1

HIGHLY SENSITIVE,
RAPID, RELIABLE, AND AUTOMATIC
CARDIOVASCULAR DISEASE DIAGNOSIS WITH
NANOPARTICLE FLUORESCENCE ENHANCER
AND MEMS

Bin Hong,
a
Junhai Kai,
b
Yongjie Ren,
a
Jungyoup Han,
b
Zhiwei Zou,
b

Chong H. Ahn,
b
and Kyung A. Kang
a


1. INTRODUCTION

Cardiovascular diseases, especially acute myocardial infarction (MI; heart attack),
have been the leading threat to human life.
1
Thus, in the emergency situation, a rapid and
accurate diagnosis of cardiovascular diseases is important to save lives. It can be realized
by rapidly and sensitively measuring cardiac markers that are released from injured
cardiac muscles. Creatine Kinase-MB (CK-MB), myoglobin (MG), and cardiac troponin
I (cTnI) are important markers for early diagnosis of heart attack.
2
B-type natriuretic
peptide (BNP) and C-reactive protein (CRP) are crucial markers for diagnosis and
prognosis of congestive heart failure (CHF) and acute coronary syndromes (ACS),
respectively.
3-4
Our effort is focused on developing highly sensitive, accurate, and
reliable instrument using nanoparticle reagents and Microelectromechanical Systems
(MEMs).
The main challenge in developing biosensor is the low concentrations of cardiac
markers in plasma (only a few tens pico-moles) at an early stage of cardiovascular
diseases.
5
Since our sensing is interrogated by fluorescence, fluorescence enhancement
can, therefore, improve the sensitivity. Nanogold particles (NGPs) hold strong plasmon
polariton field on their surface, which can couple the lone-pair electrons (normally used
for self-quenching) of the fluorophore for fluorescence enhancement.
6-7
Some

a
Bin Hong, Yongjie Ren, and Kyung A. Kang, Department of Chemical Engineering, University of Louisville,
Louisville, KY 40292.
b
Junhai Kai, Jungyoup Han, Zhiwei Zou, and Chong H. Ahn, Department of Electrical
and Computer Engineering and Computer Science, University of Cincinnati, Cincinnati, OH 45221.

B. HONG ET AL.
2

biocompatible solvents were also found to enhance fluorescence, by shifting the
fluorophore excitation/emission wavelength and increasing the amount of trans carbon
double bonds.
6-7
To maximize the enhancement effect, NGPs and the solvent were
combined, forming nanogold particle reagents (NGPRs). According to our previous
results, the mixture of 5 nm sized NGPs coated with 2 nm thick self-assembled
monolayer (5nmNGP-SAM2nm) and 1-butanol has shown to be an excellent enhancer.
7

MEMs improves biosensors by ultra-small sample usage, high sensing consistency,
high compactness, as well as mass production and low energy consumption. Thus, for a
reliable and fully automated sensing performance with a minimal system size, MEMs
was integrated to the sensing system. Our research group has previously developed a
prototype, semi-automatic, four-marker sensing system.
8
Using this system simultaneous
four-cardiac marker quantification was completed within 10 minutes. The average
signal-to-noise (S/N) was also as high as 20, which was excellent.
In this paper, based on our previous system, a more sensitive and accurate
simultaneous four-cardiac marker sensing system with the application of NGPR and
MEMs is reported.


2. MATERIALS, INSTRUMENTS, AND METHODS

2.1. NGP, solvent and NGPR related study

5 nm nanogold particles coated with tannic acid (Ted Pella, Redding, CA) and 16-
mercaptohexadecanoic acid (MHA; Sigma/Aldrich, St. Louis, MO) were used to
synthesize 5nmNGP-SAM2nm by self-assembling MHA on NGP.
6
For the NGPR with
butanol basis, 5nmNGP-SAM2nm was then dispersed in 1-butanol (Sigma/Aldrich).

2.2. Cardiac marker sensors and assay protocol

Human BNP was purchased from Bachem (Torrance, CA) and monoclonal IgGs
against human BNP, from Strategic Biosolutions (Newark, DE). cTnI, MG, CRP from
human heart and their respective monoclonal antibodies were obtained from Fitzgerald
Industries (Concord, MA). The fluorophore, Alexa Fluor

647 (AF647; max.

excitation/emission wavelengths, 649/666 nm) was from Invitrogen (Carlsbad, CA).
Four cardiac marker biosensors were constructed following the protocol established by
Tang et al.
5
The fluorometer with four sensing channels (Analyte 2000
TM
) was purchased
from Research International (Monroe, WA). The main principle of our sensing system is:
The first monoclonal antibodies (1
o
Mab) against the respective marker are covalently
immobilized on the sensor surface. During the assay, the sample is injected to the
sensing chamber and the target marker binds specifically to the 1
o
Mab. All other
unbound molecules are washed away from the sensing chamber. Next, the second
monoclonal antibody tagged with AF647 (AF647-2
o
Mab) is applied to the sensor and
binds to the 1
o
Mab-marker complex, forming sandwich complex. When the bound
AF647s are excited by the light source, the fluorescence is emitted from the sandwich
complex and the intensity is correlated with the marker concentration in samples.
Sensing with NGPR follows the protocol established by Hong and Kang:
6
NGPR is
applied before the sample incubation. During the incubation of the NGPR, the
MULTI-CARDIAC MARKER BIOSENSOR

3


Figure 1. Schematic diagram of hot embossing using PDMS mold and thermal bonding. The plastic substrate
is heated to a temperature above the glass transition temperature (Tg) of the polymer material and a master
PDMS mold is pressed against the molten side of the substrate. The substrate cools as the force is applied and
the pattern on the PDMS mold is transferred to the plastic substrate.


fluorescence signal is measured for the base line. NGPR is also applied after the
incubation of AF647-2
o
Mab and sensor washing. The fluorescence is measured again for
the sample. Then, the difference of these two measurements is correlated to the marker
concentration in sample.

2.3. Microfluidic sensing system utilizing MEMs

Labview
TM
(version 7.1) and data acquisition (DAQ) card (USB-6008, 8 inputs, 12
bits, 10 ks/s, multifunctional I/O) from National Instruments (Austin, TX) were used to
control the entire sensing system. Electronically controllable micro-solenoid pump (12 v,
50 L per stroke, 2 W) and micro-solenoid valves (12 v, 280 mW) from The Lee Co.
(Westbrook, CT) execute microfluidics. Drive circuit with power plug, switch and LED
were customized by our research group. The plastic microchannel network and
serpentine sensing module were microfabricated as follows:
9
Thick photoresist (SU-8) on
nickel disc (diameter, 3 inch) was patterned by UV-LIGA process; Polydimethylsiloxane
(PDMS) was cast on nickel disc for PDMS mold with desired pattern (Fig. 1); Hot
embossing between the plastic wafer and PDMS mold was performed to transfer the
pattern on plastic wafer; Plastic wafers were sliced; Thermal bonding of multiple plastic
wafers was conducted for microchannels. To generate micro-turbulence, bumps or
baffles were added on the upper and bottom sides of a microchannel with a square cross-
section for a serpentine sensing chamber.


3. RESULTS AND DISCUSSION

Our fluorophore mediated, fiber-optic immuno-sensor is a highly sensitive detection
tool. As an example, from our comparative study for BNP quantification, it has a
sensitivity 100 times higher than Enzyme-Linked Immunosorbent Assay (ELISA) (data
not shown). Because of its sensitivity and accuracy, it may be used for various human
disease diagnosis/ prognosis.
5,8,10
However, for rapid and simultaneous multi-cardiac
marker quantification, additional sensitivity improvement is needed, especially for BNP
and cTnI due to their extremely low concentrations in plasma at an early disease stage.

3.1. Cardiac marker sensing using NGPR

As previously stated, 5nmNGP-SAM2nm in 1-butanol was an excellent NGPR. Its
enhancement effect was, therefore, tested in a 3-cm BNP sensor in a smooth capillary
B. HONG ET AL.
4












(a) (b)

Figure 2. Sensing performances of (a) BNP sensor at the BNP sensing range and (b) four cardiac marker
sensors in the microfluidic sensing system without using NGPR and with NGPR. [Experimental conditions: for
(a), 3-cm BNP sensor, 3/4 minutes for BNP and AF647-2Mab incubation, respectively, flow velocity at 1.2
cm/sec, pump frequency at 0.6 Hz, NGPR, 5nmNGP-SAM2nm in 1-butanol, capillary microchannel, automatic
sensing; For (b), size of four sensors, 3 cm, plasma sample with 0.1 ng/mL BNP; 0.7 ng/mL cTnI; 70 ng/mL
MG; and 700 ng/mL CRP, AF647-2Mab mixture, 20 g/ml in PBS buffer for all markers, serpentine sensing
module, other conditions were same as (a).]


microchannel. Figure 2a shows the sensing performance of BNP sensor with and without
the NGPR. With the NGPR, the sensors sensitivity was found to be 410% more than
that without NGPR in average. This NGPR was also tested with four sensors encased in
a four-microchannel serpentine sensing module (Fig. 2b). To examine the enhancement
effect of the NGPR for four different sensors simultaneously, the mixture of four cardiac
markers in plasma was used as a sample. The concentration of each marker was selected
to be at its lower limit in its sensing range, as this condition is most difficult to measure
and the signals are most needed to be enhanced. Results showed that NGPR is able of
increasing the sensitivities of BNP, cTnI, MG, and CRP sensors by 60, 50, 180, and
230%, respectively, indicating that signals from the sensors for the markers with higher
concentration ranges were enhanced more than those for lower concentration markers. It
is probably because relatively larger amount of fluorophore tagged protein complexes
bound on sensor surface due to high concentration of the marker in sample provide more
opportunity of interaction with NGPR for fluorescence enhancement. In our studies,
NGPRs have shown their potential as nano-fluorescence enhancers in various fluorophore
mediated, fiber-optic, biomarker sensors whether the marker is protein (e.g., MG) or
peptide (e.g., BNP).
5,8


3.2. Cardiac marker sensing using MEMs

For automatic sample and fluid delivery, circulation, and washing in our system, a
computer assisted flow control unit was developed (Fig. 3ab). In the unit, there are a
micro-pump, seven micro-valves, and a serpentine sensing module which are connected
by the microchannel network forming an open (washing)/ close (sample and fluid
delivery, circulation) flow pattern. A computer code written by Labview
TM
software was
then added to electronically control these electronic parts for an easy operation.
Therefore, a MEMs was constructed for simultaneous four-cardiac marker quantification
0
100
200
300
400
500
600
700
0 50 100 150 200 250 300
BNP concentration (pM)
S
i
g
n
a
l

i
n
t
e
n
s
i
t
y

(
p
A
)
Without NGPR
With NGPR
0
100
200
300
400
500
600
700
800
900
BNP cTnI MG CRP
S
i
g
n
a
l

i
n
t
e
n
s
i
t
y

(
p
A
)
Without NGPR
With NGPR
MULTI-CARDIAC MARKER BIOSENSOR

5




























with our sensing protocols (Fig. 3c). This system is beneficial not only for a user-
friendly diagnostic tool, but also improving the reliability of the system since it
minimizes human errors. Automated assays showed similar signal intensities to those of
the manual operation (less than 5% discrepancy) but doubled average S/N ratio from 18
to 35 (data not shown). In addition, by using microchannel network, the sample volume
required for each assay was decreased from 1 mL to 300 L.
Effective analyte mass transport from bulk media to sensor surface is important for
mass-transfer-limited, fiber-optic, immuno-biosensing. Convective application of liquid
samples/ reagents to the sensor surface was proven to improve the sensitivity of this type
of sensors.
10
However, the laminar flow in the smooth surfaced capillary channels still
limits the analyte transport.
5
The microchannel with the inner surface that can create
local turbulence may enhance the analyte transport to the surface.
8,11
For this purpose, a
series of bumps (radius, 600 m; height, 400 m; spacing between adjacent bumps, 1200
m) were microfabricated on the inner surface of the microchannel (i.e., serpentine
microchannel) (Fig. 4abc). Out of several bump shapes, half-circular bump was found to
be very effective.
8
By applying this serpentine microchannel, the sensing performance of
3-cm BNP and MG sensors were studied and the results were compared with those using
smooth capillary microchannel. BNP and MG were selected because one has a low
concentration range (26~260 pM) and the other, a high concentration range (4~40 nM).
For BNP sensing (Fig. 4d), serpentine microchannel presented approximately 30~90%
Figure 3. BioMEMs: (a) Schematic
diagram and (b) the top view of the
microfluidic sensing unit: plastic plate
with imbedded microchannel network,
micro-pump, micro-valves, serpentine
sensing module; (c) BioMEMs
components: laptop computer with
LabVIEW control panel, microfluidic
sensing unit with serpentine sensing
module, Analyte 2000 (fluorometer)
with four sensing channels.

2-way microvalves
3-way
microvalve
Micropump
Microchannel
network
Plastic plate Waste outlet
(a)
Optical sensors
Serpentine sensing module
2-way microvalves
Optical sensors
Serpentine
sensing module
(b)
LabVIEW
control panel
Microfluidic
sensing unit
Fluorometer
(c)
B. HONG ET AL.
6


































higher signals than capillary microchannel. While for MG sensor, a slight signal increase
(0~6%) was exhibited (Fig. 4e). It may indicate that serpentine structure showed better
performance for BNP since BNP has low sensing range and, therefore, the mass transport
was more limited than MG.
In order for our sensing system to be used for rapid disease diagnosis, a shorter assay
time is desired. Initially, the incubation times for the sample and the AF647-2
o
Mab with
convection were 3 and 4 min, respectively, which lead to a total assay time of 10 min
including washing and regeneration steps. To increase the sensitivity and minimize the
assay time, NGPR and MEMs were incorporated into the sensing system. The sensing
performance of four sensors was, therefore, studied simultaneously in the four-channel
plastic serpentine sensing module (Fig. 4e) at various incubation times for the sample and
the AF647-2
o
Mab mixture for four markers. For the effect of the sample incubation time
on the sensing performance, plasma sample with four markers at their lower
concentration limits was incubated for 1, 2 or 3 min. For this test, the incubation time for
(b)
Serpentine bumps
4 -channel
sensing
module
Microchannels (c)
Figure 4. Schematic diagram (a) and the top view (b) of a
serpentine microchannel [Structure: a series of half-
circular bumps at 600 m radius, 400 m height and 1200
m spacing between bumps and 1.4 1.4 mm square
cross-section]; and (c) a four-channel serpentine sensing
module; and the sensing performance of (d) BNP and (e)
MG sensors using capillary microchannel () and
serpentine microchannel (). [Experimental conditions: 3-
cm BNP or MG sensor, 3/4 minutes for sample and
AF647-2Mab incubation, respectively, flow velocity at
1.2 cm/sec, pump frequency at 0.6 Hz, NGPR, 5nmNGP-
SAM2nm in 1-butanol , automatic sensing, single channel
serpentine sensing module with the structure same as (a).]



600 m
1200 m


1.4 mm
1.4 mm
400 m
(a)
0
100
200
300
400
500
600
700
800
900
0 50 100 150 200 250 300
BNP concentration (pM)
S
i
g
n
a
l

i
n
t
e
n
s
i
t
y

(
p
A
)
Capillary microchannel
Serpentine microchannel
(d)
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
0 10 20 30 40 50
Myoglobin concentration (nM)
S
i
g
n
a
l

i
n
t
e
n
s
i
t
y

(
p
A
)
Capillary microchannel
Serpentine microchannel
(e)
MULTI-CARDIAC MARKER BIOSENSOR

7






















Figure 5. The effects of (a) sample reaction time and (b) AF647-2Mab reaction time on the sensing
performance of four sensors. [Experimental conditions: four cardiac markers with concentrations at the lower
limit in their sensing ranges, 0.1 ng/ml BNP, 0.7 ng/ml cTnI, 70 ng/ml MG, 700 ng/ml CRP in plasma; sensor
size, 3 cm; incubation time of AF647-2Mab mixture was fixed as 4 min for the effect of sample reaction time,
while sample mixture incubation time was fixed as 3 min for the effect of AF647-2Mab reaction time; flow
velocity, 1.2 cm/sec; NGPR, 5nmNGP-SAM2nm in 1-butanol; serpentine sensing module; automatic sensing.]


the AF647-2
o
Mab mixture remained constant at 4 min (Fig. 5a). Due to the similar
sensing range, results of BNP and cTnI sensors were shown in one figure, while MG and
CRP in another figure. Sensing of cTnI (Fig. 5a1, ) and MG (Fig. 5a2, ) presented
similar exponential increase of signals as the sample reaction time increased. The
reaction between the analyte and its corresponding 1Mab was minimal before 2 min
while increased significantly after 2 min. While for BNP (Fig. 5a1, ) and CRP (Fig.
5a2, ) sensors, signals increased nearly linearly with the increase of the sample
incubation time, indicating a faster reaction for CRP and BNP. Although the signal may
increase after 3 min for all sensors, the signal intensities from four sensors were much
higher than those without using NGPR and serpentine sensing module. When the
reaction time was 2 min with NGPR and serpentine sensing module, the signals were
1.3~6.4 times of the signals without using NGPR or serpentine sensing module.
Therefore, 2 min reaction time was enough for the sample incubation for accurate
quantification of four markers simultaneously. Similarly, with a constant sample
incubation time of 3 min, the effect of the AF647-2Mab incubation time (1, 2, 3, or 4
min) on the sensing performance was studied. All four sensors showed a similar signal
profile with the increase of the AF647-2Mab incubation time (Fig. 5b). At 3 min, the
signal increase was reduced significantly. 2 min showed signal intensities 110~450% of
those by 4 min reaction time without NGPR or serpentine sensing module. Therefore, the
0
100
200
300
400
500
600
700
800
900
0 1 2 3 4
Sample mixture incubation time (minute)
S
i
g
n
a
l

i
n
t
e
n
s
i
t
y

(
p
A
)
MG
CRP
0
100
200
300
400
500
600
700
800
900
0 1 2 3 4 5
AF647-2Mab mixture incubation time (minute)
S
i
g
n
a
l

i
n
t
e
n
s
i
t
y

(
p
A
)
MG
CRP
(a2)
(b2)
0
10
20
30
40
50
60
70
0 1 2 3 4
Sample mixture incubation time (minute)
S
i
g
n
a
l

i
n
t
e
n
s
i
t
y

(
p
A
)
BNP
cTnI
0
10
20
30
40
50
60
70
0 1 2 3 4 5
AF647-2Mab mixture incubation time (minute)
S
i
g
n
a
l

i
n
t
e
n
s
i
t
y

(
p
A
)
BNP
cTnI
(a1)
(b1)
B. HONG ET AL.
8

total assay time was shortened by 3 min from the previous protocol of without using
NGPR and serpentine sensing module.


4. CONCLUSIONS

An automated, fluorophore mediated, fiber-optic, multi-analyte, immuno-sensing
MEMs system was developed to quantify four important cardiac markers simultaneously
in blood plasma. To improve the system sensitivity, fluorescence enhancing NGPR was
applied and the sensitivities of BNP, cTnI, MG and CRP sensors were increased by
approximately 60%, 50%, 180% and 230%, respectively. Serpentine structure was
microfabricated on the surface of the sensing microchamber and the structure could also
improve the sensitivity. MEMs was incorporated to the system for a reliable detection
and user-friendly operation. As a result, the consistency of the system was doubled; the
assay time was shortened to be 7 min and; the sample volume decreased to 300 L.
Our multi-analyte Biosensing MEMs is potential for quantifying any group of human
disease representing biomarkers rapidly, simply, and cost-effectively.


ACKNOWLEDGEMENT

Authors acknowledge the financial supports from Kentucky Science and Engineering
Foundation (KSEF-148-502-03-55) for fluorescence enhancement studies and National
Science Foundation (BES-0330075) for cardiac marker biosensing. Sigma Xi honor
society is acknowledged for Bin Hongs Grants-in-Aid of Research award for NGPs
related studies and the Institute for Molecular Diversity and Drug Design (IMD
3
) at the
University of Louisville for Bin Hongs Graduate Fellowship.


REFERENCES

1. American Heart Association. Heart Disease and Stroke Statistics, Update, 10-12 (2005).
2. F. S. Apple, R. H. Christenson, R. Valdes, A. J. Andriak, K. Mascotti, and A. H.B. Wu, Simultaneous rapid
measurement of whole blood myoglobin, creatine kinase MB, and cardiac troponin I by the triage cardiac
panel for detection of myocardial infarction, Clin. Chem. 45(2), 199-205 (1999).
3. A. S. Maisel, P. Krishnaswamy, H. C. Herrmann, and P. A. McCullough, Rapid measurement of B-type
natriuretic peptide in the emergency diagnosis of heart failure, New Engl. J. Med. 347, 161-167 (2002).
4. M. S. Sabatine , D. A. Morrow, C. P. Cannon, and E. Braunwald, Multimarker approach to risk stratification
in non-ST elevation acute coronary syndromes: Simultaneous assessment of troponin I, c-reactive protein,
and b-type natriuretic peptide, Circ. 105, 1760-1763 (2002).
5. L. Tang, Y. J. Ren, B. Hong, and K. A. Kang, A fluorophore-mediated, fiber-optic, multi-analyte, immuno-
sensing system for rapid diagnosis and prognosis of cardiovascular diseases, J. Biomed. Optics 11, 021011 (2006).
6. B. Hong and K. A. Kang, Biocompatible, nanogold-particle fluorescence enhancer for fluorophore mediated,
optical immunosensor, Biosens. Bioelectron. 21(7), 1333-1338 (2006).
7. K. A. Kang and B. Hong, Biocompatible nano-metal particle fluorescence enhancers, Crit. Rev. Eukar. Gene
Expres. 16(1), 45-60 (2006).
8. L. Tang, 2005. Multi-analyte, fiber-optic immuno-biosensing system for rapid disease diagnosis: model
systems for anticoagulants and cardiac markers. Dissertation. Chemical Engineering, University of
Louisville, Louisville, KY
9. Y. Sohn, J. H. Kai, C. H. Ahn, Protein array patterning on Cyclic Olefin Copolymer (COC) for disposable
protein chip, Sensor Lett. 2, 171-174 (2005).
MULTI-CARDIAC MARKER BIOSENSOR

9
10. L. Tang, H. J. Kwon, and K. A. Kang, Theoretical and experimental analysis of analyte transport in a fiber
optic, protein C immuno-biosensor, Biotech. Bioeng. 88, 869-879 (2004).
11. J. S. Rossier, G. Gokulrangan, H. H. Girault, S. Svojanovsky, G. S. Wilson, Characterization of protein
adsorption and immunosorption kinetics in photoablated polymer microchannels, Langmuir 16, 8489-
8494 (2000).