c).119,120 The least squares solution to the discretized model in terms of the particle size distribution (PSD) is defined as T T lsA A A f1) ( ˆ (Eq. 6.2) However, small errors in the quadrature or e xperimental measurements can translate into large errors for lsf ˆ. Regularization of this equati on leads to a constrained solution defined by T T lsA H A A f1) ( ˆ (Eq. 6.3) where is a regularization parameter and H is a covariance matrix that helps to narrow the possible solution sets that arise from th e propagation of error. There exists an acceptable value for such that the error of lsf ˆ is minimized. This value is estimated using the Generalized Cross Validation (G CV) technique described by the equation 2 2)] ( [ )] ( [ ) ( Z I trace Z I m V (Eq. 6.4)

PAGE 179

158 where Z()=A(ATA+I)-1AT.120 The variable m represents the number of turbidity measurements with respect to wavelength.117,121 The result of these equations is a reliable equivalent sphere particle size distribution for use with the turbidity equation in the interpretation model. Variants of the Interpretation Model The interpretation model was modified in three ways to provid e slightly different approaches to the interpretation of the resealed red cell system.69 Table 6.1 identifies and describes the three versions. Version Name Description RBCHb01a Choice of one encapsu lated chromophore (oxyhemoglobin) RBCHb01b Choice of two encapsulated chromoph ores (oxyand another hemoglobin derivative such as methemoglobin) RBCHb02a Choice of one chromopho re but with the option to put a fraction of the chromophore outside of the cell in the medium Table 6.1: The three versions of the interpreta tion models and the va riations in their functions. Version 01a is the simplest of the th ree with 01b and 02a introducing additional parameters for the prospect of elucidating the better solution set. However, increasing the number of parameters adds more complexity to the calculations as well as amplifying the potential error in the solution set such th at the ideal situation would be if RBCHb01a proved to be an adequate model for a majority of the measured data. Therefore, the data was first analyzed with 01a while implementi ng the subsequent versions on an as-needed basis.

PAGE 180

159Corroboration of the Estimated Parameters While the interpretation model is a powerful tool for extracting particle information from a single spectrum, its deve lopment requires valida tion of the observed parameters using separate anal ytical methods. These methods were previously described in the Chapter 4. Briefly, the hemato logy analyzer (Serono-Baker system) gives impedence-based values for the cell counts and the cell volume, with the volume being readily calculated to the equi valent sphere diameter using the equation for the volume of a sphere (v=4r3/3). The hemoglobin concentration can be calculated in two independent manners albeit both are based on the same Drabkin’s principle (d iscussed in Chapter 4).46,47 The first is the modified Drabkin’s assay auto mated by the hematology analyzer. The second is the manual Drabkin’s assay, used as a ve rification tool for the values obtained via instrumental analysis. A lthough the two methods share th e same principle, their executions bring about a systematic offset that needs to be addressed. Figure 6.2 is a scatter plot comparison of hemoglobin concen trations (HGB in g/dl) acquired from the two methods using an array of data points obt ained throughout this st udy. The solid line represents a 45 degree reference line wh ere a perfect correlation between the two methods of measurement would put the data poin ts on the reference line. It is evident, however, that there is a relatively consistent bias in the data. The offset shows the Drabkin’s values being slightly greater than the hematology analyzer values. Using the Serono numbers as the reference, the calculate d percent (%) offset fo r the large (~20 g/dl) and medium (~10 g/dl) hemoglobin concentratio ns are within 10%. Although the best fit

PAGE 181

160 line tapers towards the juxtaposed reference line at the lower values as expected, the numerical value for the % offset increases due to the decrease in the numerical values. Furthermore, the data points at the larg er values (>20 g/dl) show a more pronounced deviation from the reference line. The di screpancy in the two techniques can be y = 1.0527x + 0.5674 R2 = 0.9974 0.0 5.0 10.0 15.0 20.0 25.0 0510152025 Serono-Baker (g/dl)Drabkin's (g/dl)45 degree reference Best-fit line for data points Figure 6.2: Comparison of hemoglobin concentrati on (HGB) values (g/dl) measured by the Serono-Baker and the Drabkin’s assay. Th e solid line is a 45 degree reference line, the scattered points represent the data points, and the dotted line is a linear regression of the data points. attributed to a systematic error which is not uncommon when comparing two such methods.122,123 As long as the bias is considered, the Drabkin’s method can still be used as a validation for the Serono-Baker and the estimated values from the interpretation model.

PAGE 182

161 The hemoglobin concentration values co mpared in Figure 6.2 are raw numbers obtained directly from the measured spectra. In the case of both methods, the red cells are lysed, the optical density is taken at a predetermined wavelength and then translated into the hemoglobin concentration using a ca libration curve. Since the Serono-Baker is an automated system, the error associated with the HGB reading is minimized. This is in contrast to the manual Drabkins method in wh ich one must take into account the human y = 0966x + 00332 R2 = 0996 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 00.050.10.150.20.250.30.350.4 Serono-Baker (mass fraction)Drabkin's (mass fraction)45 degree reference Best-fit line for data points Figure 6.3: Comparison of the mean corpuscula r hemoglobin concentration (MCHC) values (mass fraction) measured by the SeronoBaker and the Drabkins assay. The solid line is a 45 degree reference line, the scattered points repres ent the data points, and the dotted line is a linear regre ssion of the data points.

PAGE 183

162 pipetting error as well as instrumental errors. The difference is illustrated in an analysis where six replicates of the same sample gave an error of 0.7% (standard deviation). Replicates of the Drabkin’s assay showed a higher error of 1.4%. The offset of the two methods can be further represented as a comparison of MCHC in Figure 6.3. It must be consider ed, however, that the MCHC is a calculated value instead of a raw number like the HGB In the equation MCHC = HGB/HCT, it is evident that the computation introduces the hematocrit (HCT) value that is obtained by the Serono-Baker, an additional source of error. Thus this error is introduced into the MCHC obtained via both analytic al formats. The plot shows that the data points at the lower MCHCs have a relatively tight fit to the linear regression whereas the higher MCHC values become slightly more scattered. Testing the Interpretation Model Thirty-four samples of purified and reseal ed red blood cells were analyzed to give good estimates red cell parameters. This section presents analyzed results for representative samples in the high, medium and low MCHC ranges. In order to explain the inputs and outputs of the model, a specifi c example of a resealed red cell sample in the low range MCHC (sample ID: r41603g) w ill be used. The program inputs for the RBCHb01a interpretation model with resp ect to this sample are as follows: Problem number: 1 Cell pathlength (cm): 1 Particle density (g/ml): 1 Sample filename: r41603g.dat Wavelength range for analysis (nm): 240 900 Optical properties file or partic les and suspending medium: ophbo2.01

PAGE 184

163 Enter filename for estimation results: r41603g.txt Number of particle populations: 1 Fraction of chromophores in particles: 0.05 Estimate of Dn (nm): 4500 Estimate of sigma: 0.1 Enter filename for calculated PSD: r41603g.psd Enter filename for calcula ted spectrum: r41603g.clc Some of the parameters are self-explan atory. The optical properties file contains the absorption coefficient and the refractive index of oxyhemoglobin, and also the refractive index of the suspending medium (water). The “particle population” refers to the particle size population defined by the mean, where up to three distributions are allowed by the program. The majority of our analyses kept the population at one since this provided successful outcomes. The frac tion of chromophores fi eld is an estimation of the hemoglobin mass fraction that provides a starting point for the model to facilitate convergence. The value 0.05 was chosen as a round number close the Serono-Baker output (0.062). Dn is the estimated number based diameter of the sphere which was chosen to be 4500 nm, slightly smaller than th e equivalent sphere di ameter of a cell of volume 80 fl (5350 nm). The breadth of the variance () for an equivalent lognormal size distribution was estimated to be 0.1. The spectral output file is illustrated in Figure 6.4 along with some of the key estimated parameters. The data includes the measured spectrum fitted to the calculated spectrum. The residuals indicate the differe nce between the measured and calculated. The Beer-Lambert spectrum of the chromophore shows th e absorption spectrum of oxyhemoglobin if the total mass of the encap sulated protein remained constant but existed free in solution in the same sample vol ume. An important feature to note in the

PAGE 185

164 calculated spectrum is its clos e fit to the measured data The scattering region (>600 nm), devoid of any absorption effects, shows a n ear-perfect overlay. Slight differences in the absorbing region of the spect rum may be attributed to either assumptions made in the modeling, or perhaps there ar e aspects of the biological sy stem not being considered. Such issues will be discusse d later in the chapter. -0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 1902903904905906907908909901090 Wavelength (nm)OD Measured Calculated Residual Abs of Hb Calculated: MCHC = 0.084 Dn = 5.46 microns MCV = 85.2 fl Count = 2.12E6 cells/ l RSSQN = 0.96 Measured: MCHC = 0.062 D = 5.29 microns MCV = 79.8 fl Count = 3.03E6 cells/ lr41603g Figure 6.4: Output of the spectral estimate by the RBCHb01a interpretation model for low range MCHC. The measured and calculate d spectra closely overlay each other. The residuals represent the difference between th e calculated and measured spectra, and the absorption of hemoglobin illustrates the abso rption spectrum if the hemoglobin resided free in solution while keeping the protein mass and sample volume constant. Figure 6.4 is representative of successf ul interpretations achieved for resealed cells with the MCHC in the low range (< 0.10 mass fraction). In the case of the MCHC

PAGE 186

165 value for this sample (r41603g), the measured quantity is reported to be 0.062 by the Serono-Baker (as shown in the figure) and 0. 098 by the Drabkin’s as say. The estimated value of 0.084 falls in the range between the two measur ed numbers. The MCV value itself is estimated to be 85.2 fl, slightly diffe rent from the measured quantity of 79.8 fl. The difference could be related to the instrumental error of the measured value and/or the flexibility of the numerical solutions within th e statistical tolerance of the interpretation model. To elaborate this latter point, a few requirements need to be met to arrive at a reasonable solution: 1) the solution must provide a good fit to the empirical spectrum across the entire wavelength range of analysis, and 2) the values assigned to each of the important parameters (diameter, cell count, hemo globin concentration, etc) must all make MCHCMCV Figure 6.5: A representation of the algorithmic pro cess of the interpretation model.

PAGE 187

166 sense within physiologic constr aints. Figure 6.5 illustrate s a simplified example where statistically viable solutions exist within th e limits of the oblong el lipse defining the error tolerance along a reference line. However, relevant solution sets reside within the smaller oval and the regularization technique narrows the possible solutions within these limits. -0.2 0.1 0.3 0.5 0.7 0.9 1.1 1902903904905906907908909901090 Wavelength (nm)OD Measured Calculated Residual Abs of Hb Calculated: MCHC = 0.241 Dn = 557 microns MCV = 90.4 fl Count = 2.16E6 cells/ml RSSQN = 0.99 Measured: MCHC = 0.228 D = 5.49 microns MCV = 89.1 fl Count = 3.10E6 cells/mlr6363b Figure 6.6: Output of the spectral estimate by the RBCHb01a interpretation model for medium range MCHC. The measured and cal culated spectra rese mble one another, particularly in the scattering region > 600 nm The residuals repr esent the difference between the calculated and measured sp ectra, and the absorption of hemoglobin illustrates the absorption spectrum if the hemoglobin resided free in solution while keeping the protein mass and sample volume constant. The outputs for the hemoglobin concentration, size, and number paramete rs match well with the measured values. Figure 6.6 is a representative result of a spectral interpretation by RBCHb01a for a resealed red cell population in the medi um range of hemoglobin concentrations. In

PAGE 188

167 terms of protocol, the length of time the ce lls were subject to permeabilization in the hypotonic buffer was reduced to approximately one minute so that less hemoglobin was released. The result was a higher encapsula ted hemoglobin concentration compared to the data set represented by Figure 6.4. Higher hemoglobin conten t raised the overall refractive index of the particle, hence increasing the scattering component of the particle. -0.2 0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5 1.7 1902903904905906907908909901090 Wavelength (nm)OD Measured Calculated Residual Abs of Hb Calculated: MCHC = 0.311 Dn = 5.47 microns MCV = 85.5 fl Count = 3.93E6 cells/ l RSSQN = 1.12 Measured: MCHC = 0.32 D = 5.73 microns MCV = 101.7 fl Count = 4.06E6 cells/ l101698 Figure 6.7: Output of the spectral estimate by the RBCHb01a interpretation model for physiological range MCHC. The measured and calculated spectra resemble one another, particularly in the scattering region > 600 nm The residuals repr esent the difference between the calculated and measured sp ectra, and the absorption of hemoglobin illustrates the absorption spectrum if the hemoglobin resided free in solution while keeping the protein mass and sample volume constant. The outputs for the hemoglobin concentration, size, and number paramete rs match well with the measured values.

PAGE 189

168 Increase in the scattering pattern adds to the complexity of modeling the total spectrum, particularly in the wavelength range < 600 nm that contains significant absorption information. This is reflected in the mode l’s difficulty in closel y reproducing all of the detailed features of the experimental spect rum, especially at the lower wavelengths. From a quantitative persp ective, however, the three major parameters, MCHC, diameter/volume, and cell count all yield good results. Similar results were seen in samples with MCHC values in the physiol ogical range (Figure 6.7). Experimentally, these erythrocytes were is olated from the major components of whole blood with no modification to the erythrocytes themselves. The interpreted spectral fit shows excellent correlation in the scattering-only region at wavelengths above 600 nm. The estimated MCHC and the cell counts s how remarkable correspondence. The size values obtained from the experimental and calculated spectra were less similar, however, the calculated volume still falls within the physiological range (80 – 100 fl) and is a plausible value. 6.2 Quantitative Results of the Interpretation Model As mentioned previously the turbidity equation depends on not only the optical properties of the er ythrocytes, but also a number of important parameters describing the particles: the cell size (equivalent sphere diameter/volume), the hemoglobin content, and the cell count. Provi ded that the refractive index of the medium is properly accounted for,73 the interpretive results of the model generally showed satisfactory values for each of the para meters compared to their corroborative experimental values.

PAGE 190

169 The three different versions of the interpretation model listed in Table 6.1 offered varying options in the way the data was evaluated. Program version RBCHb01a simulated the simplest case, assuming that all of the hemoglobin was encapsulated inside the cells in the form of oxyhemoglobin. We found that this ve rsion of the program converged to the most physiologically consis tent solutions (43 of 59 samples analyzed converged to adequate values). The other tw o versions of the interpretation model only 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 00.050.10.15020.2503 0.35 0.4 Calculated MCHC (mass fraction)Measured MCHC (mass fraction) Serono-Baker Drabkin's Linear (Serono-Baker) Linear (Drabkin's) Figure 6.8: Comparison of the calculated MCHC to the MCHC values measured by the hematology analyzer and the manual Drabkins assay. The scatter plots were fitted with a linear regression curve. Both methods show a high level of corroboration to the estimated data. The separation of the two lines is attributed to the systematic offset inherent between the two methods of measurement.

PAGE 191

170 improved the estimations in isolated cases. The bulk of the results represented herein were obtained from RBCHb01a. Quantitative Results Obtained with RBCHb01a The mean corpuscular hemoglobin concentration (MCHC) value was reported by the Serono-Baker hematology anal yzer and estimated by the interpretation model. Additionally, the MCHC was calculated from the hemoglobin concentration value obtained from the Drabkin’s assay. Figure 6.8 shows a comparison of the estimated MCHC to those acquired from the tw o independent experime ntal procedures. Sample Serono-Baker Estimated Drabkin's 6399a 0.036 0.050 0.065 6399b 0.041 0.056 0.077 6399c 0.043 0.045 0.074 6399d 0.047 0.047 0.084 6999a 0.048 0.051 0.077 r5103d 0.084 0.084 0.125 r5103e 0.090 0.083 0.111 r41703d 0.095 0.128 0.168 r51403c 0.199 0.241 0.226 r6303b 0.228 0.241 0.247 r51403d 0.229 0.251 0.246 r6303a 0.238 0.251 0.271 r6303g 0.306 0.333 0.355 r41703a 0.333 0.335 0.365 r51403a 0.343 0.358 0.372 Table 6.2: Table of estimated and measured values (g/dl) for representative erythrocyte samples containing various amount s of hemoglobin. The calculated values generally fall between the two measured values with a few exceptions.

PAGE 192

171 A perfect corroboration woul d be represented by a 45o line passing through the origin. The separation of the two best-fit lines is related to the systematic offset of the Drabkin’s assay and the Serono-Baker discu ssed earlier in the chapter (F igures 6.2 and 6.3). Table 6.2 offers an abbreviated look at some represen tative values plotted in Figure 6.8. The estimated (calculated) values largely fall be tween the two measured values with a few exceptions, indicating a general consistency in the results of the model. Interestingly, both measured data sets in Figure 6.8 show an increased scatter in the points as the 0 5 10 15 20 25 30 0510152025 Calculated HGB (g/dl)Measured HGB (g/dl) Serono-Baker Drabkin's Linear (Drabkin's) Linear (Serono-Baker) Figure 6.9: Comparison of the calculated hemoglobi n concentration (HGB) to the HGB values measured by the hematology analyzer and the manual Drabkin’s assay. The scatter plots were fitted with a linear regression curve. Both methods show a higher level of corroboration to the estimated values than the MCHC data. The separation of the two lines is attributed to the systematic o ffset inherent between the two methods of measurement.

PAGE 193

172 hemoglobin values become larger. However, the same trend is seen in the direct comparison between the two physical methods in Figure 6.3 hence the increased scatter at the high hemoglobin concentra tions is a result of measurement variability. A similar comparison was made in terms of the total hemoglobin concentration (HGB) expressed in g/dl. This term represents the hemoglobin concentration quantified as the cells in su spension were lysed and the hemoglobin was diluted into solution.47 The Drabkin’s assay was th e standardized method for the acquisition of this parameter. The hematol ogy analyzer used a modi fied Drabkin’s assay to collect the same information. The plot (Figure 6.9) illustrates a linear regression for both methods, each of which shows a high co rrelation with the calculated values. Consistent with the MCHC data, there is more scatter at the higher values than the low range. The tighter distribution of the point s in Figure 6.8 is suggestive of the HGB parameter’s unprocessed nature whereas MC HC is a value calculated from HGB and hematocrit (HCT) with an implied propagation of error. Importantly, the estimation successfully accounted for the hemoglobin mass balance without the implementation of any correction for molecular hypochromism. Furthermore, comparisons were made between measured and calculated sizes of the cells, both expressed as equivalent sphere diameter and volume. The hematology analyzer provided the validation for the estima ted sizes. The equivalent sphere diameters estimated by the interpretation model seemed to be reasonable estimates of the measured values for a large proportion (43/59) of sample s (Figure 6.10). The calculated values for particle were close to or in the ra nge of physiological values (5.35 – 5.76 m). Greater

PAGE 194

173 00 10 20 30 40 50 60 70 12345678910111213141516171819202122232425262728293031323334 Sample #Equivalent Sphere Diameter (micrometers) Calculated Measured (S-B) Figure 6.10: Calculated vs measured equivalent sphe re diameters. The measured values were obtained from the Serono-Baker hemato logy analyzer and the calculated values were estimated by the interpretation model. than 90% of the calculated estimates were within a 10% error of their measured counterparts. When these values were tran slated into volumes, even small differences were magnified due to the exponent in th e spherical volume equation (Figure 6.11). Nevertheless, the differences between th e calculated and measured MCVs are proportionate to their counterpa rts reported in Figure 6.10. Moreover, the phase contrast micrographs (Chapter 4) indica ted that the sizes were consis tent with both measured and estimated values. Hence it can be generall y stated that the estimated volumes for a majority of our 59 experimental samples are in or approaching the physiological range of the erythrocytes.

PAGE 195

174 0 20 40 60 80 100 120 140 12345678910111213141516171819202122232425262728293031323334 Sample #MCV (fl) Calculated Measured (S-B) Figure 6.11: Calculated vs measured cell volumes. The measured values were obtained from the Serono-Baker hematology analyzer a nd the calculated values were estimated by the interpretation model. Figure 6.12 illustrates the comparison be tween calculated and measured cell counts of purified and modified red cells. Although some of the smaller counts show significant variability, there ar e many samples that show very good agreement similar to the comparisons of cell volume in the previous figure. Of the 34 samples analyzed, the estimated values of approximately 70% of the samples were within 20% error of the measured values. It is important to note th at the measured MCV and the cell counts have instrument-associated error as defined in Tabl e 4.2 (Chapter 4). Moreover, the calculated values can fluctuate within the limits of accep table solution sets as indicated previously in Figure 6.5.

PAGE 196

175 0.0E+00 1.0E+06 2.0E+06 3.0E+06 4.0E+06 5.0E+06 6.0E+06 7.0E+06 8.0E+06 9.0E+06 12345678910111213141516171819202122232425262728293031323334 Sample #RBC (cells/ul) Calculated Measured (S-B) Figure 6.12: Calculated vs measured cell counts. The measured values were obtained from the Serono-Baker hematology analyzer a nd the calculated values were estimated by the interpretation model. To further explore the variability of the measured values of cell count and cell size in comparison to the calculated numbers, a Coulter Z2 counter was used. Although the counts were obtained using elect rical impedence measurements like the Serono-Baker hematology analyzer, it still pr ovided another level of corroboration to validate the interpreted numbers. The c ounter is reported to give reproducible RBC counts within 5% and has an upper linearity limit of 8 x 106 cells/l.124 Tables 6.3 presents counts for six resealed cell samp les measured on the Coulter and the SeronoBaker, and their calculated counterparts. Samples 1 – 3 were replicate samples of resealed cells prepared in parallel as were samples 4 – 6. Within each sample set, the

PAGE 197

176 counts and sizes were relatively consistent fo r each instrument. Interestingly however, there were constant systematic offsets be tween the two impedance counters for each set of parameters. The calculated values obtai ned from the interpre tation model generally fell close to the range of both measured values, but the precision appeared samplerelated. The first sample set (1 – 3) yiel ded calculated values for both parameters that were closer to the measured values than the second sample set (4 – 6). The calculated equivalent sphere diameters (Table 6.4) s howed better agreement to the experimental values. Coulter (106/ l) Serono-Baker (106/ l) Calculated (106/ l) Sample 1 3.21 4.01 3.30 Sample 2 3.59 4.08 3.38 Sample 3 3.64 4.15 3.28 Sample 4 1.73 2.23 2.74 Sample 5 1.69 2.16 3.60 Sample 6 1.60 2.20 3.52 Table 6.3: Comparison of cell counts of resealed cells between measured (Coulter Z2 and Serono-Baker 9110) and calculated values. Coulter (m) Serono-Baker(m) Calculated(m) Sample 1 5.18 5.27 5.06 Sample 2 5.18 5.25 5.13 Sample 3 5.32 5.23 5.16 Sample 4 5.58 5.44 5.82 Sample 5 5.58 5.44 5.50 Sample 6 5.59 5.44 5.65 Table 6.4: Comparison of the equivale nt sphere diameter (m) of resealed cells between measured (Coulter Z2 and Serono-Ba ker 9110) and calculated values.

PAGE 198

1776.3 Discussion It has long been debated over the last 50 years whether or not molecular hypochromism plays a role in the optical charac teristics of encapsulated systems such as red blood cells and chloroplasts.6,85 Considering the success of the interpretation model, however, it can be stated with confidence that molecular hypochromism is insignificant in the spectral analysis of red blood cells, a nd most probably in similar systems. The extension of Mie theory to accurately quan tify the total hemoglobi n concentration in a suspension of red cells is a la rge contribution to both fundame ntal and applied science. Moreover, the success in estimating other im portant red cell parameters including cell count, MCHC, MCV, and cell diameter streng thens the validity of the interpretation model. When the interpretation model was subj ected to multiple evaluations of the same red cell data, the algorithm consistently returned the same HGB value, although the MCHC and MCV fluctuated s lightly and dependently betw een trials. This can be explained by examining how the interpretati on model calculates these values. Consider the relationship Hb pf D N HGB 63 where the middle term in parentheses is the volume of a sphere, is the density of the cell, and fHb is the mass fraction of hemoglobin. Th e interpretation program first attempts to fit a normalized spectrum, hence the particle concentration (Np) is eliminated from the equation. Therefore, the two estimated values are D and fHb. These two parameters

PAGE 199

178 fluctuate in accordance with each other to give a consistent HGB value that correlated well with the measured value. Furthermore, the success of the estimated parameters was accompanied by good estimated fits of the measured spectra. The fits were best across the entire wavelength range (the data was fitted from 240 to 900 nm) for the low MCHC values (~0.05 – 0.10 mass fraction). This could be a result of one or more combined effects. First, the resealed cells appeared to take on a more spherical shape as evident in the phase contrast micrographs, thus approaching the s pherical approximation used in Mie theory. Second, and perhaps more important, the lower contrast of resealed cells to medium due to the low hemoglobin concentration decreased whole particle scattering characteristics, simplifying the interpretation of the spectrum. As the MCHC increased, the scattering component increased adding to the complexity of the spectrum. At the wavelengths > 600 nm where mainly the scattering component is represented, a good overlap between the measured and calculated spectra is seen. At the low wavelengt hs, a combination of scattering and absorption components are repr esented. Here, there are differences in estimated and calculated spectra however, the areas under the cu rves are similar. This appears to reflect a compensation for variati ons in spectral match that allows us to achieve relevant values for red cells. The experimental spectrum of the high MCHC (Figure 6.7) shows a smoothed curve compared to the estimated spectrum. This effect may be the result of distribution of cell or ientations since it has been shown that a broadening of the distribution causes curve smoothening (unpublished work).69 Thus accounting for cell shape and orientation in th e calculations should further improve the

PAGE 200

179 interpretation model, which is already providing good estimates of red cell parameters. Additionally, there are ot her distributions that merit cons ideration. The particle size distribution for the resealed ce lls could be slightly irregular compared to that of the unmodified red cells. Microscopy pictures suggest that the sizes range in the approximate diameters from 4 to 8 m. In some cases, the dist ribution could be bimodal. Moreover, the MCHC distribution in the res ealed cell samples could be heterogeneous, but to what degree remains to be seen. Examination of such issues should further improve upon an already satisfactory interpretati on model. To conclude, in spite of the fact that we are using a simp lified spherical approximation, the fact that there is good agreement between spectrosc opy-based estimates and sta ndard measurement strongly suggests that improvement of the scattering m odel is the right direc tion in completing the interpretation of th e red blood cell.

PAGE 201

180 Chapter 7: Conclusions Multiwavelength UV-visible spectrophotometry is a versatile method for the characterization of suspended particles of various sizes. Optical analysis of particles such as red blood cells have been attempted on di fferent levels (single wavelength or small wavelength range) and problems such as phot ometer design and the possibility of hypochromism have been identified. Howeve r, this is the first body of work that addresses the full array of f undamental questions associated with macroscopic particle characterization and successfully implements a reliable interpretation model for the quantification of the im portant parameters of red blood cells The contributi ons are listed as follows. The concept of hypochromicity was revised in the context of obtaining precise estimates of hemoglobin concentration in red blood cells from spectroscopic measurements. Hypochromism (a decreas e in the absorption spectrum with the increase in the concentration of a strong chromophore) in the past had only been defined as a phenomenon cause by electronic molecular interactions. In order to characterize red blood cells us ing the light scattering th eory, it was necessary to adjust the perception of hypochromicity. Two levels of hypochromicity were establishe d to better describe the optical behavior of macroscopic particles. Molecular hypochromicity is th e aforementioned effect due

PAGE 202

181 to particle interactions on the molecular level. Macroscopic hypochromicity is the attenuation of the absorption component as a result of increased lig ht scattering as the refractive index of the part icle increases. Red blood cells with high physiological concentrations of hemoglobin exhibit a particularly pronounced macroscopic hypochromicity. Instrumental configuration was shown to have a significant effect on the spectrum and the way it is interprete d. In previous studies,6,7 the effect of photometer design (with respect to the angle of acceptance) on macroscopic particle spectra has been confused with molecular hypochromism. Th is study examined the importance of the acceptance angle on the red cell spectrum and in the process, we established that a small angle transmission spectrum containe d the best balance of absorption and scattering information. A modified protocol for th e hypotonic modification of re d blood cells was developed, with the ability to control the resulting encapsulated hem oglobin concentration. The resealed cells were essential in the i nvestigation of the optical behavior of red cells as the refractive index and the size of the particles changed. The light scattering theory of electromagnetic radiation was used to successfully account for the scattering and absorption components of red cell spectra. The capability to analyze both the scattering a nd absorption components of a large particle suspension allowed for the close examinati on of the two types of hypochromicity. In the context of light scattering theo ry, the Mie theory was implemented and extended to account for chemical compositi on across the entire wavelength range

PAGE 203

182 (190 – 1100 nm). Although the theory assumes spherical particles, results showed its capability to reliably interpret red blood cell spectra in physiological and modified states. Using the Mie theory, trends and features of experimental red cell spectra were successfully simulated. Simulations were helpful in predicting the outcomes of experiments, and also to suggest the direction of future experiments. The interpretation model base d on the Mie theory yielded realistic values of MCHC, MCV, Np, and HGB for purified and resealed red cell samples. The success of the interpretation model in giving reasonable values for the hematologic parameters, particularly th e total hemoglobin co ncentration (HGB) indicated that there was no si gnificant molecular hypochromic effect as suggested in previous studies. The implications of this work encompa ss the understanding of fundamental light scattering concepts and the possibilities of a rapid, inexpensive and reliable analysis of whole blood in a clinical sett ing. Since the contributions of red blood cells dominate a spectrum of whole blood, a good ch aracterization of red cells was necessary in order to examine the effects of lesser optical contribu tors such as platelets and leukocytes. Furthermore, since the basic ideas are in pl ace, multiwavelength light scattering analysis can be applied to encapsulated particle systems similar to red blood cells.

PAGE 204

183Future Work The spherical restriction of the Mie theory may have been the reason for our inability to perfectly fit all of the features of the red cell spectrum, particularly for the cells containing a high concentration of hemogl obin. At high refractive indices of red cells, the scattering becomes more domina nt and parameters such as shape and orientation could become important to the features of the spectrum. Current work involves other members of the research group a ttempting to incorporate form factors into the analysis to examine the effects of shape and orientation on the spectrum.69,125 Moreover, with regards to the characte rization of resealed cells, there were physical details of the cells that were not accounted for by inspecting the sample with the hematology analyzer. For example, when the cells were permeabilized and resealed, the resulting MCHC most probabl y did not represent a uniform distribution of hemoglobin concentration across the entire population. The Serono-Bake r reports an average value and is incapable of accounting for multiple populations. Simulations can be used to determine the impact of hypothe tical distributions of hemogl obin concentrations. If the calculated model suggests that it is significant, then the resealed cells could be subjected to density separations to better characterize the cell spectra. Experimental model systems can be used to further validate the results of the resealed cells. Liposomes were proposed in our studies as an alternative system, but results showed difficulty in producing unila mellar vesicles (Appendi x C). A different protocol should be chosen to refine the quality of the liposomes to facilitate their characterization. Moreover, aggregation of proteins using high salt media provides

PAGE 205

184 possibilities of mimicking encapsulated systems. Preliminary work with the aggregation of serum albumin and hemoglobin (Appendix H) showed drastic spectral changes when comparing preand post-aggrega tion. If the density and the size of the aggregates can be characterized, their spectr a can be quantified.

PAGE 206

185 References 1 Garcia-Rubio, L. H. and Ro, N. Detailed copolymer characteriza tion using ultraviolet spectroscopy. (1985) Can J Chem. 63 (1): 253-63. 2 Garcia-Rubio, L. H. Characte rization of proteins during a ggregation using turbidimetry. (1989) Chem Eng Commun. 80 : 193-210. 3 Elicabe, Guillermo E. and Garcia-Rubio, Luis H. The selection of the regularization parameter in inverse problems: estimation of particle size distribution from turbidimetry. (1988) Polymeric Materials Sc ience and Engineering. 59 : 165-8. 4 Brandolin, A., Garcia-Rubio, L. H., Provder, T., Koehler, M. E., and Kuo, C. Latex particle size distribution from turbid imetry using inversion techniques; experimental va lidation. (1990) Polymeric Materials Science and Engineering. 62 : 306-11. 5 Weissbluth M. Hypochromism. (1971) Quart Rev Biophys. 4 (1): 1-34. 6 Vekshin, N. L. Screening hypochromism in molecular aggregates and biopolymers. (1999) J Biol Phys. 25 (4): 339-354. 7 MacRae, Robert A., McClure, Joseph A., and Latimer, Paul. Spectral transmission and scattering properties of red blood cells. (1961) J. Opt. Soc. Am. 51 : 1366-72. 8 Bohren, C. F. and Huffman, D. R. Ab sorption and Scattering of Light by Small Particles. (1983) New York, NY: John Wiley and Sons, Inc. 9 Mattley, Y., Leparc, G., Potter, R., and Garc ia-Rubio, L. Light sca ttering and absorption model for the quantitative in terpretation of human blood platelet spectral data. (2000) Photochem Photobiol. 71 (5): 610-619. 10 Alupoaei, Catalina E., Olivares, Jose A., and Garcia-Rubio, Luis H. Quantitative spectroscopy analysis of prokaryotic cells: vegetativ e cells and spores. (2004) Biosensors & Bioelectronics. 19 (8): 893-903. 11 McKenzie, S. B. Textbook of hematology. (1996) Baltimore, MD: Williams and Wilkins.

PAGE 207

186 12 Alkire K. and Collingwood J. Physio logy of blood and bone marrow. (1990) Seminars in Oncology Nursing. 6 (2): 99-108. 13 Alberts. B, Bray, D, Lewis, J, Raff, M, Roberts, K, and Watson, J. D. Molecular biology of the cell. (1989) New Yo rk, NY: Garland Publishing, Inc. 14 Beutler, E., Lichtman, M. A., Coller, B. S., and Kipps, T. J. Williams Hematology, Sixth ed. (2001) New York, NY: McGraw-Hill, Inc. 15 Harmening, D. M. Clinical hematol ogy and fundamentals of hemostasis. (1997) Philadelphia, PA: F. A. Davis Co. 16 Hoffman, R., Benz, Jr. E. J., Shattil, S. J ., Furie, B., and Cohen, H. J. Hematology: Basic principles and practice. (1991) New York, NY: Churchill Livingstone Inc. 17 Lee, G. R., Bithell, T. C., Foerster, J., Athens, J. W., a nd Lukens, J. N. Wintrobe's Clinical Hematology, Volume 1. (1992) Philadelphia, PA: Lea and Febiger. 18 Besa, E. C., Catalano, P. M., Kant, J. A., and Jefferies, L. C. Hematology. (1992) Malvern, PA: Harwal Publishing Company. 19 Dailey, J. F. Dailey's notes on blood. (1991) Somerville, MA: Medical Consulting Group. 20 Kragh-Hansen, Ulrich. Structure and lig and binding properties of human serum albumin. (1990) Dan. Med. Bull. 37 (1): 57-84. 21 Meloun, B., Moravek, L., and Kostka, V. Complete amini acid sequence of human serum albumin. (1975) FEBS Lett. 58 (1): 134-137. 22 Schreiber, G. Synthesis, pr ocessing and secretion of plas ma proteins by the liver and other organs and their regulation. In: Pu tnam, F. W. editor.The plasma proteins: Structure, function and genetic control. (1987) Orlando, FL: Academic Press, Inc. 23 Ueno, A., Hong, Y. M., Arakaki, N., and Ta keda, Y. Insulin-stimulating peptide from a tryptic digest of bovine serum albumin: Purification and characterization. (1985) J. Biochem. 98 : 269-278. 24 Narayanan, S. Aggregation a nd structural changes in biolog ical systems; An ultraviolet visible spectroscopic approach for analys is of blood cell aggr egation and protein conformation. Ph.D. dissertati on. (1999) Department of Chemistry, University of South Florida.

PAGE 208

187 25 Ness, P. M. and Stengle, J. M. Historical introduction. Editor: Surgenor, D. M.The Red Blood Cell. Second edition. (1974) New York, NY: Academic Press. 26 Hillman, R. S. and Finch, C. A.Red cell manual, 7th Ed. (1996) Philadelphia, PA: F. A. Davis Co. 27 Steck, T. L. The organization of protei ns in the human red blood cell membrane. A review. (1974) J Cell Biol. 62 (1): 1-19. 28 Voet, D.; Voet, J. G., and Pratt, C. W. Fundamentals of Biochemistry. Upgrade ed. (2002) New York, NY: John Wiley and Sons, Inc. 29 Winkelmann, J. C. and Forget, B. G. Er ythroid and nonerythroid spectrins. (1993) Blood. 81 (12): 3173-85. 30 Marchesi, V. T. Structure and function of the erythrocyte membrane skeleton. Editors: Kruckeberg, W. C., Eaton, J. W., Aste r, J., and Brewer, G. J. Erythrocyte membranes 3: Recent clinical and expe rimental advances. (1983) New York, NY: Alan R. Liss, Inc. 31 Agre, P., Smith, B. L., Saboori, A. M., and Asimos, A. The red cell membrane skeleton: A model with ge neral biological relevance but pathological significance for blood. Gunn, R. and Parker, J. C., Ed itors. Cell Physiology of Blood. (1988) New York, NY: Rockefeller University Press 32 Liu, S. C. and Derick, L. H. Molecular anatomy of the red blood cell membrane skeleton: structurefunc tion relationships. (1992) Semin Hematol. 29 (4): 231-43. 33 Peters, L. L. and Lux, S. E. Ankyrins: structure and function in normal cells and hereditary spherocytes. (1993) Semin Hematol. 30 (2): 85-118. 34 Brewer, George J. General red cell metao lism. Surgenor, D. M., Editor. The Red Blood Cell, 2nd Ed. (1974) 1: 387-433. 35 Weinstein, R. S. The mor phology of adult red cells. Surgenor, D. M., Editor. The Red Blood Cell, 2nd Ed. (1974) 1 : 213-268. 36 Voet, D. and Voet, J. G. Biochemistry: Bi omolecules, mechanisms of enzyme action, and metabolism. Vol. 1. Third ed. (2 004) New York, NY: John Wiley and Sons, Inc. 37 Antonini, Eraldo and Brunori, Mauriz io. Hemoglobin and Myoglobin in their Reactions with Ligands (Frontiers of Biology, Vol. 21). (1971) New York, NY: American Elsevier.

PAGE 209

188 38 Devlin, T. M. Textbook of Biochemistry with Clinical Correlations. Fourth ed. (1997) New York, NY: John Wiley and Sons, Inc. 39 Kaplan, L. A. and Pesce, A. J. Clinical Chemistry: Theory, Analysis, and Correlation. Third ed. (1996) St. Louis, MO: Mosby. 40 Stryer, L. Biochemistry, Fourth ed. (1995) New York, NY: W. H. Freeman. 41 Baldwin, Joyce and Chothia, Cyrus. Hemogl obin: the structural changes related to ligand binding and its allosteric mechanism. (1979) J Mol Biol. 129 (2): 175-220. 42 Zwart, A., Van Kampen, E. J., and Zijlstra, W. G. Results of routine determination of clinically significant hemoglobin deriva tives by multicomponent analysis. (1986) Clin Chem. 32 (6): 972-8. 43 Petragnani, Nicola, Nogueir a, Otilia C., and Raw, Is aias. Methemoglobin reduction through cytochrome b5. (1959) Nature. 184 (Suppl. No. 21): 1651. 44 Mrvos, R. The spectrophotometric meas urement of methemoglobin. (1997) Clin Lab Sci. 10 (3): 119-121. 45 Gordy, Edwin, Drabkin, David L., and Ma rsh, Julian B. Spectr ophotometric studies. XVI. Determination of the oxygen satura tion of blood by a simplified technique, applicable to standard equipment. (1957) J Biol Chem. 227 : 285-99. 46 International Committee for Standardization in Haematology. Recommendations for haemoglobinometry in human blood. (1967) Br J Haematol. 13 : 71-5. 47 International Committee for Standardization in Haematology. Recommendations for reference method for haemoglobinometry in human blood (ICSH standard EP 6/2: 1977) and specifications for internati onal haemiglobincyanide reference preparation (ICSH standard EP 6/3: 1977). (1978) J Clin Pathol. 31 (2): 139-43. 48 Senozan, N. M. and Devore, J. A. Ca rbon monoxide poisoning: some surprising aspects of the equilibrium between he moglobin, carbon monoxide, and oxygen. (1996) J Chem Ed. 73 (8): 767-770. 49 Cox C J, Habermann T M, Payne B A, Klee G G, and Pierre R V. Evaluation of the Coulter Counter model S-Plus IV. (1985) Am J Clin Pathol. 84 (3): 297-306. 50 Mayer K, Chin B, and Baisley A. Evaluation of the S-Plus IV. (1985) Am J Clin Pathol. 83 (1): 40-6. 51 Wintrobe, M. M. The size and hemoglobin content of the erythrocyte. Methods of determination and clinical application. (1932) J Lab Clin Med. 17 : 899-912.

PAGE 210

189 52 Narayanan, S., Orton, S., Leparc, G. F., Garcia-Rubio, L. H., and Potter, R. L. Ultraviolet and visible li ght spectrophotometric a pproach to blood typing: objective analysis by aggl utination index. (1999) Transfusion. 39 (10): 10511059. 53 Narayanan, S., Galloway, L., Nonoyama, A ., Leparc, G. F., Garcia-Rubio, L. H., and Potter, R. L. UV-visible spectrophotometric approach to blood typing II: Phenotyping of subtype A2 and weak D and whole blood analysis. (2002) Transfusion. 42 (5): 619-626. 54 Lane, T. A.; Anderson, K. C.; Goodnough, L. T.; Durtz, S.; Moroff, Gary; Pisciotto, Patricia T.; Sayers, Merlin, and Silberst ein, Leslie E. Leukocyte Reduction in Blood Component Therapy. (1992) Ann Int Med. 117: 151-162. 55 Raftos, J. E., Stewart, I. M., and Lovric, V. A. Supernatant hem oglobin determinations after prolonged blood storage. (1986) Pathology. 18 (1): 123-6. 56 Tinoco, I. Jr. Hypochromism in polynucleotides. (1960) J Am Chem Soc. 82 : 4785-90. 57 Heiz, C., Raedler, U., and Luisi, P. L. Spectroscopy and Recognition Chemistry of Micelles from Monoalkyl Phos phoryl Nucleosides. (1998) J Phys Chem B. 102 (44): 8686-8691. 58 Latimer, P. The deconvulation of absorpti on spectra of green pl ant material-improved corrections for the sieve effect. (1983) Photochem Photobiol. 38 (6): 731-734. 59 Silverstein, R. M., Bassler, G. C., and Mo rrill, T. C. Spectrometric Identification of Organic Compounds, Fifth Ed. (1991) Ne w York, NY: John Wiley and Sons, Inc. pp. 289-315. 60 Christian, G. D. Analytic al Chemistry, Sixth Ed. ( 2004) Hoboken, NJ: John Wiley and Sons, Inc.: 457-521. 61 http://micro.magnet.fsu.edu/primer/java/electromagnetic/ 62 Garcia-Rubio, L. H. Refractive index effects on the absorption spectra of macromolecules. (1992) Macromolecules. 25 (10): 2608-13. 63 Tinoco, I., Sauer, K., Wang, J. C., and Puglis i, J. D. Physical Chemistry: Principles and Applications in Biological Systems. (2002) Upper Saddle River, NJ.: Prentice Hall, Inc. 64 Levine, Rodney L. and Federici, M. Marc ia. Quantitation of aromatic residues in proteins: model compounds for se cond-derivative spectroscopy. (1982) Biochem. 21 (11): 2600-6.

PAGE 211

190 65 Ichikawa, Tetsuo and Terada, Hiroshi. Second derivative spec trophotometry as an effective tool for examining phenylal anine residues in proteins. (1977) Biochim Biophys Acta. 494 (1): 267-70. 66 Wetlaufer, D. B. Ultrav iolet spectra of proteins and amino acids. (1962) Advan. Protein Chem. 17 : 303-90 67 Creighton, T. E. Proteins: Structures and Molecular Pr operties, Second Ed. (1993) New York, NY: W. H. Freeman and Company. 68 Latimer, P., Brunsting, A., Pyle, B. E., and Moore, C. Effects of asphericity on single particle scattering. (1978) Appl Opt. 17 (19): 3152-3158. 69 Personal communication with Pr of. Luis H. Garcia-Rubio, College of Marine Science, University of South Florida, St. Petersburg, FL. 70 Reynolds, L. O. Optical Diffuse Reflectance and Transmittance from an Anisotropically Scattering Finite Blood Medium. (1957) Ph.D. Dissertation, University of Washington. 71 Kerker, M. The scattering of light and ot her electromagnetic ra diation. (1969) New York, NY: Academic Press. 72 van de Hulst, H. C. Light Scattering by Small Particles. (1981) New York, NY.: Dover Publications, Inc. 73 Garcia-Rubio, L. H. Refractive index effects on the absorption spectra of macromolecules. (1992) Macromolecules. 25 (10): 2608-13. 74 Garcia-Rubio, L. H., LopezMenacho, C. A., and Grossman, S. Characterization of proteins during aggregation. II. Us e of model molecules for spectroscopy analysis. (1993) Chem Eng Commun. 122 : 85-101. 75 Alupoaei, C. E. Modeling of the transm ission spectra of microorganisms. M.S. Thesis. (2001) Department of Chemi cal Engineering, University of South Florida. 76 Yaroslavsky, A. N., Priezzhev, A. V., Rodri guez, J., Yaroslavsky, I. V., and Battarbee, H. Optics of blood. Tuchin, V. V., Ed itor. Handbook of Optical Biomedical Diagnostics. (2002) Bellingham, WA : SPIE Press. pp. 169-216. 77 Jay, Alfred W. L. and Canham, Peter B. Sedimentation of single human red blood cells. Differences between normal a nd glutaraldehyde fi xed cells. (1972) J Cell Physiol. 80 (3): 367-72.

PAGE 212

191 78 Smallwood R H, Tindale W B, and Trow bridge E A. The physics of red cell sedimentation. Physics Med Biol. 30 (2): 125-37. 79 Steinke, J. M. and Shepherd, A. P. Compar ison of Mie theory and the light scattering of red blood cells. (1988) Appl Opt. 27 (19): 4027-4033. 80 Hammer, M., Schweitzer, D., Michel, B., Th amm, E., and Kolb, A. Single scattering by red blood cells. (1998) Appl Opt. 37 (31): 7410-7418. 81 Tycko, D. H., Metz, M. H., Epstein, E. A., and Grinbaum, A. Flow-cytometric light scattering measurement of red blood cell volume and hemoglobin concentration. (1985) Appl Opt. 24 (9): 1355-65. 82 Latimer, P. The influence of photometer design on optical-conformational changes. (1975) J Theor Biol. 51 (1): 1-12. 83 UV Atlas of Organic Compounds, Volume II. (1966) New York: Plenum Press. 84 Latimer, P. The deconvulation of absorpti on spectra of green pl ant material-improved corrections for the sieve effect. (1983) Photochem Photobiol. 38 (6): 731-734. 85 Duysens, L. N. M. Flattening of the abso rption spectrum of suspensions, as compared to that of solutions. (1956) Biochim Biophys Acta. 19 : 1-12. 86 Tinoco, I. Jr. Hypochromism in polynucleotides. (1960) J Am Chem Soc. 82 : 4785-90. 87 Vekshin, N. L. Screening hypochrom ism of biological macromolecules and suspensions. (1989) J Photochem Photobiol, B: Biology. 3 (4): 625-30. 88 Lawley, P. D. Interaction studies with deoxyribonucleic acid. III. Effect of changes in sodium-ion concentration, pH, and temp erature on the ultraviolet absorption spectrum of sodium thymonucleate. (1956) Biochim Biophys Acta. 21 : 481-8. 89 Vekshin, N. L. Screening hypochromis m of chromophores in macromolecular biostructures. (1999) Biofizika. 44 (1): 45-55. 90 Vekshin, N. L. Screening hypochr omism stacked chromophores. (1987) Opt Spectrosk. 63 (3): 517-19. 91 Bateman, J. B., Hsu, S. S., Knudsen, J. P., and Yudowitch, K. L. Hemoglobin spacing in erythrocytes. (1953) Arch Biochem Biophys. 45 : 411-22. 92 Garcia-Rubio, L. H. The effect of the molecular size on the absorption spectra of macromolecules. (1987) Macromolecules. 20 (12): 3070-5.

PAGE 213

192 93 Kramer, Kurt, Elam, James O., Saxton, Geo. A., Elam, Wm. N. Jr., and Heob, Dorothy. Influence of oxygen saturation, erythrocyt e concentration, and optical depth upon the red and near-infrared light tr ansmittance of whole blood. (1951) Am J Physiol. 165 : 229-46. 94 Anderson, N. M. and Sekelj, P. Ligh t-absorbing and scattering properties of nonhaemolysed blood. (1967) Phys Med Biol. 12 (2): 173-84. 95 Latimer, P. The influence of photometer design on optical-conformational changes. (1975) J Theoret Biol. 51 (1): 1-12. 96 Borovoi, A. G., Naats, E. I., and Oppen, U. G. Scattering of light by a red blood cell. (1998) J Biomed Opt. 3 (3): 364-372. 97 Amesz, J., Duysens, L. N. M., and Brandt, D. C. Methods for measuring and correcting the absorption spectrum of sca ttering suspensions. (1961) J Theoret Biol. 1 : 5974. 98 Leach, S. J. and Scheraga, H. A. Effect of light scattering on ultraviolet difference spectra. (1960) J Am Chem Soc. 82 : 4790-2. 99 Whitaker, T. R., Sartiano, G. P., Hamelly, L. J. Jr., Scott, W. L., and Glew, R. H. Hypotonic exchange loading of human eryt hrocytes with iron-59-labeled rabbit hemoglobin. (1974) J Lab Clin Med. 84 (6): 879-88. 100 Marsden, N. V. B. and Ostling, S. G. Accu mulation of dextran in human red cells after hemolysis. (1959) Nature. 184 (Suppl. No. 10): 723-4. 101 Sartiano, G. P. and Hayes, R. L. Hypo tonic exchange-loading of erythrocytes. II. introduction of hemoglobins S a nd C into normal red cells. (1977) J Lab Clin Med. 89 (1): 30-40. 102 Ihler, G. M., Glew, R. H., and Schnure, F. W. Enzyme loading of erythrocytes. (1973) Proc Natl Acad Sci U S A. 70 (9): 2663-6. 103 Baker, R. F. Entry of ferritin into huma n red cells during hypot onic hemolysis. (1967) Nature. 215 (5099): 424-5. 104 Seeman, P. Transient hole s in the erythrocyte membra ne during hypotonic hemolysis and stable holes in the membrane afte r lysis by saponin and lysolecithin. (1967) J Cell Biol. 32 (1): 55-70. 105 Agilent 8453 UV-visible Spectroscopy System : Operator’s Manual. (2002) Agilent Technologies.

PAGE 214

193 106 Hematology Analyzer 9110+ Operator’s Manual. Serono-Baker, PA. 107 Jones, O. T., Earnest, J. P., and McNamee, M. G. Solubilization and reconstitution of membrane proteins. Findlay, J. B. C. and Evans, W. H. Editors. Biological Membranes: A Practical Approach. ( 1987) Washington DC: IRL Press Ltd. pp. 139-177. 108 Norbert P. Zemankiewicz (1995), Univ ersity of South Florida, Tampa. 109 Personal communication with Dr. Yvette Mattley, Tampa, FL. 110 High concentrations of mannitol and dext rose maximize preservation. Brochure. Baxter Healthcare Corp, Fenwal Division, Deerfield, IL. 111 Prihoda, L. A., Kelley, A., and Shah, C. Comparison of leukocyte reduction filters for red blood cells. (1997) Transfus. 37 (Suppl): 17S. 112 Lane T A, Anderson K C, Goodnough L T, Ku rtz S, Moroff G, Pisciotto P T, Sayers M, and Silberstein L E. Leukocyte reduc tion in blood component therapy. (1992) Ann Int Med. 117 (2): 151-62. 113 Bodemann, H. and Passow, H. Factors cont rolling the resealing of the membrane of human erythrocyte ghosts afte r hypotonic hemolysis. (1972) J Membr Biol. 8 (1): 1-26. 114 Hoffman J. F. Physiological characteri stics of human red blood cell ghosts. (1958) J. Gen. Physiol. 42 (1): 9-28. 115 Latimer, P. and Eubanks, C. A. H. Absorption spectrophotometry of turbid suspensions: a method of correcting for large systematic distortions. (1962) Arch Biochem Biophys. 98 : 274-85. 116 Thormaehlen, I., Straub, J., and Grigu ll, U. Refractive index of water and its dependence on wavelength, temperature, and density. (1985) Journal of Physical and Chemical Reference Data. 14 (4): 933-45. 117 Mattley, Y. D. An investigation of th e spectroscopic propertie s of platelets during activation and storage: Implementation of a new interpretation model. Ph.D. Dissertation (2000) Departme nt of Chemistry, Univer sity of South Florida. 118 Narayanan, S. Aggregation and structur al changes in biological systems: An ultraviolet visible spectroscopic approach for analysis of blood cell aggregation and protein conformation. Ph.D. Dissert ation (1999) Department of Chemistry, University of South Florida.

PAGE 215

194 119 Elicabe, G. E. and Garcia-Rubio, L. H. Latex particle size distribution from turbidimetry using inve rsion techniques. (1989) J Colloid Interf Sci. 129 (1): 192200. 120 Elicabe, G. E. and Garcia-Rubio, L. H. Latex particle size distribution from turbidimetric measurements combining regularization and generalized crossvalidation techniques. (1990) Adv Chem Ser. 227 : 83-104. 121 Golub, G. H., Heath, M., and Wahba, G. Generalized cross-vali dation as a method for choosing a good Ridge parameter. (1979) Technometrics. 21 (2): 215-223. 122 Personal communication with Dr. German Leparc, Florida Blood Services, Tampa, FL. 123 Personal communication with Dr. Debbie Hu ffman, College of Marine Science, University of South Florida, St. Petersburg, FL. 124 Coulter Z Series particle count and size analysers. User Manual #9914591-C. (1997) Beckman Coulter Inc., CA. 125 Personal communication with A licia Garcia-Lopez. (2004). 126 New, R. R. C. Liposomes: A Practical Approach. (1990) New York: IRL Press. 127 Monnard, P.-A., Oberholzer, T., and Luisi, P. Entrapment of nucleic acids in liposomes. (1997) Biochim Biophys Acta. 1329 (1): 39-50. 128 Djordjevich, L. and Miller, I. F. Synt hetic erythrocytes from lipid encapsulated hemoglobin. (1980) Exp Hemat. 8 (5): 584-92. 129 Hunt C. A. and Burnette R R. Li pid microencapsulation of hemoglobin. Appl Biochem Biotech. 10 : 147-9. 130 Domokos, G., Jopski, B., and Schmidt, K. H. Preparation, pr operties and biological function of liposome-encapsu lated hemoglobin. (1992) Biomat Art Cells Immob Biotech. 20 (2-4): 345-54. 131 Mobed, M., Nishiya, T., and Chang, T. M. S. Purification and characterization of liposomes encapsulating hemoglobin as potential blood substitutes. (1992) Biomat Art Cells Immob Biotech. 20 (1): 53-70. 132 Dimitriadis, G. J. Entrapment of ribonucleic acids in liposomes. (1978) FEBS Letters. 86 (2): 289-93.

PAGE 216

195 133 Brochu, H., Polidori, A., Pucci, B., and Vermette, P. Drug delivery systems using immobilized intact liposomes: A comp arative and critic al review. (2004) Current Drug Delivery. 1 (3): 299-312. 134 Ohki, N., Kimura, T., and Ogata, Y. The reduction of methemoglobin in Neo Red Cell. (1998) Art Cells Blood Subs Immob Biotech. 26 (5 & 6): 477-485. 135 Szebeni, J., Di Iorio, E. E., Hauser, H ., and Winterhalter, K. H. Encapsulation of hemoglobin in phospholipid liposomes: ch aracterization and st ability. (1985) Biochemistry. 24 (12): 2827-32. 136 Shew, R. L. and Deamer, D. W. A novel method for encapsulation of macromolecules in liposomes. (1985) Biochim Biophys Acta. 816 (1): 1-8. 137 Liu L and Yonetani T. Pr eparation and characterization of liposome-encapsulated haemoglobin by a freeze-thaw method. J Microencapsulation. 11 (4): 409-21. 138 Gregoriadis, G. Liposome Technol ogy: Liposome Preparation and Related Techniques. Vol. 1. (1993) Boca Raton: CRC Press. 139 Schurtenberger, P. and Hauser, H. Char acterization of the size distribution of unilamellar vesicles by gel filtration, qua si-elastic light scattering and electron microscopy. (1984) Biochim Biophys Acta. 778 (3): 470-80. 140 Ostrowsky, N. Liposome size measurem ents by photon correlation spectroscopy. (1993) Chemistry and Physics of Lipids. 64 (1-3): 45-56. 141 Deutscher, M. P. Methods in Enzymology, Vol 182: Guide to Pr otein Purification. (1990) New York: Academic Press, Inc. 142 Scopes, R. K. Protein Puri fication: Principles and Pract ice, Third Ed. (1994) New York: Springer-Verlag Inc.

PAGE 217

196 Appendices

PAGE 218

197 Appendix A: Mie Theory Formulae It has been established that the Mie efficiency coefficient8,71,72 for extinction (Qext) is related to the optical density with the turbidity equation revisited below, dD D f m Q D N dD D f m Q D Nabs p sca p) ( )) ( ( 4 ) ( )) ( ( 42 0 2 0 where Qext = Qsca + Qabs thus the turbidity equation is divided into its scattering and absorption components. Other important parame ters of the equation include the particle number (Np), the equivalent sphere diameter (D), the size distribution function (f(d)), the size parameter ( ) and the complex refractive index (m). The latter two parameters were defined in Chapter 3 as D ) ( ) ( ) ( ) ( ) (0 0 n i n n m Qsca and Qext are defined as } { ) 1 2 ( 22 2 1 2 n n n scab a n Q )} {Re( ) 1 2 ( 21 2n n n extb a n Q where Qabs can be inferred from these two know values. Mie scattering coefficients, an and bn are ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( m m m m m m an n n n n n n n n

PAGE 219

198 Appendix A (continued) ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( m m m m m m bn n n n n n n n n n(z), ’n(z), n(z), and ’n(z) are Riccati-Bessel functions defined in terms of Bessel functions (Jn), ) ( ) 1 ( ) ( 2 ) (2 1 1 2 1 2 1z J i z J z zn n n n n z zn n ) ( ) ( '1 ) ( 2 ) (2 1 2 1z J z zn n n z zn n ) ( ) ( '1

PAGE 220

199Appendix B: Concentration-Related Instru mental Limitations of Hemoglobin and Erythrocytes It is important to understand and to operate within the limitations of the instrument (Agilent 8453 UV-vis spectrophoto meter) when analyzing data. Two major concerns that required attenti on were both related to the conc entration of the erythrocytes in the suspension: linearity of the instrument and multiple scattering. The former can be discussed in terms of the ab sorbance equation, A = log (I0/It). If the recovered transmitted light is only 5% of the incident lig ht intensity due to absorption, the value for A would be 1.3. At 1% recovery, A would be 2. Thus as absorption increases, the sensitivity of the detector is tested due to the decreasing light tr ansmission. We have chosen a conservative 1.2 optical density units as our upper limit to ensure the reliability of our data. Figure B1 demons trates the linearity of the optical density measurements as a function of concentration for a hemoglobi n solution taken with a 1 cm pathlength cuvette. The small peak at 540 nm stays below an OD of one and thus maintains good linearity over the observed concentrations. The hemoglobin Soret band at 417 nm keeps linearity at higher OD and it can be inferred by the plot that the linearity breaks down between 2.5 and 3.0 OD. Therefore, it can be seen that our self-i mposed limit of 1.2 OD is well within the linear limits and will gi ve reliable data. Due to the high extinction coefficient of hemoglobin, the 417 nm band r eaches and OD of 1.2 at a concentration of ~1.3 mg/ml and the 540 nm band at a concentrat ion of ~14 mg/ml. These values are not remotely close to physiological encapsulated hemoglobin concentrations of 330 mg/ml. Thus it is easy to see the difficulty in exam ining the effect of molecular hypochromism in a free hemoglobin solution approaching physiologi cal concentrations. For suspensions of

PAGE 221

200 Appendix B (continued) red cells, it was determined experimental ly that a concentration of 4000 cells/ l gives an OD of ~1.2. 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 00.20.40.60.8112 Concentration (mg/ml)OD 540 nm 417 nm Figure B1: Assessment of the optical density linearity limits of the Agilent 8453 spectrophotometer as a function of hemoglobi n (solution) concentr ation. The 540 nm band maintains linearity due to its low optical density over the observed concentrations. The 417 nm band, however, loses its linearity between OD of 2.5 and 3.0. Multiple scattering effects also need to be considered in a particle suspension. The Mie theory predicts light scattered by a sing le particle. If the cells are concentrated enough to scatter light by multiple particles, the task of in terpreting the spectra becomes complex. Figure B2 examines if there are any multiple scattering effects at the working sample dilution of 4000 cells/ l with serial dilutions of w hole blood. The presence of

PAGE 222

201 Appendix B (continued) multiple scattering would manifest itself by deviating from linearity as the cell concentration increases. According to our observations, the measurement at approximately 4000 cells/ l shows negligible deviations at wavelengths of 417 and 540 nm. Thus it can be concluded that multiple scattering effects are not a factor in our simulations and interpretations of spectra based on the Mie theory. 0 0.2 0.4 0.6 0.8 1 1.2 1.4 050010001500200025003000350040004500 RBC (cells/microliter)OD 417 nm 540 nm Linear (540 nm) Linear (417 nm) Figure B2: Serial dilutions of whole blood for the examination of multiple scattering. The linearity of the spectrum at wavele ngths of 417 and 540 nm imply negligible multiple scattering e ffects at 4000 cells/ l.

PAGE 223

202Appendix C: Spectral Characterizati on of a Liposome Model System Multiwavelength spectra of modified red blood cells were successfully characterized and interpreted in this dissertation. As an alternate experimental model system parallel to the red cells, liposomes were used to encapsulate hemoglobin (hemosomes) and other model proteins. The advantages to a good liposome model system are that they are easy to manipulat e, a variety of model molecules can be encapsulated within the membrane, and thei r spherical nature fits well in the Mie framework. Liposomes are lipid vesicles that encl ose an aqueous volume. When lipids are introduced to aqueous media at the appropriate concentrations, they form spontaneously into vesicles with the membrane in a bilayer structure similar to t hose of biological cells (Figure C1).126 Liposomes are typically made with common phospholipids such as phosphatidylcholine, phosphatidylserine and sphingomyelin in combination with cholesterol (~ 1:1 molar rati o of lipid:cholesterol). Th e physical structure of the liposomes membranes can be unilamellar or multilamellar (multiple concentric membranes within the same liposome) de pending on the manner in which they are prepared.126,127 Moreover, sizes of the liposomes can vary in the range of 0.02 – 20 m in diameter.126,128 Liposome systems have differe nt clinical and experimental applications such as artificial red blood cell substitutes,128,129,130,131 nucleic acid carriers,127,132 drug delivery systems,133 and systems observing enzymatic activity under encapsulated conditions.134

PAGE 224

203 Appendix C (continued) Figure C1: Diagram of a liposome. Like bi ological phospholipids bilayers, the membranes adopt tail-to-tail configurati on and encapsulate an aqueous volume. The objective of this research was to perform a multiwavelength UV-vis spectral analysis/chararacterization of liposomes encapsulating hemoglobin (hemosomes) to parallel our studies of the m odified red cells. Given that the properties of the hemosomes are accurately elucidat ed, it should be possible to interprete the spec tra of the hemosomes, considering the success the modified red ce ll characterization. The liposomes vary greatly in size, however this should not pose a problem with the interpretation since there are no size restrictions with the Mie theory Moreover, liposomes encapsulating albumin were examined due to the accessibility of the model protein. Materials and Methods: For the preparation of the protein-en capsulated liposomes, a commonly described thin film method was used.128,135 Lyophilized egg yolk phosphatidylcholine (PC) (Sigma-Aldrich, MO) was reconstituted to a stock solution of

PAGE 225

204 Appendix C (continued) 100 mg/ml with chloroform. Using an appr oximate molecular weight of the PC (MW ~768 Da) given by the manufacturer, a 1:1 mola r ration of PC:cholesterol was calculated. In a 250 ml round-bottom flask, 200 ml of PC stock solution and 10 mg of cholesterol were measured and dissolved in approximately 3 ml of chloroform. The flask was fitted to a rotary evaporator (Buchi Rotovapor R-3000 with a Buchi 461 water bath) and subjected to a vacuum with the wa ter bath set to approximately 55oC, slightly lower than the boiling point of chloroform (61.5oC). The flask was rotated until the chloroform was completely evaporated and a thin layer of lipid/cholesterol was distributed evenly on the bottom of the flask. The proteins encapsulated were bovine serum albumin (Sigma-Aldrich MO) (typically ~ 5% concentration was used), a nd hemoglobin harvested from red blood cells (Figure C2). For the hemoglobin, whole bl ood samples were obtained from the Florida Blood Services (St. Petersburg, FL). The whole blood sample was washed by centrifugation in three cycles with isotonic PB S as previously described in Chapter 4. The washed cell sample was then lysed by a fr eeze-thaw method. The cells were frozen in a -80oC freezer for approximately 10 minutes, then immediately thawed in a 56oC water bath. The process was repeated three times until inversion of the sample in its holding tube no longer showed any turbidity of intact cells. The sample was then centrifuged at 100,000 x g in a Beckman Optima TL ultr acentrifuge with a TLA-100.4 rotor to rid the suspension of cell debris. The resulting hemoglobin solution was assayed for concentration using the Drabkin’s assay (Chapter 4).

PAGE 226

205 Appendix C (continued) Next, the hemoglobin soluti on diluted to a desired c oncentration was measured (~10 ml) into the round-bottom flask containi ng the thin lipid/choles terol layer and the mixture was sonicated for approximately 15 minutes at a frequency of 47 KHz (Branson 5120 sonicator) (Figure C3). The resulting liposomes were washed three times by centrifugation and resuspended with PBS to approximately 2 ml. A spectrum of the Obtain blood samplesCount on SeranoBaker Determine dilution Take whole blood spec Wash blood 4x by centrifugation Resuspendcells with PBS Freeze-thaw lyse4x (-80/57oC) Spec of lysedsuspension Spin down cell debris Determine concentration of Hb solution with Drabkin’sassay Dilute Hbsolution to desired concentration (5%)** Confirm concentration with Drabkin’sassay Figure C2: Flowchart of the method for hemoglobi n harvesting prior to the production of hemoglobin-encapsulated liposomes (hemosomes).

PAGE 227

206 Appendix C (continued) Lecithin/cholesterol in chloroform (1:1 molar ratio) Prepare thin lipid film in rotoevaporator (vacuum dry) Add prepared Hbsolution (~10 ml) Sonicate(15 min) Wash hemosomes Spec of crude hemosomes SmallMediumLarge 0.22 m0.45 m Figure C3: Protocol for the generation of hemoglobin-encapsulated liposomes (hemosomes). Other model proteins such as albumin can be substituted for hemoglobin. crude sample suspension was obtained. The suspension was then mechanically extruded with syringe filters (Whatman polyethersulfone disc filters for low protein adsorption) with pore sizes of 0.22 and 0.45 m to narrow the size range s of the liposomes for analysis. Spectra were obtained for each size range. Results and Discussion Figure C4 shows normalized spectra of the crude, filtered and free albumin spectra from a representative albumin lipos ome experiment. The spectrum of the free albumin solution shows a clearly defined Soret band at ~280 nm but in contrast, none of

PAGE 228

207 Appendix C (continued) the liposome spectra show this peak. This eff ect is a scattering-related effect comparable to the one seen with hemoglobi n and red cells. The crude sa mple, which represents the 0.0E+00 1.0E-03 2.0E-03 3.0E-03 4.0E-03 5.0E-03 6.0E-03 1902903904905906907908909901090 Wavelength (nm)Normalized OD Crude Liposomes < 0.22 um 0.22 0.45 um > 0.45 um Albumin Solution Figure C4: Normalized spectra of crude and filtered albumin liposomes compared to a spectrum of free albumin in solution. unfiltered liposomes, the spectrum seems to be dominated by the scattering properties of the larger liposomes, judging from the flattened spectrum. This conjecture is further supported by the similarities of the crude spec trum to the that of the filtered liposomes containing sizes above 0.45 m. The low (< 0.22 m) and medium (0.22 – 0.45 m)

PAGE 229

208 Appendix C (continued) range sizes of liposomes show spectra that ne arly overlap. The small differences in the two spectra reflect a trend resulting from size disparities, consistent with our results from red cells. That is, as the size increases, a region of the small wavelengths (< 300 nm) decreases and the spectral inte nsity increases at the larger wavelengths (> 300 nm). 0.0E+00 10E-03 20E-03 30E-03 40E-03 50E-03 60E-03 1902903904905906907908909901090 Wavelength (nm)Normalized OD Crude Hemosome < 0.22 um 0.22 0.45 um > 0.45 um Hemoglobin Solution Figure C5: Normalized spectra of crude and filtered hemosomes compared to a spectrum of free hemoglobin in solution. Figure C5 shows normalized spectra of crude and filtered hemosomes and free hemoglobin in solution. The hemosomes were prepared using a hemoglobin solution with a concentration of ~10%. The spectral characteristics of each of the samples are similar to those of the albumin-containing liposomes. The spectrum for the low size

PAGE 230

209 Appendix C (continued) range (< 0.22 m) showed a characteristic hemoglobin band at 417 nm. This could be an indication of encapsulated hemoglobin, or free hemoglobin in solution outside the liposomes. The latter case is more probable considering that if the hemoglobin were encapsulated, the 417 nm peak would be visible in the other filtered fractions to some degree as well. For future reference, an a dditional step of centr ifugation and inspection of the supernatant would verify this suspicion. Both the albumin liposomes and hemosome s showed similar spectral qualities and trends. Visual examination of the lipos omes using light microscopy showed large variations in particle size Additionally, it was noted that the liposomes were multilamellar to varying degrees and tended to ag gregate (Figure C6). The lamellarity of the vesicles would change the opt ical properties of the particles in the context of spectral analysis. It was therefore no surprise wh en the interpretation model did not give reasonable values for size and protein con centration for the albumin liposomes and the hemosomes. In order to achieve a good ch aracterization of the liposome system, a few issues must be considered: 1) lamellarity of the liposomes, 2) the particle size distribution, 3) how much of the protein was encapsulated. Recommendations for future work in this area should start with the refinement of the protocol to make liposomes. Different methods have been reported that control the characteristics of the liposomes during produc tion, such as lipid dispersion in water,136 freeze-thaw method137 and emulsification.138 Furthermore, corroboration of particle

PAGE 231

210 Appendix C (continued) characteristics require analytical methods, such as electron microscopy139 and photon correlation spectroscopy.140 Figure C6: Light microscope picture of albumin liposomes magnified to 400x. Many of the liposomes are multilamellar and tend to aggregate. Hemosome showed similar qualities.

PAGE 232

211Appendix D: RBC Swelling Under physiological conditions, red blood cells typically exist as biconcave disks. Changes in ionic strength can cause the cells to undergo morphological changes such as swelling, shrinking or crena ting. UV-visible spectrophotometry was used to examine whole blood in varying medium tonicities to observe how these changes are represented in the spectra. Although whol e blood was used, the red cells in particular act as the osmometers thus any changes in the spectra would reflect the cha nges in the red cells. Materials and Methods Whole blood samples were obtained from the Florida Blood Services, Tampa, FL and the sample was diluted to a concentration of ~4000 cells/ l as described in Chapter 4. Two dilutions were performed with the firs t being a 1:50 using 0.9% PBS. For the second dilution needed to achieve the final ce ll concentration, the io nic strengths of the media were varied. The different tonicities of media were obtained by diluting the 0.9% stock PBS accordingly to obtain concentrations of 0.7%, 0.5%, and 0.4%. When mixed with the blood cell sample, the final saline concentrations were calculated to be 0.899%, 0.707%, 0.515%, and 0.419%. OD spectra of th e mixture was obtained as previously described. Results and Discussion Figure D1 illustrates normalized spectra of red cells in the various tonicities of saline each after a 15 minute incubation time where decreasing the decreasing tonicity of

PAGE 233

212 Appendix D (continued) the PBS medium causes the cells to swell, and eventually burst. The differences between the 0.899% and 0.707% curves are subtle. The small changes are a result of the swelling of the cells in the 0.707% medium reflecting a slight increase in the volume (size) coupled with a small decrease in the hemoglobi n concentration. As the ionic strength of the medium is decreased to 0.515%, the slope of the linear portion of the spectrum (> 600 nm) changes, denoting larger a lterations in size. Also, the emergence of a peak at 417 nm denotes the lysis of a small population of cells and the presence of free hemoglobin in the medium. Saline with a tonicity of 0. 417% causes a more significant bursting of the 0.0E+00 10E-03 20E-03 30E-03 40E-03 50E-03 60E-03 1902903904905906907908909901090 Wavelength (nm)Normalized OD 0.899% 0.707% 0.515% 0.419% Saline tonicity Figure D1: Spectra of red cells in varying tonicities of PBS media. As the tonicity decreases, the cells swell, eventually bursting and releas ing hemoglobin into solution (0.419% curve).

PAGE 234

213 Appendix D (continued) cells. This is evident as the spectrum is dominated mainly by the hemoglobin absorption spectrum. There is indication of a populati on of intact cells because the spectrum is significantly elevated from the baseline across the entire wavelength range. Thus it can be concluded that subtle changes in the swelling of the red ce lls can be detected experimentally. This supports the swelling simulations shown in Chapter 5. Moreover, UV-visible spectroscopic measurements can dete ct lysis of cells. Further work should allow for the quantification of lysed cells, an application useful in clinical settings.

PAGE 235

214Appendix E: Spiking Red Cell Suspensions with Hemoglobin Within the scope of this study, it has b ecome evident that free hemoglobin in the medium of the red cell suspension is detectable in the multiwavelength spectrum. The spectral manifestation of the free hemoglobin in the suspension is most profound at the 417 nm Soret band. As a clini cal application, simple spec troscopic detection can prove useful for the detection of disorders such as hemolytic anemia14 and the quan tification of blood substitutes like polymerized hemoglobin.14 Herein, we show a preliminary study of free hemoglobin in a red cell suspension by ar tificial hemoglobin spiking and how it is spectrally represented. Materials and Methods Human hemoglobin A0 was obtained from Sigma (c at. # H-0267) and it was reconstituted in PBS for a final concentrati on of approximately 0.5 g/dl. A sample of purified red blood cells was diluted 1:50 (2.45 ml PBS, 0.05 ml ce ll suspension), then a second dilution was performed (2.95 ml PBS, 0.05 ml 1:50 suspension) to achieve an OD maximum of approximately 0.5. To the 3 ml suspension, 10 l of stock Hb solution was added and a spectrum was taken. A subseque nt spectrum was taken after the addition of another 10 l of stock Hb. The suspension was centrifuged for 10 minutes at 1500 x g and the supernatant was spectropho tometrically analyzed in an attempt to retrieve the 20 l of Hb added to the suspension. As a c ontrol, 10 ml of hem oglobin stock was added twice to 3 ml of PBS and a spectrum was take n at each step. The same procedure was

PAGE 236

215 Appendix E (continued) performed with a sample of resealed cel ls. The hemoglobin concentrations were estimated using 417 = 7786 cm2/g. Results Figure E1 shows the control samples of free hemoglobin in 3 ml of PBS. The addition of 10 l and 20 l of stock Hb to the PBS was quantified by the Drabkin’s assay to be 0.0178 mg/ml and 0.0365 mg/ml. Fi gure E2 illustrates spectra of purified red blood cells spiked with the same volumes of stock hemoglobin solution as the control. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1902903904905906907908909901090 Wavelength (nm)OD +20 ul Hb +10 ul Hb0.0365 mg/ml 0.0178 mg/ml Figure E1: Spectrum of the control PBS spiked w ith two volumes of free hemoglobin. The concentrations were estimated usi ng the extinction coefficient at 417 nm.

PAGE 237

216 Appendix E (continued) The increasing amounts of hemoglobin is represen ted with an increase in the intensity of the 417 nm peak. The absorption characterist ics of the free hemoglobin largely changes the spectrum in the absorption region < 600 nm. The scattering region >600 nm shows a slight change. Since this wavelength range is mostly affected by size and refractive index changes, perhaps there is some surface adso rption of the free hemoglobin on the outside of the cells, affecting the refractive index of th e particle to a small extent. Quantification of the Hb concentration in the supernatant ga ve a result of 0.0352 mg/ml, a value that is close to that of the control. 0 0.2 0.4 0.6 0.8 1 1.2 20030040050060070080090010001100 Wavelength (nm)OD resealed red blood cells +10 ul Hb (A1) +20 ul Hb (A2) supernatant0.0352 mg/ml Figure E2: Spectrum of a purified red cell suspensi on spiked with free hemoglobin. The presence of the free hemoglobin is represente d by the peak at 417. The spectrum of the supernatant of the spiked cell sample wa s quantified for concentration using the extinction coefficient at 417 nm.

PAGE 238

217 Appendix E (continued) Similar results are shown for a spiked resealed cell sample (Figure E3). The MCHC of the resealed sample was reporte d to be 0.215 mass fract ion (Serono-Baker). Once again, the 417 nm peak significantly in creased with increasing free hemoglobin. The supernatant was quantified to be 0.0348 mg/ml, a value that is extremely close to that of the spiking experiment of th e purified red cells above. Both values are slightly smaller than the control, a fact that supports the hypothesis of surface adsorption. 0 0.2 0.4 0.6 0.8 1 1.2 20030040050060070080090010001100 Wavelength (nm)OD red blood cells +10 ul Hb (B1) +20 ul Hb (B2) supernatant0.0348 mg/ml Figure E3: Spectrum of a resealed cell suspension spiked with free hemoglobin. The presence of the free hemoglobin is represente d by the peak at 417. The spectrum of the supernatant of the spiked cell sample wa s quantified for concentration using the extinction coefficient at 417 nm.

PAGE 239

218 Appendix E (continued) Attempts to quantify the free hemoglobin for these samples using RBCHb02a.exe interpretation model (which takes into account free Hb) is covered in Appendix F.

PAGE 240

219Appendix F: Detailed Examination of Experimental and Simulated Spectra of Resealed Cells Chapter 5 showed a comparison of data between measured and simulated spectra of resealed red blood cells. In this comp arison, the simulation model proved to be successful at reproducing features and trends of the spectrum with changing MCHC and MCV. This appendix further examines the effectiveness of the simulations in predicting the features and trends of measured spectra that are grouped into narrower MCHC ranges. 0.0E+00 5.0E-04 1.0E-03 1.5E-03 2.0E-03 2.5E-03 3.0E-03 3.5E-03 1902903904905906907908909901090 Wavelength (nm)Normalized OD Units 0.084/71.0 0.093/80.0 0.117/81.1 0.092/96.6 Hemoglobin MCHC/MCV Figure F1: Experimental data set of resealed red cells in the low MCHC range with varying MCV values. The data set is normalized in the 230-900 nm range using the area under the curve method. The MCHC is expre ssed in mass fractions and the MCV in fl. The free hemoglobin solution spectrum is included as a reference.

PAGE 241

220 Appendix F (continued) 0.0E+00 5.0E-04 1.0E-03 1.5E-03 2.0E-03 2.5E-03 3.0E-03 3.5E-03 1902903904905906907908909901090 Wavelength (nm)Normalized OD Units 0.084/71.0 0.093/80.0 0.117/81.1 0.092/96.6 Hemoglobin MCHC/MCV Figure F2: Simulated spectra of the low MCHC range data set normalized on a per cell basis. The MCHC is expressed in mass fractions and the MCV in fl. The free hemoglobin solution spectrum is included as a reference. Figures F1 and F2 are measured and simu lated spectra respectively of resealed red cells. Here, the theoretical model simulates the trends well. The sample with the highest MCHC (0.117 mass fraction) shows the lowest 417 nm hemoglobin band and the lowest MCHC (0.084) has the highest 417 nm peak in both figures. Although the samples vary in both MCHC and MCV, it has been shown in the simulations in Chapter 5 that the effect of MCHC changes is a more dominant one.

PAGE 242

221 Appendix F (continued) 0.0E+00 5.0E-04 1.0E-03 1.5E-03 2.0E-03 2.5E-03 1902903904905906907908909901090 Wavelength (nm)Normalized OD Units 0.238/86.2 0.220/88.0 0.232/91.2 0.229/93.2 Hemoglobin MCHC/MCV Figure F3: Experimental data set of resealed red cells in the medium MCHC range with varying MCV values. The data set is normalized in the 230-900 nm range using the area under the curve method. The MCHC is expre ssed in mass fractions and the MCV in fl. The free hemoglobin solution spectrum is included as a reference. Figures F3 and F4 represent measured and simulated data respectively for a medium range of MCHCs. The calculated spectr al features and trends in Figure F4 agree well with the experimental spectra from Figure F3. The lowest MCHC (0.220 mass fraction) exhibits the highest spectral intensity at the lower wavelengths (<600 nm), a trend that is consistent in bot h figures. There are small discre pancies in the features of

PAGE 243

222 Appendix F (continued) the measured and calculated spectra, however it must be noted that the simulation model does not take into account some details of the experimental red cell sample. For example, 0.0E+00 5.0E-04 1.0E-03 1.5E-03 2.0E-03 2.5E-03 1902903904905906907908909901090 Wavelength (nm)Normalized OD Units 0.238/86.2 0.220/88.0 0.232/91.2 0.229/93.2 Hemoglobin MCHC/MCV Figure F4: Simulated spectra of the medium MCHC range data set normalized on a per cell basis. The MCHC is expressed in ma ss fractions and the MCV in fl. The free hemoglobin solution spectrum is included as a reference. the spherical approximation may pose limits on the simulation of the measured data. Furthermore, when the cells reseal, we are assuming a homogeneous distribution of MCHC which is probably not the case. Desp ite such unanswered questions, the model

PAGE 244

223 Appendix F (continued) does a good job simulating spectra based on known parameters of the particle suspension (i.e. MCHC and MCV). 0.0E+00 2.0E-04 4.0E-04 6.0E-04 8.0E-04 1.0E-03 1.2E-03 1.4E-03 1.6E-03 1.8E-03 2.0E-03 1902903904905906907908909901090 Wavelength (nm)Normalized OD Units 0.336/89.9 0.333/92.2 0.329/96.1 0.332/97.8 0.320/101.7 Hemoglobin MCHC/MCV Figure F5: Experimental data set of resealed re d cells in the high MCHC range with varying MCV values. The data set is normalized in the 230-900 nm range using the area under the curve method. The MCHC is expre ssed in mass fractions and the MCV in fl. The free hemoglobin solution spectrum is included as a reference. Figures F5 and F6 show measured and simulated spectra respectively in the high MCHC range. Here, the simulated values show small changes in the spectra compared to the experimental spectra. This can be attrib uted to the above mentioned issues of the spherical approximation, or the assumpti on of hemoglobin homogeneity. At higher

PAGE 245

224 Appendix F (continued) hemoglobin concentrations, the refractive index of the particle increases, and scattering becomes prominent enough to exacerbate the eff ects of factors (such as shape) that may not have been a problem at lower re fractive indices of the particle. One peculiar sample in Figures F5 and F6 was the spectrum with an MCHC of 0.329. Upon close examination of experiment al notes, it was evident that the sample used was approximately 10 days old. This was in stark contrast to the rest of the samples (50+ samples) which were used within 24-48 hours post-donation. It has been 0.0E+00 2.0E-04 4.0E-04 6.0E-04 8.0E-04 1.0E-03 1.2E-03 1.4E-03 1.6E-03 1.8E-03 2.0E-03 1902903904905906907908909901090 Wavelength (nm)Normalized OD Units 0.336/89.9 0.333/92.2 0.329/96.1 0.332/97.8 0.320/101.7 Hemoglobin MCHC/MCV Figure F6: Simulated spectra of the high MCHC ra nge data set normalized on a per cell basis. The MCHC is expressed in mass fractions and the MCV in fl. The free hemoglobin solution spectrum is included as a reference.

PAGE 246

225 Appendix F (continued) demonstrated by another project member that th e aging of red cells in storage resulted in morphologic changes that translated into si gnificant alterations of spectral features (unpublished data).108 Over time, the morphology of th e cells turned into a spherical crenated form. An examination of th e sample under a light microscope at 100x magnification did indeed reveal a significan t population of crenated cells. Such alterations manifested in the spectra could e xplain the irregularity of this particular sample.

PAGE 247

226Appendix G: Alternate Versions of the Interpretation Model In Chapter 6, two alternate versions of the interpretation model were introduced: RBCHb01b and RBCHb02a. The kernels of bot h models are identical to that of RBCHb01a (the version used for the main anal ysis in this work), all functioning on the basis of the Mie theory-based turbidity calc ulations. Small altera tions in the program design allow RBCHb01b to account for the pr esence of two derivatives of hemoglobin (typically oxyhemoglobin and methemoglobin in the case of resealed cells), and RBCHb02a introduces the capability to quant ify any free hemoglobi n existing in the medium outside the cells. The main scheme of this investigation used an interpretation model that did not consider such details incl uded in the alternate versions because the simplified interpretation model generally yielded good results. However, when scrutinized on an individual basis, some samples were identified that were better suited to be interpreted using the alternate versions. So me examples are cited in this appendix. A comparison of the results of RBC Hb01a and RBCHb01b generally yielded similar results in the values of paramete rs such as cell number, cell size and cell hemoglobin concentration. However, ther e were a few select samples where the interpreted values improved with the us e of RBCHb01b, accounting for the presence of methemoglobin. The inputs for this version were similar to those of RBCHb01a (as described in Chapter 6) with the major di fference being the desi gnation of the optical properties file of a second hemoglobin deri vative (ophbfe3.01 file containing the complex refractive index of methemoglobin and the refrac tive index of water). In an example of a comparison of the two versions, Figure G1 shows a resealed cell with low MCHC

PAGE 248

227 Appendix G (continued) evaluated with RBCHb01a. The MCHC is es timated well, however, the cell counts and the MCV values do not agree well with the measured. The same data interpreted with RBCHb01b (incorporating methemoglobin) sh ows improved estimations of both counts and MCV (Figure G2). The sample cont aining a low MCHC could very well have elevated amounts of oxidized hemoglobin compared to those of higher MCHC considering the manner in which it is prepared After permeabilizi ng the cells by way of hypotonic shock, the -0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 1902903904905906907908909901090 Wavelength (nm)OD Measured Calculated Residual Beer of Chromo Estimated: MCHC = 0.045 Dn = 6.59 microns MCV = 150 fl count = 1.90e6 /ul RSSQN = 7.45 Measured (S-B): MCHC = 0.043 D = 5.67 microns MCV = 98.5 fl count = 2.85e6 /ul6399c Figure G1: Resealed cell data interp reted using RBCHb01a interp retation model. The MCV is estimated to be substantially higher than the measured value.

PAGE 249

228 Appendix G (continued) suspension is equilibrated for an extended period before restoration (see Chapter 4 for details). This subjects the oxyhemoglobin to a highly oxygenated medium increasing the chances of conversion to methemoglobin. -0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 1902903904905906907908909901090 Wavelength (nm)OD Measured Calculated Residual OxyHb MetHb Estimated: MCHC (oxyHb) = 0052 MCHC (metHb) = 0.0023 Dn = 5.45 microns MCV = 84.8 fl count = 2.86e6 /ul RSSQN = 7.36 Measured (S-B): MCHC = 0.043 D = 567 microns MCV = 98.5 fl count = 2.85e6 /ul Figure G2: Data of same resealed sample as Figure A1 interpreted with RBCHb01b, accounting for the presence of methemoglobin. The MCV and cell count estimations are closer to the measured values. Analysis of data using RBCHb02a is a wo rk in progress. This version of the interpretation model accounts fo r any free hemoglobin in the medium, outside of the red cells. The test sample is a suspension of purified red cells spiked with a known amount of free hemoglobin (as describe d in Appendix E). Figure G3 shows that the estimated fit of the spectrum is not accurate although the estimated values are relatively close to the

PAGE 250

229 Appendix G (continued) measured values. The calculated spectrum also does not reflect th e large amount of free hemoglobin with a peak at 417 nm. The data shown represents work that has not been completed, and further improvement of this model will provide us with the capability to detect small amounts of lysis in red cell suspensions and whole blood. 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1902903904905906907908909901090 Wavelength (nm)OD Measured Calculated Calculated Measured D (um) 5.57 5.52 MCV (fl) 90.4 88.0 MCHC (mass frac) 0. 306 0.333 counts ( /ul) 3.76e6 6.75e6 free Hb (mg/ml) 0.0312 0.0529 Figure G3: An RBCHb02a interpretation of a purifi ed red cell suspension spiked with free hemoglobin. The fit of the curve is not good although the estimated values compare well with the measured values.

PAGE 251

230Appendix H: UV-Visible Analys is of Molecular Aggregates The focus of the encapsulated systems men tioned in this work (resealed red cells, liposomes) was to achieve a spectral understa nding of the effects of changing chemical composition. A similar model system for th e optical characterization of macroscopic particles is molecular aggreg ates. For example, a comparison of hemoglobin in solution and hemoglobin in an aggregated state should s how drastic changes in the spectral profile. Light scattering characteristics of the aggr egated form should approach the optical properties of encapsulated systems containing similar concentrations of hemoglobin. Provided that there are good corroborative me thods of analyzing the physical properties of the aggregates, a reliable interpretation of the spectral data is sure to be achieved. Large molecules on the order of prot eins can be salted out at high salt concentrations by taking adva ntage of their hydrophobicity.141,142 At low salt concentrations, water molecules beco me ordered around hydrophobic patches on the protein, a thermodynamically unstable situation. With the addition of a salt, the hydrated ions of the salt disrupt the ordered wate r molecules, exposing the hydrophobic patches and allowing the proteins to aggregate. Th e concentration of salt required to salt out a protein depends on the amount of hydrophobic resi dues residing on the protein of interest. The most effective salts are ones containi ng multivalent anions and possess a high solubility, such as ammonium sulfate [(NH4)2SO4].142 Alternate preci pitation methods use water miscible organic solvents (ethanol actone) or a soluble high molecular weight organic polymer (polyethylene glycol) to decrease the solvation power of the water thereby aggregating the protein.

PAGE 252

231 Appendix H (continued) Our preliminary studies on protein a ggregation were performed on purified human hemoglobin A0 (Sigma-Aldrich) and bovine serum albumin (Sigma-Aldrich). The latter was chosen because it was readily available and it had been previously characterized spectrally. Ammonium sulfate was used as the precipitating agent and the concentrations (percent saturation) re quired to aggregate the proteins were experimentally determined. The spectral charac teristics of the proteins were compared in their dissolved and precipitated states for any detectable differences. Materials and Methods The buffer used in these experiments was a standard 0.2 M phosphate recipe.141 The ionic strength of the buffer is not much of an issue here as it is when preparing the hypotonic shock phosphate buffer (0.007 M) for th e resealing experiments. Briefly, 100 ml of 0.2 M solutions of monobasic sodi um phosphate and dibasic sodium phosphate were prepared separately. To obtain a final pH of 7.4, 19.0 ml of the monobasic and 81.0 ml of the dibasic solutions were combined and the final volume was br ought up to 200 ml. For the preparation of th e ammonium sulfate soluti ons, the % saturation was calculated by the equation 2 1 23 0 100 ) ( 533 S S S g where S1 and S2 are the initial and final saturation percentages and g is grams of ammonium sulfate to be added to 1 L of water or buffer at 20o C.142 Different saturation levels of the ammonium sulfate were test at 10% increments to determine the

PAGE 253

232 Appendix H (continued) concentration at which hemoglobin and album in aggregated. Incubation times were typically around 2 minutes. Th e concentration of the stock solution of the commercial hemoglobin sample was 20 mg/ml and the albumin stock was 10 mg/ml with both proteins being reconstituted with phosphate buffer. An example of concentration calculations for the hemoglobin sample is as follows (with albumin concentrations being calculated in the same manner). Here, the desired concentration for the sample to be measured was 0.1 mg/ml hemoglobin in 80% saturated ammonium sulfate. With the final sample volume to be 2 ml, 0.01 ml of the 20 mg/ml stock solution was added to 1.99 ml of 80% phosphate buffered ammonium sulfate solution (PB/AMS). If the PB/AMS was 80% sa turated at the time of addition, the final saturation would be 79.6%, which is a neg ligible difference for our purposes. Results For both albumin and hemoglobin, it was determined that 90% ammonium sulfate saturation was needed to maximize precipita tion of the proteins. Figure H1 shows a precipitation experiment of a 0.2 mg/ml albumin solution. The dashed line represents the albumin solution prior to aggreg ation. The spectrum shows a characteristic peak at 280 nm with a low baseline at th e wavelengths > 300 nm indicat ing a lack of macroscopic scattering. When the albumin is subject ed to a 90% AMS solution, there is an appreciable elevation in the entire wavele ngth range suggesting the formation of

PAGE 254

233 Appendix H (continued) aggregates. Centrifugation (15,000 x g ) to isolate the supernatant shows no significant presence of free albumin as evident by the lack of a 280 nm peak. 0 0.5 1 1.5 2 2.5 200300400500600700800 Wavelength (nm)OD albumin albumin in PB/AMS supernatant Figure H1: Ammonium sulfate precip itation of bovine serum albumin. There is a noticeable difference in the spectra between free albumin and aggregated albumin. The aggregated suspension shows considerable sca ttering with an elevated OD spectrum. The supernatant spectrum after cen trifugation shows no indica tion of an albumin peak, indicating the absence of a significa nt amount of non-aggretated albumin. Figure H2 shows a precipita tion of a 0.10 mg/ml solution of hemoglobin. Much like the encapsulated form of hemoglobin, the aggregated form exhibits significant scattering elements marked by the elevation of the OD spectrum. The preliminary data indicates that an aggregated system can be used as another model system to support the results of the resealed cells. Contrary to th e liposomes, there is no need to worry about

PAGE 255

234 Appendix H (continued) multiple membrane layers changing the refractive index of the aggregates, since there is no membrane. Hence it seems that the inte rpretation model could be successfully implemented as it was with the resealed cells. It will however be necessary to corroborate the size distribution and hem oglobin content of the aggregates. 0 0.2 0.4 0.6 0.8 1 1.2 1.4 200300400500600700800 Wavelength (nm)OD Hb Hb in PB/AMS Figure H2: Ammonium sulfate prec ipitation of 0.10 mg/ml hemoglobin. Upon aggregation, the spectrum show s larger scattering elements.

PAGE 256

About the Author Akihisa Nonoyama received his Bachelor’s degree in Chemistry from Case Western Reserve University in Cleveland, OH. During his undergraduate years, he gained laboratory experience doing resear ch in the Molecula r and Microbiology Department at the CWRU School of Medicine There, he learne d skills such as DNA mutagenesis and sequencing. He entered th e graduate program in Chemistry at the University of South Florida in 1995 where he began work on a multi-disciplinary project characterizing red blood cells using multiw avelength UV-visible spectroscopy. During his tenure as a graduate student, he held a position as a Teaching Assistant for General Chemistry, Organic Chemsitry, and Biochemistry laboratories.