Clinical Chemistry 54:9
1481–1488 (2008)
Disposable Reagentless Electrochemical
Immunosensor Array Based on a
Biopolymer/Sol-Gel Membrane for
Simultaneous Measurement of Several
Tumor Markers
Jie Wu,1 Feng Yan,2* Xiaoqing Zhang,3 Yuetian Yan,1 Jinhai Tang,2 and Huangxian Ju1,3*
BACKGROUND: A reagentless sensor array for simulta-
neous multianalyte testing (SMAT) may enable accu-
rate diagnosis and be applicable for point-of-care
testing. We developed a disposable reagentless immu-
nosensor array for simple immunoassay of panels of
tumor markers.
METHODS: We carried out SMAT with a direct capture
format, in which colloidal gold nanoparticles with
bound horseradish peroxidase (HRP)-labeled antibod-
ies were immobilized on screen-printed carbon elec-
trodes with biopolymer/sol-gel to trap their corre-
sponding antigens from sample solution. Upon
formation of immunocomplex, the direct electro-
chemical signal of the HRP decreased owing to increas-
ing spatial blocking, and the analytes could be simulta-
neously determined by monitoring the signal changes.
RESULTS: The proposed reagentless immunosensor ar-
ray allowed simultaneous detection of carcinoma anti-
gen 153, carcinoma antigen 125, carbohydrate antigen
199, and carcinoembryonic antigen in clinical serum
samples in the ranges of 0.4 –140 kU/L, 0.5–330 kU/L,
0.8 –190 kU/L, and 0.1– 44 ␮g/L, respectively, with de-
tection limits of 0.2 kU/L, 0.5 kU/L, 0.3 kU/L, and 0.1
␮g/L corresponding to the signals 3 SD above the mean
of a zero standard. The interassay imprecision of the
arrays was ⬍9.5%, and they were stable for 35 days. The
positivity detection rate of panels of tumor markers
was ⬎95.5% for 95 cases of cancer-positive sera.
CONCLUSIONS: The immunosensor array provides a
SMAT with short analytical time, small sampling vol-
1
Key Laboratory of Analytical Chemistry for Life Science (Ministry of Education of
China), Department of Chemistry, Nanjing University, Nanjing, P.R. China;
2
Jiangsu Institute of Cancer Prevention and Cure, Nanjing, P.R. China; 3 Key
Laboratory of Laboratory Medical Diagnostics (Ministry of Education of China),
Department of Laboratory Medicine, Chongqing Medical University, Chongqing,
P.R. China.
* Address correspondence to these authors at: Key Laboratory of Analytical
Chemistry for Life Science (Ministry of Education of China), Department of
Chemistry, Nanjing University, Nanjing 210093, P. R. China. Fax ⫹86-25-
Automation and Analytical Techniques
ume, no need for substrate, and, no between-electrode
cross-talk. This method not only proved the capability
of the array in point-of-care testing, but also allowed
simultaneous testing of several tumor markers.
© 2008 American Association for Clinical Chemistry
Cancer is one of the leading causes of mortality, and
early clinical diagnosis is crucial for successful treat-
ment of the disease. Many immunosensors and immu-
noassay methods have been developed for the determi-
nation of a single tumor marker, whose concentration
in human serum is associated with the stages of tumors
(1– 4 ). Because many cancers express 1 marker [e.g.,
breast cancer is associated with carcinoma antigens 153
and 125 (CA 153 and CA 125)4 and carcinoembryonic
antigen (CEA)], and concentrations of several tumor
markers often increase in the serum of a patient, accu-
rate simultaneous multianalyte test (SMAT) of combi-
nations of tumor markers may improve the diagnosis
of certain types of tumor (5– 8 ).
SMAT may offer a shorter analytical time, higher
sample throughput, lower sampling volume, and lower
cost per assay compared with traditional single-analyte
tests (9, 10 ). Thus, multilabel assays and spatially re-
solved assay systems have been developed as the main
modes to perform SMAT (11–25 ). Application of the
multilabel assays has been limited by difficulty in accu-
rate quantification due to different optimal assay con-
ditions and the signal overlap of different labels (11–
13 ). Although a set of substrate zone-resolved
techniques have been proposed to overcome these
83593593; e-mail hxju@nju.edu.cn or yanfeng2007@sohu.com.
Received December 19, 2007; accepted June 9, 2008.
Previously published online at DOI: 10.1373/clinchem.2007.102350
4
Nonstandard abbreviations: CA 153, carcinoma antigen 153; CA 125, carcinoma
antigen 125; CEA, carcinoembryonic antigen; SMAT, simultaneous multianalyte
test; HRP, horseradish peroxidase; SPCE, screen-printed carbon electrode; POCT,
point-of-care testing; CA 199, carbohydrate antigen 199; SEM, scanning elec-
tron micrograph; DPV, differential pulse voltammetry.
1481
drawbacks (14, 15 ), the restriction in the number of
available labels still greatly limits their application.
Spatially resolved assays with a single label seem
well suited for performing SMAT, and optical SMAT,
relying on fluorescence emission and optical reflec-
tance, has been developed into mature technology.
Optical SMAT, however, often needs an expensive ar-
ray detector, such as a charge-coupled device camera
(16, 17 ). Electrochemical array, which is distinguished
by its convenient miniaturization for high-throughput
systems, low assay cost, and absence of sophisticated
and expensive array detectors, shows promising appli-
cation in cancer screening (18, 19 ). Electrochemical
cross-talk caused by diffusion of the detectable enzy-
matic products is the main problem in the fabrication
of electrochemical array. To solve this problem, many
approaches have been followed. One approach, for
example, is to ensure that the distance between adja-
cent electrodes is larger than the diffusion distance of
enzymatic product (20 –23 ), but such an approach con-
flicts with the goal of miniaturization. Another simple
method to completely avoid the electrochemical cross-
talk can be achieved by immobilizing the electron-
transfer mediator on an individual immunosensor to
shuttle electrons (24, 25 ), but this approach requires
the addition of hydrogen peroxide, leading to limited
practical application.
A reagentless electrochemical immunosensor is an
attractive strategy (26 ). In our previous work, we pre-
pared several reagentless immunosensors by using
sol-gel matrix to immobilize immunoreagents and
detected the direct electron transfer of labeled enzyme,
horseradish peroxidase (HRP) (27–31). To achieve SMAT,
this work further fabricated a reagentless immunosen-
sor array by individually embedding 4 kinds of HRP-
labeled antibody-modified gold nanoparticles in a
newly designed biopolymer/sol-gel matrix formed on
screen-printed carbon electrodes (SPCEs), where the
HRP-Ab-Au nanoparticles were limited in the holes of
the biopolymer/sol-gel film. Chitosan, a biopolymer
with excellent film-forming ability, biocompatibility,
nontoxicity, and high mechanical strength, acted as the
adhesion frame in the synthesis of the sol-gel and made
the electrical communication between redox sites of
the enzyme and sensing surface easier due to the coop-
erative effort of chitosan and sol-gel matrix. The pres-
ence of gold nanoparticles accelerated the electron
transfer between immobilized HRP and the electrode
and increased the hole size for improving the perme-
ability of the sol-gel matrix so that the antigens in so-
lution could easily penetrate into the sol-gel film for
immunoreaction. Upon formation of immunocom-
plexes, the electrochemical responses decreased due to
increasing spatial blocking, leading to the reagentless
immunosensing to corresponding antigens without
1482 Clinical Chemistry 54:9 (2008)
cross-talk. The proposed electrochemical immunosen-
sor array had high analyte throughput, showed accept-
able comparability to conventional methods for mea-
suring several tumor markers, could be fabricated with
mass production techniques, and thus provided the
potential for application in point-of-care testing (POCT).
Materials and Methods
REAGENTS
We purchased CA 153, CA 125, carbohydrate antigen
199 (CA 199), and CEA ELISA kits from CanAg Diag-
nostics AB. They consisted of a series of CA 153, CA
125, CA 199, and CEA standard solutions with concen-
trations from 0 –250 kU/L, 0 –500 kU/L, 0 –240 kU/L,
and 0 –75 ␮g/L, respectively, and the stock solutions
of HRP-labeled CA 153, CA 125, CA 199, and CEA
monoclonal antibodies with concentrations of 50, 30,
40, and 60 mg/L. We purchased bovine serum albu-
min, chitosan (molecular weight approximately 1 ⫻
106 Da; approximately 85% deacetylation), and
(3-aminopropyl)triethoxysilane from Sigma-Aldrich
Chemical Co. Tetraethoxysilane (analytical reagent
grade) was from Shanghai Chemical Company. All
other reagents were of analytical reagent grade and
used without further purification. We used 0.1 mol/L
Tris-HCl (pH 7.2) containing 1% bovine serum al-
bumin blocking buffer. By mixing the stock solutions
of NaH2PO4 and Na2HPO4 and adjusting the pH with
0.2 mol/L NaOH and H3PO4, we prepared 0.2 mol/L
PBS of various pHs. Doubly distilled water was used
throughout the experiments. Serum specimens from
Jiangsu Institute of Cancer Prevention and Cure were
stored at 4 °C.
APPARATUS
We performed electrochemical measurements on an
Epsilon electrochemical analyzer (Bioanalytical Sys-
tems Inc.) with a 3-electrode configuration, in which 4
working electrodes were connected with a switch. We
obtained scanning electron micrographs (SEMs) of the
working electrode surfaces using a JSM-5610LV scan-
ning electron microscope (JEOL), and we measured
concentrations of the tumor markers in sera with an
electrochemiluminescent analyzer used for routine
clinical laboratory testing (Elecsys 2010; Roche).
FABRICATION OF IMMUNOSENSORS ARRAY
Construction of the multianalyte immunosensor array
is shown in Fig. 1. The 6-electrode array was fabricated
as reported (32 ), and it contained 4 graphite working
electrodes (2 mm in diameter, 1 mm edge-to-edge sep-
aration), 1 graphite auxiliary electrode, and 1 Ag/AgCl
reference electrode. All working electrodes shared the
same reference and auxiliary electrodes. The insulating
Reagentless Electrochemical Immunosensor Array
Fig. 1. Schematic diagrams of immunosensors array and multianalyte electrochemical immunoassay system.
(a), Nylon sheet, (b), silver ink, (c), graphite auxiliary electrode, (d), Ag/AgCl reference electrode, (e), graphite working electrode,
(f), insulating dielectric.
layer printed around the working area constituted a
reservoir of the electrochemical microcell.
Chitosan solution (1% by weight) was prepared by
ultrasonically dissolving chitosan powder in 1% acetic
acid. We prepared the biopolymer/sol-gel by mixing
50 ␮L (3-aminopropyl)triethoxysilane, 25 ␮L tetra-
ethoxysilane, 10 ␮L of 10 mmol/L HCl as catalyst, and
500 ␮L of 1% acetic acid and 200 ␮L of 1% (wt) chi-
tosan solution in a small test tube under stirring for
5– 6 min at room temperature. We optimized these
ratios of (3-aminopropyl)triethoxysilane, tetraethoxy-
silane to chitosan according to the direct electrochem-
ical signal of the labeled HRP. We prepared 24-nm di-
ameter colloidal gold nanoparticles as described (33 )
and HRP-labeled CA 153, CA 125, CA 199, and CEA
monoclonal antibody–modified gold nanoparticles by
mixing the as-prepared nanoparticles with the HRP-
labeled antibodies at the same volume ratio of 1:1. After
these mixtures were stored at 4 °C for 12 h, they were
mixed with biopolymer/sol-gel at the volume ratio of
2:1. We then dropped 0.5 ␮L of these mixtures indi-
vidually on the working electrodes and dried them
under ambient conditions for 2–3 h to form an im-
munosensor array. The immunosensor array was in-
cubated with blocking buffer for 20 min to block the
sites against nonspecific adsorption (which has been
accepted as a common blocking method in immuno-
sensor fabrication), thoroughly rinsed with doubly dis-
tilled water, and stored in air before use.
ASSAY PROTOCOL
We carried out SMAT with an assay format of direct
capture, in which the HRP-labeled antibodies immo-
bilized on working electrodes reacted with the antigens
in sample solution, respectively (Fig. 1). The electro-
chemical response was derived from the direct reduc-
tion of immobilized HRP from its resting state [Fe(III)].
Formation of the HRP-Ab/Ag complex blocked elec-
tron transfer between HRP and electrode, leading to
signal reduction, which was proportional to the amount
of formed immunocomplex. The amounts of immu-
nocomplexes produced on immunosensor surfaces de-
pended on the concentration of antigens in standard or
sample solutions applied to the electrode. The incuba-
tion step of 40 min could be performed in batch by
dropping 20-␮L mixtures of standard CA 153, CA 125,
CA 199, and CEA solutions or samples on the entire
immunosensor array at room temperature and a rela-
tive humidity of 100% to avoid evaporation of solvent.
After the residual of the incubation solution was re-
moved with doubly distilled water, the immunosensor
arrays were detected in 50 ␮L anaerobic 0.2 mol/L PBS,
pH 6.9, as supporting electrolyte with differential pulse
voltammetry (DPV) from ⫺100 to ⫺800 mV (vs Ag/
AgCl) at pulse amplitude of 50 mV and pulse width of
50 ms in nitrogen atmosphere. The dissolved oxygen
could be reduced on the underlying carbon electrode in
the potential window used, and this effect was pre-
cluded by placing the system in a nitrogen atmosphere.
When performing the POCT, the reduction of dis-
solved oxygen could be precluded by adding Na2SO3
into the detection solution to replace the nitrogen
atmosphere.
Results and Discussion
CHARACTERIZATION OF SOL-GEL MEMBRANES ON SPCES
SPCEs are one of the most desirable techniques for the
development of POCT. Immobilization of proteins on
SPCEs, however, is more difficult than on graphite
Clinical Chemistry 54:9 (2008) 1483
 Fig. 2. SE images of bare SPCE (A), sol-gel matrix (B), and biopolymer/sol-gel membrane–modified SPCE (C) at 2
different magnifications.

electrodes owing to the insulating polymers mixed in
printing inks. Furthermore, the randomly distributed
carbon particles produce inhomogeneous and uneven
surface of SPCE (Fig. 2A), leading to irreproducibility
of sensor fabrication. The chemical modification of
the SPCE surface with different biocompatible mem-
branes provides a way to prepare immunosensors
with good reproducibility (34 ). As shown in Fig. 2B,
however, when the organically modified silicate sol-
gel membrane, a commonly used matrix for protein
immobilization on graphite electrode (30, 31 ), was
formed on the SPCE surface, the membrane was unsta-
ble and showed some fissures, which resulted in mem-
brane falling off from the SPCE surface. For preparing
an entirely uniform and stable membrane, chitosan
was introduced into the sol-gel matrix as the adhesion
frame, and the formed biopolymer/sol-gel membrane
on the SPCE surface showed homogeneous porous
structure (Fig. 2C). It provided good biocompatibility
and high capability for protein loading.
HRP-antibody conjugate increased 2.1-fold (curve e,
Fig. 3). The cyclic voltammetric experiments at differ-
ent gold electrodes showed the same appearance, and
upon incorporation of gold nanoparticles into the
biopolymer/sol-gel at SPCEs, the reduction peak cur-
rent at the same scan rate increased 1.98-fold (see
Supplemental Fig. 1 in the Data Supplement that accom-
panies the online version of this article at http://www.
clinchem.org/content/vol54/issue9). Thus the Au nano-
particles could accelerate the direct electrochemistry of
HRP to further amplify the detectable signal. This peak
was also 2.7 times higher than that of HRP–anti-CA
153-Au nanoparticles/sol-gel modified SPCE (curve f,
Fig. 3), indicating the positive effects of chitosan with
good biocompatibility and hydrophilicity (35 ), which

ELECTROCHEMICAL BEHAVIOR OF LABELED HRP IN

BIOPOLYMER/SOL-GEL
The DPV curves of both the bare and biopolymer/
sol-gel modified SPCEs in 0.2 mol/L PBS, pH 6.9, did
not show any detectable signal in the applied poten-
tial window (Fig. 3). After we embedded 0.17 ␮L of
50 mg/L HRP–anti-CA 153 in the biopolymer/sol-gel,
the modified SPCEs displayed a sensitive peak around
⫺540 mV (vs Ag/AgCl) (curve d, Fig. 3), which was
close to the reduction peak potential of HRP/biopoly-
mer/sol-gel prepared with 0.17 ␮L of 2.0 mg/L HRP
(curve c, Fig. 3), indicating the direct electron transfer Fig. 3. DPVs of bare SPCE (a), biopolymer/sol-gel (b),
between electrode and the labeled HRP with regard to HRP/biopolymer/sol-gel (c), HRP–anti-CA 153/biopolymer/
Fe(III)–Fe(II) conversion. The small difference of sol-gel (d), HRP–anti-CA 153–Au nanoparticles/biopolymer/
peak potentials between HRP–anti-CA 153 and HRP sol-gel (e), HRP–anti-CA 153–Au nanoparticles/sol-gel
resulted from the change of microenvironment around modified SPCE in pH 6.9 PBS (f), and panel e in pH 6.9
HRP molecules because of the presence of antibody. In PBS after incubation in 20 ␮L of 50 kU/L CA 153 at room
the presence of gold nanoparticles in the biopolymer/ temperature for 40 min (g).
sol-gel, the reduction peak of the equal amount of

1484 Clinical Chemistry 54:9 (2008)

Reagentless Electrochemical Immunosensor Array
Fig. 4. Dependences of DPV responses of immu-
nosensors on incubation time (A) and pH of detection
solution (B) for CA 153, CA 125, CA 199, and CEA.
enhanced water uptake and swelling of the film and led
to better permeability of the film for the transfer of
counter ions to neutralize the charge change during the
redox process and a favorable microenvironment for
electron hopping or electron self-exchange between
immobilized HRP molecules (36 ). Thus electron
transfer kinetics and direct electrochemical signal in-
creased. After the modified SPCE was incubated with
CA 153, the direct electrochemical signal decreased
markedly due to the increased barrier that resulted
from the formation of immunocomplex (curve g, Fig.
3), leading to a reagentless immunosensing method for
antigen detection.
OPTIMIZATION OF IMMUNOSENSOR PERFORMANCE
The formation of immunocomplex depended on the
incubation temperature and time. For the sake of con-
venient manipulation, the incubation step was per-
formed with 20 ␮L antigen solution or the mixture of
antigens for SMAT at room temperature, after which
the DPV response of the labeled HRP decreased with
increasing incubation time and reached a relatively
stable value at 30 – 40 min (Fig. 4A), indicating satu-
rated formation of immunocomplex in the membrane.
Thus, 40 min was chosen as the optimal incubation
time for SMAT.
The acidity of detection solution also exerted a
marked effect on the direct electrochemical signal of
the immobilized HRP. The maximum responses for
CA 153, CA 125, and CEA occurred at pH 6.9, whereas
the optimal response for CA 199 occurred at pH 7.3
(Fig. 4B). Thus pH 6.9 was selected as the optimal pH
value for the SMAT.
EVALUATION OF CROSS-REACTIVITY
The immunosensing method was based on monitoring
the direct electrochemical signal of the labeled HRP,
and no substrate or peroxide was added; thus electro-
chemical cross-talk that generally comes from the dif-
fusion overlap of signal molecule on 1 electrode to
neighboring electrodes was not seen in the proposed
system. The lack of cross-talk was an advantage for fur-
ther miniaturization of the array, as the distance be-
tween neighboring electrodes could be reduced.
The cross-reactivity between the HRP-labeled
antibody immobilized on the SPCE and other noncor-
responding tumor markers at 1 immunosensor or be-
tween 1 tumor marker and other noncorresponding
HRP-labeled antibodies at the other 3 immunosensors
was another factor affecting the performance of the im-
munosensor array for SMAT. The experimental results
indicated that only the immunosensor corresponding
to the tumor marker in the incubation solution showed
a sharp decrease of direct electrochemical signal, and
others could retain at least 90% of their original re-
sponses in the detectable concentration ranges. Thus
the cross-reactivity at the array was negligible, suggest-
ing that the 4 tumor markers could be assayed individ-
ually in a single run without interference from each
other.
SMAT OF 4 TUMOR MARKERS
After the immunocomplexes were formed on the im-
munosensor array with an incubation step, the SMAT
was performed on a single-channel potentiostat with a
sequential detection format without interval between 2
assays. Under optimal conditions, with the increasing
concentrations of tumor markers, the DPV peak cur-
rents of the immunosensors decreased (Fig. 5), which
resulted from the increasing amount of immuno-
complexes formed and the increasing barrier to the
electron transfer. The calibration curves were obtained
with several sensor arrays to measure the detection sig-
nals corresponding to a series of mixtures of the 4 an-
tigens at different concentrations. The calibration
curves showed linear relationships in the concentra-
tion ranges of 0.4 –140 (CA 153, R2 ⫽ 0.9916), 0.5–330
(CA 125, R2 ⫽ 0.9959), 0.8 –190 kU/L (C〈 199, R2 ⫽
0.9936), and 0.1– 44 ␮g/L (CEA, R2 ⫽ 0.9930), re-
spectively, in the incubation solution with the slopes
of ⫺1.63, ⫺1.03, and ⫺1.95 nA/(kU/L) and ⫺3.81 nA/
(␮g/L). The detection limits corresponding to the sig-
nals 3 SD above the mean for a zero standard for the 4
analytes were 0.2, 0.5, and 0.3 kU/L and 0.1 ␮g/L, re-
spectively. The proposed detection ranges were suffi-
cient for detection of CA 153, CA 125, C〈 199, and
CEA concentrations in clinical samples with cutoff
values of 25, 35, and 37 kU/L and 3 ␮g/L for a positive
result.
APPLICATION IN DETECTION OF SERUM TUMOR MARKERS
The 4 tumor markers in human serum samples were
detected after the immunosensor array was incubated
Clinical Chemistry 54:9 (2008) 1485
 Fig. 5. DPVs (A, C, E, and G) and calibration curves (B, D, F, and H) for simultaneous immunoassay of CA 153 (A, B),
CA 125 (C, D), CA 199 (E, F), and CEA (G, H).
with 20 ␮L serum samples. When the decrease of peak
current went beyond the linear range shown in Fig. 5,
appropriate dilution of the serum samples was neces-
sary. The concentrations of the 4 tumor markers were
measured by this approach in serum specimens of
95 patients with pathologically diagnosed cancers, in-
cluding 53 specimens with colorectal or gastric cancer,
22 with epithelial ovarian cancer, 8 with breast cancer,
and 12 with lung cancer, and 20 serum specimens from
healthy volunteers. The SMAT results were compared
with results produced by a commercial electrochemi-
luminescent single-analyte measurement system. Only
6 of a total 215 results (2.8%) showed some difference
from the results of the commercial assay. When con-
sidering the other 209 results using orthogonal regres-
sion analysis, the plots of the SMAT results vs the
control results for CA 153, CA 125, CA 199, and CEA
gave the slopes of 0.99, 0.97, 1.04, and 0.99, intercepts
of 0.42 kU/L, 6.1 kU/L, ⫺12.6 kU/L, and 0.55 ␮g/L, Sy兩x
of 3.88 kU/L, 21.2 kU/L, 83.7 kU/L, and 23.6 ␮g/L,
and R2 values of 0.990, 0.998, 0.997, and 0.999, respec-
tively (see Supplemental Fig. 2 in the online Data Sup-
plement). The regression parameters indicated accept-
able agreement of this proposed method with an
established clinical laboratory method.
When the SMAT was used for POCT, nitrogen at-
mosphere was inconvenient. We used a detection solu-
tion of 0.2 mol/L PBS, pH 6.9, containing 0.1 mol/L
1486 Clinical Chemistry 54:9 (2008)
Na2SO3 without presence of nitrogen atmosphere for
POCT. We examined the relative errors of the obtained
antigen concentrations in 20 standard solutions with
different concentrations from those using 0.2 mol/L
PBS, pH 6.9, as detection solution in the presence of
nitrogen atmosphere, and the maximum value was
6.0%. Thus, this SMAT system could be used for POCT
by using 0.2 mol/L PBS, pH 6.9, containing 0.1 mol/L
Na2SO3 as detection solution.
The positive detection rates in 115 serum samples
using the SMAT method are shown in Table 1. Accord-
ing to the panels of tumor markers for combination
diagnosis, CA 199 –CEA panel for 53 cases of colorectal
and gastric cancer, CA 153–CA 125–CEA panel for 8
cases of breast cancer, and CA 199 –CEA panel for 12
cases of lung cancer showed positive detection rates of
100%, whereas CA 125–CA 199 –CEA panel for 22
cases of epithelial ovarian cancer showed positive de-
tection rates of 95.45%. As controls, 20 healthy volun-
teers showed 2 positive results. Thus, the present im-
munosensor array appears promising for SMAT of CA
153, CA 125, C〈 199, and CEA concentrations in clin-
ical applications.
REPRODUCIBILITY AND STABILITY OF THE ARRAYS
Because 0.5 ␮L sol-gel matrix was used to fabricate the
sensor, the fabrication reproducibility was first exam-
ined by measuring the direct electrochemical signals in
Reagentless Electrochemical Immunosensor Array

Table 1. Positivity detection rates of clinical sera.

Positive cases, Positivity detection rate,

Sample n Associated tumor markers n %

Colorectal or gastric cancer 53 CA 199, CEA 53a 100

Epithelial ovarian cancer 22 CA 125, CA 199, CEA 21a 95.5
Breast cancer 8 CA 153, CA 125, CEA 8a 100
Lung cancer 12 CA 199, CEA 12a 100
b
Normal serum 20 CA 153, CA 125, CA 199, CEA 2 10
a
All concentrations obtained with the proposed method are higher than the clinical cutoff values of individual tumor markers.
b
Cases with concentrations of ⱖ1 tumor markers are higher than their clinical cutoff values.

the absence of analyte. The CVs for the responses of
HRP labeled to CA 153, CA 125, CA 199, and CEA
antibody in 10 immunosensor arrays were ⬍5.1%,
2.5%, 5.6%, and 5.1%, respectively (see Supplemental
Fig. 3 in the online Data Supplement), indicating ac-
ceptable fabrication reproducibility. We examined the
interassay imprecision of the immunosensor arrays
with 2 panels of tumor markers at various concentra-
tions. Each panel was measured 5 times using 5 arrays.
The CVs were 4.6%, 7.8%, 9.5%, and 7.1% for 13.5
kU/L CA 153, 6.7 kU/L CA 125, 8.0 kU/L CA 199, and
5.0 ␮g/L CEA, and 3.3%, 5.6%, 7.5%, and 3.7% for 20.9
kU/L CA 153, 24.9 kU/L CA 125, 32.0 kU/L CA 199,
and 11.3 ␮g/L CEA. These results indicated acceptable
imprecision and fabrication reproducibility.
The immunosensor arrays could be stored in dry
air at room temperature. The DPV responses were
87.9%, 89.6%, 97.6%, and 94.7% of initial responses
for CA 153, CA 125, C〈 199, and CEA after a storage
period of 35 days. Thus, the storage stability of the im-
munosensor arrays was acceptable, and the proposed
array was suitable for clinical diagnostics.
In comparison with previous reports (24, 25 ), this
array avoids the addition of mediator to shuttle elec-
1. Soper SA, Brown K, Ellington A, Frazier B, Garcia-
Manero G, Gau V, et al. Point-of-care biosensor
systems for cancer diagnostics/prognostics. Bio- 6.
sens Bioelectron 2006;21:1932– 42.
2. Rasooly A, Jacobson J. Development of biosen-
sors for cancer clinical testing. Biosens Bioelec- 7.
tron 2006;21:1851– 8.
3. Wang J. Electrochemical biosensors: towards 8.
point-of-care cancer diagnostics. Biosens Bioelec-
tron 2006;21:1887–92.
4. Wu J, Fu ZF, Yan F, Ju HX. Biomedical and clinical
applications of immunoassays and immunosen- 9.
sors for tumor markers. Trac-Trend Anal Chem
2007;26:679 – 88.
5. Tsao KC, Wu TL, Chang PY, Hong JH, Wu JT. 10.
Detection of carcinomas in an asymptomatic Chi-
nese population: advantage of screening with
trons, and thus can exclude the electrochemical cross-
talk at the electrode dimensions used here. Further-
more, the measurement of the direct electrochemical
signal of HRP labeled to immunoreagents also avoids
the need for other reagents in the detection process.
Although the measurements show acceptable results,
adding sulfite in the detection solution is not the best
solution for the removal of oxygen. Thus, a system has
been developed for POCT to exclude oxygen from the
detection solution (see Supplemental Fig. 3 in the on-
line Data Supplement).
Grant/Funding Support: This project was funded by
National Science Funds for Creative Research Groups
(20521503), the Key Program (20535010) and Major
Research Plan (90713015) from NNSFC, Science
Foundation of Jiangsu (BS2006006, BS2006074), and
Outstanding Medical Talents Program (RC2007069)
from the Department of Health of Jiangsu.
Financial Disclosures: None declared.
Acknowledgments: We thank Yanan Chen for her
help.
References
multiple tumor markers. J Clin Lab Anal 2006;20: immunoassay with detection by inductively cou-
42– 6. pled plasma mass spectrometry. Clin Chem 2004;
Sanchez-Carbayo M. Antibody arrays: technical 50:1214 –21.
considerations and clinical applications in cancer. 11. Kricka LJ. Multianalyte testing [editorial]. Clin
Clin Chem 2006;52:1651–9. Chem 1992;38:327– 8.
Haab BB. Antibody arrays in cancer research. Mol 12. Goldman ER, Clapp AR, Anderson GP, Uyeda HT,
Cell Proteomics 2005;4:377– 83. Mauro JM, Medintz IL, Mattoussi H. Multiplexed
Zheng GF, Patolsky F, Cui Y, Wang WU, Lieber toxin analysis using four colors of quantum dot
CM. Multiplexed electrical detection of cancer fluororeagents. Anal Chem 2004;76:684 – 8.
markers with nanowire sensor arrays. Nat Bio- 13. Liu G, Wang J, Kim J, Jan MR, Collins GE. Elec-
technol 2005;23:1294 –301. trochemical coding for multiplexed immunoas-
Brecht A, Abuknesha R. Multi-analyte immunoas- says of proteins. Anal Chem 2004;76:7126 –30.
says application to environmental analysis. Trac- 14. Fu ZF, Liu H, Ju HX. Flow-through multianalyte
Trend Anal Chem 1995;14:361–71. chemiluminescent immunosensing system with
Zhang SC, Zhang C, Xing Z. Simultaneous deter- designed substrate zone-resolved technique for
mination of alpha-fetoprotein and free beta-hu- sequential detection of tumor markers. Anal
man chorionic gonadotropin by element-tagged Chem 2006;78:6999 –7005.

Clinical Chemistry 54:9 (2008) 1487

15. Fu ZF, Yang ZJ, Tang JH, Liu H, Yan F, Ju HX.
Channel and substrate zone two-dimensional res-
olution for chemiluminescent multiplex immuno-
assay. Anal Chem 2007;79:7376 – 82.
16. FitzGerald SP, Lamont JV, McConnell RI, Benchikh
EO. Development of a high-throughput auto-
mated analyzer using biochip array technology.
Clin Chem 2005;51:1165–76.
17. Knecht BG, Strasser A, Dietrich R, Martlbauer E,
Niessner R, Weller MG. Automated microarray
system for the simultaneous detection of antibi-
otics in milk. Anal Chem 2004;76:646 –54.
18. Dill K, Montgomery DD, Ghindilis AL, Schwarz-
kopf KR, Ragsdale SR, Oleinikov AV. Immunoas-
says based on electrochemical detection using
microelectrode arrays. Biosens Bioelectron 2004;
20:736 – 42.
19. Tang D, Yuan R, Chai Y. Magnetic control of an
electrochemical microfluidic device with an ar-
rayed immunosensor for simultaneous multiple
immunoassays. Clin Chem 2007;53:1323–9.
20. Kojima K, Hiratsuka A, Suzuki H, Yano K, Ike-
bukuro K, Karube I. Electrochemical protein chip
with arrayed immunosensors with antibodies im-
mobilized in a plasma-polymerized film. Anal
Chem 2003;75:1116 –22.
21. Wilson MS. Electrochemical immunosensors for
the simultaneous detection of two tumor mark-
ers. Anal Chem 2005;77:1496 –502.
22. Wilson MS, Nie WY. Electrochemical multianalyte
immunoassays using an array-based sensor. Anal
Chem 2006;78:2507–13.
1488 Clinical Chemistry 54:9 (2008)
23. Wilson MS, Nie WY. Multiplex measurement of
seven tumor markers using an electrochemical
protein chip. Anal Chem 2006;78:6476 – 83.
24. Wu J, Yan F, Tang JH, Zhai C, Ju HX. A disposable
multianalyte electrochemical immunosensor array
for automated simultaneous determination of tu-
mor markers. Clin Chem 2007;53:1495–502.
25. Wu J, Zhang ZJ, Fu ZF, Ju HX. A disposable
two-throughput electrochemical immunosensor
chip for simultaneous multianalyte determination
of tumor markers. Biosens Bioelectron 2007;23:
114 –20.
26. Kerman K, Nagatani N, Chikae M, Yuhi T, Taka-
mura Y, Tamiya E. Label-free electrochemical im-
munoassay for the detection of human chorionic
gonadotropin hormone. Anal Chem 2006;78:
5612– 6.
27. Dai Z, Yan F, Chen J, Ju HX. Reagentless ampero-
metric immunosensors based on direct electro-
chemistry of horseradish peroxidase for determi-
nation of carcinoma antigen-125. Anal Chem
2003;75:5429 –34.
28. Chen J, Yan F, Dai Z, Ju HX. Reagentless ampero-
metric immunosensor for human chorionic gona-
dotrophin based on direct electrochemistry of
horseradish peroxidase. Biosens Bioelectron
2005;21:330 – 6.
29. Chen J, Tang JH, Yan F, Ju HX. A gold nanopar-
ticles/sol-gel composite architecture for encapsu-
lation of immunoconjugate for reagentless elec-
trochemical immunoassay. Biomaterials 2006;27:
2313–21.
30. Tan F, Yan F, Ju HX. A designer ormosil gel for
preparation of sensitive immunosensor for carci-
noembryonic antigen based on simple direct elec-
tron transfer. Electrochem Commun 2006;8:
1835–9.
31. Tan F, Yan F, Ju HX. Sensitive reagentless
electrochemical immunosensor based on an or-
mosil sol-gel membrane for human chorionic
gonadotrophin. Biosens Bioelectron 2007;22:
2945–51.
32. Wu J, Tang JH, Dai Z, Yan F, Ju HX, Murr NE. A
disposable electrochemical immunosensor for
flow injection immunoassay of carcinoembryonic
antigen. Biosens Bioelectron 2006;22:102– 8.
33. Liu SQ, Ju HX. Renewable reagentless hydrogen
peroxide sensor based on direct electron transfer
of horseradish peroxidase immobilized on colloi-
dal gold-modified electrode. Anal Biochem 2002;
307:110 – 6.
34. Yu H, Yan F, Dai Z, Ju HX. A disposable ampero-
metric immunosensor for ␣-1-fetoprotein based
on enzyme-labeled antibody/chitosan-membrane
modified screen-printed carbon electrode. Anal
Biochem 2004;331:98 –105.
35. Huang H, Hu NF, Zeng YH, Zhou G. Electrochem-
istry and electrocatalysis with heme proteins in
chitosan biopolymer films. Anal Biochem 2002;
308:141–51.
36. Willner I, Willner B, Katz E. Biomolecule–
nanoparticle hybrid systems for bioelectronic ap-
plications. Bioelectrochemistry 2007;70:2–11.