Biosensors and Bioelectronics 49 (2013) 105–110

Contents lists available at SciVerse ScienceDirect

Biosensors and Bioelectronics

0956-56http://d

n CorrE-m

journal homepage: www.elsevier.com/locate/bios

Sensitive quantification of vascular endothelial growth factor (VEGF)using porosity induced hydrogel microspheres

Mohammad Ali Al-Ameen, Gargi Ghosh n

Bioengineering Program, Department of Mechanical Engineering, University of Michigan-Dearborn, 4901 Evergreen Road, Dearborn, MI 48128, USA

a r t i c l e i n f o

Article history:Received 4 March 2013Received in revised form26 April 2013Accepted 2 May 2013Available online 10 May 2013

Keywords:VEGFPoly(ethylene) glycol diacrylate (PEGDA)Hydrogel microspheresPorosityPorogen

63/$ - see front matter & 2013 Elsevier B.V. Ax.doi.org/10.1016/j.bios.2013.05.004

esponding author. Tel.: +313 593 5013; fax: +ail address: gargi@umich.edu (G. Ghosh).

a b s t r a c t

Vascular endothelial growth factor (VEGF) plays a crucial role in vasculogenesis (blood vessel formation)and angiogenesis (capillary formation from a pre-existing blood vessel). Dysregulation of VEGF has beenassociated with several diseases including cancer, rheumatoid arthritis, and psoriasis. As a result, serumlevel of VEGF has important implications as biomarker for different clinical disorders as well as forsubsequent therapy monitoring. A simple detection method capable of rapid and sensitive analysis ofVEGF in serum of patients with different clinical disorders is of paramount importance. Here, we reportthe fabrication and utilization of capture-antibody immobilized macro-porous poly(ethylene) glycoldiacrylate (PEGDA) hydrogel microspheres for quantitative and reproducible measurement of VEGF.We demonstrate that induction of porosity using PEG porogen improves the sensitivity of this simplehydrogel microsphere based system with a detection limit of 2.5 pg/ml; indicating that the sensitivity ofthe assay exceeds that of the conventional technologies.

& 2013 Elsevier B.V. All rights reserved.

1. Introduction

Vascular endothelial growth factor (VEGF) is a homodimericglycoprotein which plays a crucial role in vasculogenesis (bloodvessel formation) and angiogenesis (capillary formation from apre-existing blood vessel) (Byrne et al., 2005; Farrera, 2004;Hoeben et al., 2004). Binding of VEGF to its receptors initiates asignaling cascade which results in migration and proliferation ofendothelial cells and formation of new blood vessels from existingvascular beds (Farrera, 2004). Dysregulation of VEGF is associatedwith several diseases. For example, VEGF secretion by tumor cellsin response to oxygen deprivation or hypoxia triggers the sprout-ing of new capillaries, and thereby regulates the progression fromavascular to vascular stage of tumor as well as metastasis todistant organs (Carmeliet, 2005; Rapisarda and Melilo, 2012).VEGF is commonly expressed in many solid tumors includinglung, thyroid, breast, kidney, and bladder (Gasparini, 2000; Klubo-Gwiezdzinska et al., 2007; Nakanishi et al., 2009; Rozman et al.,2012). As a result, serum level of VEGF is often used as a cancerbiomarker in clinical practice. In addition, elevated levels of VEGFare also associated with other diseases including rheumatoidarthritis, diabetes mellitus, and psoriasis (Crawshaw et al., 2012;Lee et al., 2001; Prabhulkar et al., 2009). Therefore, serum level of

ll rights reserved.

313 593 3851.

VEGF has important implications as biomarker for different clinicaldisorders as well as biomarker for subsequent therapy monitoring.

Many VEGF detection systems have been reported and are usedin clinical settings. Traditional techniques for VEGF quantificationinclude enzyme linked immunosorbent assays (ELISA) and Lumi-nex (Cho et al., 2012; Prabhulkar et al., 2009). Sandwich ELISAinvolves attachment of capture-antibody for the target protein tothe surface of a microwell plate. Introduction of clinical samplefacilitates binding of VEGF with capture-antibody. A fluorophorelabeled second reporter antibody is then utilized to quantify VEGFconcentration in the sample. However, it is a time consuming andlabor intensive method (Prabhulkar et al., 2009). Luminex, invol-ving doped polystyrene microspheres, provides an alternativeplatform for high throughput analysis of samples. However,spectral overlap limits their sensitivity. Other analytical platformsreported for quantification of VEGF include label free and electricaldetection of VEGF using anti-VEGF aptamer modified Si nanowirefield effect transistors (Lee et al., 2009), biosensor based onferrocene monocarboxylic acid labeled anti-VEGF immobilized onthe microelectrode surface (Prabhulkar et al., 2009), enhancedresonance light scattering by gold nanoparticles modified VEGFimmunocomplex (Chen et al., 2013), VEGF detection by rollingcircle amplification and quantum dots tagging (Cheng et al., 2010),fluorescent peptide conjugated nanopillar chip (Suzuki andYokoyama, 2011) and electrochemical and optical sensors fordetection of protein–aptamer complexes (Abe et al., 2012; Choet al., 2012; Freeman et al., 2012). However, these methods involveexpensive instrumentation and complicated sample preparation.

M.A. Al-Ameen, G. Ghosh / Biosensors and Bioelectronics 49 (2013) 105–110106

Additionally, direct spotting of capture antibody or aptamer ontothe surface of microwell plates, polystyrene microspheres, orelectrode surfaces leads to their denaturation and deactivationwhich in turn results in low sensitivity and high background signalor noise. Hence, there is a critical need for a simple detectionstrategy which will enable rapid and sensitive analysis of VEGF inserum of patients with different clinical disorders.

To address the challenge associated with direct spotting ofantibodies onto solid substrates, in the present study, capture-antibody was immobilized within poly(ethylene) glycol diacrylate(PEGDA) hydrogel microspheres. Hydrogels are three dimensionalpolymeric networks that are able to swell and retain aqueoussolutions. The hydrated environment within the hydrogels permitsthe immobilized antibody to retain its activity and structure.In addition, the resistance of PEGDA to protein adsorption(Nemir et al., 2010) reduces the non-specific protein binding andthus improves the signal to noise ratio. An important designconsideration for sensitive analysis of VEGF in sample involvesthe transport of VEGF and reporter antibody within the hydrogelnetwork. Mesh size, i.e. the distance between the crosslinks, ofhydrogel relative to the solute dictates the transport withinhydrogel matrices. Hydrogels fabricated via crosslinking of poly-meric chains creates tight knit mesh network. Even though soluteswith hydrodynamic radii smaller than the mesh size can diffusefreely, the tight knit mesh network limits efficient diffusion ofmolecules with hydrodynamic radii comparable to hydrogel meshsize (Slaughter et al., 2009). Tuning hydrogel properties so as toincrease mesh size or porosity of hydrogel matrix can profoundlyimpact the diffusion characteristics of solutes. In the recent years,porosity has been introduced in hydrogels by several methodsincluding use of porogen, freeze-drying, foaming agent, and phaseseparation (Courtois et al., 2006; Kabiri et al., 2003; Lin-Gibsonet al., 2007; Okay, 2000; Omidian et al., 2007). Void space can becreated within the hydrogel matrix by reacting carbonate saltswith acids to create foam (Wachiralarpphaithoon et al., 2007).The porogen method uses salts, organic solvents, and inertmaterials including sodium chloride, calcium carbonate, glucose,and polyethylene glycol (PEG) to create pores within the matrix(Courtois et al., 2006; Mohan et al., 2006; Tran et al., 2011;Wachiralarpphaithoon et al., 2007). These porogens exclude grow-ing polymer chains from regions within the hydrogel matrix andupon removal from hydrogel following polymerization leavebehind a network of interconnected pores.

In this study, we demonstrate the development of VEGFdetection method by leveraging the advantages of immobilizationof capture-antibody within macro-porous hydrogel microsphereswith the immunoassay approach. We have investigated the impactof addition of porogen on structural properties of the hydrogel anddiffusion kinetics of large dextran molecules within hydrogelmatrix to identify the appropriate concentration and molecularweight of PEG porogen. We further demonstrated that inducedmacro-porosity augmented vascular endothelial growth factor(VEGF) detection sensitivity of the hydrogel microspheres due toenhanced diffusion of protein and target antibody within spheresand increased accessibility of capture antibody immobilized withinthese microspheres as opposed to those microspheres formedwithout porogen.

2. Materials and methods

2.1. Materials

Poly(ethylene) glycol diacrylate (PEGDA) (Mn∼575, 1000, and6000, hence forth referred as PEG575DA, PEG1000DA, andPEG6000DA), poly(ethylene glycol) (PEG) (Mn∼400, 10,000, 20,000,

and 35,000, hence forth referred as PEG 400, PEG 10K, PEG 20K,and PEG 35K), photo-initiator (2-hydroxy-4′-(2-hydroxyethoxy)-2-methylpropionphenone), fluorescein isothiocyanate-dextran (FITC-Dextran 250 kDa), glycerol, ethylene glycol, mineral oil, pluronicF-127, sodium chloride (NaCl), phosphate buffer saline (PBS), albuminfrom bovine serum (BSA), IgG from human serum were procuredfrom Sigma-Aldrich (St. Louis, MO). Vascular endothelial growthfactor (VEGF), anti-VEGF antibody, and Alexa Fluor 488 conjugatedanti-VEGF antibody were purchased from R&D systems (Minneapolis,MN). Fetal bovine serum (FBS) was procured from Gibco, trypan bluefrom MP Biomedicals and fluorescein PEG 20,000 Da from NanocsInc. (Boston, MA).

2.2. Fabrication and characterization of hydrogels

PEGDA hydrogel matrices were fabricated by adding 500 mlpre-polymer solution consisting of 6–10% (w/v) of PEGDA, 50%glycerol, 2% (w/v) photoinitiator with and without PEG to eachwell of 24 well plates and then exposing the plates to UV (CL-1000UV Crosslinker (UVP), 365 nm, 400 μW/cm2) for 3 min. The differ-ent molecular weight of PEGDA investigated were 575, 1000, and6000 and that of PEG porogen were 400, 10K, 20K, and 35K. Theconcentration of PEG porogen was also varied from 4%, 10%,and 20%.

2.2.1. Swelling ratio and mesh sizeFor measuring the mass swelling ratio (Qm), the hydrogel disks

were incubated in PBS for 72 h in a 37 1C shaker (50 rpm).Following incubation, excess fluid was removed from the swollenhydrogels and weights of the samples (Wwet) were measured. Thesamples were then dried at 60 1C for 24 h and the dry weightswere measured (Wdry). Qm of the hydrogel was calculated from theequation: Qm¼Wwet/Wdry. The average and standard deviation ofthree independent experiments each carried out in triplicate werecalculated. The mesh sizes (ξ) of the hydrogels were estimatedfrom Qm as described by Zustiak and Leach (2010).

2.2.2. DegradationFor measuring degradation of the hydrogel, the weights of the

hydrogel disks were measured following fabrication (day 0). Thedisks were then incubated in PBS in a 37 1C shaker (50 rpm) andweights of the hydrated hydrogels were measured on days 2, 6, 9,and 20 after fabrication. Degradation of the hydrogels was thenestimated from the ratio of weights at different time points to theweights on the day of fabrication.

2.3. Fabrication of microspheres

Hydrogel microspheres were fabricated by the oil emulsionmethod. 100 μL of polymer solution was added to 1 ml of oil withand without the surfactant (0.05% pluronic F-127). The solutionwas vortexed for 45 s, and then exposed to the UV cross-linker for3 min. Spheres were collected by centrifugation and stained withtrypan blue for visualization. Images were taken using Zeiss AxioObserver A1 microscope with integrated CCD camera. The sizedistribution was calculated by using Axio-Vision software.

2.4. Characterization of porosity

A functional analysis was used to determine the impact ofinduced porosity on macromolecular diffusion. Hydrogel micro-spheres fabricated with and without PEG porogen were incubatedwith FITC-Dextran 250 kDa and diffusion within spheres wasmonitored for 60 min by taking time series images at every5 min intervals. The fluorescence intensities from the microspheres

M.A. Al-Ameen, G. Ghosh / Biosensors and Bioelectronics 49 (2013) 105–110 107

and the background were measured by NIH ImageJ software. Thefluorescent signal from each sphere was calculated as the ratio ofintensity from the sphere to that of background.

2.5. Detection of VEGF

Microspheres containing capture-antibody (anti-VEGF anti-body) were then fabricated with and without PEG and incubatedin PBS for 48 h at 4 1C. Following which, the spheres wereincubated with VEGF with concentration varying from 1 to1000 pg/ml for 60 min. Then they were washed with TBST threetimes and incubated for another 60 min in Alexa Flour 488-anti-VEGF antibody. Following this, the spheres were washed withTBST and images of the microspheres were taken. Fluorescentintensities were analyzed using ImageJ software. The fluorescentsignal from each sphere was calculated as the ratio of intensityfrom the sphere to that of background. Mean of fluorescent signalfrom at least ten microspheres was calculated for each condition.To optimize incubation time, the spheres were incubated in100 pg/ml VEGF for time ranging from 0 to 120 min.

2.6. Statistical analysis

For all experiments, data reported as mean7S.E.M of threeindependent experiments. Statistical analyses were carried outwith one way ANOVA. Differences between two sets of data wereconsidered significant at p-valueo0.05.

3. Results and discussion

3.1. Effect of concentration and molecular weight of PEGDA on meshsize

Since the mesh size of the hydrogel is regulated by the cross-linking density of the precursor molecule, the structural propertiesof hydrogel can be regulated by altering the molecular weight andconcentration of polymer precursors. In this study, three differentmolecular weights of PEGDA at three different concentrationswere examined to understand the impact of PEGDA precursorson mesh size of hydrogel matrices. As demonstrated in Fig. 1,increasing molecular weight of PEGDA (which increases the lengthof PEG chain between crosslinking points) resulted in an increasein mesh size of hydrogels from 2.7 nm for PEG575DA to 3 nm and9 nm for PEG1000DA and PEG6000DA respectively at 6% polymerconcentration. On the other hand, increasing the precursor con-centration from 6% to 10%, which increases the potential cross-linking groups per volume, the mesh size of hydrogels decreased

Fig. 1. Effect of molecular weight and concentration of the PEGDA on the meshsize. Increase in molecular weight (chain length) and decrease in polymerconcentration increase the mesh size of the hydrogel. Error bars are S.E.M (N¼3).

from 2.7 to 1.9 nm, 3 to 2.4 nm, and 9 to 7 nm for PEG575DA,PEG1000DA, and PEG6000DA respectively (Fig. 1). Taking intoaccount the physical integrity of the hydrogels and consideringthat the hydrodynamic radii of many biomolecules includingcytokines, growth factors and globular proteins in blood are inthe range of 5–15 nm (Armstrong et al., 2004), 8% PEG6000DA waschosen for further studies.

3.2. Effect of PEG porogens on structural properties of hydrogel

From the wide range of PEG available, four different molecularweights (PEG 400, PEG 10K, PEG 20K, and PEG 35K) were selected toevaluate their impact on hydrogel properties including swelling ratioand stability. Swelling ratio of the hydrogel increased with increasein the molecular weight (chain length) of the porogens. As theporogens leave behind pores, water diffuses inside hydrogels. Longerthe chain length or molecular weight of the porogen, larger meshsizes are created within these gels. As demonstrated in Fig. 2A, theswelling ratios of hydrogels increased by 9, 21, 28, and 7 percentrespectively upon incorporation of 4% porogen, PEG 400, PEG 10K,PEG 20K, and PEG 35K, within pre-polymeric solution as comparedto hydrogel without porogen. Similar trend was observed at higherconcentration of porogen (data not shown). Minimal effect of PEG35K porogen on swelling ratio is probably due to difficulties inremoval of the porogen from hydrogel. Since, incorporation of PEG400 and PEG 35K had least effect on swelling ratio of hydrogelmatrices, further studies were carried out with different concentra-tion of PEG 10K and PEG 20K only.

To determine the stability of porous hydrogel matrices, hydro-gels were incubated with PBS at 37 1C and degradation over timewas measured. As hydrogels degrade over time, bonds are brokenwhich in turn facilitates diffusion of more aqueous solution inside

Fig. 2. Effect of incorporation of PEG porogen on swelling ratio and degradation ofhydrogels. (A) Swelling ratio of 8% PEG6000DA hydrogels increased upon incor-poration of 4% PEG porogen of different molecular weight. (B) Stability of theporous hydrogels compared to that of hydrogels fabricated in absence of porogen.Error bars are S.E.M (N¼3).

Fig. 3. (A) Comparison of size distribution of hydrogel microspheres formed in theoil phase with and without surfactant. Error bars are S.E.M (N¼3). Hydrogelmicrospheres are shown in inset. (B) Comparison of transport of FITC-250 kDadextran within the hydrogel microspheres fabricated with and without PEGporogen. Error bars are S.E.M (N¼3). (C) Fluorescence images of hydrogel micro-spheres fabricated with 10% and 20% FITC-PEG 20K immediately after polymeriza-tion (Day 0) and after incubation in PBS for 48 h (Day 2).

M.A. Al-Ameen, G. Ghosh / Biosensors and Bioelectronics 49 (2013) 105–110108

the gel. Thus, increased swollen weight of the hydrogel indicatesdegradation of the gel. Fig. 2B demonstrates the stability of thehydrogels with increased porosity over a time ranging from 2 to 20days. Stability of porous hydrogels created with 4% PEG 10K wasnot significantly different from that of control i.e. hydrogels with-out porogen. Porous hydrogels, fabricated by incorporation of PEGporogens of different molecular weight and concentrations,demonstrated significantly higher rate of degradation as comparedto the control hydrogels. However, all the porous hydrogelmatrices retained their physical integrity during this time period.Taken together these studies indicate that porosity of hydrogelmatrices can be tuned by incorporating PEG porogens within pre-polymer solution. Higher porosity leads to increased interactionbetween water and hydrophilic groups of the hydrogel.

3.3. Effect of surfactant on fabrication of microspheres

PEGDA microspheres were generated by vortexing an emulsionof hydrogel precursor solution (PEGDA, photo-initiator, and PEGporogen) in mineral oil followed by photo-polymerization. Theprocess yielded a suspension of polymer droplets which werecollected by centrifugation, washed and incubated in PBS. Tofacilitate visualization and characterization of these microspheres,trypan blue, a stain commonly used to assess cell viability and stainPEG based hydrogels (Franco et al., 2011), was used. These stainedspheres were then imaged and analyzed to determine the sizedistribution of these spheres. As demonstrated in Fig. 3A, vortexingin the absence of surfactant yielded a heterogeneous mixture ofspheres with sizes ranging from 100 to 400 mm radii with a mean ofaround 260 mm. To better control the size distribution of PEGDAmicrospheres, PEG based stabilizing surfactant pluronic F-127 wasused. Addition of 0.05% of the surfactant shifted the sphere sizedistribution to the left such that the average sphere size wassignificantly reduced to 170 mm. In addition, heterogeneity ofspheres was also regulated. This is likely caused by the lowering ofsurface tension by the surfactant which prevented the microspheresfrom coalescing. Henceforth, surfactant was used in all experimentsto fabricate hydrogel microspheres.

3.4. Effect of PEG porogens on diffusion of macromolecules

To investigate the effect of induced porosity on solute accessi-bility within hydrogel matrices, hydrogel microspheres wereincubated with solutions of FITC-Dextran 250 kDa and fluores-cence intensity was monitored over time. Relative intensities werecalculated from ratio of average fluorescent signal within eachsphere to the average signal in the surrounding area. Increasedfluorescent signal highlights enhanced diffusion of macromole-cules within the spheres. As demonstrated in Fig. 3B, solutepenetration increased with increased incorporation of PEG 10Kwithin hydrogel microspheres. However, when the concentrationof PEG 20K was increased from 10% to 20%, diffusion of dextranmacromolecules reduced. To investigate this phenomenon, wefabricated hydrogel microspheres containing 10% and 20% FITClabeled PEG 20K to monitor diffusion of PEG from the micro-spheres. Following fabrication, these microspheres were incubatedin PBS and images were captured to monitor diffusion of PEG atdifferent time points. As shown in Fig. 3C, while majority ofporogen diffused out of hydrogel microspheres fabricated with10% PEG 20K following 2 day incubation in PBS leaving behindinterconnected pores, some porogens remained within the hydro-gel microspheres fabricated with 20% PEG 20K after 2 days.Reduced diffusion of FITC-Dextran 250 kDa within hydrogel micro-spheres is likely due to increased retention of PEG moleculeswithin the spheres. Since, significantly improved transport ofmacromolecules were observed in microspheres fabricated with

20% PEG 10K and 10% PEG 20K, these porogen compositions wasused for further studies.

3.5. Detection of VEGF using porosity tuned hydrogel microspheres

After demonstration of increased accessibility due to inducedporosity, hydrogel microspheres were fabricated with covalentlyimmobilized VEGF antibody. These spheres were then incubatedwith varying concentration of VEGF for 1 h, following whichmicrospheres were washed and incubated with Alexa Fluor 488conjugated anti-VEGF antibody for 1 h. After washing the micro-spheres to remove the unbound reporter antibody, the fluorescentintensities from the microspheres were accessed using ZeissAxio Observer microscope. Average fluorescent signal from the

Fig. 4. Detection of VEGF using hydrogel microspheres fabricated in the presenceand absence of PEG porogen. Induced macroporosity enhanced the fluorescentsignal (ratio of fluorescent intensity of the microspheres to that of the background)due to enhanced diffusion of VEGF and target antibody within the spheres. Errorbars are S.E.M (N¼3).

Fig. 5. (A) Optimization of incubation time of VEGF with capture antibody immobi-lized hydrogel microspheres. (B) Determination of the specificity of the assay byincubating the microspheres with FBS, PBS (blank), 100 mg/ml BSA, 1 M NaCl, 100 pg/ml IgG, and 100 pg/ml VEGF for 1 h. (C) Determination of the detection limit of theassay. The hydrogel microspheres were incubated with 0–100 pg/ml VEGF for 1 h. Thedetection limit of this assay was found to be 2.5 pg/ml. Error bars are S.E.M (N¼3).

M.A. Al-Ameen, G. Ghosh / Biosensors and Bioelectronics 49 (2013) 105–110 109

microspheres were obtained by calculating the mean of the ratioof the average signal from the microspheres to the averagefluorescent signal in the area surrounding the spheres. As illu-strated in Fig. 4, sensitivity of VEGF detection was enhanced uponintroduction of porosity within spheres as compared to the micro-spheres fabricated without the porogen (8% PEGDA). This is due toincreased diffusion of VEGF and target antibody within thehydrogel microspheres. Even though there was no statisticaldifference (p-value40.05) in fluorescent intensities obtained frommicrospheres with induced porosity at low VEGF concentration(1 pg/ml), signal obtained from spheres generated with 10% PEG20K at concentrations of VEGF (10–1000 pg/ml), was significantlyhigher (p-valueo0.05) than that obtained from spheres with 20%PEG 10K. Incubation with higher concentration of VEGF drovemore VEGF molecules into the microspheres leading to increasedformation of complex with immobilized VEGF antibody. Thisphenomenon probably resulted in inhibited diffusion of fluoro-phore conjugated VEGF antibody in case of microspheres withsmaller mesh size (7.670.09 nm), formed upon addition of 20%PEG 10K. In comparison, in microspheres with larger mesh size(8.670.5 nm) formed upon addition of 10% PEG 20K, increasednumber of VEGF antibody molecules could diffuse within themicrospheres leading to enhanced signal. Henceforth, all theexperiments were carried out with VEGF antibody immobilizedporous microspheres generated with 10% PEG 20K porogen.

To optimize the reaction time, microspheres were incubatedwith 100 pg/ml VEGF for time ranging from 0 to 120 min. Fig. 5Aillustrates the time dependent changes in fluorescent signal high-lighting the interaction between VEGF and antibody. As expected,protein-antibody interaction increased with increase in reactiontime from 15 to 60 min. However, the fluorescent signal obtainedfollowing incubation of the spheres with VEGF for 120 min wasnot statistically different (p-value40.05) from that obtained upon60 min incubation indicating that the fluorescent signal saturatedby 1 h. Since, significantly higher fluorescent intensity was observedfollowing 1 h incubation, all the further experiments were carriedout with an incubation time of 1 h.

3.6. Determining specificity of the assay

Serum contains multiple proteins, electrolytes, antibodies, andantigens which can potentially give rise to fluorescent signal eitherdue to nonspecific interaction with immobilized antibody or dueto nonspecific adsorption within microspheres. This is one of themajor challenges in developing the protein assay. Our hydrogelbased protein detection method can leverage panels of antibody

already validated by commercial sources for their high specificitytowards the target protein. To further address whether othercomponents of serum can affect fluorescent signal, VEGF antibodyimmobilized microspheres were incubated with 100 mg/ml BSA inPBS, 1 M NaCl in PBS, 100 pg/ml IgG in PBS, FBS, and 100 pg/mlVEGF for 1 h. A high signal was obtained (11.170.6) for 100 pg/mlVEGF and low signal for FBS (3.270.3) (Fig. 5B). Signal obtainedupon incubation with BSA, NaCl, and IgG ((1.270.22),(1.0970.033), and (1.1670.09) respectively) was not significantlydifferent from that obtained from blank i.e. PBS with no VEGF(1.0670.07). The study indicated that proteins other than VEGFcontribute very little towards the fluorescent signal. Moreover, italso established that immobilization of antibody within hydrogelmicrospheres aids in eliminating the protein blocking and thusreduces the complexity of the assay.

3.7. Determining the detection limit

After confirming the fluorescent signal observed was primarilydue to interaction between VEGF and antibody, the detection limit of

Table 1Comparison of hydrogel microsphere based VEGF detection assay with ELISA and Luminex assays.

Assays Detection limit (pg/ml) Detection time (h) CV% intra-run variation CV% inter-run variation

ELISAa 6.4 5.5 7.9 8.8Luminexb o10 pg/ml 3.5 4.6 7.6Hydrogel microspheres 2.5 2 7.7 8.2

a QuantiGlo Human VEGF Chemiluminescent Immunoassay.b VEGF Human Singleplex Bead Kit.

M.A. Al-Ameen, G. Ghosh / Biosensors and Bioelectronics 49 (2013) 105–110110

the method was estimated by exposing the microspheres to0–100 pg/ml VEGF solutions for 1 h. As demonstrated in Fig. 5C,fluorescent signal increased with increase in VEGF concentration. Theoverall signal increased from 1.9 to 11.5 upon varying VEGF concen-tration from 1 pg/ml to 100 pg/ml. The minimum concentration ofVEGF detectable by this assay was considered to be the concentrationat which the signal deviated from the average signal obtained inabsence of VEGF by three times the standard error. The detectionlimit of this assay was found to be 2.5 pg/ml. This highlights that thehydrogel based VEGF detection method can be utilized not only forestimating VEGF levels in serum of patients with clinical disordersbut also for detecting VEGF levels in healthy individuals or formonitoring treatment progression. Intra- and inter-run variationswere used to calculate the coefficient of variation (CV). As demon-strated in Table 1, while the assay performance was comparable, thesensitivity of the hydrogel microsphere based VEGF quantificationmethod exceeded that of the conventional assays including ELISAand Luminex. In addition, elimination of protein blocking step notonly significantly reduced the complexity of the assay but also thetime required for protein detection.

4. Conclusion

The present study demonstrated the development of a novel VEGFdetection method by integrating the standard immunoassay with theadvantages of immobilization of capture antibody within macro-porous hydrogel microspheres. The hydrated environment withinthe hydrogels permitted the immobilized antibody to retain its activityand structure and thereby improved the detection sensitivity. Utiliza-tion of bio-inert PEG hydrogel attenuated the background signalthereby further improving the sensitivity of the detection method aswell as eliminated the need for blocking the proteins. In this study, weestablished that tuning the porosity of hydrogel microspheres byincorporation of 10% PEG 20K within pre-polymer solution improvedquantification of VEGF. We further showed that while this simplehydrogel microsphere based VEGF detection system provided compar-able assay performance (CV), sensitivity of the assay exceeded that ofthe conventional assays including ELISA and Luminex.

Acknowledgment

The authors would like to thank University of Michigan—Dearborn and University of Michigan—Ann Arbor: Office of theVice President for Research for the financial support.

References

Abe, K., Hasegawa, H., Ikebukuro, K., 2012. Electrochemistry 80, 348–352.Armstrong, J.K., Wenby, R.B., Meiselman, H.J., Fisher, T.C., 2004. Biophysical Journal

87, 4259–4270.Byrne, A.M., Bouchier-Hayes, D.J., Harmey, J.H., 2005. Journal of Cellular and

Molecular Medicine 9, 777–794.Carmeliet, P., 2005. Oncology 69 (Suppl. 3), 4–10.Chen, Z., Lei, Y., Gao, W., Liu, J., 2013. Plasmonics 8, 605–611.Cheng, W., Yan, F., Ding, L., Ju, H., Yin, Y., 2010. Analytical Chemistry 82, 3337–3342.Cho, H., Yeh, E.-C., Sinha, R., Laurence, T.A., Bearinger, J.P., Lee, L.P., 2012. ACS Nano 6,

7607–7614.Courtois, J., Bystrom, E., Irgum, K., 2006. Polymer 47, 2603–2611.Crawshaw, A.A., Griffiths, C.E.M., Young, H.S., 2012. Expert Opinion on Investiga-

tional Drugs 21, 33–43.Farrera, N., 2004. Endocrine Reviews 25, 581–611.Franco, C.L., Price, J., West, J.L., 2011. Acta Biomaterialia 7, 3267–3276.Freeman, R., Girsh, J., Jou, A.F.-J., Ho, J.A., Hug, T., Dernedde, J., Willner, I., 2012.

Analytical Chemistry 84, 6192–6198.Gasparini, G., 2000. Oncologist 5 (Suppl. 1), 37–44.Hoeben, A., Landuyt, B., Highley, M.S., Wildiers, H., Van Oosterom, A.T., De Bruijn, E.A.,

2004. Pharmacological Reviews 56, 549–580.Kabiri, K., Omidian, H., Hashemi, S.A., Zohuriaan-Mehr, M.J., 2003. European

Polymer Journal 39, 1341–1348.Klubo-Gwiezdzinska, J., Junik, R., Kopczynska, E., Juraniec, O., Kardymowicz, H.,

2007. European Journal of Endocrinology 157, 521–527.Lee, H.S., Kim, K.S., Kim, C.-J., Hahn, S.K., Jo, M.-H., 2009. Biosensors and

Bioelectronics 24, 1801–1805.Lee, S.S., Joo, Y.S., Kim, W.U., Min, D.J., Min, J.K., Park, S.H., Cho, C.S., Kim, H.Y., 2001.

Clinical and Experimental Rheumatology 19, 321–324.Lin-Gibson, S., Cooper, J.A., Landis, F.A., Cicerone, M.T., 2007. Biomacromolecules 8,

1511–1518.Mohan, Y.M., Murthy, P.S.K., Raju, K.M., 2006. Journal of Applied Polymer Science

101, 3202–3214.Nakanishi, R., Oka, N., Nakatsuji, H., Koizumi, T., Sakaki, M., Takahashi, M.,

Fukumori, T., Kanayama, H.O., 2009. Urology International 83, 98–106.Nemir, S., Hayenga, H.N., West, J.L., 2010. Biotechnology and Bioengineering 105,

636–644.Okay, O., 2000. Progress in Polymer Science 25, 711–779.Omidian, H., Park, K., Rocca, J.G., 2007. Journal of Pharmacy and Pharmacology 59,

317–327.Prabhulkar, S., Alwarappan, S., Liu, G., Li, C.-Z., 2009. Biosensors and Bioelectronics

24, 3524–3530.Rapisarda, A., Melilo, G., 2012. Advances in Cancer Research 114, 237–267.Rozman, A., Silar, M., Kosnik, M., 2012. Radiology and Oncology 46, 354–359.Slaughter, B.V., Khurshid, S.S., Fisher, O.Z., Khademhoseini, A., Peppas, N.A., 2009.

Advanced Materials 21, 3307–3329.Suzuki, Y., Yokoyama, K., 2011. Biosensors and Bioelectronics 26, 3696–3699.Tran, R.T., Naseri, E., Kolasnikov, A., Bai, X., Yang, J., 2011. Biotechnology and Applied

Biochemistry 58, 335–344.Wachiralarpphaithoon, C., Iwasaki, Y., Akiyoshi, K., 2007. Biomaterials 28, 984–993.Zustiak, S.P., Leach, J.B., 2010. Biomacromolecules 11, 1348–1357.