close

Вход

Забыли?

вход по аккаунту

?

Rapid Assessment of Biological and Environmental Samples in Resource-Limited Settings Using Microwave-Accelerated Bioassay Technique

код для вставкиСкачать
ABSTRACT
Title
of
Dissertation:
RAPID ASSESSMENT OF BIOLOGICAL AND
ENVIRONMENTAL SAMPLES IN RESOURCELIMITED
SETTINGS
USING
MICROWAVEACCELERATED BIOASSAY TECHNIQUE
Enock Bonyi, Doctor of Philosophy, December 2017
Dissertation Advisor:
Kadir Aslan, Ph.D.
Department of Civil Engineering
Detection and quantification of analytes that are related to human health in
biological and environmental samples have been the primary focus of scientists,
health
providers,
and
allied
policy-makers
for
several
decades.
Most
commercialized bioassays for the detection of pathogens or pollutants are carried
out in controlled environments, hence, take long time to get results, and in most
cases, have low sensitivities. In addition, current instrumentation used in
pathogen detection is expensive to acquire and to maintain, require trained and
certified personnel.
In this work, we report the preparation and stability studies of next
generation of bioassay platforms by surface modification of circular, poly(methyl
methacrylate) (PMMA) discs, paper and polyethylene terephthalate (PET) with
nanoparticles (silver and indium tin oxide, ITO). The use of these bioassay
platforms in combination with microwave heating (i.e., microwave-accelerated
bioassays) have improved bioassay sensitivity, nanoparticle film stability, and
reduced the colorimetric bioassay time for HRP-2 assay (used in malaria
diagnosis) and colorimetric MC-LR bioassay (used in MC-LR toxin diagnosis)
from 2 hours and 90 minutes at room temperature to 15 minutes under
microwave heating. In addition, we have developed a MATLAB algorithm, and
window and mobile-based applications, which can convert colored images of
bioassays into numerical (pixel) values for rapid assessment of biological and
environmental samples in resource-limited settings.
RAPID ASSESSMENT OF BIOLOGICAL AND ENVIRONMENTAL SAMPLES IN
RESOURCE-LIMITED
SETTINGS
USING
MICROWAVE-ACCELERATED
BIOASSAY TECHNIQUE
by
Enock Bonyi
A Dissertation Submitted in Partial Fulfillment
of the Requirements for the Degree
Doctor of Philosophy
MORGAN STATE UNIVERSITY
December 2017
ProQuest Number: 10690773
All rights reserved
INFORMATION TO ALL USERS
The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.
ProQuest 10690773
Published by ProQuest LLC (2018 ). Copyright of the Dissertation is held by the Author.
All rights reserved.
This work is protected against unauthorized copying under Title 17, United States Code
Microform Edition © ProQuest LLC.
ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, MI 48106 - 1346
RAPID ASSESSMENT OF BIOLOGICAL AND ENVIRONMENTAL SAMPLES IN
RESOURCE-LIMITED SETTINGS USING MICROWAVE-ACCELERATED
BIOASSAY TECHNIQUE
by
Enock Bonyi
has been approved
December 2017
DISSERTATION COMMITTEE APPROVAL:
______________________________________
Kadir Aslan, Ph.D., Committee Chair
______________________________________
Dereje Seifu, Ph.D., Committee Member
______________________________________
Timothy Akers, Ph.D., Committee Member
______________________________________
Birol Ozturk, Ph.D., Committee Member
______________________________________
Mahmudur Rahman, Ph.D.¸ Committee Member
ii
DEDICATION
I dedicate this work to my parents (Newton Bonyi and Hellen Bonyi),
beloved spouse (Ruth Nyamongo), sons (Caleb and Lincoln), brother (Keffah
Bonyi), Sister (Maureen Bonyi), late uncle (Nelson Mosomi) and the Aslan
Research Group (ARG) family.
iii
ACKNOWLEDGEMENT
First and foremost, I thank God for the gift of life and protection for the
entire period of my dissertation research.
I earnestly express my heartfelt appreciation to my mentor, Dr. Kadir
Aslan, who through his wisdom, grace, and great insight, has steered me through
my dissertation work. His patience and decisiveness can never be overlooked.
Thank you, Dr. Aslan! Also, sincere thank you to my dissertation committee
members: Dr. Dereje Seifu, Dr. Timothy Akers, Dr. Mahmudur Rahman, and Dr.
Birol Ozturk who have provided progressive critique to my work and spent time to
read and offered suggestions on the best ways to improve on the subject matter.
I cannot forget Morgan State University, School of graduate studies who
extended me a three-year tuition scholarship award.
Subsequently, I am grateful to Mr. Edward Constance, Ms. Zeenat Kukoyi,
Ms. Oluseyi Daodu, Mr. Frank Coleman, and Mr. Oreoluwa Adesina who in a
significant way committed their quality time to help me throughout my dissertation
research and who have persistently offered encouragement when things seemed
bleak.
Lastly, I am grateful to my wife (Ruth) and sons (Caleb and Lincoln) who
have not only been a source of inspiration for me to push hard and promised of a
better day each day but who have also accepted my limited time at home.
iv
TABLE OF CONTENTS
DEDICATION ...................................................................................................... iii
ACKNOWLEDGEMENT ..................................................................................... iv
TABLE OF CONTENTS ....................................................................................... v
LIST OF FIGURES .............................................................................................. ix
LIST OF PUBLICATIONS ................................................................................. xxi
ABBREVIATIONS ........................................................................................... xxiii
CHAPTER I .......................................................................................................... 1
INTRODUCTION .................................................................................................. 1
CHAPTER II ......................................................................................................... 7
LITERATURE REVIEW ........................................................................................ 7
2.1. Conventional ‘Gold Standard’ Methods for Biological and Environmental
Sample Analysis ........................................................................................ 7
2.2. Rapid Diagnostic Tests (RDTs).................................................................. 9
2.3. Point-of-Care Testing (POCT).................................................................. 11
2.4.Recent Developments in Hand-held Devices in Diagnostics and
Environmental Monitoring ........................................................................ 13
2.5. Enzyme-Nanoparticle Hybrid Structures for Enhanced Enzymatic Activity
................................................................................................................. 15
2.6. Metal-Assisted Microwave-Accelerated Fluorescence (MAMEF)............. 17
2.7. Microwave-Accelerated Bioassays (MABs).............................................. 19
CHAPTER III ...................................................................................................... 24
MATERIALS AND METHODS ........................................................................... 24
3.1. MATERIALS................................................................................................ 24
3.1.1.
Chemicals and Reagents ................................................................... 24
3.1.2.
Proteins and Antibodies ..................................................................... 25
3.2. Instrumentation ......................................................................................... 25
v
3.3. METHODS .................................................................................................. 27
3.3.1. AIM 1: TO DEVELOP THE NEXT GENERATION OF CIRCULAR
BIOASSAY PLATFORMS FOR RAPID DETECTION OF TARGET
ANALYTES USING MICROWAVE-ACCELERATED BIOASSAYS ......... 27
3.3.1.1 Modification of PMMA by a Chemical Method. .................................... 27
3.3.1.2. Deposition of Silver Thin Films (STFs) on Modified PMMA Surfaces. 27
3.3.1.3. Generation and Attachment of ITO onto Silicone Isolators. ............... 29
3.3.1.4. Incorporation of Silver Nanowires on Paper. ...................................... 29
3.3.1.5. Surface Analysis. ............................................................................... 29
3.3.1.6. Assessment of Nanofilm Stability on Next Generation Platforms using
b-BSA Protein Assay. .............................................................................. 30
3.3.1.7. The performance of b-BSA Assay on Next Generation Platforms at
Room Temperature and Low-Power Microwave Heating......................... 30
3.3.1.8. Characterization of Platform Surfaces after b-BSA Protein Assays. .. 33
3.3.1.9. Application of Next Generation Assay Platforms in Real-Life Assays:
Rapid Detection of Ki-67 by Colorimetric Method. ................................... 33
3.3.2 AIM 2: TO DEVELOP SOFTWARE TO CONVERT COLORIMETRIC
AND CHEMILUMINESCENCE RESPONSES INTO NUMERICAL
VALUES FOR THE QUANTIFICATION OF TARGET ANALYTES. ......... 35
3.3.2.1 Development of Novel Software for Quantification of Target Analytes.
................................................................................................................. 35
3.3.2.2. Novel Software using MATLAB .......................................................... 35
3.3.2.3. Novel Software using C++ and C# Programming .............................. 36
3.3.2.4. Graphical User Interface .................................................................... 37
3.3.3. AIM 3: TO DEMONSTRATE THE USE OF CIRCULAR BIOASSAY
PLATFORMS AND THE NOVEL SOFTWARE FOR RAPID
ASSESSMENT OF BIOLOGICAL AND ENVIRONMENTAL SAMPLES IN
A RESOURCE LIMITED SETTING.......................................................... 39
3.3.3.1. Institutional Review Board (IRB) Approval. ........................................ 39
3.3.3.2. Collection of Blood Samples. ............................................................. 39
3.3.3.3. Processing and Storage of Serum. .................................................... 40
3.3.3.4. Comparison in Detection of Analytes from Biological and
Environmental Samples using Gold Standard Methods and Novel
Software................................................................................................... 40
3.3.3.4.1 Evaluation of Performance of our Software using Model Protein
Assay ....................................................................................................... 40
3.3.3.4.2. Dilution of Monoclonal Antibodies against HRP-2 and MC-LR in
Buffer (PBS), Serum, and Blood .............................................................. 41
vi
3.3.3.4.3. Preparation of HRP-2 and MC-LR Antigens.................................... 42
3.3.3.4.4. Indirect and Competitive Bioassay for HRP-2 and MC-LR on ITO
using Our Software .................................................................................. 42
3.3.3.4.5. Indirect and Competitive Assay for HRP-2 and MC-LR using
Commercial Immunoassay Kits and UV-Vis Spectrophotometer. ............ 43
3.4. Statistical Analysis ................................................................................... 43
CHAPTER 4 ....................................................................................................... 46
RESULTS AND DISCUSSION ........................................................................... 46
4.1 Development of next generation circular bioassay platforms for rapid
detection of target analytes using microwave-accelerated bioassays. ..... 48
4.1.1. Surface modification of PMMA and paper. ............................................ 49
4.1.2. Colorimetric response from a model bioassay for b-BSA on planar
platforms. ................................................................................................. 50
4.1.3. Colorimetric response of real life bioassay on 10 nm STFs and ITO on
PET. ......................................................................................................... 62
4.1.4. Evaluation of physical stability of the STFs and ITO on PET during reallife bioassay. ............................................................................................ 67
4.2 Development of software to convert colorimetric, and chemiluminescence
responses into numerical values for the quantification of target analytes. 73
4.3 Demonstration of the use of circular bioassay platforms and the software
for the rapid assessment of biological and environmental fields in
resource-limited setting ............................................................................ 94
4.3.1 Application of developed software in detection of biological samples .... 96
4.3.2 Application of the developed software in the detection of environmental
samples ................................................................................................. 146
CHAPTER 5 ..................................................................................................... 163
CONCLUSION AND FUTURE WORK ............................................................. 163
5.1 The development of the next generation of circular bioassay platforms for
rapid detection of target analytes using microwave-accelerated bioassays
………………………………………………………………………………..163
5.2 Development of software to convert colorimetric responses into numerical
values for the quantification of target analytes ....................................... 164
5.3 The potential application of circular bioassay platforms and the software
for the rapid assessment of biological and environmental samples in
resource-limited settings. ....................................................................... 165
5.4 FUTURE WORK .................................................................................... 167
vii
REFERENCES ................................................................................................. 168
APPENDIX I: Features of MATLAB-based Software ........................................A
APPENDIX II: Use of Novel Diagnostic Software (MATLAB version) with a
Real-Life Assay (HRP-2 Assay) .........................................................................C
APPENDIX III: Procedure for Window-based Application ...............................K
APPENDIX IV: Procedure for Smartphone-based Application .......................N
APPENDIX V: iCrystal System .......................................................................... O
viii
LIST OF FIGURES
Figure 1.Overview of the specific aims of this study. ........................................... 6
Figure 2. Immobilization of enzyme to Silver Island Films using biotin-avidin
protein assay. ..................................................................................................... 17
Figure 3. AFM image of silver island film on a glass substrate and the intensity of
fluorescence glass substrate (control experiment, No SiFs) and glass substrate
impregnated with silver island films (SiFs). ......................................................... 18
Figure 4. 3D plots of Acridian fluorescence emission vs. time vs. wavelength for
glass substrate and silvered substrates, coated with BSA-Biotin and Streptavidin,
exposed to low power microwave. ...................................................................... 19
Figure 5. Depiction of the detection of STX and GFAP proteins using 21-well
circular iCrystal platforms and colorimetric and fluorescence detection methods.
........................................................................................................................... 22
Figure 6. Colorimetric response for GFAP and STX using MAB technique [47] 23
Figure 7: Schematic depiction of surface modification of PMMA by chemical
method. .............................................................................................................. 28
Figure 8: Assessment of physical stability of next generation assay platforms .. 32
Figure 9: Schematic depiction of a real-life bioassay Ki-67 carried out on indium
tin oxide (ITO) on PET, and chemically modified PMMA with 10 nm STFs. All
control experiments were performed using low power MW heating and at RT
Total bioassay time: MW = 30 mins .................................................................... 34
Figure 10: Graphical user interface and software framework ............................ 38
ix
Figure 11: Indirect bioassay for HRP-2 assays on ITO at RT and MW heating
using blood, serum, and buffer as diluents. ........................................................ 44
Figure 12: Competitive bioassay for microcystin-LR (MC-LR) in creek water and
buffer on ITO at room temperature and under microwave heating. .................... 45
Figure 13: Water contact angles for chemically modified PMMA surfaces and
paper. (A). Scanning electron microscopy (SEM) images for (B) unmodified or
blank PMMA (C) chemically modified PMMA, (D) blank paper and (E) silver
nanowires functionalized on paper. .................................................................... 51
Figure 14: Colorimetric response for b-BSA bioassay carried out on (A)
chemically modified PMMA with STFs of different thickness (1 nm, 5 nm and 10
nm), (B) paper functionalized with AgNWs and (C) ITO using low power MW
heating and RT, a control experiment. Total bioassay time = 15 mins ............... 54
Figure 15: Colorimetric response for b-BSA bioassay at low-power MW heating
and at RT (a control experiment). ....................................................................... 55
Figure 16: Comparison of b-BSA bioassay performance using low power MW
heating on three platforms: chemically modified PMMA sputtered with 10 nm
STFs (black circles), indium tin oxide (ITO) (white circles), and silver nanowires
(SNW) functionalized on paper (triangles). (A) Colorimetric response and (B)
real-color pictures of the platforms. .................................................................... 56
Figure 17: Optical absorbance spectrum of (A) 1 nm STFs, (B) 5 nm STFs, and
(C)10 nm STFs before (initial) and after (final) the completion of Ki 67 bioassay
using low-power MW heating.............................................................................. 59
x
Figure 18: Optical absorbance spectrum of (A) 1 nm STFs, (B) 5 nm STFs, and
10 nm STFs before (initial) and after (final) the completion of Ki 67 bioassay at
room temperature (RT) ....................................................................................... 60
Figure 19: Optical absorbance spectrum of ITO platforms before, during and
after model protein bioassay was carried out using (A) microwave (MW) heating
and (B) at room temperature (RT). Concentrations of b-BSA in each well are as
follows: 1: 10-6 M, 2: 10-7 M, 3: 10-8 M, 4: 10-9 M, 5: 10-10 M, 6: 10-11 M, B: No bBSA. Each experiment was repeated three times and average values were
presented. .......................................................................................................... 61
Figure 20: Colorimetric response for Ki 67 bioassay on (A) chemically modified
PMMA with 10 nm STFs, (B) ITO using low-power MW heating and RT, a control
experiment. ......................................................................................................... 64
Figure 21: Colorimetric response for Ki-67 on 10 nm STFs using low-power MW
heating and RT (a control experiment). .............................................................. 65
Figure 22: Colorimetric response for Ki-67 on ITO using low-power MW heating
and RT (a control experiment). ........................................................................... 66
Figure 23: Optical absorbance spectrum of (A) STFs and (B) ITO before and
after the completion of Ki 67 bioassay using low-power MW heating and RT (a
control experiment). ............................................................................................ 70
Figure 24: SEM images of PMMA platforms with 10 nm STFs before and after
the completion of a real-life bioassay, Ki 67 using low-power (MW) heating and
RT, a control experiment. Scale bar = 20 µm. .................................................... 71
xi
Figure 25: SEM images of ITO platforms before and after the completion of a
real-life bioassay, Ki 67 using low-power (MW) heating and RT, a control
experiment. ......................................................................................................... 72
Figure 26: The layout for the MATLAB application with four sections: Image
upload, Graphical and histogram output, Tabular output and Settings sections. 75
Figure 27: Computer generated profile for shades of yellow color and their
corresponding grayscale (A), graphical and tabular representation of the
computed pixels upper and lower panels, respectively (B) using MATLAB
software. ............................................................................................................. 76
Figure 28: Pixels across rows (A) and down the columns (B) for empty high
thoroughput screening (HTS) wells (Inset, A) lit by an LED light source in a dark
room (cartoon sketch, inset (B)). ........................................................................ 77
Figure 29: Pixels across rows (A) and down the columns (B) for empty high
thoroughput screening (HTS) wells lit by LED light source on a laboratory bench
with incident light (cartoon sketch, inset (B)). ..................................................... 78
Figure 30: Pixels across rows (A) and down the columns (B) for empty high
thoroughput screening (HTS) wells lit by LED light source in a black box (cartoon
sketch, inset (B))................................................................................................. 79
Figure 31: Distribution of pixels for samples in HTS wells generated once the
samples are exposed to direct sunlight (red solid outline) and under a shade
(blue solid outline) on either an LED white board or regular white paper, cartoon
sketch, shown on the left. ................................................................................... 81
xii
Figure 32: Normalized pixels for enzymatic product o-phenylenediamine
dihydrochloride (OPD) generated by varying volumes of OPD while maintaining
the enzyme, Streptavidin-Horseradish peroxidase (Strep-HRP) constant (left
panel) carried out in direct sunlight and under a shade. Real color images of
OPD product samples exposed to direct sunlight (black and green solid line
enclosure) and under a shade (red and blue solid line enclosure). .................... 83
Figure 33: Pixels for b-BSA assay in the presence (panel B) or absence (panel
A) of 10 nm silver thin films in room with incident light (cartoon sketch, upper left
side) and in a dark room. The samples were lit by LED light in both environments
(cartoon sketch, bottom left side) for Trial 1. ...................................................... 84
Figure 34: Pixels for b-BSA assay in the presence (panel B) or absence (panel
A) of 10 nm silver thin films in room with incident light (cartoon sketch, upper left
side) and in a dark room. The samples were lit by LED light in both environments
(cartoon sketch, bottom left side) for Trial 2. ...................................................... 85
Figure 35: Pixels for b-BSA assay in the presence (panel B) or absence (panel
A) of 10 nm silver thin films in room with incident light (cartoon sketch, upper left
side) and in a dark room. The samples were lit by LED light in both environments
(cartoon sketch, bottom left side) for Trial 3. ...................................................... 86
Figure 36: Pixels for b-BSA assay in the presence (panel B) or absence (panel
A) of 10 nm silver thin films in room with incident light (cartoon sketch, upper left
side) and in a dark room. The samples were lit by LED light in both environments
(cartoon sketch, bottom left side) for trial 1. ........................................................ 90
xiii
Figure 37: Pixels for b-BSA assay in the presence (panel B) or absence (panel
A) of 10 nm silver thin films in room with incident light (cartoon sketch, upper left
side) and in a dark room. The samples were lit by LED light in both environments
(cartoon sketch, bottom left side) for trial 2. ........................................................ 91
Figure 38: Pixels for b-BSA assay in the presence (panel B) or absence (panel
A) of 10 nm silver thin films in room with incident light (cartoon sketch, upper left
side) and in a dark room. The samples were lit by LED light in both environments
(cartoon sketch, bottom left side) for trial 3. ........................................................ 92
Figure 39: Pixels for b-BSA assay in the presence (panel B) or absence (panel
A) of 10 nm silver thin films in room with incident light (cartoon sketch, upper left
side) and in a dark room (cartoon sketch, bottom left side). The samples were lit
by LED light in both environments. ..................................................................... 93
Figure 40: Colorimetric response for HRP-2 assay in buffer on commercial HTS
wells (A) at room temperature and pixel values computed using the novel
diagnostic software (B, bottom panel). The experimental samples (black solid
enclosure, B top panel) and control samples are shown by different colors (B, top
panel); positive control (red solid enclosure), cut off (green solid enclosure), and
negative control (blue solid enclosure). .............................................................. 98
Figure 41: Colorimetric response for HRP-2 assay in buffer on modified ITO
platform (A) under low power microwave heating and grayscale pixel values
computed using the novel diagnostic software (B, bottom panel). The
experimental samples (black solid enclosure, B top panel) and the control
xiv
samples (B, top panel); positive control (red solid enclosure), cut off (green solid
enclosure), and negative control (blue solid enclosure). ................................... 101
Figure 42: Colorimetric response for HRP-2 assay in buffer on modified ITO
platform (A) under low power microwave heating and grayscale pixel values
computed using the novel diagnostic software (B, bottom panel). The
experimental samples (black solid enclosure, B top panel) and the control
samples (B, top panel); positive control (red solid enclosure), cut off (green solid
enclosure), and negative control (blue solid enclosure). ................................... 102
Figure 43: Colorimetric response for HRP-2 assay in buffer on modified ITO
platform (A) under low power microwave heating and grayscale pixel values
computed using the novel diagnostic software (B, bottom panel). The
experimental samples (black solid enclosure, B top panel) and the control
samples (B, top panel): positive control (red solid enclosure), cut off (green solid
enclosure), and negative control (blue solid enclosure). The substrate volume
was increased 3-fold......................................................................................... 103
Figure 44: Colorimetric response for HRP-2 assay in buffer on modified ITO
platform (A) at room temperature and grayscale pixel values computed using the
novel diagnostic software (B, bottom panel). The experimental samples (black
solid enclosure, B top panel) and the control samples (B, top panel): positive
control (red solid enclosure), cut off (green solid enclosure), and negative control
(blue solid enclosure). The substrate volume was increased 3-fold. ................ 106
Figure 45: Colorimetric response for HRP-2 assay in serum on modified ITO
platform (A) under low power microwave heating and grayscale pixel values
xv
computed using the novel diagnostic software (B, bottom panel). The
experimental samples (black solid enclosure, B top panel) and the control
samples (B, top panel): positive control (red solid enclosure), cut off (green solid
enclosure), and negative control (blue solid enclosure). The substrate volume
was increased 3-fold......................................................................................... 110
Figure 46: Colorimetric response for HRP-2 assay in serum on modified ITO
platform (A) at room temperature and grayscale pixel values computed using the
novel diagnostic software (B, bottom panel). The experimental samples (black
solid enclosure, B top panel) and the control samples (B, top panel); positive
control (red solid enclosure), cut off (green solid enclosure), and negative control
(blue solid enclosure). The substrate volume was increased 3-fold. ................ 111
Figure 47: Colorimetric response for HRP-2 assay in blood on modified ITO
platform (A) under low power microwave heating and grayscale pixel values
computed using the novel diagnostic software (B, bottom panel). The
experimental samples (black solid enclosure, B top panel) and the control
samples (B, top panel); positive control (red solid enclosure), cut off (green solid
enclosure), and negative control (blue solid enclosure).The substrate volume was
increased 3-fold. ............................................................................................... 113
Figure 48: Colorimetric response for HRP-2 assay in blood on modified ITO
platform (A) at room temperature and grayscale pixel values computed using the
novel diagnostic software (B, bottom panel). The experimental samples (black
solid enclosure, B top panel) and the control samples (B, top panel); positive
xvi
control (red solid enclosure), cut off (green solid enclosure), and negative control
(blue solid enclosure). The substrate volume was increased 3-fold. ................ 116
Figure 49: Real-picture images of the enzymatic product for HRP-2 assay in
blood, serum and buffer at room temperature and under low power microwave
heating. ............................................................................................................. 117
Figure 50: RGB format pixel output for random test samples on a 21-well iCrystal
Plate (A) and individual pixel intensity values for Green, Blue and Red colors (B).
......................................................................................................................... 119
Figure 51: RGB channel pixel intensities for HRP-2 assay images taken on 21well iCrystal plates. The HRP-2 assay in buffer was completed using a kitchen
microwave oven. ............................................................................................... 124
Figure 52 :Colorimetric response for HRP-2 assay (A), real color pictures (B) for
the enzymatic product inside HTS wells and their corresponding RGB channel
pixel intensities (D) and the grayscale pixel intensities (C). HRP-2 assay in buffer
(dynamic range 0.565-0.00565 mg/mL) was completed under microwave heating.
......................................................................................................................... 125
Figure 53: Colorimetric response for HRP-2 assay in buffer completed under
microwave heating (iCrystal system) (A) and grayscale (C) and blue pixel
computation (C) for the sample picture image (D) representing test samples (16), controls (7-9) and blank experiments. Legend: 1= 0.113, 2 = 0.0565, 3 =
0.0113, 4 = 0.00565, 5 = 0.00113, 6 = 0.000565 mg/mL. Controls: 7 = positive
control, 8 = cut-off, 9 = negative ....................................................................... 131
xvii
Figure 54: Colorimetric response for HRP-2 assay in serum completed under
microwave heating (iCrystal system) (A) and grayscale (C) and blue pixel
computation (C) for the sample picture image (D) representing test samples (16), controls (7-9) and blank experiments. Legend: 1= 0.113, 2 = 0.0565, 3 =
0.0113, 4 = 0.00565, 5 = 0.00113, 6 = 0.000565 mg/mL. Controls: 7 = positive
control, 8 = cut-off, 9 = negative control and B = blank sample. ....................... 132
Figure 55: Colorimetric response for HRP-2 assay in blood completed under
microwave heating (iCrystal system) (A) and grayscale (C) and blue pixel
computation (C) for the sample picture image (D) representing test samples (16), controls (7-9) and blank experiments. Legend: 1= 0.113, 2 = 0.0565, 3 =
0.0113, 4 = 0.00565, 5 = 0.00113, 6 = 0.000565 mg/mL. Controls: 7 = positive
control, 8 = cut-off, 9 = negative control and B = blank sample. ....................... 133
Figure 56: Comparison in red, green and blue channel pixel intensity
computation for sample images taken on iCrystal platform (Top) and HTS-wells
(Bottom) using window-based application. ....................................................... 137
Figure 57: Comparison in red, green and blue channel pixel intensity
computation for sample images taken on iCrystal platform (Top) and HTS-wells
(Bottom) using MATLAB application. ................................................................ 138
Figure 58: Comparison in blue pixel intensity computation for sample images
taken on iCrystal platform and HTS wells using window-based app (B and C) and
MATLAB software (D and E). ........................................................................... 144
xviii
Figure 59: Comparison in performance for window-based application and matlab
software to compute blue pixel intensity for sample images taken on iCrystal
platforms and HTS wells. .................................................................................. 145
Figure 60: Colorimetric responses for MC-LR assay in buffer of MC-LR
standards performed on HTS wells under microwave heating (total assay time =
15 mins) and analyzed using MATLAB software (A) and UV-spectrophotometer
(B and C). ......................................................................................................... 147
Figure 61: Colorimetric responses for MC-LR assay in buffer of MC-LR
standards performed on HTS wells at room temperature and analyzed using
MATLAB software (A) and UV-spectrophotometer (B and C). .......................... 150
Figure 62: Colorimetric responses for MC-LR assay in buffer and MSU creek
water of MC-LR standards and samples (S1 and S2) performed on iCrystal
plates at room temperature (A) and under microwave heating (B) and analyzed
using UV-spectrophotometer ............................................................................ 151
Figure 63: Pixel computation for MC-LR assay in buffer and MSU creek water of
MC-LR standards and samples (S1 and S2) performed on iCrystal plates at room
temperature (A) and under microwave heating (B) using MATLAB software. .. 155
Figure 64: Colorimetric responses and pixel computation for MC-LR assay in
buffer and MSU creek water of MC-LR standards and samples (S1 and S2)
performed on HTS wells under microwave heating (iCrystal system) using
MATLAB software. ............................................................................................ 156
Figure 65: Colorimetric responses and pixel computation for MC-LR assay in
buffer and MSU creek water of MC-LR standards and samples (S1 and S2)
xix
performed on HTS wells under microwave heating (iCrystal system) using
MATLAB software. The assay pictures were obtained using high resolution
camera and smartphone camera. ..................................................................... 157
Figure 66: Real color pictures (A), colorimetric response (B) and blue pixel
computation using iOS and MATLAB application (C) for HRP-2 assay in buffer
(concentration range: 0.113 mg/mL – 0.000113 mg/mL) performed under iCrystal
microwave system (Total assay time = 15 min). ............................................... 161
Figure 67: Real color pictures (A) and RGB histograms (B) for HRP-2 assay in
buffer (concentration range: 0.113 mg/mL – 0.000113 mg/mL) performed under
iCrystal microwave system (Total assay time = 15 min). .................................. 162
xx
LIST OF PUBLICATIONS
The following publication(s) that are directly related to this dissertation:
1. Bonyi, E., Kukoyi, Z., Daodu, O., Boone-Kukoyi, Z., Coskun, S.,
Unalan, H.E. and Aslan, K. "Metal oxide surfaces for enhanced
colorimetric response in bioassays." Colloids and Surfaces B:
Biointerfaces 154 (2017): 331-340.
The following publication(s) that are not directly related to this dissertation:
1. Constance, E.N., Zaakan, A., Alsharari, F., Gordon, B., Syed, F.,
Mauge-Lewis, K., Bonyi, E., Boone-Kukoyi, Z. and Aslan, K., "Effect of
Microwave Heating on the Crystallization of Glutathione Tripeptide on
Silver Nanoparticle Films." The Journal of Physical Chemistry C 121.10
(2017): 5585-5593.
2. Thompson, N., Boone-Kukoyi, Z., Shortt, R., Lansiquot, C., Kioko, B.,
Bonyi, E., Toker, S., Ozturk, B. and Aslan, K, "Decrystallization of
Crystals Using Gold “Nano-Bullets” and the Metal-Assisted and
Microwave-Accelerated Decrystallization Technique." Molecules 21.10
(2016): 1388.
3. Mauge-Lewis K, Gordon B, Syed F, Syed S, Bonyi, E, Mohammed M,
Toth EA, Seifu D, Aslan K. "Crystallization of Lysozyme on Metal
Surfaces Using a Monomode Microwave System." Nano Biomed. Eng
8.2 (2016): 60-71.
xxi
4. Barare B, Babahan I, Hijji YM, Bonyi E, Tadesse S, Aslan K. "A Highly
Selective Sensor for Cyanide in Organic Media and on Solid Surfaces."
Sensors 16.3 (2016): 271.
5. Bonyi, E., Onuk, Z., Constance, E., Boone-Kukoyi, Z., Gordon, B.,
Kioko, B., Daodu, O. and Aslan, K. "Metal-assisted and microwaveaccelerated
evaporative
crystallization:
an
approach
to
rapid
crystallization of biomolecules." CrystEngComm 18.30 (2016): 56005610.
6. Ettinoffe, Y.S., Kioko, B.M., Gordon, B.I., Thompson, N.A., Adebiyi, M.,
Mauge-Lewis, K., Ogundolie, T.O., Bonyi, E., Mohammed, M. and
Aslan,
K.
"Metal-Assisted
and
Microwave-Accelerated
Decrystallization." Nano Biomedicine & Engineering 7.4 (2015).
xxii
ABBREVIATIONS
HTS - High throughput screening
PMMA – Poly(methyl methacrylate)
PET - Polyethylene terephthalate
MABs - Microwave-accelerated bioassays
MAMEF - Microwave-accelerated metal-enhanced fluorescence
MA-SPCL - Microwave-accelerated surface plasmon-coupled luminescence
MT-MEC - Microwave-triggered metal-enhanced chemiluminescence
ECL - Electroluminescence
SNP - Single nucleotide polymorphism
MGB - Minor groove binding
MLST - Multiple-locus sequence typing
MLVA - Multiple-locus variable-number tandem repeat analysis
NGS - Next generation sequencing methods
LAMP - Loop-mediated isothermal amplification
RDT - Rapid diagnostic test
MRDD - Malaria rapid diagnostic devices
pLDH - Plasmodium lactate dehydrogenase
HRP-2 - Histidine rich protein 2
POCT - Point-of-care testing
FENO - Fractional exhaled nitric oxide
AFM – Atomic force microscopy
MATLAB - Matrix laboratory
xxiii
b-BSA- Biotinylated bovine serum albumin
ELISA - Enzyme-linked immunosorbent assays
GFAP - Glial fibrillary acid protein
STX - Shigella-like toxin
PCR - Polymerase chain reaction
APTES - Aminopropyltrimethoxysilane
ITO - Indium tin oxide
SNW – Silver nanowires
PBS - Phosphate buffered saline
OPD - o-phenylenediamine
BSA - Bovine serum albumin
Strep-HRP - Streptavidin-conjugated horseradish peroxidase
SEM – Scanning electron microscope
DSA - Drop shape analysis system
STFs - Silver thin films
LiAH – Lithium aluminum anhydride
EG - Ethylene glycol
PVP - Poly(vinylpyrrolidone)
DAP - Diaminophenazine
GUI – Graphic user interface
IRB - Institutional review board
MC-LR – Microcystin leucine (L) arginine (R)
SAMs - Self-assembled monolayers
xxiv
LLOD - Lowest level of detection
SPR - Surface plasmon resonance
EDS – Energy dispersive X-ray spectroscopy
xxv
CHAPTER I
INTRODUCTION
Detection and quantification of biological analytes and environmental
pollutants are carried out by medical scientists, health providers, and the
outcomes of these studies are utilized by policy-makers to affect the quality of
human life. Biological samples can be defined as a representative portion(s) of a
whole tissue, organ, or body of an individual, which include blood, urine, saliva,
hair, amniotic fluid, nails, and sputum. Categories of environmental samples are
soil, water, and air. For decades, numerous sample analysis methods and tools,
for example, blood cell analyzers, molecular methods (polymerase chain reaction
or PCR) are the developed conventional methods. Traditional methods are
culture based techniques used in microbiological analysis, biochemical
identification methods used as confirmatory tests for samples that test positive in
conventional methods (e.g., mass spectroscopy, gas chromatography, highperformance liquid chromatography, intact cell spectrometry), affinity based
methods that utilize ligands for specific analytes of interest in the quest of
providing precise disease diagnosis and analysis of environmental pollutants to
safeguard human life.
Even though the above mentioned established methods are well
established, some factors, such as, sensitivity, specificity, portability, rapidity,
repeatability, complexity, efficiency, and cost, still limit their use. For instance,
conventional methods take a long time to yield results (1-2 days), are labor
intensive and subject to confirmatory tests, such as, biochemical and molecular
1
techniques. Furthermore, most of these methods require large quantities of
sample, highly trained and certified personnel. Biochemical methods also suffer
from shortcomings similar to the conventional methods. Affinity-based methods
suffer only sensitivity and specificity deficiencies. The nucleic acid-based
detection methods are pricey in regards to equipment cost, installation, and
maintenance.
In the past three decades, there has been a renewed drive to develop new
technologies and tools that can overcome some of the limitations of the
instruments above and methods mentioned earlier. Many resources are being
used for enhanced portability, easy operability, affordability, and increased
sensitivity. In this regard, the emergence of nanotechnology has revolutionized
the field of biosensing. Numerous reports on this topic, partially summarized in
Chapter 2 of this dissertation, can be found in the literature. For the sake of
brevity, we mention the most relevant ones to the study described here. The
Aslan Research Group has demonstrated that metal nanoparticles, such as,
silver island films impregnated on traditional platforms (high throughput screening
(HTS) wells and glass substrates), improved bioassay sensitivities and
significantly reduced bioassay time when coupled with low power microwave
heating
[1,
2].
electroluminescence
Moreover,
(ECL),
other
technologies
microwave-accelerated
(immuno-magnetic
metal-enhanced
fluorescence (MAMEF), aptamer-magnetic bead ECL, biosensors) were also
shown to improve the sensitivity of bioassays. However, these new techniques
still suffer from lack of stability of the metal films and repeatability. Subsequently,
2
there is a continued need to stabilize the nanoparticle films on planar surfaces for
increased bioassay sensitivity, and automation in the analysis of analytes using
compact mobile computing devices, especially in resource-limited settings.
In this Ph.D. dissertation, we investigated the preparation of next
generation bioassay platforms by surface modification of circular poly(methyl
methacrylate) (PMMA) discs, paper, and polyethylene terephthalate (PET) with
non-magnetic nanoparticles (that is, silver and indium tin oxide, ITO). The use of
these bioassay platforms in combination with low power microwave heating (i.e.,
microwave-accelerated bioassays, MABs) and compact computing devices have
improved bioassay sensitivity, nanoparticle film stability, and reduced the assay
time from several hours to <10 minutes. Moreover, these bioassay platforms
afford for the multiplexed assessment of different biological and environmental
samples in a semi-quantitative manner in a resource-limited setting and more
advanced laboratories. Specifically, we have three clearly defined specific aims,
as listed below, and summarized in Figure 1:
Aim 1: To develop the next generation of circular bioassay platforms for
rapid detection of target analytes using microwave-accelerated bioassays.
Hypothesis: Surface modification of bioassay platforms stabilizes nanoparticle
thin films and leads to enhanced enzymatic signal in microwave-accelerated
bioassays.
a.
Chemical surface modification of the blank circular bioassay platforms to
create primary NH2 groups using lithium aluminum anhydride and 3aminopropyltrimethoxysilane.
3
b.
Deposition of metal/metal oxide thin films (silver thin films, silver
nanowires, indium tin oxide) on modified blank circular bioassay platforms for use
in microwave-accelerated bioassays.
c.
Evaluation of the effect of the physical stability of metal/metal oxide thin
films on enzyme activity in microwave-accelerated bioassays by optical
absorption spectroscopy.
Aim 2: To develop software to convert colorimetric and chemiluminescence
responses into numerical values for the quantification of target analytes.
Hypothesis: Software developed using MATLAB; C++ and Visual Studio and
installed in compact mobile computing devices converts colored pictures into
numerical pixel values in the range 0-255.
a.
Creation of an algorithmic framework using Matrix Laboratory (MATLAB),
C++, and Visual Studio that can convert colorimetric images into numerical
values and further process the data to give graphical traces.
b.
Creation of a graphical user interface using MATLAB Toolkit, and Visual
Studio to enable sample input and collection of information related to the sample.
c.
Comparison of the performance of the developed software with
competitors regarding low levels of detection (LLOD).
4
Aim 3: To demonstrate the use of the circular bioassay platforms and the
software for the rapid assessment of biological and environmental samples
in resource-limited settings.
Hypothesis: Novel software and microwave-accelerated bioassays affords
rapidity, sensitivity, affordability, and convenience in environmental sensing and
disease detection in low cost settings.
a.
Development of point-of-care kits comprised of circular bioassay
platforms, enzyme-modified antibody-detection solution (colorimetric) and wash
buffer solutions for microwave-accelerated bioassays for use in the detection of
analytes in a laboratory setting.
b.
Development of point-of-care kits, comprised of circular bioassay
platforms, enzyme-modified antibody-detection solution (colorimetric) and wash
buffer solutions for use in the detection of analytes in a laboratory setting without
power supply.
5
CHAPTER II
LITERATURE REVIEW
2.1.
Conventional
‘Gold
Standard’
Methods
for
Biological
and
Environmental Sample Analysis
A plethora of traditional methods has been developed over the past
decades to detect biological analytes and environmental contaminants. These
conventional
methods
include:
(i)
culture
based
techniques
used
in
microbiological analysis [3] (ii) biochemical identification methods used as
confirmatory tests for samples that test positive in conventional methods (e.g.,
gas chromatography and intact cell spectrometry) [4, 5], (iii) affinity based
methods that utilize ligands that specifically attach to analytes of interest (iv) flow
cytometry assay, (v) sandwich-based assay ELISA [6-9], (vi) nucleic acid-based
detection methods (e.g., PCR) (vii) locked nucleic acid (LNAs) probes, (viii)
single nucleotide polymorphism (SNPs), (ix) real-time assays + minor groove
binding (MGB) probes, (x) multiple-locus sequence typing (MLST) and multiplelocus variable-number tandem repeat analysis (MLVA), (xi) next generation
sequencing methods (NGS), and (xii) loop-mediated isothermal amplification
(LAMP)) [10-14].
Despite the current and widespread use of the well-known methods, some
inadequacies limit their efficacy in quantitative and qualitative detection of target
7
molecules from biological or environmental samples. For instance, Hobbs and
Seña observed that while culture based methods displayed improved sensitivity
over microscopy in the detection and testing of bacterial susceptibility in
Trichomonas vaginalis, significant setbacks, such as, requirement of an
incubator, trained personnel and long assay times (up to one week) to obtain
results [15]. Moreover, Barber et. al., have highlighted that in addition to the
requirement of trained and experienced personnel, microscopy technique, the
gold
standard
method
for
malaria
diagnosis,
showed
insufficiency
in
differentiating different malaria-causing Plasmodium species in endemic zones,
that is, plasmodium vivax, p.ovale, and p.malariae. Also, microscopy is labor
intensive, time-consuming and there are high chances of human error, which in
most cases leads to misdiagnosis [16].
Similarly, biochemical methods and nucleic acid-based detection methods
require seasoned and highly trained staff, which in most cases, must be certified.
Subsequently, some of these methods work best when the molecules of interest
are in moderately high concentrations. Moreover, several critical factors, such as,
the initial cost for equipment acquisition, installation, and high maintenance cost
make biochemical and nucleic acid analysis a burden to patients and
environmental agencies found in regions that are less endowed with resources.
Additionally,
most
of
these
methods
are
sophisticated
and
in
most
8
circumstances, require electricity to operate, which limits their use in field studies,
emergency situations, and in smaller laboratories found in resource-limited
settings.
Due to the challenges mentioned above, inadequacies, and limitations of
the conventional methods of analysis, the World Health Organization (WHO) has
acted upon its mission to increase the rate of disease detection and early
disease management across all populations with an intention to spur growth in
medical access and improve health outcomes. The international body has
recommended that the scientific community should innovate technologies that
can provide results, which are either similar or better to those offered by the
traditional methods. In that respect, among the desirable features suggested by
the WHO is for the new tools to be highly sensitive, rapid, specific, portable, easy
to perform, cost-effective, and smaller in size.
2.2. Rapid Diagnostic Tests (RDTs)
RDTs achieve qualitative detection of molecules of interest in biological
samples and yield results within 20 minutes. Most of the RDTs utilize the vertical
or lateral flow of aqueous samples on a hydrophilic substrate. Ideally, a patient
sample is applied to a sample pad together with appropriate reagents, and after
some time, the presence of specific bands in the test card window signifies the
9
presence or absence of tested disease. Typical samples for RDTs include saliva,
capillary blood, and urine. Most RDTs are used in patients to test for conditions
like diabetes, anemia, human immunodeficiency virus (HIV), malaria, and
pregnancy [17].
Numerous research papers have reported on the current use and benefits
of RDTs to date. Murray et. al. discussed the importance of malaria rapid
diagnostic devices (MRDD) and their reliability compared to available alternative
methods for malaria diagnosis. The authors have reported that recent trials
revealed wide variability in the diagnosis and detection of infections in endemic
and non-endemic countries. Moreover, the authors identified various aspects of
these devices, which need further improvement [18]. For instance, MRDDs were
unable to differentiate between the species that cause infections. The presence
of histidine-rich protein 2 (HRP-2) and Plasmodium lactate dehydrogenase
(pLDH) causes differences in the results as well as certain failures. The presence
of pLDH is known to distinguish parasite viability and useful in determining
therapeutic response, while the presence HRP-2 is known to limit the use of this
antigen as an indicator of therapeutic response. Limitations in the study design
cause the clinical application of MRDDs to be incapable of fulfilling the worldwide
expectations due to the limited study population and controlled conditions.
MRDDs can provide comprehensive health impact, but its practicality depends on
10
the production of reproducible, accessible, and acceptable assays with ease of
performance and interpretation, its stability when stored, and its ability to
differentiate species, all at an affordable price. Moreover, current methods of
malaria diagnosis also include laser desorption mass spectrometry, fluorescent
microscopy, flow cytometry, automated blood cell analyzers, molecular methods
and serology-antibody detection. Several other papers have proven that
diagnosis by clinical symptoms alone is highly unreliable and subjective.
However, light microscopy is labor-intensive, requires significant technical skills,
and can result in therapeutic delays.
2.3. Point-of-Care Testing (POCT)
POCT is a laboratory testing done at the point of patient care; away from
the central laboratory [19]. Other synonymous words that are used to describe
POCT include near-patient testing, ancillary testing, decentralized testing,
patient-focused testing, and alternate-site testing [20]. The most ancient and
shared examples of POCT are urine dipsticks and blood glucose testing [19].
Due to an increase in patient volume, reduced reimbursements, and the desire
for improved patient health care, POCT has grown exponentially. Its growth is not
bound only to catering to bedside patients in hospitals but also to testing nonblood samples, for example, strep testing, toxicology testing, and sexually
11
transmitted diseases [19]. In other words, the scope of use for POCTs has
expanded to include emergency departments, accident sites, home health care
setting, and nursing homes. Also, POCTs get used in day care facilities, law
enforcements agencies, ambulance services (vehicles and helicopter), doctors’
offices, and surveillance of disease at ports of entry at airports [20-22].
Mor and Waisman divided POCTs into two groups; critical care profile
tests (e.g., blood gasses, acid-base, electrolytes, and hematology) and patient
focused tests. The patient-focused tests include blood coagulation test, renal and
liver function test, cardiac injury markers, therapeutic drug levels, drugs of abuse
and ethanol levels, pregnancy-human chorionic gonadotropin, urine pregnancy
test, and fecal occult blood [20]. However, in a recent review paper, Jan et. al.
categorized POCT into three generations [17]. The first generation of POCT
involved affinity based structures and biomarkers such as antibodies and
antigens coupled with biochemical reactions. Under this category, the most
common test format includes lateral flow test, vertical flow test, manually read
dipsticks, manually read cartridge based strips, and automated reading.
Examples of instruments that utilize these technologies include glucometers and
hemoglobin meters [17]. The second generation of POC tests involves the
detection of more complex and less accessible biomolecules (biomarkers), which
include molecules such as nucleic acids and cell-surface markers [17]. In this
12
category, samples are held using test cartridges. Sample processing and
readouts are done using small machines plugged to display screens. The
technology takes advantage of the advances in microfluidics, microelectronics,
optical systems, and laboratory-on-a-chip, nucleic acid test based amplification
and detection techniques [17]. Third generation POC, which are on the horizon,
will afford simultaneous detection of multiple targets by the use of certain
biomarkers. Another peculiar feature of this category of POCs is their projected
use to transmit results to relevant recipients (i.e., patients, physicians,
environmental agencies) remotely through internet, wireless, and Bluetooth
connectivity. The benefits will include antiviral and antibiotic-resistance
screening, differential screening and home-based self-testing [17].
In the last two decades, POCT has improved patient health care services
and reduced turnaround time. Hence, patients are encouraged to do follow-ups
especially in TB and HIV testing. Therefore, this leads to saving of lives due to
POTC’s rapidity and infectious diseases are better controlled and managed
through self-testing (e.g., HIV) [23-25].
2.4. Recent Developments in Hand-held Devices in Diagnostics and
Environmental Monitoring
In 2010, researchers at the University College Dublin School of Medicine
and Medical Science evaluated whether the diagnostic accuracy of handheld
13
computing devices is comparable to that of monitors used in emergency
teleconsultation [26]. The study was prompted because handheld disease
diagnosis devices are becoming increasingly prevalent in modern society and
researchers have gone to great lengths to discuss the prominence of these
devices. The results from their study demonstrated that the handheld devices
that were investigated are comparable to second class monitors. Future research
is still essential to make a precautionary approach and to understand the
importance of both physical and clinical evaluation of new technologies or
radiological procedures. The overall study suggested that the devices are great
tools for radiological reporting and disease diagnosis. Furthermore, Nemiroski et.
al., have recently innovated a handheld analytical device that utilizes
electrochemical reactions in the detection of (i) glucose in the blood for personal
health, (ii) trace heavy metals (lead, cadmium, and zinc) in water for in-field
environmental monitoring, (iii) sodium in urine for clinical analysis, and (iv) a
malarial antigen (Plasmodium falciparum histidine-rich protein 2) for clinical
research, and dissemination of the test results to the ‘cloud’ thereby enabling the
access of the data by either physicians or agencies concerned [27].
Also, Kuo has invented and patented a new device equipped with a test
head, a handle having a capillary tube fitted with a sensor and liquid reagent
dispensing opening. The device is utilizable for the diagnosis of body fluids such
14
as saliva, body fluid, blood and vaginal fluid. Additionally, Yetisen et. al., from the
University of Cambridge, developed an algorithm that enables all smartphones
brands to become potential diagnostic tools. The researchers have demonstrated
that the combination (smartphone and algorithm) can quantify commercial
colorimetric urine tests with high accuracy and reproducibility in measuring pH,
protein, and glucose [28]. Moreover, Gaster et. al., in their research paper have
developed an ultra portable handheld device, dubbed nanoLAB, which is
powered by a battery and avoids the washing steps common in many of the
ELISA assays [29]. Also, new hand-held techniques, electrochemical sensor
NIOX MINO and fractional exhaled nitric oxide (FENO) have been demonstrated
to be effective in the measurement of exhaled nitric oxide, which is an essential
compound in the control and management of asthma [30, 31].
2.5. Enzyme-Nanoparticle Hybrid Structures for Enhanced Enzymatic
Activity
Abel et. al. investigated the interactions between metal nanoparticles and
enzymes to identify modalities that can be used to enhance the enzymatic signal
in colorimetric and luminescent-based detection methods. In their research paper
published in 2011, Abel et. al., utilized protein assay, b-BSA, silver island films
loaded at different extents (i.e., low, medium, and high) on glass slides,
15
horseradish peroxidase enzyme, and varying the distance between the enzyme
and nanoparticle using strategies such as self-assembled monolayers (SAMs),
polymer layer, and biotin-avidin protein assay layers (Illustrated in Figure 2). The
authors demonstrated that the distance between the plasmonic nanolayer and
enzyme and the extent of loading affects plasmon-enhanced enzymatic
reactions. In particular, the researchers observed that as the loading of SiFs
increased so did the intensity of the colorimetric signal. On the other hand, the
authors discovered that there is an improvement in the signal output when the
distance between the enzyme and nanoparticle is ~ 4-8 nm [32, 33].
Abel et. al, also demonstrated that there was ~4.4 fold and 67%
enhancement of color in colorimetric response of horseradish peroxidase based
on biotin-avidin model protein assay, and glial fibrillary protein (GFAP) assay
respectively, when plasmonic thin films (silver, gold, copper and nickel)
impregnated on glass substrates [34]. Consequently, Abel et.al described a rapid
method to immobilize oligonucleotides onto silver nanoparticle films hence
creating hybrid nanostructures which upon selective microwave heating thereby
reduced the incubation time from 24 hours in conventional oligonucleotide
hybridization to 10 seconds [35].
16
films sprayed onto glass slide substrates improved fluorophore stability and
showed high fluorescence enhancement in MEF-based bioassays as shown in
Figure 3 [40]. Melendez et. al., have recently demonstrated the use of the
MAMEF technique in the detection of Chlamydia trachomatis [41], and Neisseria
gonorrhoeae [42].
Figure 3. AFM image of silver island film on a glass substrate and the intensity
of fluorescence glass substrate (control experiment, No SiFs) and glass
substrate impregnated with silver island films (SiFs).
In another rapid detection technique based on the use of metal surfaces
and microwave heating, Aslan et. al., have described a novel technique,
Microwave-Accelerated Surface Plasmon-Coupled Luminescence (MA-SPCL).
This technique is used to yield sensitive and faster results from assays
completed in the buffer, serum and whole blood [43]. Using biotin-streptavidin as
model protein assay, low-power microwaves, gold coated glass slides, and
quantum dots as fluorophores, the authors demonstrated that the bioassay was
completed within 1 minute and was able to detect proteins within 200 nm due to
18
increased chemiluminescence caused by the plasmonic effects of gold
nanolayers.
Also, Previte et. al. described another technique known as the MicrowaveTriggered Metal-Enhanced Chemiluminescence (MT-MEC), which utilizes
microwaves, silvered glass slides, acridan and peroxide and chemiluminescent
reagents, and b-BSA assay as a model assay [44]. These authors demonstrated
that the MT-MEC technique afforded for rapidity and sensitivity, that is, the
protein was detected faster (<2 min) and in very low concentration (femtomole
scale) (Figure 4).
Figure 4. 3D plots of Acridian fluorescence emission vs. time vs. wavelength for
glass
substrate and silvered
substrates,
coated with BSA-Biotin and
2.7. Microwave-Accelerated
Bioassays
(MABs)
Streptavidin, exposed to low power microwave.
The Aslan Research Group described and presented a detailed
description of the MABs technique for the rapid detection of target molecules
from biological and environmental samples [45-47]. The MABs technique
involves the combined use of PMMA plates-also referred to as iCrystal plates
19
sputter coated with silver thin films (STFs) and microwaves generated by a 900
W conventional microwave oven. Using a model protein assay, b-BSA, and
colorimetric and fluorescence as methods of quantifying the protein, these
authors demonstrated that MABs technique reduced assay time from three hours
to less than 2 minutes, and improved the lowest detection limit 100-fold. Since
then, The Aslan Research Group has continued to improve the MABs technique
with bioassay platforms to reduce the total assay time in comparison to wellknown traditional methods (e.g., conventional enzyme-linked immunosorbent
assays (ELISAs), typified with long processing time (~2-4 hours). For example, in
2014, Mohammed and Aslan employed MABs technique to rapidly assay and
quantify protein p53, one of the marker protein in cancer diagnosis, using
colorimetric and fluorescence detection schemes [48]. The authors observed that
when using MABs technique, the total assay time reduced to less than 10
minutes compared to room temperature that took several hours to complete.
Also, p53 assay completed using MABs was ~100 fold sensitive compared those
done at room temperature [48].
More recently, Mohammed et. al., have employed MABs in the detection
of two proteins; glial fibrillary acid protein (GFAP) and Shigella-like toxin (STX)
(Figures 5 and 6) and also compared MABs performance using a 900 W
conventional microwave oven and the iCrystal monomode microwave system
20
[47]. MABs technique was rapid (yielded results in ~<10 minutes) and assay
sensitivity was improved ~10 fold [47]. MABs experiments performed using the
iCrystal system yielded stable (variability <0.1%) and improved sensitivities
(~1,000 fold) compared MABs experiments completed using a 900 W
conventional microwave [47]. From these experimental observations, authors
demonstrated that the MABs technique has the potential to be used as a POC
testing set-up where rapid processing times for samples are critical for patients
with life-threatening injuries or in environmental monitoring where speedy
environmental remediation programs are needed in case there is environmental
contamination. Despite these developments, there exist a few challenges, which
when addressed, will cause MABs to have a more far-reaching impact on the
detection and quantification of biological and environmental samples. The
present challenges include; (1) automation in the acquisition and processing of
test results using hand-held computational devices thereby circumventing the
need for expensive instrumentation (2) addressing the loss of nanofilms from the
iCrystal plates during succeeding assay steps.
21
Figure 6. Colorimetric response for GFAP and STX using MABs technique
[47]
23
CHAPTER III
MATERIALS AND METHODS
3.1. MATERIALS
3.1.1. Chemicals and Reagents
PMMA discs (diameter = 5 cm, thickness = 0.2 cm) were acquired from
McMaster-Carr (Elmhurst, GA, USA). 21 well press-to-seal silicone isolators used
to mark the discs into 21 wells for multiplex diagnostics and sensing were
purchased from Grace Biolabs (Bend, OR, USA). Ethanol 190 proof, diethyl ether
(reagent grade, ACS), and 2-propanol were purchased from Pharmco AAPER.
Lithium aluminum hydride powder, a reducing agent utilized to etch PMMA
surfaces (reagent grade, (95%), 3-aminopropyltrimethoxysilane (APTES),
hydrogen peroxide (30%), and sulfuric acid were purchased from Sigma-Aldrich
Corp. (Milwaukee, WI, USA). Indium tin oxide (ITO)-coated plastic PET films
(height = 0.175 mm, width = 300 mm, length = 1 m, resistivity = 14 Ω/sq), were
purchased from MTI Corporation (Richmond, CA, USA). Deionized water with
18.0 mΩ.cm resistivity at 25 0C was achieved using Millipore Direct-Q UV3
system with a 0.22 μm filter. Copy paper (8.5” x 11”) was bought from Staples
(Towson,
MD,
USA),
phosphate
buffered
saline
(PBS)
pellets,
o-
phenylenediamine HCl (OPD), and sodium phosphate-citric buffer was prepared
using deionized water.
24
3.1.2. Proteins and Antibodies
Bovine
serum
albumin
(BSA),
streptavidin-conjugated
horseradish
peroxidase (strep-HRP), biotinylated bovine serum albumin (b-BSA), and protein
A were bought from Sigma-Aldrich Corp. (Milwaukee, WI, USA). Antigen KI-67
recombinant protein and rabbit anti-mouse KI-67 polyclonal antibody (HRP
conjugated) were bought from MyBioSource (San Diego, CA USA). Recombinant
falciparum histidine rich protein-2 (HRP-2) was bought from Bio-Rad (Raleigh,
NC, USA). Mouse monoclonal antibody against KI-67, HRP-2 and ELISA kit for
HRP-2 were obtained from Abcam (Cambridge, MA, USA). Whole blood was
acquired from Good Samaritan Hospital (Baltimore, MD, USA). ELISA kit for MCLR was purchased from Abnova (Walnut, CA, USA).
3.2.
Instrumentation
Thermal test chamber with a digital temperature controller, heavy duty
vacuum pump with exhaust filter was purchased from MTI Corporation
(Richmond, CA, USA), Scanning Electron Microscope (SEM) (S-5500) used to
obtain images of silver nanowires, and ITO on PET was purchased from Hitachi
High-Technologies
Corporation
(Tokyo,
Japan).
Subsequently,
surface
characterization of unmodified and chemically modified PMMA discs was
performed using SEM at the University of Maryland Dental School Core Imaging
Facility (Baltimore, MD, USA); microwave heating was performed using a 700 W
Emerson kitchen microwave and an in-house built 100 W iCrystal system.
Absorbance values were measured using a Varian spectrophotometer from
25
Varian Inc. and Apollo plate reader (Berthold technologies). SigmaPlot version
12.5 was used for statistical analysis of all data. Water contact angle
measurement for chemically modified and unmodified PMMA discs, silanized and
non-silanized ITO sheets were done using drop shape analysis system (DSA)
purchased from Kruss USA (Matthews, NC, USA). The shaker used to agitate
the PMMA during chemical modification was acquired from Bellco Biotechnology
(Vineland, NJ, USA) and 1-hole punch was acquired from Staples (Towson, MD,
USA). An EMS 150RS sputter coater and silver target (diameter = 57 mm) were
used to introduce silver thin films (1 nm, 5 nm, 10 nm) on both modified and
unmodified PMMA discs were purchased from Electron Microscopy Sciences
(Hatfield, PA, USA). Harrick plasma cleaner used to clean the ITO surfaces
before surface modification using APTES was purchased from Harrick Plasma
Inc. (Ithaca, NY, USA).
26
3.3. METHODS
3.3.1. AIM 1: TO DEVELOP THE NEXT GENERATION OF CIRCULAR
BIOASSAY PLATFORMS FOR RAPID DETECTION OF TARGET ANALYTES
USING MICROWAVE-ACCELERATED BIOASSAYS
3.3.1.1 Modification of PMMA by a Chemical Method.
We adopted the chemical modification method for PMMA as proposed by
Cheng et. al. with minor modifications. Ideally, the circular PMMA discs were
thoroughly cleaned with 2-propanol and rinsed in deionized water and left to air
dry. PMMA discs were immersed in lithium aluminum hydride solution in diethyl
ether (16 g of LAH in 100 mL diethyl ether) for 24 hours under constant agitation
for chemical reduction to take place. Subsequently, PMMA discs were rinsed in
2-propanol and then immersed into a solution containing APTES in 2-propanol
(5% v/v) for 1 hour. PMMA discs were then rinsed in 2-propanol and later cured
in a vacuumed thermal test chamber at 125 °C for a range of hours (i.e., 1, 3, 5,
10, and 16 hours). The summary of the entire process is shown in Figure 7.
3.3.1.2. Deposition of Silver Thin Films (STFs) on Modified PMMA Surfaces.
Silver thin films of thicknesses 1 nm, 5 nm, and 10 nm were introduced on
the surfaces of chemically-modified PMMA platforms using a 21-well mask using
a plasma sputter coater loaded with a silver target. A 21-well silicone isolator
distinctly delineate the wells was carefully aligned on top of each sputtered spot
and the platforms covered with paraffin film and stored in a desiccator until
experiment time.
27
3.3.1.3. Generation and Attachment of ITO onto Silicone Isolators.
ITO dots were punched from the ITO-coated PET films using a 1-hole
punch and securely aligned with each of the individual 21-wells of the silicone
isolator. The 21-well silicone isolators with the attached ITO dots were positioned
onto the blank circular PMMA discs [Note: PMMA was used as a solid support
base and not as a platform].
3.3.1.4. Incorporation of Silver Nanowires on Paper.
Silver nanowires were synthesized by the Unalan Research Group (Middle
East Technical University, Ankara, Turkey) using polyol method, where ethylene
glycol (EG) was used as a solvent and reducing agent, poly(vinylpyrrolidone)
(PVP) as a stabilizing agent and silver nitrate as a source of silver. The prepared
silver nanowires were sprinkled on paper and left to dry. SEM was used to
validate the deposition of silver nanowires on paper. A 1-hole punch was used to
obtain silver nanowire dots which were securely attached to the 21-well silicon
isolator. A 21-well silicon isolator, now conjugated with silver nanowire dots was
pressed firmly on the circular PMMA discs [Note: The PMMA was used as a solid
support base and not as a platform].
3.3.1.5. Surface Analysis.
To ascertain whether the PMMA discs and ITO sheets were modified to
have NH2 groups, water contact angles were performed using the water contact
angle goniometer. In that regard, 20 µL of deionized water was introduced on
29
modified surfaces of PMMA surfaces and ITO and water contact angles
recorded. We compared the data obtained (water contact angle) with that
published in the literature to choose surfaces that modified to have NH2 groups.
SEM was employed to characterize the surface morphologies of circular PMMA
discs before and after the chemical modification.
3.3.1.6. Assessment of Nanofilm Stability on Next Generation Platforms
using b-BSA Protein Assay.
An assay for a model protein (b-BSA) was used to evaluate the physical
stability of the nanofilms on modified PMMA (impregnated with silver thin films of
thicknesses (1 nm, 5 nm, and 10 nm), paper functionalized with silver nanowires,
and ITO on PET platforms. The optical absorbance of the modified platforms
except for paper + silver nanowires were measured before the running the b-BSA
assay. We used the scanning electron microscope to obtain SEM micrographs
for all the platforms. The experimental scheme is as summarized in Figure 8.
3.3.1.7. The performance of b-BSA Assay on Next Generation Platforms at
Room Temperature and Low-Power Microwave Heating.
A stock solution of b-BSA (prepared in PBS) at a concentration of 10-5 M
was developed from which serial dilutions were performed to obtain
concentrations in the range 10-6 M - 10-11 M. A 30 µL solution of b-BSA of
different concentrations (10-6 M - 10-11 M) was transferred into six different wells
on three sets of platforms (i.e., ITO, PMMA modified with NH 2 groups (sputtered
with silver thin films), and SNWs impregnated on paper) as shown in Figure 9.
30
The platforms were either incubated at room temperature for one hour or
transferred into a conventional microwave oven operating at 700 W (power level
= 1) for 5 min. The wells were washed using deionized water and kept out to air
dry. 30 µL 5% (w/v) solution of blocking agent, bovine serum albumin (BSA) was
added to the wells to prevent nonspecific binding and incubated at room
temperature for 1 hour or microwave heated for 5 min. The wells were washed
using deionized water and kept out to air dry. 30 µL solution of streptavidinconjugated horseradish peroxidase was added to the wells and incubated for one
hour at room temperature or microwave heated for 5 min. The wells were
washed with deionized water and left to air dry. 30 µL of substrate solution
comprised of 4 mg o-phenylenediamine HCl dissolved in 10 mL of sodium
phosphate-citric acid buffer (pH = 5) and four µL of hydrogen peroxide was
added to the wells and incubated at room temperature for 1 hour or heated in the
microwave oven for 5 min. The reaction was stopped by transferring the contents
of the wells into HTS wells containing 2 M sulfuric acid and the absorbance of the
enzymatic product, diaminophenazine (DAP) measured at 492 nm using an UVvisible spectrophotometer.
31
3.3.1.8. Characterization of Platform Surfaces after b-BSA Protein Assays.
Optical spectroscopy measurements were done on PMMA + 1 nm, 5 nm,
10 nm, and ITO on PET using an UV-Visible spectrophotometer. Due to the
opacity of paper+silver nanowires platform, optical spectroscopy was not
performed.
3.3.1.9. Application of Next Generation Assay Platforms in Real-Life
Assays: Rapid Detection of Ki-67 by Colorimetric Method.
From the initial evaluations of our three bioassay platforms, we chose two
of the platforms 1) PMMA + 10 nm and 2) ITO on PET on which to perform
bioassays for the rapid detection of Ki-67 antigen and the steps are represented
in the scheme in Figure 9. In the first of the bioassay for Ki-67, protein A was
added to the wells and incubated overnight. After washing the surfaces, mouse
monoclonal antibodies (capture antibody) against Ki-67, dilution 1:1000 in PBS,
were added and microwave heated for 5 min at power level 1. After three
consecutive washes of the wells, a solution of bovine serum albumin (BSA) (1%
w/w) was added to block off the surfaces that unreacted with the capture
monoclonal antibodies. After microwave heating BSA and washing excess
solution off the surfaces, we added a solution containing Ki-67 antigen and
microwave heated. After washing off the antigen and addition of secondary
antibody (monoclonal mouse antibody against Ki 67 (dilution 1:1,000)), which
was then microwave heated for 5 min. Subsequently, the detection antibody
conjugated to horseradish peroxidase (dilution 1:10,000) was added to the wells,
33
microwave heated, and wells washed. A substrate for HRP, OPD, was added
and left at room temperature for 3 min for color to develop. We stopped the
reaction by transferring the contents of the wells into HTS wells containing 2 M
sulfuric acid and the absorbance of the enzymatic product, diaminophenazine
(DAP) measured at 492 nm using an UV-visible spectrophotometer. Note: All
assay wash steps were performed using PBS (pH = 7.4) and all microwave
heating was done using power level 1 for 5 minutes in the microwave settings.
Also, the control assay experiments were completed at room temperature with a
one hour waiting period for each step.
3.3.2 AIM 2: TO DEVELOP SOFTWARE TO CONVERT COLORIMETRIC AND
CHEMILUMINESCENCE RESPONSES INTO NUMERICAL VALUES FOR THE
QUANTIFICATION OF TARGET ANALYTES.
3.3.2.1 Development of Novel Software for Quantification of Target
Analytes.
MATLAB, C++, and C# were used to create novel software to quantify the
colored response of the bioassays. We used MATLAB to develop the software
under the commercial license of Mathworks.
3.3.2.2. Novel Software using MATLAB
We developed an algorithm (see Appendix II-III), which converted the
color intensity in optical images from bioassay experiments into numerical pixel
values with precision, that is, the high intensity colors produced low pixels (<50
pixels) whereas low intensity colors produced high pixels (>180 pixels). The
35
program performed digital analysis of images stored in four varieties of image file
formats (JPEG, BMP, GIF, and PNG). The imported image was converted from
colored image to grayscale where the pixels for the darkest (black) and lightest
(white) shades were ~ 0 and ~ 255 pixels, respectively. A crosshair function is
then used to select the wells of interest by manually picking points near the
center of each well. The code lets the user set a radius about these points and
reads the pixel intensities of the image within that radius. The range of the radius
is established by measuring the diameter of a single well using a ruler function.
Once the grayscale pixel intensities within each well are collected and the mean
value calculated. Each well is labeled with a number in chronological order, and
these labels are displayed in a table along with the associated mean grayscale
intensities. Thus, the pixel intensities within a specified region in each well are
averaged and shown alongside the well number and associated concentration in
tabulated format. Simultaneously, the grayscale pixel intensities are plotted
against the respective concentrations for each well, and the plot displayed in
another section of the GUI. Successive plots can be displayed on the same axes
in different colors to distinguish each line. The results of analysis can be saved
as a Matlab .m file or exported to an Excel spreadsheet.
3.3.2.3. Novel Software using C++ and C# Programming
Imread function from OpenCV was integrated into the initial part of the
program, which made the importation and conversion of picture file format (e.g.,
JPEG, PNG, TIFF) into a 3D matrix. The imported RGB image was converted to
grayscale. A mouse cursor was used to set an anchor position in the picture
36
image and using a set of parameters, that is, the size of the radius, a circle
smaller in circumference compared to that of the test wells is drawn inside the
established neighborhood. Once the boundaries are determined, each pixel point
within the boundary gets computed, and once all the points are measured, they
are averaged to give out one-pixel value within the range 0 to 255. Depending on
the number of samples being analyzed, the output is given in both tabular and
graphical formats. In graphical format, a plot of the number of pixels versus
analyte concentration is displayed with clear line demarcations for controls (i.e.,
positive control, negative control and cut-off). Any data point above the cutoff line
was considered negative while those below were considered to be positive. The
window-based application is displayed in appendix IV
3.3.2.4. Graphical User Interface
Using Java programming language in conjunction with Android and iOS
libraries, a mobile GUI platform was developed, and integrated into the main
program from section 3.3.2.3. Figure 10 shows the components of the GUI. The
GUI allowed for easy acquisition of images from samples images using camera
on Android or iOS mobile phones. Using user-friendly control tools, the images
were loaded into the mobile application, processed and the results in the form of
pixels intensity for each sample displayed both in tabular and graphical formats.
For the graphical format, a cut-off line was drawn across the graph, which
delineates positive samples from negative specimens. That is, samples with pixel
intensity above the cut-off line were deemed negative while those below the line
37
3.3.3. AIM 3: TO DEMONSTRATE THE USE OF CIRCULAR BIOASSAY
PLATFORMS AND THE NOVEL SOFTWARE FOR RAPID ASSESSMENT OF
BIOLOGICAL AND ENVIRONMENTAL SAMPLES IN A RESOURCE LIMITED
SETTING.
The overall experimental procedure for the use of novel software
application and next generation of circular bioassay platforms for detection of
biological analytes (HRP-2) and environmental toxicant (MC-LR) are given in
Figures 11 and 12, respectively.
3.3.3.1. Institutional Review Board (IRB) Approval.
Since our research involved the use of human samples, blood, and as part
of ethical research guidelines, we applied for IRB approval from Morgan State
University IRB board.
3.3.3.2. Collection of Blood Samples.
Blood samples with purple caps and without patient identifiers were
obtained from the Good Samaritan Hospital located in Baltimore, MD, USA. It is
important to note that blood samples were specimens drawn by hospital
phlebotomists via venipuncture for patient diagnostics and were destined for
disposal after seven days as per the hospital policy. After the expiry of the policy
timeline, the blood samples were transported to our laboratory (The Aslan
Research Group, Dixon Research Center Room 123) at Morgan State University,
in an insulated ice bucket according to IRB approved procedures.
39
3.3.3.3. Processing and Storage of Serum.
Upon receipt of the blood samples from the Good Samaritan Hospital, the
samples tubes were centrifuged at 2500 rpm for 10 minutes using Eppendorf
5810R centrifuge and the supernatant, serum, aspirated using a pipette into
plastic screw capped vials while the remaining blood samples were frozen (- 82°
C). The serum was aliquoted and kept in Ziploc bags labeled bio-hazard and
frozen at -82° C until when needed for experiments.
3.3.3.4. Comparison in Detection of Analytes from Biological and
Environmental Samples using Gold Standard Methods and Novel Software
In this section, we considered several target molecules: b-BSA (a model
protein) and HRP-2 in buffer, serum and whole blood (biological samples) and
microcystin-LR toxin in stream water collected from a creek (environmental
sample) adjacent to Morgan State University. The following section describes the
procedure for each target molecule.
3.3.3.4.1 Evaluation of Performance of our Software using Model Protein
Assay
Protein assay, b-BSA (a model protein) was carried out, and colorimetric
detection method was used to report the magnitude of the enzymatic signal. After
the development of the colored product and the reaction stopped with a stopping
solution, optical images of the wells were taken using an Apple smartphone (12
MP) under the following environments; (1) On the laboratory bench with incident
40
light (e.g., room light) and light emitting diode (LED) light source lighting the
backside of the HTS wells, (ii) in a dark room with only the LED light source
lighting the backside of the HTS wells. [Note: HTS wells were painted black on
the sides of the wells, and the bottom side left clear to avoid cross-talk in the
wells]. Optical images were saved using JPEG file format and later uploaded into
the GUI window of the main program using the insert button. Once the
parameters (i.e., the number of rows, the number of columns, diameter range,
and pixel range) were set, pixel computation in each well was done by pressing
the run button, and the output in the form of graph and table was yielded. A
standard curve with a cut-off line (from control experiments) was constructed and
saved using three set of experiments. A full protocol on how to use the software
is outlined in the Appendices II-V.
3.3.3.4.2. Dilution of Monoclonal Antibodies against HRP-2 and MC-LR in
Buffer (PBS), Serum, and Blood
The frozen serum and refrigerated blood were left to thaw and warm to
room temperature. First, the mouse monoclonal antibodies against HRP-2 and
MC-LR were diluted in PBS in ratios of 1:200, 1:400, 1:600, 1:800, 1:1000, and
1:2000. To mimic the human body environment, the same setup was repeated
except that the mouse monoclonal antibodies were diluted in whole blood and
serum. Antibodies were stored at +4°C.
41
3.3.3.4.3. Preparation of HRP-2 and MC-LR Antigens.
The antigens were diluted ten-fold using PBS (pH = 7.4) and stored in a
refrigerator maintained at +4°C.
3.3.3.4.4. Indirect and Competitive Bioassay for HRP-2 and MC-LR on ITO
using Our Software
Figures 12 and 13 shows the schematic depiction of the indirect HRP-2
and MC-LR assays, respectively and the details of the bioassays are given in the
sections below. Antigens were incubated separately overnight at +4°C on
silanized ITO. Unbound antigens were washed off using PBS buffer. Monoclonal
antibodies against HRP-2 and MC-LR were added and microwave heated for
heated for 5 minutes. After washing the wells three times with washing buffer,
detection antibody, goat anti-mouse conjugated to HRP was added to the wells
and microwave heated for 5 minutes. The substrate (i.e., OPD) was added and
left to incubate at room temperature for 3-5 minutes. The well contents were
transferred into the HTS wells with walls painted black and containing stop
solution of sulfuric acid. An Apple smartphone (iPhone 7) was used to take
optical images of the wells in a dark room. The wells were back-lit with LED light
source. Pixel count on the picture images was performed using the diagnostic
software.
42
3.3.3.4.5. Indirect and Competitive Assay for HRP-2 and MC-LR using
Commercial Immunoassay Kits and UV-Vis Spectrophotometer.
Each assay for HRP-2 (Figure 12) and MC-LR (Figure 13) was performed
as per the manufacturer's’ recommendations. In brief, the kits and its contents
were left to attain room temperature. HTS wells were already pre-coated with
corresponding antigens for HRP-2 and antibodies against MC-LR. Samples and
controls were added and left to incubate at 37°C for one hour. The wells were
washed thrice with a wash buffer after which an HRP-labelled detection antibody
was added and left to incubate for one hour at room temperature. After washing,
a substrate, TMB was added to each well and left to incubate at room
temperature for 15 minutes at room temperature. A stop solution was added, and
this yielded a blue color which later changed to yellow. Absorbance
measurements of the colored product were done at 450 nm using an UV-Visible
spectrophotometer.
3.4. Statistical Analysis
Data generated from UV-Visible spectrophotometer was analyzed using
SigmaPlot (version 12.5). Also, data generated by our novel software was
processed in-house and the output files stored in the memory hand-held
computing devices.
43
CHAPTER 4
RESULTS AND DISCUSSION
The output signal from colorimetric bioassays is significantly influenced by
the physical and chemical surface properties of planar assay platforms
comprised of metal nanoparticle films. Physical stability of metal nanoparticle
films on planar surfaces can be improved by employing appropriate surface
modification techniques and type of metal nanoparticles, and subsequently, the
enzymatic response of colorimetric bioassays can be increased for improved
dynamic range for the detection of biomolecules.
Solid planar surfaces are widely employed in the field of biosensors. Solid
planar surfaces, such as ceramics[49], paper [50, 51], glass [52] and plastic [53]
offer the advantage of simple adsorption for the bioassay components. In
biosensing applications, plastic, glass, and paper are the predominantly used
base support material due to their functional characteristics (i.e., optical
transparency within the visible range, light in weight, easy to manufacture,
inexpensive and availability) [54]. On the other hand, several drawbacks still
exist, mainly the lack of availability of active surface area for the attachment of
target molecules of interest, thereby limiting the extent of analyte detection.
Subsequently, bioassays performed on traditional planar surfaces suffer from
lower sensitivity (lowest concentration of analyte detectable).
To address the issue of lower assay sensitivity, current efforts are focused
on the optimization of the surface properties of the planar surfaces without
compromising the integrity of bulk [34, 35, 55-57]. In this regard, the retention
46
capacity of the surfaces can be increased by the introduction of metal
nanoparticles on the surface of platforms by either physical or by chemical
adsorption methods [58-60]. For instance, Abel et. al. demonstrated a ~4-fold
increase in the colorimetric response signal on glass surfaces deposited with
plasmonic thin films compared to the blank non-modified glass substrates, a
clear indication of retained complexes [34]. Subsequently, Abel et. al., have
established that high-throughput screening (HTS) microplates engineered with
silver island films (SIFs) exhibited higher sensitivities compared to the unmodified
control experiments [1]. Also, one can further modify the nanoparticles with selfassembled monolayers (SAMs) of alkanethiols to covalently link the bioassay
components to the surface [57, 61]. However, the use of metal nanoparticles with
the planar platforms presents an additional challenge of loss of nanoparticle films
from the surface, and anchor proteins during wash steps of the bioassays, which
compromise the efficacy of the bioassays on these platforms [34, 48].
In this work, we first investigated the stability of silver thin films (thickness
= 1 nm, 5 nm, and 10 nm) on chemically modified PMMA, silver nanowires
functionalized on paper and metal oxide (indium tin oxide) on terephthalate using
b-BSA protein assay as a model bioassay carried out at room temperature and
under microwave heating. Furthermore, we ascertained whether the surface
properties of the platforms were affected during the bioassay steps through
optical absorbance spectroscopy and scanning electron microscopy. The
absorbance values for the colored product of the enzymatic reactions were
47
determined using UV-Vis spectrophotometer, where high absorbance units
correlated with higher amounts of the captured biomolecule of interest.
Moreover, we have developed a novel software, which is easy to use,
sensitive, and able to convert colorimetric and fluorescence signal for the model
assay (b-BSA protein assay) and real-life assays (HRP-2 and MC-LR assays)
into numerical values using MATLAB and C++/C# languages. The software can
give both qualitative and quantitative results for the molecules of interest and can
be used in desktop computers and hand-held computational devices, such as,
smartphones and tablets and hence can be used in field without the need for
electrical outlets. Detailed discussions of the data generated in this work in all
three specific aims are presented in the following sections.
4.1 Development of next generation circular bioassay platforms for rapid
detection of target analytes using microwave-accelerated bioassays.
In the first specific aim, using a bioassay for a model protein, i.e., b-BSA,
three different planar assay platforms (1) poly(methyl methacrylate) (PMMA):
chemically modified to have NH2 groups before introducing silver thin films
(STFs) of thicknesses; 1 nm, 5 nm and 10 nm, (2) silver nanowires (SNW)
functionalized paper and (3) indium tin oxide (ITO) on polyethylene terephthalate
(PET) were evaluated to investigate the extent of improvements to the enzymatic
signal. Bioassays for b-BSA and Ki-67 antigen (a real-life bioassay) in buffer
were performed using low power microwave heating (total assay time is 25-30
min) and at room temperature (a control experiment, total assay time is 3 hours).
The extent of the loss of STFs and ITO were assessed by scanning electron
48
microscopy (SEM) and optical spectroscopy. Significant loss of STFs from
PMMA platforms during the execution of the bioassays using microwave heating
and control experiments were observed. The extent of ITO on PET remained
unchanged at the end of all bioassays. The lowest level of detection (LLOD) for
b-BSA bioassays were observed to be: 10-10 M for 10 nm STFs on PMMA and
paper + SNW and 10-11 M for ITO. Bioassays for Ki-67 antigen yielded an LLOD
of <10-9 M on ITO platforms, while STFs platforms were deemed unusable due to
significant loss of STFs from the surfaces.
4.1.1. Surface modification of PMMA and paper.
Figure 13A shows the contact angles and the pictures of a water drop on
chemically modified PMMA platforms, which were silanized and incubated at
125°C for different time periods, i.e., 1, 3, 5, 10, and 16 hours. In the control
PMMA surface (no silanization, no baking), a larger water contact angle of 72 ± 2
degrees was observed as compared to the modified PMMA platforms (i.e., water
contact angle range 19 ± 2 to 41 ± 10 degrees). The observed changes in the
water contact angle provided evidence for the surface modification of PMMA. It
was previously reported that the average water contact angle for amine groups
are measured to be 33 ± 4 degrees [62]. Subsequently, PMMA surfaces that
were silanized and baked for 1 hour are used in the bioassays described in this
work. SEM images of PMMA platforms show that the surface of PMMA platforms
appears smooth and roughened before and after chemical surface modification,
respectively (Figure 13B and 13C). Figure 13D and 13E show SEM images of
regular paper and SNW-functionalized paper. During the collection of SEM
49
images, the electron beam damaged regular paper due to the poor electric
conductivity of paper. Paper functionalized with SNW exhibited a dense network
of SNW and was not damaged due to the excellent electric properties of silver
[63, 64].
4.1.2. Colorimetric response from a model bioassay for b-BSA on planar
platforms.
To evaluate the effect of surface modification of PMMA platforms on
bioassay performance, the colorimetric signal from a model bioassay for b-BSA
was carried out on each of the platforms was measured and compared. Figure
14A (absorbance values for the colored product at 495 nm) and Figure 15
summarize the results of the bioassay carried out on PMMA platforms coated
with STFs of thickness 1 nm, 5 nm, and 10 nm at room temperature and using
low power microwave heating. Under microwave heating, PMMA platform coated
with 10 nm STFs (Figure 14A, top panel, solid rectangles) showed the highest
colorimetric signal output and the lowest level of detection (LLOD) of [b-BSA=1010
M], followed by PMMA platform coated with 1 nm STFs (Figure 14A, top panel,
solid circles) which yielded LLOD of [b-BSA=10-9 M] and PMMA platforms with 5
nm STFs produced the lowest signal (Figure 14A, inverted solid triangles) and
LLOD of [b-BSA=10-7 M]. At room temperature, colorimetric response and LLOD
for all the platforms [b-BSA=10-8 M] from bioassays were significantly lower
(highest absorbance value was ~0.2) as compared to bioassays carried out using
microwave heating (highest absorbance value was ~0.8).
50
Figure 14B shows the results of the model bioassay for b-BSA carried out
on the blank paper and paper functionalized with SNWs both at room
temperature and in microwave heating. Colorimetric response from bioassays on
paper with SNW (i.e., absorbance = 0.4) was increased as compared to blank
paper (absorbance = 0.1) using microwave heating. Bioassay carried out using
microwave heating on paper functionalized with SNW, and blank paper yielded
LLOD of [b-BSA=10-11 M] and [b-BSA=10-9 M], respectively (Figure 14B, top
panel). No marked variations in the colorimetric response from the bioassays
carried out on blank paper and paper functionalized with SNWs at room
temperature environment were observed. The LLOD in both platforms was
determined to be [b-BSA=10-10 M] (Figure 14B, bottom panel).
Figure 14C shows the results of the model bioassay for b-BSA carried out
on the ITO films using microwave heating (Figure 15C, top panel) and at room
temperature (Figure 14C, bottom panel). We observed no significant differences
in the colorimetric response for the bioassay under the two experimental
conditions. However, LLODs of [b-BSA=10-10 M] and [b-BSA=10-8 M] were
observed for bioassays carried out using microwave heating and at room
temperature, respectively. [Note: The horizontal lines represent the absorbance
values of the control experiments and the total assay time was 15 mins and 3
hours under microwave heating and at room temperature, respectively].
To effectively compare the bioassay performance on planar surfaces,
colorimetric response from the model bioassay from STFs on PMMA, SNW on
paper and ITO and their corresponding control experiments plotted together and
52
real-Color pictures of the planar surfaces before and after the bioassay were
provided in Figure 16. Colorimetric response at 495 nm from 10 nm STFs on
PMMA was the highest (~0.8 for 10-6 M) among all three planar surfaces,
followed by ITO (>0.4 for 10-6 M) and SNW on paper (0.4 for 10-6 M), as shown in
Figure 16A. Background colorimetric responses (control experiments) were
similar in all planar surfaces (varied between ~0.05 and 0.1) and were 4-order of
magnitude less than the absorbance values observed for various concentrations
of b-BSA.
In our previous publications related to the use of silver island films in
colorimetric bioassays, we noted that a fraction of the colored product remains on
the planar surface and/or silver island films are removed from the surface during
washing steps that result in the removal of enzyme from the surface and loss of
colorimetric response [34, 48]. To assess whether the observations mentioned
above still occur on the planar surfaces studies here, optical absorbance
spectrum of these surfaces were measured and real color pictures visually
inspected. Figure 16B clearly shows that the colored product (yellow color) is
present after the completion of the bioassay on all surfaces. On the other hand,
the color of STFs surface is darker than the other surfaces, which can be
attributed to marked differences in the initial thickness of the silver films.
53
Subsequently, the absorption spectrum of the planar surfaces before and
after the completion of the model bioassay was measured to verify and assess
the extent of the colored product present on surfaces, as shown in Figures 17-19.
Real-color pictures of ITO after the bioassays (using microwave heating) show
yellow color on the surfaces (Figure 19), however, the absorption spectrum of
ITO does not change during the implementation of the model bioassay for b-BSA
using microwave heating and at room temperature. For the sake of brevity, we
note data only for the highest (10-6 M) and the lowest concentration (10-10 M) of
b-BSA. These observations imply that the extent of colored product was
significantly less than that in the solution, and does not affect the outcome of the
colorimetric response of bioassays carried out on ITO surfaces [65].
Optical absorption spectrum of STFs (1, 5 and 10 nm thick) before and
after the completion of a model bioassay using low-power microwave heating
and at room temperature are given in Figures 17 and 18, respectively. Both
Figures 17 and 18 show a broadening and a blue-shift in the spectra for 1 nm
and 5 nm thick STFs after the completion of the bioassay, and the absorption
peak for the colored at 495 nm does not appear on both silvered surfaces. These
observations are directly attributed to the loss of silver from the surface and
change in the dielectric properties of the silver surface due to the absorption of
bioassays components [65-67]. In contrast, 10 nm STFs display marked changes
in the absorption spectra after the completion of the bioassay using microwave
heating (Figure 17).
57
Absorption spectra for STFs show the characteristics of a typical thin film
before the bioassay: high absorbance values >600 nm and no sharp surface
plasmon resonance (SPR) peak for silver around 420 nm [68]. After the
completion of the bioassay, SPR peak for silver between 400-450 nm appears,
and the absorbance values >600 nm are significantly reduced, which implies that
STFs are converted into nanoparticle films due to significant loss of silver from
the surface when exposed to microwave heating. It is also important to note that
absorbance peak for the colored product at 495 nm is detectable on the silver
surface, especially for b-BSA concentrations >10-7 M (Figure 17).
Figure 18 shows that the loss of silver from 10 nm STFs surface was
significantly less for bioassays carried out at room temperature as compared to
bioassays completed using microwave heating. These observations can be
explained by the interactions of 10 nm STFs with microwaves, where increased
coupling of electromagnetic energy with thick metal films as compared 1 nm and
5 nm STFs are expected [65]. As 10 nm STFs are repeatedly exposed to
microwave heating during the execution of bioassays, an electric charge builds
up on the surface that results in cracks in the silver films and silver can be
washed away during washing steps of the bioassay. In addition, physical damage
to STFs by micropipette tips during washing steps can result in the loss of silver
from the surface [65]. At room temperature, 10 nm STFs mostly retain the thin
film characteristics and the only source of the observed changes in the
absorption spectra for STFs is the physical damage as mentioned earlier [65].
58
4.1.3. Colorimetric response of real life bioassay on 10 nm STFs and ITO on
PET.
Once we established that 10 nm STFs and ITO surfaces could enhance
the colorimetric response of a model protein bioassay for b-BSA using
microwave heating, we employed these two planar surfaces in a real-life
bioassay for a biologically relevant antigen (i.e., Ki-67). Ki-67 is a nuclear protein
associated with cellular proliferation and prognostic marker in cancer patients
[69]. Ki-67 is present in all stages of cell division; S, G1, G2, and M phases,
except in Go [70, 71]. Figures 20-22 show the colorimetric response for the Ki-67
antigen concentration range of 10-7 -10-12 M using microwave heating and at
room temperature. Figures 20A and 21 indicates that colorimetric response at
495 nm from Ki-67 bioassay on 10 nm STFs were indistinguishable from the
background response for both bioassays carried out using microwave heating
and at room temperature.
Colorimetric response from Ki-67 bioassay conducted on ITO using
microwave heating and at room temperature was clearly distinguishable from the
background response and the LLOD was determined to be [Ki-67 = 10-10 M] in
both bioassay conditions [65]. It is important to note that the colorimetric
response on ITO using microwave heating was significantly larger than those
measured after the completion of the bioassay carried out on the same ITO
surface at room temperature (Figures 20B and 22). The observed increase in the
colorimetric response from the bioassays performed using microwave heating
can mainly be attributed to the reduction of non-specific interactions of primary
62
and secondary antibodies and HRP-labeled avidin with the ITO surface. Since
each bioassay step in the Ki-67 bioassay is completed within 5 min during
microwave heating, the occurrence of specific biological recognition events
between the antibodies and antigen are more likely than those non-specific
binding of antibodies to the ITO surface: mass transfer of antibodies towards the
surface is increased due to microwave heating [46, 48], and as the antibodies
approach the surface specific binding events are more likely to occur between
antibodies and antigens that are located ~4 to 16 nm (thickness of protein A and
primary antibody and Ki-67) away from the surface and in to the solution. In
bioassays carried out at room temperature, each step of the bioassays takes 1
hour, and subsequently, the extent of non-specific binding of antibodies to the
surface can significantly increase (despite the presence of blocking agents on the
surface).
63
4.1.4. Evaluation of physical stability of the STFs and ITO on PET during
real-life bioassay.
To investigate the effect of surface properties of STFs and ITO on the
colorimetric response of bioassays as described above, we characterized these
surfaces before and after completion of the bioassay. Real-color photographs of
ITO after the completion of Ki-67 bioassays using microwave heating and a room
temperature show no discernable color difference as compared to another,
however, are darker than the ITO surface before their use in the bioassays
(Figure 20). Real-color photographs of STFs after the completion of the
bioassays show a significant color change on the surface (Figure 23). To
investigate the reasons for the above observations, optical absorption spectrum
of 10 nm STFs and ITO before and after completion of Ki-67 bioassays for 10-7 M
and 10-12 M using microwave heating and at room temperature were measured
and are shown in Figure 23. Optical absorbance spectrum of STFs after the
completion of bioassays using microwave heating and at room temperature
(Figure 23A) reveal a significant loss of silver thins films from the surfaces as
evidenced by the decrease in the absorbance values at wavelengths >600 nm
and the appearance of the SPR peak around 420 nm [65]. Subsequently, our
earlier observation of indistinguishable colorimetric response at 495 nm from Ki67 bioassay on 10 nm STFs from the background response is directly attributable
to the loss of silver from the surface. The optical absorbance spectrum of ITO
after the completion of bioassays using microwave heating and at room
temperature (Figure 23B) reveal no loss of ITO from the surface. These
67
observations imply that accurate colorimetric response can be measured from
bioassays carried out on ITO. We note that minor differences in the absorption
spectra for ITO are due to changes in the dielectric constant of the medium
around ITO [58, [65].
To visualize the physical changes described above, SEM images with
EDS analysis of STFs and ITO surfaces before and after the completion of the
bioassay was collected (Figures 24 and 25). Figure 24 shows that STFs surface
appears as a roughened thick film before the commencement of the Ki-67
bioassay. After the completion of the Ki-67 bioassay steps, STFs appear as silver
island films (i.e., discontinuous silver film), where a significant loss of silver can
be visually observed. However, the extent of loss of silver from the STFs surface
is also dependent on the concentration of Ki-67 antigen: that is, the loss of silver
is higher for [Ki-67] = 10-7 M than that observed for [Ki-67] = 10-12 M. EDS
analysis of the STFs before (Ag: 0.7±0.1 %) and after the completion of the
bioassay using microwave heating (Ag: 0.4-0.60 %) and at room temperature
(Ag: 0.1-0.20 %) confirmed the significant loss of silver in a quantitative manner
(Figure 24-insets). In contrast to STFs, the surface of ITO after the bioassay was
completed using microwave heating appeared to be identical to the ITO surface
after the completion of the bioassay at room temperature and before the
bioassay (Figure 25).
In contrast to STFs, the surface of ITO after the bioassay was completed
using microwave heating appeared to be identical to the ITO surface after the
completion of the bioassay at room temperature and before the bioassay (Figure
68
25). EDS analysis of ITO surfaces agree with the visual evidence that ITO
surfaces are physically stable during the execution of the bioassays and no ITO
was removed from the surface to affect the outcome to the colorimetric response
from the bioassays carried out on ITO [65].
69
4.2
Development
of
software
to
convert
colorimetric,
and
chemiluminescence responses into numerical values for the quantification
of target analytes.
In this second aim, we have developed a new method to analyze
biological and environmental samples, and the technique involves the use of
image processing of colorimetric signals. This approach has afforded accurate
detection of colorimetric signals, therefore, proven to be a potential tool for
application in biomedical and environmental monitoring efforts and at the same
time providing an inexpensive alternative to the current expensive tools and
instruments. The theory behind this method is based on the observation that
color intensity of a solution varies directly with the concentration of the
biomolecule of interest. Therefore, critical levels of sample analytes can be
detected by analyzing images of infected specimens. This project required the
use of MATLAB computing, C++/C# coding to collect digital information from
images of biological samples and display the results in a presentable format.
Figure 26 shows the layout of the MATLAB application displaying four
salient sections: image upload area section, graphical and histogram output
section, tabular output section and settings section. The image upload area
section provides space to display the sample images whereas the graphical
output section displays the pixel output for each respective sample and provides
a plot against each known concentration. The tabular output section displays the
pixels results for test samples together with their corresponding concentration.
73
The settings section affords for the user to change preferences (i.e., diameter
range, pixel range) to optimize the output results.
To evaluate the efficiency of our software to detect different color
intensities, a computer generated standard profile for shades of yellow color was
uploaded to the software and analyzed. Figure 27A displays the colored and
grayscale profiles of yellow color, and it should be noted that the work presented
here (second aim) involved the conversion of colored images to grayscale color
format. Figure 27B shows the graphical output (top panel) and tabular
presentation (bottom panel) for different shades of yellow color, which we used
as hypothetical representatives for concentrations of biomolecules of interest and
their respective pixel intensity. From the output, we observed that pixel intensity
increased with decreasing concentration and vice versa; a clear indication that
our software is accurate.
Also, we evaluated whether there exists variation in pixel intensity across
rows and down the columns for empty HTS wells in different indoor
environments: (1) HTS wells on LED light source in a dark room (Figure 28), (2)
HTS wells on LED light source on a lab bench with incident light (Figure 29), and
(3) HTS wells on LED light source on the lab bench inside a black box (Figure
30). In all the three environments, we observed that the HTS wells at the center
exhibited a higher pixel intensity compared to those at the periphery. This
observation can be attributed to cross-talk between the wells which arises from
internal reflections given that the wells are transparent thereby making the center
wells appear brighter and hence higher pixels.
74
We assessed the performance of the software in an outdoor environment,
that is, exposure to direct sunlight, using different extents of the enzymatic
substrate, OPD, added onto a constant volume of streptavidin conjugated to the
enzyme horseradish peroxidase (HRP) to generate a gradient of yellow color
intensities. Figure 31 depicts the pixels intensities, and real-color pictures for
three sets of samples in HTS wells rested on either LED board or white paper
and masked with or without a shade. For all the three columns in four
experiments, the extent of OPD reduces down the columns, i.e., from eight drops
to 1 drop of OPD. Analyzed images for samples taken on LED board and white
paper and exposed to direct sunlight produced pixel ranges of ~60-220 pixels
and ~50-140 pixels, respectively (Figure 31, solid red outline). The real-color
images for samples exposed to direct sunlight and on LED board without a shade
appear brighter compared to other three set-ups. This observation can be
explained that, because the LED board is smooth and polished, it reflected the
light from the sun and given the possibility of constructive interference of the
reflected light and the incoming light, the wells appear brighter and hence higher
pixel intensity. On the other hand, the surface of the white paper is not as uniform
as that of LED board hence yielding not as much constructive interference as
that of the LED board surface. Therefore, sample colors appear darker
translating into lower pixel intensity. However, samples exposed to direct
sunlight-LED board with shade and direct sunlight-white paper with shade
(Figure 31, solid blue outline) displayed uniform pixel intensity range of ~60-180
pixels with minimum variation since the effect of direct sun’s rays are
80
blocked by the shade and the insignificant difference in their pixel intensity can
be attributed to the internal reflections and close-talk between wells.
Figure 32A shows normalized pixels for images of colored enzymatic
product taken under four different environments: (1) under direct sunlight on LED
board (black solid line with filled circles), (2) under direct sunlight on LED board
with a shade (red solid line with unfilled circles), (3) under direct sunlight on a
white paper (green line with filled inverted triangles) and (4) under direct sunlight
on a white paper with a shade (blue solid line with unfilled upright triangles).
Figure 32B illustrates their corresponding real-color pictures for each of the four
environments. From these normalized results, it was also evident that there was
minimal variation in pixels for sample images taken under a shade (blue and red
solid lines) compared to those acquired in direct sunlight (black and green solid
lines).
Following the observations from previous experiments reported on Figures
28-30, where we noted that there existed differences in pixel intensities due to
inter-well communication we painted the walls of the wells black to minimize the
inter-well reflections before taking picture images for pixel intensity processing.
The pictures were taken in the following environments: (1) a room with incident
light (Figures 33-35, cartoon sketch, top left corner) or (2) in a dark room (Figures
33-35, cartoon sketch, bottom left corner) and in both environments the samples
were lit with LED light source.
82
In order to assess the effect of darkened wells on pixel computation of
captures images, we performed three trials of pixel analysis using a model
protein assay, bBSA, in the concentration range 10-6 M to 10-12 M on a blank
PMMA platform (i.e., no silver) and on PMMA sputtered with 10 nm STFs and the
results are as illustrated in Figures 33-35. As shown in Figure 33 (i.e., trial 1),
there was insignificant pixel intensity variation for sample images taken in a room
with incident light both for samples on blank PMMA (~2.6 pixels) and PMMA + 10
nm STFs (~2.2 pixels) (Figures 33A and 33B, top panel). However, there were
pixel disparities in images taken in a dark room. While samples completed on
PMMA + 10 nm silver displayed an insignificant difference in pixel intensity (1.5 ~
pixels), the samples completed on blank PMMA showed a slight variation of
pixels (~4.6 pixels) in three columns (Figures 33B and 33A bottom panels,
respectively). We also noted a minimal change in pixel intensity, that is, 2.2
pixels versus 1.5 pixels for samples completed on PMMA + 10 nm STFs and
pictures taken in a room with incident light and a dark room, respectively.
Figure 34 shows the grayscale pixel output for bioassay for b-BSA assay
(trial 2) in the presence or absence of 10 nm STFs in a room with incident light
and in a dark room with the samples back-lit by LED light source. In the presence
of incident light, we observed minimal pixel variation between sample images for
b-BSA assay completed on PMMA platforms with and without STFs (Figure 34B
and A, top panel). However, the grayscale pixels were elevated by ~ 3 pixels on
sample images taken on PMMA platforms without STFs in a room with incident
light (Figure 34A, top panel) compared to pixels for samples images taken on
87
PMMA platforms modified with 10 nm STFs under the same environment (Figure
34B, top panel). This observation can be explained that at the end of the
bioassay process, a significant of silver is lost from the PMMA surface translating
the once uniform films of silver to silver nanoparticles which have the property to
scatter the light. Since light scattering is diffuse, the amount of reflected light
reaching the camera lenses from PMMA + 10 nm platforms is less compared to
PMMA platforms without silver. In the dark room environment, the images for
samples completed on blank PMMA and PMMA+ 10 nm STFs exhibited the
same grayscale pixel range of 162-189 pixels. However, the pixel variation
between different sample concentrations being ~4 pixels and ~0.8 pixels
completed on blank PMMA and PMMA+10 nm STFs, respectively.
Figure 35 displays the pixel values for b-BSA assay (trial 3) completed on
blank PMMA and PMMA+10 nm STFs platforms and samples images taken in a
room with incident light and in a dark room. Samples images taken in a room with
incident light and on blank PMMA platforms showed a pixel range of 172-193
pixels (Figure 35A, top panel). Under the same physical environment (i.e., a
room with incident light), we witnessed a pixel range of 172-198 pixels for sample
images obtained from PMMA +10 nm STFs (Figure 35B, top panel).
Normalized results for the colorimetric responses and the lowest limit of
detection (LLOD) for three trials of b-BSA assay completed on blank PMMA and
PMMA sputtered with 10 nm STFs under low power microwave heating are
illustrated in Figures 36-39. It should be noted that the experiments were
completed in triplicates to minimize experimental error and bias. Also, the
88
pictures were taken in a dark room and under incident light, and the samples
backlit with a LED light in both environments. Figure 36 shows trial 1 results for
b-BSA assay performed in the concentration range of 10 -6 M – 10-11 M on blank
PMMA (Figure 36A) and silver modified PMMA platforms (Figure 36B). We
observed an LLOD of 10-8 M for sample pictures obtained in a dark room
environment (Figure 36A and B, bottom panels). Under incident light illumination,
LLODs of 10-7 M (Figure 36A, top panel) and 10-8 M (Figure 36B, top panel) were
observed for b-BSA assay samples completed on blank PMMA and PMMA + 10
nm STFs, respectively. Trials 2 and 3, both completed on blank PMMA and
PMMA passivated with 10 nm STFs platforms, yielded an LLOD of 10 -8 M both in
incident light and dark room environments (Figures 37 and 38). Figure 39 shows
a summary of results from trials 1, 2, and 3 for b-BSA assay. Regardless of
whether the PMMA platforms were blank or modified with 10 nm STFs,a LLOD of
10-8 M was observed in all the samples pictures taken under incident light and in
a dark room (Figure 39A bottom panel and Figure 39B top and bottom panel)
except for samples images obtained from experiments completed on blank
PMMA platform and under incident light illumination, where we witnessed an
LLOD of 10-7 M (Figure 39A , top panel).
89
4.3 Demonstration of the use of circular bioassay platforms and the
software for the rapid assessment of biological and environmental fields in
resource-limited setting
In the third specific aim, we demonstrate the applicability of the novel
software and the circular bioassay platforms in the detection and quantification of
environmental and biological samples in a way that is accurate and with a short
turn-around time. To investigate the efficacy of our software, we considered
malaria and Microcystin-LR toxin as model disease and environmental
contaminant, respectively.
Malaria is a disease that is common in the tropical region, and it is mainly
caused by a protozoan parasite, Plasmodium. There exist four species of
Plasmodium that are known to cause disease in humans; plasmodium vivax,
Plasmodium malariae, Plasmodium ovale and Plasmodium falciparum and
amongst them Plasmodium falciparum is the most dangerous species known to
cause severe malaria epidemics in human populations.
Mosquitoes are the main vector of the malaria parasite, and it is through
mosquito bites that the malaria parasites get transmitted from infected persons to
healthy individuals. The most notorious mosquito is the female anopheles that is
known to fly long distance per single night in search of a blood meal (mosquito
magnet) a necessary material for malaria parasite development. Other
compounding factors that further promote malaria epidemic are poor drainage,
hot and wet climate, all prevalent in the tropics year-round.
94
Malaria remains among the major killer diseases in most tropical
countries. From the WHO statistics, over half million people, predominantly
children succumb to malaria scourge every year. For instance, in 2013, it was
estimated that about 198 million people contracted malaria and in as much as
mortality rates seem to be declining each year, still malaria is a predominant killer
disease in Africa and Asia.
On the other hand, Microcystin (MC) is cyclic heptapeptide cyanotoxins
produced by the blue-green algae, and there are ten different MC variants (MCLA, MC-LL, MC-LF, MC-LR, MC-LY, MC-YL, MC-LW, MC-FR, MC-RR, and MCYR) with molecular weights in the range of 909-1044 Daltons. Of these variants,
MC-LR is the most potent variant, and it causes severe toxicity in both humans
and animals, and according to W.H.O., the current allowed levels of MCs in water
is 1 µg/L for humans. Water contaminated with MCs pose a surmountable threat
to aquatic life such as fish and amphibians and members of higher trophic levels.
Among the documented adverse effects associated with MC-LR include
reproductive toxicity and DNA damage.
The results from malaria and MC-LR bioassays reported in this section
were completed on ITO at room temperature and using microwave heating.
Colorimetric responses were measured using UV-vis spectrophotometer, and
pixel intensities compute using our novel diagnostic software.
95
4.3.1 Application of developed software in detection of biological samples
In the HRP-2 assay, we considered different concentrations of anti-HRP-2
antibodies diluted in buffer, blood, and serum as summarized in the scheme in
Figure 12. For trial 1 of HRP-2 assay, the anti-HRP2 antibodies were diluted in
the range 0.565-0.0113 mg/mL in buffer and the assay completed at room
temperature (Figure 40). Figure 40A upper panel, shows the colorimetric spectral
output of the samples containing anti-HRP-2 antibodies in different proportions.
Grayscale pixel intensities for the picture image (Figure 40B, top panel) are
illustrated in Figure 40B, bottom panel. Figure 40B; bottom panel (Inset)
represents the normalized pixels obtained from the picture image. The control
experiments; positive control, negative control, and cut-off results are displayed
in red, green, and blue colors, respectively both in line graphs (Figure 40A and B
bottom panels, horizontal lines) and on picture image (Figure 41B, upper panel,
red, green, and blue circles). By visual inspection, samples 0.565 mg/mL and
0.113 mg/mL appear dark yellow as compared to 0.0565 mg/mL and 0.0113
mg/mL which are lighter yellow. From the sample spectra output (Figure 41A, top
panel), 0.565 mg/mL and 0.113 mg/mL samples exhibited higher absorbance
values (~1.0 and ~0.8) as compared to 0565 mg/mL and 0.0113 mg/mL samples
(~0.4 and ~0.2). This can be explained that in principle, there are more
antibodies in 0.565 mg/mL and 0.113 mg/mL samples than in 0565 mg/mL and
0.0113 mg/mL samples. Therefore, the antibodies bond to the antigens
specifically and exhaustively to form antigen-antibody complexes in samples with
higher concentration of primary antibodies than to those samples with lower
96
primary antibody concentrations. The amount of these complexes in a sample is
directly proportional to the color reported by a chromogen annexed to them using
a detection antibody; in this case strep-HRP. In the control experiments, the
positive and cut-off are dark yellow compared to the negative control (Figure
40B, top panel) and their spectrophotometric absorbance values were ~2.8, 1.5,
and 0.4 respectively (Figure 40A). Samples above the negative control, that is
0.565 mg/mL and 0.113 mg/mL, are positive and samples 0565 mg/mL and
0.0113 mg/mL are considered negative.
Using our diagnostic software, we measured the grayscale pixel intensity
for the real color picture on Figure 40B (top panel). The pixel intensities ranged
from 163 to 182 pixels with the darker wells recording lower pixels and less dark
samples having higher pixel count (Figure 40B bottom, panel). The zoomed in
version of the normalized pixel graphical plot (Figure 40, Inset) shows the points
that can be considered positive and negative using the negative control cut-off.
The samples below the negative control line are considered positive and those
above the cut-off are regarded as negative. Therefore, from this observation,
samples 0.565 mg/mL and 0.113 mg/mL are positive and samples 0.565 mg/mL
and 0.0113 mg/mL.
In both spectrophotometric and pixel intensity measurements, the samples
0.565 mg/mL and 0.113 mg/mL exhibited absorbance and pixel intensities above
and below the negative control (Figure 41 A and B, bottom panel), respectively
and as such were positive or diseased samples.
97
Although the samples were considered positive their readings were below the
diagnostic threshold, that is, the cut-off measurement.
We repeated the same experiment twice and instead of completing the
HRP-2 assay at room temperature, we used microwave heating (power level = 1,
time = 5 mins). Colorimetric responses obtained using UV-vis spectroscopy and
pixel intensity computation achieved through novel diagnostic software for HRP-2
assay are illustrated in Figures 41 and 42 representing trials 2 and 3,
respectively. The absorbance values for the enzymatic product revealed that all
the samples in the dilution range 0.565 mg/mL and 0.0113 mg/mL, were
diagnostically positive because they exceeded the cut-off limit line (Figure 40A)
depicted by a magenta spectrum line (Figure 41A, top panel) and a solid green
line (Figure 41A, bottom panel). Thus, trial 2 of HRP-2 assay yielded an LLOD of
0.0113 mg/mL under UV-vis spectrophotometry. Strikingly, only one sample (i.e.,
0.565 mg/mL) was diagnostically positive (Figure 41B) under the diagnostic
software which computes grayscale pixel intensity, despite samples 0.1130.0113 mg/mL appearing darker than the controls. Consequently, the magnitude
of the positive control was unusual in its output as compared to how it appears in
real-color pictures (Figure 41B, bottom panel). This observation was anomalous
especially with our software and we decided to perform a third trial for HRP-2
assay where we expanded the dynamic range by addition of one sample (i.e.,
0.00565 mg/mL). Figure 42 shows colorimetric responses and grayscale pixel
intensity computation for HRP-2 assay completed under microwave heating.
Through UV-vis measurements, we observed an LLOD of 0.0113 mg/mL (Figure
99
42A) whereas no sample was recognized as positive by our software (Figure
42B, bottom panel, Inset) since the output from our software depicted control
samples as darker than test samples. These results were against the principle of
color computation using grayscale format which stipulates that the darker the
image the lower the pixels and vice versa. From the real-color pictures (Figure
42B, upper panel), it is evident that the sample images are darker in color, that is,
with yellow pigmentation (black solid line enclosure) as compared to the controls
(red, blue and green solid line enclosures), which are lighter in appearance.
These observations led us to hypothesize that during the conversion of colored
pictures to grayscale; there could be a significant loss of pixels. This is
preferentially predominant in samples with a lighter color than those samples that
have darker colors leading to the former samples to be regarded as false
negatives.
Figure 43 shows results from HRP-2 assays carried out under the same
conditions as previously described, that is, microwave heating and on modified
ITO platform. However, we increased the volume of enzyme substrate by 3-fold
to completely utilize all the enzymes inside the wells to exhaustively convert the
substrate to color. In Figure 43A, top panel, we observed highest absorbance
values of ~0.6 for 0.565 mg/mL samples (solid black spectrum line) and lowest
absorbance values of ~0.3 for 0.00565 mg/mL samples (solid blue spectrum
line). In Figure 43A, bottom panel, all the samples were diagnosed as positive
because their absorbance values were higher than that of cut-off (~0.05) and
negative control (~0.1).
100
Surprisingly, the increase in substrate volume did not yield more color instead,
the intensity of colored product was reduced ~ 2-fold when compared to HRP-2
assay results reported in Figure 42 where we used the volume recommended by
the manufacturer. This observation can be explained by the fact that the excess
substrate that was never converted to a colored product combined with the
stopping solution inside the HTS wells, resulting in the dilution of the colored
product. From the software output illustrated on Figure 43B, bottom panel, we
observed higher grayscale pixel values for 0.565 mg/mL samples (> 200 pixels)
compared to 0.00565 mg/mL samples (>190 pixels). This implies that 0.565
mg/mL samples are lighter than 0.00565 mg/mL samples and this is not
compliant with the real color image on Figure 43B, upper panel, which displays
0.565 mg/mL samples as darker in color compared to 0.00565 mg/mL samples.
This observation fits into the prior explanation that there is loss of pixels during
image conversion from color format to grayscale format.
We have also evaluated the performance of our software with HRP-2
assays in buffer completed at room temperature and compared the results with
those from UV-vis spectroscopy. It should be noted that the volume of substrate
used in this assay (HRP-2 assay in buffer) and HRP-2 assays in serum and
blood is 3-fold. Figure 44 demonstrates both spectral and grayscale pixel
computation for the enzymatic product for HRP-2 assay completed at room
temperature. The spectra display is uniform with the clear distinction in
absorbance values for high concentration and lower concentration samples, that
is, 0.565 mg/mL abs = ~2.5 and 0.00565 mg/mL samples abs = ~0.7 (Figure
104
44A). Using our software to analyze the real color picture (Figure 44B, top panel)
we observed that the pixel increased with a decrease in concentration for 0.565
mg/mL, 0.113 mg/mL and 0.0565 mg/mL samples. Beyond these three sample
points, the pixel intensity for two remaining samples, 0.0113 mg/mL and 0.00565
mg/mL reduced with decrease in sample concentration. This observation can be
explained using the picture image on Figure 44B, top panel. The first three wells
(i.e., 0.565 mg/mL, 0.113 mg/mL and 0.0565 mg/mL) display a darker yellow
color than 0.0113 mg/mL and 0.00565 mg/mL and controls. Therefore, when
converted to grayscale, they appeared darker hence the lower the pixels.
Surprising was the observation for 0.0113 mg/mL and 0.00565 mg/mL and
controls. They all appear lighter in the real color image which needed them to
have translated into higher pixels. Instead, these samples (0.0113 mg/mL and
0.00565 mg/mL) recorded low grayscale pixels as 0.565 mg/mL, 0.113 mg/mL,
and 0.0565 mg/mL samples. This anomaly can be attributed to the loss of pixel
information during the conversion of color images to grayscale format.
Nonetheless, under UV-vis spectroscopy we observed a LLOD of 0.0565 mg/mL
(Figure 44A, bottom panel) whereas no sample could be regarded as positive
under grayscale pixel computation.
105
The same observation is partially captured using our computation software
as illustrated in Figure 44B, bottom panel. Samples 0.565 mg/mL, 0.113 mg/mL,
and 0.0565 mg/mL are dark yellow in picture images and they generate lesser
grayscale pixels when computed on our software. That is, the grayscale pixel
intensity increases with a decrease in sample concentration. This assertion is
represented by an almost linear relationship between three samples points for
0.565 mg/mL, 0.113 mg/mL, and 0.0565 mg/mL samples. Contrary, samples
0.0113 mg/mL and 0.00565 mg/mL which fundamentally need to have higher
grayscale pixel intensity, yielded pixel values that are marginally alike to that of
0.565 mg/mL, 0.113 mg/mL, and 0.0565 mg/mL samples (Figure 44B, bottom
panel). We observed an LLOD of 0.1 for HRP-2 in buffer at room temperature.
The color of HRP-2 assay in buffer carried at room temperature is darker
compared to the same assay completed using microwave heating. This can be
explained that microwave heating of HRP-2 bioassays for 5 minutes is not
sufficient to generate a temperature gradient to allow for the interaction and
recognition of antibodies by the antigens immobilized on the wells because from
our previous work we observed microwave heating on ITO platform beyond
power level 1 and 5 minutes led to the damaging of ITO platforms. Therefore,
fewer antigen-antibody complexes form which translates into less enzymatic
colored product. Bioassays completed at room temperature produced more
colored product which indicates that there were more antigen-antibody
complexes formed due to extended time that allowed more antibodies to
recognize the antigens to bind.
107
We have also evaluated the performance of HRP-2 assay in serum and
analyzed the results using both the UV-vis spectrophotometer and our software.
Figure 45 and 46 displays results for HRP-2 assay under microwave heating and
at room temperature, respectively. From the spectral output illustrated in Figure
45A, upper panel, we respectively observed the highest and lowest absorbance
values of ~0.4 and >0.2 and a LLOD of 0.00565 mg/mL (Figure 45A, bottom
panel). Running software grayscale pixel computation on picture image (Figure
45B, top panel), yielded an output shown in Figure 45B, bottom panel. The
results demonstrated that our software could detect up to 0.565 mg/mL of HRP-2
antibodies in serum which is 1000-fold less than UV-vis spectrophotometer.
From the optical absorption spectra for HPR-2 assay in serum completed
at room temperature, we observed a maximum and minimum absorbance values
of ~0.5 and ~0.4, respectively (Figure 46A, top and bottom panel). Subsequently,
we observed a LLOD of 0.00565 mg/mL and 0.113 mg/mL for HRP2 antibodies
in serum utilizing UV-vis spectrophotometer and our software respectively. Also,
the HRP-2 assay color at the end of the experiment at room temperature was
deep yellow as compared to the same assay under microwave heating (Figures
45 and 46). These observations are attributed to the length of the assay at room
temperature (i.e., one hour in each step), which allow for the primary antibodies
to interact with the immobilized antigens to form antigen-antibody complexes
which are readily recognized by detection antibodies conjugated to a chromogen
yielding more color in presence of a substrate.
108
After assessing the performance of our software in the detection of the
extent of HRP-2 antibodies in buffer and serum, we proceeded to evaluate its
efficacy in detecting the same HRP-2 antibodies in whole blood, and
subsequently, compared the results with the standard method, UV-vis
spectrophotometry. Figure 47 shows the colorimetric response and the grayscale
pixels for HRP-2 assay in blood and performed under low power microwave
heating. From the optical absorption spectra shown in Figure 47A, all the
samples in the concentration range 0.565-0.00565 mg/mL, exhibited absorbance
of >0.4 compared to background (control experiments: positive, cut-off, and
negative), which displayed absorbance of <0.1. We also witnessed highest (0.9)
and lowest (0.3) absorbance for samples with concentrations 0.0113 mg/mL and
0.565 mg/mL, respectively (Figure 47A, bottom panel). The real color pictures for
the enzymatic product for HRP-2 assay in blood is shown in Figure 47B top
panel. The color is intense in sample wells (solid black enclosure) compared to
control experiments: positive (red solid enclosure), cut-off (green solid enclosure)
and negative, (blue solid enclosure). When the picture of the sample was
uploaded into our software and converted into grayscale, the pixel intensity for
each well was computed and the yield is as shown in Figure 47B, bottom panel.
109
The grayscale pixel intensity was lowest (~183 pixels) for samples with a
concentration of 0.565 mg/mL and the grayscale pixel intensity increased
gradually to the highest pixels of ~215 pixels, which was associated with 0.0565
mg/mL samples. Subsequently, the pixel intensity for 0.113 mg/mL, 0.0113
mg/mL and 0.00565 mg/mL were 205, 215 and 210 pixels, respectively. Under
the UV-vis analysis, we observed a LLOD of 0.00565 mg/mL, however, using our
novel software computation, all the samples (0.565 mg/mL, 0.113 mg/mL, 0.0565
mg/mL, 0.0113 mg/mL, and 0.00565 mg/mL) were considered negative, that is,
their grayscale pixel intensities were higher compared the pixel intensity of the
negative control experiment.
112
Figure 48 displays the results of HRP-2 assay in blood conducted at room
temperature. The assay generated the highest and lowest absorbance values of
1.1 and 0.9 for 0.0113 mg/mL and 0.565 mg/mL samples, respectively under UVvis analysis. Other samples yielded absorbance values as follows: 0.565 mg/mL
(abs = 1.06), 0.00565 mg/mL (abs = 1.05) and 0.113 mg/mL (abs = 1.04).
Subsequently, the control samples displayed the following absorbance values;
positive (abs = 0.4638), cut off (abs = 0.3553) and negative (abs = 0.2386)
(Figure 48A, bottom panel). The picture image (shown in Figure 48B) for the
enzymatic product has the same organization and display as previously
described in Figure 47B. However, the control samples in Figure 48B developed
slightly deeper compared to those in Figure 47B and this can be attributed to the
longer waiting periods at room temperature which affords for more antigenantibody complex formation. Consecutively, these complexes attach to the
polyclonal antibodies bound to HRP enzyme and following the addition of a
substrate, the enzyme breakdown the substrate to yield color. The color
produced is proportional to the amount of chromogen bound, which is also
proportionate to the amount of antigen-antibody complexes. Therefore, Figure
48B, top level (black solid line enclosure) depicts that the amount of antigenantibody complexes formed in test samples are significantly predominant
compared to those formed in control experiments. Figure 47B, bottom panel,
demonstrates the output from our software in which the pixel intensity linearly
increases with a reduction in test sample concentration for the first three
samples: 0.565 mg/mL (202 pixels), 0.113 mg/mL (210 pixels), and 0.0565
114
mg/mL (218 pixels). After the third sample, we observed a decline in pixel
intensity, that is, 0.0113 mg/mL (216 pixels) 0.00565 mg/mL (212 pixels) and
these reductions in pixel intensity can be attributed to the loss of pixels when the
colored test sample picture is converted to grayscale format.
Figure 49 displays a layout for real picture images for the enzymatic
product for HRP 2 assay in blood, serum and buffer completed at room
temperature and under low power microwave heating. Although HRP-2 assays in
serum and buffer conducted under microwave heating (Figure 49, top panel)
yielded color for the enzymatic product, the color was not as predominant as that
witnessed in HRP-2 assay in blood where the color was deep yellow.
Conversely, HRP-2 assays in blood, serum and buffer and conducted at room
temperature produced a consistent intensity of yellow color for the assay product
in all test samples. These observations can be explained that due to the long
waiting period (of 1 hour) for primary antibodies and polyclonal antibodies,
unspecific interaction with the blocking agent are likely to occur that lead to the
generation of more color in the presence of enzyme substrate. On the other
hand, we observed little color change in intensity for the enzymatic product in the
control samples completed both at room temperature and under low powermicrowave heating.
115
Following the results from grayscale computation for HRP-2 assay test
sample pictures, we observed that the output was not comparable with the realcolor of the enzymatic product in the wells. That is, wells with deep yellow color
did not translate into low grayscale pixel intensity and vice versa. Subsequently,
the output from greyscale computation was compared to the standard method
(UV-vis) and the optical absorbance spectra results from UV-vis were
harmonious with colored product of HRP-2 assay (i.e. absorbance increased with
increase in antibody concentration). In that regard, we sought to investigate
whether RGB (Red, Green, and Blue) color format for pixel computation could
yield better and consistent results for HRP-2 test sample pictures to ultimately
replace grayscale pixel computation. Since the color for the enzymatic product
for our HRP-2 assay is yellow, a construct of R, G, and B proportions, we
endeavored to establish the influence of each color component in the triad had
on the saturation (luminosity) of yellow color. Figure 50 gives a summary of RGB
format pixel output (full RGB grid, Green color, Blue color, and Red color pixels)
for random samples (with an insignificant difference in saturation for yellow color
-Figure 50A) taken on a 21-well iCrystal plate using 12 MP iPhone camera.
Although the intensity of yellow color appears uniform in 21-wells (Figure 50B),
the distribution of intensity in the constituent colors (i.e., RGB) is clear (Figure
50B). The pixels values for Green and Red colors have an equivalent distribution
(~250 pixels) in all the 21 wells. But the pixel values for Blue color are varied in
all the wells. From these observations, we were convinced that the Blue color
118
constituent (not Green or Red color component) determines the extent of
brightness in the Yellow color and as such, RGB format could be utilized for
calibration in our novel software.
To quantify the above-mentioned observations that blue color regulated
the brightness of yellow color, we performed a real-life assay (HRP-2 assay) in
the concentration range 0.565-0.00565 mg/mL on 21-well iCrystal plates under
microwave heating and the results are as shown in Figure 51. Figure 51, top
panel (center), depicts a histogram of RGB pixel values for the colored enzymatic
product from HRP-2 assay in eight wells. Well numbers 1-5 and 6-8 represent
the concentration range of 0.565-0.00565 mg/mL of HRP-2 antibodies and
control experiments (i.e., 6-positive control, 7-cut-off, and 8-negative control),
respectively (Figure 51, top right corner). Figure 51, bottom panel, shows the
RGB output for Blue, Green and Red colors with the pixel values for control
samples represented by solid horizontal lines (i.e., red solid line = positive
control, green solid line = cut-off, and blue solid line = negative control). The
normalized pixel results for three colors illustrates that the pixel values for green
and red colors are constant in all the five samples (0.565-0.00565 mg/mL).
On the other hand, the blue color pixels values vary from one
concentration to another; that is, we observed low pixel values (~0.4) for high
sample concentration of HRP-2 antibodies (0.565 mg/mL) and high blue pixel
values (<0.7) for low concentration of HRP-2 antibodies (0.00565 mg/mL) as
shown in Figure 51, top left corner (closed-in version of Figure 51, bottom left
corner). Also, none of the test samples (0.565-0.00565 mg/mL) could be resolved
120
either as positive or negative using green or red color pixel values since their
pixel values were like those produced by the control experiments. So, the LLOD
for test samples was not obvious. However, when we analyzed the test samples
using the blue color constituent, the pixel values both for test samples and
control experiments were distinct from each other making the diagnosis
apparent. Subsequently, we observed an LLOD of 0.00565 mg/mL. Therefore,
we recognized that the blue pixels regulate the brightness of yellow color and can
be used to define a set of samples with a decreasing lightness of yellow color.
After establishing the descriptive effects of blue pixels on yellow color
(color of enzymatic product from HRP-2 assay) in test samples completed on
iCrystal plates, we further investigated whether the outcome was similar for the
assay pictures taken on HTS wells painted black on the outside (Figure B). We
note that the contents in HTS wells are colored enzymatic product from HRP-2
assay performed on iCrystal plates and transferred into the HTS wells to stop the
enzymatic reaction by the stop solution. Figure 52 shows results for HRP-2 assay
completed at room temperature and test sample pictures taken on HTS wells
(Figure 52B). The absorbance at 450 nm (Figure 52A), grayscale and RGB pixel
intensities (Figures 52C and 52D) for the colored product were determined using
UV-vis spectroscopy and novel software, respectively. The optical absorbance
spectra indicate that sample absorbance values increased with increase in the
concentration for test samples (Figure 52A). These observations can be
explained that high concentrations of HRP-2 antibodies readily recognize and
combine with immobilized antigens yielding a higher number of antigen-antibody
121
complexes onto which detector antibodies conjugated to HRP enzyme bind and
in the presence of substrate, the color that develops is commensurate to the
amount of enzyme present. However, the output from the grayscale pixel
computation for the image on Figure 52B, shows that the pixels of the first three
samples (0.565-0.0565 mg/mL) increased linearly with a decrease in sample
concentration and then the pixels reduced for the last two samples (0.0113
mg/mL and 0.00565 mg/mL) (Figure 52C). These observations can be described
using image on Figure 52B. The color inside the first three wells (marked 8, 7,
and 6) which represents concentrations (0.565, 0.113 and 0.05656 mg/mL),
exhibit deep color intensity than the rest of the wells and during conversion from
colored image to grayscale image, the wells in these three rows are darker hence
the low pixel intensity. For grayscale pixels intensities for 0.0113 and 0.00565
mg/mL and control samples show a reduction in intensity this can be attributed to
loss of pixels during image conversion. Conversely, the same picture image
generated different results for RGB constituents. As shown in Figure 52 D (Green
and Red), the green and red components produced results similar in trend to
grayscale pixel computation, that is, increase in pixel values with decrease in the
concentration for the first three samples. However, for the blue constituent, the
results are different: the pixel values increase with reduction in sample
concentration (Figure 52D, Blue) and these results are accommodating with the
picture image and the histogram of stacked RGB pixel values (Figure 52B). Also,
in samples where we observed a high intensity of yellow color, there were low
blue pixels and vice versa. When we compare the output of blue pixels and
122
optical absorbance from UV-vis, it is clear that the relationship is of inverse
proportionality. Subsequently, we observed an LLOD of 0.00565 mg/mL for HRP2 assay in buffer completed at room temperature and analyzed using UV-vis and
blue constituent of the RGB pixel computation.
Once we finished assessing the performance of our software in HRP-2
samples completed at room temperature and under microwave heating (kitchen
microwave) as captured in previous discussions, we advanced to evaluate the
efficacy of the software in HRP-2 samples conducted using the iCrystal system
operated at 70 W and increased the concentration dynamic range from 0.5650.00565 mg/ml to 0.113-0.000565 mg/mL. The iCrystal system was developed by
the Aslan Research Group at Morgan State University and it has five different
components (Appendix VI). Figures 53-55 shows the results for HRP-2 assay
using three different methods of analysis: UV-vis spectroscopy, grayscale, and
RGB (blue pixel component). Figure 53 displays the results for HRP-2 assay in
buffer and completed using the iCrystal system maintained at 70 W for 5
minutes. The optical absorbance values at 450 nm for the test and control
samples are as shown in Figure 53A, top panel. From these results, we observed
the highest absorbance (>0.8) for 0.00565 mg/ml sample and the lowest
absorbance (>0.2) for 0.000565 mg/mL samples and an LLOD of 0.000565
mg/mL (Figure 53A, bottom panel). The grayscale pixel computation method,
revealed a decrease in pixel intensity for the most concentrated samples
123
(0.113 mg/mL) to the least concentrated (0.000565 mg/mL) evidenced by the tilt
of slope for the normalized pixels (direction right to left) (Figure 53B). These
observations signified that the samples appeared darker as we moved from the
most concentrated samples to the least concentrated. Comparing the results
from Figure 53B with real-color image for enzymatic product for HRP-2 assay on
Figure 53D, it is evident that the samples marked 4 (i.e. 0.00565 mg/mL) had
intense yellow coloration compared to all other samples. However, the same
sample shows grayscale pixel values of ~0.9 (Figure 53B, third data point from
right) compared to sample 6 (i.e., 0.000565 mg/mL) average ~0.7 pixels, which is
the lightest in color: the greyscale computation displays samples with lighter
coloration as darker (i.e. lower pixels) and vice versa, a condition that is
inaccurate and these observations can be explained by the loss of pixels during
image conversion. For RGB pixel computation using blue color component, the
pixel output was observed to decrease at both extremes of the data range. That
is the center wells (i.e. wells 4 = 0.00565 mg/mL) showed the lowest blue color
pixels whereas wells 1 and 2 (0.113 and 0.0565 mg/mL) and wells 5 and 6
(0.00113 and 0.000565 mg/mL) displayed a decrease in blue pixels with
decrease in concentration and increase in pixels with decrease in sample
concentration, respectively. These observations can be explained using the
picture image (Figure 53D). Wells 1 have a lighter yellow coloration compared to
wells labelled 2. As previously observed, the blue pixels regulate the saturation of
the yellow color, that is, at lower saturation (i.e. light yellow), the amount of the
blue pixels is higher and at higher saturation (dark yellow) the extent of blue
126
pixels is low. Therefore, from the real-color image (Figure 53D), wells labelled
four had the lowest blue pixel value (~0.7) because they are darker (in yellow
coloration) compared to wells labelled 1 (lighter in yellow coloration) with blue
pixel value of ~0.9. Other wells have the following blue pixel values: wells marked
2 = 0.75 pixels, 5 = 0.72 pixels, and 6 = 0.78 pixels.
Figure 54 shows the optical absorbance, grayscale, and RGB (blue
component) pixel values of the enzymatic product (yellow color) for HRP-2 assay
in serum completed in the iCrystal system operated at 70 W for 5 minutes for
each assay step. Figure 54A, upper panel displays the optical absorbance values
for test samples (in the range of 0.113-0.000565 mg/mL), and control samples;
positive, cut-off, and negative control experiments. We observed highest
absorbance (~1.5) in 0.0565 mg/mL samples and lowest absorbance of ~0.6 in
0.000565 mg/mL sample. The absorbance for the other samples were 0.113
mg/mL (Abs = 1.0), 0.0113 mg/mL (Abs ~0.9), 0.00565 mg/mL (Abs = 0.9), and
0.00113 mg/mL (Abs = 0.6). The control samples exhibited the following
absorbance values; positive control (Abs = 0.19), cut-off (Abs =0.14), and
negative control (Abs =0.06). We also witnessed a LLOD of 0.000565 mg/mL
under UV-vis analysis. Subsequently, Figure 55B displays the normalized
grayscale pixel values for the sample image (Figure 54D) for the HRP-2 assay.
The normalization of pixels happens by dividing the pixels for each sample
(Figure 54D, wells 1-7) and control wells (Figure 54D, wells 7-9) with the pixels of
the blank samples (Figure 54D, wells labelled ‘B’). We witnessed the highest and
lowest grayscale pixel values of 1.11 and 1.02 for 0.0565 mg/mL and 0.000565
127
mg/mL, respectively. Grayscale pixel values for other test samples include 0.113
mg/mL (pixels = 1.04), 0.0565 mg/mL (pixels = 1.06), and 0.00565 mg/mL (pixels
= 1.1). Control samples: positive control (pixels = 1.0), cut-off (pixels = 1.01) and
negative control (pixels =1.0). Using grayscale pixel computation, all the samples
were considered negative: since all the control samples showed lower grayscale
pixel values than test samples. On the other hand, the Figure 54C gives a
summary of results for RGB pixel computation-blue component. For RGB (blue
component) analysis, we observed the highest and lowest blue pixels of 0.75 and
0.57 pixels for 0.0113 mg/mL and 0.0565 mg/mL test samples, respectively. The
blue pixel values for other test samples were 0.000565 mg/mL (pixels = 0.74),
0.00565 mg/mL (pixels = 0.72), 0.113 mg/mL (pixels = 0.62) and control samples
(positive control = 0.89 pixels, cut-off = 0.95 pixels and negative control = 0.92).
Also, RGB analysis-blue pixels displayed a LLOD of 0.000565 mg/mL. Figure
54B shows the real-color image for the enzymatic product inside HTS wells
painted black on the outside. Each row represents a given test sample
concentration repeated in triplicates (i.e., 1= 0.113, 2 = 0.0565, 3 = 0.0113, 4 =
0.00565, 5 = 0.00113, 6 = 0.000565 mg/mL. Controls: 7 = positive control (color
code: red), 8 = cut-off (color code: green), 9 = negative control (color code: blue)
and B = blank sample).
Figure 55 displays the colorimetric response obtained using UV-vis
spectroscopy, grayscale, and RGB (blue component) pixel computation
completed using our novel software for HRP-2 assay in blood accomplished
under microwave heating (the iCrystal system). The optical absorbance values
128
shown in Figures 55A and B, indicate the highest and lowest absorbance values
of ~1.0 and ~0.5 in samples 0.0565 mg/mL and 0.000565 mg/mL, respectively.
Other HRP-2 assay test samples produced the following absorbance values:
0.113 mg/mL (Abs = 0.85), 0.113 mg/mL (Abs = 0.82), 0.00565 mg/mL (Abs ~
0.7), and 0.00113 mg/mL (Abs = 0.67). We observed absorbance values 0.19,
0.08 and 0.12 for positive, cutoff, and negative control samples, respectively and
from these results we detected a LLOD of 0.000565 mg/mL for HRP-2
antibodies. Subsequently, we determined the grayscale pixels for the enzymatic
product for HRP-2 assay in blood and the results are displayed in Figure 55B.
From these results, we observed that the pixels values increased with a
decrease in test sample concentration for the first four sample trials (i.e., 0.113
mg/mL (pixels = 1.02), 0.0565 mg/mL (pixels = 1.06), 0.0113 mg/mL (1.09), and
0.00565 mg/mL (pixels = 1.1)), but the grayscale pixels reduced with a reduction
in specimen concentration in the last two samples (0.00113 mg/mL (pixels =
1.05), and 0.000565 mg/mL (pixels = 0.99)) (Figure 55B, Inset). The control
samples exhibited the lower greyscale pixel values compared to the test results.
That is, positive control (pixels = 0.92), cut-off (pixels = 0.94 and negative control
(pixels = 1.01). However, the trend was different when we subjected the
enzymatic product into RGB (blue component) pixel computation. As shown in
Figure 55C, the blue pixel values for the HRP-2 assay product increased with a
decrease in sample concentration except for the first (0.113 mg/mL) and last
(0.000565 mg/mL) sample sets which showed a slightly high (0.66 pixels vs. 0.62
pixels) and lower (0.73 pixels vs. 0.77 pixels) blue pixels values compared to
129
those in the next and previous neighbors respectively (Figure 55C, Inset). The
control samples, positive, cut-off and negative controls, showed higher blue pixel
values (positive control = 0.81, negative control = 0.90 and cut-off = 0.88) than
HRP-2 antibody samples. Using RGB – blue pixel computation, we witnessed an
LLOD of 0.000565 mg/mL for HRP-2 antibodies diluted in blood.
So far, all the pixel results (grayscale and RGB) that we have reported
have been computed using MATLAB application. But, MATLAB suffers a huge
limitation of being only available to licensed users. To that effect, we have
developed a windows-based application with not only comparable capability to
perform RGB computation as the MATLAB program but also available to all
users with a windows computer. We have compared the efficiency of both
MATLAB program and window-based application in terms of LLOD and pixel
distribution for test sample images for HRP-2 assay taken in iCrystal plates and
HTS wells.
130
Figure 56 shows the RGB pixel output for enzymatic product (in 21-well
iCrystal plates and HTS wells) for HRP-2 assay in buffer completed in a kitchen
microwave oven and the test sample picture were analyzed using windowsbased application. In Figure 56A and B (top panel), we observed that the
normalized pixels for red and green channels levelled off (formed a plateau) in
the first two samples (i.e., 0.565 mg/mL and 0.113 mg/mL) and then reduced with
a decrease in sample concentration in 0.0565, 0.0113 and 0.00565 mg/mL
samples. For Figure 56A (Inset-top panel), we observed the highest and lowest
red pixels of 1.0 and 0.95 pixels for samples 0.565 mg/mL and 0.00565 mg/mL,
respectively. Figure 56B (top panel) shows the green pixels for HRP-2 assay
where we witnessed the highest pixels of 1.03 for samples 0.565 mg/mL and low
pixels of 0.95 pixels for 0.00565 mg/mL. The blue pixel results showed a different
trend from red and green channels, that is, the blue pixels increased with a
reduction in sample concentration (Figure 56C, top panel). We witnessed the
lowest pixels of 0.7 and highest pixels of 0.9 for 0.565 mg/mL and 0.00565
mg/mL samples, respectively. In the control experiments, represented by
horizontal lines in Figures 56A-C, positive control (red solid line) showed the
highest pixels (1.17) followed by cut-off (1.0 pixels) (green solid line) and the
negative control (blue solid line) showed the lowest pixels (0.99). We noted that
the utilization of red and green and blue pixels (in our window-based software) to
analyze HRP-2 samples on 21-well iCrystal plates yielded a LLOD of 0.0565
mg/mL and 0.00565 mg/mL, respectively (Figure 56A-C).
134
Figure 56D-F presents normalized pixels of red, green and blue channels
for the enzymatic product for HRP-2 assay in HTS wells. The samples were
analyzed using our novel windows-based software. The pixels for the red
channel increased steadily with an increase in sample concentration and we
observed the least red pixel value of 1.0 in 0.00565 mg/mL and maximum red
pixel values of 1.2 in 0.565 mg/mL samples (Figure 56D). We observed the same
trend of increase of pixels with increase in sample concentration for green
channel pixels where we noted the highest and lowest green pixels of 1.1 and
1.02 in 0.565 mg/mL and 0.00565 mg/mL samples, respectively (Figure 56E).
For the red and green constituents, we observed the following pixels for their
control experiments: positive control (red = 1.01 pixels, green = 1.05 pixels), cut
off (red = 1.01 pixels, green = 1.03 pixels), and negative control (red = 1.06
pixels, green = 1.04 pixels) (Figure 56 A-B, bottom panel-Inset). We noted that
the pixels for the blue channel increased with a decrease in sample
concentration where the highest and lowest blue pixels of 0.88 and 0.72 were
observed in 0.00565 mg/mL and 0.565 mg/mL samples, respectively (Figure
56F). The blue pixels for the control samples, represented by horizontal solid
lines, showed higher pixel values than test samples: positive control (pixels =
0.85), cut-off (pixels = 0.87) and negative control (pixels = 0.89). When using red,
green, and blue channel pixels (in our window-based software) to evaluate the
enzymatic product for HRP-2 assay on HTS wells we observed a LLOD of
0.0565 mg/mL, 0.0113 mg/mL and 0.00565 mg/mL, respectively (Figure 56D-F).
135
After the evaluation of our window-based software on images of the
enzymatic product for HRP-2 assay on two platforms (i.e., 21-well iCrystal plates
and HTS wells), we advanced to investigate the performance of our MATLABbased application on the same platforms and the RGB pixels for each of the
platforms are illustrated in Figure 57. Figure 57A-C shows the red, green and
blue pixels for HRP-2 assay in buffer for the image of the enzymatic product
taken on 21-well platforms. As shown in Figure 57A, the red pixels increased with
increase in test sample concentration and we noted a value of 0.8 and 0.76 for
0.565 mg/mL and 0.0113 mg/mL samples, respectively. Other samples including
control samples showed red pixels values of 0.78 (0.113 mg/mL), 0.77 (0.0565
mg/mL), 0.77 (0.00565 mg/mL), 0.77 (positive control), cut-off (0.77), and
negative control (0.78). Also, the green pixels showed the same trend as red
pixels, that is, they (green pixels) increased as the sample concentration
increased (Figure 57B). Using the green pixels, we observed the lowest and
highest green pixels values of 0.82 and 0.84 in 0.0565 mg/mL and 0.535 mg/mL
samples respectively. The other test samples exhibited different green pixel
values as follows: 0.83 (0.113 mg/mL), 0.83 (0.0113 mg/mL) and 0.82 (0.00565
mg/mL) and the control experiments: positive, cut-off and negative control
showed green pixels of 0.81, 0.81 and 0.83, respectively.
136
On the other hand, the blue pixels increased with a decrease in sample
concentration. 0.565 mg/mL and 0.00565 mg/mL samples correspondingly,
produced the lowest and highest blue pixel values and the other samples blue
pixel values were distributed in-between the two measurements; 0.113 mg/mL
(0.60 pixels), 0.0565 mg/mL (0.60 pixels) and 0.0113 mg/mL (0.66 pixels) (Figure
57C). Subsequently, we witnessed LLODs of 0.565 mg/mL, 0.0113 mg/mL and
0.00565 mg/mL when we used red, green and blue pixels to analyze through
MATLAB application the sample pictures taken on a 21-well iCrystal plates
(Figure 57A-C).
Also, we investigated the efficacy of the MATLAB application to analyze
the images of the enzymatic product from HRP-2 assay on HTS wells and the
RGB pixels are as displayed in Figures 57D-F. From these RGB pixel results, it is
evident that the red and green pixels were dependent on the concentration of the
test samples. We observed the highest and lowest red pixels values of 1.13
pixels and 1.01 pixels in 0.565 mg/mL and 0.00565 mg/mL test samples. Other
samples recorded different red pixel values: 0.113 mg/mL (1.08 pixels), 0.0565
mg/mL (1.05 pixels), 0.0113 mg/mL (1.03 pixels) (Figure 57D). The highest and
lowest green pixels were 1.1 (0.565 mg/mL) and 1.04 pixels (0.00565 mg/mL),
respectively. Other samples showed the following green pixel values; 0.113
mg/mL (1.08), 0.0565 mg/mL (1.07 pixels), 0.0113 (1.06 pixels) (Figure 57F).
The pictures for enzymatic product for test sample in HTS wells produced blue
pixels same in trend with those taken on 21-well iCrystal plates. The blue pixels
increased with a decrease in test sample concentration. 0.88 pixels and 0.72
139
pixels were the highest and lowest blue pixel values for 0.00565 mg/mL and
0.565 mg/mL samples. The rest of the samples showed the following blue pixel
values: 0.75 pixels (0.113 mg/mL), 0.80 pixels (0.0565 mg/mL), and 0.81 pixels
(0.0113 mg/mL). Each color channel gave the following pixel values for the
control experiments. Red pixels (positive control = 1.00 pixels, cut-off = 1.00
pixels, negative control = 1.04 pixels), green pixels (positive control = 1.05 pixels,
cut-off = 1.03 pixels, negative control = 1.05 pixels) and blue pixels (positive
control = 0.85 pixels, cut-off = 0.88 pixels, negative control = 0.89 pixels). For the
images of the enzymatic product (in HTS wells) for HRP-2 assay, we observed
LLODs of 0.0565 mg/mL (for red color channel), 0.0113 mg/mL (for green color
channel) and 0.00565 mg/mL (for blue color channel).
From the results, the blue pixels produced better LLODs as compared to
red and green color channels in both 21-well iCrystal plates and HTS wells.
Therefore, we sought to find out how the LLOD for the assay compare between
the blue pixels (computed using windows and MATLAB-based software) and UVvis spectroscopy analysis. Figure 58 shows the results for the enzymatic results
for HRP-2 assay from UV-vis spectroscopy and blue pixel computation through
our novel software- MATLAB and window-based software. Figure 58A displays
the colorimetric response for HRP-2 assay using UV-vis spectroscopy and we
observed that the absorbance values for the enzymatic product increased with
increase in sample concentration with the highest and lowest being ~0.6 and
~0.3 for 0.565 mg/mL and 0.00565 mg/mL, respectively. The absorbance values
for other samples were 0.4 (0.113 mg/mL), 0.35 (0.0565 mg/mL), and 0.33
140
(0.0113 mg/mL). The control samples under UV-vis analysis displayed the
following absorbance values: positive control (abs = 0.30), negative control (abs
= 0.23) and cut-off (abs = 0.18).
The blue pixels increased with a reduction in sample concentration (Figure
58B). We witnessed the lowest pixels and highest blue pixels of 0.7 and 0.9 for
0.565 mg/mL and 0.00565 mg/mL samples, respectively (Figure 58B). In the
control experiments, positive control (red solid line) showed the highest pixels
(1.17) followed by cut-off (1.0 pixels) (green solid line) and the negative control
(blue solid line) showed the lowest pixels (0.99). We noted a LLOD of 0.00565
mg/mL for HRP-2 samples on 21-well iCrystal plates (windows-based software)
(Figure 58B).
The blue channel pixels increased with a decrease in sample
concentration in HTS wells where the highest and lowest blue pixels of 0.88 and
0.72 were observed in 0.00565 mg/mL and 0.565 mg/mL samples, respectively
(Figure 58C). The blue pixels for the control samples, represented by horizontal
solid lines, showed higher pixel values than test samples: positive control (pixels
= 0.85), cut-off (pixels = 0.87) and negative control (pixels = 0.89). Under the
blue channel pixels (in our window-based software), we observed a LLOD of
0.00565 mg/mL in the enzymatic product for HRP-2 assay on HTS wells (Figure
58C).
The blue pixels increased with a decrease in sample concentration. That
is, 0.565 mg/mL and 0.00565 mg/mL samples, correspondingly, produced the
lowest (0.48 pixels) and highest (0.67 pixels) blue pixel values. In the other
141
samples, blue pixel values were distributed in-between the two blue pixel
measurements; 0.113 mg/mL (0.60 pixels), 0.0565 mg/mL (0.60 pixels) and
0.0113 mg/mL (0.66 pixels) (Figure 58D). The blue pixels read out for the control
samples are identical to those reported in Figure 58B. Subsequently, we
witnessed LLOD of 0.00565 mg/mL when we used blue pixels to analyze HRP-2
assay’s enzymatic product through MATLAB application with the sample pictures
taken on a 21-well iCrystal plates (Figure 58D).
In Figure 58E, the blue pixels increased with a decrease in test sample
concentration. 0.88 pixels and 0.72 pixels were the highest and lowest blue pixel
values for 0.00565 mg/mL and 0.565 mg/mL samples. The rest of the samples
showed the following blue pixel values: 0.75 pixels (0.113 mg/mL), 0.80 pixels
(0.0565 mg/mL), and 0.81 pixels (0.0113 mg/mL). The blue color channel gave
the following pixel values for the control experiments: positive control = 0.85
pixels, cut-off = 0.88 pixels and negative control = 0.89 pixels. For the images of
the enzymatic product (in HTS wells) for HRP-2 assay and MATLAB analyzed,
we observed LLOD of 0.00565 mg/mL (Figure 58E).
Furthermore, we investigated whether our two application (MATLAB and
windows-based) produced same blue pixel results for images of the enzymatic
product for HRP-2 assay taken on 21-well iCrystal and HTS well platforms.
Figure 59A shows the blue pixels results for images of the enzymatic product for
HRP-2 assay taken on 21-well iCrystal and analyzed using window-based and
MATLAB application. From these results, we observed that there exist variations
in the amount of pixels computed by our windows-based and MATLAB software.
142
In particular, the 0.0565 mg/mL showed the most variation of ~ 0.5 pixels. The
other two samples that displayed a difference of ~ 0.2 pixels in blue pixels were
0.113 mg/mL and 0.00565 mg/mL. On the other hand, Figure 59B displays the
blue pixels for HRP-2 assay images taken on HTS wells and analyzed using
window-based application and MATLAB software. The results depict that the blue
pixel output from the two applications was the same. However, the extent of
pixels we observed in HTS wells were slightly elevated compared to blue pixels
from iCrystal platforms. These observations can be explained by the real-images
on the top panel in Figure 59. The yellow color in the real-picture for HTS wells
(Figure 59B, top panel) appears lighter compared to iCrystal image (Figure 59A,
top panel). Therefore, samples in HTS wells reflected more blue color pixels (~
0.9 normalized pixels) than samples in iCrystal platforms which reflected less
blue pixels (~ 0.7 normalized pixels).
143
4.3.2
Application
of
the
developed
software
in
the
detection
of
environmental samples
As noted earlier, MC-LR is our model environmental contaminant. We
performed an indirect competitive ELISA for MC-LR assay using the commercial
kit and the experiments were completed at room temperature and under
microwave heating (monomode microwave oven) as depicted in the scheme in
Figure 12. Blue channel pixels and absorbance measurements for the enzymatic
product from the MC-LR assay were determined using our software and UV-vis
spectrophotometer, respectively.
Figure 60 shows the colorimetric response and blue pixels output for MCLR assay in buffer for MC-LR antigen standards under microwave heating.
Figure 60A shows the absorbance for MC-LR antigen standards in the range of 0
- 2.5 µg/L, whereas Fig 61B gives absorbance for the same MC-LR antigen
standards at 450 nm wavelength. In both outputs (Figure 60A and B), we
observed that the absorbance of the enzymatic product reduced with increase in
MC-LR antigen concentration. These observations can be attributed to the fact
that higher concentrations of MC-LR antigen allow for fewer MC-LR conjugate to
attach to the antibodies resulting in the generation of a lighter yellow color in the
presence of a substrate. In the contrary, more MC-LR conjugate bind to the
antibodies translating into a darker color when the substrate and stop solution
are present. Therefore, we witnessed absorbance values of 0.35 and 0.10 for 0.1
μg/L and 2.5 μg/L MC-LR antigen concentrations, respectively. Implying that
146
[MC-LR] = 0.1 μg/L produced darker yellow color whereas [MC-LR] = 2.5 μg/L
yielded lighter yellow color for the enzymatic product.
Figure 60C shows the blue pixel output for MC-LR assay in buffer
completed under microwave heating and computed using the MATLAB-based
software. We observed that the blue pixels increased with increase in MC-LR
antigen concentration. That is, there were low blue pixels for [MC-LR] = 0.1 μg/L
(~190 pixels) and higher blue pixels for [MC-LR] = 2.5 μg/L (~220 pixels),
suggesting that darker and lighter yellow color were generated at low and high
MC-LR concentrations, respectively. These observations are further confirmed by
the real color image of MC-LR assay illustrated in Figure 61B. Evidently, the
intensity of yellow color reduces the columns with the highest yellow color
intensity being generated at [MC-LR] = 0 μg/L and the lowest intensity produced
by [MC-LR] = 2.5 μg/L.
Shown in Figure 61 are the colorimetric responses (Figure 61A and B) and
blue pixel computation outputs (Figure 61C) for MC-LR assay standards
performed on commercial HTS wells and completed at room temperature. We
observed that the absorbance for the enzymatic product (yellow in color as
depicted in Figure 61D) reduced with increase in MC-LR antigen concentration;
that is, [MC-LR] = 0.1 μg/L yielded absorbance of 1.5 while [MC-LR] = 2.5 μg/L
produced absorbance of ~0.5 (Figure 61A and B). Subsequently, using our pixel
computation software, we observed that the blue pixels for image in Figure 61D,
increased with increase in MC-LR antigen concentration (i.e., 0.1 μg/L = 60
pixels, 0.5 μg/L = 118 pixels, 1.0 μg/L = 146 pixels, and 2.5 μg/L = 153 pixels)
148
indicative that the yellow color was more intense at low MC-LR concentrations
compared to higher concentration. These observations fit to the same
explanation alluded earlier that lower blue pixels are recorded at darker yellow
samples and vice versa. The intensity of the yellow color for the enzymatic
product was more pronounced for MC-LR assay performed at room temperature
(A = 1.5 for [MC-LR] = 0.1 μg/L) compared to ones finished under microwave
heating (A = 0.35 for [MC- LR] = 0.1 μg/L).
Figure 62 shows the colorimetric response for MC-LR assay in buffer and
in creek water for MC-LR standards and samples (labeled S1 and S2) performed
on iCrystal plates with ITO at room temperature and under iCrystal microwave
system. The absorbance of the colored enzymatic product was analyzed using
UV-vis spectrophotometer. Figure 62A displays the results for MC-LR assay
completed at room temperature (total assay time = 90 mins). The standard curve
was constructed based on five standard samples with concentrations 0 Pg/L, 0.1
Pg/L, 0.5 Pg/L, 1.0 Pg/L and 2.5 Pg/L. Using the standard curve of y = min+(maxmin)/(1+(x/EC50)^(hillslope)) and f = a/(1+exp(-((x-x0)/b)), the estimated
concentrations of creek water spiked with MC-LR toxin (environmental samples)
were determined. The concentration for sample 1 (filled circle) was undefined as
shown by the red-line which cuts across the graph while the estimated
concentration for sample 2 (unfilled circle) was ~ 0.5 Pg/L. Figure 62B shows
results for MC-LR assay completed under microwave heating (total assay time =
15 mins). The estimated concentration for sample 1 (filled circle) and sample 2
(unfilled circle) are ~ 0.5 and ~ 0.7 Pg/L.
149
Figure 63 shows the colorimetric response for MC-LR assay in buffer and in
creek water for MC-LR standards and samples (labeled S1 and S2) performed
on iCrystal plates with ITO at room temperature and under microwave heating
and the pixel computation for the colored enzymatic product analyzed using
MATLAB software. The standard curve was constructed using five standard
samples with concentrations 0 µg/L, 0.1 µg/L, 0.5 µg/L, 1.0 µg/L, and 2.5 µg/L.
The S1 and S2 represent environmental samples (creek water spiked with MCLR toxin). Figures 63A and B displays normalized pixel results for MC-LR assay
completed at room temperature (total assay time = 90 mins) and under iCrystal
mono mode microwave system (total assay time = 15 mins), respectively. At
room temperature, both S1 (filled circle) and S2 (unfilled circle) test samples
were undefined (i.e., their concentrations could not be determined) since they
were out of range of the standard curve (Figure 63A, solid red-lines). On the
other hand, we could only resolve the concentration for the unknown S1 (Figure
63B, top solid red-line). The concentration for the unknown sample S2 was
unresolved (Figure 63B, bottom, solid red-line) using the standard curve.
Since results for the colorimetric response for MC-LR assay shown in
Figures 62 and 63, were unsatisfactory to accurately approximate the
concentration of S1 and S2, we advanced to rearrange the HTS strips in the
manufacturer’s MC-LR assay kit. That is, the layout was equivalent to that of
iCrystal plates and then attached the ITO circles at the bottom for individual wells
as shown in Figure 64 (snippet picture on the top left-hand side). Displayed in
Figure 64 is the colorimetric response for MC-LR assay using iCrystal microwave
152
system (total assay time = 15 mins). The colored enzymatic product was
analyzed using Uv-vis spectrophotometer (Figure 64A) and MATLAB software
(Figure 65B). The standard curve was obtained from 5 standard samples
mentioned in Figures 62. After performing UV-vis analysis, the concentration for
the unknown samples, S1 and S2 was estimated to be ~0.4 ug/L and ~1.2 ug/L,
respectively. (Figure 64A, solid red reference lines), whereas under microwave
heating, the estimated concentrations were undefined and ~0.5 µg/L, for S1 and
S2 samples, respectively. These results implied that the UV-vis analysis
produced accurate results for S2 sample whereas the estimated concentration
for S1 was inaccurate whereas the blue pixel computation by MATLAB software
produced less accurate result for S2 and undefined for S1 samples.
We also evaluated whether the type of camera influenced the outcome of
the assays. Therefore, we took the pictures of the enzymatic colored product of
MC-LR assay using a 12 MP iPhone camera and an 80 MP high resolution
camera (Phase One) and compared the MATLAB computed pixels (blue pixels)
with
MC-LR
assay
absorbance
values
obtained
using
an
UV-vis
spectrophotometer. Figure 65 gives a summary of the colorimetric response for
MC-LR assay in creek water performed in the iCrystal microwave system. Figure
65A shows the absorbance results for MC-LR assay using UV-vis analysis. The
standard curve shows that the low concentration of MC-LR displayed higher
absorbance values compared to higher MC-LR concentrations leading to a
reverse sigmoid standard curve. This can be explained that given that the MC-LR
assay is a competitive assay, low concentration of MC-LR toxin lead to the
153
development a deep color after assay which translates into higher absorbance
values and vice versa. From the UV-vis analysis, the concentrations for the
unknown samples S1 and S2 were estimated to be 0.5 µg/L and 1.0 µg/L,
respectively.
Figures 65B and C shows the normalized blue pixels (range ~0.8 to ~1.07
pixels) for MC-LR assay on images obtained using 80 MP high resolution and
iPhone camera, respectively. In both figures, the pixel values increase with
increase MC-LR concentration for the first three standard samples and then
levels off. This implies that at low (<0.1 µg/L) MC-LR concentration, the color of
the enzymatic product in deep, which translates to fewer blue pixels (~0.8 pixels).
At higher MC-LR toxin concentration (> 1.0 µg/L), a less intense yellow color
develops which gives rise to high blue pixel values (~> 0.9 pixels). In the
determination of the concentration for the unknown samples, S1 and S1, images
taken with HRI (80 MP) and iPhone camera (12 MP) produced the same
concentration: that is, sample S1 = 0.5 µg/L and S2 = 1.0 µg/L. From these
results, we were convinced that the type of camera does not influence the MCLR assay results.
154
Following the evaluation of Windows and MATLAB-based applications in
detecting biological (HRP-2) and environmental (MC-LR) samples using the blue
pixel computation, we were convinced of their ability to accurately identify and
quantify the extent of biomolecules of interest with comparable precision to
standard methods e.g., UV-vis spectral analysis. However, MATLAB application
is only available to licensed users and Windows software is accessible only on
desktop computers. As such, the goal of portability and cost (for MATLAB) was
unmet. To solve the inadequacy, we have developed an iOS-based smartphone
application, which is compatible with iOS devices. The iOS application operates
on the same principle of converting enzymatic colored product into numerical
values for analysis of molecules of interest in environmental and biological
samples.
Figure 66 shows the standard curves developed from the colorimetric
response and the blue pixel computation for HRP-2 assay in buffer
(concentration range: 0.113 mg/mL – 0.000113 mg/mL) completed under iCrystal
microwave system (100 W). Figure 66A displays the real-pictures for the
enzymatic product for HRP-2 assay (concentration range: 0.113 mg/mL0.000113 mg/mL) taken on iCrystal plates (Figure 66 A, top panel) and HTS
wells (Figure66 A, bottom panel). Note: HTS wells attached with ITO dots were
used for assay experiments (Refer Figure 65, left top panel). For iOS and
MATLAB pixels computation, we used images taken on iCrystal plates only
(Figure 66 A, top panel). For both iCrystal and HTS platforms, the wells
containing standard samples and control samples are shown by solid black
158
rectangle enclosure and solid circles (red circles: positive control, blue circles:
negative control, green circles: cut-off), respectively. The results displayed in
Figure 66B were analyzed using UV-vis spectrophotometer and they depict the
absorbance values for the enzymatic product at 450 nm for HRP-2 assay. From
the analysis, we observed that the absorbance increased with increase in the
concentration of HRP-2 antibodies and the control samples-positive (solid red
line), negative (solid blue line) and cut-off (solid green line) showed lower
absorbance values than the standard samples. These observations can be
explained using the real image in Fig 66A, top panel. The wells in the first row
(i.e., 11, 1, and 9) represent the HRP-2 assay samples with highest HRP-2
antibody concentration and they appear to have developed more color compared
to samples with lowest HRP-2 antibody concentration displayed in wells (8, 18,
and 17). Given that the darker the color, the higher the absorbance, we observed
higher absorbance values for samples in wells 11, 1 and 9 than in wells numbers
8, 18 and 17. Conversely, color developed in control samples was of low intensity
except for negative control (well number 12), which translated into small
absorbance values. Figure 66C illustrates the blue pixel computations for HRP-2
assay using iOS (Figure 66C, top panel) and MATLAB (Figure 66C, bottom
panel) applications. The standard curves developed from both applications were
comparable with high and low concentrations of HRP-2 yielding lower and high
blue pixels, respectively.
Figure 67 displays individual computed raw RGB values for HRP-2 assay
(concentration range: 0.113 mg/mL – 0.000113 mg/mL) for optical image (Figure
159
67A) using iOS smartphone. The standard samples occupy the area with a black
solid line while the control samples are represented with solid circles of different
colors (red circle-positive control, green circle- cut-off, and blue circle-negative
control). Each of the fours tabs on top right corner (Figure 67B) of individual
output represent quick buttons to display all channels (black), blue pixels (blue),
green pixels (green) and red pixels (red). The blue pixels data was processed
using statistical software sigma plot and are explained in the foregoing text
(Figure 66C, upper panel). Worth noting is that the smartphone application can
also display the location of the sample and the time the test was performed (Fig
67A, map data). Therefore, the application can map out the prevalence of
disease (HRP-2, malaria) or environmental contaminant (MC-LR) which makes
the surveillance of disease and environmental monitoring possible.
160
CHAPTER 5
CONCLUSION AND FUTURE WORK
5.1
The development of the next generation of circular bioassay
platforms for rapid detection of target analytes using microwaveaccelerated bioassays
We have evaluated the physical stability of STFs on chemically modified
PMMA, Ag NWs on paper and ITO on PET using a model bioassay carried out
using microwave heating to speed up the bioassay steps and at room
temperature (control experiments). SEM and EDS analysis of STFs showed that
a significant loss of silver from the PMMA surfaces, despite their prior chemical
modification with NH2 groups, during the execution of the bioassays was
observed. The extent of loss of silver from the surfaces was pronounced on 10
nm STFs when exposed to microwave heating. A ~7-fold increase in the
colorimetric response from a model bioassay for b-BSA performed on modified
PMMA with 10 nm STFs using microwave heating compared to the colorimetric
signal produced from the same platform at room temperature. On 10 nm STFs
and using microwave heating LLOD of [b-BSA] >10-10 M was observed.
Colorimetric response and LLOD for all the platforms [b-BSA=10-8 M] from
bioassays carried out at room temperature were significantly lower (absorbance
of ~0.2) as compared to bioassays performed using microwave heating
(absorbance of ~0.8). Ag NWs on paper platforms yielded similar results as
compared to STFs on PMMA; however, due to the fragile nature of paper that
results in physical damage during repeated wash steps deemed these surfaces
163
unreliable for using them in bioassays. No or insignificant loss of ITO from PET
was observed during the execution of the bioassays, where an LLOD of [b-BSA]
>10-11 M was measured. On a real-life assay for Ki-67 antigen, the colorimetric
response from bioassay carried out on ITO using microwave heating and at room
temperature was clearly distinguishable from the background response and the
LLOD was determined to be [Ki-67 = 10-10 M].
5.2
Development of software to convert colorimetric responses into
numerical values for the quantification of target analytes
In this study, we have successfully developed software which
appropriately converts the colored enzymatic product into grayscale format and
computes the pixel intensities in the range 0-255 for each assay sample. We
developed the application in three different frameworks: (1) MATLAB-based, (2)
Windows-based and (3) mobile-based with each having three architectural
segments, that is, images upload area, graphical and tabular output sections.
Protein bioassay for b-BSA was used as a model assay to study the
effectiveness of our novel computation tool. The protein assay was performed on
blank iCrystal and iCrystal plates sputtered with 10 nm STFs under microwave
heating. Initially, the real color pictures of the assay were taken on clear-walled
HTS wells and we observed that the center of the wells showed high grayscale
pixel values due to inter-well communication and we minimized this challenge by
painting the walls of the HTS wells black. The sample images were taken in a
room with incident light, a dark room with samples backlit with a LED light
source, and in an outdoor environment. The grayscale pixel intensity increased
164
with decrease in sample concentration. Samples with low color intensity when
converted to grayscale exhibited low grayscale pixels due to the loss of pixels
during the color to grayscale conversion and the presence of a shade over
samples taken in an outdoor environment minimized pixel variation compared to
samples images taken exposed to direct-sunlight. Using our software, we
observed LLOD of [b-BSA] =10-8 M for b-BSA bioassay completed on PMMA+10
nm STFs and blank PMMA and pictures taken in a room with incident light and
dark room, respectively. We observed LLOD of [b-BSA] = 10-7 M for samples
performed on blank PMMA and images obtained in a room with incident light.
5.3
The potential application of circular bioassay platforms and the
software for the rapid assessment of biological and environmental samples
in resource-limited settings.
In this work, we have demonstrated the combined use of new generation
platforms of ITO on PET and hand-held computation devices to diagnose
biological and environmental samples in regions rapidly and efficiently with
limited resources. We considered malaria (HRP-2 antibodies concentration in the
range 0.565-0.000565 mg/mL) and microcystin-LR toxin as representatives for
biological sample and environmental contaminant, respectively. The HRP-2
assay in blood, serum and buffer were performed at room temperature and under
microwave heating (kitchen microwave oven and iCrystal system). The results
were analyzed using our novel software and compared in efficiency with the
standard method-UV-vis spectrophotometry. At room temperature, HRP-2 assay
in buffer and blood yielded more color (Abs >2.0 and >1.0, respectively) as
165
compared to HRP-2 assay in serum (abs > 0.5). LLODs of 0.0113 mg/mL and
0.00565 mg/mL were observed for HRP-2 assays in buffer and serum and blood
respectively. Under microwave heating, HRP-2 in buffer and blood yielded
absorbance values >1.0 and <1.0 and LLOD of 0.0113 mg/mL and 0.00565
mg/mL, respectively whereas HRP-2 assay in serum produced absorbance of <
0.5 and LLOD of 0.00565 mg/mL. Our novel software was efficient in HRP-2
assay samples which developed deep color compared to those that appeared
lighter and we have learnt that there was loss of pixels during the conversion of
colored sample images to grayscale picture format. Also, the samples analyzed
using grayscale pixel intensity computation showed inferior LLOD (<0.113
mg/mL) compared to the results of the same samples under UV-vis analysis
(LLOD = 0.00565 mg/mL).
Following the limitation of using grayscale format to compute pixels for
samples which appear light in color, we advanced to use RGB color scheme to
calculate the pixels values for the colored enzymatic product. We noted that
green and red pixels did not vary significantly from each other and the blue pixels
regulated the lightness of the yellow color, that is, the blue pixels increased with
a reduction in sample concentration. Although, green and red color constituents
showed an increase in pixel intensity with sample concentration, they (red and
green pixels) yielded high LLODs (<0.0113 mg/mL) compared to LLOD observed
when using blue pixels: 0.00565 mg/mL. Subsequently, HRP-2 assays in blood
and serum completed using iCrystal system yielded enzymatic product with more
color compared to HRP-2 assays in buffer.
166
There was no significant difference in output between the window-based
and MATLAB-based application when computing RGB pixels for HRP-2 assays.
At the same time, the samples images obtained from either iCrystal plates
(PMMA platforms) or HTS wells with blackened walls showed similar RGB
output.
5.4
FUTURE WORK
We have successfully demonstrated that new generation platforms –ITO,
and
novel
software
(MATLAB,
Windows-based,
and
smartphone-based
applications) can be used to analyze both biological and environmental samples
with similar accuracy to standard methods such as UV-vis spectrophotometry.
However, we have future considerations to improve on: (1) evaluate the
efficiency of bioassays on thick nanoparticle films ~ 100 nm- 30 Pm, (2) study the
construct of other colors in terms of RGB constituents, (3) enhance the mobile
(iOS and Android) and window-based applications to quantity pixels for other
enzymatic product colors including luminescence, (4) perform pixel computation
using other color models such as HSL and HSV, and (5) integrate the voiceguided instructions to ease sample image processing and analysis.
167
REFERENCES
1.
Abel, B., T.C. Clement, and K. Aslan, Enhancement of enzymatic
colorimetric response by silver island films on high throughput screening
microplates. Journal of immunological methods, 2014. 411: p. 43-49.
2.
Aslan, K., P. Holley, and C.D. Geddes, Microwave-Accelerated MetalEnhanced Fluorescence (MAMEF) with silver colloids in 96-well plates:
Application to ultra fast and sensitive immunoassays, High Throughput
Screening and drug discovery. Journal of immunological methods, 2006.
312(1): p. 137-147.
3.
Knappik, M., et. al., Evaluation of molecular methods to improve the
detection of Burkholderia pseudomallei in soil and water samples from
Laos. Applied and environmental microbiology, 2015. 81(11): p. 37223727.
4.
Edwards-Jones, V., et. al., Rapid discrimination between methicillinsensitive and methicillin-resistant Staphylococcus aureus by intact cell
mass spectrometry. Journal of Medical Microbiology, 2000. 49(3): p. 295300.
5.
Hites, R.A., Gas chromatography mass spectrometry. Handbook of
instrumental techniques for analytical chemistry, 1997: p. 609-626.
6.
Cheung, H.C., Resonance energy transfer, in Topics in fluorescence
spectroscopy. 2002, Springer. p. 127-176.
7.
Paulie, S. and H. Perlmann, Enzyme-Linked Immunosorbent Assay. eLS,
2003.
168
8.
Shapiro, H.M., Practical flow cytometry. 2005: John Wiley & Sons.
9.
van der Wal, F.J., et. al., Bead-based immunoassays. 2014.
10.
Colmenero, J.D., et. al., Multiplex real-time polymerase chain reaction: a
practical approach for rapid diagnosis of tuberculous and brucellar
vertebral osteomyelitis. Spine, 2010. 35(24): p. E1392-E1396.
11.
Keim, P., et. al., Multiple-locus variable-number tandem repeat analysis
reveals genetic relationships within Bacillus anthracis. Journal of
Bacteriology, 2000. 182(10): p. 2928-2936.
12.
Liu, Z., Single nucleotide polymorphism (SNP), in Aquaculture Genome
Technologies. 2007, Blackwell USA. p. 59-72.
13.
Vester, B. and J. Wengel, LNA (locked nucleic acid): high-affinity targeting
of complementary RNA and DNA. Biochemistry, 2004. 43(42): p. 1323313241.
14.
Metzker, M.L., Sequencing technologies—the next generation. Nature
reviews genetics, 2010. 11(1): p. 31-46.
15.
Hobbs, M.M. and A.C. Seña, Modern diagnosis of Trichomonas vaginalis
infection. Sexually transmitted infections, 2013. 89(6): p. 434-438.
16.
Barber, B.E., et. al., Limitations of microscopy to differentiate Plasmodium
species in a region co-endemic for Plasmodium falciparum, Plasmodium
vivax and Plasmodium knowlesi. Malaria journal, 2013. 12(1): p. 1.
17.
Jani, I.V. and T.F. Peter, How point-of-care testing could drive innovation
in global health. N Engl J Med, 2013. 368(24): p. 2319-2324.
169
18.
Murray, C.K., et. al., Update on rapid diagnostic testing for malaria.
Clinical microbiology reviews, 2008. 21(1): p. 97-110.
19.
Altieri, M.F. and M. Camarca, Point of care testing. Clinical Pediatric
Emergency Medicine, 2001. 2(4): p. 275-279.
20.
MOR, M. and Y. WAISMAN, Point-of-care testing: a critical review.
Pediatric emergency care, 2000. 16(1): p. 45-48.
21.
Mohammed, M., Development of rapid analysis system based on
microwave-accelerated bioassay technique for point-of-care applications.
2015, MORGAN STATE UNIVERSITY.
22.
Peredy, T.R. and R.D. Powers, Bedside diagnostic testing of body fluids.
The American journal of emergency medicine, 1997. 15(4): p. 400-407.
23.
Food, U. and D. Administration, FDA approves first over-the-counter
home-use rapid HIV test. Silver Spring, MD: US Food & Drug
Administration, 2012.
24.
Pai, N.P., et. al., Impact of round-the-clock, rapid oral fluid HIV testing of
women in labor in rural India. PLoS Med, 2008. 5(5): p. e92.
25.
Pai, N.P., et. al., Point-of-care testing for infectious diseases: diversity,
complexity, and barriers in low-and middle-income countries. PLoS Med,
2012. 9(9): p. e1001306.
26.
Ast, K., Diagnostic Efficacy of Handheld Devices for Emergency
Radiologic Consultation: Toomey RJ, Ryan JT, McEntee MF, et. al. AJR
Am J Roentgenol 2010; 194: 469–74. The Journal of Emergency
Medicine, 2010. 39(2): p. 273.
170
27.
Nemiroski, A., et. al., Universal mobile electrochemical detector designed
for use in resource-limited applications. Proceedings of the National
Academy of Sciences, 2014. 111(33): p. 11984-11989.
28.
Yetisen, A.K., et. al., A smartphone algorithm with inter-phone
repeatability for the analysis of colorimetric tests. Sensors and Actuators
B: Chemical, 2014. 196: p. 156-160.
29.
Gaster, R.S., D.A. Hall, and S.X. Wang, nanoLAB: an ultraportable,
handheld diagnostic laboratory for global health. Lab on a Chip, 2011.
11(5): p. 950-956.
30.
Antus, B., I. Horvath, and I. Barta, Assessment of exhaled nitric oxide by a
new hand-held device. Respiratory medicine, 2010. 104(9): p. 1377-1380.
31.
Maniscalco, M. and J.O. Lundberg, Hand-held nitric oxide sensor NIOX
MINO® for the monitoring of respiratory disorders. Expert review of
respiratory medicine, 2010. 4(6): p. 715-721.
32.
Abel, B., et. al., Plasmon-enhanced enzymatic reactions: a study of
nanoparticle-enzyme
distance-and
nanoparticle
loading-dependent
enzymatic activity. Nano biomedicine and engineering, 2011. 3(3): p. 184.
33.
Abel, B., Enzymatic signal amplification using plasmonic nanostructures
for applications in bioenvironmental science and engineering. 2015,
MORGAN STATE UNIVERSITY.
34.
Abel, B., et. al., Enhancement of the colorimetric response of enzymatic
reactions by thermally evaporated plasmonic thin films: application to glial
fibrillary acidic protein. Analytical Methods, 2015. 7(3): p. 1175-1185.
171
35.
Abel, B. and K. Aslan, Surface modification of plasmonic nanostructured
materials with thiolated oligonucleotides in 10 seconds using selective
microwave heating. Annalen der Physik, 2012. 524(11): p. 741-750.
36.
Aslan, K., et. al., Metal-enhanced fluorescence: an emerging tool in
biotechnology. Current Opinion in Biotechnology, 2005. 16(1): p. 55-62.
37.
Aslan, K. and C.D. Geddes, A review of an ultrafast and sensitive
bioassay platform technology: microwave-accelerated metal-enhanced
fluorescence. Plasmonics, 2008. 3(2-3): p. 89-101.
38.
Aslan, K., et. al., Microwave-accelerated metal-enhanced fluorescence:
application to detection of genomic and exosporium anthrax DNA in< 30
seconds. Analyst, 2007. 132(11): p. 1130-1138.
39.
Aslan, K., Rapid whole blood bioassays using microwave-accelerated
metal-enhanced fluorescence. Nano biomedicine and engineering, 2010.
2(1): p. 1.
40.
Abel, B., et. al., Metal-Enhanced Fluorescence from Silver Nanowires with
High Aspect Ratio on Glass Slides for Biosensing Applications. The
Journal of Physical Chemistry C, 2014. 119(1): p. 675-684.
41.
Melendez, J.H., et. al., Blind evaluation of the microwave-accelerated
metal-enhanced
fluorescence
ultrarapid
and
sensitive
Chlamydia
trachomatis test by use of clinical samples. Journal of clinical
microbiology, 2013. 51(9): p. 2913-2920.
172
42.
Melendez, J. Ultra Rapid and Sensitive Detection of Neisseria Gonorrhea
and Antimicrobial Resistance Susceptibility. in 2014 National STD
Prevention Conference. 2014: CDC.
43.
Aslan, K., S.N. Malyn, and C.D. Geddes, Microwave-Accelerated Surface
Plasmon-Coupled Directional Luminescence: Application to fast and
sensitive assays in buffer, human serum and whole blood. Journal of
immunological methods, 2007. 323(1): p. 55-64.
44.
Previte,
M.J.,
et.
chemiluminescence:
al.,
Microwave
Quantitative
triggered
protein
metal
enhanced
determination.
Analytical
chemistry, 2006. 78(23): p. 8020-8027.
45.
Addae, S.A., et. al., Rapid and sensitive colorimetric ELISA using silver
nanoparticles, microwaves and split ring resonator structures. Nano
biomedicine and engineering, 2010. 2(3): p. 155.
46.
Mohammed, M. and K. Aslan, Design and proof-of-concept use of a
circular PMMA platform with 16-well sample capacity for microwaveaccelerated bioassays. Nano biomedicine and engineering, 2013. 5(1): p.
10.
47.
Mohammed, M., M.F. Syed, and K. Aslan, Microwave-accelerated
bioassay technique for rapid and quantitative detection of biological and
environmental samples. Biosensors and Bioelectronics, 2016. 75: p. 420426.
173
48.
Mohammed, M. and K. Aslan, Rapid and sensitive detection of p53 based
on DNA-protein binding interactions using silver nanoparticle films and
microwave heating. Nano biomedicine and engineering, 2015. 6(3): p. 76.
49.
Vallet-Regí, M., Ceramics for medical applications. Journal of the
Chemical Society, Dalton Transactions, 2001(2): p. 97-108.
50.
Martinez, A.W., et. al., Patterned paper as a platform for inexpensive, low volume, portable bioassays. Angewandte Chemie International Edition,
2007. 46(8): p. 1318-1320.
51.
Balbi, G.M. and P.A. Hartman, Highly sensitive paper-disc assays for
detecting penicillin in milk. Journal of Food Protection®, 1985. 48(1): p.
16-20.
52.
Hergenrother, P.J., K.M. Depew, and S.L. Schreiber, Small-molecule
microarrays: covalent attachment and screening of alcohol-containing
small molecules on glass slides. Journal of the American Chemical
Society, 2000. 122(32): p. 7849-7850.
53.
Soper, S.A., et. al., Peer Reviewed: Polymeric Microelectromechanical
Systems. Analytical chemistry, 2000. 72(19): p. 642 A-651 A.
54.
Andrady, A.L. and M.A. Neal, Applications and societal benefits of
plastics. Philosophical Transactions of the Royal Society B: Biological
Sciences, 2009. 364(1526): p. 1977-1984.
55.
Abel, B. and K. Aslan, Plasmon-enhanced enzymatic reactions 2:
optimization of enzyme activity by surface modification of silver island
174
films with biotin-poly (ethylene-glycol)-amine. Nano biomedicine and
engineering, 2012. 4(1): p. 23.
56.
Abel, B., et. al., Enhancement of the Chemiluminescence Response of
Enzymatic Reactions by Plasmonic Surfaces for Biosensing Applications.
Nano biomedicine and engineering, 2015. 7(3): p. 92.
57.
Aslan, K. and V.H. Pérez-Luna, Surface modification of colloidal gold by
chemisorption of alkanethiols in the presence of a nonionic surfactant.
Langmuir, 2002. 18(16): p. 6059-6065.
58.
Karakouz, T., et. al., Stabilization of gold nanoparticle films on glass by
thermal embedding. ACS applied materials & interfaces, 2011. 3(4): p.
978-987.
59.
De, G., et. al., Metal nanoparticle doped coloured coatings on glasses and
plastics through tuning of surface plasmon band position. Bulletin of
Materials Science, 2008. 31(3): p. 479-485.
60.
Fan, M., et. al., Silver nanoparticles on a plastic platform for localized
surface plasmon resonance biosensing. Analytical chemistry, 2010.
82(15): p. 6350-6352.
61.
Phadtare, S., et. al., Immobilization of Candida bombicola Cells on FreeStanding Organic-Gold Nanoparticle Membranes and Their Use as
Enzyme Sources in Biotransformations. Biotechnology progress, 2004.
20(6): p. 1817-1824.
175
62.
Henry, A.C., et. al., Surface modification of poly (methyl methacrylate)
used in the fabrication of microanalytical devices. Analytical chemistry,
2000. 72(21): p. 5331-5337.
63.
Kim, S., et. al., Enhancement of electrical conductivity of silver nanowiresnetworked films via the addition of Cs-added TiO2. Nanotechnology, 2015.
26(13): p. 135705.
64.
Ji, Y.-H., et. al., Significantly enhanced electrical conductivity of silver
nanowire/polyurethane
composites
via
graphene
oxide
as
novel
dispersant. Composites Science and Technology, 2016. 132: p. 57-67.
65.
Bonyi, E., et. al., Metal oxide surfaces for enhanced colorimetric response
in bioassays. Colloids and Surfaces B: Biointerfaces, 2017. 154: p. 331340.
66.
Ma, Y., et. al., Direct measurement of propagation losses in silver
nanowires. Optics letters, 2010. 35(8): p. 1160-1162.
67.
Chapman, R. and P. Mulvaney, Electro-optical shifts in silver nanoparticle
films. Chemical physics letters, 2001. 349(5): p. 358-362.
68.
Barnes, W.L., A. Dereux, and T.W. Ebbesen, Surface plasmon
subwavelength optics. Nature, 2003. 424(6950): p. 824-830.
69.
Inwald, E., et. al., Ki-67 is a prognostic parameter in breast cancer
patients: results of a large population-based cohort of a cancer registry.
Breast cancer research and treatment, 2013. 139(2): p. 539-552.
176
70.
Gerdes, J., et. al., Cell cycle analysis of a cell proliferation-associated
human nuclear antigen defined by the monoclonal antibody Ki-67. The
Journal of Immunology, 1984. 133(4): p. 1710-1715.
71.
Scholzen, T. and J. Gerdes, The Ki-67 protein: from the known and the
unknown. Journal of cellular physiology, 2000. 182(3): p. 311-322.
177
APPENDIX I: Features of MATLAB-based Software
Menu Bar for Novel Diagnostic Software
File menu contains functions such as Load image, Open
file, Save as, Export data and Exit. The functions load
image and open file perform the same task as tool icons;
. Export data function makes it possible to transfer
results in .xls file format from a tabulation section to a
folder in a documents gallery. The exit function allows the
user to c lose the novel diagnosis software.
The mode menu affords for the selection to run the assay
using either colored or grayscale pictures.
The tools menu makes it possible for the user to access
icons such as ruler, crosshair, etc.
And the help tab has links with which to use to get more
information about the novel diagnostic software.
The analysis tab permits the user to select the type of
bioassay to examine, for example, Malaria bioassay.
A
3. Use the hairpin icon
, to select singly the sample wells and the control
sample wells and then click or the accept icon
on the tools bar and RUN
button at the bottom of the GUI window. To normalize the pixels for the
experimental samples, select the normalization icon
and click on one of the
blank sample wells and then click Accept icon.
D
then click RUN. A horizontal line that cut across the graph will appear. Samples
above the line are considered negative and those below positive.
H
(4)
(5)
I
(6)
J
APPENDIX III: Procedure for Window-based Application
1. Launch the application in the visual studio software platform.
2. Click on the new file tab
and upload the assay image from a domicile
folder (e.g., documents, desktop). The real picture for HRP-2 assay gets
displayed on the on the left-hand side of the application window.
K
3. Using the mouse cursor, first click the center of a blank sample
and then
the center of each well that you need to analyze and click average pixels tab
. In case there is an error in well selection by the mouse cursor,
use the reset button
to correct the mistake. The output will be normalized
pixels. Note: The area of the well to be evaluated can be adjusted by variation of
the radius parameter
. In cases where column samples or row samples
need to be analyzed individually, click the sample button
after each
computation. In each cycle, the RGB pixel points will be displayed in different
colors.
4. Once the ‘Average pixels button’ is clicked, a result of RGB pixels for each
channel is displayed in a graphical format with pixel values and sample
concentration (in mg/mL) displayed on the y- and x-axis, respectively. Pixel
values for each data point can be established by moving the mouse cursor over
the points.
L
APPENDIX IV: Procedure for Smartphone-based Application
1. Launch the smartphone application using the icon “ARG diagnostics app”.
2. Load the assay picture using the load image icon or take a picture using the
phone camera.
3. Select auto-select button in the controls window to automatically analyze all
the wells then select the RUN button. To normalize the pixels, click the Blank
button and select RUN.
4. Using the mute buttons, silence RED and GREEN pixels.
5. To differentiate between positive and negative samples, select P, N and C
buttons (P-positive control, N-Negative control and C-Cut-off) in that order. P and
N buttons will label the location of each point in the sample data set. However,
the selection of C button draws a horizontal line across the graph. Samples
beyond the cut-off line are considered negative while those below the cut-off line
are regarded as positive. Concurrently, a shade of red and blue appear for the
region with positive and negative samples, respectively.
N
Документ
Категория
Без категории
Просмотров
0
Размер файла
45 167 Кб
Теги
sdewsdweddes
1/--страниц
Пожаловаться на содержимое документа