Potential-Resolved Electrochemiluminescence for Determination ofTwo Antigens at the Cell SurfaceFangfei Han,† Hui Jiang,§ Danjun Fang,*,‡ and Dechen Jiang*,†

†The State Key Lab of Analytical Chemistry for Life Science, School of Chemistry and Chemical Engineering, Nanjing University,Nanjing, Jiangsu 210093, China‡Department of Pharmacology, Nanjing Medical University, Nanjing, Jiangsu 210029, China§State Key Laboratory of Bioelectronics-Chien-Shiung Wu Lab, School of Biological Science and Medical Engineering, SoutheastUniversity, Nanjing, Jiangsu 210096, China

*S Supporting Information

ABSTRACT: The potential-resolved electrochemiluminescence(ECL) was achieved for the determination of two antigens at thecell surface through a potential scanning on the electrode.Luminol and Ru(bpy)3

2+ groups as ECL probes were linked withthe antibodies to recognize the corresponding antigens on thecell surface. A self-quenching of luminescence from the luminolgroup under negative potential was initialized by the introductionof concentrated aqueous luminol, which offered accuratemeasurements of the luminescence from luminol and Ru(bpy)3

2+

groups under positive and negative potentials, respectively. Usingthis strategy, carcinoembryonic (CEA) and alphafetoprotein(AFP) antigens on cells as the models were quantified seriallythrough a potential scanning. Different patterns of luminescencewere observed at MCF 7 and PC 3 cells, which exhibited that the assay can characterize the cells with a difference expression ofantigens. Compared with fluorescence measurement, the potential resolved ECL for the detection of two analytes was not limitedby the spectrum difference of probes. The strategy involving potential-induced signals required a simplified optical setup andeventually offered an alternative imaging method for multiply antigens in immunohistochemistry.

The assay of multiple biomarkers at cells is significant forthe accurate diagnosis of cancers.1,2 In the past few years,

the considerable efforts have been put in developing fluorescentimmunoassays for the detection.3,4 Typically, the biomarkers atthe cell surface are labeled with the fluorophores with differentexcitation/emission wavelengths. Under the irradiation of laser,the emissions from different fluorophores are distinguishedthrough the beam splitters and filters before collected on aphotomultiplier tube (PMT) or charge-coupled device (CCD).Although fluorescent immunoassay is advantagous for a widespectrum of fluorophores, the ease of fluorescent labeling, andhigh throughput detection,5 the fluorescence assay relies on thespectrum difference of probes. Therefore, the components ofthe laser, the filter, or the beam splitter are required thatcomplex the optical setup and increase instrumental cost.Electrochemiluminescence (ECL) is the other optical

strategy for the immunoassay that replaces the fluorophorewith the ECL probe.6−9 ECL does not need the light sourceresulting in low background, high detection sensitivity, andsimple instrumentation. Since the emission of the luminescencefrom the ECL probe was controlled by the potential, it wasfeasible to develop potential-resolved ECL that detectedmultiply analytes through a potential scanning. This potentialresolved strategy is not related with the spectrum difference of

probes, and thus, no beam splitter or filter is needed todistinguish the signals. Pioneering work on the potential-resolved ECL assay had been performed using Ir(ppy)3 and[Ru(bpy)2(L)]

2+ complexes in solution, which emitted the lightunder different potentials.10,11 For the application of potential-resolved ECL assay in immunoassay, Ru(bpy)3

2+ and luminolare chosen as ECL probes, which have been reported to belinked with the antibody for the recognition of the antigen atthe cell surface.12−14 In principle, applied with a positivepotential, the luminol anion goes through electro-oxidation todiazaquinone, which is further oxidized into the excited 3-aminophthalate species for the emission light.15−17 For theRu(bpy)3

2+ probe, S2O82− as the coreactant is reduced into

SO4•− under negative potential, which generates Ru(bpy)3

2+*to give out light.18−20 Given the fact that ECL reactions ofluminol and Ru(bpy)3

2+ occur under different potentials, it ispossible to detect two antigens using both of ECL probesthrough a potential scanning. Compared with the multiplexedECL determination of surface antigens using an electrodearray,21 the assay on one cell population can reveal the

Received: February 10, 2014Accepted: June 25, 2014Published: June 25, 2014

Article

pubs.acs.org/ac

© 2014 American Chemical Society 6896 dx.doi.org/10.1021/ac501571a | Anal. Chem. 2014, 86, 6896−6902

correlation of surface antigens from the same cell state and,most importantly, be applied for cellular imaging of multiplyantigens in immunohistochemistry.Although potential-resolved ECL based on the luminol/

Ru(bpy)32+ system is feasible in principle, we have previously

incurred challenges in applying this system to achieve potential-resolved ECL because the cross-reactions between luminol andS2O8

2− gave the luminescence under the same potential withRu(bpy)3

2+. The luminescence overlapping induced thedifficulty to correlate the luminescence with each luminescenceprobe. To overcome this challenge, a self-quenching ofluminescence from the luminol group in the presence of theRu(bpy)3

2+ group and S2O82− under negative potential was

initialized by the introduction of concentrated aqueous luminol,which achieved the restriction of the luminescence fromluminol and Ru(bpy)3

2+ groups under positive and negativepotentials, respectively. The luminescence related with specificpotentials realized the potential-resolved ECL for the serialdetection through a potential scanning. By labeling luminol andRu(bpy)3

2+ groups on AFP and CEA antibodies, the potential-resolved method was validated on AFP and CEA antigenmodified silica particles. Afterward, the assay was applied forthe quantification of AFP and CEA antigen at cells. The successin the establishment of potential-resolved ECL offered astrategy for the analysis of multiply analytes at cells without anycomplex optical setup.

■ EXPERIMENTAL SECTIONChemical. Au nanoparticles (Au NPs) with 13 nm in

diameter were purchased from Beijing Deke Daojin Science andTechnology Co., Ltd. (Beijing, China). CEA antibody andantigen were obtained from Zhengzhou Biocell BiotechnologyCo., Ltd. (Zhengzhou, China) and Beijing Bioss BiotechnologyCo., Ltd. (Beijing, China), respectively. AFP antibody andantigen were from Shuangliu Zhenglong Biochem Lab(Chengdu, China). Biotinylated CEA antibody was purchasedfrom Beijing Key-bio Biotech Co., Ltd. (Beijing, China). MCF7 cells and PC 3 cells were from the Institute of Biochemistryand Cell Biology, Shanghai Institute for Biological Sciences ofChinese Academy of Science (Shanghai, China). Silica particles(20 μm in diameter) with carboxylic acid groups were obtainedfrom Micromod Partikeltechnologie GmbH (Germany). Allother chemicals were from Sigma-Aldrich, unless indicatedotherwise. Ultrapure water with a resistivity of 18.2 MΩ/cmwas used throughout. Buffer solutions were sterilized.Cell Culture. MCF 7 cells and PC 3 cells were seeded in

DMEM/high glucose medium and F-12K medium supple-mented with 10% fetal bovine serum (FBS) and 1% antibiotics(penicillin/streptomycin), respectively. Cultures were main-tained at 37 °C under a humidified atmosphere containing 5%CO2.Preparation of Antibody Associated Luminol Com-

plex. 3-Mercaptopropionic acid (3-MPA) (1 mM) was addedinto the solution with 13 nm Au nanoparticles (Au NPs). Theamount of Au NPs was 9 × 1013/mL. The mixture was stirredat 37 °C for 12 h and centrifuged at 14 000 rpm for 30 min toget the gold conjugates (Au NPs-MPA). To activate thecarboxylic acid groups on the surface of gold conjugates, themixture of 1 mM ethyl(dimethylaminopropyl) carbodiimide(EDC) and N-hydroxysuccinimide (NHS) were reacted withAu NPs-MPA at 37 °C for 1 h. Then, the activated Au NPswere reacted with 5.4 × 1013 molecules/mL antibody (equal to12.88 μg/mL) at 37 °C for 12 h. The little excessive Au NPs

might introduce only one antibody on each particle. Au NPswithout the addition of antibody was removed through thecentrifugation after the following interaction with antigenmodified silica particle or cells. Finally, 0.25 mM N-(4-aminobutyl)-N-ethylisoluminol was applied on Au NPs-MPA-antibody for 12 h in the dark to introduce luminol group on AuNPs. The amount of N-(4-aminobutyl)-N-ethylisoluminol was1.5 × 1017 molecules/mL resulting in the averaged 1666luminol groups on each Au NP. The final product (Au NPs-MPA-antibody/luminol) was resuspended in 10 mM phos-phate-buffered saline (PBS, pH 7.4) and stored at 4 °C.

Preparation of Streptavidin Associated Ru(bpy)32+

Complex. One mg/mL streptavidin was mixed with 1 mg/mL bis(2,2′-bipyridine)-4′-methyl-4-carboxybipyridine-ruthe-nium N-succinimidyl ester-bis (hexafluorophosphate) (Ru-(bpy)3

2+) complex for 2 h at 4 °C. The complexes werepurified by ultrafiltration using Amicon Ultra filters with a 10kmolecular weight cut off membrane (Millipore). Thestreptavidin associated Ru(bpy)3

2+ complex was diluted to 20μg/mL with PBS (pH 7.4) and stored at 4 °C.

Synthesis of Antigen-Modified Silica Particles. Silicaparticles were reacted with 0.2 M NHS and 0.8 M EDC at 37°C for 1 h to activate carboxylic acid groups. Then, the particleswere linked with 12.88 μg/mL antibody at 37 °C for 24 h. Theremaining active carboxylic acid groups were blocked with 0.1%(w/v) bovine serum albumin (BSA) at 37 °C for 1 h. After theremoval of extra antibody in the solution, 2 μg/mL antigen wasadded and incubated at 37 °C for 2 h to form antigen-modifiedsilica particles.

Linkage of the Complexes on Antigen-Modified SilicaParticles and Cells. To link Au NPs-MPA-luminol/antibodyon the antigen modified silica particles, the particles wereresuspended in PBS (pH 7.4)−0.02% (w/v) Tween-20 andincubated with luminol associated antibody at 37 °C for 2 h.For the linkage of streptavidin-modified Ru(bpy)3

2+ complexon antigen modified silica particles, the silica particles wereincubated with 3 μg/mL biotinylated antibody at 37 °C for 30min. After washing for 3 times, the product was mixed with 20μg/mL streptavidin associated Ru(bpy)3

2+ complex for 2 minto form Ru(bpy)3

2+ labeled silica particles. The same procedurewas applied for the linkage of the complexes on the cells. Beforethe linkage, the cells were fixed by 2.5% glutaraldehyde.

Luminescence Detection. The indium tin oxide (ITO)electrode with a diameter of 2 cm was used as the workingelectrode for luminescence detection. Ag/AgCl and Ptelectrodes were connected as a reference and counter electrode,respectively. PMT voltage was set at 600 V. For luminolassociated analysis, the potential scanned from 0.6 to 0 V in 10mM aerated PBS and the luminescence read at 0.6 V was takenas the signal. For Ru(bpy)3

2+ associated analysis, the potentialscanned from 0 to −1.0 V in 10 mM PBS containing 20 mMluminol and 3 mM S2O8

2−, and the luminescence read at −1.0V was taken as the signal. To decrease the measurement errorinduced by different ITO electrodes, the differences in theluminescence between luminescence probe-antibody/antigenmodified particles/cells and antigen modified particles/cellswere calculated and normalized by the background lumines-cence read at 0 V as the “luminescence ratio”. Theluminescence spectra were measured using the band-pass filterswith the bandwidth of 20 nm from 400 to 720 nm betweenITO electrode and PMT.

Analytical Chemistry Article

dx.doi.org/10.1021/ac501571a | Anal. Chem. 2014, 86, 6896−69026897

■ RESULTS AND DISCUSSIONLuminescence Distinction from Each Probe under

Different Potentials. The aim of our work is the detection oftwo antigens at the cell surface using ECL probes through apotential scanning. Thus, the luminescence under differentpotential should be associated with each probe specifically. Itwas well-known that Ru(bpy)3

2+ in the presence of S2O82−

emitted the luminescence under the negative potential. For thecorrelation of the luminescence under negative potential withRu(bpy)3

2+ associated complex, the introduction of luminolgroup was not expected to generate luminescence in thispotential window. To investigate the luminescence of theluminol group under negative potential, the luminescences ofRu(bpy)3

2+/ S2O82− in the absence and presence of luminol

were compared. Although an intense luminescence wasobserved for Ru(bpy)3

2+ with the potential less than −1.0 V(Figure S1 in Supporting Information), the lowest potentialapplied for the analysis was chosen at −1.0 V to minimize theinterruption of potential on the cells. As shown in Figure 1A

curves a and b, the mixture of luminol and Ru(bpy)32+

generated more luminescence than Ru(bpy)32+ itself in the

presence of 3 mM S2O82− on the ITO electrode under negative

potential. The additional luminescence was attributed to thecross-reaction between luminol and S2O8

2− under negativepotential, as evidenced in Figure 1 curve c. Replacing theelectrode material into gold did not change the phenomenon,which indicated that reaction between luminol and S2O8

2− wasnot related with electrode material. Also, since aerated bufferwith oxygen was required for the following cell analysis, thecontribution of oxygen in the generation of luminescence wasinvestigated. As shown in Figure S2 (Supporting Information),the same luminescence was observed in the absence andpresence of oxygen exhibiting no contribution of oxygen in the

production of luminescence. The cogeneration of luminescencefrom luminol and Ru(bpy)3

2+ in the presence of S2O82− gave

the difficulty to distinguish the luminescence from theRu(bpy)3

2+ group under negative potential.To restrict the luminescence from luminol under negative

potential, self-quenching of luminescence from concentratedluminol was initialized. As shown in Figure 1B and Figure S3(Supporting Information), more aqueous luminol introduced inthe presence of S2O8

2− led to a gradual decrease in theluminescence under negative potential. When the concen-tration of luminol was over 15 mM, a constant weakluminescence from luminol was obtained under negativepotential that was independent of the amount of luminol.After obtaining the constant luminescence in the presence of 20mM luminol and 3 mM S2O8

2− ion, more Ru(bpy)32+ was

added in the solution and an increase in the luminescence wasobserved, as shown in Figure 1B. The luminescence increasewas associated with the amount of Ru(bpy)3

2+ only in thepresence of concentrated luminol and, thus, can be used toquantify the amount of Ru(bpy)3

2+ under negative potential.The same trend of luminescence with concentrated aqueousluminol was observed in the absence of oxygen, as shown inFigure S4 in the Supporting Information, which confirmed thatoxygen was not related to the quenching process.As for positive potential, Figure 2 showed that the starting

potentials for the emission of luminescence from luminol and

Ru(bpy)32+ were 0.40 and 0.60 V, respectively. The

luminescence spectra of Ru(bpy)32+ at the potential over 0.6

V, as shown in Figure S5 (Supporting Information), showedthe maximum peak wavelength near 620 nm. The resultindicated a small amount of Ru(bpy)3

2+* generated over thispotential.18 To avoid the overlapping of luminescence fromluminol and Ru(bpy)3

2+, the highest positive potential was setat 0.6 V. Hydrogen peroxide was a classic coreactant to enhanceluminol ECL. However, Figure S6 (Supporting Information)exhibited that the coexistence of luminol, Ru(bpy)3

2+, hydrogenperoxide, and S2O8

2− generated chemiluminescence, which wasindependent of the potential and quenched ECL fromRu(bpy)3

2+ and S2O82− under negative potential. The

mechanism was most likely that S2O82− as a strong oxidant

induced the production of oxygen radical from hydrogenperoxide for luminol chemiluminescence. The consumption ofS2O8

2− in chemiluminescence inhibited the conversion ofS2O8

2− into SO4•− under negative potential to generate ECL

from Ru(bpy)32+. Since luminol itself generated luminescence

under positive potential, hydrogen peroxide was excluded fromthe system to restore ECL from Ru(bpy)3

2+/S2O82−. Also, the

contribution of oxygen in the luminescence of luminol and

Figure 1. (A) The luminescence curve of (a) 100 μM Ru(bpy)32+, (b)

100 μM luminol/Ru(bpy)32+ and (c) luminol in the presence of 3 mM

S2O82− under negative potential; (B) the luminescence read at −1.0 V

from 100 μM Ru(bpy)32+ and 3 mM S2O8

2− in the presence of 0.1, 1,5, 10, 15, 20 mM luminol. The last point presents the luminescencefrom 200 μM Ru(bpy)3

2+ in 20 mM luminol and 3 mM S2O82−.

Figure 2. Luminescence curve of (a) 200 μM luminol and (b) 200 μMRu(bpy)3

2+ under positive potential.

Analytical Chemistry Article

dx.doi.org/10.1021/ac501571a | Anal. Chem. 2014, 86, 6896−69026898

Ru(bpy)32+ was investigated. As shown in Figure S7

(Supporting Information), the same luminescence increasewas observed in the presence and absence of oxygen. Thisphenomenon suggested that oxygen did not contribute to thegeneration of luminescence, which was similar to the literatureresult collected on a Pt or graphite electrode.22 Therefore,when the positive potential was scanned from 0.6 to 0 V, theluminescence in this potential range was only attributed toluminol that was independent of the amount of Ru(bpy)3

2+.Overall, the luminescences under the positive and negative

potentials were determined to be associated with luminol andRu(bpy)3

2+ groups, respectively. The potential-resolved ECLassay for the determination of two antigens at the cell surfacethrough a potential scanning was proposed as (1) under thepositive potential, only luminol associated antigen complex atthe cell surface emitted the luminescence, which was used tothe determination of the amount of luminol associated antigen;(2) under the negative potential, the addition of aqueousluminol in the solution quenched the luminescence fromluminol and luminol associated complex in the presence ofS2O8

2− resulting in a small constant luminescence; after theextracting of that constant luminescence from luminol, theadditional luminescence under the negative potential wasrelated with the amount of Ru(bpy)3

2+ associated antigen.Although the solution component was altered when thepotential reached the negative region, the same cell populationwas analyzed through a potential scanning that could offer theinformation about two antigens at the cell surface.Potential-Resolved ECL Assay for the Surface Antigen

at Silica Particles. The distinction of luminescence wasfurther validated using antigen modified silica particles. Thediameter of particles was 20 μm that was similar to the cell size.During the analysis, the particles sit on the electrode by gravityto mimic the behavior of adherent cells. To modify the antigenon silica particles, the particles with carboxylic groups werereacted with the amino group at the CEA antibody that boundCEA antigens. Then, the antigen modified particles interactedwith luminol or Ru(bpy)3

2+ labeled CEA antibody for ECLdetection. All the preparation procedures are shown in Figure 3.As for the labeling of luminol group on antibody, 13 nm goldnanoparticles (Au-NPs) with carboxylic acid groups werereacted with CEA antibody and NH2-coupling luminol in serialso that both the antibody and luminol groups were associatedon Au-NP as a complex for the recognition of antigen.23 Toassociate CEA antibody with Ru(bpy)3

2+ group, the biotiny-

lated CEA antibody was used. The biotin on the antibody waslinked with the streptavidin-Ru(bpy)3

2+ complex.12

Under positive potential from 0.6 to 0 V, Figure 4A curves aand b showed the “background luminescence” from CEA

antigen modified particles and more luminescence fromluminol-CEA antibody/antigen particles on the ITO electrode.The maximum emission wavelength near 430 nm, as shown inFigure S8 (Supporting Information), supported the generationof luminescence from the luminol group on the particles, whichwas associated with CEA antigens. The weak luminescence wasmainly caused by the limited number of luminol group on theelectrode. On the basis of the fluorescence intensity (Ex/Em280/335 nm) of antibody-Au NPs complex in the solution

Figure 3. Preparation procedure for antigen modified silica particles and luminol or Ru(bpy)32+-antibody/antigen modified silica particle.

Figure 4. (A) Luminescence curve under positive potential from (a)105 CEA antigen modified silica particles, (b) 105 luminol-CEAantibody/antigen particles, (c) mixture of 105 luminol-CEA antibody/antigen particles and 105 Ru(bpy)3

2+-CEA antibody/antigen particles,(d) 3 × 105 luminol-CEA antibody/antigen particles. (B)Luminescence curve in the presence of 20 mM luminol and 3 mMS2O8

2− under negative potential from (a) 105 CEA antigen modifiedparticles, (b) 105 Ru(bpy)3

2+-CEA antibody/antigen particles, (c)mixture of 105 Ru(bpy)3

2+-CEA antibody/antigen particles and 105

luminol-CEA antibody/antigen particles, (d) 3 × 105 Ru(bpy)32+-CEA

antibody/antigen particles.

Analytical Chemistry Article

dx.doi.org/10.1021/ac501571a | Anal. Chem. 2014, 86, 6896−69026899

before and after the linkage on the particle, the amount of AuNPs on each particle was estimated to be 1.60 × 10−19 mol.Taking into account that 1666 luminol groups were associatedwith one Au NP, ∼105 particles on the electrode had amaximum of 2.66 × 10−11 mol of luminol group. Since onlyluminol group at the interface of particles and the electrodegenerated the luminescence, the total amount of luminol wasrare for weak luminescence. To investigate the interruption ofthe Ru(bpy)3

2+ group on the luminescence under positivepotential, Ru(bpy)3

2+-CEA antibody/antigen particles wereintroduced on the electrode. No increase in the luminescencewas observed in curve c confirming that the luminescenceunder positive potential was only attributed to luminolassociated particles. The addition of more luminol-CEAantibody/antigen particles on the electrode increased theamount of antigens and an enhanced luminescence wasobserved in Figure 4 A curve d, which indicated that ourassay can monitor the amount charge of antigen on particles.For ECL under negative potential, 20 mM aqueous luminol

and 3 mM S2O82− were added into the solution after the

potential scanned from 0.6 to 0 V. Figure 4B curve a exhibitedthe “background luminescence” from CEA antigen modifiedparticles on the electrode with the potential from 0 to −1 V.The weak luminescence observed at −1.0 V came from aqueousconcentrated luminol in the presence of S2O8

2−. When theparticles were replaced by Ru(bpy)3

2+-CEA antibody/antigenparticles, more luminescence observed in curve b was generatedfrom the Ru(bpy)3

2+ group on the particles. The addition ofluminol-antibody/antigen particles did not alter the lumines-cence, as shown in curve c. These results confirmed that theluminescence from Ru(bpy)3

2+ group was determined in thepresence of luminol associated antigen. Similar to the resultunder positive potential, more Ru(bpy)3

2+-antibody/antigenmodified particles were introduced on the electrode and anincrease in the luminescence was observed, as expected in curved.After the validation of our assay, the relation of luminescence

with the amount of antigen was established for the followingquantitative measurement in cells. To correlate the lumines-cence of luminol or Ru(bpy)3

2+ with different surface antigens,luminol and Ru(bpy)3

2+ groups were labeled with AFP andCEA antibody, respectively, for the recognition of thecorresponding antigens at silica particles. For the regulationof antigen amount, the number of particles on the electrodewas controlled. The estimation process of CEA and AFPantigen on each silica particle was discussed in the SupportingInformation. Since concentrated luminol gave a weak andconstant luminescence, the background luminescence collectedon the unmodified cells needed to be excluded from theluminescence collected on luminescence probe-modified cells.This luminescence difference was further normalized by thebackground luminescence to minimize the luminescencedeviation created by different ITO slides. As shown in Figure5A,B, linear relationships of luminescence ratio on the amountof AFP and CEA antigens under positive and negativepotentials were demonstrated. The correlation supported thequantitative measurement of surface antigens at cells using ourassay.Potential-Resolved ECL Assay for the Detection of

Two Antigens at Cell Surface. For the detection of twosurface antigens at cells, the cells were fixed to ensure theinteraction between the antigens and antibodies and minimizethe interruption of concentrated aqueous luminol on cellular

activity. MCF 7 cells with high expression of CEA and AFPantigens at the cell surface were used as a model.24 Using thesame modification protocol, the cells were colabeled with AFPantibody-luminol and CEA antibody-Ru(bpy)3

2+ complexes.Figure 6A showed the luminescence with the potential scanningfrom 0.6 to −1.0 V. When the potential reached 0 V, 20 mMluminol and 3 mM S2O8

2− were added. Compared with the

Figure 5. Correlation of luminescence ratio with (A) AFP antigen and(B) CEA antigen on the electrode. The red lines were the linear fittingcurves. The error bars represent the standard deviation from fourindependent experiments.

Figure 6. (A) The luminescence curves from (a) 1.6 × 105 luminol-AFP antibody and Ru(bpy)3

2+-CEA antibody labeled MCF 7 cells, and(b) 1.6 × 105 unlabeled MCF 7 cells; (B) the bar of luminescence ratioon the measurement of AFP and CEA antigen at 1.6× 105 MCF 7 andPC 3 cells. The bars labeled with “∗” were collected from thesimultaneous detection of two antigens. The error bars were thestandard deviation from three groups of cells.

Analytical Chemistry Article

dx.doi.org/10.1021/ac501571a | Anal. Chem. 2014, 86, 6896−69026900

background luminescence on unmodified cells in curve a, theincreases in luminescence under both of positive and negativepotentials in curve b exhibited the analysis of two antigens atthe cell surface. The luminescence ratios from three groups ofcells were calculated to be 1.59 ± 0.09 for AFP antigen and 0.92± 0.13 for CEA antigen, as shown in Figure 6B. The controlexperiment was performed on bare ITO slide, which did nothave the cells cultured. The luminescence curve collected frombare ITO slide, as shown in Figure S9 in the SupportingInformation, was similar to the background luminescence curvewith a slight increase in value. Since the luminescence probeswere not located on the slide after the cell removal, and thus,the similarity of luminescence curve collected on bare ITO slideand ITO slide cultured with unmodified cells was reasonable.The slight increase in the luminescence value was ascribed to amore exposed electrode surface after the removal of cells. Thisresult confirmed that the luminescence increase observed oncurve b was attributed to the luminescence probe linked withthe antigens on the cell surface.To confirm the accuracy of our detection, MCF 7 cells were

labeled with either AFP antibody associated luminol or CEAantibody associated Ru(bpy)3

2+ to measure the luminescence.The luminescence ratios shown in Figure 6B were similar tothose from the codetection in the presence of twoluminescence probes, which suggested that the colabeling oftwo ECL probes and the detection did not give anymeasurement error. Meanwhile, PC 3 cells with no expressionof CEA and AFP antigens on the surface were used as anegative control cell model.24 No luminescence change in thewhole potential range was observed before and after thelabeling of AFP antibody-luminol and CEA antibody-Ru-(bpy)3

2+ on PC 3 cells, as shown in Figure 6B. All theseresults supported that our assay quantified surface antigens witha relative accuracy. Also, our assay needed 1.6 × 105 cells, whichwas similar to the cell number required in the clinic tests usingfluorescence, HPLC, and mass spectroscopy. The success in theanalysis of similar pool-sized cells exhibited a potentialapplication of our assay for the real diagnosis.Referring to the luminescence ratio in Figure 5A,B, the

average amount of AFP and CEA antigen at the cells weredetermined as 2.15 fg and 32.5 fg per cell. The results wereclose to the literature data, which was 0.21−1.75 fg for AFP and18−27 fg for CEA per cell using cell lysate for analysis.24−26

The small difference in the antigen amount might be attributedto our analysis process. For ECL assay, only the antigenassociated ECL probe at the boundary of cells/particles andelectrode generated the luminescence, which was used toestimate the amount of antigen on the whole cell surface. Ascompared with the direct analysis of whole cell lysate, theaddition of some measurement error might exist.

■ CONCLUSIONIn this paper, a potential resolved electrochemiluminescenceassay was achieved for the detection of AFP and CEA antigensat the cell surface. The luminescence from each probe wasrestricted in the limited potential windows so that they wereused for the quantification of surface antigens. Compared withthe fluorescence assay, the analysis strategy using the potential-controlled signals avoids the limitation of spectrum differencefor probes and simplifies the instrumental setup. Thecontinuous introduction of more ECL probes with differentpotentials will permit the strategy to analyze more surfaceantigens during one potential scanning. Also, coupled with the

electrochemiluminescence imaging system, the assay waspromising for the coimaging of multiply analytes at cells. Thesuccess of this imaging method will provide an alternativetechnology for the cellular immunochemistry.

■ ASSOCIATED CONTENT*S Supporting InformationAdditional information as noted in the text. This material isavailable free of charge via the Internet at http://pubs.acs.org.

■ AUTHOR INFORMATIONCorresponding Authors*E-mail: djf@njmu.edu.cn. Phone: 086-25-86868477. Fax: 086-25-86868477.*E-mail: dechenjiang@nju.edu.cn. Phone: 086-25-83594846.Fax: 086-25-83594846.NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThis work was supported by the 973 Program (2013 GrantCB933800), the National Natural Science Foundation of China(Grant Nos. 21327902, 21135003, 21105045, and 21105049),and the open research fund from State Key Laboratory ofAnalytical Chemistry for Life Science (Grant SKLACLS 1212),Nanjing University.

■ REFERENCES(1) Wulfkuhle, J. D.; Liotta, L. A.; Petricoin, E. F. Nat. Rev. Cancer2003, 3, 267−275.(2) Marquette, C. A.; Corgier, B. P.; Blum, L. J. Bioanalysis 2012, 4,927−936.(3) Pepperkok, R.; Squire, A.; Geley, S.; Bastiaens, P. I. H. Curr. Biol.1999, 9, 269−272.(4) Vignali, D. A. A. J. Immunol. Methods 2000, 243, 243−255.(5) Hemmila, I. Clin. Chem. 1985, 31, 359−370.(6) Bard, A. J., Ed. Electrogenerated Chemiluminescence; MarcelDekker: New York, 2004.(7) Forster, R. J.; Bertoncello, P.; Keyes, T. E. Annu. Rev. Anal. Chem.2009, 2, 359−385.(8) Miao, W. J. Chem. Rev. 2008, 108, 2506−2553.(9) Richter, M. M. Chem. Rev. 2004, 104, 3003−3036.(10) Doeven, E. H.; Zammit, E. M.; Barbante, G. J.; Francis, P. S.;Barnett, N. W.; Hogan, C. F. Chem. Sci. 2013, 4, 977−982.(11) Doeven, E. H.; Barbante, G. J.; Kerr, E.; Hogan, C. F.; Endler, J.A.; Francis, P. S. Anal. Chem. 2014, 86, 2727−2732.(12) Deiss, F.; LaFratta, C. N.; Symer, M.; Blicharz, T. M.; Sojic, N.;Walt, D. R. J. Am. Chem. Soc. 2009, 131, 6088−6089.(13) Tian, D. Y.; Duan, C. F.; Wang, W.; Cui, H. Biosens. Bioelectron.2010, 25, 2290−2295.(14) Chen, Z. H.; Liu, Y.; Wang, Y. Z.; Zhao, X.; Li, J. H. Anal. Chem.2013, 85, 4431−4438.(15) Sakura, S. Anal. Chim. Acta 1992, 262, 49−57.(16) Cui, H.; Xu, Y.; Zhang, Z. F. Anal. Chem. 2004, 76, 4002−4010.(17) Marquette, C. A.; Blum, L. J. Anal. Chim. Acta 1999, 381, 1−10.(18) White, H. S.; Bard, A. J. J. Am. Chem. Soc. 1982, 104, 6891−6895.(19) Bolletta, F.; Ciano, M.; Balzani, V.; Serpone, N. Inorg. Chim.Acta 1982, 62, 207−213.(20) Bolletta, F.; Juris, A.; Maestri, M.; Sandrini, D. Inorg. Chim. Acta1980, 44, L175−L176.(21) Wu, M. S.; Shi, H. W.; He, L. J.; Xu, J. J.; Chen, H. Y. Anal.Chem. 2012, 84, 4207−4213.(22) Wro blewska, A.; Reshetnyak, O. V.; Koval’chuk, E. P.;Pasichnyuk, R. I.; Błazejowski, J. J. Electroanal. Chem. 2005, 580,41−49.

Analytical Chemistry Article

dx.doi.org/10.1021/ac501571a | Anal. Chem. 2014, 86, 6896−69026901

(23) Yang, X. Y.; Guo, Y. S.; Wang, A. G. Anal. Chim. Acta 2010, 666,91−96.(24) Grieve, R. J.; Woods, K. L.; Mann, P. R.; Smith, S. C. H.;Wilson, G. D.; Howell, A. Br. J. Cancer 1980, 42, 616−619.(25) Biddle, W.; Sarcione, E. J. Breast Cancer Res. Treat. 1987, 10,279−286.(26) Zeskind, B. J.; Jordan, C. D.; Timp, W.; Trapani, L.; Waller, G.;Horodincu, V.; Ehrlich, D. J.; Matsudaira, P. Nat. Methods 2007, 4,567−569.

Analytical Chemistry Article

dx.doi.org/10.1021/ac501571a | Anal. Chem. 2014, 86, 6896−69026902